Retrospective Theses and Dissertations Iowa State University Capstones, Theses and Dissertations 1997 Mass transfer mechanisms in air sparging systems Washington Jose Braida Iowa State University Follow this and additional works at: https://lib.dr.iastate.edu/rtd Part of the Agriculture Commons, Civil Engineering Commons, Environmental Engineering Commons, and the Soil Science Commons Recommended Citation Braida, Washington Jose, "Mass transfer mechanisms in air sparging systems " (1997). Retrospective Theses and Dissertations. 12276. https://lib.dr.iastate.edu/rtd/12276 This Dissertation is brought to you for free and open access by the Iowa State University Capstones, Theses and Dissertations at Iowa State University Digital Repository. It has been accepted for inclusion in Retrospective Theses and Dissertations by an authorized administrator of Iowa State University Digital Repository. For more information, please contact digirep@iastate.edu. INFORMATION TO USERS This manuscript has been reproduced from the microfibn master. UMI fihns the text directly from the original or copy submitted. Thus, some thesis and dissertation copies are in typewriter &ce, i^e others may be from any type of computer printer. The quality of this reprodaction is dependent upon the quality of the copy submitted. Broken or indistinct print, colored or poor quality illustrations and photographs, print bleedthrough, substandard margins, and improper alignment can adversely affect reproduction. In the unlikely event that the author did not send UMI a complete manuscript and there are missing pages, these will be noted. Also, if unauthorized copyright material had to be removed, a note will indicate the deletion. Oversize materials (e.g., maps, drawings, charts) are reproduced by sectioning the original, beginning at the upper left­hand comer and continuing from left to right in equal sections with small overlaps. Each original is also photographed in one exposure and is included in reduced form at the back of the book. Photographs included in the original manuscript have been reproduced xerographically in this copy. BQgher quality 6" x 9" black and white photographic prints are available for any photographs or illustrations appearing m this copy for an additional charge. Contact UMI directly to order. UMI A Bell & Howell Infonnation Company 300 North Zeeb Road, Ann Aibor MI 48106­1346 USA 313/761­4700 800/521­0600 Mass transfer mechanisms in air sparging systems by Washington Jose Braida A dissertation submitted to the graduate faculty in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Major: Civil Engineering (Envirormiental Engineering) Major Professor: Say Kee Ong Iowa State University Ames, Iowa 1997 Copyright © Washington Jose Braida, 1997. All rights reserved. UMI Nxunber: 9814622 Copyright 1997 by Braida, Washington Jose All rights reserved. UMI Microform 9814622 Copyright 1998, by UMI Company. All rights reserved. This microform edition is protected against unauthorized copying under Title 17, United States Code. UMI 300 North Zeeb Road Ann Arbor, MI 48103 ii Graduate College Iowa State University This is to certify that the doctoral dissertation of Washington Jose Braida has met the dissertation requirements of Iowa State University Signature was redacted for privacy. Committee N^ember Signature was redacted for privacy. Committee Member Signature was redacted for privacy. Committee Member Signature was redacted for privacy. Committee Signature was redacted for privacy. ajorlProfessor Signature was redacted for privacy. Ee{ the Major Program Signature was redacted for privacy. For the CJradwie College lii To my lovely wife Gaye, my daughters Maria Eugenia and Constanza and my parents IV TABLE OF CONTENTS LIST OF FIGURES viii LIST OF TABLES xi ACKNOWLEDGMENTS xiii ABSTRACT xiv CHAPTER ONE. INTRODUCTION AND OBJECTIVES 1 Introduction 1 Objectives 2 Dissertation Organization 3 References 4 CHAPTER TWO. LITERATURE REVIEW 5 Air Sparging: Process Description 5 Nature of the Injected Air Flow in the Porous Media 7 Mass Transfer and Transport Mechanisms 11 Volatilization of VOCs: Air­water mass transfer coefficients 13 Sorption­desorption of VOCs onto porous media 20 Dissolution and volatilization of nonaqueous phase liquids (NAPLs) in the subsurface 24 Modeling Mass Transfer 26 VOCs transport in porous media 31 References 33 CHAPTER THREE. AIR SPARGING EFFECTIVENESS: THE AIR CHANNEL MASS TRANSFER ZONE Abstract 41 41 V Introduction 41 Materials and Methods 44 Results and Discussion 45 Mass transfer zone 45 Model correlation of mass transfer zone 47 Conclusions 50 Notation 51 Acknowledgments 52 References 52 CHAPTER FOUR. AIR SPARGING: AIR­WATER MASS TRANSFER COEFFICIENTS 65 Abstract 65 Introduction 65 Materials and Methods 67 Results and Discussion 69 VOC concentration profiles and mass transfer zone (MTZ) 69 Estimation of mass transfer coefficients 71 Model correlation of lumped mass transfer coefficients 74 Conclusions 78 Notation 79 References 79 CHAPTER FIVE. VOLATILIZATION OF VOCs UNDER AIR SPARGING CONDITIONS: MASS TRANSFER ANALYSIS 94 Inttoduction 94 Materials and Methods 94 Results and Discussion 95 Conclusions 103 References 104 VI CHAPTER SIX. FATE OF NONAQUEOUS PHASE LIQUIDS UNDER AIR SPARGING CONDITIONS 105 Abstract 105 Introduction 105 Materials and Methods 107 Results and Discussion 109 Stagnant air conditions 109 Influence of air flow rate 110 Influence of porous media 111 VOC removal efficiency 111 Conclusions 113 References 113 CHAPTER SEVEN. INFLUENCE OF SORPTION­DESORPTION PROCESSES ON AIR SPARGING EFFECTIVENESS 125 Abstract 125 Introduction 125 Materials and Methods 127 Results and Discussion 129 Conclusions 134 Notation 134 References 135 CHAPTER EIGHT. MODELING AIR SPARGED SOIL COLUMNS 146 Introduction 146 Materials and Methods 146 Description of Model 150 Results and Discussion 153 Conclusions 167 References 168 vii CHAPTER NINE. GENERAL CONCLUSIONS AND FUTURE WORK 165 Conclusions 165 Future work 167 APPENDIX A. NOTATION 172 APPENDIX B. RAW DATA AND MASS BALANCES 175 Single­air channel experiments. Gas phase and liquid phase VOC concentrations 176 Mass balance calculations 204 APPENDIX C. ONE­D DIFFUSION MODEL: COMPUTER CODES 209 Computer code for ethylbenzene, single­air channel setup 210 Computer code for ethylbenzene, single air­channel setup, adsorption 212 Computer code for styrene, soil column 214 BIOGRAPHICAL SKETCH 216 VIII LIST OF HGURES CHAPTER TWO Figure 1 Typical air sparging system 6 Figure 2 Mass transfer processes occuring during air sparging 12 Figures Two­resistance model 15 CHAPTER THREE Figure 1 Conceptual sketch of air channels and mass transfer zone 57 Figure 2 Single­air channel apparatus 58 Figure 3 VOC concentrations in the air phase for Ottawa sand at an air flow rate of 2.5 cm/s 59 VOC concentration profiles for various times during air sparging (Ottawa sand, air flow rate 2.5 cm/s) 60 VOC concentration profiles for various times during air sparging (Ottawa sand, air flow rate 2.5 cm/s) 61 o­Xylene concentration profiles for different porous media (air flow rate of 2.5 cm/s) 62 Benzene concentration profiles for different air flow rates (Ottawa sand) 62 Pore Diffusion Modulus: experimental vs. computed values with 95% confidence intervad plots 63 Comparison of experimental and computed Pore Diffusion Modulus for 1,2 Dichlorobenzene and 1,2,4 Trichlorobenzene with 95% confidence interval plots 64 Figure 4 Figure 5 Figure 6 Figure 7 Figure 8 Figure 9 CHAPTER FOUR Figure 1 Single­air channel apparams 87 Figure 2 (a) Benzene concentration (mg/L) in the exhaust air (b) Benzene concentration (mg/L) in aqueous phase for sand 70/100, and air velocity of l.l cm/s 88 Air­water interface mass transfer, two­resistance model and observed air channel conditions 89 Figure 3 IX Figure 4 Experimental and predicted benzene concentrations for sand 70/100 and air velocity of 1.1 cm/s: (a) exhaust air, (b) aqueous concentration 90 Figure 5 Experimental vs. predicted modified Sherwood number and 95% confidence interval for the population 91 Experimental vs. predicted Damkohler number and 95% confidence intervd for the population 92 Estimated lumped gas phase mass transfer coefficient using eq. 11 vs. estimated lumped gas mass transfer coefficient using eq. 12 93 Figure 6 Figure 7 CHAPTER FIVE Figure 1 Figure 2 Figure 3 Figure 4 Interfacial mass transfer resistance vs. Henry's Law constant for Ottawa sand and 95% confidence limits 96 Interfacial mass transfer resistance vs. Henry's Law constant for sand 30/50 and 95% confidence limits 97 Interfacial mass transfer resistance vs. Henry's Law constant for sand 70/100 and 95% confidence limits 98 Liquid side mass transfer coefficients as a function of air velocity for different porous media 102 CHAPTER SIX Figure 1 Single­air channel apparatus 116 Figure 2 Isoconcentration lines (mg/L) and actual VOC concentrations, (a) chlorobenzene NAPL in sand 30/50 after 24 hours with no air flow, 05) benzene NAPL in sand 30/50 after 24 hours with an air flow rate of 68 mL/min 117 Isoconcentration lines (mg/L) for benzene NAPL in sand 30/50 and zero air flow 118 Isoconcentration lines (mg/L) for chlorobenzene NAPL in sand 30/50 and zero air flow 119 Isoconcentration lines (mg/L) for benzene NAPL in sand 30/50 and an air flow rate of 27.5 mL/min 120 Isoconcentration lines (mg/L) for benzene NAPL in sand 30/50 and an air flow rate of 68 mL/min 121 Isoconcentration lines (mg/L) for cholorobenzene NAPL in sand 30/50 and an air flow rate of 27.5 mL/min 122 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 X Figure 8 Figure 9 Isoconcentration lines (mg/L) for benzene NAPL in sand 70/100 and an air flow rate of 27.5 mlVmin 123 VOC concentration in the exiiaust gas and VOC removal efficiency 124 CHAPTER SEVEN Figure 1 Single­air channel apparatus 139 Figure 2 Relative VOC concentrations in the exhaust air for different organic carbon contents 140 Figure 3 Benzene concentration profiles for different organic carbon contents 141 Figure 4 Ethylbenzene concentration profiles for different organic carbon contents 142 Figure 5 n­Propylbenzene concentration profiles for different organic carbon contents 143 Figure 6 Experimental and predicted VOC concentration profiles in the water phase: (a) organic carbon content 0.04%, (b) organic carbon content 0.45% Figure 7 144 Benzene concentration in the exhaust air: (a) organic carbon content 0.04%, (b) organic carbon content 0.45% 145 CHAPTER EIGHT Figure 1 Sketch of soil colunm 147 Figure 2 Array of cylindrical air channels and with their corresponding mass transfer zones 151 Figure 3 Cross section of air chaxmel and mass transfer zone 154 Figure 4 (a) Benzene concentrations in the exhaust air, (b) benzene mass removal for sand 30/50 at an air velocity of 0.86 cm/s 156 (a) Ethylbenzene concentration in the exhaust air, (b) ethylbenzene mass removal for sand 70/100 at an air velocity of 1.21 cm/s 157 (a) Styrene concentration in the exhaust air, (b) styrene mass removal for Ottawa sand at an air velocity of 1.34 cm/s 158 Comparison of experimental and predicted mass distribution after sparging for sand 30/50 and air velocity of 0.86 cm/s 163 Comparison of experimental and predicted mass distribution after sparging for sand 70/100 and air velocity of 1.21 cm/s 164 Comparison of experimental and predicted mass distribution after sparging for Ottawa sand and air velocity of 1.34 cm/s 165 Influence of air channel radius on benzene mass removal 166 Figure 5 Figure 6 Figure 7 Figure 8 Figure 9 Figure 10 xi LIST OF TABLES CHAPTER THREE Table 1 Physical­chemical properties of porous media 55 Table 2 Physical­chemical properties of VOCs at 20°C 55 Table 3 Dimensionless numbers used for modeling 56 Table 4 Summary of stepwise regression analysis 56 CHAPTER FOUR Table 1 Physical­chemical properties of porous media 83 Table 2 Physical­chemical properties of VOCs at 20°C 83 Table 3 Estimated air phase mass transfer coefficients for various compounds, porous media, and air velocities (cm/min) 84 Table 4 Dimensionless numbers used for modeling 85 Table 5 Summary of stepwise regression analysis 86 CHAPTER FIVE Table 1 Table 2 Table 3 Liquid and air side mass transfer resistance. Regression analysis results 99 Relative liquid side and gas side mass transfer resistance for selected VOCs 100 Estimated values for flow rates 103 (cm) for Ottawa sand at different CHAPTER SIX Table I Experimental matrix 115 CHAPTER SEVEN Table 1 Table 2 Linear partition coefficients (K<j) for different VOCs into Ottawa sand as a function of organic carbon content 137 Sorbed VOC concentration (mg/Kg) + standard deviation 137 xii Table 3 Table 4 Retardation factors (R) for different VOCs on Ottawa sand as a function of organic carbon content 138 Parameters used in the model 139 CHAPTER EIGHT Table I Parameters used in the model 155 Table 2 Surrunary of modeling results 159 xiii ACKNOWLEDGMENTS There were several people who were instrumental in the completion of my dissertation. First and foremost, I would like to express my deep gratitude to Professor Say Kee Ong, for convincing me to attend Iowa State University for my doctorate and for being my dissenation advisor and mentor. He has set high standards that have always helped me learn more and his encouragement was always there along the hills and valleys of this process. I wish to express my appreciation to Professor LaDon Jones for the helpful discussions on the analysis and modeling of my results. I also want to thank Juan Jose Goyeneche, my countrymate and friend, for his invaluable help in the statistical analysis of the data. I am also grateful to Professor Charles Oulman, Professor L. K. Doraiswamy, and Robert Horton for serving in my doctoral committee. I gratefully acknowledge my student colleagues, Pak­Hing Lee, Dr. Chul­Sung Kim, Roger Protzman, Steve Van Dyke, Todd Fryzek, and Michael Carr for the great times and encouragement through endless hours of laboratory work. I want to extend special thanks to Dr. Keh­Ping Chao for his valuable input in ±e early stages of my work. I would like to express my appreciation to my parents, daughters and relatives for all their love and support. Finally, I would like to express my greatest indebtedness to my wife, Gaye Goker, for her endless support, love, and encouragement. xiv ABSTRACT The air­water mass transfer of VOCs during air sparging was investigated using a single­ air channel air sparging semp and a 14 cm (S? in) diameter soil colunm. Three different porous media and 10 VOCs were used in the study. Air velocities ranged from 0.2 cm/s to 2.5 cm/s. Experimental results for the single­air channel setup indicated that volatilization of VOCs during air sparging was a difftision limited process. VOCs, in a thin layer of saturated porous media next to the air chaimels (identified as the mass transfer zone, MTZ), were found to deplete rapidly during air sparging resulting in a steep concentration gradient within this zone while the VOC concentrations outside the zone remained fairly constant. The rapid depletion was associated with faster initial volatilization of VOCs at the air­water interface as compared to the diffusive transport of VOCs to the air­water interface. The size of MTZ ranged from 17 to 41 mm or between 70 dp$o and 215 dp50 {dpso = mean particle size of the porous media) depending on the VOC. A general correlation predicting the size of the MTZ was developed. The size of MTZ was found to be directly proportional to the aqueous diffusivity of the VOC, the mean particle size, and the uniformity coefficient. The size of MTZ was also found to decrease with increasing organic carbon content of the porous media. This effect was larger for VOCs with low solubilities and high partition coefficients. Air­water mass transfer coefficients (ATc) for the volatilization of VOCs were estimated by fitting experimental data to a one­dimensional diffusion model. The air­water mass transfer coefficients ranged from 1.79 x 10'^ cm/min to 3.85 x 10'" cm/min for the VOCs tested. Two empirical models were developed for the prediction of mass transfer coefficients by correlating the Damkohler and modified air phase Sherwood numbers with air phase Peclet number, Henry's law constant, and reduced mean particle size of porous media. The estimated lumped mass transfer coefficients {Kg a) were found to be directly related to the air diffusivity of the VOC, air velocity, particle size, and inversely related to the Henry's law constant of the VOCs. Based on the two­resistance model, the liquid­side resistance accounted for more than 90% of the total resistance for the air­water interfacial mass transfer. Experiments with nonaqueous phase liquid (NAPLs) indicated that air sparging may control the spreading of NAPLs and were more effective for NAPLs with higher solubilities XV and lower densities. Removal efficiencies of NAPLs and dissolved VOCs were found to be greatly affected by the grain size of the porous media. The MTZ concept and the correlations developed for the single­air channel study were incorporated into a one­dimensional radial diffusion model and were found to successfully predict the air phase concentrations, final aqueous VOC concentrations, and total mass removed for a 5? in diameter air sparged soil column. CHAPTER ONE. INTRODUCTION AND OBJECTIVES Introduction The U.S. Environmental Protection Agency (EPA) has estimated that about 25% of two million underground storage tank (UST) systems located at 700,000 facilities across the U.S. may be leaking (U.S. EPA, 1988). The improper and illegal disposal of hazardous wastes together with the release of organic chemicals from UST systems have a significant enviromnental impact on groundwater resources and may be a risk to human health. Volatile organic compounds (VOCs) are generally slightly soluble in water and may volatilize into the air pores of the unsaturated zone. After their release, the VOCs may migrate downward under the influence of gravity, capillary forces, and viscous forces. VOCs commonly found in contaminated groundwater include benzene, creosote mixture, ethylbenzene, xylenes, toluene, and chlorinated hydrocarbons. Various physical, chemical, and biological treatment methods have been developed to remediate contaminated sites. The most common remedial approach employed in the control and remediation of contaminated aquifers is the pump­and­treat system. The pump­and­treat system uses a series of wells placed downstream of the flow path of the plume to capture the contaminated water. The pumped water is then treated above ground. The effectiveness of the pump­and­treat approach as a remedial technology is uncertain when dealing with contaminants sorbed to the samrated soils or when the site has extensive heterogeneity. A remediation approach frequently used is in situ biodegradation in which indigenous microorganisms present in the aquifer are used for the degradation of the organic compounds. The success of bioremediation depends on the enhancement of the degradation capability of the indigenous population through the addition of nutrients and oxygen. As such, bioremediation is sensitive to many physical and chemical variables at a given site. Another remedial approach for contaminated aquifers is the use of physical means such as aeration technologies. One emerging soil aeration technology, air sparging, involves the injection of contaminant­free air, directly below the water table for the stripping/volatilization of volatile ­) organic compounds (VOCs) and the promotion of biodegradation (Brown et al., 1994). Contaminated air is then removed with extraction wells located in the unsaturated zone. Even though air sparging has been successfully applied at several contaminated sites, Johnson et al. (1993) pointed out that the mechanism of air flow in saturated porous media and the physical­chemical/biological processes involved during air sparging operations are not yet well understood. Similarly, Reddy et al. (1995) and Chao and Ong (1995) pointed out that the current design and implementation of air sparging systems are based on empirical approaches from field experience. The physical­chemical processes involved during air sparging is a complex combination of mass transport and transformation processes which include volatilization, dissolution, diffusion, advection, hydrodynamic isolation, and desorption of contaminants. The transformation processes include biodegradation of VOCs from an increase in dissolved oxygen in the contaminated aquifer. The extent of mass transfer during air sparging is dependent on the advective­dispersive characteristics of the air flow, dispersive, and possibly advective characteristics of the groundwater, and the diffusive properties of the VOCs. Field studies on air sparging are usually not exhaustive and results are not well documented. As a result, it is difficult to thoroughly elucidate the fundamental processes involved in air sparging and the long term performance of this remediation technology. To understand air sparging processes with the eventual goal of designing air sparging systems in an economical and efficient way, it is essential to have a better comprehension of the fundamental contaminant transport and transformation mechanisms involved in the application of this technology. Objectives The study will focus on the physical­chemical interactions occurring during air sparging. Biological processes associated with the use of this technology will not be included. The objectives of this research are divided into three parts as follows: (i) To investigate and quantify the major and limiting mass transfer processes affecting volatilization of VOCs during air sparging by using a laboratory­scale single­air channel 3 experimental setup. Different types of VOCs, air flow rates, porous media and the presence of NAPLs were used as variables to provide a better understanding of the mass transfer of VOCs. Mass transfer coefficients for different VOCs were determined for different experimental conditions. (ii) To correlate the mass transfer coefficients found in objective (i) with macroscopic properties of the system using dimensional analysis. The correlation derived may be used for the prediction of mass transfer for different experimental situations. (iii) Based on results of objectives (i) and (ii), a model to predict mass Uransfer was developed and applied to a complex system such as an air sparged soil column. The approach taken by this dissertation was to investigate the "microscale" effects of air sparging by using an experimental setup with a single­air channel. The results from the single­air charmel experiments were then used to predict the results of a more complex system such as air sparged soil columns. Dissertation Organization The dissertation is divided into nine chapters. Chapters three, four, six, and seven are in the form of journal papers. Chapter 2 presents a literature review of the different issues related with the operation and application of air sparging as a remediation technology. In Chapter 3, results firom well controlled laboratory experiments provided a quantitative insight on the factors influencing the volatiUzauon of VOCs under air sparging conditions. The governing mechanisms for the volatilization of VOCs at the "microscale" level were discussed and based on these observations, a correlation between the mass transfer zone and macroscopic properties of the system was developed. In Chapter 4, an advection­diffusion model was developed and used for the estimation of the air­water mass transfer coefficients for the VOCs under sparging conditions. In this chapter, a dimensionless model for the prediction of mass transfer coefficients was developed. Chapter 5 presents an analysis of the dominant mass transfer resistance affecting the volatilization of VOCs under sparging conditions. In Chapter 6, the effects of air sparging on NAPLs were investigated. The influence of adsorption processes on the performance of air sparging systems is presented in 4 Chapter 7. Chapter 8 presents the application of a one­dimensional diffusion model to predict the results of air sparged soil columns. Parameters estimated in the earlier chapters were used as inputs for the model. Chapter 9 presents the general conclusions, recommended fixture work, and closing statement of the study. References Brown, R.A., R.J. Hicks, and P.M. Hicks, Use of air sparging for in situ bioremediation, in Air Sparging for Site Remediation, edited by R.E. Hinchee, pp. 38­55, Lewis Publishers, Boca Raton, FL, 1994. Chao, K.P., and S.K. Ong, Air Sparging: effects of VOCs and soil properties on VOC volatilization, in In Situ Aeration: Air Sparging, Bioventing, and Related Remediation Processes, R.E. Hinchee, R. Miller, and P. Johnson (eds.), pp. 103­110, Battelle Press, Columbus, OH, 1995. Johnson, R.L., P.C. Johnson, D.B. McWhorter, R.E. Hinchee, and I. Goodman, An overview of in situ air sparging, GWMR, vol. 13(4), pp. 127­135, 1993. Reddy, K.R., S. Kosgi, and J. Zhou, A review of in­sim sparging for the remediation of VOC­contaminated saturated soils and groundwater. Hazardous Wastes & Hazardous Materials, vol. 12(2), pp. 97­117, 1995. U.S. EPA, Underground storage tanks: technical requirements. Federal Register 53:37082, Sept. 23, 1988. 5 CHAPTER TWO. LITERATURE REVIEW Air Sparging: Process Description Contamination of soil and groundwater as a result of accidental spills, leaking underground storage tanks, uncontrolled waste disposal, and leaking landfills is one of the major environmental concerns in the U.S.. Increasingly, irmovative strategies are being proposed for the remediation of contaminated soil and groundwater. The direct injection of air into contanainated groundwater was first proposed for the cleanup of contaminated sites as early as 1974 (Raymond, 1974). However, field application of air sparging has only been applied in the last 10 years. The reduction in remediation time and cost by air sparging may be as much as 50% as compared to a pump­and­treat and soil vapor extraction (SVE) system (Control News, 1992). Air sparging consists of the injection of air into the water saturated zone for the purpose of removing organic contaminants by a combination of volatilization and aerobic biodegradation processes (Johnson et al., 1993). The injection of air under pressure in the saturated zone displaces water from the porous matrix and creates a transient air­filled porosity. The contaminants are transported to the top of the saturated zone and into the unsaturated zone where they are captured by SVE wells. A typical air sparging system has one or more wells where air is injected into the saturated zone (Figure 1). Different mass transport and transformation processes occur during air sparging operations. The transport processes include volatilization, dissolution, diffusion, advection, hydrodynamic isolation and desorption of contaminants (Reddy et al., 1995). The transformation processes include enhanced biodegradation of VOCs from an increase in dissolved oxygen in the contaminated aquifer. The advective­dispersive characteristic of the air flow plus the dispersive and, possibly, advective characteristics of the groundwater will affect the extent of mass transfer during air sparging. Mass transfer is also affected by the diffusive properties of the VOCs. 6 Treated Air Air Treatment Unit Collection Manifold Air Compressor Ground Level t Vacuum Svstem t Vapor Extraction Well Vapor Extraction Well Vadose Zone Injection Well Water Table J Air Channels Contaminated Aquifer Well Screen Impervious Layer Figure 1. Typical air sparging system 7 Several treatment technologies which are similar in concept to the conventional air sparging system have been developed. These technologies include the horizontal trench sparging system, in­well air sparging, and biosparging (Suthersan, 1997). Horizontal trench sparging was developed for low permeability conditions and for contamination located less than 30 feet below ground level. Horizontal trench sparging has generated much interest recently as a result of a successful application of the technology at the Savannah River Integrated Demonstration Project (Plummer et al., 1997). In­well air sparging or the vacuum­ vaporizer­well technology (UVB) was developed by B. Bemhart (Buermann and Bott­ Breuning, 1994). Air is injected into the inner casing of the well containing the contaminated groundwater but the sparged air does not come in contact with the contaminated soil matrix. The injected air produces an air­lift effect which volatilizes the VOC and reoxygenate the groundwater and, at the same time, circulates the groundwater through the well casing and into the aquifer. Biosparging, is a modification of air sparging, where air is injected into the aquifer at very low flow rates (0.5 cfm to less than 2­3 cfm per injection well) to promote biodegradation (Suthersan, 1997). Despite the successful application of the air sparging technology at several contaminated sites, Chao and Ong (1995) pointed out that the current design and implementation of air sparging systems were based on field experiences only. Some areas where information is lacking and further research are needed to provide a scientific basis for the design of air sparging system include: (i) field methods and/or empirical correlations for the estimation of air saturation and the size and density of air channels (ii) the importance of different mass transfer mechanisms during air sparging (iii) the influence of saturated soil hydrogeology on the mode and behavior of air movement. Nature of the Injected Air Flow in the Porous Media Researchers in the field of fluid flow through porous media have used the continuous approach to explain the macroscopic transport processes in porous media. However, fluid 8 flow and diffusion in porous media may be viewed as taking place within extremely complicated microscopic boundaries that make any rigorous solution of the transport equation in the capillary network of the porous media practically impossible. The continuous approach does not consider the microscopic structure of the porous media channels and fails to explain the behavior of the fluids on the microscopic scale (Dullien, 1979). The exact nature of air flow in a saturated porous media is not completely understood. Research work in this area is fairly scarce. At the present time, there is no comprehensive study on the pathway and nature of air flow in an aquifer. Several conceptual models on air flow in saturated porous media under air sparging conditions have been proposed. Brown and Fraxendas (1991) proposed that air flow in the saturated zone is in the form of air bubbles which move upwards through the saturated soil column into the unsaturated zone. They suggested that the air movement was analogous to air bubbles formed in an aeration basin and that air sparging in an aquifer may be viewed as a crude air stripping tower in the subsurface with the soil acting as the packing. Other authors such as Middleton and Miller (1990) suggested that the movement of air was through irregular air channels in the saturated soil. The rationale for suggesting a continuous slug of air instead of bubbles was that air tends not to form bubbles in a tightly packed porous medium. In a laboratory study on air flow in porous media, Ji et al. (1993) found that the most probable air flow behavior in medium to fine grained water saturated porous media was in the form of discrete charmels. The authors reported that for grain diameters of 4 mm and larger (medium to coarse gravel), bubbly air flow was observed; for grain sizes of 0.75 mm or less (corresponding to sands, sUts, and clays) air flow in channels was prevalent. The transition between these two flow regimes appeared to occur for grain sizes between 4 mm and 0.75 mm. The authors concluded that under natural subsurface conditions, the air chaimel flow regime would likely occur during air sparging. The asymmetric pattern of air channels formed for a porous media with a mixture of grain sizes suggested that the air channels were sensitive to the medium heterogeneity. The pathway and nature of air flow in saturated porous media are affected by the air flow rate, air pressure, soil characteristics (particle size and shape), soil permeability, presence of 9 active surface agents, and the size of well screen or air orifice. The type of soil, soil permeability, and the so­called "air entry pressure" would determine the ease at which air will flow in the porous media. When air is pumped into a well, it must first displace the standing water within the well to the top of the screen opening of the well. The minimum air pressure. Py^; to achieve this is equal to the hydrostatic pressure of the standing water: Pw=pwghw (I) where p^. (ML'^) is the density of the water, (L) is the water level above the top of the well screen, and g (LT") is the acceleration due to gravity. After reaching the top most opening of the well screen, the injected air will penetrate the aquifer only if the air pressure exceeds the head loss through the well screen (P^) and the capillary pressure of the porous media. At the interface of a steady state, two­fluid system in porous media, the difference between the pressure of the two fluids is balanced by the surface tension of the interface. This interfacial tension may be quantified by the capillary pressures within the pores of the porous media. The capillary pressure is inversely proportional to the mean radius of curvature of the interface between the two fluids and may be quantified using the Laplace's equation. P,= P"­P'=— (2) r.. I "y where Pg (ML' T") is the capillary pressure, P" and P' are the pressures of the two different fluids, ct(MT") is the surface or interfacial tension, and (L) is the mean radius of curvature at the interface between the two fluids. The fluid with the concave side of the surface must have a pressure P " which is greater than the pressure P 'on the convex side. In general, the radius of curvature at the interface is a function of the porous media, the particle size, and the contact angle between water and the porous media. A decrease in the mean particle size diameter will correspond to a decrease in the mean radius of curvature. 10 Capillary entry pressure may be approximated using the following expression (Reddy et al. 1995): P,^ Cocosj y, where C (dimensionless) is the shape factor which is a function of the particle shape and distribution, <5is the contact angle, £ (dimensionless) is the void ratio, dpso (L) is the mean particle size, and (J(MT") is the interfacial tension previously defined. When the injected air pressure is greater than P^+Pe+Pd, the air displaces part of the pore water and creates a transient air porosity within the media. Air flow will initially be in a horizontal direction but the bouyancy of air will tend to cause the air to flow upwards. Therefore based on the above evaluation, the likely pathway for air flow in an aquifer will be through the pores with the lowest capillary pressure or pathways of least resistance. Changes in die permeability of the soil or in the soil structure will easily cause a change in the path of the air flow. In addition, when air channels hit a less permeable media, lateral dispersion of air may occur resulting in the entrapment or accumulation of a thin continuous air layer under the less permeable media. As a consequence of this phenomenon, some field data seemed to suggest that air flow may be restricted in soils with a hydraulic conductivity of 0.001 cm/sec or less (Middleton and Miller, 1990). Air injected into the aquifer may travel laterally and vertically depending on the injected air pressures (Brown et al., 1993). The furthest lateral displacement of the air in the saturated zone from the injection wells is defined as the radius of influence (ROI). Four common methods are used to determine the ROI of the sparging well: (i) pressure changes below the water table, (ii) increase in dissolved oxygen concentrations in the groundwater, (iii) mounding of the water table, and (iv) increase in VOC vapor concentration in the unsaturated zone (McCray and Falta, 1997). None of the four methods mentioned above has been shown to determine the ROI definitely. Given the asymmetric nature of the air channel distribution, some researchers have used the zone of influence or areal extent of influence {AEI) instead of 11 the ROI. The AEI is defined as an inverted cone or paraboloid centered at the sparging well (Reddy et al. 1995, Suthersan, 1997). For effective cleanup, the contaminant zone should be placed within the ROI or AEI. Although a knowledge of the nature of air flow in saturated porous media and the ability to predict air flow is central in understanding the physical­chemical processes occurring during air sparging, this work will not investigate the pathway and nature of air flow in saturated porous media. Instead, based on the above discussion and current work done by other researchers, the flow pattern of air will be assumed to occur in the form of discrete air channels. Mass Transfer and Transport Mechanisms VOCs in the subsurface may be present as a free phase, adsorbed phase, vapor phase, and dissolved phase. In an air sparged aquifer, the bulk liquid away from the air channel is expected to be quiescent rather than completely mixed. Therefore, mass transfer from the quiescent bulk solution to the air­water interface must be evaluated as a possible limiting mass transfer mechanism. Depending on the air flow rates, it is probable that soil particles adjacent to the air channels have most of their mobile water pushed away except for a thin layer of water attached to the soil particle. Therefore, on the onset of air sparging, the majority of the VOCs volatilized will come from this thin layer of water. Several mass transfer processes (Figure 2) may occur during air sparging: (i) volatilization of contaminants from groundwater into the sparged air (ii) desorption of contaminants from soils into groundwater (iii) advection and diffusion of contaminants in the liquid and gas phases (iv) dissolution of contaminants from free nonaqueous liquid phase (NAPLs) into groundwater (v) direct volatilization of NAPLs into the sparged air. 12 Advective VOCs NAPL Transpon in the Volatilization Air Phase VOC Diffusion Groundwate: Flow yOC Diffusion VOC Desorption NAPL NAPL Dissolution bissolutioi iC Diffusion ( VOC Volatilization O OnO/2 loil Particle VOC Desorption Air Channels Figure 2. Mass transfer processes occuring during air sparging 13 Volatilization of VOCs: Air­water mass transfer coefficients Volatilization of VOCs across the air­water interface for surface waters and in treatment process has been well researched. In these systems, several conceptual theories have been proposed to explain the mass transfer phenomenon. Conceptual theories developed for these systems may be adapted and applied for VOC mass transfer in saturated porous media during air sparging. The net flux of a chemical in a medium may be described by Pick's first law: F = ­D fi where F is the net flux of chemical per unit surface area (ML"^' ), D is the diffusion coefficient of the chemical in water or air (L^'), C is the concentration of the chemical (ML'^), z is the distance in the medium (L). Under equilibrium conditions, the distribution of VOCs between the water and air phases may be represented by Henry's law: KH=Ca/C^ (5) where Ca and C„­ are concentrations (ML'^) in the air and water phases, respectively, and Kh is the dimensionless Henry's law constant. Henry's law constants are temperature­dependent and their values for a wide variety of compounds are available in the literature (Mackay and Shiu, 1981). Henry's law constant predicts the relative distribution of VOCs in the air and water phases. VOCs with higher Henry's law constant will have a larger fraction of the VOC present in the gas phase. Several researchers have shown that under advective gas flow conditions such as air sparging, the movement of air across the air­water interface may not provide enough residence time for local equilibrium to be achieved (Mendoza and Frind, 1990). On the other hand, some researchers have found that the assumption of local equilibrium for air­water mass transfer is applicable (Brusseau et al., 1990; Gierke et al., 1992). If the local equilibrium assumption is not appHcable, a nonequilibrium mass transfer expression such as 14 the first order mass transfer relationship used by Sleep and Sykes is usually applied (1989). The nonequilibrium mass transfer expression commonly used is as follows: dCa/dt = Kl a^MCa/KH) ­ CJ (6) where Kl is the overall mass transfer coefficient (LT"'), and a^a represents the specific interfacial area (L"'). In the above expression, the driving force for the transfer of VOC across the air­water interface is equal to the difference between the equilibrium water concentration (Co/Kh) and the actual water concentration (Ch). The physical phenomenon occurring at the air­water interface has been investigated by several researchers. As a result of their research, several concepmal and hydrodynamic models have been proposed. A conceptual model which is presently accepted as describing the diffusive exchange of chemicals between water and air is the two­resistance model. The model was first described by Whitman in 1923 and first applied to environmental situations by Liss and Slater in 1974 (Mackay et al., 1990). The model agrees with the mass flux expressed in equation 4 except that the model assumes the existence of two thin stagnant films on each side of the gas­liquid interface and that a chemical or solute must diffuse in series through the two layers of water and air or viceversa. The rates at which mass are transferred are characterized by the mass transfer coefficients, k, which are essentially mass transfer velocities. Based on this model, the inverse of the overall mass transfer coefficient can be assumed to be the sum of the liquid and gas phase film resistance as given by: 1 _ 1 1 K ~ '^L k ^ ^KH ' k^ A where ki and kA are the Uquid and gas film mass transfer coefficients (LT'), respectively. For most chemicals with high Henry's law constant, the term KnkA is usually larger than k^, making the liquid film mass transfer, the controlling mass transfer mechanism. According to this model, the concentration gradient would be linear in each film (Figure 3) and the liquid and gas film coefficients may be given by: 15 Cu Bulk Aqueous Cone. Cwi Interfacial Aqueous Cone. Ca Bulk Gas Phase Cone. Cai Interfacial Gas Phase Cone. At the interface Well Mixed Air (Turbulent Transport) Distance from Interface Cai = K H C « i \ \ Air­Water Interface Distance from Interface Air Stagnant Layer. Sa NetVOCFlux^^ Aqueous Stagnant Layer. 5,, Well Mixed Water . Ca Cai Cwi Cw Figure 3. Two­resistance model Increasing Concentrations 16 kL = D^./S^. where kA = D A / S ^ (8) and Da are the diffiisivities of the chemical in water and air, respectively, and (L) and Sa (L) are the thickness of the stagnant layers or films in the water phase and the air phase, respectively. In many applications with tnrbulent flow, mass transport by diffusion is relatively small in comparison with other modes of mass transfer. Consequently, the two­resistance model may not be the best conceptual model to explain the air­water mass transfer under turbulent flow conditions. The concept of a stagnant layer at the air­water interface is probably the weakest assumption in the two­resistance model and to overcome this weakness, Danckwerts (1951) proposed the surface renewal theory. This theory is a conceptual expansion of the penetration theory developed by Higbie in 1935. The main concept is that the fluid at the surface is periodically renewed by turbulent eddies impinging on the air­water interface. Because of the turbulent nature, the thickness of the liquid phase film changes over time and space. When an eddy reaches the air­water interface, the eddy is exposed to the gas phase and nonsteady state diffusion of the compound into and out of the liquid phase occurs. Higbie proposed that the exposure time of the eddy, 6, is short enough that diffiasion into the liquid is slow and the dissolving compound is never able to reach the full depth of the eddy. Under these conditions, the average liquid film mass transfer coefficient may be expressed as: ki = 2(D^./enf^ Experience shows that the exponent on (9) may range from near zero to 0.9 (Treybal, 1968). Unlike the two­resistance model where mass transfer is dependent on the film thickness, mass transfer using the surface renewal theory is dependent on the frequency of renewal by the eddies. Dankwerts assumed that the renewal of the surface elements has a Gaussian distribution and found that the liquid and gas transfer coefficients may be expressed as a 17 function of the diffusivity (D) of the chemical in the liquid and air phase, respectively, as given below: (10) where and are the mean renewal rates on the water and air side, respectively. Several other conceptual theories have been proposed since the surface renewal theory was put forth by Danckwerts. O'Connor and Dobbins (1956) proposed the film renewal model which combines the two­resistance model and Danckwerts' renewal model. Arguing that the liquid content of the liquid film at the interface is continuously replaced in a random manner by liquid from the bulk region, O'Connor and Dobbins proposed that the liquid transfer coefficient may be given by: kt = (Dy, coth [(r^:Sj)/D^J (II) Equation 11 becomes equations 9 or 10 when the renewal velocity Tw approaches infinity or zero, respectively. Other models include the random eddy model of Harriot (1962) and the surface divergence model of Brumley and Jirka (1988). Brumley and Jirka model uses the divergence of the horizontal liquid velocity at the surface plane to interpret the mechanism of gas transfer at the near surface region. Several models which combine convective fluid flow in the interfacial region and measurable turbulence properties in the bulk region have been proposed. These models include the large­eddy model postulated by Fortescue and Pearson (1967), the small­eddy model of Lament and Scott (1970) and the model suggested by Theofanous et al. (1976). The last model combines the large­eddy model for low turbulent Reynolds number (<500) with the small­eddy model for high Reynolds number (>5(X)). Theofanous et al. found that the liquid film coefficient may be correlated with various dimensionless parameters as follows: 18 lcL=fn( kL=fn( Sc­^­^. Re­^­^^) for Re < 500 (12) for Re > 500 (13) where Sc is the Schmidt number (//„, /pw D^) and Re is the turbulent Reynolds number. Considerable amount of research has been devoted in determining the limiting mass transfer layer and estimating the mass transfer coefficients for gases and organic chemicals in wastewater and in natural bodies of water. In the field of wastewater, various types of aeration and air stripping equipment have been evaluated to assess the transfer of oxygen from air to water phase and the removal efficiency of VOCs from wastewater. Matter­Muller et al. (1981) concluded that the transfer of volatile substances from water to the air phase was strongly dependent on die type of gas­liquid contacting operation (bubble aeration, surface aeration, stripping towers, etc.). The authors also found that the mass transfer coefficients and the degree of saturation of the exit gas with respect to the VOC were important parameters in determining the transfer of the VOC from water to air. This conclusion supports the use of equation 6 to describe the mass transfer rate across the two phases. Munz and Roberts (1989) showed using a laboratory­scale mechanical surface aeration unit that the mass transfer coefficient for a volatile organic compound may be related to that of oxygen by a proportionality factor ifr. V ­ (^Li / f<^L02) = ( Du /Di,02f = (Ku /K[_02) (14) where ±e Ku and Kl,o2 are the overall mass transfer coefficients for the VOC and oxygen, respectively, and Du and Dlo2 are the aqueous molecular diffiision coefficients for the VOC and oxygen, respectively. Equation 14 is exttemely useful and has become increasingly popular because it can be used to predict the mass transfer coefficients of different VOCs based on their diffusivities. However, it must be emphasized that the proportionality factor y/ is a constant only when the transport mechanism for both compounds, i.e., VOC and oxygen, is controlled by the liquid phase resistance of the mass transfer. If the gas phase resistance would in fact contribute to the overall mass uransfer resistance in a significant way, the y/ 19 values will be smaller, and the potential of volatilization of a given compound will be overestimated accordingly. Munz and Roberts found significant deviations in the values of for the least volatile compounds and concluded that the gas phase mass transfer resistance is more important than previously assumed. As a consequence of these findings, the authors concluded that mass transfer for compounds with dimensionless Henry's law constant > 0.55 was probably completely liquid phase controlled. The value of the constant n was found to vary between 0.5 and 1.0. The lower value of n, 0.5, is for turbulent conditions closely representing the surface renewal theory while the higher value, 1.0, represents the stagnant films as in the two­resistance model. In their study, the liquid mass transfer coefficient ki was found to be proportional to the square root of the diffusivity of the VOC in accordance with the surface renewal theory. The diffusivity or diffiision coefficient D, is a property of the chemical compound and is dependent on temperature, pressure and the physical­chemical nature of the components of the system such as the type of media. In the absence of experimental data, gas phase diffusivities can be estimated based on the kinetic theory of gases using the Hirschfelder­ Bird­Spotz correlation (Perry and Chilton, 1978). Estimation of the diffusivity in liquids in the absence of experimental data cannot be made with the same accuracy as for gases due to a lack of an adequate theory for the liquid structure. However, for dilute solutions of nonelectrolytes, the empirical correlation of Wilke and Chang may be used (Treybal, 1968): Dab = 7.4 xW' T/fi (15) where D^b is the diffusivity of A (L^', em's"') in dilute solution in solvent B, Mb is the molecular weight (M, g) of the solvent B, T is the absolute temperature (K), /n is the solution viscosity in centipoise (ML"'T'), va is the solute molal volume (cm^g"') at the normal boiling point (L^M"'), and (p is the association factor for the solvent (2.6 for water as a solvent). The value of may be estimated from tables based on the individual atomic and molecular conuibutions to the molal volume. Due the changes in viscosity with concentration and the 20 changes in the degree of ideality of the solution, equation 15 cannot be applied for concentrated solutions. In natural bodies of water, wind speed along with the presence of wave fields and water currents have been shown to be important parameters controlling mass transfer at the air­ water interface (Janhe et al., 1984). Studies by Mackay and Yuen (1983) have shown that the mass transfer coefficients for various VOCs were strongly dependent on the wind speed. Volatilization rates were measured using a 6 meters wind­wave tank for 11 VOCs. Experimental data confirmed the validity of the two­resistance model for high wind speeds and showed that there were no interaction amongst the VOCs when the VOCs were volatilized simultaneously. The authors also found that the mass transfer coefficient was a fiinction of the wind friction velocity (the hydrodynamic parameter) and the Schmidt number which characterized the solute diffusion properties and temperature effect. As previously stated, an air sparged aquifer may be visualized as an array of air channels with the volatilization of VOCs occurring at the air­water interface located at the wall of each air channel. The effectiveness of air sparging as a remediation technology is related to the ability of the system to provide maximum contact between the contaminated aquifer materials and the air flow. So far, information on mass transfer coefficients for the volatilization of VOCs under sparging conditions is fairly scarce (Szatkowski et al., 1995). In a later section of this chapter, correlations for mass transfer coefficients of VOCs under different experimental conditions will be presented. Sorption­desorption of VOCs onto porous media Much of the information obtained for the volatilization of VOCs from bulk water may be applicable for air sparging systems. However, the presence of the porous media introduces new factors which may affect the volatilization of VOCs. These factors include the increase in the distance that VOCs must travel to reach the air­water interface and sorption of the VOCs onto the soil matrix. For soil media with a high affinity for VOCs, desorption of VOCs from the soil to the groundwater may play a role in mass uransfer. VOCs with high 21 partition coefficients are less likely to volatilize even when the VOC has a high Henry's law constant. Adsorption of the VOCs is defined as the accumulation of VOCs at the water­solid interface while absorption is defined as the partition of VOCs between two phases. The term sorption is used when both adsorption and absorption are occurring and cannot be distinguished from one another. Partitioning between the dissolved and sorbed solutes for a given temperature may be modeled using linear or nonlinear isotherms as presented below: Cs = Kd Cw (16) Cs=Kfc;, (17) where Cs represents the concentration of the VOCs sorbed onto the soil matrix (MM"'), Cv is the concentration of VOC in the hquid phase (ML'^), Kd is the linear partition coefficient (L^M"'), and /iT/represents the Freundlich partition coefficient (L^"M""). Due to the nonpolar nature of most VOCs, sorption of VOCs onto soils may be correlated with the organic matter content in the porous media. While the linear partition coefficient, Kd, of a chemical varies significantly from soil to soil, the ratio between Kd and the mass fraction of organic carbon in the soil (foe) is less variable. This ratio results in a proportionality constant for each organic compound known as the organic carbon partition coefficient (Koc)­ This relationship can be represented as (Karickhoff et al., 1979; Karickhoff, 1984): Kd — foe Koc (18) The above relationship is considered to be valid for organic carbon contents greater than 0.1%. When contaminants are in contact with low organic carbon aquifer material (foe < 0.001), adsorption of contaminants may play a significant role. Sorption of organic products may be two to four times higher than the values predicted by equation 18 (Piwoni and Baneije, 1989). Most organic contaminants are rapidly sorbed onto soils but desorption 22 can be very slow. This phenomenon tends to reduce the efficiency of remediation technologies such as air sparging (Suthersan, 1997). The linear isotherm (equation 16) has been widely used to describe the transport of VOCs in the subsurface (Gierke et al., 1990, Ong and Lion, 1991). Karickhoff (1984) suggested that use of linear isotherm is a reasonable assumption if the contaminant concentrations were lower than 10'^ M or less than half the water solubility of the organic contaminant. Based on experimental results, Karickhoff (1979) derived an empirical correlation between the organic carbon partition coefficient and the octanol­water partition coefficient (Kok)'­ log Koc = log Kow ­ 0.21 (19) The above equation is useful for estimating the organic carbon partition coefficient of an organic compound from the chemical properties of the compound. As presented in the previous paragraph, sorption of VOCs to soil particles may play an important role in the distribution of the VOCs in the solid­liquid­gas phase system such as air sparging. Ong and Lion (1991) showed that the sorption of TCE in different soils and under different moisture conditions may be accounted for by: (a) dissolution of TCE in the water bound to the soil particle as governed by Henry's law (this contribution can constitute a large fraction of the total mass of VOC for soils with low organic carbon content), and (b) sorption of TCE at the solid­liquid interface as governed by the saturated partition coefficient. VOCs with high partition coefficients are less likely to volatilize even when the VOCs have high Henry's law constants. Under a dynantiic air flow situation, removal of VOCs at the air­water interface may occur so fast that the desorption of the VOCs could become the rate limiting step in the transfer of VOCs into the gas phase. Pavlostathis and Mathavan (1992) conducted batch desorption experiments with field­ contaminated soils and found that a substantial portion of the sorbed contaminant resisted desorption. The desorption pattern showed an initial fast desorption phase followed by a slow desorption phase. The authors also found that the rate and extent of desorption were independent of the soil and VOC properties such as organic carbon content, cation exchange 23 capacity, specific surface area, and water solubility. Several researchers have modeled the soq)tion­desorption phenomenon as a first order kinetic process (Brusseau and Rao, 1989; Armstrong et al., 1994). The rate limited sorption­desorption model may be expressed as: ^ = ks(c.­c:) (20) where C5 is the concentration of VOC sorbed onto the soil particles (MM"'), C,v is the acmal aqueous concentration of the VOC (ML'^), C*»­ is the equilibrium aqueous concentration of the VOC in contact with the soil surface (ML"^), and ks is the first order sorption­desorption coefficient (L^M"'T'). Using the linear isotherm described by equation 16, equation 20 can be rearranged as follows: ^ = kAK,C.­C,) at (21) where the dimension of ks is T Weber et al. (1991) studied concepmally the sorption phenomenon in subsurface systems and their effects on contaminant fate and transport. The emphasis of this study was the development of concepts involved in the sorption­desorption phenomenon and the translation of these concepts into functional models for characterizing sorption­desorption rates and equilibrium. Several conceptual mass transfer models applicable to simations such as the volatilization of volatile compounds from solids and liquids, the dissolution of nonaqueous phase liquid into water and treatment process were presented in this study. The authors stated that under the typical fluid flow conditions found in subsurface systems, molecular diffusion generally dominates the microscopic mass transfer. This molecular diffusion has two main components: (a) random or Fickian diffusion, and (b) Knudsen diffusion or "constrained diffusion". Knudsen diffusion occurs when molecular velocities and ratios of longitudinal to radial pore lengths are high, such as in the gas phase. However, according to 24 the authors, Fickian motion as described by equation 4 is the dominant diffusion mechanism in liquid phase. Finally, Nadim and coworkers (1997) studied the mass transfer limitations for the desorption of VOCs and the liquid phase diffusion of organic molecules in unsaturated soils during soil vapor extraction operations. The authors concluded that the gas phase effluent concentrations were a result of: (i) removal of compounds initially dissolved in the liquid layer which was responsible for the rapid exponential decrease of the VOCs concentration in the effluent air, and (ii) slow desorption for the VOCs from the solid phase which was the reason for the long tailing in the effluent concentration observed during soil vapor extraction. Dissolution and volatilization of nonaqueous phase liquids (NAPLs) in tiie subsurface In general, mass transfer limitations observed in the field have been attributed to the slow diffusion of the dissolved pollutants from low permeability porous media to the mobile groundwater or to the gas phase of the vadose zone. Releases of organic products from hazardous waste sites may also result in the migration of contaminants as a separate organic phase. The behavior of these organic phases is controlled by three major forces: capillary forces, viscous forces, and buoyancy forces. In a typical aquifer, organic liquid is trapped in samrated porous media as a result of capillary forces. Depending of their densities, NAPLs are classified as light nonaqueous phase liquids (LNAPLs) and dense nonaqueous phase liquids (DNAPLs). LNAPLs are less dense than water and are vertically transported through soil until they reach the capillary fringe where they spread horizontally forming floating pools. DNAPLs, which are more dense than water, are vertically transported through soil by gravity and capillary forces until they reach an impermeable layer. Conrad et al. (1992) in a visualization study of residual organic liquid trapped in aquifers found that large amounts of organic contaminants were trapped as isolated microscopic blobs. The size, shape, and spatial distribution of these blobs of residual organic liquid affected the dissolution of the organic contaminants into the groundwater. For an air sparged NAPL­contaminated aquifers. Baker et al. (1996) proposed that the rate of mass removal was time dependent and depended on the 25 degree of contact between the air and the NAPL ganglia. A fast removal rate is achieved initially as contaminant closest to the air channels are volatilized. Subsequent mass removal is slower because contaminants must diffuse to the air channels before being removed. Wilkins et. al (1995) smdied the volatilization of NAPLs in unsamrated sandy porous media and found that the effluent concentration of the contaminant in vapor phase deviated from local equilibrium values by 10­40 % for pore velocities ranging from 0.25 to 1.5 cm/s. In addition the mass transfer rates were found to decrease with decreasing soil mean grain size. Powers et al. (1992) investigated the steady state dissolution of NAPLs in saturated subsurface systems and found that the dissolution rate of NAPLs was dependent on the distribution pattern of the NAPLs and the aqueous phase velocity. In addition. Powers et al. (1994) found that the length of time required to dissolve NAPLs was substantially longer than the times predicted by equilibrium calculations. The dissolution of NAPLs may be modeled using a first order kinetic expression: (22) where is the water filled porosity (dimensionless), S represents the aqueous solubility of the organic chemical (ML'^), Cw represents the acmal concentration of the organic chemical in the liquid phase (ML'^), and or represents the mass transfer coefficient (T"'). Powers et al. (1994) reported that longer cleanup times were associated with coarse or graded media and speculated that this was due to the larger and more amorphous NAPL blobs present in the media. Miller et al. (1990) successfully used a local equilibrium assumption to compute mass transfer rates between a toluene NAPL and flowing water in a laboratory study using glass beads as a porous media. The findings of this study were in agreement with the work by Powers and her coworkers, except that mass transfer rate was not significantly related with the mean particle size. 26 In a series of modeling studies on the application of horizontal S VE systems, Gomez­ Lahoz et al. (1994a, and 1994b) and Rodriguez Maroto et al. (1994) showed that remediation by SVE may be diffusion controlled, limited by the dissolution of the NAPL into the aqueous phase and/or desorption controlled. Their modeling studies showed that the remediation time was dependent on the length of ±e clay lenses which the organic contaminants have to cross before reaching into the gas phase. In addition, remediation time was dependent on the diameter of the NAPLs and the initial concentration of the contaminants. These authors found that at the start of the extraction, the VOC concentration in the soil gas was very high but dropped rapidly to a fairly low constant tailing concentration. The time of the tailing was observed to increase with an increase in the diffusion layer thickness. The information presented above, suggested that the presence of NAPLs may introduce new limitations in the transport and removal of VOCs during air sparging operations. Modeling Mass Transfer Due to the complexity associated with air flow through saturated porous media, a fundamental approach in predicting the mass transfer rates of VOCs for air sparging is currently not available in the literature. Several empirical air­water mass transfer correlations derived in the field of chemical engineering and for soil vapor extraction systems will be presented to provide a quantitative idea of mass transfer coefficients for air sparging systems. Mass transfer coefficients for the volatilization of organic solutes from water were estimated by many researchers. Among them, Mackay and Yuen (1983) studied the volatilization of 11 different organic compounds from aqueous phase and concluded that the two­resistance model may be applied to model mass transfer. They also found that the Schmidt number, Sc (dimensionless ratio between viscosity and the product of density times molecular diffusivity) correlated well with the mass transfer coefficients. Linton and Sherwood (1950) developed a dimensionless correlation that may be applied for the estimation of the overall liquid side mass transfer coefficient for air flowing through porous media in the form of discrete air channels. The correlation was developed using data from Sherwood and Gilliland for wetted­wall towers and their own mass transfer smdies for 27 water flow through pipes with soluble walls. The correlation developed by Linton and Sherwood is presented below. The Reynolds number for the correlation ranged from 2,000 to 35,000 while the Schmidt number ranged from 0.6 to 3000. Sh = 0.023 Re°­^^ Sc'^^ (23) where Sh represents the Sherwood number (dimensionless mass transfer coefficient), and Re and Sc represents the Reynolds and Schmidt numbers, respectively. Mathematical description of these dimensionless parameters are presented in Appendix A. Using soil columns packed with Ottawa sand and other aggregate porous media. Gierke and coworkers (1992) showed that the vapor transport of toluene in the unsaturated soil columns was a nonequilibrium transport process. With the aid of mathematical models, mass transfer was shown to be impacted by the dispersion, film transfer, and intraaggregate diffusion. The mass transfer coefficient used in the model was derived from that of Crittenden and coworkers (1986) and is given as follows: K,a= 15 DS r (24) where Kta is the overall liquid side lumped mass transfer coefficient (T"'), Dp represents the intraaggregate diffusion coefficient (L^"'), S is the degree of saturation (dimensionless), Kh is the dimensionless Henry's law constant, and Ra is the aggregate radius (L). Work done by Cho and Jaffe (1990) showed that during infiltration into a soil column, equilibrium conditions for the volatihzation of VOCs between the dissolved and gas phase cannot always be used. Szatkowski et al. (1995) used an experimental semp with a constant interfacial area between phases to investigate the resistance to aqueous­vapor mass transfer at the air­water phase boundary. Using dimensional analysis, Szatkowski and coworkers found that the modified Sherwood number correlated to several macroscopic properties of the system. The correlation obtained by Szatkowski et al. was as follows: 28 Sh'.. = flow (0.0235c°^ + 0.849 Sc°J ) for 0.001< Re^­ <0.1 (25) The modified Sherwood number for the aqueous phase (5/2^ ') was defined as: Sh ­ K, a„^ d' L p D k d' —p_ D ,26) Definition of the variables may be found in Appendix A. Szatkowski et al. reported that predictions from the correlation described in equation 25 were in good agreement with work done by other researchers (Cho et al., 1994; and Turek and Lange, 1981). A shortcoming of this smdy was that only one VOC and one type of porous media was used. Because of this shortcoming, the mass transfer coefficients were correlated with the hydrodynamic conditions (air and water velocities) of the experiment only, but not with the chemical properties of the VOCs and physical properties of the porous media. Based on experimental and computational simulations of TCE extraction in sandy soils, Armstrong et al. (1993) concluded that the mass uransfer processes may not be represented by a simple first order rate constant but rather by a time dependent mass transfer coefficient. They used a decreasing power law relationship for the mass transfer coefficient and studied the influence of the Damkohler number on mass ttansfer. The Damkohler number is the ratio between mass partitioning and mass removal by advection. A shortcoming of this study and probably a reason why a time dependent mass transfer coefficients was considered, was that diffusion effects were not included in the transport equation for the dissolved phase. Mohr (1995) assessed the influence of groundwater flow on air­water mass transfer rate by using an empirical correlation. The correlation was developed for water flowing around a sphere of air and is given by: = 2 + 0.6 Re'^Sc'^^ (27) 29 where Nu is the Nusselt number representing the ratio of mass transfer with and without convection. Re and Sc are the Reynolds and Schmidt numbers, respectively. Definition of these dimensionless numbers may be found in Appendix A. For typical groundwater conditions of Sc = 1000 and a groundwater flow of 0.3 m/day, Mohr showed that the Nu number, which may be viewed as dimensionless mass transfer coefficient, increased by less than 35% in comparison to the stagnant water solution. This implies that groundwater flow has minimum impact on mass transfer at the air­water interface. Chao (1997) derived a correlation for the dimensionless Sherwood number (which includes the mass transfer coefficient) with the air phase Peclet number (Pe), the dimensionless mean particle size (do), and the Henry's law constant {Kh). The correlation was statistically obtained from a series of nonequilibrium air­water mass transfer experiments using a bench scale air sparging system. The correlation was: Sh = 10­'­^' Pe°­^ do'­^' (28) Miller and coworkers (1990) studied the dissolution of NAPLs in glass beads with toluene as a contaminant. The experiments were conducted using packed columns and at various NAPLs fluid samration levels. The authors found that the interfacial mass transfer rate was directly related to the aqueous phase velocity and nonaqueous phase fluid samration while no significant dependence on the mean particle size was found. Powers et al. (1992), investigating the steady­state dissolution of NAPLs trapped within saturated porous media, developed a phenomenological model correlating the modified Sherwood number with the grain size (dso), uniformity coefficient (C/C), and water phase Reynolds number. The correlation is given by: Sh' = 57.7Re°­^' dso"'^ (29) 30 where Sh' is the modified Sherwood number as defined in Appendix A. In a later study Powers and coworkers (1994) extended the correlation model (equation 29) by incorporating changes in the interfacial area as a result of NAPL dissolution. The new model can be represented as: 0.S98 . 0.673 (30) where the ratio i9,/6o) is the ratio between the acmal and initial NAPL volumetric fraction. The exponent yS/ varied between 0.75 and 0.96 depending on the porous media. The authors were able to correlate the value of the exponent with the dimensionless mean grain size and the uniformity coefficient of the porous media. Wilkins et al. (1995) correlated the volatilization of NAPLs with several physical properties of the porous media. The mass transfer coefficient is given by: «<nv = 10 0,38..0.62 jO.44 d. flso (31) where kaw is the lumped mass transfer coefficient (T"'). Other variables in equation 31 are defined in Appendix A. Wilkins and coworkers compared the results of their smdies for the volatilization of NAPLs in unsamrated porous media with that of Powers et al. (1992) for the dissolution of NAPLs in saturated porous media. Wilkins and coworkers pointed that for saturated systems, the rate of NAPLs dissolution was inversely correlated to the soil mean size while in unsaturated porous media, volatilization of NAPLs was positively correlated with soil mean grain size. In addition, the dissolution rate was strongly correlated with the uniformity coefficient for unsaturated porous media while the uniformity coefficient had no effect in saturated porous media. Wilkins and coworkers attributed the differences to the differences in NAPL configuration and the differences in the flowing phase in saturated and unsamrated porous media. 31 VOCs transport in porous media Several models explaining the mass transfer of oxygen and VOCs at the air­water interface may be found in the literature. These models are greatly simplified and do not exactly represent the conditions for a wide range of air flow rates during air sparging. Among them are the simple steady­state and nonsteady­state models developed by Wilson and coworkers (Clark et al., 1996). A more elaborate model was developed by Marley et al. (1992) which described the distribution of air pressure and air velocity during the air sparging operation. The model assumed that: (i) stagnant conditions for water, (ii) steady state mass transfer conditions, (iii) no biodegradation, and (iv) air wells acted as line sources with air flow obeying Darcy's law. Sellers and Scherber (1992) derived an exponential equation describing the change in groundwater concentration when air sparging was used. They assumed that the rate of pollutant diffusing into the sparging bubbles was balanced by the loss of dissolved phase contaminant from the groundwater. A model developed by Sleep and Sykes (1993a, 1993b) included three­phase flow and interfacial mass transfer due to volatilization and dissolution but assumed equilibrium conditions between the phases. Other attempts to model air sparging systems included the model by Gvirtzman and Gorelik (1992) and the work of Ostendorf et al. (1993). In a review paper on air sparging, Reddy et al. (1995) presented a model from the work of Abriola and Finder (1985) on multiphase transport of organic compounds in porous. The modified model included microscale volatilization and biodegradation processes, macroscale advection and dispersion processes, and interfacial mass transfer processes such as adsorption, desorption, dissolution, and volatilization. A more refined model which incorporated the diffusion of die VOCs from the porous matrix to the air channels was proposed by Hein et al. (1994). The model showed that the mass transfer was dominated by gaseous dispersion and advection, and liquid diffusion. The model equations are as follows: 32 SC^.(z,rj) ^ fdC^{z,rjy dr \ r J^ry dr + 2St[C^.(z, r, r) ­ C(z, r)] (32) (33) where Cw and C are the dimensionless water and air concentrations of the pollutant, respectively; z and r are the dimensionless axial and radial coordinates, respectively; t is the time elapsed; Ed is the pore diffusion modulus (rate of transport by radial diffusion/rate of transport by advection); Pe is the air phase Peclet number (rate of transport by advection/rate of transport by axial dispersion); and St is the Stanton number (rate of transport by air/water mass transfer/rate of transport by advection). The description of all these dimensionless parameters may be found in Appendix A. The numerical model was shown to be more sensitive to changes in the air phase Peclet number and pore diffusion modulus. Changes in the Stanton number did not affect the model output and this was in accordance with the previous work done by Gierke et al. (1992) for SVE systems. The work done by Hein et al. supported the hypothesis that the mass transfer was controlled by molecular diffusion. Hein et al. model was developed by assuming a vertical air charmel and that VOCs diffused from the porous media to the air channel. Under these assumptions, the term representing the interfacial mass transfer between the liquid phase and the gas phase as presented in equation 33 should be stated as a boundary condition linking the transport between the two phases rather than as defined in the model. Drucker and Di Julio (1996) developed a diffusion limited model which is similar to that of Wilson et al. (1994). Predictions of the model seemed to agree with results from a soil column experiment using TCE as the contaminant. No adsorption effects were considered in the development of the model and its validity for other VOCs was not proven. Sensitivity analysis with the model indicated that the diffusion coefficient and the air channel radius had the biggest impact on remediation rate. The influence of Henry's law constant was limited to values between 0.01 and 0.35. For values smaller than 0.01, the rate of volatilization is the controlling factor in the remediation while for values larger than 0.35 diffusion of VOC in the 33 liquid phase controlled the process. The analysis showed that beyond a certain air flow rate, no further improvement in the removal rate was achieved. In general, current models addressed the removal of VOC by air sparging on a macroscale. The influence of microscale processes such as the diffusion of VOCs through the porous media to the air channels has not been fioUy addressed yet. This research will study the influence of "microscale" processes governing the volatilization of VOCs and the results from a single­air chaimel system will be incorporated into a model with the aim of explaining the behavior of mass transfer in more complex systems. References Abriola, L.M., and G.F. Finder, A multiphase approach to the modeling of porous media contamination by organic compounds. I. Equation Development, Water Resources Research, vol. 21(1), pp. 11­18, 1985. Armstrong, J.E., J. Croise, and V. Kaleris, Simulation of rate­limiting processes controlling the vapour extraction of trichloroethylene in sandy soils, in Proceedings of the International Conference on the Environment and Geotechnics, Paris, France, pp. 327­334, April 6­8, 1993. Baker, D.M., and C.H. Benson, Review of factors affecting in situ air sparging, in Proceedings of the Non­Aqueous Phase Liquid (NAPLs) in Subsurface Environment: Assessment and Remediation, edited by L.N. Reddi, Washington, D.C., pp. 292­310, November 12­14,1996. Brown, R.A., and R. Fraxendas, Air sparging extending volatilization to contaminated aquifers, Prepublication draft presented at the Symposium on Soil Venting, Robert S. Kerr Environmental Research Laboratory, Houston, TX, April 29­May 1, 1991. Brown, R.A., R.J. Hicks, and P.M. Hicks, Use of air sparging for in situ bioremediation in Air Sparging for Site Remediation, edited by R.E. Hinchee, pp. 38­55, Lewis Publishers, Boca Raton, FL, 1994. 34 Brumley, H.B., and G.H. Jirka, Air­water transfer of slightly soluble gases: turbulence, interfacial processes and conceptual models, J. Physico­Chemical Hydro., vol. 10(3), pp. 295­319, 1988. Brusseau, M.L., and P.S.C. Rao, Sorption nonideality during organic contaminant transport in porous media, CRC Critical Reviews in Environmental Control, vol. 19(1), pp. 33­99, 1989. Brusseau, M.L., Transport of organic chemicals by gas advection in structured or heterogeneous porous media: development of a model and application to column experiments. Water Resources Research, vol. 27(12), pp. 3189­3199, 1991. Brusseau, M.L., R.E. Jessup, and P.S.C. Rao, Sorption kinetics of organic chemicals: evaluation of gas purge and miscible­displacement techniques. Environ. Sci. TechnoL, vol. 24(5), pp. 727­735, 1990. Buermann, W., and G. Bott­Breuning, Bioremediation by groundwater circulation using the vacuum­vaporizer­well (UVB) technology: basics and case smdy, in Air Sparging for Site Remediation, edited by R.E. Hinchee, pp. 97­107, Lewis Publishers, Boca Raton, FL, 1994. Chao, K. P., Aqueous­vapor mass transfer of VOCs in saturated porous media under air sparging conditions, Ph.D. Dissertation, 176 pp.. Polytechnic University, Brooklyn, NY, 1997. Chao, K.P., and S.K. Ong, Air Sparging: effects of VOCs and soil properties on VOC volatilization, in In Situ Aeration: Air Sparging, Bioventing, and Related Remediation Processes, R. Hinchee, R. Miller, and P. Johnson (eds.), pp. 103­110, Battelle Press, Columbus, OH, 1995. Cho, H.J., and P.R. Jaffe, The volatilization of organic compounds in unsaturated porous media during infiltration. Journal of Contaminant Hydrology, vol. 6, pp. 387­410, 1990. Cho, H.J., P.R. Jaffe, and J.A. Smith, Simulating the volatilization of solvents in unsaturated soils during laboratory and field experiments. Water Resources Research, vol. 29(10), pp. 3329­3342, 1994. Clark, A.N., D.J. Wilson, and R.D. Norris, Using models for improving in­situ cleanup of groundwater. Environmental Technology, July/August, pp. 34­41, 1996. 35 Conrad, S.H., J.L. Wilson, W.R. Mason, and W.J. Peplinski, Visualization of residual organic liquid trapped in aquifers. Water Resources Research, vol. 28(2), pp. 467­478, 1992. Control Technology News, VOC remediation costs, time frames reduced up to 50% witii air sparging, J. Air Waste Management Assoc., vol. 42(5), p. 729, 1992. Crittenden, J.C., N.J. Hutzler, D.G. Geyer, J.L. Oravitz, and G. Friedman, Transport of organic compounds with saturated groundwater flow: model development and parameter sensitivity. Water Resources Research, vol. 22(3), pp. 271­284, 1986. Danckwerts, P.V., Significance of liquid­film coefficients in gas adsorption, Ind. Eng. Chem., vol. 43(6), pp. 1460­1467, 1951. Drucker, A.S., and S.S. Di Julio, Groundwater clean up by In Situ air sparging: development of a model and application to saturated soil column experiments, in Proceedings of the Water Environmental Federation 69 th. Annual Conference & Exposition. Dallas, Texas, October 5­9, 1996. Dullien, F.A.L., Porous Media. Fluid Transport and Pore Structure, Academic Press, Inc.. NY, 1979. Fortescue, G.E., and F.R.E. Pearson, On gas absorption into a turbulent liquid, Chem. Eng. Sci., vol. 22, p. 1163, 1967. Gierke, J.S., N.J. Hutzler, and J.C. Crittenden, Modeling the movement of volatile organic chemicals in columns of unsaturated soil. Water Resources Research, vol. 26(7), pp. 1529­ 1547, 1990, Gierke, J.S., N.J. Hutzler, and D.B. McKenzie, Vapor transport in columns of unsaturated soil and implications for vapor extraction. Water Resources Research., vol. 28(2), pp. 323­ 335, 1992. Gomez­Lahoz, C., J.M. Rodriguez­Maroto, and D.J. Wilson, Soil cleanup by in­situ aeration. XVn. Field scale model with distributed diffusion. Separation Science and Technology, vol. 29(10), pp. 1251­1274, 1994a. Gomez­Lahoz, C., J.J. Rodriguez, J.M. Rodriguez­Maroto, and D.J. Wilson, Biodegradation phenomena during soil vapor extraction, m. Sensitivity smdies for two substrates. Separation Science and Technology, vol. 29(10), pp. 1275­1291, 1994b. 36 Gvirtzman, H., and S. Gorelik, The concept of in situ vapor stripping for removing VOCs from groundwater, Transport in Porous Media, vol. 8, pp. 71­92, 1992. Harriott, P., A random eddy modification of the penetration theory, Chem. Eng. Sci.. vol. 17, pp. 149­154, 1962. Hein, G.L., N.J. Hutzler, and J.S. Gierke, Quantification of the mechanisms controlling the removal rate of volatile contaminants by air sparging, in Proc. of the 1994 National Conference on Environmental Engineering, J.N. Ryan and M. Edwards (eds), pp. 556­563, ASCE, NY, NY, 1994. Jahne, B., K.H. Fisher, J. Ilmberger, P. Libner, W.Weis, D. Imboden, U. Lemin, and J.M. Jaquet, Parameterization of air/lake exchange, in Gas Transfer at Water Surfaces, W. Brutsaert and G.H. Jirka (eds.), D. Reidel, Boston, pp. 459­466, 1984. Ji, W., A. Dahmani, D.P. Ahlfeld, J.D. Lin, and E. Hill, Laboratory study of air sparging: Air flow visualization, GWMR, vol. 13(4), pp. 115­126, 1993. Johnson, R.L., P.C. Johnson, D.B. McWhorter, R.E. Hinchee, and L Goodman, An overview of in situ air sparging, GWMR, vol. 13(4), pp. 127­135, 1993. Karickhoff, S.W., D.S. Brown, and T.A. Scott, Sorption of hydrophobic pollutants on natural sediments. Water Research, vol. 13, pp. 241­248, 1979. Karickhoff, S.W., Organic pollutant sorption in aquatic systems, /. Hydraulic Eng., ASCE, vol. 110, pp. 707­735, 1984. Lamont, J.C. and D.S. Scott, An eddy­cell model of mass transfer in the surface of a turbulent liquid, AIChE J., vol. 16, p. 513, 1970. Linton, W.H. and T.K. Sherwood, Mass transfer from solid shapes to water in stream line and turbulent flow. Chemical Eng. Progress, vol. 46(5), pp. 258­264, 1950. Lyman, W.J., W.F. Reehl, and D. H. Rosemblatt, Handbook of Chemical Properties Estimation Methods, American Chemical Society, Washington D.C., 1990. McCray, J.E., and R.W. Falta. Numerical simulation of air sparging for remediation of NAPL contamination. Ground Water, vol. 35(1), pp. 99­110,1997. Mackay, D., and W.Y. Shiu, A critical review of Hemy's law constant for chemicals of environmental interest, J. Phys. Chem. Ref Data, vol. 10(4), pp. 1175­1199, 1981. 37 Mackay, D., W.Y. Shiu, K.T. Valsaraj, and L.J. Ttiibodeaux, Air­water transfer: the role of partitioning, in Air Water Mass Tranter, Selected Papers from the Second International Symposium on Gas Transfer at Water Surfaces, S.C Wiltielms and J.S. Gulliver (eds.), ASCE, NY, NY, 1990. Mackay, D., and A.T.K. Yuen, Mass transfer coefficients correlations for volatilization of organic solutes from water. Environ. Sci. and TechnoL, vol. 17(4), pp. 211­233, 1983. Marley, M.C., F. Li, and S. Magee, The application of a 3­D model in the design of air sparging systems, in Proc. of the 1992 Conference on Petroleum Hydrocarbons and Organic Chemicals in Groundwater: Prevention, Detection and Restoration, NWWA, Houston, Texas, pp. 377­392, 1992. Matter­MuUer, C., W. Guer, and W. Giger, Transfer of volatile substances from water to the atmosphere. Water Research, vol. 15, pp. 1271­1279, 1981. Mendoza, C., and E.O. Frind, Advective­dispersive transport of dense organic vapors in the unsaturated zone: I. Model development. Water Research, vol. 23(3), pp. 379­387, 1992. Middleton, A.C., and D. Miller, In situ aeration of groundwater ­ a technology overview, in Proceedings of the Conference on Prevention and Treatment of Soil and Groundwater Contamination in the Petroleum Refining and Distribution Industry, Montreal, Canada, October 16­17, 1990. Miller, C.T., M.M. Poirier­McNeill, and A.S. Mayer, Dissolution of trapped nonaqueous phase Uquids: mass transfer characteristics, Water Resources Research, vol. 26( 11), pp. 2783­2796, 1990. Mohr, D.H., Mass transfer concepts applied to in siw air spai­ging, Paper presented at the In Situ and On­Site Bioremediation Third International Symposium, San Diego, California, April 24­27, 1995. Munz, C., and P.V. Roberts, Gas­liquid phase mass transfer resistances of organic compounds during mechanical aeration. Water Research, vol. 23(5), pp. 589­601, 1989. 38 Nadim, F., A. Nadim, G.E. Hoag, and A.M. Dahmani, Desorpiion rate limitation in the extraction of organic molecules from unsaturated soils during soil venting operations. Journal of Contaminant Hydrology, vol. 25, pp. 21­37, 1997. O'Connor, D.L., and W.E. Dobbins, The mechanism of reaeration in natural streams, J. Sanit. Engrg. Div., ASCE, vol. 82(6), Paper No. 215, 1956. Ong, S.K., and L.W. Lion, Effects of soil properties and moismre on the sorption of TCE vapor. Water Research, vol. 25(1), pp. 29­36,1991. Ostendorf, D.W., E.E. Moyer, and E.S. Hinlein, Petroleum hydrocarbon sparging from intact core sleeve samples, in Proc. of the 1993 Conference on Petroleum Hydrocarbons and Organic Chemicals in Groundwater: Prevention, Detection and Restoration, NWWA, Houston, Texas, pp. 415­427, 1993. Pavlostathis, S.G., and G.M. Mathavan, Desorption kinetics of selected organic compounds from field contaminated soils. Environ. Sci. and Tech., vol. 26(3), pp. 532­538, 1992. Perry, R. H. and C. H. Chilton, Chemical Engineer's Handbook, 5'*' Edition, pp. 3.231­ 3.234, McGraw­ffill, NY, NY, 1978. Piwoni, M.D., and P. Baneijee, Sorption of volatile organic solvents from aqueous solution onto subsurface solids. Journal of Contaminant Hydrology, vol. 4, pp. 169­173, 1989. Powers, S.E., L.M. Abriola, and W.J. Weber Jr., An experimental investigation of nonaqueous phase liquid dissolution in saturated subsurface systems: steady state mass transfer rates. Water Resources Research, vol. 28(10), pp. 2691­2705, 1992. Powers, S.E., L.M. Abriola, and W.J. Weber Jr., An experimental investigation of nonaqueous phase liquid dissolution in saturated subsurface systems: transient mass transfer rates. Water Resources Research, vol. 30(2), pp. 321­332, 1994. Plummer, C. R., J.D. Nelson, and G.S. Zumwalt, Horizontal and vertical well comparison for in sim air sparging, GWMR, vol. 17(1), pp. 91­96, 1997. Raymond, R.L, Reclamation of Hydrocarbon Contaminated Waters, U.S. Patent Office 3,846,290, 1974. 39 Reddy, K.R., S. Kosgi, and J. Zhou, A review of in­situ sparging for the remediation of VOC­contaminated saturated soils and groundwater. Hazardous Wastes & Hazardous Materials, vol. 12(2), pp. 97­117, 1995. Rodriguez­Maroto, J.M., C. Gomez­Lahoz, and D.J. Wilson, Soil cleanup by in­situ aeration. XVIII. Field­scale models with diffusion from clay structures. Separation Science and Technology, vol. 29(11), pp. 1367­1399, 1994. Sellers, K., and R. Schreiber, Air sparging model for predicting groundwater cleanup rate, in Proc. of the 1992 Conference on Petroleum Hydrocarbons and Organic Chemicals in Groundwater: Prevention, Detection and Restoration, NWWA, Houston, Texas, pp. 365­ 376, 1992. Sleep, B.E., and J.F. Sykes, Modeling the transport of volatile organics in variable saturated media. Water Resources Research, vol. 25(1), pp. 81­92, 1989. Sleep, B.E., and J.F. Sykes, Compositional simulation of groundwater contamination by organic compounds. 1. Model development and verification. Water Resources Research, vol. 29(6), pp. 1697­1708, 1993a. Sleep, B.E., and J.F. Sykes, Compositional simulation of groundwater contamination by organic compounds. 2. Model applications. Water Resources Research, vol. 29(6), pp. 1709­ 1718, 1993b. Suthersan, S.S., Remediation Engineering: Design Concepts, pp. 91­121, CRC Press, Inc., Boca Raton, FL, 1997. Szatkowski, A., P.T. Imhoff, and C.T. Miller, Development of a correlation for aqueous­ vapor phase mass transfer in porous media, Journal of Contaminant Hydrology, vol. 18, pp. 85­106, 1995. Theofanous, T.G., R.N. Houze, and L.K. Brumfield, Turbulent mass transfer at free, gas­ liquid interface, with applications to open­channel, bubble and jet flows. Int. J. of Heat and Mass Transfer, vol. 19, pp. 613­624, 1976. Treybal, R.E., Mass Transfer Operations, pp. 29­31. McGraw­Hill Kogakusha Ltd., 2"'' Edition, Tokyo, Japan, 1968. 40 Turek, F., and R. Lange, Mass transfer in trickle­bed reactors at low Reynolds number. Chem. Eng., vol. 36, pp. 569­579, 1981. U.S. EPA, Underground storage tanks: technical requirements. Federal Register 53:37082, Sept. 23, 1988. Weber, W.J. Jr., P.M. McGinley, and L.E. Katz. Sorption phenomena in subsurface systems: concepts, models and effects on contaminant fate and transport. Water Research, vol. 25(5), pp. 499­528, 1991. Wilkins, M.D., L.M. Abriola, and K.D. Pennel, An experimental investigation of rate­ limited nonaqueous phase liquid volatilization in unsaturated porous media: steady state mass transfer. Water Resources Research, vol. 31(9), pp. 2159­2172, 1995. 41 CHAPTER THREE. AIR SPARGING EFFECTIVENESS: THE AIR CHANNEL MASS TRANSFER ZONE A paper submitted to Water Resources Research Washington J. Braida and Say Kee Ong Abstract Air sparging is one of the many innovative technologies being developed for the remediation of contaminated groundwater. Using a single air channel apparatus, mass transfer of VOC were found to occur within a thin layer of saturated porous media next to the air channel. In this zone, the VOCs were found to rapidly deplete during air sparging resulting in a steep concentration gradient while the VOC concentration outside the zone remained fairly constant. The size of this mass transfer zone was between 17 mm and 41 mm or 70 dpso and 215 dpso {dpso = mean particle size). The size of the mass transfer zone was found to be proportional to the aqueous diffusivity of the VOC, the mean particle size, and the uniformity coefficient. A general correlation predicting the size of the mass transfer zone was developed. The correlation incorporated the properties of the porous media, the air phase Peclet number, and the Pore Diffusion Modulus. The model was developed using data from eight different VOCs. The tailing effect of the air phase concentration and the rebound in the VOC concentration after the sparging system is turned off are some of the consequences of the existence of this mass transfer zone. Introduction The U.S. Environmental Protection Agency (EPA) has estimated that about 25% of the two million underground storage tank (UST) systems located at 700,000 facilities may be leaking (U.S. EPA, 1988). The release of volatile organic compounds (VOCs) from UST systems has a significant environmental impact on groundwater resources and may pose a risk to human health. VOCs commonly found in contaminated aquifers include aromatic 42 hydrocarbons such as benzene and xylenes, and chlorinated hydrocarbons such as trichloroethylene. A remedial approach for VOC­contaminated aquifer is the in situ air sparging technology. Air sparging involves the injection of contaminant­free air below the water table. The air flow through the aquifer results in the volatilization of VOCs from the aqueous phase. At the same time, oxygen is transferred from air to the contaminated groundwater which in turn may promote the biodegradation of VOCs (Brown et al., 1994). Contaminated air is then removed by an extraction well in the unsaturated zone. Even though air sparging has been successfully applied at several contaminated sites, Johnson et al. (1993) pointed out that the mechanism of air flow in saturated porous media and the physical­chemical processes involved during air sparging operations are not well understood. The injection of air into the aquifer creates complex transient physical­chemical conditions within the subsurface environment. The exact nature of air flow in a saturated porous media is not completely understood and research work in this area is fairly limited. In an air flow visualization study using glass beads as a porous media, Ji et al. (1993) found that in medium to fine grained water saturated porous media, air flows in discrete pore­scale channels. With a vertical air sparging well, the network of air channels formed may be visualized as the roots of a tree (Drucker and Di Julio, 1996). Based on this pattern of air channels network, many remedial engineers/scientists have used the radius of influence (ROD of the sparging well in the design of air sparging systems despite the lack of agreement among researchers and engineers on the definition and field estimation of the ROI (McCray and Falta, 1997). Even if the ROI is an important design parameter, the ROI basically provides a macroscopic estimate of the volume of the aquifer impacted by the well. However, the contaminated water impacted by the air channels is that which are nearest to the air­water interface. Volatilization of VOCs at the air­water interface will result in a VOC concentration gradient in the aqueous phase causing VOCs in the bulk water to diffuse to the interface (see Figure 1). If the rate of transport by liquid diffusion is slower than the volatilization rate of the VOCs at the interface a situation would arise in which the actual volume of the contaminated aquifer impacted by the sparging well would be less than ±e 43 total volume defined by the ROI. Therefore, within the volume as defined by the ROI, it is probable that a portion of this volume would be directly impacted by the air channels while a portion of the volume would not be impacted by the air channels. Ji (1994) proposed in a theoretical study that volatilization occurred by the diffusion of VOCs through saturated porous media to a cylindrical air channel. In his modeling efforts, he showed that a steep aqueous concentration profile was formed close to the air­water interface. To model mass transfer of VOCs during air sparging, Hein et al. (1994) proposed an arbitrarily defined cylindrical boundary away from the air channel where the VOC aqueous concentration remained constant. Based on this selected boundary condition, Hein and coworkers showed that the air flow rate and the aqueous diffusion rate were the main parameters controlling VOC volatilization. To model air sparging in soil columns, Drucker and Di Julio (1996) assumed that the air channels in the soil column can be represented by a composite of evenly spaced cyUndrical air channels. Each of the air channels was surrounded by a nonadvective aqueous region. Sensitivity analysis of the proposed model showed that the time needed to achieve 90% removal of the initial mass of VOC was inversely proportional to the diameter of the air channels but was directly proportional to the aqueous diffusivity of the VOCs. Using a one­dimensional model and results from soil column experiments, Chao (1997) estimated that the volume impacted by the air channels (called the "mass transfer zone") was as low as 15% to as much as 40% of the total air sparged volume depending on the air flow rate and porous media used. The work of the researchers cited above showed that aqueous diffusion is an important controlling mechanism during air sparging and the differences in the volatilization of various VOCs may be due to the distinctive "mass transfer zone" which may be a limiting factor for the proper operation of air sparging systems. If the size of the mass transfer zone is small compared to the distance between two air channels, remediation times will increase dramatically due to diffusion transport limitations. The objective of this work was to experimentally confirm, for dissolved contaminant plumes, the existence of a mass Uransfer zone surrounding the air channels and to study the influence of air sparging conditions and the physical­chemical properties of VOCs and 44 porous media on the size of the mass transfer zone. Experiments were conducted using a single air­channel experimental setup. An estimate of the size of the air channel mass transfer zone will provide further insights into the physical­chemical phenomena controlling the VOCs volatilization during air sparging. Materials and Methods To investigate and quantify the mass uransfer zone associated with air channels during air sparging, an experimental apparatus as shown in Figure 2 was built. The experimental setup consisted of an air channel of approximately 1.58 mm above the saturated porous media. The size of the air charmel was within the range of the sizes of air channels observed by others (Ji, 1994). The single­air channel apparams removed some of the complexities associated with using a typical air sparging system such as a soil column and may be used to assess the major mass transfer processes Umiting the volatilization of VOCs under controlled conditions. The apparatus was made of thick acrylic sheets with dimensions 17.5 cm long, 5 cm wide, and 11 cm depth. The apparatus was covered with a flat acrylic piece which provided a gap of approximately 1.58 mm (1/16") for the circulation of humidified air. In­house compressed air was used as the air source. The air was filtered to remove particulates and oil droplets, and humidified before being introduced into the experimental apparatus. Air flow was measured with a Gilmont model 11 flowmeter (Barrington, IL). Fifteen sample points as shown in Figure 2 were included to allow water samples to be collected throughout the porous media profile. A sample point to measure the VOC concentration in the effluent gas was included. Since the air flow rates under air sparging conditions are much higher than the water flow rate in an aquifer, stagnant water conditions were used for the experimental runs. The experimental runs were conducted at room temperature (21°C ± 2°C). Three different types of sand, graded Ottawa sand, sand 30/50, and sand 70/100 from U.S. Silica Company (Ottawa, IL), were used in the study. The properties of the porous media are summarized in Table 1. Specific surface areas were measured using the ethylene glycol monoethyl ether (EGME) procedure (Chihacek and Bremmer, 1979) and organic carbon 45 content was determined using the Walkley­Black procedure (Nelson and Sommers, 1982). Eleven VOCs with saturated solubilities ranging from 30 mg/L to 1800 mg/L and Henry's law constants ranging from 0.07 to 0.37 were studied. A summary of the physical­chemical properties of the VOCs is shown in Table 2. The aqueous solutions of various VOC concentrations were prepared from HPLC grade chemicals purchased from Sigma­Aldrich Chemical Company, Inc. (Milwaukee, WI). Concentrations of the VOCs used in the experiments ranged from 8 mg/L to 150 mg/L. A slurry was made by carefully mixing the porous media with the aqueous solution of the VOCs. The reactor was then packed layer by layer with the slurry to avoid entrapment of air bubbles and immediately sealed to minimize loss of VOCs. For each type of porous medium, four different air velocities: 0.2, 0.5, 1.1, and 2.5 cm/s were used. Values of the velocity of air in the air channels during air sparging operations have not been reported in the literature. The measurement of this value requires the knowledge of the injected air flow, the ROI, and the air saturation. The last two values are very difficult to determine in the field and may be the reason of the lack of information about air velocities in field applications. The range of air velocities selected for this study is consistent with typical field vapor extraction rates in which air pore velocities are generally expected to be less than 2 cm/s (Baehr et al., 1989) The VOC concentration in the air phase and in the liquid phase were measured with a Hewlett­Packard 5890 Series II gas chromatograph (Avondale, PA) equipped with a HP­5 capillary column and a flame ionization detector. Air phase concentration was determined by direct injection of 1 ml sample while liquid phase concentration was measured using the head­space technique. For the head space technique, 25 p.1 of an aqueous sample was placed in a 1.8 ml aluminum crimp cap vial and the aqueous concentration was estimated from the measured head space concentration after equilibrium was reached. Results and Discussion Mass Transfer Zone A typical set of results showing the change in the VOC concentrations in die exhaust air over time is presented in Figure 3. The corresponding changes in the aqueous concentrations 46 for various distances away from the air­water interface are shown in Figures 4 and 5. The results presented are for the center row of sampling points on the experimental setup. The air velocity for Figures 3,4, and 5 was 2.5 cm/s. The change in the concentration of VOCs in the exhaust air, presented in Figure 3, typically represents the behavior of the effluent air concentration for field­scale air sparging systems with an initial rapid decrease in the VOC concentration followed by a slower change in the VOC concentration and a finally with the VOC concentration remaining fairly constant. The asymptotic VOC concentration in the exhaust air was reached between 2 and 3 hours after the start of the sparging, implying that a quasi­steady condition for the volatilization of the VOCs was reached for all the experimental runs. Discussion on the relative volatility of each individual VOC will not be presented since our main focus is on the existence of the mass transfer zone and the effects of physical­ chemical properties of the soil and VOC on the mass transfer zone. Measurement of aqueous VOC concentrations indicated that there was a concentration gradient at the start of the experiments as a result of losses of VOCs during the packing of the reactor (Figure 4 and 5). Separate experiments indicated that when the apparatus was left alone for 18 hours, VOC losses, other than losses due by the initial packing, were less than 2%. Experimental data showed that during air sparging the VOC aqueous concentration was rapidly depleted within a thin layer of porous media next to the air channel. The rapid depletion is the result of a faster volatilization of VOCs at the air­water interface than the diffusive transport of the VOCs to the air­water interface (Figiires 4 and 5). After 4 hours of sparging the concentration gradient for each VOC became fairly constant suggesting that a quasi­steady state condition for the diffusion of the VOC through the porous media was reached. After quasi­steady state conditions were reached, a distinctive zone with a steep concentration gradient was found in the porous media. In all experiments, the 8­hour concentration profiles of the aqueous phase corresponded to the fairly constant VOC concentration in the air phase. Therefore, the aqueous concentration profiles strongly suggest that the volatilization of VOCs during air sparging may be controlled by the aqueous diffusion of VOCs through the porous media to the air channels. The aqueous concentration 47 profile showed that the air channels have a mass transfer zone beyond which the effects of the air flow were strongly reduced. The size of the mass transfer zone appeared to be somewhat dependent on the type of VOC (see Figures 4 and 5), and the porous media (Figure 6) but marginally dependent on the air velocity (Figure 7). For convenience, the size of the mass transfer zone was assumed to be the distance from the air­water interface to where the VOC concentration was 90% of the bulk VOC concentration. The size of the zone for the various experimental conditions was estimated to be between 17 mm and 41 mm. For the experiments using Ottawa sand, which has the largest uniformity coefficient, the size of the mass transfer zone was between 22 mm (115 dpso) and 40 mm (210 dpso) depending on the contaminant and the air flow rate. For sand 30/50, with an average grain size of 0.305 ram and an uniformity coefficient of 1.41, the size of the mass transfer zone ranged between 22 mm (70 dpso) and 41 mm (130 dpso)­ For sand 70/100, the size of the mass transfer zone was between 17 mm (100 dpso) to 36 mm (215 dpso)­ Unformnately, examination of these values cannot direcdy show the influence of the mean particle size and uniformity coefficient of the porous media on the size of the mass transfer zone. Other parameters which may have a potential influence on the size of the mass transfer zone included the aqueous and air diffusivities of the VOCs, the porosity of the porous media, and the Henry's law constant of the VOCs. A tool which may be used to assess the influence of the various physical­chemical factors on the size of the mass uransfer zone is the dimensionless number modeling approach. Model Correlation of Mass Transfer Zone To characterize the influence of different physical­chemical parameters on the size of the mass transfer zone, a regression analysis of various dimensionless numbers which incorporated the size of the mass transfer zone and the various physical­chemical parameters of the VOCs and the porous media was conducted. The regression analysis had two objectives: (a) to determine which parameters most strongly affect the size of the air channel mass transfer zone, and (b) to generate a model which may predict the size of the mass transfer zone for different conditions. 48 The dimensionless numbers used in the regression analysis are presented in Table 3 and they were selected because they included most of physical­chemical parameters which potentially have an effect on the size of the mass transfer zone. The dimensionless numbers to be modeled included the air phase Peclet number {Pe), which is the ratio of the rate of transport by advection and the rate of transport by diffusion in the air phase, the Pore Diffusion Modulus {Ed), which is the ratio between the mass transport by radial diffusion in the aqueous phase and the mass transport by advection in the gas phase, and the Henry's law constant {KH)- The Pe numbers used in the determination of this correlation ranged from 0.052 to 1.523. The width of the mass transfer zone (AfTZ) was included in the Pore Diffusion Modulus {Ed). Properties of the porous media used in the dimensionless analysis included the porosity (£), the uniformity coefficient (C/C), and the dimensionless average particle size do = dpso /dm as used by Wilkins et al. (1995). The term dm is the mean grain size of a "medium" sand as defined by the USDA and is equal to 0.05 cm (Driscoll, 1986). A multiple, stepwise regression analysis was conducted to determine the best fit for Ed with the other dimensionless numbers using a log linearized correlation as shown below. The regression analysis was conducted for 8 of the 10 VOCs. The 1,2 dichlorobenzene and 1,2,4 trichlorobenzene were reserved for verification. log(Ed) = Po + PilogiPe) + PzlogiUC) ­^pslogi£) + j3jog( KH ) + fislog(do) (1) The stepwise regression procedure evaluated the least square residual (/^) and the F statistical parameter of each predictor to determine the most appropriate model parameters. The regression analysis was performed using the statistical software SAS 6.10 (SAS Instimte Inc., 1993). The variables in the model which were found to be significant at the F = 0.15 level are presented in equation (2). Ed = 10'^ ^° Pe UC do°­^^ r^= 0.9032 (2) 49 Both the porosity of the media and the Henry's law constant for the VOCs were not selected to be included in the empirical model. A summary of the stepwise analysis is shown in Table 4. The Peclet number {Pe) explained most of the variation of the Pore Diffusion Modulus i^Ed) followed by the uniformity coefficient (C/C) and the dimensionless mean particle size (jdo)­ The lack of correlation of the Henry's law constant with the Ed (F< 0.15) in the empirical model seemed to suggest that the size of the MTZ does not depend of the volatility of the VOC and this strongly implies that air sparging is a diffusion controlled process. The experimental and predicted Pore Diffusion Modulus are plotted as shown in Figure 8 along with the 95% confidence intervals. To test the validity of the model, the correlation (equation 2) was used to predict the Ed values of 1,2 dichlorobenzene and 1,2,4 trichlorobenzene which were not included in the regression analysis. Results of the predicted values are shown in Figure 9. Despite the differences in solubility and Henry's law constant values especially for 1,2,4 trichlorobenzene, the correlation was found to predict the size of the mass transfer zone for these two organic compounds. A direct comparison of the effects of the physical­chemical parameters of the system on the size of the mass transfer zone can be made by expressing equation (2) in terms of each individual physical­chemical property: 0.87 MTZ =10 0.05 air D"' (3) Examination of equation (3) revealed that the size of the mass transfer zone was directly proportional to the square root of the aqueous diffusivity of the VOC but was inversely related with the air diffusivity of the VOC. The mean particle size and the uniformity coefficient of the porous media have a direct impact on the size of the mass transfer zone. Intuitively, equation (3) is correct since the size of the mass transfer zone will be affected by the aqueous phase diffusivity, the size of the particle present, and how well disUibuted was the porous media. The velocity of the sparged air was included in the Pore Diffusion 50 Modulus and in the air phase Peclet number but its influence on the size of the mass transfer zone was marginal as indicated by the low value of its exponent. A one order of magnitude change in the air velocity (0.2 to 2.5 cm/s) for the experiments conducted resulted in only a 14% change of the size of the mass transfer zone. The existence of a mass transfer zone surrounding the air channels during air sparging operations implies that VOC removal by volatilization during air sparging is a diffusion limited process. An important observation which can be drawn from this study is that if the distance between two air channels is larger than twice the size of the mass transfer zone then portions of the aquifer within the ROI would not be affected by the air flow, i.e. portions of the air sparged volume would not be remediated. This conclusion is in agreement with the results reported by Ahlfeld et al. (1994). These authors, using a simple diffusion model, showed that decreasing the average channel spacing from 480 to 20 mm resulted in a three order of magnitude reduction in the cleanup time. The existence of this mass transfer zone explains the long tailing effect in the VOC concentration in the gas phase typically seen after the aquifer is sparged for sometime and the rebound in aqueous VOC concentration after the sparging system is turned off. In addition, work done by Bass and Brown (1997) indicated that even when the VOCs were shown to be removed in the gas stream, measurements of VOC concentrations in the soil core seemed to be statistically unchanged. By quantifying and assessing the various physical­chemical parameters affecting the size of the mass transfer zone along with an understanding of distribution of air channels in the aquifer, more accurate air sparging models may be developed to predict the performance of air sparging systems under various operation conditions. Conclusions A mass transfer zone was shown to be associated with air channels during air sparging operations. In the mass transfer zone, a steep VOC concentration gradient was found to form after several hours of air sparging. The size of this zone ranged between 17 mm and 41 mm or between 70 and 215 dpso. The presence of the mass transfer zone strongly suggests that the volatilization of VOCs by air sparging is a diffusion limited process. 51 An empirical model using the Pore Diffusion Modulus (Ed) which included the size of the mass transfer zone was found to correlate well with the air phase Peclet number (f e), the uniformity coefficient {UO, and the dimensionless mean particle size (^do). Based on the correlation, the size of the mass transfer zone was found to be proportional to the aqueous diffiasivity of the VOC, the unifomtiity coefficient, and the mean particle size of the porous media. Air velocity had a marginal effect on the size of the mass transfer zone under the experimental conditions tested. The existence of the mass transfer zone under air sparging conditions implies that for remediation to be successful, the air channels during air sparging must be as close as possible with the mass transfer zones of the two adjacent air channels overlapping each other. In other words, the larger the mass transfer zone, the higher are the chances that air sparging will be effective. For small mass transfer zones associated for example with tight aquifer materials, a larger volume of contaminated soil within the radius of influence of the well will remain unaffected by the air flow and remediation will take a longer time. Notation Ca/r Cwaier VOC concentration in the air and aqueous phases, ML'^ VOC aqueous diffusivity, L'T'' Da VOC air diffusivity, L'T'' do dimensionless mean particle size dpso mean particle size, L Ed Pore Diffusion Modulus, dimensionless e porous media porosity, dimensionless K„ Henry's law constant, dimensionless Ko.. octanol­water partition coefficient, dimensionless MTZ width of the mass transfer zone, L Pe air phase Peclet number, dimensionless UC uniformity coefficient, dimensionless ^air air velocity ( air flow rate/air channel cross sectional area), LT 52 Acknowledgments The authors would like to thank Juan Jose Goyeneche for his technical assistance in the statistical analysis of the data. References Ahfeld, D., A. Dahamani, and W. Ji, A conceptual model for field behavior of air sparging and its implications for application, GWMR, vol 14(4), pp. 132­139, 1994. Baehr, A.L., G.E. Hoag, and C. Marley, Removing volatile contaminants from the unsaturated zone by inducing advective air phase transport. Journal of Contaminant Hydrology., vol. 4, pp. 1­26, 1989. Bass, H.D., and R.A. Brown, Performance of air sparging systems a review of case smdies, in In Situ and On­Site Bioremediation: Volume /, edited by B. C. AJleman and A. Leeson, pp. 117­122, Battelle Press, Columbus, OH, 1997. Brown, R.A., R.J. Hicks, and P.M. Hicks, Use of air sparging for in situ bioremediation, in Air Sparging for Site Remediation, edited by R.E. Hinchee, pp. 38­55, Lewis Publishers, Boca Raton, FL, 1994. Chao K.P., Aqueous­vapor mass transfer of VOCs in saturated porous media under air sparging conditions, Ph.D. Dissertation, 176 pp., Polytechnic University, Brooklyn, NY, 1997. Chihacek L.J., and J.M. Bremmer, A simplified ethylene glycol monoethyl ether procedure for assessment of soil surface area. Soil Sci. Soc. Am. J., vol. 43, pp. 821­822, 1979. Driscoll, F.G., Groundwater and Wells, 2"^^ Edition, Johnson Filtration Systems, St. Paul, Mirm., 1986. Drucker A.S., and S.S. Di Julio, Groundwater clean up by In Situ air sparging: development of a model and application to satiu­ated soil column experiments, in Proceedings of the Water Environmental Federation 69 Th. Annual Conference & Exposition, Dallas, TX, October 5­9, 1996. 53 Hein, G.L.. N.J. Hutzler, and J.S. Gierke, Quantification of the mechanisms controlling the removal rate of volatile contaminants by air sparging, in Proc. of the 1994 National conference on Environmental Engineering, edited by J.N. Ryan and M. Edwards, pp. 556­ 563, ASCE, NY, NY, 1994. Ji, W., A. Dahamani, D.P. Ahlfeld, J. D. Lin, and E. Hill, Laboratory study of air sparging; Air flow visualization, GWMR, vol. 13(4), pp. 115­126, 1993. Ji, W., Air sparging: Experimental and theoretical analysis of flow and numerical modeling of mass transfer, Ph.D. Dissertation, 154 pp.. The University of Connecticut, Storrs, CT, 1994. Johnson, R.L., P.C. Johnson, D.B. McWorther, R.E. Hinchee, and I. Goodman, An overview of in situ air sparging, GWMR, vol. 13(4), pp. 127­135, 1993. Lyman, W.J., W.F. Reehl, and D.H. Rosemblatt, Handbook of Chemical Properties Estimation Methods, American Chemical Society, Washington D.C., 1990. McCray, J.E. and R.W. Falta. Numerical simulation of air sparging for remediation of NAPL contamination. Ground Water, vol. 35(1), pp. 99­110, 1997. Nelson, D.W., and L.E. Sommers, Total carbon, organic carbon, and organic matter, in Methods of Soil Analysis, Part 2, Chemical and Microbiological Properties, Second Edition, edited by A.L. Page, pp. 570­571, American Society of Agronomy, Inc., Soil Science Society of America, Inc., Madison, WI, 1982. Perry, R.H. and C.H. Chilton, Chemical Engineer's Handbook, 5"' Edition, pp. 3­231­ 3.234, McGraw­Hill, NY, NY, 1978. Treybal, R.E., Mass Tranter Operations, 2"*^ Edition, pp. 29­31, McGraw­Hill Kogakusha Ltd., Tokyo, Japan, 1968. SAS Institute Inc., 5A5® Software: Changes and Enhancements. Release 6.10, Carey, NC. 1993. U.S. EPA, Underground storage tanks: technical requirements. Federal Register 53:37082, Sept., 1988. 54 Wilkins, M.D., L.M. Abriola, and K.D. Pennel, An experimental investigation of rate­ limited nonaqueous volatilization in unsaturated porous media: steady state mass uransfer. Water Resources Research, vol. 31(9), pp. 2159­2172, 1995. 55 Table I. Physical­chemical properties of porous media Type of Sand Organic Carbon Porosity 2.16 Specific Surface Area (m­/g) 1.99 0.377 0.0066 0.0305 1.41 1.17 0.370 0.0062 0.0168 1.64 2.73 0.400 0.0063 Uniformity Coefficient Ottawa Sand Mean Particle Size (cm) 0.0190 Sand 30/50 Sand 70/100 (%) Table 2. Physical­chemical properties of VOCs at 20°C (Lyman et al., 1990) compound Molecular Solubility Da DwXlO^ KH Weight (cm"/s) (mg/1) (cm"/s) * Benzene 78.12 0.195 1780 9.59 0.0923 log Kow 2.12 Toluene 92.15 0.233 515 0.0830 8.46 2.73 Ethylbenzene 106.18 0.291 152 0.0732 7.63 3.15 o­Xylene 106.18 0.178 130 0.0759 7.63 2.95 m­Xylene 106.18 0.247 175 0.0759 7.63 1.38 p­Xylene 106.18 0.256 198 0.0759 7.63 3.26 Chlorobenzene 112.56 0.137 500 0.0725 8.52 2.84 n Propylbenzene 120.21 0.369 60 0.0544 6.99 3.87 1,2 Dichlorobenzene 147.00 0.118 145+ 0.0829 7.97 3.60 1,2,4 Trichlorobenzene Styrene 181.44 0.069 30+ 0.0686 7.09 4.30 104.16 0.0967 300 0.0746 7.89 2.95 Estimated firom Hirschfelder, Bird, and Spotz Estimated from Wilke and Chang correlation (Treybal,1968) at 25°C 56 Table 3. Dimensionless numbers used for modeling Comments Dimensionless Number Equation Mass transport bv radial aqueous diffusion Pore Diffusion Modulus DJPso Mass transport by gas advection (Ed) Air Phase Peclet Number (Pe) Dimensionless Mean Grain Size (do) Porosity (£) Henry's Law Constant KH VaJPsO Da dPso Rate of transport bv gas advection Rate of transport by molecular gas diffusion VTotal ­Vsolid V^ Total Void volume/total volume c^war^r Uniformity Coefficient (UC) dni= 0.05 cm is the mean grain size of a "medium" size sand Ratio of air phase concentration to aqueous phase concentration Measured of the grain size distribution Table 4. Summary of stepwise regression analysis Variable Parameter Estimate Standard Partial Error r" Intercept ­7.6972 0.0572 Model r Pe ­1.1014 0.0388 0.7526 0.7526 UC ­1.7335 0.2573 0.1396 0.8895 do 0.6549 0.1814 0.0137 0.9032 Unaffecled Aquifer Air Compressor Material Mass Transfer Zone ^ \L Ground Level l)in\is|on of VOCsjio Air Channels 1 Water Table Aquifer MTZ Air Channels M I'Z: Mass Transfer Zone Impacleil by Air I'low I'igure 1 Conceptual sketch of air channels and mass transfer zone Sample Ports Air Channel Gap 1.58 mm Clean Humidified Air Contaminated Air to Gas Sampling Point Saturated Porous Media Contaminated with VOCs l­igurc 2. Singlc­air channcI apparatus 59 0.75 0.75 J 0.50 0.50 ­ 'ai) E w o c m,p­Xylenes I I I 11 I I i I I I I I 0 100 200 300 U O > Ethylbenzene / 0.00 400 o­Xylene 0.25 ­ I I I i­1 I I I t I 1 1 I i"° I 0 500 100 Time (min) 200 I I9I I^ 300 400 500 Time (min) 0.75 0.75 2 0.50 ­ ^ 0.50 s S u s e I... Toluene U O > Styrene Chlorobenzene n­Propylbenzene 0.25 ­ "O O 0.00 "Tr r^Ti I'i~j" I "I"! "i j I i I t I i J r I 0 100 200 300 Time (min) 400 500 0.00 ^ i I I r I » I » 0 100 ~ ^ ^ r J I > ; I I > > i 1 I I > I 1 200 300 400 Time (min) Figure 3. VOC concentrations in the air phase for Ottawa sand at an air flow rate of 2.5 cm/s 500 60 Air­Water Interface Air­Water Interface T=8h ­20 ­ ­g­ ­40 ­ £ j: s. T=Oh ­20 ­ T=Oh T=8h 'g ­40 ­ T=4h ­ ­60 ­ S­ ­60 ­ ­80 ­ ­80 ­ T=4 h Benzene Ethylbenzene ­100 ­100 50 0 150 100 0 VOC CoDc. (mg/L) 50 25 "^=8 h ­g­ S ­40 ­ ­ f. Q44 ­60 100 VOC Cone. (mg/L) Air­Water Interface ­20 75 Air­Water Interface T=Oh T=Oh ­20 ­ T=4 h ­40 ­ T=4 h JS ­80 ­ ­60 ­ ­80 ­ o­Xylene ­100 ­100 0 25 50 VOC CoDc. (mg/L) 75 100 0 25 50 VOC Cone. (mg/L) Figure 4. VOC concentration profiles for various times during air sparging (Ottawa sand, air flow rate 2.5 cm/s) 75 61 ir­Water Interface Air­Water Interface ­20 ­ T=8h ­20 ­ T=8 h T=4h ­40 ­ ­40 ­ T=0 h T=4 h T=0h £ ­60 ­ ­60 ­ ­80 ­ ­80 ­ Chiorobenzene Toluene ­100 ­100 0 10 20 30 40 50 0 10 VOC Cone. (mg/L) 30 20 40 VOC Cone. (mg/L) 0 Air­Water Interface T=8h Air­Water Interface T=0h ­20 T=8h T=0h T=4 h T=4h ? B A a ­40 H ­60 H a ­80 H Styrene ­100 VOC Cone. (mg/L) n­Propylbenzene I I I I VOC Cone. (mg/L) Figure 5. VOC concentration profiles for various times during air sparging (Ottawa sand, air flow rate 2.5 cm/s) 50 62 •. o. ­20 ­ Ottawa Sand Sand 70/100 Sand 30/50 ­40 ­ E E tia. a ­60 ­ ­80 ­ 90% of Bulk Concentration ­100 I I I r I I I I I I I I I I I I I I' I r I I i' I I I I 0.0 0.2 0.4 0.6 0.8 1.0 VOC Relative Concentration (C/CJ Figure 6. o­Xylene concentration profiles for different porous media (air flow rate 2.5 cm/s) ­20 ­ ^ E ­40 ­ g­ ­60 H E " ­80 ­ 0.2 cm/s 0.5 cm/s 1.1 cm/s 2.5 cm/s 90% of Bulk Concentration ­100 0.0 0.2 0.4 0.6 0.8 1.0 Relative VOC Concentration (C/CJ Figure 7. Benzene concentration profiles for different air flow rates (Ottawa sand) 63 1e­6 / / / t •a o s / / / / A° / / /° O / / / /o o / rf" / / an 3 1e­7 ­ *5 3 / o/ / //A /o°°/ / / o As im O O. ^/ °r7Vo 0 ^ /„ rr°v •5 1e­8 ­ V E *fcZ a ST U] / / / / / / / r—I—I I I IIII 1e­9 1e­9 1e­8 1 1—I—I 1 I 111 1 1—I I I I I r 1e­7 1e­6 Computed Pore Diffusion Modulus {Ed) Figure 8. Pore Diffusion Modulus: experimental vs. computed values with 95% confidence interval plots 64 1e­6 ­J I "i o ~^ s '5 .3 a u k o 1e­8 ­ eu s i* S *>Z a. w ie­9 / / _/ / y/ ° *• X* • / X • / T — I — I I I 1111 1e­9 ° 1e­8 1 * 1,2 Dichlorobenzene ° 1,2,4 Trichlorobenzene 1—r I I IIII 1 1—I r I I ri 1e­7 1e­6 Computed Pore Diffusioo Modulus (Ed) Figure 9. Comparison of experimental and computed Pore Diffusion Modulus for 1,2 Dichlorobenzene and 1,2,4 Trichlorobenzene with 95% confidence interval plots 65 CHAPTER FOUR. AIR­WATER MASS TRANSFER COEFnCIENTS A paper submitted to Water Resources Research Washington J. Braida and Say Kee Ong Abstract Experiments investigating the mass transfer of several volatile organic compounds (VOCs) across the air­water interface were conducted using a single­air channel air sparging system. Three different porous media were used in the study. Air velocities ranged from 0.2 cm/s to 2.5 cm/s. The tormosity factor for each porous media and the air­water mass transfer coefficients were estimated by fitting experimental data to a one­dimensional diffusion model. Tortuosity factors found were 0.52 for Ottawa sand, 0.51 for sand 30/50, and 0.47 for sand 70/100. The estimated mass transfer coefficients {Kg) ranged from 1.79 x 10"^ cm/min to 3.85 X 10'" cm/min. The estimated lumped mass transfer coefficients {Kg a) were found to be directly related to the air diffusivity of the VOC, air velocity, particle size, and inversely related to the Henry's law constant of the VOCs. Of the four parameters, the parameter which controlled or had a dominant effect on the lumped mass transfer coefficient was the air diffusivity of the VOC. Two empirical models were developed by correlating the Damkohler and the modified air phase Sherwood numbers with the air phase Peclet number, Henry's law constant, and the reduced mean particle size of the porous media. The correlation developed in this study may be used to obtain better predictions of mass transfer fluxes for field conditions. Introduction In situ air sparging has been used for more than 10 years with varying success for the remediation of aquifers contaminated with dissolved volatile organic compounds (VOCs) and nonaqueous phase liquids (NAPLs). In a typical air sparging system, clean air is injected below the water table to strip or volatilize VOCs from the aqueous phase. In addition, aerobic biodegradation is enhanced through the transfer of oxygen from air to the 66 contaminated groundwater (Brown et al., 1994). Air sparging is usually used in combination with a vacuum extraction system located in the unsaturated zone. Despite the widespread use of air sparging as a remedial technology, an understanding of the different operadve processes present during air sparging has lagged behind the practical application of the technology. Injection of air into the aquifer creates complex transient physical conditions within the subsurface environment. Laboratory air flow visualization studies showed that the prevalent air flow in fine to medium grained saturated porous media was in the form of discrete air channels (Ji et al., 1993). Under air sparging conditions, VOCs in the aqueous phase must travel by diffusion through the porous media to the air channels before being volatilized at the air­water interface. Several researchers have shown that for advective gas flow conditions such as air sparging and soil vapor extraction, the movement of air across the air­water interface may not provide enough residence time for local equilibrium to be achieved between the VOCs in the water and air phases (Mendoza and Friend, 1992; Armstrong et al., 1993). Nonequilibrium mass transfer relationships such as the first order mass transfer equation have been used to describe the mass transfer of VOCs in porous media (Sleep and Sykes, 1989; Szatkowski et al., 1995; Brusseau, 1991; Gierke et al., 1992; Armstrong et al., 1993; and Hein et al., 1994). The first order mass transfer equation is represented by: J = K^{K„C^.­C^)=K, ^ c C ^ ^ (1) where J represents the mass transfer flux across the air­water interface (ML"^"'), Ca and Cu are the gas and liquid phase concentrations (ML'^), respectively at the air­water interface, Kh is the Henry's law constant (dimensionless), and KG and KL are the gas phase and liquid phase overall mass transfer coefficients (LT'), respectively. Volatilization of VOCs across the air­water interface is a two­phase phenomenon driven by the difference in the chemical potential of the species in both phases. In equation 1, the driving force for the transfer of VOC across the air­water interface is given by the difference between the equilibrium air or water concentration {KhC^ or Co/Kh) and the actual air or liquid phase concentration (Ca or 67 Ch). Careful examination of equation 1 indicates that the value of the mass transfer coefficient will be determined by the liquid and gas phase concentrations selected to define the potential gradient. For a given mass flux, J, a different mass transfer coefficient value will be needed if the bulk concenUration of the VOC in each phase is used instead of the VOC concentrations at the air­water interface. Even though the VOC concenurations at the air­ water interface may more accurately represent the physics of the system, measurement of the VOC concentrations at the air­water interface is not possible. Therefore, the standing issue in the estimation of the mass transfer coefficients between gas and liquid phases during air sparging is to define in a practical way, the driving force (i.e., concentration gradient) for the volatilization of VOCs. Since an air sparged porous media is not a completely mixed system, the task of defining the concentration gradient is more challenging. The objective of this work was to investigate the air­water mass transfer process for a wide range of system parameters using an experimental setup where the interfacial area between the air and liquid phases was held constant. Mass transfer coefficients determined in the study were then correlated with various dimensionless numbers. The correlations developed in this study may be used to predict the mass transfer coefficients for different operating conditions. Materials and Methods The approach taken was to use an experimental setup which would simulate air flow in a single discrete air channel in saturated porous media. Figure 1 shows the experimental apparatus used for the investigation of air­water mass transfer processes during air sparging. The apparams consisted of a single air chaimel of approximately 1.58 mm located above the saturated porous media. Using a single air channel would provide "microscopic" information such as changes in the aqueous phase concentrations as a function of distance from the air­ water interface with time. The size of the air channel used was within the range of the sizes of air channels observed by others (10 to 20 pore size diameters) during air sparging (Ji, 1994). The apparatus was made of 6.35 mm (1/4") thick acrylic sheets and was 17.5 cm long, 5 cm wide, and 11 cm deep. The apparams was covered with a flat acrylic piece which provided a gap of approximately 1.58 mm (1/16") for the circulation of humidified air. 68 Three columns of five sample points each, as shown in Figure 1, were included to allow water samples to be collected throughout the porous media profile. A glass T connector in the exhaust line was used as a gas sampling point to measure the VOC concentration in the effluent gas. In­house compressed air was used as the air source. Prior to its introduction into the experimental setup, air was filtered to remove particulates and oil drops and then humidified. Air flow was measured with a Gilmont Model 11 flowmeter (Harrington, H­). Stagnant water conditions were used for all the experimental runs. Since the air flow rates under typical sparging conditions are much higher than the flow rates of water in an aquifer, use of a stagnant flow condition was reasonable within the experimental test period. The experimental runs were conducted at room temperature (21°C ± 2°C). To assess the influence of the physical­chemical properties of VOCs and the porous media on the rate of volatilization of VOCs, three different porous media, eleven VOCs, and four air flow rates were used. The three different types of sand used in the study were: (a) graded Ottawa sand, (b) sand 30/50, and (c) sand 70/100. All three sands were obtained from U.S. Silica Company (Ottawa, IL). The properties of the porous media are summarized in Table 1. Specific surface areas were measured using the ethylene glycol monoethyl ether (EGME) procedure (Chihacek and Bremmer, 1979) and organic carbon content was determined using the Walkley­Black procedure (Nelson and Sommers, 1982). The VOCs used had saturated solubilities ranging from 30 mg/L to 1800 mg/L and Henry's law constants ranging from 0.07 to 0.37. A summary of the physical­chemical properties of the VOCs is presented in Table 2. Partition coefficients of the VOCs for the three porous media were determined using the head space technique described by Garbarini and Lion (1985). Linear partition coefficients were found to range between 0.025 mL/g and 0.082 mL/g. Aqueous solutions of the different VOCs with concentrations ranging from 8 mg/L to 150 mg/L were prepared using HPLC grade chemicals. All chemicals were purchased from Sigma­Aldrich Chemical Company Inc. (Milwaukee, WI). A slurry was made by carefully mixing the porous media with the aqueous solution of the VOCs. The reactor was then rapidly packed layer by layer with the slurry to avoid entrapment of air bubbles and immediately sealed to minimize VOCs losses. Aqueous samples were then taken from the various sampling points to determine the 69 initial aqueous phase concentration in the porous media. Air was then circulated through the air channel. For each type of porous medium, four different air velocities, 0.2, 0.5, 1.1, and 2.5 cm/s, were used. There are not reported values of air velocity during field scale air sparging operations. This lack of information may be due to the difficulties associated with the determination of the ROI and air saturation in the field. These two values together with the air injection rate are needed to estimate the air pore velocity. The range of air velocities selected for this study was within the range of air velocities typically found for field applications of soil vapor extraction in which air pore velocities are expected to be less than 2 cm/s (Baehr, 1989). To check for possible losses through adsorption of VOCs onto the reactor walls or leaks in the experimental semp, an 18­hour long experiment without air flow was conducted. The reactor was filled as described above and the VOC concentrations across the porous media were measured at the start and end of the 18­hour experiment. VOC losses were found to be less than 2%. The VOC concentrations in the air phase and liquid phase were measured with a Hewlett­ Packard 5890 Series 11 gas chromatograph (Avondale, PA) equipped with a HP­5 capillary column and a flame ionization detector. Air phase concentration was determined by direct injection of 1 mL sample into the gas chromatograph. The liquid phase concentration was determined using the head­space technique where 25 |iL of an aqueous sample was placed in a 1.8 mL aluminum crimp cap vial. After equilibrium was reached, the head space concentration was measured by direct injection of a head space sample into the gas chromatograph. The aqueous concentration was then estimated from the measured head space concentration. Results and Discussion VOC Concentration Profiles and Mass Transfer Zone (MTZ) A typical set of results showing the change in benzene concentrations over time in the exhaust air and in the water phase is presented in Figure 2. Air flow rate was 1.1 cm/s. Concentration profiles for o±er VOCs were similar to that of Figure 2 and are not presented here. The results presented are for the center column of sampling points on the experimental 70 setup. Experimental measurements showed that liquid phase concentrations at 5 and 10 mm below the air­water interface were slightly lower in the first column of sampling points but slightly higher in the last column of sampling points. The aqueous phase concentration at larger depths, i.e., greater than 10 mm below the air­water interface were similar for all the sampling rows. Therefore, concentration profiles of center column generally represented the average value of the VOC aqueous concentration at any given depth. The change in the benzene concentration of the exhaust air is shown in Figure 2a. The results of Figure 2a typically represents the behavior of the exhaust air concentration for field­scale air sparging systems where the VOC concentrations in the exhaust air was characterized by a rapid decrease in VOC concentration followed by a fairly constant low VOC concentration. The asymptotic concentration in the exhaust air for the experimental apparatus was reached after 2 to 3.5 hours of sparging. This implies that a quasi­steady state condition was reached for the volatilization of the VOCs. A steep aqueous concentration gradient with very low benzene concentration at the air­water interface was evident as shown in Figure 2b. However, a short distance away from the interface, the benzene concentrations were fairly constant and did not change even after 8 hours of sparging. Braida and Ong (1997) have defined this concentration gradient zone as the mass transfer zone {MTZ). At the onset of air sparging, the majority of the VOC volatilized came from the MTZ adjacent to the air channel while the VOC in the bulk liquid several centimeters away from the air channel remained almost unaffected by the air flow. The development of the MTZ may be the result of a faster rate of VOC volatilization at the air­water interface as compared to the diffusive transport of VOCs through the porous media to the air­water interface. For all the experiments, the MTZ (i.e., the concentration gradient zone) for each VOC became fairly constant after four hours of sparging suggesting that a quasi­steady state condition was reached whereby the mass transfer to the air phase was controlled by ±e diffusion of die VOCs through the MTZ. The presence of a concentration gradient at the start of the experiments was due to VOC losses during the packing of the material into the reactor. 71 Estimation of Mass Transfer CoefHcients To describe the transport of VOCs throughout the porous media and their volatilization across the air­water interface, a one­dimensional diffusion model was used. Because of the very low organic carbon content and low partition coefficients of the porous media, sorption of VOCs was not considered in the model for the estimation of mass transfer coefficients. Assuming stagnant conditions for the aqueous phase, the equation describing diffusive u­ansport of VOCs through the porous media can be stated as follows: dC d'C e—^ = TD ­—(2) where z (L) is the depth of the porous media with the air­water interface located at z = 0 and the bottom of the reactor at z = L, the aqueous diffusivity of the VOC is the aqueous concentration of the VOC (ML'^), is f is the porosity of the porous media (dimensionless), and r is the tortuosity factor (dimensionless) which accounts for the change in the length of the diffusion path of the VOCs in the porous media. The initial and boundary conditions (IC and BC) for the experimental setup were: IC Cy, is known for all z at time zero (experimental data) BC ac rD^­^ = ­K^{K„C^­C^) dz dC ——^ = 0 dz atz = Oforallt (3) at z = L for all t (4) Based on equation 3, the driving force for the volatilization of VOCs was the difference in the chemical potential between the VOCs in the two phases. Since the VOC concentrations Ch and Ca at the interface were not known, estimating the mass transfer coefficients based on the VOC concentration at the interface would not be useful for field applications. According to the two­resistance model of Lewis and Whitman (1924) for mass transfer, two laminar sublayers exist at each side of the air­water interface where molecular diffusion controls the 72 transport of compounds. Beyond these thin laminar regions, turbulent conditions for the transport of chemical compounds were assumed to be prevalent. Using the two­resistance model, the concentration gradient will be based on the bulk concentrations of the chemicals in each phase. However, for air channels in porous media, the assumption of turbulent conditions in the aqueous phase for the transport of contaminants is not a valid assumption. The difference between the assumptions of the two­resistance model and the situation found in saturated porous media during air sparging operations is shown in Figure 3. Because of the concentration gradient in the vicinity of the air chaimels, bulk concentration in the aqueous phase cannot be used to define the driving force for the mass transfer. To overcome this difficulty, a volumetric weighted average concentration, C„., of the mass transfer zone (A/7Z) was proposed and used for the aqueous concentration term on the right hand side of equation 3. The volumetric average concentration, , was computed as follows: — =±—I!—22. TV C r C. ^ ("51 (5) where V„ is the volume of the n layer (L^), C^.„(ML'^) is the aqueous concentration of VOC in the n layer, and the summation extends to all the layers included in the MTZ. The reason for selecting this concentration was that the MTZ was the region direcdy impacted by the air channel but beyond the MTZ, air flow had negligible impact on the dissolved contaminants. Movement of contaminants from the bulk phase into the MTZ was by diffusion only as described by equation 2. To simplify the estimation of the mass transfer flux, VOCs volatilized from the air­water interface were assumed to be rapidly swept away firom the air­water interface. Therefore, the air concentration, Ca, at the air water interface may be assumed to be negligible. With these assumptions, the fost boundary condition (equation 3) can be rewritten as: rD,^ = ­K^K„C, (6) 73 The VOC concentration in the effluent air was computed using the u­anspon equation for the VOCs in the air phase: V.^ = ­Q.C.+K„AK„C, at (7) where Q is the measured VOC concentration in the effluent air (ML'^), Va represents the air volume (L^) in the experimental setup, Qa represents the air flow (L^T*), A is the air­water interfacial area (L") of the experimental setup (87.5 cm"). The objective of the modeling approach was to estimate the gas side overall mass transfer coefficient {Kc) by numerically solving equation 2 with boundary conditions defined by equations 4 and 6. Equation 2 was solved numerically using the Continuous System Modeling Program (CSMP) software (IBM, 1972). The value of KG determines the flux of VOCs at the air­water interface while r determines the final shape of the VOC concentration profile in the porous media. The VOC concentration profile was divided into 26 horizontal layers of variable thickness (thinner layers were used closer to the air­water interface to compute more accurately the VOC concentrations in this region), x is a property of the porous media and was estimated by fitting the computed VOC concentration profile to the experimentally determined aqueous concentration profile. With an initial value of x, a KG value was estimated by matching the acmal mass volatilized with the computer's estimate of the total mass volatilized. The value of x was then adjusted to recompute the aqueous phase concentration profile and the KG was estimated again by matching the computed mass volatilized wi± the actual mass of VOC volatilized. The iteration was completed when the two successive sets of estimated values and the actual values for the VOC aqueous phase concentration profile, air phase concentration, and the total mass of VOC volatilized differed by less than 2%. The x value determined for a given porous media was then used to estimate the KG values for all the other experimental mns involving ±e same porous media. Figure 4 shows the experimental and the estimated results for air phase and aqueous phase concentration profile for benzene after eight hours of sparging. The approach taken for this 74 Study is different from other studies. For this study, both air phase and liquid phase VOC concentrations were used to estimate the mass transfer coefficients allowing for a more accurate estimation of the mass transfer coefficients. Other smdies used only the air phase VOC concenuration to estimate the mass transfer coefficient. As such, simplifications to the mathematical models had to be made in other studies and usually a lumped model is used. As expected, the tormosity factor was found to be dependent on the porous media. The values estimated were: 0.52 for Ottawa sand, 0.51 for sand 30/50 and 0.47 for sand 70/100. The values estimated were in general agreement with the values reported elsewhere (Schaefer et al., 1995). Estimated mass transfer coefficients values {KG) are shown in Table 4 and they ranged from 1.79 x 10"^ cm/min for n­propylbenzene in sand 70/100 with an air velocity of 0.2 cm/s to 3.85 x 10'" cm/min for 1,2,4 trichlorobenzene in Ottawa sand with an air velocity of 2.5 cm/s. The overall mass transfer coefficients (cm/min) may be transformed to lumped mass transfer coefficients (s"') by multiplying with the specific interfacial area (0.09 cm"') of the air­water interface of the experimental setup. For the experimental conditions, the lumped mass transfer coefficients ranged from 2.7 x 10"^ s"' to 5.8 x 10'^ s"'. These values compared favorably well with the values of 2 to 6 x 10"^ s"' reported by Fisher et al. (1996) for the soil vapor extraction of four chlorinated VOCs in a sand box and the values (2 x 10'^ s"' to 2 X 10"^ s"') reported by Cho and Jaffe (1990) for the volatilization of TCE in soil columns during infiltration. The last two studies used experimental setups that were very different from the single­air channel reactor used in this study and were designed to depict the unsaturated zone. The similarities in the range of values for the air phase lumped mass transfer coefficients may suggest that not all the volume subjected to air sparging or soil vapor extraction was affected by the advecting air. Model Correlation of Lumped Mass Transfer Coefficients A visual inspection of the mass transfer coefficients (ATc) showed that the coefficients were inversely dependent on the Henry's law constant of the VOC but directly dependent on air velocity and some of the physical­chemical properties of the porous media. Other parameters which may correlate with the mass transfer were the concentration of the VOC present in the reactor and the aqueous and air diffusivities of the VOCs. To assess the impact 75 of the physical­chemical properties of the VOCs and porous media on the mass transfer coefficients, a regression analysis of various dimensionless numbers which incorporated the physical­chemical properties of the VOCs and the porous media were conducted. The regression analysis had two objectives: (a) to determine the parameters which may be correlated with the lumped mass transfer coefficient, and (b) to generate a model which may predict the lumped mass transfer coefficient for different conditions. Four dimensionless expressions for the mass transfer coefficient were used in the regression analysis. They were: (a) Sherwood number (5/i») for the aqueous phase, (b) Sherwood number (Sha) for the air phase, (c) modified Sherwood number (Sh^) for the air phase (Szatkowski et al., 1995), and (d) Damkohler number (CT) as described by Armstrong et al. (1993). The independent variables used were the Peclet number {Pe) for the air phase, uniformity coefficient (t/C), dimensionless particle size {do), and porosity (£) for the porous media, Henry's law constant ( KH ) for the VOCs and the dimensionless VOC concentration (X = VOC bulk concentrationA'^OC aqueous solubility). A description of all the dimensionless parameters used is presented in Table 3. The Peclet number based on the experimental conditions ranged from 0.052 to 1.523. Multiple stepwise regression analysis was conducted to determine the best fit for the four expressions of the dimensionless mass transfer coefficient with the other dimensionless numbers as shown below: log(Y) = /3o+Pilog(Pe)+fi2log(UC)+fi3log(£)+/34log(KH)+/35log(do)+P6log(X) (8) where Y represents any of the four mass transfer dimensionless numbers cited previously and are coefficients determined by the regression analysis. The stepwise regression procedure evaluated the least square residual (r) and the F statistical parameter of each predictor to determine the most appropriate model parameters. To be included in the model, variables must be significant at the F=0.15 level. The regression analysis was performed using the statistical software SAS 6.10 (SAS Institute, 1993). A summary of the stepwise analysis is shown in Table 4. The aqueous phase 76 Sherwood number correlated very poorly with the independent variables (r = 0.48), but the correlation improved when the air phase Sherwood number (Sha) was used {r = 0.6948). The modified air phase Sherwood number correlated well with the reduced mean particle size, the Henry's law constant for the VOC, the air phase Peclet number and the porosity of the porous media. Since the standard error for the estimated exponent for porosity was greater than 50% and the total improvement for the correlation was almost negligible (r changed from 0.794 to 0.799), porosity was dropped from the correlation. Both the uniformity coefficient and the reduced concentration were not selected by the regression analysis. The correlation was as follows: = 10 Pe ° dj­^ KH r = 0.7940 (9) Likewise, porosity did not improve the overall correlation for the Damkohler model and v/as dropped from the correlation. In addition to porosity, the reduced particle size was excluded from the correlation for the same reasons and neither uniformity coefficient nor the reduced concentration was selected by the regression analysis. The correlation was as follows: j j GJ= lO­*­^' = 0.8535 (10) r I Uniformity coefficient may be considered as a surrogate parameter for the particle size distribution of the porous media. The lack of influence of the uniformity coefficient on the mass transfer coefficients was in agreement with the work conducted by Wilkins et al. (1995) on the volatilization of NAPLs in unsaturated porous media. These authors found that mass transfer rate for the volatilization of NAPLs in unsaturated porous media was positively correlated with the soil grain size but was negligibly impacted by variations in grain size distribution. For the correlation described in equation 9, the reduced mean particle size and the Henry's law constant explained most of the variation of the modified Sherwood number. For the correlation shown in equation 10, Pe and KH explained all the variation of the Damkohler number. The experimental and predicted modified Sherwood and Damkohler numbers are shown in Figure 5 and Figure 6, respectively. Equations 9 and 10 may be 77 expressed in terms of KG a and ttie other variables as shown in equations 11 and 12, respectively: Kg a = 10 dpso'^'^ KH'^'^ (11) Kca = dpso­"''KH­'­'' (12) Equations 11 and 12 showed that the air phase lumped mass transfer coefficient was dependent on gas diffiisivity of the VOC but was inversely related to the Henry's law constant of the VOC and the mean particle size of the porous media. Both equations showed similar exponents for all the independent variables except for the constant term and the mean particle size. The lumped mass transfer coefficient values predicted by equation 11 were plotted against the values predicted by equation 12 as shown in Figure 7. Figure 7 shows that both correlations were equivalent and gave similar predictions of the gas side lumped mass transfer coefficients. The low value of the exponent for the air velocity indicated that air velocity have limited impact on the lumped mass transfer coefficients although the creation of more turbulent conditions in the air channel may enhance mass transfer. Likewise, VOCs with higher air diffusivities will be more favorably volatilized into the air phase. The inverse relationship of the air phase lumped mass transfer coefficient to the Henry's law constant of the VOC is correct since the air phase mass transfer coefficient is related to the liquid phase mass transfer coefficient by the following: (13) By substituting equation 13 into equations 11 and 12, the liquid side lumped mass transfer coefficient was directly dependent on the 0.17 power of the Heruy's law constant. The small positive exponent for the Henry's law constant suggests that even though the volatility of VOC may influence the transfer of VOCs from the liquid phase to the air phase, other physical­chemical properties such as the diffusion of the VOCs have control or have a more 78 dominant effect on the mass transfer process. Conclusions A one­dimensional diffusion model was used to estimate tortuosity factors and mass transfer coefficients for the volatilization of VOCs under controlled conditions. Tortuosity factors were 0.52 for Ottawa sand, 0.51 for sand 30/50, and 0.47 for sand 70/100. The overall gas phase mass transfer coefficients {Kc) were estimated from experimental data to range from 1.79 x 10"^ cm/min to 3.85 x 10'" cm/min. Two empirical models for the quantification of the lumped mass transfer coefficients of VOCs under sparging conditions were developed. The modified Sherwood number {Sh 'a) and the Damkohler number (®) were found to cortelate well with the air phase Peclet number, the Henry's law constant, and the dimensionless particle size of the porous media. Based on the correlations, the lumped mass transfer coefficient for the air phase was found to be proportional to the gas diffusivity of the VOC but was inversely proportional to the VOC Henry's law constant. Lumped air phase mass transfer coefficients did not correlate well with the particle size distribution (uniformity coefficient) and dimensionless VOC concentration. The dominant and controlling parameter in the cortelations for the prediction of the mass transfer coefficients was the gas diffusivity of the VOC. Results of this smdy provided direct experimental evidence that air sparging is a diffusion controlled process. It is expected that the correlation developed in this study for the air phase mass transfer coefficient and the physical­chemical properties of the system will provide a basis for better predictions of mass transfer fluxes in field applications. Notation Co, C„ VOC concentration in the air and aqueous phases, ML'^ VOC weighted concentration average in the MTZ, ML'^ measured VOC concentration in the exhaust air phase, ML'^ Dk­ VOC aqueous diffusivity, Da VOC air diffusivity, L?T'' 79 do dimensionless mean particle size DPSQ mean particle size, L G5 Damkohler number, dimensionless £ porous media porosity, dimensionless X tortuosity coefficient, dimensionless KH Henry's law constant, dimensionless Kow octanol­water partition coefficient, dimensionless MTZ width of the mass transfer zone, L Pe air phase Peclet number, dimensionless UC uniformity coefficient, dimensionless ^AIR air velocity, LT'' Sha air phase Sherwood number, dimensionless Shw water phase Sherwood number, dimensionless Sh'a air phase modified Sherwood number, dimensionless X reduced VOC concentration, dimensionless. a specific interfacial area, U' KG overall air side mass transfer coefficient, LT' KG a lumped air side mass transfer coefficient, 7*' KL overall aqueous side mass transfer coefficient, LT' Kia lumped aqueous side mass transfer coefficient, T' References Armstrong, J.E., J. Croise, and V. Kaleris, Simulation of rate­limiting processes controlling the vapour extraction of trichloroethylene in sandy soils, in Proceedings of the International Conference on the Environment and Geotechnics, Paris, France, pp. 327­334, April 6­8, 1993. Baehr, A.L., G.E. Hoag, and C. Marley, removing volatile contaminants from the unsaturated zone by inducing advective air phase transport. Journal of Contaminant Hydrology, vol 4. pp 1­26, 1989. 80 Braida, W.J., and S.K. Ong, Air sparging effectiveness: the air channel mass transfer zone, submitted to Water Resources Research, 1997. Brown, R.A., R.J. Hicks, and P.M. Hicks, Use of air sparging for in situ bioremediation, in Air Sparging for Site Remediation, edited by R. E. Hinchee, pp. 38­55, Lewis Publishers, Boca Raton, FL, 1994. Brusseau, M.L., Transport of organic chemicals by gas advection in structured or heterogeneous porous media: development of a model and application to column experiments. Water Resources Research, vol. 27(12), pp. 3189­3199, 1991. Chihacek L.J., and J.M. Bremmer, A simplified ethylene glycol monoethyl ether procedure for assessment of soil surface area. Soil Sci. Soc. Am. J., vol. 43, pp. 821­822, 1979. Cho, H.J., and P. Jaffe, The volatilization of organic compounds in unsaturated porous media during infiltration. Journal of Contaminant Hydrology, vol. 6, pp. 387­410, 1990. Driscoll, F.G., Groundwater and Wells, 2"^* Edition, Johnson Filuration System, St. Paul, Minn., 1986. Fisher, U., R. Schulin, M. Keller, and F. Stauffer, Experimental and numerical investigation of soil vapor extraction. Water Resources Research, vol. 32(12), pp. 3413­ 3427, 1996. Garbarini, D.G., and L.W. Lion, Evaluation of sorptive partitioning of nonionic pollutants in closed systems by headspace analysis. Environ. Sci. TechnoL, vol. 19(1), pp. 1122­1129, 1985. Gierke, J.S., N.J. Hutzler, and D.B. McKenzie, Vapor transport in columns of unsaturated soil and implications for vapor extraction. Water Resources Research, vol. 28(2), pp. 323­ 335, 1992. Hein, G.L., N.J. Hutzler, and J.S. Gierke, Quantification of the mechanisms controlling the removal rate of volatile contaminants by air sparging, in Proc. of the 1994 National conference on Environmental Engineering, edited by J.N. Ryan and M. Edwards, pp. 556­ 563, ASCE, NY, NY, 1994. IBM, System/360 Continuous System Modeling Program user's manual. Program Number 360A­CX, 5th ed., IBM Corp., Technical Publ. Dept., White Plains, New York, 1972. 81 Ji, W., A. Dahamani, D.P. Ahlfeld, J.D. Lin, and E. Hill, Laboratory study of air sparging; Air flow visualization, GWMR, vol. 13(4), pp. 115­126, 1993. Ji, W., Air sparging: Experimental and theoretical analysis of flow and numerical modeling of mass transfer, Ph.D. Dissertation, 154 pp.. The University of Connecticut, Storrs, CT, 1994. Lewis, W.K., and W.G. Whitman, Principles of gas adsorption. Industrial and engineering Chemistry, 16, pp. 1215­1220, 1924. Lyman, W.J., W.F. Reehl, and D.H. Rosemblatt, Handbook of Chemical Properties Estimation Methods, American Chemical Society, Washington D.C., 1990. Mendoza, C., and E.O. Frind, Advective­dispersive transport of dense organic vapors in the unsaturated zone: L Model development, Warer/?e50urces, vol. 23(3), pp. 379­387, 1992. Nelson, D.W., and L.E. Sommers, Total carbon, organic carbon, and organic matter, in Methods of Soil Analysis, Part 2, Chemical and Microbiological Properties, Second Edition, edited by A.L. Page, pp. 570­571, American Society of Agronomy, Inc., Soil Science Society of America, Inc., Madison, WI, 1982. Perry, R.H., and C.H. Chilton, Chemical Engineer's Handbook, 5''' Edition, pp. 3­231­ 3.234, McGraw­ffill, NY, NY, 1978. Treybal, R.E., Mass Transfer Operations, 2"'^ Edition, pp. 29­31, McGraw­Hill Kogakusha Ltd., Tokyo, Japan, 1968. Schaefer, C.E., R.R. Arands, H.A. van der Sloot, and D.S. Kosson, Prediction and experimental validation of liquid­phase diffusion resistance in unsaturated soils. Journal of Contaminant Hydrology, vol. 20, pp. 145­166, 1995. Sleep, B.E., and J.F. Sykes, Modeling the transport of volatile organics in variable saturated media. Water Resources Research, vol. 25(1), pp. 81­92, 1989. SAS Institute Inc., 5A5® Software: Changes and Enhancements. Release 6.10, Carey, NC, 1993. Szatkowski, A., P.T. Imhoff, and C.T. Miller, Development of a correlation for aqueous­ vapor phase mass transfer in porous media. Journal of Contaminant Hydrology, vol. 18, pp. 85­106, 1995. 82 Wilkins, M.D., L.M. Abriola, and K.D. Pennel, An experimental investigation of rate­ limited nonaqueous volatilization in unsaturated porous media: steady state mass transfer. Water Resources Research, vol. 31(9), pp. 2159­2172, 1995. 83 Table 1. Physical­chemical properties of porous media Type of Sand Specific Surface Area (m"/g) Porosity Organic Carbon Uniformity Coefficient Ottawa Sand Mean Particle Size (cm) 0.0190 2.16 1.99 0.377 0.0066 Sand 30/50 0.0305 1.41 1.17 0.370 0.0062 Sand 70/100 0.0168 1.64 2.73 0.400 0.0063 (%) Table 2. Physical­chemical properties of VOCs at 20°C (Lyman et al., 1990) Compound Molecular Solubility DwXlO^ Da KH (mg/L) (cmVs)* (cm'/s)** Weight Benzene 0.0923 78.12 1780 9.59 0.195 log Kow 2.12 Toluene 92.15 0.233 515 0.0830 8.46 2.73 Ethylbenzene 106.18 0.291 152 0.0732 7.63 3.15 o­Xylene 106.18 0.178 130 0.0759 7.63 2.95 m­Xylene 106.18 0.247 175 0.0759 7.63 1.38 p­Xylene 106.18 0.256 198 0.0759 7.63 3.26 Chlorobenzene 112.56 0.137 500 0.0725 8.52 2.84 n Propylbenzene 120.21 0.369 60 0.0544 6.99 3.87 1,2 Dichlorobenzene 147.00 0.118 145+ 0.0829 7.97 3.60 1,2,4 Trichlorobenzene 181.44 0.069 30+ 0.0686 7.09 4.30 Styrene 104.16 0.0967 300 0.0746 7.89 2.95 * Estimated from Hirschfelder, Bird, and Spotz (Perry and Chilton, 1978) ** Estimated from Wilke and Chang correlation (Treybal,1968) + at 25°C Table 3. Estimated air phase mass transfer coefficients for various compounds, porous media, and air velocities (cm/min). Media odawa Ottawa oKawa Ottawa 30/50 30/50 30/50 30/50 70/l(X) 70/l(K) 70/100 70/100 Air Velocity (cm/s) 0.2 0.5 I.I 2.5 0.2 0.5 1.1 2.5 0.2 0.5 I.I 2.5 Benzene Ethyl­ benzene in.p­ Xylene o­Xylene Toluene Chloro­ bcnzene Styrene n­Propyl­ benzene 0.00648 0.00410 0,00486 0.00594 0.00668 0.00635 0.00391 0.00510 0.00558 0.00876 0.00517 0.00689 0,00338 0,00298 0.00436 0.00464 0.00483 0.00406 0.00336 0.00439 0,00361 0,00633 0.00398 0,00514 0,00361 0,00385 0,00467 0,00394 0,00465 0,00432 0,00319 0,00444 0,00339 0,00594 0,00605 0,00510 0.00563 0.00528 0.00436 0.00595 0.00527 0.00717 0,00534 0.00598 0,00434 0,01097 0,00746 0,01034 0,00379 0.00879 0.00637 0.00721 0.00376 0,00524 0,(M)467 0.(K)526 0.00417 0.(K)537 0,00637 0,00551 0,00589 0.01328 0,01156 0,01243 0.00707 0.00897 0.00784 0.00797 0.00511 0,00782 0.00802 0,00787 0,00597 0.01599 0.01707 0.01956 0.00996 0,01180 0,01105 0.01087 0.00743 0,00967 0,01435 0,01781 0,00329 0.00504 0.00463 0.00487 0.00280 0.(K)209 0,(KI466 0.00412 0,00179 0,00257 0,00221 0,00232 1.2 Dichloro­ benzcne 0.00597 0.01019 0,00923 0,01253 0,00393 0.00245 0,01283 0.00736 0.00629 0.00672 0.00766 0,(K)78I 1.2.4 Trichloro­ benzcne 0,01377 0,00965 0,02487 0.03854 0,01361 0.01016 0,01913 0,00913 0,00671 0,01445 0,(X)822 0,01995 85 Table 4. Dimensionless numbers used for modeling Dimensionless Number Equation Comments Mass transDort bv water­air partition Damkohler Number KoCiK (GJ) Mass transport by gas advection Va,r Aqueous Phase Sherwood Number ( S h w ) KL^PSO Mass transDOrt bv water­air partition Mass transport by liquid diffusion Air Phase Sherwood Number ( S h a ) ^C^Pso Mass transport bv water­air partition Mass transport by gas diffusion Modified Air Phase Sherwood Number (Sh 'a) KQQCIP^ Mass transDOrt bv water­air partition Mass transport by gas diffusion Sol. ^airdPsO Actual VOC bulk concentration VOC aqueous solubility Rate of transport bv sas advection Rate of transport by molecular gas diffusion Dimensionless Concentration Air Phase Peclet Number (Pe) Dimensionless Mean Grain Size ( d o ) Porosity (e) Henry's Law Constant Kh Uniformity Coefficient (UC) dPso d. V Toted ­V Sclid vTotal 0.05 cm is the mean grain size of a "medium" size sand (Driscoll, 1986) dn,= Void volume Total volume Ratio of air phase concentration to aqueous phase concentration Measure of the grain size distribution ^.0 86 Table 5. Summary of stepwise regression analysis Modified air phase Sherwood number Variable Parameter Estimate Standard Error ­7.1387 Intercept 0.0646 Partial r Model r ­) do 1.6598 0.1217 0.4329 0.4329 KH ­0.8264 0.0603 0.3186 0.7515 Pe ­0.1626 0.0332 0.0425 0.7940 Damkohler number Variable Parameter Estimate Intercept ­4.8104 Standard Error 0.0568 Partial r Model r Pe ­0.7875 0.0356 0.6614 0.6614 KH ­0.8332 0.0672 0.1921 0.8535 Sample Ports Air Channel Gap 1.58 mm Clean Humidified Air ~ZL Contaminated Air \o Gas Sampling Point Saturated Porous Media Contaminated with VOCs 17.5 cm I'igure 1. Single­air channel apparatus OO 88 0 100 200 300 400 500 Time (min) Air­Water Interface T = Oh : T = 8h T = 4h ­20 ­ ^ ­40 ­ • ­60 H ­80 ­ •100 0 25 50 75 100 Aqueous Benzene Cone. (mg/L) Figure 2. (a) Benzene concentration (mg/L) in the exhaust air (b) Benzene concentration (mg/L) in aqueous phase for sand 70/100 and air velocity of 1.1 cm/s Cu Bulk air phase concentration Cai Interfacial air phase concentration Cw Bulk liquid phase concentration Cwi Interfacial liquid phase concentration Air Flow Completely Mixed Air Phase (bulk) Air Layer Interface Completely Mixed Interface Aqueous Layer Mass Transfer Zone Liquid Phase (bulk) (MTZ) Bulk Porous Media Completely Mixed Two­Resistance Model (Diffusion Controlled Transport) Air Channels I'igure 3. Air­water interface mass transfer, two­resistance model and air channel conditions 90 1.5 •J (a) ac S ^ 1.0 H Model Experimental 0 100 200 300 400 500 Time (min) Air­Water Interface (b) ­20 ­ o ­40 ­ • ­60 ­ 0 ® ­80 ­ Time = 8 h (Exp.) Time = 8 h (Model) c 0 25 50 75 100 Aqueous Benzene Concentration (mg/L) Figure 4. Experimental and predicted benzene concentration for sand 70/100 and air velocity of 1.1 cm/s: (a) exhaust air, (b) aqueous concentration 91 1e­6 oog oo , (CO o oo ^ o 1e­8 1e­8 1e­7 1e­6 Experimental Modified Sherwood Number Figure 5. Experimental vs. predicted modified Sherwood number and 95% confidence interval for the population 92 1.0e­2 ­q 1.0e­3 ­ 1.0e­4 ­ 1.0e­5 1.0e­5 1.0e­4 1.0e­3 1.0e­2 Experimental Damkohler Number Figure 6. Experimental vs. predicted Damkohler number and 95% confidence interval for the population 93 4e­5 3e­5 oo 2e­5 1e­5 OqO oo 1e­5 2e­5 3e­5 4e­5 K(~a in s'^ (equation 12) Figure 7. Estimated gas side lumped mass transfer coefficient using eq. 11 estimated lumped gas side mass transfer coefficient using eq. 12 94 CHAPTER FIVE. VOLATILIZATION OF VOCs UNDER AIR SPARGING CONDITIONS: MASS TRANSFER ANALYSIS Introduction Nonequilibrium mass transfer between air and water and vice versa has been extensively studied and many models have been proposed to explain the transport mechanisms governing the transfer of molecules across the interface. Chapter 2 presented a summary of the different conceptual and hydrodynamic models which have been developed to explain the volatilization and adsorption of chemicals at the air­water interface. Although a fair amount of research has been done in this area, studies on the controlling mechanisms in the volatilization of VOCs in saturated porous media during air sparging is fairly scarce. In previous chapters, diffusion was identified as the controlling mechanism in the nonequilibrium mass transfer of VOCs from groundwater to the air­water interface during air sparging operations. Although the overall mass transfer of VOCs seemed to be controlled by their diffusion through the porous media to the air channels, the mechanisms controlling the interfacial phenomenon of volatilization have not been elucidated yet. The objective of this chapter is to study the controlling mechanism for the volatilization of VOCs at the air­water interface during air sparging operations. This objective was achieved by using the estimated mass transfer coefficients in Chapter 4 to test the applicability of the two­resistance model of Whitman (1923). The contributions of the liquid and air phase resistances to the overall mass transfer resistance were estimated and the influence of parameters such as air velocity and mean particle size on the relative resistance to the mass transfer was assessed. Materials and Methods A description of the single­air channel experimental setup, experimental procedures, and computational methods used for the determination of mass transfer coefficients was presented in Chapter 4. Appendix C includes a sample of the CSMP computer program developed for the estimation of the mass transfer coefficients. 95 Results and Discussion A conceptual model, describing the diffusive exchange of chemicals between water and air is the two­resistance model. The model was first described by Whitman in 1923 and first applied to environmental transfer by Liss and Slater (1974). The model agrees with the mass flux expressions (equation 1) used in Chapter 4 for the estimation of mass transfer coefficients except that the model assumes the existence of two stagnant thin films on each side of the air­water interface. Transfer of chemicals or solutes from one phase to the other must diffuse through the two thin layers in series. The rate at which mass is transferred in each layer is characterized by the mass transfer coefficients k which is essentially the transfer velocities. Based on the two­resistance model, the inverse of the overall mass transfer coefficient is equal to the sum of the liquid and gas phase film resistance: K k k ^ where kL and k^ are the liquid and gas film coefficients (LT'), respectively. For most chemicals with high Henry's law constant, the term Ku/kt is usually larger than Hk^, making the liquid film mass transfer the controlling mass transfer mechanism. Overall gas side mass transfer resistances were computed by taking the reciprocal of the overall gas side mass transfer coefficients estimated in Chapter 4. For each experimental condition (i.e., porous media type and air velocity), the values of the overall gas side mass transfer coefficients (ATG"') were plotted against the Henry's law constant of the VOCs. As seen in equation 1, this plot should result in a straight line with slope equal to Hki and intercept of llk^. Figures I to 3 show plots of the estimated overall gas side mass transfer resistances against the Henry's law constant and the regression lines with their 95% confidence intervals. Ten different VOCs were initially used in each plot with Henry's law constants ranging from 0.069 to 0.369. However, to obtain a good linear relationship, outliers were not included in the regression analysis. The good linear relationships in Figures 1 to 3 suggest that the two­resistance model of Whitman may be applied to the volatilization of VOCs under air sparging conditions. The surface renewal model posmlated by 96 o 400 400 300 ­ 300 ­ 200 ­ 100 ­ 100 ­ (b) 0.5 cm/s (a) 0.2 cm/s 0 —1—I—I—[­T—I—I—I—I—I—I—I—I—I—I—I—I—I—I—r 0.0 0.1 0.2 0.3 Kh 400 400 300 ­ • o • o 200 ­ 100 ­ 200 100 ­ (c) 1.1 cm/s (d) 2.5 cm/s 1—I—I—I—I—I—I—I—I—I—I—I—I—I—I—I I I I 0.0 0.1 0.2 0.3 0.4 Kh Kh Figure I. Interfacial mass transfer resistance vs. Henry's Law constant for Ottawa sand and 95% confidence limits 0.4 97 400 400 300 ­ 300 ­ 200 ­ • o 100 ­ 200 ­ 100 ­ (b) 0.5 cm/s (a) 0.2 cm/s Kh Kh 400 400 300 ­ 300 ­ • o 200 ­ 100 ­ 200 100 ­ (d) 2.5 cm/s (c) 1.1 cm/s Kh Kh Figure 2. Interfacial mass transfer resistance vs. Henry's law constant for sand 30/50 and 95% confidence limits 98 c3 400 400 300 ­ 300 ­ o 200 ­ 100 ­ 100 ­ (a) 0.2 cm/s 0 (b) 0.5 cm/s I r~T I I r~T I r—j i i i i [ n i—p 0.0 0.1 0.2 0.3 "T—I—rn—I—I—!—I I I I I—I—I—pT—I—I—r 0.4 0.0 0.1 Kh • c3 0.2 0.3 0.4 Kh 400 400 300 ­ 300 ­ 200 ­ CD 100 ­ 200 ­ 100 ­ (c) 1.1 cm/s (d) 2.5 cm/s ­T—I—rn—I—I—;—rn—|—i—\—\—i—|—i—i—i—rJ 0.0 Kh 0.1 0.2 0.3 Kh Figure 3. Interfacial mass transfer resistance vs. Henry's law constant for sand 70­100 and 95% confidence interval 0.4 99 Danckwerts (1951) may show a similar behavior if the overall resistance was plotted against the Henry's law constant (Schwarzenbach et al., 1993). The Danckwerts model assumed a continuous turnover of bulk fluid at the interface as a consequence of the level of turbulence in the fluid. Although this situation may be acceptable for the highly turbulent air phase in the air channels during air sparging, the aqueous phase surrounding the air channels is rather quiescent. Therefore, the turbulence effects may be assumed to be marginal as compared to the diffusional transport of the VOCs. Assuming that the liquid side (1/^z.) and gas side (1//:^) contributions to the overall resistance were only dependent on the porous media and air velocity, {Hki) and (1/^^) may be estimated. The estimated values for (l/A:^) and (I/^a) along with their standard deviations are presented in Table 1. The correlation coefficients for all the plots except for two were close to one. Table 1. Liquid and gas side mass transfer resistance. Regression analysis results. 5 Porous Air Velocity r~ (cm/s) (min/cm) Media (min/cm) 0.90 Ottawa 0.2 653 ±108 88 ±26 Ottawa 0.94 0.5 1401±165 ­75 ± 32 0.86 Ottawa 634± 117 1.1 17 ±27 0.93 Ottawa 2.5 770 ± 97 ­17 ±20 Szuid 30­50 0.90 0.2 928 ±140 8 ±32 Sand 30­50 0.98 0.5 870 ± 61 ­7± 13 Sand 30­50 0.91 1280^184 l.l ­42 ±36 Sand 30­50 2.5 0.92 750 ± 99 31± 18 Sand 70­100 0.2 0.90 748 ±142 78 ±22 Sand 70­100 0.5 0.77 904 ± 223 ­5 ±49 Sand 70­100 0.65 1.1 573 ±187 50 ±39 Sand 70­100 2.5 1+29 712 + 139 0.84 Examination of the values for the gas side mass transfer resistances showed a larger variation in the data with five negative values. Since is not possible to have a negative resistance, for these cases, a value equal to zero, i.e., no gas side resistance, was assumed for comparative purposes. The large variation in the data was expected considering the accumulation of uncertainties after two transformation of the experimental data (i.e., concentration to mass transfer coefficients to mass transfer resistances). Table 1 showed that 100 the liquid side mass transfer rates {ki) were smaller than the corresponding gas side mass transfer rates This result is in agreement with the findings reported in Chapter 3 and Chapter 4 whereby liquid diffusion controls the volatilization of VOCs under air sparging operations. The liquid mass transfer resistance is given by ±e product of the Henry's law constant (Kfi) and the reciprocal of the liquid side mass transfer rate (ki). Its relative influence in the overall resistance is dependent on the type of chemical. Table 2 shows the relative influence of the gas and liquid side mass transfer resistance on the overall gas side mass transfer resistance for 1,2,4 trichlorobenzene (KH = 0.069), and n­propylbenzene {KH = 0.369). 1.2,4 trichlorobenzene and n­propylbenzene were the least volatile and most volatile VOCs used in the study, respectively. Table 2. Relative liquid side and gas side mass transfer resistance for selected VOCs. Porous Media Air Velocity (cm/s) 1,2. Trichlorobenzene Gas Side Liquid Side Resistance Resistance (%) (%) 58 ­ 7 2 2 8 ­• 4 2 n­Propyibenzene Gas Side Liquid Side Resistance Resistance (%) (%) 6 8 ­• 8 0 20 ­ 3 2 Ottawa 0.2 Ottawa 0.5 0 100 0 100 Ottawa I.l 0 ­50 5 0 ­ 100 0 ­ 12 8 8 ­ 100 Ottawa 2.5 0 ­5 9 5 ­ 100 0­1 9 9 ­100 Sand 30/50 0.2 0 ­38 6 2 ­ 100 0 ­ 10 9 0 ­ 100 Sand 30/50 0.5 0 ­9 91 ­ 100 0­ 2 9 8 ­ 100 Sand 30/50 1.1 0 100 0 100 Sand 30/50 2.5 20 ­ 4 9 51 ­ 80 4­­ 15 8 5 ­ 96 Sand 70/100 0.2 51 ­ 6 6 3 4 ­ 49 17 ­ 2 7 73 ­ 87 Sand 70/100 0.5 0 ­42 5 8 ­ 100 0­• 11 8 9 ­ 100 Sand 70/100 1.1 18 ­ 6 4 3 6 ­ 82 5 •­ 3 0 7 0 ­ 95 Sand 70/100 2.5 0­­ 4 0 6 0 ­ 100 0­­ 10 9 0 ­ 100 The relative resistance included in Table 2 show that for VOCs with high values of Henry's law constant, most of the resistance to the mass transfer was located in the liquid phase. In general, liquid side resistance accounted for more than 90% of the total resistance to the mass transfer. For VOCs with low volatility such as 1,2,4 trichlorobenzene, a higher 101 portion of the overall resistance was in the gas phase for low air velocity. The air velocity and porous media characteristics seemed to have an impact on the relative magnitude of the two resistance. For example, the gas side mass transfer resistance for Ottawa sand seemed to decrease from 60% to 5% when the air speed increased from 0.2 cm/s to 2.5 cm/s. Sand SO­ SO showed a similar general trend for 1,2,4 trichlorobenzene. No consistent trends were observed to quantify the influence of porous media in the relative distribution of the mass transfer resistance. According to the Whitman (1923) model, the liquid and gas phase mass transfer rates may be related to the diffusivity of the VOCs as follows: kt = D^./S^. kA=DA/S^ (2) where D^. and DA are the diffusivities of the chemical in bulk water and air, respectively, and (L) and (L) are the thicknesses of the stagnant layers in the water phase and the air phase, respectively. Schwarzenbach et al. (1993) reported values of 5x10'^ to 5xlO'~ cm for and 0.1 to 1 cm for Sa for the transport of chemicals in water reservoirs to the atmosphere. In saturated porous media under stagnant conditions and with minimum mixing, the size of the stagnant liquid side film, 5v, may be assumed to be larger than the values reported by Schwarzenbach and coworkers. Using equation 2, the thickness of the water phase stagnant layer may be computed as presented in Table 3. The estimated thicknesses of the liquid phase stagnant film were one to two orders of magnimde larger than the corresponding values for water reservoirs. The results imply that mrbulent eddy diffusion was either not present or was very small in the liquid side of the interface. This suggests that the air flow in the air channels did not introduce any significant mixing effect on the surrounding saturated porous media. Table 3 shows that in general decreased with increasing air velocities. Because of the errors involved in the analysis, a 10% to 20% difference in the values of for a 10 fold increase in the air velocity, would not indicate that there were some air mixing effects. To further clarify this point. Figure 4 presents a plot of the liquid side mass transfer coefficient as a function of the air velocity for the three porous media used. Figure 4 suggests that the ki were fairly constant if the 0.004 A o o Ouawa Sand Saiid 30/50 0.000 ­ ~i —I—I—I—I—I—I—I—I—I—I—I—r ­|—I—I—r 0.003 ­ "E" E E Sand 70/100 0.002 ­ 0.001 ­ 0 T () 12 3 4 Air Velocity (cm/s) Figure 4. Liquid side mass transfer coefficients as a function of air velocity for different porous media 103 experimental errors were considered. Another way of visualizing the size of the liquid side film thickness was to compare the <5L with the grain size of the media. Values presented in Table 3 inferred that were approximately 15 to 50 times the grain size of Ottawa sand. Data in Table 3 shows that the liquid phase layer did not penetrate more than 0.9 cm into the porous media. As a result of this lack of mixing, the movement of the VOCs towards the air­ water interface will be controlled by the liquid diffusion. In comparison to the MTZ as reported in Chapter 3, the thickness of the stagnant liquid layer was between 20% to one order of magnitude smaller than the size of the MTZ. Table 3. Estimated values for <5u­ (cm) for Ottawa sand at different flow rates voc 0.2 cm/s 0.5 cm/s 1.1 cm/s 2.5 cm/s Benzene 0.38 ± 0.06 0.81 ±0.09 0.36 ± 0.07 0.44 ± 0.06 Ethylbenzene 0.30 ± 0.05 0.64 ±0.08 0.29 ± 0.05 0.36 ± 0.05 m,p­Xylene 0.30 ± 0.05 0.64 ±0.08 0.29 ± 0.05 0.36 ± 0.05 o­Xylene 0.30 ± 0.05 0.64 + 0.08 0.29 ±0.05 0.36 ± 0.05 Toluene 0.33 ± 0.05 0.72 ± 0.08 0.32 ±0.06 0.39 ± 0.05 Styrene 0.31 ±0.04 0.66 ±0.08 0.30 ± 0.05 0.37 + 0.05 n­Propylbenzene 0.29 ± 0.04 0.59 ±0.07 0.28 ± 0.05 0.32 ± 0.04 Chlorobenzene 0.34 ± 0.06 0.72 ± 0.08 0.27 ± 0.05 0.39 ± 0.05 1.2 DCB 0.31 ±0.05 0.66 ±0.08 0.30 ± 0.05 0.37 ± 0.06 1.2,4 TCB 0.28 ± 0.05 0.59 ± 0.07 0.27 ± 0.05 0.33+0.04 Conclusions The controlling mechanisms in the volatilization of VOCs at the air­water interface during air sparging operations were analyzed using the nonequilibrium mass transfer coefficients determined by using a single­air channel setup. Regression analysis of the data showed that the two­resistance model of Whitman may be used to describe the volatilization of VOCs. Unlike the bulk water in reservoirs and rivers, the presence of porous media resulted in the lack of mixing in the liquid phase. This may favor the use of the two­resistance model for the description of liquid phase mass transfer. For VOCs with large Henry's law constants (such as n­propylbenzene), the liquid side film controlled the volatilization at the interface 104 and may account for more than 90% of the total resistance. For VOCs with low Henry's law constants (such as 1,2,4 trichlorobenzene), air velocity and the mean particle size of the porous media have some impact on the relative magnimde of the gas side film resistance to mass transfer. The air phase resistance may range from 0% to 72% of the total resistance for a compound such as 1,2,4 trichlorobenzene with a low Henry's law constant. Even though gas phase resistance seemed to decrease with higher air velocities, no consistent trends were observed to assess the influence of porous media characteristics on the relative magnitude of the gas side resistance to mass transfer. Experimental data implies that there was a lack of mixing effect on the liquid side of the air channel. As a result of this, any technological modification of air sparging that would improve the mixing in the liquid phase would likely improve the efficiency of air sparging as a remediation technology. For example, intermittent air injection may result in improvements in VOC removal. References Danckwerts, P. V., Significance of liquid­film coefficients in gas adsorption, Ind. Eng. Chem., vol. 43(6), pp. 1460­1467, 1951. Liss, P.S. and P.G. Slater, Flux gases across the air­sea interface. Nature, vol. 247, pp. 181­184, 1974. Schwarzenbach, R.P., P.M. Gschwend, and D.M. Imboden, Environmental Organic Chemistry, p. 218, John Wiley & Sons, Inc., New York, 1993. Whitman, W.G., The two­film theory of gas absorption, Chem. Metal Eng., vol. 29, pp. 146­150, 1923. 105 CHAPTER SIX. FATE OF NONAQUEOUS PHASE LIQUIDS UNDER AIR SPARGING CONDITIONS A paper to be submitted to Journal of Contaminant Hydrology Washington J. Braida and Say Kee Ong Abstract Nonequilibrium air­water mass transfer experiments using a single­air channel setup were conducted to investigate the fate of nonaqueous phase liquids (NAPLs) under air sparging conditions. Chlorobenzene was used as a dense NAPL (DNAPL) while benzene was used as a light NAPL (LNAPL), respectively. Two dimensional isoconcentration profiles drawn from experimental results showed that air sparging may effectively control the spreading of NAPLs plumes and may be used as a remediation technology. The control of the spreading and remediation of the contaminant plume seemed to be more effective for NAPLs with higher solubilities and diffusivities. Removal efficiency for the VOCs was affected by the grain size of the porous media. More than 50% reduction in the removal rate of benzene was found when a porous medium of mean particle size of 0.168 mm was used instead of a porous medium with a mean particle size of 0.305 mm. Experimental results suggest that diffusion of VOC from the NAPL to the air­water interface controlled the mass ttansfer. Removal efficiencies were independent of air flow rate. According to the results of this study, air sparging may be an effective remedial tool for controlling the contaminant plume. Effective remediation of NAPLs can only be realized if the advective air flow can interact with the NAPL and diffusion limitations can be overcome. Introduction In the past few years, the growth of innovative technologies such as air sparging for the remediation of contaminated soil and groundwater has created the need for a deeper understanding of the physical/chemical/biological processes involved during remediation. Organic contaminants may be found dissolved in the groundwater, adsorbed onto the soil 106 matrix, or present as a separate phase. Free product of organic compounds in the form of nonaqueous phase liquids (NAPLs) may be present in contaminated groundwater close to the source of contamination. NAPLs may be categorized as: dense nonaqueous phase liquid (DNAPLs) which are denser than water and light nonaqueous phase liquid (LNAPLs) which are less dense than water. Organic compounds forming DNAPLs include chlorinated hydrocarbons while petroleum products and low molecular weight hydrocarbons tend to form LNAPLs. As a result of gravity and capillary forces NAPLs move through the unsaturated zone downward and laterally. Once they reach the water table, LNAPLs will spread on the freatic surface while DNAPLs may continue to migrate below the water table and into the aquifer. Pump­and­treat technology has been shown to be ineffective in the remediation of NAPLs. Thus, alternative remedial technologies such as in situ air sparging has been the focus of much research recently for the remediation of NAPLs. Although air sparging has been used as a remedial technology for sites containing NAPLs, not much is known about the effect of air sparging on the fate and spreading of NAPLs in the saturated porous media. Using a laboratory­scale air sparging system, Johnson et al. (1997) reported that the removal efficiency of NAPLs increased with pulsed air injection and for compounds with high solubilities but was not affected by an increase in air velocities. As expected, the authors observed low removal rates for chemicals located away from the air channels. McCray and Falta (1997) searching for more accurate means to define the radius of influence of sparging wells, reported that the effective zone for NAPL remediation would overlap the regions of high air saturations. Wilkins et al. (1995) reported that the rate of volatilization of NAPLs in unsaturated porous media was dependent on the air velocity, the diffusivity of the organic compound in air, and the porous media mean particle size. The physical­chemical behavior of NAPLs in an aquifer is expected to be different from the behavior of the dissolved product. For example, under air sparging conditions, partitioning of organic compounds between NAPLs and the advective air stream is expected to be governed by Raoult's law rather than Henry's law. Besides the direct volatilization of NAPLs, the fate and distribution of NAPLs in porous media is also dependent on the dissolution of the NAPL. Several researchers have conducted simple visualization 107 experiments to study the distribution of NAPLs in porous media (Kennedy and Lennox. 1997; Schroth et al., 1995; Zhou and Blunt, 1997; Conrad et al., 1992), while others have investigated the dissolution of NAPLs in saturated porous media. Miller et al. (1990) found that the local equilibrium assumption may be used to describe the dissolution of toluene NAPLs in glass beads over a wide range of NAPL samration and groundwater flow. Powers et al. (1992, 1994) reported that NAPLs dissolution rates for steady state conditions were dependent on the distribution pattern of the NAPLs as well as the aqueous phase velocity. For the transient conditions. Powers and coworkers found that mass transfer was related to die porous media properties, aqueous phase Reynolds number, and volumetric fraction of the NAPL. In summary, the fate and transport of NAPLs in the saturated zone is a complex issue. Mass removal of NAPLs using air sparging may be affected by the dissolution characteristics of the NAPL and by the diffiisivity of the dissolved product in the aqueous phase. Research on the behavior of NAPLs under air sparging conditions is fairly scarce in the Uterature. The objective of this study was to improve our conceptual understanding of the behavior of NAPLs under air sparging conditions by qualitatively investigate the dissolution, diffusion, and volatilization behavior of LNAPLs and DNAPLs by using a single­air channel setup. Experiments were designed to show the impact of air flow rate, type of porous media, and chemical type on the removal and control of plume spreading during air sparging. Materials and Methods The dissolution and volatilization patterns of NAPLs under air sparging conditions were investigated using a single­air channel setup (Braida and Ong, 1997a, 1997b). The experimental setup consisted of a box made of thick Plexiglas ™ in which air was circulated in a single air channel, 1.58 mm thick, over saturated porous media. A sketch of the experimental setup is shown in Figure 1. Even though the air flow was directed horizontally over the porous media, the experimental setup was designed to reproduce the air flow in air chaimels. The experimental setup allowed a 2­D visualization of NAPL dissolution and spreading in porous media under air sparging conditions. The VOCs concentrations in water were measured by using 15 sampling points placed across the experimental setup as shown. 108 A sampling point to measure the VOC concentration in the exhaust line of the experimental setup was also included. In­house compressed air was filtered and then humidified before being used as sparging gas. Air flow was measured using a Gilmont Model 11 flowmeter (Barrington, EL). The porous media used were sand 30/50 and sand 70/100 from U.S. Silica Company (Ottawa, IL) with mean particle size of 0.305 mm and 0.168 mm, and uniformity coefficients of 1.41 and 1.64, respectively. Organic carbon contents were determined using the Walkley­Black procedure (Nelson and Sommers, 1982) and were found to be less than 0.007% for both porous media. In a typical experiment, the porous media were mixed with distilled water and the slurry was then carefiilly packed layer by layer into the reactor to minimize the formation of air pockets. The resulting porosity of the packed media was 0.37 for the sand 30/50 and 0.4 for the sand 70/100. The top cover was then secured to provide an air gap of 1.58 mm above the porous media. The air gap was within the size of air channels (10­20 pore sizes) observed by others (Ji et al., 1993) Two organic compounds were used. They were benzene for LNAPL and chlorobenzene for DNAPL. Both chemicals were of HPLC grade purchased from Sigma­Aldrich Chemical Company Inc. (Milwaukee, WI). NAPLs were created by injecting 60 |iL of the chemical into the reactor. This represents a total mass of 52 mg for benzene and 66 mg for chlorobenzene. The NAPL was placed at approximately 19 mm below the air­water interface, 88.9 mm from the air inlet, and 25 mm from the front and back side walls. Experiments were conducted with no aeration, and with air flow rates of 27.5 and 68 mL/min. The experimental setup was kept a constant temperature of 21 + 2°C. A summary of the experimental matrix is shown in Table 1. The VOC concentration in the air phase and in the liquid phase were measured with a Hewlett­Packard 5890 Series 11 gas chroraatograph (Avondale, PA) equipped with a HP­5 capillary column and a flame ionization detector. Air phase concentration was determined by direct injection of 1 mL sample of exhaust air while the liquid phase concentration was measured using the head­space technique. For the head­space technique, 25 |j.L of an aqueous sample was placed in a 1.8 mL aluminum crimp cap vial. After equilibrium was 109 reached, the head­space concentration was measured and the aqueous concentration was estimated from the measured head space concentration. Aqueous phase samples were collected at the start of the experimental run, 24 hours, 48 hours, and 72 hours. Results and Discussion Based on the measured aqueous concentrations, two dimensional isoconcentration lines for the porous media were drawn by kriging using the Surfer 5.03 software package (Golden Software Inc., Golden, CO). Figure 2 shows two examples of isoconcentration lines and the measured concenU:ations. In this chapter, isoconcentration lines will be used to provide a qualitative assessment of the behavior of NAPLs under different air sparging conditions. The isoconcentration lines at various times throughout the experimental run are presented in Figures 3 to 8. An infinite point source with a concentration equal to the VOC solubility was assumed for the computations. Stagnant air conditions Figures 3a, 3b, and 3c show the changes in the dissolved benzene concentrations for sand 30/50 and zero air flow for 0, 24 and 48 hours, respectively. Figures 4a, 4b, 4c, and 4d show the results for chlorobenzene under the same conditions. For benzene, an almost symmetrical diffusion of benzene from the NAPL was seen after 24 hours. However, after 48 hours, an asymmetric distribution was evident. The vertical spreading was found to be less than the lateral spreading. Benzene concentrations at the air­water interface was between 100 to 500 mg/L directly above the location of the NAPL. At 48 hours, the lower isoconcentration line (1 mg/L) of the plume did not extend any further and was similar to that at 24 hours. Lateral diffusion clearly dominated over the downward vertical movement of benzene. This may indicate that volatilization at the air­water interface and, to certain extent, the buoyancy characteristics of the NAPL had an influence on the diffusion of the dissolved benzene in saturated porous media. On the other hand, chlorobenzene, a DNAPL, under the same experimental conditions, seemed to show a different behavior (see Figures 4a, 4b, 4c, and 4d). Chlorobenzene, with a solubility three times lower than benzene and a diffusivity 12% lower than benzene, seemed to diffuse further and more uniformly than benzene. Figure 110 4b shows that after 24 hours, the isoconcentration lines were fairly symmetrical around the point source. The 1 mg/L isoconcentration line covered more than 85 % of the experimental setup volume after 24 hours but unlike benzene NAPL, the maximum concentration at the air­water interface was not greater than 50 mg/L (compare Figure 3c with Figure 4c). The 25 mg/L isoconcentration line were located further laterally but the downward spreading seemed to be less extensive which may be due to the bottom wall of the reactor. After 72 hours (see Figure 4d), the 25 mg/L isoconcentration line covered approximately 65% of the experimental semp. The heavier density but less volatile nature of chlorobenzene appeared to have an influence on the diffusion of the dissolved product resulting in a more symmetrical diffusion around the point source. Influence of air flow rate The impact of air flow rate on the distribution of benzene and chlorobenzene is presented in Figure 5 and Figure 7, respectively. The isoconcentration line without air flow and with an air flow rate of 27.5 mL/min were fairly similar (compare Figures 3b and 5b). However, with an air flow rate, the plume did not change much with time and diffusion of benzene was limited to the upper central part of the experimental semp (see Figure 5b and Figure 5c). The influence of air flow at the air­water interface was evident and can be seen to be controlling the migration of the LNAPL plumes. Even after 72 hours, the plume remained the same, indicating that a quasi steady­state condition was reached whereby a balance was achieved between the dissolution and diffusion of benzene and the volatilization at the air­water interface. With a higher flow rate (68 mL/min) the distribution of benzene isoconcentration lines was similar to that for an air flow rate of 27.5 mL/min (compare Figure 5b with Figure 6b). This shows that air flow rate had negligible influence on the distribution and dissolution of the NAPL. This may be due to the lack of physical contact between the NAPL and the advective air stream. Because of the lack of physical contact with the LNAPL, volatilization of benzene would be controlled by the diffusion of benzene. Therefore, as seen in Figures 6c and 6d, the distribution of the isoconcentration lines was similar to that of Figures 5c and 5d. The above results showed that when a LNAPL is located several centimeters from the air Ill channel. LNAPL volatilization was independent of the air flow rate. This behavior seemed to be in agreement with the work done by Johnson et al. (1997). Further evidence supporting this observation will be presented later. Similarly, as for benzene, air sparging seemed to control the spreading of chlorobenzene NAPL. Examination of the isoconcentration lines for 1 mg/L and 25 mg/L in Figures 4b (no air flow) and Figure 7b (with an air flow rate of 27.5 mL/min) showed that chlorobenzene diffused farther from the point source in the absence of air flow. Figures 7c and 7d showed that the 50 mg/L isoconcentration line for chlorobenzene remained fairly the same. At the air­water interface ±e concentration of chlorobenzene was lower with air flow. This is due to volatilization at the air­water interface. Influence of porous media The influence of porous media on the distribution of benzene can be seen by comparing Figure 5 and Figure 8. Figure 8 presents the result for sparging benzene NAPL at a flow rate of 27.5 mL/min with sand 70/100 as a porous media. Sand 70/100 has a smaller mean particle size (0.168 mm) as compared to sand 30/50 (0.305 mm) but with a higher porosity (0.4 as compared to 0.37 for sand 30/50). In general, the experimental results showed that benzene tend to diffuse further laterally and vertically downward from the air­water interface in sand 70/100 (Figures 8b and 8c) as compared with sand 30/50. Figure 8 clearly showed that the size of the porous media affected the diffusion of the benzene to the air­water interface and therefore resulted in less volatilization at the air­water interface. This situation resulted in more spreading of the dissolved phase away from the air­water interface. Further evidence in terms of mass removed, supporting this observation will be presented later in Figure 9. VOC removal efficiency Figures 9a and 9b present the concentration of benzene and chlorobenzene in the exhaust air for the experimental conditions tested. The influence of a porous media particle size on the volatilization is evident in Figure 9a. For sand 30/50, the peak benzene concentration appeared approximately 60 hours after air sparging started. In contrast, the peak benzene 112 concentration for the sand 70/100 was attained only after 90 hours of sparging. In addition, the peak concentration for sand 30/50 was almost 3 times larger than the sand 70/100. The peak concentration for an air flow rate of 68 mL/min was 3 times smaller than for an air flow rate of 27.5 mL/min. The chlorobenzene concentration in the exhaust air was lower than that for benzene even though the porous media and air flow rate were the same (see Figure 9b). The concentration of chlorobenzene in the exhaust air ranged between 0.08 mg/L and 0.12 mg/L after 30 hours of sparging. The lower chlorobenzene concentrations may be related to four possible factors: (i) the higher density of chlorobenzene, (ii) a lower solubility compared with benzene (515 mg/L against 1780 mg/L), (iii) a lower Henry's law constant (0.137 against 0.197), and (iv) a lower diffusion coefficient (8.52 x 10"® em's"' against 9.59 x lO"*^ cmV) Mass removal efficiency over time for the different experiments are shown in Figure 9c. For sand 30/50 with an air flow rate of 27.5 mL/min, mass removal efficiency after 120 hours was approximately 82%. With a higher air flow rate of 68 mL/min, an initial higher removal efficiency was obtained but after 144 hours the mass removal efficiencies for both air flow rates were similar. This result clearly shows that mass removal for NAPL located close to an air charmel was independent of the air flow rate but depends on the liquid diffusion of the dissolved phase. For a finer porous media (sand 70/100), the mass removed was only 34%. The increase in the diffusion pathway for benzene in a porous media with a smaller mean particle size may be a reason for the lower removal efficiency. For chlorobenzene the removal efficiency after 120 hours of sparging was found to be just 29%. These results along with the isoconcentration plots suggest that the solubility of the compound may also have an influence on the effectiveness of air sparging of NAPLs. The visualization study showed that LNAPLs and DNAPLs behave differently under air sparging conditions. The spreading of LNAPLs, in this case benzene, seemed to be better controlled by air sparging than DNAPL, in this case chlorobenzene. This behavior may be due to the differences in solubility and diffusivity between benzene and chlorobenzene rather than the differences in density. Results fi­om this study seemed to indicate that air flow may enhance diffusion and therefore indirectly affect NAPL dissolution but increasing the air flow rate did not increase the overall mass removal. 113 Conclusions The fate of NAPLs under air sparging conditions was investigated using a laboratory­scale single­air channel setup. The study showed that the presence of advective air flow in air channels during air sparging has an effect and may control the spreading of the dissolved phase of NAPLs. The control of the spreading and remediation of the contaminant plume seemed to be more effective for NAPLs with higher solubilities and diffusivities. Removal efficiency was seriously impacted by the grain size of the porous media. More than 50% reduction in the removal of benzene was found when the mean particle size for the porous media was reduced from 0.305 mm to 0.168 mm. Removal efficiency was found to be independent of air flow rate. This suggests that for a NAPL located several centimeters away from the air­water interface, diffusion of VOC will control the volatilization and remediation of the NAPL. In summary, this smdy showed that if a NAPL is not in direct contact with the air channel but several centimeters away firom the air channel, air sparging may be an effective remedial tool in controlling the size of the dissolved phase around the NAPL. References Braida, W.J., and S.K. Ong, Air sparging effectiveness: the air channel mass transfer zone, submitted to Water Resources Research, 1997a. Braida, W.J., and S.K. Ong, Air sparging: air­water mass transfer coefficients, submitted to Water Resources Research, 1997b. Conrad, S.H., J.L. Wilson, W.R. Mason, and W.J. Peplinski, Visualization of residual organic trapped in aquifers. Water Resources Research, vol. 28(2), pp. 467­478, 1992. Ji, W., A. Dahmani, D.P. Ahlfeld, J.D. Lin, and E. Hill, Laboratory study of air sparging: Airflow visualization, GWMR, vol. 13(4), pp. 115­126, 1993. Johnson, P.C., A. Das, R.L. Johnson, A. Leeson, D. McWhorter, and R. Hinchee, Effects of LVS process changes on the removal of immiscible­phase hydrocarbons, in Proceedings of the Four International In Situ and On­Site Bioremediation Symposium, vol. 1, pp. 135­140, New Orleans, LA, April 28­May 1", 1997. 114 Kennedy, A.C., and W. Lennox, A pore­scale investigation of mass transport from dissolving DNPL droplets. Journal of Contaminant Hydrology, vol. 24, pp. 221­246, 1997. McCray, J.E., and R.W. Falta, Numerical simulation of air sparging for remediation of NAPL contamination. Ground Water, vol. 35(1), pp. 99­110, 1997. Miller, C.T., M.M. Poirier­McNeil, and A.S. Mayer, Dissolution of trapped nonaqueous phase liquids: mass transfer characteristics. Water Resources Research, vol. 26(11), pp. 2783­2796, 1990. Nelson, D.W., and L.E. Sommers, Total carbon, organic carbon, and organic matter, in Methods of Soil Analysis, Part 2, Chemical and Microbiological Properties, Second Edition, edited by A. L. Page, pp. 570­571, American Society of Agronomy, Inc., Soil Science Society of America, Inc., Madison, WI, 1982. Powers, S.E., L.M. Abriola, and W.J. Weber Jr., An experimental investigation of nonaqueous phase liquid dissolution in saturated subsurface systems: steady state mass transfer rates. Water Resources Research, vol. 28(10), pp. 2691­2705, 1992. Powers, S.E., L.M. Abriola, and W.J. Weber Jr., An experimental investigation of nonaqueous phase liquid dissolution in saturated subsurface systems: transient mass transfer rates. Water Resources Research, vol. 30(2), pp. 321­332, 1994. Schroth, M.H., J.D. Istok, S.J. Aheam, and J.S. Selker, Geometry and position of light nonaqueous­phase liquid lenses in water­wetted porous media. Journal of Contaminant Hydrology, vol. 19, pp. 269­287, 1995. Wilkins, M.D., L.M. Abriola, and K.D. Pennel, An experimental investigation of rate­ limited nonaqueous volatilization in unsaturated porous media: steady state mass transfer. Water Resources Research, vol. 31(9), pp. 2159­2172, 1995. Zhou, D., and M. Blunt, Effect of spreading coefficient on the distribution of light non­ aqueous phase liquid in the subsurface, Journal of Contaminant Hydrology, vol. 25, pp. 1­19, 1997. 115 Table 1. Experimental matrix Experiment Number NAPL Aeration Rate (mL/min) Porous Media 1 Benzene Stagnant 30/50 2 Benzene 27.5 30/50 3 Benzene 27.5 70/100 4 Benzene 68 30/50 5 Chlorobenzene Stagnant 30/50 6 Chlorobenzene 27.5 30/50 i Sample Ports Air Channel Gap 1.58 mm Clean Humidified Air £ z: 4.80 cm> 2.54 cm 3.81 cm 3.81 cm to Gas Sampling Point 5.08 cm cm 2.54 cm Saturated Porous Media Contaminated with VOCs Contaminated Air 1.91 cm ~ Z 2.11 cm 5cm 17.5 cm l­igure 1, Single­air channel apparatu.s 117 Air­Water Interafce •15.4 *0.0 *0.0 •15.3 .1.0 28.8 ­50.­ ­ioa *20.4 *17.8 *16.3 *15.4 *13.8 *13.5 *0.C *0.0^ 175 Length (mm) Air­Water Interafce *0.0 1.94 *0.0 1.18 1.54 Z49 1.97 *0.88 *1.24 *0.47 *0.0 *0.74 *0.29 *0.0 ­10& 175 Length (mm) Figure 2. Isoconcentration lines (mg/L) and actual VOC concentrations, (a) chlorobenzene NAPL in sand 30/50 after 24 hours with no air flow, (b) benzene NAPL in sand 30/50 after 24 hours with an air flow rate of 68 mUmin Air­Water Interface OJ I I Air­Water Interlace I I I L 04 ­25­. ­2&. ­50.. ­50.. ­7S. ­7S. I U­. I (a) Time = 0 (b) Tinic = 24 h •lOOl I I I I J.I I I 1^1 I I 0 25 Length (mm) M Length (mm) Air­Water Interface OJ « ' ' I ' ' ((78^ ­25.. ­50.. ­75.. (c) Iinie48h ­10011 • 0 iXi • • 25 I • ' Length (mm) I'igiirc 3. Is(x;oncciitration lines (nig/I,) lor Ix­n/enc NAPl. in siuitl 30/50 iuid zero air How. 11 ,, Air­Water Interlace 04 I U Air­Water Interface ^ 04 I 4­^ I ­2S­ ­50­­ I ­7S­ (b) !inie=24h (a) 1 inie = 0 0 I I I iXl " . jpl I I 25 I I M I I M 0. .J­' 25 " 'ii" . 'is' " 'ilxj " ' i k ' ' ' i i c C ' ' ' Length (mm) Length (mm) Air­Water Interface Air­Water Interface 1.0 ­2S. ­2S­ ­50­. ­5a. ­7&. ­75.. (c) Tiine = 48 li •100],,>1 ,.t,,, 0 25 ,, ,^5,. 1 Length (mm) . I (d) I'inx: = 72 li , ­10a 'i'" 'i"''h'" 'ifirf" 'ik " Length (mm) i'iglue 4. Isoccinccntratidii lines (nig/I.) (or chlon)lx;n/.ene NACl. in siuid 30/50 iuid /jcri air (low. 175 Air­Water Interlace 04 1 Air­Water Interface J L 04 L ­25­. ­25­. ­50­• ­50­. ­75­. ­75.. I I... , I (b) Time = 24 h (a) Time = 0 •''"I' ' '^6' " 'Al" ' 'k " 'itrf " 'lk " " '175 Length (mm) Length (mm) Air­Water Interface Air­Water Interlace oi 1 ­25­• ­5a. ­75.. (ci) I inK' ­ 72 li (c) rime = 48 h IJ . , , • ' ^5' " Length (nrm) • • 'iJd • • ­100 io''' ' i s ' ' "iM " Sk '' 'i^ Length (mm) I'igiuc 5. Isoconccntralion lines lor bcn/cne NAPl. in siuid 30/30 anil for jui air How rate of 27.5 ml ^min. Air­Water Interface I I 0­1 J L I L 04 Air­Water Interface I I I (17^ ­25­­ ­50­ ­50.. ­75­ ­7S. (b) Tinie = 24 h (a) Time = 0 Length (nm) Length (nrm) Air­Water interface 04 Air­Water Interface 1 Oi a ­2&. I .1... J L.,^ .1... I 8^ ­50­­ ­7&. (c) I inic = 48 h . . .^5. • • Length (mm) • • 'iJtf • • 'lis (d) nnie = 72 h 'i'' "i'" '^5'" 'i4d" ' "lid" '1I5 Length (mm) I'igurc 6. l.S(K'()ncenlnili()ii lines for ben/cne NAPI, in Siuici M)/5i) for ;ui air How rate of 68 ml ymin. Air­Water Interface Air­Water Interface ­25­. ­2S. ^ ­50­• ­75­. (b) I'inie = 24 li ­100. 175 ­loa T75 Length (mm) Length (mm) Air­Water Interface Air­Water Interface ­25­. (d) finie = 72 h ­100­ 175 Length (mm) ­100­ 175 Length (mm) I'igiirc 7. ls(x;onccntration lines (nig/L) for chloroben/ciie NAPI, in siuid 30/50 luid foriui air How rate of 27.5 niiyniin. Air­Water Interface 04 1 Air­Water Interface L Oi M ­2S. 1 1 ­25­. ­50­ ­5a­ ­7S­ ­7S­ (a) Tinie = 0 0 , , 25 ^5,,, , ,^5,.. Length (mm) (b) Tinie = 24 h ­ n •'"js" •'jb" "A" Length (mm) Air­Water Interlace 10 u» Air^Wcitcr IntcrUicc ­25­­ ­25­ ­50­­ ­50­ ­7S­ ­75­ (c) Tinie = 48 h 0' "s ' 25 ' " ' J b " ' ' k " 'iid " Length (mm) (<J)linxi^72 h " 'lirf " Length (mm) l igiue 8. Isociincentration lines Cor Ix'n/one NAl'l. in siuid 70/KK) and lor im air llow rate of 27.5 nil /niin. 124 (i) Sand 30/50, 27.5 mL/min (ii) Sand 70/100, 27.5 mL/min (HI) Sand 30/50, 68 mL/nun 0.1 ­ Sand 30/50. 27.5 mL/min 0.3 ­ 0.2 ­ 0.1 ­ 75 ­ Benzene, sand 30/50, 27.5mL/min Benzene, sand 70/100, 27.5 mL/min Benzene, sand 30/50, 68 mL/min Chlorobenzene, sand 35/50, 27.5 mL/min 50 ­ (c) ­1 I 50 Time (h) 100 r 150 Figure 9. VOC concentration in the exhaust gas and VOC removal efficiency 125 CHAPTER SEVEN. INFLUENCE OF SORPXION­DESORPTION PROCESSES ON AIR SPARGING EFFECTIVENESS A paper to be submitted to Journal of Contaminant Hydrology Washington J. Braida, and Say Kee Ong Abstract Air sparging is a remediation technology which has been used for more than 10 years for ±e remediation of VOC­contaminated aquifers. Although it has been applied extensively in the field, an understanding of the controlling processes during air sparging is lacking. Using a single­air channel apparatus, the size of the air channel mass transfer zone {MTZ) was found to decrease with increasing organic carbon content of the porous media. This effect was larger for VOCs with low solubilities and high partition coefficients. A one­dimensional diffiision model, modified to include the retardation in the VOC transport as a result of partitioning between the liquid and solid phases, was found to predict the concentration curves for the liquid phase and the exhaust air fairly well. Experimental evidence suggest that sorbed VOCs were not substantially remediated using air sparging. Introduction Air sparging is a remediation technique which is potentially applicable for the remediation of aquifers contaminated with volatile organic compounds (VOCs). Air sparging involves the injection of air under pressure into the saturated zone to volatilize VOCs and at the same time, promote the biodegradation of contaminants present beneath the water table and in the unsaturated zone. Much of the information on the volatilization of VOCs from bulk water may be applicable for air sparging systems. However, the presence of porous media introduces several barriers which may affect the volatilization of VOCs. These barriers include the extra distance VOCs must travel to reach the air­water interface, and the sorption­desorption of the VOCs onto the soil matrix. Braida and Ong (1997a) showed experimentally that mass transfer under air 126 sparging conditions was diffusion limited and that after a certain period of time, a mass transfer zone (M7Z) with a steep concentration gradient developed around the air channels. Beyond the MTZ, the VOC concentration was almost unaffected by the air flow. For soil media with high VOC af^inity, sorption­desorption of VOCs onto the soil particles may also play a role in controlling the rate of mass transfer. Sorption of VOCs onto soil may be modeled using a linear equilibrium isotherm: Cs=KdC^­ (1) where Cs represents the concentration of the VOCs sorbed onto the soil matrix (MM"'), C. is the concentration of VOC in the liquid phase (ML"^), and Kd is the linear partition coefficient (L^M"'). The linear isotherm has been widely used to describe the sorption of VOCs in the subsurface (Gierke et al., 1990, Ong and Lion, 1991). Karickhoff (1984) suggested that use of linear isotherms is a reasonable assumption for contaminant concentrations less than 10'^ M or less than half the water solubility of the organic contaminant. Sorption of VOC onto soils has been related with the organic matter content of the porous media as follows (Karickhoff, 1984): Kd=focKoc (2) where/oc is the mass fraction of organic carbon in the soil and Koc is the organic carbon partition coefficient (L^M''). Equation 1 is valid forfac> 0.001. Most organic contaminants are rapidly sorbed onto soils but desorption can be very slow. Batch desorption experiments by Pavlostathis and Mathavan (1992) showed that a substantial fraction of sorbed contaminant resisted desorption in deionized water­soil mixtures for more than 7 days. The sorption­desorption phenomenon tends to reduce the overall removal efficiency of remediation technologies such as air sparging. Air sparging smdies by Chao (1997) using bench­scale soil columns showed that after the VOC concentration in the exhaust gas fall below detection limits, a large fraction of initial mass of VOCs remained sorbed to the 127 porous media. In an air sparging smdy at a site contaminated with ethylbenzene and xylenes. Kraus et al. (1997) reported that even though the aqueous phase VOC concentration in monitoring wells was reduced by four orders of magnitude, analysis of the soils indicated that there was no statistically significant reduction of ethylbenzene and xylenes. The objective of this study was to assess the influence of the sorption phenomenon on the performance of air sparging systems. A single­air channel experimental setup was used to study the influence of the organic carbon content of the porous media on the effectiveness of air sparging by evaluating changes in size of the MTZ and mass transfer parameters. Materials and Methods To investigate the influence of the sorption­desorption phenomenon under air sparging conditions, a single air channel setup was used (Braida and Ong, 1997a, 1997b, and 1997c). The experimental setup consisted of a box made of thick Plexiglas ^sheets in which air was circulated in a single air channel over saturated porous media. The size of the air channel was fixed at 1.58 mm. The experimental setup reproduced the physical nature of the air flow in saturated porous media and allowed a visualization of the VOC concentration profiles near the air channel. In order to monitor the aqueous phase concentration of the VOCs in the porous media, 15 sampling points across the porous media were included. A sampling point to measure the VOC concentration in the exhaust gas was also included. A sketch of the experimental setup is shown in Figure 1. In­house compressed air was filtered and then humidified before being used as sparging gas. Air flow was measured using a Gilmont Model 11 flowmeter (Barrington, IL). The experiments were conducted using three VOCs, and type of porous media with three different organic carbon contents. The VOCs used were benzene, ethylbenzene, and n­propylbenzene. Ottawa sand (U.S. Silica Company, Ottawa, IL) was used as porous media with an organic carbon content of less than 0.01%. The mean particle size of the Ottawa sand was 0.190 mm with an uniformity coefficient of 2.16. Using Ottawa sand, two other organic contents were prepared by coating the Ottawa sand with humic acid (Sigma­Aldrich Chemical Co, Inc., Milwaukee, WI). The organic carbon contents of the porous media were 0.04%a and 0.45%. The coating procedure followed that of Garbarini and Lion (1985). Eleven grams of humic 128 acid were dissolved in one liter of Nanopure™ water at a pH of 10. The pH was raised with a 0.25M NaOH solution. About 950 grams of Ottawa sand was added and the mixture was stirred continuously with the pH of the mixture gradually lowered to 4.0 with 0.25 M nitric acid. The suspension was allowed to stand for 48 hours after which the supernatant was decanted and the porous media rinsed several times with Nanopure™ adjusted to pH 4. The porous media was then air dried. Organic carbon content of the sand was determined using the Walkley­Black procedure (Nelson and Sommers, 1982). Partition coefficients of the three VOCs for the three porous media were determined using the head­space technique described by Garbarini and Lion (1985). Linear partition coefficients ranged between 0.025 mL/g and 1.00 mL/g and are presented in Table 1. Aqueous solutions containing the three VOCs were prepared using HPLC grade chemicals and their concentrations ranged between 4 mg/L and 76 mg/L. All chemicals were purchased from Sigma­Aldrich Chemical Company Inc. (Milwaukee, WI). A slurry was made by carefully mixing the porous media with the aqueous VOC solution and allowed to equilibrate for 16 hours. After the equilibration time, the slurry was transferred to the reactor and was rapidly packed layer by layer to avoid entrapment of air bubbles. The reactor was sealed immediately to minimize VOCs losses. A small quantity of the slurry was retained and analyzed for VOCs. The resulting porosity of the packed media was 0.377. Aqueous samples were then taken from the various sampling points to determine the initial aqueous phase concentrations in the porous media. Air was then circulated through the air channel. The air velocity used was 1.1 cm/s. The temperamre of the experimental setup was maintained at (21 + 2)°C. The VOC concentrations in the air phase and in the liquid phase were measured with a Hewlett­Packard 5890 Series II gas chromatograph (Avondale, PA) equipped with a HP­5 capillary column and a flame ionization detector. Air phase concentration was determined by direct injection of a 1 mL sample while liquid phase concentration was measured using the head­space technique. For the head space technique, 25 |iL of an aqueous sample was placed in a 1.8 mL aluminum crimp cap vial and the aqueous concentration estimated from the measured head space concentration after equilibrium was reached. At the end of the experiments, samples of the porous media samples were collected from 129 different depths in the reactor and analyzed for sorbed VOC concentration. VOCs in the porous media was extracted using hexane. A 1:3 (w/w) mixture of hexane and wet soil was used with an extraction time of 12 hours. An aliquot of the supernatant was placed in a 1.8 mL glass vial capped with PTFE­faced silicone septa. VOCs in the sample were quantified by direct injection of the supernatant into the gas chromatograph. Mass of VOC sorbed on the porous media was estimated by subtracting the mass in the aqueous phase remaining in the wet soil from the mass extracted from the wet soil. Results and Discussion Figure 2 shows the change in the relative concentration of benzene, ethylbenzene, and n­ propylbenzene in the exhaust air for the three porous media. The relative air phase concenUration of benzene were not significantly affected by the presence of organic matter in the porous media. However, the different air phase concentration profiles for ethylbenzene and n­propylbenzene indicated that volatilization of these compounds were affected by the organic carbon content in the porous media. N­propylbenzene which has the lowest solubility and the highest water­solid partition coefficient of the three VOCs showed the most variation in the exhaust air concentration. Table 2 presents the concentration of VOCs sorbed initially and after air sparging. As seen in Table 2, the concentration of VOCs sorbed to the porous media remained fairly constant even after 10 hours of sparging and that the mass sorbed were fairly the same throughout the depth of the porous media. This observation seemed to suggest that desorption was negligible over the experimental period. However, desorption kinetics may play an important role in the remediation of VOC contaminated aquifers once the dissolved VOCs are removed. This behavior seemed to reflect the work reported by ECraus et al. (1997) where the soil concentrations before and after air sparging were the same even though the VOC concentrations in the groundwater were reduced by 3 to 4 orders of magnitude. Rebound in the aqueous concentration of VOCs after the air sparging system is turned off may be due to the slow desorption of VOCs. Concentration profiles in the aqueous phase for each VOC are presented in Figures 3 to 5. Measurement of initial aqueous concentrations indicated that at the start of the experiment 130 there was a concentration gradient near the air­water interface as a result of losses of VOCs during the packing of the reactor. Separate experiments indicated that when the experimental setup was left alone for 18 hours without any air flow, VOC losses, other than losses due to the initial packing were found to be less than 2%. When air was passed through the reactor, the VOC aqueous concentration near the air­water interface was rapidly depleted. This depletion was confined to a thin layer of porous media near the air­water interface. The rapid depletion is associated with a faster volatilization of VOC at the air­water interface as compared to the diffusive transport of VOCs to the air­water interface. After four hours of sparging the concentration gradient of the VOCs in the liquid phase became fairly constant suggesting that a quasi­steady state condition for the diffusion of the VOCs was reached. The distinctive zone with a steep concentration gradient may be defined as the mass transfer zone (MTZ) (Braida and Ong, 1997a). Beyond the MTZ, the impact of the air channels was strongly limited. For convenience, the size of the MTZ was taken as the distance from the air­water interface to where the VOC concentration was equal to 90% of the bulk concentration. Figures 3,4 and 5 showed the influence of the organic carbon content on the size of the MTZ for benzene, ethylbenzene, and n­propylbenzene, respectively. Benzene, the most soluble and least sorbable of the three VOCs studied, showed the smallest change in the size of the MTZ (see Figure 3). The size of the MTZ for benzene varied from approximately 35 mm to 26 mm when the organic carbon content increased from 0.01% to 0.45%. The size of the MTZ for ethylbenzene ranged from approximately 38 mm for organic carbon content less than 0.01% to 28 mm for an organic carbon content of 0.45% (see Figure 4). The size of the MTZ for n­propylbenzene showed the largest variation with the organic carbon content. For this VOC, the size of the MTZ ranged from approximately 37 mm for the sand with less than 0.01% organic carbon content to approximately 17 mm for a sand with an organic carbon content of 0.45% (see Figure 5). Based on Figures 3 to 5, the size of the MTZ was most affected for the VOC with the lowest solubility and the highest liquid­solid partition coefficient. Transport of VOCs to the air­water interface is retarded by the partitioning of the VOC between the liquid phase and the solid phase. This retardation is given by the retardation 131 factor R : (3) R = 1 + Adsorbed concentration Dissolved concentration Equation 3 may be written in terms of the linear partition coefficient Kd (L^M"'), the bulk density of the porous media pb (ML'^), and the porosity of the porous media £ (dimensionless) as follows: (4) R= l+ K,^ £ Using the equilibrium partition coefficients determined earlier, the retardation factor, R, for the experimental conditions tested may be computed as presented in Table 3. Transport of VOCs from the liquid phase to the air­water interface may be described using a one­dimensional diffusion model (Braida and Ong, 1997c). The model is as follows: where z (L) is the depth of the porous media with the air­water interface located at z = 0 and the bottom of the reactor at z = L, is the aqueous concentration of the VOC (ML"^), Dw is the aqueous diffiasivity of the VOC (L^"'), £is the porosity of the porous media (dimensionless), and ris the tortuosity factor (dimensionless) which accounts for the change in the length of the diffusion path of the VOCs in the porous media. In the presence of partitioning, this model may be modified by including the retardation factor as shown in equation 6. (6) 132 The initial condition (IC) and the boundary conditions (BC) are given by: IC Cw is known for all z at time zero (experimental data) BC R at2 = 0forallt dz 3C —— = 0 at £ = L for all t (7) (8) dz where Q, is the volumetric weighted average VOC concentration, of the mass transfer zone (MTZ). The volumetric average concentration, Q., was computed as follows: _ YV C (9) where V„ is the volume of the n layer (L^), Cn (ML"^) is the aqueous concentration of VOC in the n layer, and the summation extends to all the layers included in the MTZ. Ca is the air phase concentration close to the air­water interface and was assumed to be zero. The air phase concentration, , in the exhaust air (outlet of the reactor) was computed using the transport equation for the VOCs in the air phase: K^=­Q.C.*IC^ak„C^. (10) where Q is the measured VOC concentration in the effluent air (ML'^), Va represents the air volume (L^), Qa represents the air flow (L^T*), A is the air­water interfacial area (L") of the experimental setup (87.5 cm"). Equation 6 and equation 10 were solved numerically using the Continuous System Modeling Program (CSMP) software (IBM, 1972). The porous media profile was divided into 26 horizontal layers of variable thickness (thinner layers were used closer to the air­water interface to compute more accurately the VOC concentrations in the region where the largest changes in the concentration occurred). To solve the equations. 133 used closer to the air­water interface to compute more accurately the VOC concentrations in the region where the largest changes in the concentration occurred). To solve the equations, the value for the mass transfer coefficient (ATc) and the tortuosity factor (t) were taken from Braida and Ong (1997c) while the retardation factor used are presented were in Table 3. Table 4 lists the values of the parameters used in the model for the estimation of the aqueous phase concentration and the air phase concentration. Figure 6a and Figure 6b showed the experimental and predicted liquid phase concentrations for the three VOCs in the same porous media with different organic carbon contents. The inclusion of the retardation factor in the diffusion model seemed to predict fairly well the concentration profiles after 10 hours of sparging. The experimental and predicted exhaust air concentrations for benzene are presented in Figures 7a and 7b. As seen in Figures 7a and 7b, the model slightly overpredicted the air phase concentration. The model can be assessed further by comparing the actual mass volatilized and the predicted mass volatilized. Depending on the VOC, the predicted total mass volatilized was approximately 25 to 35% larger than the actual mass volatilized. It must be pointed out that the predictions were made based on mass transfer coefficients determined for a porous medium with less than 0.01% organic carbon and by assuming equilibrium partitioning. The discrepancy in the predictions of the model and the actual values may be due to the above two assumptions. If the mass transfer coefficients were assumed to be constant for a given porous medium but regardless of the organic carbon present, then the overprediction of the mass volatilized may be due to the equilibrium assumption of VOC partitioning. However, by assuming equilibrium partitioning, sorption would be maximized and therefore mass volatilized would be minimized. It is therefore possible that the overprediction may be due to the selection of the mass transfer coefficients. Instead of using the mass transfer coefficients for porous medium with less than 0.01% organic carbon, the model may be used to determine the mass transfer coefficients by curve fitting the model predictions with the actual aqueous phase and air phase concentrations and the total mass volatilized (Braida and Ong, 1997b). If this approach was taken, the estimated mass transfer coefficients were found to be approximately 30­35% smaller than the mass transfer coefficients for porous medium with less than 0.01% organic carbon. Possible 134 to the presence of the organic matter. In addition to the above possible reasons, the tortuosity factor used could be different due to the higher organic matter present. Conclusions The effects of the adsorption of VOCs on the performance of air sparging systems were studied using a single­air­charmel experimental setup. The size of the MTZ was found to be affected by the amount of organic carbon present in the porous media. The increase in the organic carbon content of the soil resulted in a decrease in the size of the MTZ. This effect is larger for VOCs such as n­propylbenzene with low solubilities and high partition coefficients. The decrease in the size of the MTZ may be due to the retardation in the diffusion of the VOCs to the air­water interface as a result of aqueous­solid partitioning. Sorbed VOC concentration remained fairly constant along the soil profile during the sparging process. This finding seems to indicate that slowly desorption kinetics may seriously limit the remediation of contaminated aquifers. A one­dimensional diffusion model that included the retardation factor (/?) was found to predict fairly well the liquid and air phase concentrations. This show that equilibrium partitioning and a first order volatilization term at the air water interface may be used to predict concentrations and mass changes during air sparging. Notation Ca C» VOC concentration in the air and aqueous phases, ML VOC weighted concentration average in the MTZ, ML' C. VOC measured concentration in the air phase C5 VOC concentration sorbed onto soil matrix, MM' D^. VOC aqueous diffusivity, L^T'' Da VOC air diffusivity, L^T'' foe organic carbon fraction, dimensionless e porous media porosity, dimensionless T tortuosity coefficient, dimensionless Koc organic carbon partition coefficient, L^M' 135 Kd linear partition coefficient, L^M'' KH Henry's law constant, dimensionless MTZ width of the mass transfer zone, L Qa air flow rate, L^T' Va air volume, R retardation factor, dimensionless Kc overall ah" side mass transfer coefficient, LT References Braida, W.J., and S.K. Ong, Air sparging effectiveness: the air channel mass transfer zone, submitted to Water Resources Research, 1997a. Braida, W.J., and S.K. Ong, Nonaqueous liquid phase fate under air sparging conditions, submitted to Journal of Contaminant Hydrology, 1997b. Braida, W.J., and S.K. Ong, Air sparging: air phase­liquid phase mass transfer coefficients, submitted to Water Resources Research, 1997c. Chao, K.P., Aqueous­vapor mass transfer of VOCs in saturated porous media under air sparging conditions, Ph.D. Dissertation, 176 pp.. Polytechnic University, Brooklyn, NY, 1997. Garbarini, D.G., and L.W. Lion, Evaluation of sorptive partitioning of nonionic pollutants in closed systems by headspace analysis. Environ. Sci. TechnoL, vol. 19(1), pp. 1122­1129, 1985. Gierke, J.S., N.J. Hutzler, and J.C. Crittenden, Modeling the movement of volatile organic chemicals in columns of unsaturated soil. Water Resources Research, vol. 26(7), pp. 1529­ 1547, 1990, IBM, Systeni/360 Continuous System Modeling Program user's manual. Program Number 360A­CX, 5th ed., IBM Corp., Technical Publ. Dept., White Plains, New York, 1972. Karickhoff, S.W., Organic pollutant sorption in aquatic systems, J. Hydraulic Eng., ASCE, vol. 110, pp. 707­735, 1984. 136 Kraus, J., S. Nelson, P. Boersma, and A. Maciey, Comparison of pre­Zpost­sparging VOC concentration in soil and groundwater, In Situ and On­Site Bioremediation: Volume I, B. Alleman and A. Leeson (eds.), Battelle Press, Columbus, OH, 1997. Nelson, D.W., and L.E. Sommers, Total carbon, organic carbon, and organic matter, in Methods of Soil Analysis, Part 2, Chemical and Microbiological Properties, Second Edition, edited by A. L. Page, pp. 570­571, American Society of Agronomy, Inc., Soil Science Society of America, Inc., Madison, WI, 1982. Ong, S.K., and L.W. Lion, Effects of soil properties and moisture on the sorption of TCE vapor. Water Research, vol. 25(1), pp. 29­36, 1991. Pavlostathis, S.G., and G.M. Mathavan, Desorption kinetics of selected organic compounds from field contaminated soils. Environ. Sci. and Tech., vol. (26), pp.532­538, 1992. 137 Table I. Linear partition coefficients {Kd) in niL/g for different VOCs into Ottawa sand as a function of organic carbon content. VOC Organic Carbon Organic Carbon Organic Carbon <0.01% 0.04% 0.45% Benzene 0.054 0.226 0.228 Ethylbenzene 0.029 0.464 0.877 n­Propylbenzene 0.023 0.474 1.001 Table 2. Sorbed VOC concentrations (mg/Kg) ± standard deviation Organic Carbon Sample Benzene Ethylbenzene Content Description Initial Average 2.9 ±0.1 i.0±0.1 n­Propylbenzene 0.5+0.1 Final, Surface 2.9±0.1 1.0±0.1 0.4±0.1 Final, Depth 38.1 mm 3.0+0.1 1.0±0.1 0.5±0.1 Final, Depth 63.5 nmi 2.9±0.1 1.0±0.1 0.5± 0.1 Final, Depth 88.9 mm 3.0+0.1 1.0±0.1 0.5±0.1 Initial Average Final, Surface Final, Depth 19.1 mm 3.7 + 0.1 3.6 + 0.2 3.7+.0.1 1.3 ±0.1 1.3 ±0.1 1.4 ±0.1 0.3 + 0.1 0.3 ±0.1 0.3 + 0.1 Final, Depth 38.1 mm 3.7 + 0.1 1.3 ±0.1 0.3 ±0.1 Final, Depth 63.5 mm 3.7 + 0.1 1.3 ±0.1 0.3 ±0.1 Final, Depth 88.9 mm 3.7 + 0.1 1.3±0.1 0.3 ±0.1 Computations: C (sorbed) = C{measured in hexane){Vol. of hexane)­C(water){Vol. of water in sample) Dry Mass (sorbent) 138 Table 3. Retardation factors { R ) for different VOCs on Ottawa sand as a function of organic carbon content. VOC Organic Carbon Organic Carbon Organic Carbon <0.01% 0.04% 0.45% Benzene 1.24 1.99 2.00 Ethylbenzene 1.13 3.03 4.84 n­Propylbenzene 1.10 3.07 5.38 Table 4. Parameters used in the model Parameter benzene ethylbenzene 0.377 Porosity (e) 0.377 Tortuosity (T) 0.52 0.52 KG (cm/min) 0.00486 0.00436 Dw (cm* /min) 0.0005754 0.0004578 Qa (mlVmin) 55 55 0.197 0.294 KH R(0c<0.01%) 1.13 1.24 R (OC = 0.04%) 3.03 1.99 R (OC = 0.45%) 2.00 4.84 propylbenzene 0.377 0.52 0.00463 0.0004194 55 0.369 1.10 3.07 5.38 / Sample Ports Air Channel Gap 1.58 mm Clean Humidified Air Contaminated Air to Gas Sampling Point Saturated Porous 11 cm Media Contaminated with VOCs 17.5 cm l­'igurc I. Singlc­air channel apparatus U» 140 OC <0.01% OC= 0.04% OC=0.45% 0.8 u u "u" c u C 0.6 04 0.2 0.0 1.0 OC <0.01% 0C= 0.04% OC=0.45% 0.8 0.6 0.4 0.2 0.0 0.8 0.6 0.4 o. 0.2 0.0 400 Time (min) 200 600 Figure 2. Relative VOC concentrations in the exhaust air for different organic carbon contents 141 Air­Water Interface T = Oh T = 8h ­20 T = 4h f= P a. (U Q ­60 ­80 Organic Carbon < 0.01% ­100 T = Oh T = lOh T = 4h ­20 ­40 a. u Q ­60 ­80 Organic Carbon = 0.04% ­100 T = Oh T = lOh ­20 E E ­40 Qu u Q ­60 ­80 ­100 T = 4h Organic Carbon = 0.45% I I I I I I I I—r I I I I I I I—r 0.0 0.2 0.4 0.6 0.8 1.0 Relative VOC Concentration (C/Cj,^y.) Figure 3. Benzene concentration profiles for different organic carbon contents 142 0 ­20 Air­Water Interface T = 4h T = Oh T = 8h ­40 a. ­60 ­80 Organic Carbon < 0.01% ­100 0 = 0h T= 10 h T = 4h ­20 ­40 ­60 Organic Carbon = 0.04% ­80 ­100 0 T = Oh T= lOh ­20 T = 4h ­40 ­60 Organic Carbon = 0.45% ­80 ­100 0.0 0.2 0.4 0.6 0.8 1.0 Relative VOC Concentration (C/Cj,^jj.) Figure 4. Ethylbenzene concentration profiles for different organic carbon contents 143 Air­Water Interface T= Oh T = 8h ­20 T = 4h S £ ­40 u" ­60 Organic Carbon < 0.01% C ­80 ­100 = 0h T= lOh T = 4h ­20 p p ­40 ­60 Organic Carbon = 0.04% ­80 ­100 T=a T= lOh ­20 p P & Q T = 4h ­40 ­60 Organic Carbon = 0.45% ­80 ­100 0.0 0.2 0.4 0.6 0.8 1.0 Relative VOC Concentration (C/C^^,^) Figure 5. n­Propylbenzene concentration profiles for different organic carbon contents 144 Air­Water Interface (a) ­20 ­ E t= ­40 ^ ­60 Ethylbenzene Propylbenzene ­80 ­ I I I I—r I I r­| I I I I ­100 0 10 20 30 40 50 VOC Cone. (mg/L) Air­Water Interface ­20 ­ £ S ­40 «u Q ­60 Ethylbenzene Propylbenzene ­80 ­ ­100 1 r I I 'I I I I I I I I I I I 'I I I 20 30 40 50 60 I I I I I I I I I 0 10 VOC Cone. (mg/L) Figure 6. Experimental and modeled VOCs concentration profiles in the water phase; (a) organic carbon content 0.04%, (b) organic carbon content 0.45% 1\ I 145 0.75 (a) ^ p 0.50­if Experimental Data Model u c 0 U u 1 0.25 H c u CQ 0.00 %O 0 1.00 O O O O O O O O O l l l l | I M I | M r l | I M I | T M T | I M i y i l l l 100 200 300 400 500 600 700 Time (min) X 0.75 ­ (b) Experimental Data Model P £ u c Vs c 0.50 H 0.25 ­ .'J 0.00 I1I O I I IO I I I IOl i tOI I IQI T oI I r nI I IoI I Tn r i 9 i I I I 0 100 200 300 400 500 600 700 Time (min) Figiire 7. Benzene concentration in the exhaust air, (a) organic carbon content 0.04%, (b) organic carbon content 0.45% 146 CHAPTER EIGHT. MODELING AIR SPARGED SOIL COLUMNS Introduction Several researchers have attempted to model air sparging operations by using different mathematical modeling approaches. Of particular interest are the models which assumed air flow in the form of finger­like air channels (Hein, 1996; Ji, 1994; Drucker and Di Julio, 1996). Most of the models are limited in their application since they were optimized by curve fitting with the experimental results. However, application of a model would require "a priori" selection of appropriate model parameters and this had been the hardest task faced by modelers. Volatilization of VOCs under air sparging conditions occurs at the air­water interface located at the air channels wall. Using a single­air channel experimental setup, Braida and Ong (1997a, 1997b) showed that, at the air channel level, the volatilization of VOCs during air sparging was controlled by the aqueous diffusion of the VOCs through a mass transfer zone adjacent to the air channels. In Chapter 3, a one­dimensional diffusion model was used to estimate the tormosity and the mass transfer coefficients for the volatilization of VOCs at the air­water interface during air sparging operations. In this chapter, the one­dimensional diffusion model developed in Chapter 3 for a single air­channel was modified from Cartesian to radial coordinates and applied for air sparged soil columns. The objective of this study was to assess whether the various parameters determined by using a single air­channel may be applied to a more complex system with many air channels such as air sparging of a soil column. Successful application of the model and the correlations developed in Chapter 3 will strongly suggest that the various physical­ chemical phenomena observed at the air charmel level or "microscale" level may be used to explain and predict the performance of large scale air sparging systems. Materials and Methods A Plexiglass™ column, 13.97 cm (5 1/2 inches) in diameter and 55.88 cm (22 inches) in length, was used for the soil column (see Figure 1). The column contained approximately 35 147 Sampling Ampoules Exhaust Exhaust Air to Sampling Ampoules Porous Media (35­38 cm) I Humidified Air Figure 1. Sketch of soil colunan 148 cm of porous media sandwiched between layers of glass beads. The bottom layer was 3­5 cm deep and contained 3 mm diameter glass beads. Air was introduced into the column through a 1.1 cm diameter cylindrical air stone (Nature's Playland Co., Hicksville, NY) located at the base of the column. The bottom layer of glass beads provided an even distribution of air across the soil column. The top layer of glass beads consisted of a 3 ­ 5 cm lower layer of 3 mm diameter glass beads and a upper layer of 12 ­ 14 cm of 5 mm diameter glass beads. The upper layer provided an overburden pressure onto the porous media and minimized the fluidization of the porous media during sparging. The porosity of the glass beads were 0.40 and 0.42 for the 3 mm and 5 mm diameter glass beads, respectively. The exhaust air line was a PTFE hose with a pair of 125 mL gas sampling ampoules (Alltech, Deerfield, IL). In­house humidified air was used as the sparging air. Air flow was measured using a Gilmont Model 13 flowmeter (Harrington, EL). Three sets of experiments with three different porous media were conducted. The porous media used were: (a) sand 30/50 with a porosity of 0.377, (b) sand 70/100 with a porosity of 0.40, and (c) Ottawa sand with a porosity of 0.37 (U.S. Silica Company, Ottawa, IL). The VOCs used were benzene, ethylbenzene, n­propylbenzene with sand 30/50 and sand 70/100, while toluene, chlorobenzene, styrene, and 1,2 dichlorobenzene were used with Ottawa sand. Concentrations ranging between 0.7 and 100 mg/L. All chemicals were HPLC grade and were purchased from Sigma­Aldrich Chemical Company Inc. (Milwaukee, WI). To pack the colunm with porous media, the 3 mm glass beads bottom layer, saturated with contaminated water, was first put in place. A slurry was made by carefully mixing the porous media with the aqueous solution of the VOCs. The column was then rapidly packed layer by layer with the slurry to avoid entrapment of air bubbles. Aqueous samples were taken from various depths along the packing to determine the initial aqueous phase concentration in the porous media. Slurry samples were also taken during packing to determine the mass sorbed onto the porous media. The composite top layer of glass beads was put in place and the column was inmiediately sealed to minimize VOCs losses. The column was allowed to equilibrate for 16 hours before being sparged. The air flow rates were 0.8,0.7, and 1.1 L/min for sand 30/50, sand 70/100, and Ottawa sand, respectively. The air saturation in the soil was estimated by measuring the volume of 149 water displaced into the top layer of glass beads. For the above air flow rates, the air saturation were 10%, 6.3%, and 8.1% for sand 30/50, sand 70/100, and Ottawa sand, respectively. The average air pore velocities may be estimated from the air flow rate divided by the cross sectional area of the column and the air saturation factor. The estimated air pore velocities were 0.86 cm/s for sand 30/50, 1.21 cm/s for sand 70/100, and 1.34 cm/s for Ottawa sand. This range of velocities was within the range of velocities for which the mass transfer zone and mass transfer coefficient correlations were determined and was similar to air velocities reported by others (Hein, 1996). Stagnant water conditions were assumed and the water temperature was maintained at 21 + 1 °C. At the end of the experimental run, both aqueous and wet porous media samples were taken and analyzed for the VOCs remaining. A total of 16 samples were taken at depths of 0, 10,20, and 30 cm from the top of the porous media (2 aqueous and two wet porous media samples per depth). VOC concentrations in the exhaust air were measured by withdrawing air samples from the ampoules. The initial and final VOC concentrations in liquid phase were measured with a Hewlett­Packard 5890 Series II gas chromatograph (Avondale, PA) equipped with a HP­5 capillary column and a flame ionization detector. Air phase concentration was determined by direct injection of a 1 mL sample into the gas chromatograph. The liquid phase concentration was determined using the head space technique where 25 {iL of an aqueous sample was placed in a 1.8 mL aluminum crimp cap vial. After equilibrium was reached, the head space concentration was measured by direct injection into the gas chromatograph. The aqueous concentration was then estimated from the measured head space concentration. To perform a mass balance, VOCs sorbed to the porous media were determined using a hexane extraction procedure at the end of each experimental run. A 1:3 ratio (w/w) of hexane and wet soil was mixed for 12 hours in a wrist shaker (Burrell Scientific, Pittsburgh, PA). The solids were then allowed to settle. A portion of the clean supernatant was placed in a 1.8 mL glass vial capped with PTFE­faced silicone septa. VOCs in the sample were quantified by direct injection of the supernatant into the gas chromatograph. The mass of sorbed VOC was estimated by subtracting the mass of VOC present in the liquid phase remaining in the wet soil from the mass of VOC present in the hexane extract. 150 Description of Model Under air sparging conditions, the air channels in the soil column may be represented by a composite of evenly spaced cylindrical air channels surrounded with a nonadvective water saturated region as seen in Figure 2 (Hein, 1996; Ji, 1994; Dmcker and Di Julio, 1996). For convenience, the circumference of the nonadvective region for each air channel was assumed to touch the circumference of the adjacent nonadvective regions. As presented in Chapter 2 the air channel was found to be surrounded by a mass transfer zone (Af7Z) where air flow has an impact on the mass transfer within the zone. The nonadvective region may be bigger or smaller than the MTZ depending on the number and diameter of the air channels present in the system. Using this idealized picture of air chaimels in a soil coluimi, a one­dimensional diffusion model with cylindrical coordinates was used to model the air sparging of the soil colunm. The diffusion equation for each channel was given by: dC^. £• TD^ d r dr r (or R<r<c dr (1) were Cw is the concentration of VOC in the water phase (ML"'), r (L) is the radial distance from the center of the air chaimel, /? is the air charmel radius (L), c is the radius of the non advective saturated porous media region surrounding the air channel (L), £ is the water filled porosity of the porous media (dimensionless), Dy^. is the water diffusivity of the VOC and ris the tortuosity factor (dimensionless). To determine the number of chaxmels an estimation of the air filled porosity (air saturation) and the diameter and length of the air channels was needed. The air saturation was estimated from the volume of water displaced into the top layer of glass beads after air was turned on. The average air charmel diameter was estimated by direct visual inspection of several air chaimels close to the colunm wall. The average air channel diameters were estimated to be 0.2 cm, 0.35 cm, and 0.30 cm for sand 30/50, sand 70/100, and Ottawa sand, respectively. The length (Lo) of the air channels was assumed to be equal to the depth of the bottom glass beads layer and the depdi of the porous media. Under these conditions, the number of air channels may be estimated as follows: Cylindrical Air Channels igurc 152 # of air channels = volume of water displaced (2) and the total interfacial area may be estimated as follows: A = (# of air channels) 2K R (3) The radius of the nonadvective saturated region, c, was estimated as follow: / c= Cross sectional area of the column \0.5 K (# of air channels) (4) Based on the air saturation in the column for the air sparging experiments, the value of c was always smaller than the size of the MTZ. As explained later, to run the model, the nonadvective saturated zone or the MJZ was subdivided in several concentric rings. The number of rings was dependent on the value of c. The initial and boundary conditions for equation 1 are as follows (see Chapter 4): Cw is known for all r at time zero (experimental data) TD^­^ = ~KcKfjC, atr = /?fort>0 VOC flux = zero at r = c for t > 0 dr (5) The volumetric average concentration, Q,, was computed as follows; ­ C TV C (6) where V„ is the volume of the n concentric ring (L^), CH„ (ML"^) is the aqueous concentration of VOC in the n ring, and the summation extends to all the rings with r<c. The air phase concentration was computed using the ttansport equation for the VOCs in the air phase as follows: 153 V.­^ = ­Q.C.+KaAK„C, O) where Q is the measured VOC concentration in the effluent air (ML'^), Va represents the air volume (L^), Qa represents the air flow (L^T'), A is the total air­water interfacial area (L~) of the experimental setup. Equation 1 and equation 7 were solved numerically by using the Continuous System Modeling Program (CSMP) software (IBM, 1972). The VOC concentration profile was divided into several concentric layers of equal thickness as seen in Figure 3. The number of concentric layers was chosen depending on the value of c. In general, the larger the value of c, the larger the number of layers were used. In this study, four layers with a thickness of 0.05 cm were used for sand 30/50 (c = 0.3 cm, = 0.10 cm), six layers with a thickness of 0.08725 cm were used for sand 70/100 (c = 0.7 cm, /? = 0.175 cm), and seven layers with a thickness 0.05 cm were used for Ottawa sand (c = 0.5 cm, /? = 0.15 cm). ATc values were estimated using equation 11 in Chapter 4. Table 1 lists the values of all the parameters used in the model. Results and Discussion Examples of the modehng results are shown in Figures 4 to 7. Figure 4 shows the measured and the predicted vapor concentrations and the total mass removed over time for benzene. In Figure 4, sand 30/50 was used as the porous medium and the estimated air pore velocity was 0.86 cm/s. The average diameter of the air channels was visually estimated to be approximately 2 mm and the air saturation was estimated to be 10%. As seen in Figure 4, the predicted air phase concentrations of benzene by the model were approximately close to the experimental values. Similar predictions were obtained for the other two VOCs, ethylbenzene and n­propylbenzene, using sand 30/50 as the porous media. A summary of the modeling results is presented in Table 2. The mass of sorbed VOC at the end of the experimental run was found to be negligible. The predicted final masses of VOCs remaining in the soil colunm after sparging were close to the experimental values. The data presented in Table 2 showed that after 10 hours of sparging the difference between predicted and Air Channel Figure 3. Cross section of air channel and MTZ Table 1. Parameters used in the model Porous Media Sand 30/50 Sand 70/100 Ottawa Sand Air Flow Rale (L/min) Air Channel Radius (cm) Torluo.sity 0.8 0.7 I.I 0.100 0.175 0.150 0.51 0.47 0.52 K(j (cni/min) benzene 0.00664 0.00774 N/A elhyibenzene n­propyl­ benzene 0.0039.1 0.00457 N/A 0.00253 0.00295 N/A toluene N/A N/A 0.(K)6I2 Henry's law constant and liquid phase diffusivities were taken from Table2 in Chapter 2 chlorobenzcne .styrene N/A N/A 0.00849 N/A N/A 0.01173 1.2 dichloro­ benzcnc N/A N/A 0.01086 156 0 150 ­T 100 200 300 400 500 600 700 Time (min) =0 £ 0 'I I I 1 I I I I I j I I [ I I I I I I I I I I I I I I I i I I I I I 0 100 200 300 400 500 600 700 Time (min) Figure 4. (a) Benzene concentrations in the exhaust air, (b) benzene mass removal for sand 30/50 at an air velocity of 0.86 cm/s 157 1.00 c tn a 0.75 ­ O tn 3 ca JCx UJ 0.50 ­ u c o O u 0.25 ­ c uN c 1) >1 JZ UJ 0.00 1 I I I I I I I I I I I I I I I r I rI 0 100 200 300 400 500 600 Time (min) g u 10 r i I I I I I 1 1 I T' 1"i I I 1 I I I I I I i i I I I i 0 100 200 300 400 500 600 Time (min) Figure 5. (a) Ethylbenzene concentrations in the exhaust air, (b) ethylbenzene mass removal for sand 30/50 at an air velocity of 1.21 cm/s 158 1.00 0.75 ­ 0.50 ­ 0.25 ­ I 0 I I—I—p­i—I—I 100 I I I—I 200 I I I I 300 I—I r I I I—r—r 400 500 400 500 Time (min) =P 40 p a 30 20 C/3 10 0 100 200 300 Time (min) Figure 6. (a) Styrene concentrations in the exhaust air, (b) styrene mass removal for Ottawa sand and air velocity of 1.34 cm/s Table 2. Summary of modeling results Porous Media Air Velocity (cm/s) VOC Expcrimcntal Initial Mass (mg)* Expcrimenial Ma.ss Removed (mg) Predicted Ma.s.s Removed (mg) Sand 30/50 Sand 30/50 Sand 3(V50 Sand 70/100 Sand 70/100 Sand 70/100 Oduwa Sand Oitawa Sand Ottawa Sand Ottawa Sand 0.86 0,86 0.86 1.21 1.21 1.21 1.34 1.34 1.34 1.34 Benzene Ethylben/.enc n­Propylbenzene Benzene Elhylbcn/.cnc n­Propylbenzene Toluene Chlorobcnzenc Styrene 1.2­ Dichlorobcn/ene 118.7 ± 1.02 96.98 ±0.83 15.62+0.44 36.32 ± 1.92 25.40 ± 1.23 1.88 ±0.01 69.58 ± 2.75 239,9 ± 19.7 61.18 ±3.25 45,72 ± 1.79 99.67 91.69 11.74 23.56 14.87 1.39 39.96 148.43 41.66 37.97 114.83 92.69 14.45 26.22 17.17 1.13 55.86 179.84 45.21 35.66 * including standard deviation Experimental Final Average Aqueous Concentration (mg/L)* 0.7 ± 0.7 1.6 ±0.4 0.2 ±0.1 5.43 ± 1.02 4.42 ± 0.42 0.30 ±0.01 14.3 ± I.I 28.8 ± 8.6 10.6 ± 1.6 4.9 ± 1.7 Predicted Final Average Aqueous Concentration (mg/L) Experimental Final Average Mass Remaining (mg) * 1.15 1.47 0.45 3.95 3.17 0.29 5.64 24.56 6.61 4.14 1.51 ± I..54 3.63 ±0.92 0.49 ± 0.22 I3.85± 2.60 n.27± 1.07 0,77 ± 0.02 34.33 ± 2.53 69.04 ± 19.8 25.24 ± 3.68 11.81 ±3.91 Predicted Final Average Mass Remaining (mg) 2,62 3,34 1,02 10,22 8,20 0,76 13,62 59,24 15,95 10,00 160 experimental mass removed for sand 30/50 was 15 mg for benzene (13% of the total mass of benzene), 1 mg for ethylbenzene (1% of the total mass of ethylbenzene), and 2.7 mg for n­ propylbenzene (17% of the total mass of n­propylbenzene). For the three VOCs, the model overpredicted the mass removed. For this experiment, the difference between the experimental and modeled final VOC concentrations in the liquid phase ranged between 0.3 mg/L and 1.1 mg/L. Figure 5 shows the experimental and predicted ethylbenzene vapor concentrations and the cumulative mass removed using sand 70/100 as porous medium and an estimated average air pore velocity equal to 1.21 cm/s. The average air channel diameters were visually estimated to be 3.5 mm, and air saturation was estimated to be 6.3%. Just as in Figure 4, Figure 5 shows that for early sparging times, the predicted cumulative mass removed was lower than the experimental mass removed. This may be due to the initial volatilization of VOCs from the displaced water located at the top of the sparged porous media and the VOCs in the headspace of the soil column which were not considered in the formulation of the model. The difference between predicted and experimental mass removed may be estimated from Table 2. For sand 70/100, the difference between predicted and experimental mass removal was 2.7 mg for benzene (7.6% of the total mass of benzene), 2.3 mg for ethylbenzene (9% of the total mass of ethylbenzene), and 0.26 mg for n­propylbenzene (14% of the total mass of n­propylbenzene). The sensitivity of the model can be seen by comparing the results for the sand 30/50 and the sand 70/100. For sand 30/50, the initial VOCs concentrations (benzene, ethylbenzene, and n­propylbenzene) were about four times higher than the initial concentrations used in the experiment for sand 70/l(K). On the other hand, the average air pore velocity for sand 70/100 was approximately 1.5 times higher than the average air pore velocity used for sand 30/50. If the mass removed or the mass remained were compared, it can be seen that despite the low concentration of the VOCs present in sand 70/100, the mass remaining was higher than of sand 30/50. This implies that the porous media had a strong effect on the mass volatilized by air sparging independent of the air flow rate once a quasi steady­state of aqueous VOC diffusion was established. The ability of the model to predict the experimental results for different porous media and air velocities indicates that the model was sensitive to the various 161 experimental conditions. In addition, the model also provides validation to the concept of the mass transfer zone {MTZ) surrounding the air channels. Figure 6 presents the measured and predicted styrene concentrations and the cumulative mass removed using Ottawa sand as porous medium and an estimated air pore velocity of 1.34 cm/s. For this experimental run, the diameter of the air channels was approximately 3 mm and the air saturation was estimated to be 8.1%. The model predicted fairly well the measured experimental vapor concentrations for styrene. Table 2 presents the modeling results for the other three VOCs. The model predicted very well the final mass of chlorobenzene and 1,2 dichlorobenzene but the prediction was poor for toluene. After 8 hours of sparging the mass removed, experimentally and modeled, were fairly similar for styrene and 1,2 dichlorobenzene (less than 10% difference). For chlorobenzene, the difference was 30 mg or 20%. Even though this value seemed high it represented just 12.5% of the total initial mass of chlorobenzene present in the reactor. For toluene, the predicted average aqueous concentration was 5.64 mg/L while the experimentally determined aqueous concentration was 14.3 mg/L. A probable reason for the less than desirable prediction may be due to the cosolvent interactions between toluene and other VOCs. For this experimental run, the model over predicted the mass of VOC removed after 8 hours for all VOCs except for 1,2 dichlorobenzene. From Table 2, the difference between experimental and predicted mass of VOC remaining in the colunrn after 8 hours was estimated to be 30% of the total mass for toluene, 4% of the total mass for chlorobenzene, 15% of the total mass for styrene, and 4% of the total mass for 1,2 dichlorobenzene. For early sparging times. Figure 6 shows a similar behavior as seen in the two previous experimental runs with respect to the cumulative mass removal. The mass removed experimentally was higher than the predicted mass removal for early sparging times but the situation was reversed after 300 minutes of sparging. For all three experimental runs using different porous media, the model underpredicted the vapor concentrations for early sparging times but after 100 minutes of sparging the model slightly overpredicted the vapor concentration. As explained earlier, this behavior may be due to the initial displacement of head space above the saturated porous media and the volatilization of VOCs from the initial air­displaced water in the top layer of glass beads. These two processes were not considered 162 in the formulation of the model. Given the experimental errors associated with the measurement of the VOC concentrations in the different phases, experimental mass balances were conducted and the differences were found to range from 1% of the total mass for ethylbenzene to 20% of the total mass for n­propylbenzene (sand 30/50). The experimental mass balance difference was computed for each VOC as follows: Experimental VOC Mass Balance Difference = (XO + SO) ­(Y + XF +5/) (8) where Y is the experimentally measured total mass of VOC removed (mg), X is the experimentally determined mass of VOC present in the aqueous phase (mg), S is the experimentally determined mass of VOC sorbed onto the solid phase (mg), and the subscripts o and/represent initial and final conditions, respectively. Xwas determined by multiplying the average value of the VOC aqueous concentration along the column depth times the volume of water present in the colunm. The average difference between initial and final mass for ail VOCs tested was 10%. The measured mass of VOC remaining in the liquid phase and the experimentally determined mass of VOC removed were plotted and compared with the predicted values as shown in Figures 7 to 9 for the three experimental runs. Also plotted in the figures was the experimental VOC mass balance difference as defined in equation 8. Figures 7 to 9 showed that when the experimental mass balance differences were taken into account the results of the model, which conserved all the mass of the VOCs, compared fairly well with the experimental results and were within 5% of the initial mass present. Of all the different parameters used in the model, the radius of the air channels was the most difficult parameter to be estimated. To assess the influence of the air channel radius on the overall benzene mass removal for sand 30/50, simulations were conducted by using air channel radius of 1.25 mm and 0.0875 mm. The results of the simulations are presented in Figure 10. Figure 10 shows that mass removal was inversely related to the radius of the air chaimel. When the radius of the air channel was smaller, the overall mass of benzene removed was larger due to an increase in the air­water interfacial area of the system. The 163 140 3 Experimental Mass Balance Difference D Mass Removed 120 ] Remaining Dissolved Mass 100 50 £ O O > (/i V5 C3 80 60 40 20 ­20 Expt. Predicted Benzene Expt. Prediaed Ethylbenzene Expt. Prediaed n­Propylbenzene ­40 Figure 7. Comparison of experimental and predicted mass distribution after sparging for sand 30/50 and air velocity of 0.86 cm/s (Experimental mass balance difference = initial (sorbed + dissolved) VOC mass in the column ­ mass of VOC volatilized ­ final (dissolved + sorbed) mass of VOC in the column) 164 40 {in"''''"! Experimental Mass Balance Difference I I Mass Removed I I Remaining Dissolved Mass 30 ­ 50 £ U o > "o C/1 {/J A 10 ­ Expt. Predicted Benzene Expt. Predicted Ethylbenzene Expt. Predicted n­Propylbenzene ­10 Figure 8. Comparison of experimental and predicted mass distribution after sparging for sand 70/100 and air velocity of 1.21 cm/s (Experimental mass balance difference = initial (sorbed + dissolved) VOC mass in the column ­ mass of VOC volatilized ­ final (sorbed + dissolved) mass of VOC in the column) 165 300 275 250 I I I I Experimental Mass Balance Difference Mass Removed Remaining Dissolved Mass 225 200 'so a u o > 'o C/3 et 2 175 150 125 100 75 50 25 0 ­25 ­50 Expt. Predicted Expt. Prediaed Expt. Predicted Expt. Predicted Toluene Chlorobenzene Styrene 1,2 Dichlorobenzene Figure 9. Comparison of experimental and predicted mass distribution after sparging for Ottawa sand and air velocity of 1.34 cm/s (Experimental mass balance difference = initial (sorbed + dissolved) VOC mass in the column ­ mass of VOC volatilized ­ final (sorbed + dissolved) mass of VOC in the column) 166 125 11 0 —I I I I I I I I I [ I I I I I I I I I I I I I I [ ­I I I I [ I I I I 0 100 200 300 400 500 600 700 Time (min) Figure 10. Influence of air channel radius on benzene mass removal 167 predicted mass of benzene remaining in the system after 10 hours of sparging was 1.36 mg for an air channel radius of 0.0875 mm and 4.11 mg for an air channel radius of 1.125 mm, respectively. When compared with the measured mass remaining (1.51 ± 1.54) mg it can be seen that model predictions were within the actual mass remaining in the soil column. The predictions of the one­dimensional diffusion model which incorporated the independendy determined mass transfer coefficients and mass transfer zone concept as determined in the single­air charmel smdies, showed that the model was capable of describing fairly well the VOC concentrations in the exhaust air, and final VOC concentrations in the liquid phase. It is possible that the successful application of the results from a single­air chaimel apparams to a complex system such as a soil column may be extended by scaling up the model for the prediction of air sparging on a field scale. Further testing using saturated soil tanks, simulating field­scale sparging process, should be conducted to test and improve the model. Conclusions A one­dimensional diffusion model was used to model the air sparging of soil columns contaminated with VOCs. The tortuosity factor and mass transfer coefficients determined by single­air channel smdies were used as input parameters for the model. In addition, the concept of mass transfer zone {MTZ) was incorporated into the model. The model predicted fairly well the VOC concentrations in the exhaust air and the total mass removed for nine out of the ten cases tested. Except for toluene, the predictions for mass removal were ranged from 1% to 17% of the initial VOC mass present in the column. The results of the model seemed to suggest that air sparged soil columns may be represented by a composite of individual air channels surrounded by a nonadvective region of porous media or MTZ. The extent of the nonadvective region will depend on the air saturation and the physical­chemical properties of the air sparging system. Under these conditions, the closer the air channels, the better is the removal of VOCs. Application of the model to predict field­scale air sparging operations may yield similar results. However, more laboratory testing is needed to verify the model and the validity of the mass transfer correlations for different air sparging conditions. 168 References Braida, W.J., and S.K. Ong, Air sparging effectiveness: the air channel mass transfer zone, submitted to Water Resources Research, 1997a. Braida, W.J., and S.K. Ong, Air sparging: air ­water mass transfer coefficients, submitted to Water Resources Research, 1997b. Drucker, A.S., and S.S. Di Julio, Groundwater clean up by In situ air sparging: Development of a model and application to saturated soil column experiments, in Proceedings of the Water Environmental Federation 69 th Annual Conference and Exposition, Dallas, Texas, October 5­9, 1996. Hein, G.L., Air Sparging as a remediation technique: modeling and experimental analysis. Ph.D. Dissertation, 177 pp., Michigan Technological University, Houghton, MI, 1996. Ji, W., Air Sparging: Experimental and theoretical analysis of flow and numerical modeling of mass transfer, Ph.D. Dissertation, 155 pp.. The University of Connecticut, Storrs, CN, 1994. 169 CHAPTER NINE. GENERAL CONCLUSIONS AND FUTURE WORK Conclusions This chapter summarizes the conclusions presented in the preceding chapters. A single­air channel experimental setup was used to study the controlling mechanisms in the volatilization of VOCs during air sparging. Correlations between estimated mass transfer parameters and the physical­chemical properties of the system were developed. Once the controlling mechanisms for ±e volatilization of VOCs during air sparging were established, a one­dimensional diffusion model was developed. The model using the information determined by the single air channel setup was then used to predict the VOC concentrations in the exhaust air and in the liquid phase for the air sparging of soil columns. The conclusions of this study are as follows: (i) According to the results of the single­air channel setup, mass transfer of VOCs during air sparging was found to be a diffusion­limited process. A steep concentration gradient zone was shown to exist in the samrated porous media next to the air channels during air sparging operations. The steep concentration zone was named as the mass transfer zone (MTZ). Depending on the VOC and the porous media used, the size of the MTZ ranged between 17 mm and 41 mm or between 70 and 215 dpso. (ii) A correlation was developed to predict the size of the MTZ. The correlation incorporated the Pore Diffusion Modulus, the air phase Peclet number, the uniformity coefficient and the dimensionless mean particle size of the porous media. The size of the MTZ, was found to be proportional to the aqueous diffiisivity of the VOC, the uniformity coefficient, and the mean particle size of the porous media. Air velocity was found to have marginal effect on the size of the MTZ. (iii) The size of the MTZ was found to be affected by the amount of organic carbon present in the porous media. An increase in the organic carbon content of the soil resulted in a decrease in the size of the MTZ. The effect was larger for VOCs with low solubilities and high partition coefficients (such as n­propylbenzene). The decrease in the size of the MTZ may be due to sorption onto the solid phase which retarded the diffusion of the VOCs to the air­water interface. 170 (iv) The overall gas phase mass transfer coefficients {Kc) for the volatilization of VOCs were found to be between 1.79 x 10"^ ctn/min and 3.85 x 10'" cm/min. Two correlations incorporating the modified Sherwood number and Damkohler number with Peclet number, dimensionless mean particle size, and dimensionless Henry's law constant were developed. KG was found to be proportional to the gas phase diffusivity of the VOC but inversely proportional to the Henry's law constant of the VOC. The doniinant and controlling factor in the determination of the overall mass transfer coefficients was the diffusivity of the VOC in the air phase. (v) The two­resistance model of Whitman may be used to describe the controlling resistance for the volatilization of VOCs during air sparging. For VOCs with large values of the dimensionless Henry's law constant, the liquid side layer controlled the volatilization of VOCs at the air­water interface. For VOCs with low values of the Henry's law constant, air velocity and the mean particle size of the porous media seemed to have an effect on the relative magnitude of the liquid or gas resistance to mass transfer. (vi) For a NAPL located several centimeters away from the air­water interface, air sparging has a positive effect on controlling the spreading of the NAPL. The effect was larger for compounds with high water solubilities and high diffusivity. An increase in the air flow rate did not produce any improvement in the removal efficiency of the VOCs. Porous media with larger particle size than fine sand (= 0.2 mm) showed higher removal rates. This suggests that removal rate was controlled by the diffusion of the dissolved VOCs to the air­ water interface. (vii) A modification of the one­dimensional diffusion model which included a retardation factor (/?) was found to predict fairly well the concentration curves for the liquid and air phases by using the mass transfer coefficients from the correlations developed in Chapter 4. Experimental results showed that the sorbed concentration in the porous media did not change after sparging for 8 hours. This indicates that sorption plays a significant role in retarding the VOCs and much of the VOCs volatilized were from the aqueous phase. (viii) A one­dimensional diffusion model using cylindrical coordinates and the mass transfer coefficients estimated using the single air channel semp was found to predict fairly well the VOC concentrations in the exhaust air and the final VOC concentrations in the 171 liquid phase for air sparged soil columns. Recommended Future Work Even though the results of this research have answered some questions about the controlling mechanisms in the volatilization of VOCs during air sparging operations, some questions still remain unanswered. The following recommendations are provided for further research of this technology; (i) Understand the exact nature of the air flow within sparged aquifers. Predict the size and number of channels within the sparged aquifer as a fiinction of physical properties and air saturation. (ii) Study the influence of organic carbon content of the soil and the water­solid partition coefficient of the VOCs on the size of the MTZ. (iii) Study the rate of dissolution of NAPLs during air sparging operations and determine how the location of the NAPL with respect to the MTZ affects mass transfer. (iv) hivestigate the mass transfer of oxygen from the sparged air into the groundwater and assess the role of bioremediation in the cleanup of contaminated aquifers during air sparging operations. (vi) Apply the one­dimensional diffusion model for the prediction of the volatilization of VOCs in air sparged soil columns containing different organic carbon contents. (vii) Apply the diffusion model and the MTZ and KQ empirical correlations for the prediction of the volatilization of VOCs during air sparging in actual field conditions. 172 APPENDIX A. NOTATION 173 Dimensionless Numbers DeHnition and Parameters Nomenclature Shi, overall i phase Sherwood number, K d '', where dp is the geometric mean particle Di diameter. K­a, d; Shioverall i phase modified Sherwood number, ————, where dp is the geometric mean particle diameter. Oaw, specific interfacial area (L"'). Rei, i phase Reynolds number, Sci, i phase Schmidt number, Nui, i phase Nusselt number, dp Vi Oi —. [M . PiDi K,d„ . A e., porosity. fi\, i phase viscosity {MT'L'). Pi^ i phase density (ML'^). Di,j, for i = a,w and j =x,z,r, VOC's diffusion coefficient in liquid and air phase in the x z and r directions, respectively {L^T'). So, air saturation, dimensionless. KH '., dimensionless Henry's law constant, dimensionless. G L , where G is the air mass flow rate, L is the length of Pa Ap Da sparging regime and Ap is the pore area. Pe, air phase Peclet number, K a L, pa Ap St, Stanton number, — , where KiOcn,­ 2 KhG Ed, pore diffusion module, G Roi' is the lumped mass transfer coefficient. ^ ^ where R^ is the influence radius. Ka L GJ; Damkholer number, ———, where Z, is the air flowpath length, and v is the specific V air discharge. k„ for i a and w, liquid and gas side mass transfer coefficients (UT). 174 KL , overall liquid side mass transfer coefficient between liquid and air phases (Z/D. KG , overall air side mass transfer coefficient between liquid and air phases {L/T). Kow, octanol water partition coefficient, dimensionless. mass transfer coefficient between adsorbed and dissolved VOCs (T'). Vi,fori=a,s,w is the volume of the cited phase (air, solid, or liquid, respectively, L^). Kd, partition coefficient (adsorbed­dissolved, L^/M). Ktjis, mass transfer coefficient between NAPL and dissolved VOC (T'). Ca, VOC air­phase concentration {M/L^). CW, VOC liquid­phase concentration (M/L^). CS, VOC sorbed­phase concentration {M/M). Qa, air flow rate (L^/T). 175 APPENDIX B. RAW DATA AND MASS BALANCES 176 Single­Air Channel Experiments Gas Phase and Aqueous Phase VOC Concentrations Ottawa Sand, Air Flow Rate 10 mL/min, Temperature 2l°C Air Phase Concentrations (mg/L) Time (min) 1 3 5 10 20 30 45 60 75 90 105 120 150 180 210 240 270 300 360 420 480 Benzene 0.039 0.068 0.116 1.583 2.135 1.855 1.881 1.815 1.614 1.563 1.442 1.387 1.020 0.890 0.649 0.548 0.508 0.487 0.487 0.485 0.483 Ethylbenzene 0.006 0.051 0.102 0.190 0.299 0.389 0.431 0.433 0.371 0.408 0.389 0.433 0.324 0.324 0.236 0.212 0.236 0.160 0.157 0.154 0.154 m.p­Xylene 0.002 0.001 0.026 0.194 0.329 0.388 0.462 0.485 0.410 0.427 0.446 0.428 0.349 0.304 0.270 0.221 0.253 0.185 0.154 0.152 0.151 o­Xylene 0.000 0.000 0.053 0.140 0.197 0.153 0.278 0.368 0.254 0.262 0.276 0.248 0.189 0.190 0.161 0.145 0.125 0.103 0.101 0.100 0.100 Liquid Phase Concentrations (mg/L) Depth (mm) 5.0 10.0 19.1 38.1 63.5 88.9 Benzene (0 h) 67.056 88.793 121.049 120.818 115.037 121.550 Benzene (8 h) 21.659 30.522 59.135 118.747 115.033 120.786 o­Xylene (0 h) 21.719 25.054 29.864 28.118 32.431 32.340 o­Xylene (8 h) 9.774 11.868 23.581 24.968 32.215 32.046 Liquid Phase Concentrations (mg/L) Depth (mm) 5.0 10.0 19.1 38.1 63.5 88.9 Ethylbenzene (0 h) 24.394 28.767 38.230 37.816 39.965 38.597 Ethylbenzene (8 h) 9.916 15.835 30.690 35.327 38.916 38.200 m,p­Xylene (0 h) 26.575 32.642 43.925 46.013 46.202 48.025 m,p­Xylene (8h) 10.288 17.312 36.256 45.291 46.408 47.091 177 Ottawa Sand, Air Flow Rate 10 mlVmin, Temperature IVC Air Phase Concentrations (mg/L) Time (min) 0 I 3 5 10 20 30 45 60 75 90 105 120 135 150 180 210 240 300 360 420 480 Toluene 0.000 O.OOG 0.000 0.095 0.411 0.583 0.587 0.590 0.552 0.538 0.514 0.481 0.446 0.423 0.400 0.392 0.367 0.331 0.321 0.307 0.293 0.292 Chlorobenzene 0.000 0.000 0.000 0.000 0.227 0.357 0.434 0.442 0.468 0.462 0.446 0.420 0.410 0.406 0.362 0.358 0.347 0.346 0.309 0.263 0.245 0.232 Siyrene 0.000 0.000 0.000 0.000 0.143 0.197 0.221 0.259 0.337 0.252 0.255 0.248 0.240 0.220 0.218 0.209 0.210 0.200 0.200 0.189 0.179 0.165 Propylbenzene 0.000 0.000 0.000 0.000 0.095 0.170 0.195 0.214 0.293 0.329 0.336 0.337 0.344 0.284 0.279 0.274 0.224 0.219 0.137 0.126 0.084 0.085 1.2 DCB 0.000 0.000 0.000 0.000 0.046 0.086 0.132 0.248 0.271 0.285 0.376 0.334 0.268 0.219 0.209 0.202 0.180 0.200 0.185 0.190 0.188 0.178 1.2.4 TCB 0.000 0.000 0.000 0.000 0.000 0.027 0.027 0.042 0.056 0.131 0.160 0.125 0.103 0.108 0.093 0.089 0.067 0.060 0.063 0.051 0.041 0.029 Liquid Phase Concentrations (mg/L) t = Oh Depth (mm) 5.0 10.0 19.1 38.1 63.5 88.9 Toluene 49.132 65.001 64.244 68.205 69.887 69.883 Chlorobenzene 48.614 66.051 70.247 70.302 71.313 71.313 Styrene 34.501 38.865 42.924 44.370 43.981 44.370 n­propylbenzene 13.896 18.005 19.649 22.712 25.846 25.775 1,2 DCB 39.824 42.882 45.456 46.019 46.160 46.600 1.2,4 DCB 6.548 7.329 9.879 9.750 10.420 10.831 1,2 DCB 26.261 41.948 40.957 43.392 45.297 46.511 1,2.4 DCB 5.534 7.067 9.804 9.357 10.179 10.733 Liquid Phase Concentrations (mg/L) t = 4b Depth (mm) 5.0 10.0 19.1 38.1 63.5 88.9 Toluene 40.281 46.397 57.070 63.923 68.995 69.435 Chlorobenzene 42.521 46.579 63.117 69.312 70.228 70.234 Styrene 27.644 29.743 41.130 41.889 43.497 44.245 n­propyibenzene 11.279 15.379 18.978 20.974 24.650 24.905 178 Liquid Phase Concentrations (mg/L) t = Sh Depth (mm) 5.0 10.0 19.1 38.1 63.5 88.9 Chlorobenzene 34.022 41.759 58.527 68.210 70.224 70.224 Toluene 26.778 44.097 56.870 62.823 68.856 69.430 Styrene 19.844 31.043 37.630 41.881 43.500 44.235 n­propylbenzene 7.368 15.166 17.902 20.196 23.655 24.035 1.2 DCB 23.861 36.928 40.937 43.388 45.097 46.501 1,2.4 DCB 4.534 7.067 9.004 9.250 9.879 10.723 Ottawa Sand, Air Flow Rate 25 mL/min, Temperature 21°C Air Phase Concentrations (mg/L) Time (min) 0 1 2 4 11 20 30 45 60 75 90 105 120 150 180 240 300 360 420 480 Benzene 0.1600 1.0325 1.2897 1.1155 0.8325 0.6653 0.4858 0.2395 0.2028 0.1793 0.1623 0.1533 0.1485 0.1493 0.1473 0.1178 0.1145 0.1087 0.1078 0.1098 Ethylbenzene 0.0054 0.0718 0.3525 0..3330 0..1718 0.1225 0.0830 0.0750 0.0635 0.0605 0.0503 0.0508 0.0550 0.0465 0.0430 0.0468 0.0410 0.0311 0.0275 0.0247 ni.p­Xylene 0.0043 0.0735 0.3530 0.3350 0.2965 0.2260 0.1250 0.0972 0.0790 0.0744 0.0694 0.0585 0.0565 0.0510 0.0378 0.0329 0.0272 0.0269 0.0215 0.151 o­Xylene 0.0045 0.0513 0.2968 0.2228 0.1475 0.1298 0.1088 0.0828 0.0790 0.0728 0.0637 0.0528 0.0480 0.0406 0.0345 0.0237 0.0174 0.0193 0.0198 0.0100 Liquid Phase Concentrations (mg/L) Depth (mm) 5.0 15.0 19.1 38.1 63.5 88.9 Benzene (0 h) 42.342 83.513 103.258 106.840 108.484 109.387 Benzene (4 h) 16.289 67.357 93.125 100.183 108.694 109.301 Benzene (8 h) 15.357 56.196 86.765 96.951 108.691 109.294 o­Xylene(Oh) 13.963 17.066 23.366 23.396 23.517 23.467 o­Xylene (4 h) 5.472 15.119 23.315 23.311 23.457 23.368 o­Xylene(8h) 1.427 13.089 20.116 22.194 23.298 23.366 179 Liquid Phase Concentrations (mg/L) Depth (mm) 5.0 15.0 19.1 38.1 63.5 88.9 Ethylbenzene (Oh) 9.264 26.843 30.252 31.385 31.225 31.589 Ethylbenzene (4 h) 3.217 18.713 23.561 30.613 31.224 31.580 Ethylbenzene (8 h) 1.065 13.170 22.044 30.549 31.165 31.563 m.p­Xylene (0 h) 10.592 25.682 30.010 30.307 31.434 31.492 m.p­XyIene (4h) 3.978 14.012 26.117 30.105 31.569 31.489 m.p­Xylene (8h) 2.858 13.845 23.336 27.140 31.723 31.468 Ottawa Sand, Air Flow Rate 25 mL/min, Temperature 2l'C Air Phase Concentrations (mg/L) Time (min) 1 5 10 20 30 45 60 75 90 105 120 150 180 210 240 300 360 420 Toluene 0.0235 0.7210 0.6550 0.6275 0.5495 0.5145 0.4630 0.3915 0.3600 0.2775 0.2476 0.2260 0.1880 0.1487 0.1207 0.1015 0.1006 0.1016 Chlorobenzene 0.0161 0.4415 0.5400 0.5535 0.5335 0.4550 0.3750 0.3600 0.3270 0.3160 0.2950 0.2334 0.2070 0.1410 0.1201 0.1082 0.1058 0.1024 Styrene 0.0865 0.2195 0.2610 0.2865 0.2505 0.2310 0.1755 0.1715 0.1595 0.1335 0.1238 0.1245 0.1105 0.0978 0.0975 0.0875 0.0825 0.0823 Propylbenzene 0.0469 0.2425 0.2875 0.3345 0.3580 0.3690 0.3720 0.2310 0.2130 0.1835 0.1045 0.1055 0.0925 0.0752 0.0745 0.0742 0.0742 0.0739 1.2 DCB 0.0115 0.0157 0.0253 0.0288 0.0344 0.0413 0.0407 0.0620 0.0530 0.0580 0.0440 0.0447 0.0407 0.0332 0.0287 0.0219 0.0189 0.0161 1.2.4 TCB 0.0006 0.0006 0.0008 0.0148 0.0138 0.0128 0.0118 0.0118 0.0142 0.0171 0.0167 0.0166 0.0160 0.0138 0.0109 0.0093 0.0093 0.0095 Liquid Phase Concentrations (mg/L) t = Oh Depth (mm) 5.0 10.0 19.1 38.1 63.5 88.9 Toluene 29.908 30.255 32.687 35.357 35.417 35.521 Chlorobenzene 22.958 31.582 34.547 35.013 35.550 35.709 Styrene 14.439 20.503 22.603 22.767 24.114 24.579 n­propylbenzene 18.063 23.339 25.731 28.390 28.614 29.333 1.2 DCB 5.672 7.080 7.402 9.961 9.746 9.915 1.2.4 DCB 1.682 2.127 4.756 4.902 4.904 4.999 180 Liquid Ptiase Concentrations (mg/L) t=4h Depth (mm) 5.0 10.0 19.1 38.1 63.5 88.9 Toluene 9.581 19.639 32.552 34.971 34.918 35.209 Chlorobenzene 10.114 19.157 29.541 34.816 34.176 34.830 Styrene 6.918 13.071 21.123 21.727 23.238 24.049 n­propylbenzene 6.195 11.917 23.983 26.464 27.318 28.798 1.2 DCB 3.012 3.231 7.302 9.098 9.819 9.939 1.2,4 DCB 1.091 2.034 4.658 4.813 4.968 4.969 1.2 DCB 0.495 3.170 6.517 8.998 9.839 9.959 1.2.4 DCB 0.991 1.635 4.458 4.804 4.868 4.959 Liquid Phase Concentrations (mg/L) t = 8h Depth (mm) 5.0 10.0 19.1 38.1 63.5 88.9 Toluene 3.773 10.108 28.522 32.412 32.532 32.565 Chlorobenzene 2.794 10.072 27.392 31.966 33.776 33.901 Styrene 4.711 7.081 20.034 21.121 22.038 24.079 n­propylbenzene 1.595 7.919 19.853 25.986 26.448 28.392 Ottawa Sand, Air Flow Rate 55 mL/min, Temperature 21"'C Air Phase Concentrations (mg/L) Time (min) 0 I 2 4 11 20 30 45 60 75 90 120 150 180 240 310 360 420 480 Benzene 0.0370 1.5383 1.3205 0.8738 0.6600 0.4620 0.3200 0.2390 0.1578 0.0953 0.0895 0.0758 0.0695 0.0635 0.0570 0.0557 0.0503 0.0518 0.0505 Ethylbenzene 0.0087 0.4210 0.3730 0..2775 0..1026 0.0775 0.0585 0.0467 0.0383 0.0393 0.0385 0.0343 0.0258 0.0219 0.0204 0.0203 0.0157 0.0172 0.0164 m,p­Xylene 0.0000 0.4110 0.3870 0.1960 0.0895 0.0750 0.0598 0.0390 0.0353 0.0303 0.0273 0.0272 0.0274 0.0225 0.0215 0.0127 0.0123 0.0164 0.0122 o­Xylene 0.0097 0.3870 0.2216 0.1943 0.0883 0.0672 0.0494 0.0345 0.0274 0.0137 0.0109 0.0112 0.0106 0.0123 0.0112 0.0107 0.0097 0.0103 0.0108 181 Liquid Phase Concentrations (mg/L) Depth (mm) 5.0 10.0 15.0 19.1 38.1 63.5 88.9 Ben zene (Oh) 65.650 90.846 106.150 110.095 116.000 119.274 119.557 Benzene (3 h) 28.174 57.257 78.967 87.625 112.483 119.974 119.978 Benzene (8 h) 20.160 48.654 78.645 86.197 109.627 118.826 120.603 o­Xylene(Oh) 10.519 16.963 21.216 21.981 26.619 28.145 28.674 o­Xylene(3h) 3.272 5.421 18.997 21.495 26.317 28.007 28.668 o­Xylene(8h) 2.959 3.503 18.732 20.595 24.786 27.868 28.786 Liquid Phase Concentrations (mg/L) Depth (mm) 5.0 10.0 15.0 19.1 38.1 63.5 88.9 Ethylbenzene (0 h) 13.917 17.840 25.406 27.228 29.920 31.084 32.392 Ethylbenzene (3 h) 4.987 8.543 16.121 19.345 28.737 31.074 32.380 Ethylbenzene (8 h) 3.345 6.193 12.047 19.251 26.918 30.988 32.376 m,p­Xylene (Oh) 12.888 17.416 24.709 29.466 32.281 32.405 33.566 m.p­Xylene (3 h) 5.912 7.644 14.323 26.917 30.005 32.369 33.489 m.p­Xylene (8h) 3.642 6.105 12.243 19.602 29.831 32.338 33.500 Ottawa Sand, Air Flow Rate 55 mL/min, Temperature 21°C Air Phase Concentrations (mg/L) Time (min) 0 1 3 5 10 20 30 45 60 75 90 105 120 135 150 180 210 240 300 360 420 480 Toluene 0.0000 0.7355 0.5915 0.4915 0.4000 0.3275 0.2650 0.2550 0.2175 0.1840 0.1880 0.1720 0.1730 0.1525 0.1590 0.1485 0.1285 0.1235 0.1231 0.1138 0.1225 0.1210 Chlorobenzene 0.0000 0.5900 0.6035 0.5485 0.5305 0.3720 0.4115 0.3810 0.2975 0.3005 0.2400 0.2450 0.2320 0.2140 0.1945 0.1790 0.1620 0.1365 0.1250 0.1270 0.1280 0.1310 Styrene 0.0000 0.3180 0.3405 0.2840 0.2360 0.2385 0.2275 0.2225 0.2180 0.2030 0.1910 0.1560 0.1285 0.1160 0.1195 0.1190 0.1085 0.1028 0.1105 0.0920 0.0940 0.0955 Propylbenzene 0.0000 0.4855 0.5690 0.4835 0.3885 0.2645 0.1835 0.1345 0.1258 0.1208 0.1145 0.1065 0.0932 0.0800 0.0794 0.0564 0.0524 0.0528 0.0435 0.0425 0.0422 0.0418 1.2 DCB 0.0000 0.2893 0.2475 0.3603 0.2180 0.1915 0.1880 0.1468 0.1155 0.1048 0.1105 0.0838 0.0718 0.0628 0.0553 0.0538 0.0498 0.0373 0.0160 0.0153 0.0157 0.0150 1.2.4 TCB 0.0000 0.0000 0.2473 0.2620 0.2578 0.1608 0.1148 0.0823 0.0716 0.0656 0.0558 0.0543 0.0504 0.0402 0.0304 0.0184 0.0074 0.0052 0.0044 0.0039 0.0039 0.0039 182 Liquid Phase Concentrations (mg/L) t = Oh Depth (mm) 5.0 lO.O 19.1 38.1 63.5 88.9 Toluene 48.903 61.568 76.063 82.243 84.006 85.218 Chlorobenzene 52.262 63.761 88.107 92.469 94.421 95.562 Styrene 40.630 48.288 53.193 56.098 59.857 60.627 n­propylbenzene 26.579 26.821 32.948 34.442 40.097 40.505 1.2 DCB 44.258 46.028 55.237 55.593 59.214 59.053 1.2.4 DCB 12.085 15.446 17.378 18.332 18.988 19.171 1.2 DCB 27.127 42.149 43.327 50.911 59.157 58.971 1.2.4 DCB 7.328 9.996 12.174 17.296 18.947 19.069 1,2 DCB 24.505 35.820 38.230 50.867 58.757 58.870 1.2,4 DCB 2.356 4.610 9.480 17.074 18.752 18.969 Liquid Phase Concentrations (mg/L) t = 4h Depth (mm) 5.0 10.0 19.1 38.1 63.5 88.9 Toluene 18.437 46.037 62.317 76.591 81.126 83.379 Chlorobenzene 25.218 50.873 66.127 87.128 92.123 93.834 Styrene 25.119 31.846 42.321 56.119 56.320 59.176 n­propylbenzene 11.843 22.567 23.019 34.409 38.453 39.665 Liquid Phase Concentrations (mg/L) t = 8b Depth (mm) 5.0 10.0 19.1 38.1 63.5 88.9 Toluene 16.937 34.637 46.080 76.460 80.720 82.079 Chlorobenzene 18.399 38.621 53.689 84.898 89.216 92.088 Styrene 11.563 21.820 34.254 52.806 56.209 57.713 n­propylbenzene 3.189 12.810 18.304 34.362 38.350 38.816 183 Ottawa Sand, Air Flow Rate 120 mL/inin, Temperature Zl'C Air Phase Concentrations (mg/L) Time (min) 0 1 3 5 10 20 30 45 60 75 90 105 120 150 180 240 300 360 420 480 Benzene 0.0000 0.9450 0.3705 0.3327 0.2003 0.1583 0.1332 0.1300 0.1055 0.0868 0.0793 0.0703 0.0558 0.0630 0.0530 0.0425 0.0428 0.0413 0.0425 Ethylbenzene 0.0000 0.3260 0.1978 0.1915 0.1163 0.1005 0.0835 0.0785 0.0615 0.0645 0.0490 0.0440 0.0370 0.0333 0.0395 0.0231 0.0147 0.0200 0.0174 0.0124 0.041 m.p­Xylene 0.0000 0.3348 0.1803 0.1915 0.1275 0.1100 0.0750 0.0778 0.0515 0.0585 0.0440 0.0393 0.0393 0.0323 0.0285 0.0265 0.0185 0.0139 0.0149 0.0129 o­Xylene 0.0045 0.1755 0.1790 0.1658 0.0770 0.0778 0.0690 0.0513 0.0375 0.0338 0.0305 0.0283 0.0280 0.0269 0.0249 0.0249 0.0209 0.0098 0.0078 0.0075 Liquid Phase Concentrations (mg/L) Depth (nun) 5.0 10.0 19.1 38.1 63.5 88.9 Ben zene (0 h) 64.359 99.622 122.978 127.000 128.643 134.025 Benzene (4 h) 29.974 61.257 116.625 124.483 126.974 132.978 Benzene (8 h) 15.454 48.905 104.015 116.270 124.183 130.018 o­Xylene (Oh) 19.541 39.637 39.792 39.055 41.421 42.222 o­Xylene (4 h) 17.292 23.021 35.012 36.977 41.237 42.221 o­Xylene (8 h) 10.937 18.655 30.019 32.226 40.870 42.141 Liquid Phase Concentrations (mg/L) Depth (mm) 5.0 15.0 19.1 38.1 63.5 88.9 Ethylbenzene (0 h) 20.171 29.747 52.143 58.401 61.267 63.942 Ethylbenzene (4h) 12.287 25.001 41.023 55.876 57.567 63.645 Ethylbenzene (8 h) 9.424 18.424 30.652 52.633 57.422 62.891 m,p­Xylene (Oh) 33.484 56.263 66.020 70.424 70.518 70.654 m,p­Xylene (4h) 23.012 33.644 69.523 68.917 70.105 70.489 m.p­Xylene (8h) 12.642 29.043 54.858 63.450 69.095 70.490 184 Ottawa Sand, Air Flow Rate 120 mL/min, Temperature 21°C Air Phase Concentrations (mg/L) Time (min) 0 1 3 5 10 20 30 45 60 75 90 105 120 150 180 210 240 300 360 420 480 Toluene O.OOOO 0.2923 0.1665 0.1393 0.1043 0.0930 0.0843 0.0813 0.0675 0.0645 0.0622 0.0600 0.0550 0.0550 0.0538 0.0560 0.0530 0.0453 0.0454 0.0448 0.0446 Chlorobenzene 0.0000 0.2773 0.1843 0.1550 0.1208 0.1178 0.0995 0.0875 0.0805 0.0698 0.0738 0.0705 0.0705 0.0623 0.0510 0.0515 0.0510 0.0510 0.0510 0.0496 0.0448 Styrene 0.0000 0.1605 0.0900 0.0810 0.0813 0.0833 0.0593 0.0538 0.0456 0.0448 0.0415 0.0438 0.0343 0.0313 0.0340 0.0310 0.0335 0.0330 0.0313 0.0315 0.0312 Propylbenzene 0.0000 0.3510 0.1350 0.1255 0.1220 0.1005 0.0790 0.0780 0.0775 0.0535 0.0493 0.0444 0.0303 0.0288 0.0241 0.0210 0.0205 0.0203 0.0208 0.0186 0.0167 1.2 DCB 0.0000 0.2805 0.2145 0.1540 0.1445 0.1015 0.0970 0.0730 0.0610 0.0540 0.0340 0.0171 0.0152 0.0121 0.0045 0.0036 0.0025 0.0025 0.0025 0.0023 0.0022 1.2,4 TCB 0.0000 0.1080 0.0960 0.0910 0.0720 0.0650 0.0590 0.0675 0.0449 0.0412 0.0321 0.0301 0.0301 0.0262 0.0144 0.0110 0.0107 0.0108 0.0090 0.0057 0.0024 Liquid Phase Concentrations (mg/L) t = Oh Depth (mm) 5.0 10.0 19.1 38.1 63.5 88.9 Toluene 23.114 25.402 32.840 35.052 36.662 37.730 Chlorobenzene 32.518 32.518 36.599 38.397 39.364 39.676 Styrene 19.856 19.856 23.068 23.665 24.797 24.814 n­propylbenzene 11.543 18.674 19.761 24.094 24.766 26.672 1.2 DCB 14.281 19.671 21.642 25.021 25.062 25.066 1.2.4 DCB 7.479 7.893 8.422 8.832 9.088 9.079 1,2 DCB 4.217 18.987 20.978 24.211 24.213 25.058 1.2.4 DCB 0.227 4.129 6.332 8.141 8.567 9.069 Liquid Phase Concentrations (mg/L) t=:4h Depth (mm) 5.0 10.0 19.1 38.1 63.5 88.9 Toluene 9.539 17.737 30.983 34.998 34.174 36.315 Chlorobenzene 12.618 20.073 30.127 35.328 37.723 37.834 Styrene 8.014 13.214 18.645 21.978 23.567 24.071 n­propylbenzene 6.007 9.078 13.176 22.215 24.054 26.665 00 u» ^ po o> Ui 00 vo ^ O U) '\0 KJ\ o *0 00 ON 00 U» so ^^ 00 UJ o^ ^0*^^00 U) Ln Ui VO ­J VO U) 00 VO O VO N) OJ VO ­J NJ VO 00 L»J u> UJ ro p p LA bo OONN LOAN to o VO LA 00 VO U) § ) Kn) i^> J > — O 00ON)O\L<JQ4i.>— 0 0 0 0 0 0 0 0 0 <O­O' >l /0l NU) O) 0v C < ~J J<O0\ 40i0UV LA L/* 00 OJ LA fo N) ON LA Ov OJ b to uj LK) LA LA LA 4^ ­J ON 00 N> LA ON ON ON LA N> ON 00 •— 'o p bo ON O ON oo \o N) 00 u> LO N) 4^ U» K> 4^ N) U) OO O ­J O U) VO NO oo 4^ In ^ N) ON vO ­4 4a. LA •sj LA 00 ­J K) vO VO 00 U) bv b VO KJ o ro •­4 VO OroN 00 ­­4 L»i pppppppppppppooopppppo i o ^ t o N >—i suj iK( j­ Nr t)oUoJoW u » ^ ^ ! f k 4 i . L f t L n b v b \ ^ io>^j o^ j( ­D­ (j D O O^oo­^O'./I — — i>j<­no~JkOK)*oo\ivir— NOOlO o o oO O O LA 2s b ON VO LA OJ O LA O b LA L»J LA O <D LA Kji 00 o o o oo o p o b LA b o LA b g g g s LA o ro 4:^ ON 1­n o 00 o 00 00 oOoooOo b bb b u> to ro b ro b to ro b ro ro 00 o 00 VO 4^ ON ON u> LA LA LA o O O O O p o o O o oo oo b b ON bbbQb b bob' LA NO 00 00 b ON 2 NO ON VO ^5 to U) U» LA LA LA o to 8 OJ OOooOoOoOOoO b LA b LA b ObN LA b LA b ObN ObN ObN ON b •o bb LA U> ON ­J O LA •­J •­J u» 00 UJ O 00 00 00 O LA O to ro 4^ Ooo b b u>N OJ u> Ob ­­4 O o 4^ o b U) ON O ooO b LA oo O OJ g 2 ON OpOOo b b LA o ro b to b ­4 ON ­4 o iiM O U) U) •o U) N) VO ­­4 4^ LA p bo b b CA ON •O 2 GI ON o LA 00 N> to ro ^ N> to ^ Q *4^ ^ ^ 00 OS CK ON L/l VO U> 00 3 to to to ro 4i^ L»J .o 4^ Ui 00 'o N) F 00 VO o VO o VO 4^ NO vO •O to srt N O3 a lO L»i to ro LA ro 00 lO o bo Ui VO OVON a 'O 1­^ LA VO o ro L»J o ON Ov n C0 »— O vO 00 oo — P L/i O L/1 oo vo o •t' — O4in ?; Ln IM cx '— , , v>j 00 ­­4 VO 00 ­o 4^ p p lo to to 4^ ON vO to ro 0\ D VO vO O 03 •— 00 00 UJ UJ 00 00 OV )0 O VO La i— b (ta r3 & oo o* u> VO —o — .S" 00 <jJ 00 so l/> ~ i­ O o 4I^ 4^ »>J ^ I— ^ OON)On4^00L/)K)VO«^ L*J O •— On U) O O O O O O O O L f i OS O OO UJ OJ U) U) U) ro U) O VO 'to ­­J to Ln VO LA C/l u> 00 VO LA LA •— hJ 00 O fO i>j N) ON ^ vo ' Ov OS LA p oc fO k> LA b LA NJ LA o »— •— hJ — Ni VO p U ^ NO LA N> a\ p\ LA b LA b U) U) o\ N> O ET O fr 3 9r»cv 00 L*J U) to too VO LA p UJ p to P *00 00 La UI La oo o to 2 a oo On 9N Q9 a Ln U) Ln ^ u> to p P P OJ Ui ^ On L/i Ipt. bo SO C/) ^ O Q U) n vO U) LA 4^ U> CO p ­o 4^ LA p pooP ppoo p pP Ppo k) k) •"4 N) LA N) 00 U) LO to o k) 00 o »— io 4:>^ L*J ­O oo to LA vO k) VO OO Ui ­o Ov to U) £ Ov Ov o oo o La LA O o LA LA o LA « O CA O LA LA pp popppo oooppp ON ON ON bv bo 00 00 b — O N> LA OS 00 ov L»> 4^ oo L4^a OJ 4I>. U) U) CA VO 8 p VO ON to 00 to U> Ov Ui O o o o LA LA LA O o o LA LA o O O LA ppppppO I­­* to UJ r— OS OO k> VO VO o o ­O 00 LA LA LA O O 00 O opoPPoo k) U> LA 4:^ LJ b OO b to o to VO U) LA ON bl LA LA o LA o o 9 m o so Iti o LA ­J LA OS ­J bo ON U> oooooooooooooo bb b b b to b V>j b LA b VO b VO $ ON bb t— to U) to OS OS to LA VO U) o 00 oo VO CA 45. to ON OS ro LA LA •O to 00 s­ 4^ yi 'OS •O Lo . a LA OJ 9Nn 9 a • m 5­ o VO "P Ui to <> b ro . 9 00 LA t u> LA oo OS 4i­ to ro o n9N a B =1 3 o VO OS OS b 4^ bo VO 2 X 9* ^ LA Os LA LA VO LA OJ to 9n a <— »—• ^ JO VO LA ro 3 a LOsa bo b IO t X ON L»J to LA o 00 ­J LA s^a 9 a L >— oOoooooOooOoo VO 00 ­0 O •— OS LA o ? (T bs *4^ U) ro W 9 N Os o n tn pppppppppppoopooo N0>0K0 j0K0 0j N0 0>^l >­ (j^LOj ll >OJO^N4 ^O4O2Uh U 0U)»— i N) >^ ^Hf^stjO^ ^N)OJ»— t­0UiU»Lfit/iOOU»OOU»UiOUiUj b 8 b b Ov bb 8 8888o u» b b LA LA La LA Ox Ov vO 00 00 OS ­o ­J OS UJ to o vO ­J U) b 4:^ 4X bv LA \o o> LA ro G 4^ o vO o LA ti0C/) a u o so VO Lo to bo 00 b o U> OS to »— LA m 5­ »< •— O 00 VO LA o 00 Os 3 *^3 ZI )0 00 yj ^ 00 4* UJ L»J SO 4J^ o 4* 1­^ vO o 9^ Os 5'< VO Ov 00 u 9C> •— r £' E. cu *•0 cr K n oD n an a o g 00 0\ 187 Liquid Phase Concemrations (mg/L) t = 4h Depth (nun) 5.0 10.0 19.1 38.1 63.5 88.9 Toluene 20.068 21.215 32.176 34.645 33.718 34.584 Chlorobenzene 49.671 56.174 113.238 122.568 123.189 123.419 Styrene 26.015 38.756 51.176 56.343 60.489 62.212 n­propylbenzene 9.211 17.011 25.613 29.753 30.537 30.840 1.2 DCB 21.813 30.712 43.117 50.475 50.551 50.599 1.2,4 DCB 5.217 7.771 9.627 10.987 12.382 13.451 1.2 DCB 14.725 29.340 40.287 50.468 50.369 50.588 1.2.4 DCB 3.013 6.023 8.949 10.895 12.268 13.265 Liquid Phase Concentrations (mg/L) t = 8h Depth (nun) 5.0 lO.O 19.1 38.1 63.5 88.9 Toluene 16.021 17.623 28.593 33.364 33.644 34.284 Chlorobenzene 45.045 53.689 104.547 116.895 123.242 123.388 Styrene 18.939 30.079 44.955 54.369 60.451 62.221 n­propylbenzene 5.305 14.034 23.356 27.564 30.528 30.823 Sand 30­50, Air Flow Rate 25mL/min, Temperature 21°C Air Phase Concentrations (mg/L) Time (min) 0 1 3 5 10 20 30 45 60 75 90 105 120 150 180 240 300 360 420 480 Benzene 0.0000 0.0239 0.9818 2.4120 1.9175 1.2842 1.0733 0.8923 0.6950 0.6873 0.6190 0.5828 0.5435 0.4670 0.4405 0.3678 0.3178 0.3023 0.2653 0.2572 Ethylbenzene 0.0000 0.0000 0.1718 0.7103 0.5315 0.4890 0.4928 0.4110 0.3598 0.3433 0.3420 0.3168 0.3083 0.2630 0.2233 0.1803 0.1505 0.1238 0.1138 0.1168 m,p­Xylene 0.0000 0.0000 0.1577 0.6670 0.5265 0.4340 0.3998 0.3773 0.3340 0.2925 0.2878 0.2615 0.2585 0.2413 0.1913 0.1748 0.1328 0.1015 0.1058 0.1042 o­Xylene 0.0000 0.0106 0.0735 0.3295 0.4820 0.3090 0.2860 0.2918 0.1783 0.1675 0.1570 0.1525 0.1475 0.1410 0.1320 0.1023 0.0983 0.0895 0.0853 0.0863 188 Liquid Phase Concentrations (mg/L) Depth (mm) 5.0 lO.O 19.1 38.1 63.5 88.9 Ben zene (Oh) 120.818 123.091 129.990 131.000 131.799 132.456 Benzene (4 h) 77.815 94.114 123.259 130.134 129.431 130.123 Benzene (8 h) 42.889 75.727 116.541 124.901 126.911 128.795 o­Xylene(Oh) 23.060 30.872 32.074 34.074 35.448 35.684 o­Xylene (4 h) 18.117 24.315 28.274 31.973 35.121 235.069 o­Xylene(8h) 12.885 16.622 22.218 31.911 34.244 34.360 Liquid Phase Concentrations (mg/L) Depth (mm) 5.0 10.0 19.1 38.1 63.5 88.9 Ethylbenzene (0 h) 40.667 43.923 49.867 52.696 53.847 54.103 Ethylbenzene (4 h) 27.511 39.347 46.640 48.014 51.996 51.976 Ethylbenzene (8 h) 14.056 24.686 43.300 47.998 51.838 51.853 m,p­Xylene (0 h) 47.444 51.973 55.200 57.187 59.199 59.160 m,p­Xylene (4h) 37.174 44.384 52.615 55.973 58.763 59.012 m,p­Xylene (8 h) 19.474 33.837 48.366 55.946 56.351 56.977 Sand 30­50, Air Flow Rate 25 mL/min, Temperature 21°C Air Phase Concentrations (mg/L) Time (min) 0 1 3 5 10 20 30 40 60 75 90 120 180 240 300 360 420 480 Toluene 0.0000 0.2560 0.2485 0.3135 0.6510 0.3340 0.3230 0.2950 0.2840 0.2800 0.2780 0.2570 0.2210 0.2100 0.2050 0.2051 0.2015 0.2023 Chlorobenzene 0.0000 0.4731 0.5201 0.6425 1.0975 0.8745 0.8611 0.7670 0.7230 0.7025 0.7245 0.7335 0.7015 0.7315 0.7150 0.7101 0.7065 0.7059 Styrene 0.0000 0.1530 0.1975 0.4405 0.3385 0.3115 0.3015 0.3025 0.3002 0.2750 0.2570 0.2761 0.2885 0.2685 0.2700 0.2570 0.2270 0.2287 Propylbenzene 0.0000 0.0535 0.1915 0.3705 0.2495 0.1845 0.1386 0.1025 0.0805 0.0774 0.0745 0.0725 0.0736 0.0743 0.0725 0.0715 0.0728 0.0721 1.2 DCB 0.0000 0.0000 0.0000 0.0083 0.0211 0.1483 0.0905 0.0790 0.0655 0.0547 0.0344 0.0188 0.0114 0.0080 0.0081 0.0085 0.0083 0.0082 1.2,4 TCB 0.0000 0.0000 0.0000 0.0000 0.0147 0.0174 0.0184 0.0302 0.0215 0.0072 0.0062 0.0000 O.OOOO 0.0000 0.0000 O.OOOO 0.0000 O.OOOO 189 Liquid Phase Concentrations (mg/L) t = Oh Depth (mm) 5.0 10.0 19.1 38.1 63.5 88.9 Toluene 41.884 46.462 50.582 50.810 51.568 51.718 Chlorobenzene 157.68 181.05 198.66 200.40 200.29 200.59 Styrene 60.649 78.554 85.423 90.081 90.054 90.010 n­propylbenzene 32.457 44.745 48.410 53.823 54.306 54.030 1.2 DCB 38.457 48.755 51.410 53.023 54.306 55.030 1.2.4 DCB 5.379 8.985 9.068 10.363 10.334 10.527 1.2 DCB 25.917 46.181 49.871 53.001 55.309 55.028 1.2.4 DCB 2.501 6.153 8.415 10.327 10.298 10.513 1.2 DCB 22.263 45.160 49.630 52.905 54.305 55.029 1.2.4 DCB 0.212 6.535 8.303 10.312 10.285 10.509 Liquid Phase Concentrations (mg/L) t = 4h Depth (mm) 5.0 10.0 19.1 38.1 63.5 88.9 Toluene 18.412 28.128 41.374 47.728 51.501 51.691 Chlorobenzene 76.58 131.11 154.81 190.45 199.12 200.11 Styrene 24.821 49.123 72.066 86.474 90.023 90.001 n­propylbenzene 5.018 39.021 46.039 50.075 54.296 54.018 Liquid Phase Concentrations (mg/L) t = 8h Depth (mm) 5.0 10.0 19.1 38.1 63.5 88.9 Toluene 12.708 24.490 38.451 45.844 50.988 51.571 Chlorobenzene 32.04 98.73 143.30 188.60 198.52 198.96 Styrene 16.962 32.443 69.121 80.998 89.848 89.982 n­propylbenzene 3.987 38.534 44.905 48.346 54.305 54.029 190 Sand 30­50, Air Flow Rate 55 mL/min, Temperature 21°C Air Phase Concentrations (mg/L) Time (min) 0 1 3 5 10 20 30 45 60 75 90 105 120 150 180 240 300 360 420 480 Benzene 0.0000 2.8318 0.9700 0.6410 0.3928 0.3340 0.2458 0.1858 0.1637 0.1413 0.1360 0.1218 0.1160 0.1103 0.0905 0.0818 0.0790 0.0705 0.0650 0.0630 Ethylbenzene 0.0000 1.5158 0.4053 0.3345 0.1933 0.1830 0.1413 0.1388 0.1318 0.1265 0.1088 0.1058 0.0798 0.0785 0.0613 0.0525 0.0375 0.0363 0.0293 0.0253 m,p­Xylene 0.0000 0.9168 0.4383 0.3515 0.2145 0.2048 0.1288 0.1355 0.0973 0.0858 0.0858 0.0778 0.0665 0.0590 0.0475 0.0355 0.0325 0.0320 0.0275 0.0258 o­Xylene 0.0000 0.5070 0.3585 0.2185 0.1100 0.1075 0.0973 0.0978 0.0613 0.0595 0.0565 0.0525 0.0500 0.0453 0.0390 0.0378 0.0223 0.0209 0.0172 0.0159 Liquid Phase Concentrations (mg/L) Depth (nun) 5.0 lO.O 19.1 38.1 63.5 88.9 Ben zene (0 h) 86.173 111.203 134.683 146.000 147.062 147.531 Benzene (4 h) 51.876 74.212 130.123 135.127 146.567 147.012 Benzene (8 h) 32.257 73.300 128.880 130.959 145.541 146.831 o­Xylene(Oh) 20.445 31.548 40.770 42.640 43.338 43.610 o­Xylene (4 h) 17.713 23.271 37.613 42.512 42.918 43.599 o­Xylene(8h) 11.885 17.622 28.212 40.110 42.624 43.595 Liquid Phase Concentrations (mg/L) Depth (mm) 5.0 10.0 19.1 38.1 63.5 88.9 Ethylbenzene (0 h) 38.100 60.475 65.602 66.750 68.504 68.629 Ethylbenzene (4h) 32.712 41.477 58.456 62.631 68.312 68.218 Ethylbenzene (8 h) 16.174 28.692 50.936 62.476 68.022 68.547 m,p­Xylene (0 h) 45.037 58.553 66.095 72.445 74.409 75.762 m,p­Xylene (4h) 40.118 48.023 62.215 72.234 74.401 75.567 m,p­Xylene (8h) 29.210 37.750 51.091 67.046 74.327 75.726 191 Sand 30­50, Air Flow Rate 55 mL/min, Temperature 21°C Air Phase Concentrations (mg/L) Time (min) 0 1 3 5 10 20 30 40 60 90 120 180 240 300 360 420 Toluene 0.0000 0.4455 0.2120 0.2085 0.1815 0.1715 0.1545 0.1315 0.1085 0.0975 0.0845 0.0790 0.0650 0.0635 0.0630 0.0632 Chlorobenzene 0.0000 1.0545 0.6745 0.5515 0.4900 0.4315 0.4255 0.4095 0.3650 0.3075 0.2865 0.2125 0.1804 0.1785 0.1675 0.1675 Styrene 0.0000 0.4280 0.2500 0.2270 0.2145 0.2015 0.1805 0.1770 0.1690 0.1370 0.1045 0.0895 0.0735 0.0695 0.0687 0.0685 Propylbenzene 0.0000 0.2471 0.1285 0.1250 0.1220 0.1151 0.1045 0.1000 0.0930 0.0885 0.0730 0.0530 0.0404 0.0435 0.0425 0.0425 1.2 DCB 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 1,2.4 TCB 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 Liquid Phase Concentrations (mg/L) t = Oh Depth (mm) 5.0 10.0 19.1 38.1 63.5 88.9 Toluene 33.187 37.917 45.394 45.775 45.928 45.978 Chlorobenzene 108.49 119.35 133.79 148.76 147.90 148.38 Styrene 45.288 57.228 58.627 66.038 66.954 66.901 n­propylbenzene 11.207 18.977 21.068 21.442 23.982 23.630 1.2 DCB 6.308 7.345 7.826 8.448 9.052 9.083 1.2.4 DCB 4.472 4.756 5.266 8.639 9.204 9.230 1.2 DCB 2.137 5.989 6.979 8.127 9.038 9.003 1,2.4 DCB 3.012 3.251 4.267 8.529 9.003 8.998 Liquid Phase Concentrations (mg/L) t=4h Depth (mm) 5.0 10.0 19.1 38.1 63.5 88.9 Toluene 22.619 31.143 39.497 42.467 45.512 45.513 Chlorobenzene 60.07 92.54 127.48 146.98 147.52 148.12 Styrene 31.854 42.073 57.096 66.012 66.871 66.902 n­propylbenzene 5.861 8.012 15.324 21.021 23.137 23.571 192 Liquid Phase Concentrations (mg/L) t = 7h Depth (imn) 5.0 lO.O 19.1 38.1 63.5 88.9 Toluene 18.095 26.016 38.451 41.732 44.500 44.878 Chlorobenzene 49.96 80.60 114.27 131.89 147.29 147.60 Styrene 24.785 38.206 50.852 59.074 65.320 65.685 1.2 DCB 0.113 2.180 6.771 7.895 8.991 8.993 n­propylbenzene 3.130 7.770 14.195 19.188 22.518 23.266 1.2.4 DCB 1.764 2.214 4.038 8.445 8.897 8.918 Sand 30­50, Air Flow Rate 120 mL/min, Temperature 21''C Air Phase Concentrations (mg/L) Time (min) 0 1 3 5 10 20 30 45 60 75 90 105 120 150 180 240 300 360 420 480 Benzene 0.0000 0.6123 0.2485 0.1620 0.1245 0.1100 0.1008 0.0840 0.0802 0.0758 0.0747 0.0718 0.0698 0.0660 0.0535 0.0410 0.0408 0.0408 0.0403 0.0383 Ethylbenzene 0.0000 0.2460 0.1265 0.0973 0.0773 0.0723 0.0715 0.0658 0.0465 0.0448 0.0390 0.0363 0.0338 0.0290 0.0283 0.0260 0.0227 0.0243 0.0213 0.0223 m,p­XyIene 0.0000 0.2268 0.0928 0.0743 0.0625 0.0490 0.0480 0.0385 0.0278 0.0244 0.0242 0.0216 0.0188 0.0183 0.0173 0.0157 0.0123 0.0121 0.0105 0.0102 o­Xyiene 0.0000 0.1493 0.1193 0.0863 0.0653 0.0640 0.0440 0.0425 0.0363 0.0370 0.0315 0.0197 0.0225 0.0180 0.0118 0.0046 0.0030 0.0025 0.0023 0.0019 Liquid Phase Concentrations (mg/L) Depth (mm) 5.0 10.0 19.1 38.1 63.5 88.9 Ben zene (0 h) 72.067 72.911 89.908 90.000 91.286 91.616 Benzene (4 h) 49.761 61.451 83.079 89.071 90.123 91.134 Benzene (8 h) 25.196 51.834 77.154 81.507 89.840 90.253 o­Xylene (0 h) 14.195 15.126 20.982 23.301 24.240 24.553 o­Xyiene (4 h) 9.812 13.117 18.879 21.978 24.102 23.068 o­Xylene (8 h) 5.416 10.527 17.569 21.952 23.319 23.517 193 Liquid Phase Concentradons (mg/L) Depth (mm) 5.0 10.0 19.1 38.1 63.5 88.9 Ethylbenzene (0 h) 20.153 28.771 37.345 39.645 40.354 40.643 Ethylbenzene (4 h) 12.612 23.107 33.211 35.749 39.714 39.898 Ethylbenzene (8 h) 5.391 16.681 28.616 35.758 38.345 39.130 m.p­Xylene (0 h) 18.614 26.613 39.635 40.181 42.951 43.190 m,p­Xylene (4h) 15.237 20.071 33.203 40.054 42.911 40.175 m.p­Xylene (8 h) 4.935 12.226 25.246 38.258 42.866 43.103 Sand 30­50. Air Flow Rate 120 mL/min, Temperature 21°C Air Phase Concentrations (mg/L) Time (min) 0 1 3 5 10 20 30 40 60 75 120 180 240 300 360 480 Toluene 0.0000 0.2163 0.1093 0.0908 0.0773 0.0675 0.0633 0.0568 0.0543 0.0473 0.0463 0.0457 0.0452 0.0468 0.0478 0.0416 Chlorobenzene 0.0000 0.8205 0.3268 0.2933 0.2330 0.2290 0.1945 0.1760 0.1950 0.1553 0.1243 0.1170 0.0935 0.0920 0.0913 0.0720 Styrene 0.0000 0.3463 0.1295 0.1145 0.1108 0.0990 0.0930 0.0938 0.0823 0.0785 0.0655 0.0655 0.0538 0.0528 0.0490 0.0490 Propylbenzene 0.0000 0.2685 0.2435 0.2270 0.0915 0.0670 0.0660 0.0650 0.0484 0.0381 0.0279 0.0343 0.0275 0.0221 0.0212 0.0209 1.2 DCB 0.0000 0.3120 0.2015 0.1185 0.0693 0.0421 0.0335 0.0286 0.0184 0.0152 0.0142 0.0102 0.0096 0.0095 0.0094 0.0089 1.2.4 TCB 0.0000 0.0000 0.0373 0.0805 0.0642 0.­327 0.0171 0.0126 0.0084 0.0064 0.0025 0.0000 0.0000 0.0000 0.0000 0.0000 Liquid Phase Concentrations (mg/L) t = Oh Depth (mm) 5.0 10.0 19.1 38.1 63.5 88.9 Toluene 42.892 47.436 54.505 55.998 56.451 56.456 Chlorobenzene 132.84 183.43 197.07 211,58 217.06 219.59 Styrene 60.768 84.400 87.035 89.622 92.657 92.774 n­propylbenzene 28.764 37.355 41.838 44.154 44.575 45.034 1.2 DCB 33.147 40.939 46.582 49.478 50.033 51.899 1.2.4 DCB 10.057 17.986 19.718 21.004 21.445 21.904 194 Liquid Phase Concentrations (mg/L) t = 4h Depth (mm) 5.0 iO.O 19.1 38.1 63.5 88.9 Toluene 20.064 37.147 48.794 54.617 54.856 56.123 Chlorobenzene 59.89 130.39 192.56 207.31 216.39 218.15 Styrene 28.117 57.234 83.274 86.129 91.993 92.435 n­propylbenzene 6.879 22.321 32.641 41.503 44.901 44.987 1.2 DCB 4.023 36.683 39.241 49.471 50.029 51.798 1.2.4 DCB 5.067 15.214 17.251 20.996 21.405 21.876 1,2 DCB 3.725 35.600 38.802 49.000 49.983 51.883 1,2.4 DCB 2.034 14.037 17.945 20.993 21.435 21.900 Liquid Phase Concentrations (mg/L) t = 7h Depth (mm) 5.0 10.0 19.1 38.1 63.5 88.9 Toluene 9.929 33.034 48.522 52.489 54.015 55.898 Chlorobenzene 54.32 95.56 178.72 201.12 215.35 215.43 Styrene 22.479 46.441 79.494 85.493 91.069 91.551 n­propylbenzene 5.872 12.626 30.856 40.018 44.626 44.597 Sand 70­100, Air Flow Rate 10 mL/min, Temperature 21°C Air Phase Concentrations (mg/L) Time (min) 0 1 3 5 11 20 30 45 60 80 90 120 150 180 240 300 360 420 480 Benzene 0.0000 0.1700 0.6633 1.0665 1.5525 1.6368 1.2728 1.1580 1.0565 0.9275 0.8948 0.8333 0.7250 0.7080 0.6168 0.5667 0.5387 0.5368 0.5240 Ethylbenzene 0.0000 0.0007 0.0443 0.0983 0.1838 0.1938 0.1923 0.1855 0.1873 0.1855 0.1848 0.1820 0.1790 0.1728 0.1718 0.1685 0.1270 0..1353 0.1225 m,p­Xylene 0.0000 0.0007 0.0425 0.0958 0.2005 0.2165 0.2140 0.1993 0.1955 0.1873 0.1928 0.1960 0.1990 0.1980 0.1835 0.1723 0.1450 0.1558 0.1432 o­Xylene 0.0000 0.0001 0.0227 0.0440 0.0980 0.1058 0.1058 0.0935 0.0090 0.0920 0.0945 0.0930 0.0943 0.0943 0.0915 0.0883 0.0790 0.0608 0.0654 195 Liquid Phase Concentrations (mg/L) Depth (mm) 5.0 10.0 15.0 19.1 38.1 63.5 88.9 Ben zene (Oh) 81.586 82.511 84.322 91.220 92.261 95.729 95.699 Benzene (4 h) 52.871 78.341 81.269 89.614 91.897 95.613 95.456 Benzene (8 h) 31.062 34.496 77.616 82.026 89.480 94.644 95.114 o­Xylene(Oh) 12.566 14.156 15.009 16.716 17.918 17.849 17.938 o­Xylene(4h) 8.727 11.613 14.174 16.112 17.645 17.843 17.932 o­Xylene(8h) 5.895 8.330 13.079 15.512 16.949 17.841 17.927 Liquid Phase Concentrations (mg/L) Depth (mm) 5.0 10.0 15.0 19.1 38.1 63.5 88.9 Ethylbenzene (0 h) 18.512 19.304 20.171 22.735 26.279 26.731 26.616 Ethylbenzene (4 h) 12.217 15.217 20.007 21.329 25.714 25.929 26.501 Ethylbenzene (8 h) 7.126 10.068 16.137 21.264 23.866 25.846 26.411 m,p­Xylene (0 h) 23.006 24.601 26.157 27.106 33.333 33.446 33.255 m,p­Xylene (4 h) 15.126 22.147 25.789 26.692 31.879 32.862 32.989 m.p­Xylene (8 h) 10.288 16.300 24.056 26.195 30.010 32.812 32.890 Sand 70­100. Air Flow Rate 10 mL/min, Temperature 21°C Air Phase Concentrations (mg/L) Time (min) 0 1 3 5 10 20 30 40 60 90 105 120 150 180 240 300 360 420 480 Toluene 0.0000 0.1078 0.1162 0.1200 0.1615 0.1560 0.2650 0.2845 0.3035 0.3780 0.3740 0.3605 0.3015 0.2810 0.2795 02780 0.2600 0.2470 0.2435 Chlorobenzene 0.00(X) 0.1280 0.1195 0.1610 0.1620 0.2880 0.3085 0.4735 0.4965 0.5205 0.4088 0.4075 0.4075 0.3965 0.3659 0.3364 0.2995 0.2745 0.2237 Styrene 0.0000 0.0760 0.0765 0.0820 0.0875 0.1235 0.1275 0.1365 0.1920 0.2190 0.2360 0.2020 0.1885 0.1860 0.1985 0.1775 0.1675 0.1654 0.1570 Propylbenzene 0.0000 0.0166 0.0244 0.0266 0.0369 0.0396 0.0540 0.0845 0.1200 0.1330 0.1485 0.1417 0.1367 0.1315 0.1285 0.1040 0.0925 0.0875 0.0798 1,2 DCS 0.0000 0..0000 0.0000 0.0000 0.0212 0.0212 0.0267 0.0419 0.0425 0.0447 0.0425 0.0402 0.0390 0.0373 0.0339 0.0287 0.0281 0.0127 0.0097 1,2,4 TCB 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0042 0.0089 0.0142 0.0118 0.0036 0.0000 0.0000 0.0000 0.0000 196 Liquid Phase Concentrations (mg/L) t = Oh Depth (mm) 5.0 10.0 15.0 19.1 38.1 63.5 88.9 Toluene 37.534 37.534 38.070 41.579 41.860 41.809 41.148 Chlorobenzene 59.558 65.902 68.043 68.995 70,185 72,643 72.722 Styrene 36.971 39.030 40.831 44.629 49,569 50.085 50.149 n­propylbenzene 23.206 25.327 31.154 31.154 31.522 33.597 33.692 1.2 DCB 22.849 22.909 22.949 27.937 28.492 28.733 28.894 1.2.4 DCB 14.610 15.447 16.741 21.111 22.269 22.218 22.201 1.2 DCB 13.281 18.129 21.873 26.317 28.328 28.339 28.792 1.2.4 DCB 10.211 14.689 15.417 20.128 21.818 22.150 22.161 1.2 DCB 10.585 13.953 21.722 25.196 28.182 28.320 28.818 1.2.4 DCB 9456 14.579 15.376 19.702 21.798 22.147 22.165 Liquid Phase Concentrations (mg/L) t=4h Depth (mm) 5.0 10.0 19.1 38.1 63.5 88.9 Toluene 30.008 34.834 37.962 40.867 40.656 40.718 40.079 Chlorobenzene 44.917 58.012 67.249 68.331 70.021 71.871 72.678 Styrene 29.972 34.127 38.861 40.173 48.911 49.912 50.121 n­propylbenzene 18.013 24.274 25.318 29.213 31.139 33.415 33.716 Liquid Phase Concentrations (mg/L) t = 8h Depth (nun) 5.0 10.0 19.1 38.1 63.5 88.9 Toluene 18.558 27.007 31.356 37.532 40.305 40.509 40.947 Chlorobenzene 30.807 45.045 61.039 66,299 69.809 71.020 72,556 Styrene 25.149 28.490 32.937 37.242 48.253 49.418 49.956 n­propylbenzene 14.195 17.4823 24.459 27.269 30.781 33.407 33.645 197 Sand 70­100, Air Flow Rate 25 mL/min, Temperature 21°C Air Phase Concentrations (mg/L) Time (min)1 0 1 3 5 II 20 30 40 60 75 90 120 150 180 240 300 360 420 Benzene 0.0000 0.7748 1.0095 1.1685 1.0182 0.9460 0.8040 0.6930 0.6162 0.5537 0.4555 0.4058 0.3180 0.2120 0.1475 0.1375 0.1341 0.1335 m,p­Xylene 0.0000 0.0625 0.1020 0.1440 0.1460 0.1923 0.1943 0.1923 0.1925 0.1673 0.1550 0.1183 0.1077 0.1016 0.0921 0.0871 0.0851 0.0798 Ethylbenzene 0.0000 0.0685 0.1218 0.1345 0.1560 0.1938 0.1898 0.1788 0.1705 0.1548 0.1370 0.1130 0.1003 0.0927 0.0912 0.0906 0.0880 0.0825 o­Xylene 0.0000 0.0189 0.0613 0.0845 0.0978 0.1255 0.1615 0.1603 0.1190 0.1014 0.1045 0.0858 0.0743 0.0625 0.0490 0.0483 0.0458 0.0365 Liquid Phase Concentrations (mg/L) Depth (mm) 5.0 10.0 15.0 19.1 38.1 63.5 88.9 Ben zene (0 h) 48.924 50.398 60.603 63.426 65.743 66.517 67.255 Benzene (4 h) 28.791 44.819 50.816 60.039 63.211 65.198 67.021 Benzene (7 h) 6.257 18.218 46.443 53.428 63.049 63.361 66.890 o ­Xylene (0 h) 9.851 9.925 10.029 10.833 11.460 11.699 11.937 o­Xylene (4 h) 4.561 7.373 9.913 10.476 10.984 11.012 11.871 o­Xylene (7 h) 1.372 4.904 8.425 9.525 10.782 10.937 11.458 Liquid Phase Concentrations (mg/L) Depth (mm) 5.0 10.0 15.0 19.1 38.1 63.5 88.9 Ethylbenzene (0 h) 13.450 14.214 15.948 17.089 17.735 17.946 17.933 Ethylbenzene (4 h) 6.256 12.961 14.003 16.511 17.129 17.728 17.471 Ethylbenzene (7h) 2.635 5.500 8.632 15.545 16.763 17.650 17.305 m,p­Xylene (Oh) 14.857 15.175 20.097 21.950 22.551 22.803 23.158 m,p­Xylene (4 h) 7.778 12.496 14.691 21.687 22.098 22.758 23.123 m.p­XyIene (7h) 2.138 7.465 11.870 19.985 21.640 22.665 23.081 198 Sand 70­100, Air Flow Rate 25 mL/min, Temperature 21°C Air Phase Concentrations (mg/L) Time (min) 0 1 3 5 12 21 30 45 60 90 120 150 180 210 300 360 420 480 Toluene 0.0000 0.3235 0.3555 0.3395 0.2980 0.2685 0.2520 0.2340 0.2125 0.1880 0.1765 0.1645 0.1525 0.1445 0.1165 0.1145 0.1107 0.1105 Chlorobenzene 0.0000 0.1280 0.3705 0.4555 0.5340 0.4375 0.4295 0.3905 0.3760 0.3740 0.3215 0.2930 0.2560 0.2355 0.2025 0.1895 0.1765 0.1734 Styrene 0.0000 0.0750 0.1585 0.2035 0.22130 0.2090 0.1955 0.1935 0.1930 0.1705 0.1540 0.1345 0.1315 0.1150 0.1014 0.1001 0.0954 0.0875 Propylbenzene 0.0000 0.0450 0.0820 0.1230 0.1270 0.1190 0.1185 0.1010 0.0995 0.0965 0.0935 0.0705 0.0675 0.0655 0.0615 0.0595 0.0475 0.0456 1.2 DCB 0.0000 0.0436 0.0447 0.0476 0.0487 0.0453 0.0419 0.0402 0.0367 0.0321 0.0333 0.0321 0.0203 0.0150 0.0127 0.0120 0.0114 0.0123 1.2.4 TCB 0.0000 0.0000 0.0000 0.0600 0.1533 0.0142 0.0136 0.0112 0.0107 0.0107 0.0081 0.0089 0.0042 0.0000 0.0000 0.0000 0.0000 0.0000 Liquid Phase Concentrations (mg/L) t = Oh Depth (mm) 5.0 10.0 15.0 19.1 38.1 63.5 88.9 Toluene 37.536 39.061 39,290 39.748 41.503 41.732 41.642 Chlorobenzene 59.559 68.123 69.233 70.264 70.264 70.310 70.666 Styrene 36.971 38.063 39.783 40.924 41.418 41.476 41.556 n­propylbenzene 25.925 26.821 29.660 29.959 34.890 35.131 35.029 1.2 DCB 25.159 28.239 29.446 30.814 31.457 31.593 31.652 1,2,4 DCB 17.554 17.943 18.029 18.419 18.764 19.055 19.004 1.2 DCB 15.793 22.769 26.054 29.347 30.684 31.491 31.562 1.2,4 DCB 10.134 16.758 17.673 18.312 18.761 19.038 19.001 Liquid Phase Concentrations (mg/L) t = 4h Depth (mm) 5.0 10.0 15.0 19.1 38.1 63.5 88.9 Toluene 17.839 33.219 36.114 38.093 40.713 40.825 40.914 Chlorobenzene 38.271 62.311 70.003 64.617 67.121 70.128 70.617 Styrene 21.756 33.214 35.105 37.586 40.002 41.419 41.512 n­propylbenzene 20.023 25.275 29.295 29.839 32.073 33.991 35.017 199 Liquid Phase Concentrations (mg/L) t = 8h Depth (mm) 5.0 10.0 15.0 19.1 38.1 63.5 88.9 Toluene 8.9392 28.838 32.119 35.010 39.290 40.214 40.596 Chlorobenzene 28.111 41.063 50.028 60.531 63.109 69.568 69.809 Styrene 16.855 25.596 27.984 35.242 39.753 41.206 41.272 1.2 DCS 10.459 18.562 25.297 27.113 30.320 31.592 31.630 n­propylbenzene 14.195 17.109 23.011 27.143 31.614 33.792 34.913 1.2.4 DCS 6.826 8.389 16.133 17.238 18.651 19.043 19.985 Sand 70­100, Air Flow Rate 55 mL/min, Temperature IfC Air Phase Concentrations (mg/L) Time (min) 0 1 3 5 13 23 32 45 60 75 90 120 180 240 300 360 420 480 Benzene 0.0000 1.4158 0.7283 0.5193 0.3468 0.2528 0.2233 0.1718 0.1248 0.1165 0.1170 0.0878 0.0900 0.0665 0.0678 0.0600 0.0515 0.0424 Ethylbenzene 0.0000 0.2078 0.1655 0.0933 0.0893 0.0703 0.0610 0.0485 0.0435 0.0423 0.0363 0.0325 0.0325 0.0278 0.0248 0.0195 0.0152 0.0131 m.p­Xylene 0.0000 0.2288 0.1485 0.0998 0.0960 0.0835 0.0578 0.0523 0.0473 0.0465 0.0418 0.0327 0.0280 0.0265 0.0183 0.0171 0.0169 0.0144 o­Xylene 0.0000 0.0820 0.1245 0.0640 0.0540 0.0595 0.0535 0.0428 0.0380 0.0373 0.0370 0.0318 0.0313 0.0258 0.0155 0.0042 0.0040 0.0035 Liquid Phase Concentrations (mg/L) Depth (mm) 5.0 10.0 15.0 19.1 38.1 63.5 88.9 Ben zene (0 h) 45.540 69.099 73.878 79.963 82.279 82.819 83.050 Benzene (4 h) 13.147 42.182 73.213 76.021 78.763 81.027 82.917 Benzene (8 h) 9.981 41.776 65.130 69.716 77.748 80.813 81.778 o­Xylene (0 h) 7.693 10.735 12.915 16.691 17.328 17.314 17.315 o­Xylene (4 h) 2.247 6.147 11.675 15.204 15.906 17.098 17.101 o­Xylene (8 h) 1.559 5.973 9.703 14.604 15.991 16.871 17.003 200 Liquid Phase Concentrations (mg/L) Depth (mm) 5.0 10.0 15.0 19.1 38.1 63.5 88.9 Ethylbenzene (Oh) 8.521 19.455 20.736 20.759 20.887 21.264 21.312 Ethylbenzene (4h) 4.766 11.319 19.073 20.341 20.779 21.931 21.012 Ethylbenzene (8 h) 1.109 8.370 12.970 18.323 20.508 20.736 20.736 m,p­Xylene (Oh) 9.643 21.412 24.069 25.170 25.321 25.511 25.967 m,p­Xylene (4h) 3.231 11.412 19.746 23.011 24.812 24.915 25.786 m.p­Xylene (8h) 1.298 11.313 17.198 22.513 24.158 24.247 25.260 Sand 70­100, Air Flow Rate 55 mL/min, Temperature 21°C Air Phase Concentrations (mg/L) Time (min) 0 1 3 5 12 21 30 40 60 120 180 240 300 360 420 480 Toluene 0.0000 0.3250 0.2220 0.2085 0.1760 0.1690 0.1430 0.1320 0.1320 0.1285 0.1115 0.1025 0.1017 0.1010 0.0901 0.0906 Chlorobenzene 0.0000 0.2198 0.1589 0.1583 0.1215 0.1060 0.0990 0.0970 0.0956 0.0873 0.0848 0.0838 00798. 0.0745 0.0695 0.0671 Styrene 0.0000 0.1790 0.1235 0.1150 0.1140 0.1075 0.0980 0.0955 0.0895 0.0875 0.0770 0.0765 0.0730 0.0725 0.0721 0.0700 Propylbenzene 0.0000 0.0630 0.0525 0.0478 0.0378 0.0374 0.0279 0.0218 0.0190 0.0108 0.0097 0.0047 0.0046 0.0041 0.0040 0.0040 1.2 DCB 0.0000 0.0228 0.2150 0.0228 0.0168 0.0159 0.0141 0.0128 0.0018 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 1.2,4 TCB 0.0000 0.0000 0.0000 0.0600 0.0010 0.0009 0.0007 0.0006 0.0005 0.0004 0.0003 0.0003 0.0003 0.0000 0.0000 0.0000 Liquid Phase Concentrations (mg/L) t = Oh Depth (mm) 5.0 10.0 15.0 19.1 38.1 63.5 88.9 Toluene 33.113 35.449 37.399 39.162 40.664 41.411 41.563 Chlorobenzene 42.269 46.869 59.320 60.113 63.920 63.753 63.777 Styrene 28.891 30.196 33.366 33.678 35.007 34.977 35.029 n­propylbenzene 7.787 9.792 16.959 18.454 20.023 20.741 20.965 1,2 DCB 5.294 6.026 8.045 12.631 12.953 13.275 13.619 1.2.4 DCB 4.696 10.636 13.317 15.095 16.689 17.467 17.445 201 Liquid Phase Concentrations (mg/L) t=4b Depth (mm) 5.0 10.0 15.0 19.1 38.1 63.5 88.9 Toluene 19.934 24.117 28.412 33.789 38.234 41.119 41.321 Chlorobenzene 27.713 32.219 39.782 57.890 63.713 63.561 63.681 Styrene 22.619 25.117 27.431 32.129 34.349 34.458 34.992 n­propylbenzene 6.313 8.813 12.747 18.407 18.399 20.713 20.899 1.2 DCB 3.475 5.001 7.096 10.819 13.009 13.198 13.602 1.2.4 DCB 3.569 8.545 11.832 15.602 15.978 17.418 17.431 1.2 DCB 0.291 2.711 5.524 10.659 12.913 13.172 13.599 1,2.4 DCB 0.147 8.007 9.858 15.457 15.985 17.394 17.421 Liquid Phase Concentrations (mg/L) t = 8b Depth (mm) 5.0 10.0 15.0 19.1 38.1 63.5 88.9 Toluene 9.174 15.844 20.888 28.865 37.705 40.901 41.029 Chlorobenzene 12,223 20.426 27.374 53.482 63.444 63.709 63.057 Styrene 10.681 17.233 20.115 26.808 30.408 33.995 37.747 n­propylbenzene 2.090 3.736 12.439 17.643 18.408 20.686 20.950 202 Sand 70­100, Air Flow Rate 120 mL/min, Temperature 2l°C Air Phase Concentrations (mg/L) Time (min) 0 I 2 5 11 23 30 40 60 90 120 180 240 300 360 420 480 Benzene 0.0000 0.7180 0.4368 0.2555 0.2003 0.1113 0.1030 0.0665 0.0588 0.0580 0.0480 0.0373 0.0368 0.0373 0.0308 0.0308 0.0301 m.p­Xylene 0.0000 0.1318 0.0743 0.0660 0.0388 0.0235 0.0236 0.0260 0.0201 0.0157 0.0107 0.0076 0.0071 0.0633 0.0058 0.0057 0.0047 Ethylbenzene 0.0000 0.1143 0.0735 0.0680 0.0490 0.0453 0.0218 0.0170 0.0174 0.0137 0.0122 0.0115 0.0111 0.0105 0.0100 0.0099 0.0091 o­Xylene 0.0000 0.0923 0.1173 0.0950 0.0698 0.0550 0.0345 0.0328 0.0304 0.0278 0.0237 0.0109 0.0087 0.0071 0.0069 0.0065 0.0067 Liquid Phase Concentrations (mg/L) Depth (mm) 5.0 10.0 15.0 19.1 38.1 63.5 88.9 Ben zene (0 h) 56.728 61.469 81.933 84.052 85.478 86.437 88.368 Benzene (4 h) 27.091 46.371 63.761 82.189 84.319 84.231 88.127 Benzene (8 h) 12.178 18.344 40.273 78.368 83.819 84.091 87.983 o­Xylene (0 h) 8.555 9.375 10.693 12.131 12.533 12.793 12.909 o­Xylene (4 h) 2.137 4.891 5.991 10.914 11.812 12.217 12.605 o­Xylene (8 h) 1.757 3.176 4.227 10.455 11.403 12.156 12.600 Liquid Phase Concentrations (mg/L) Depth (mm) 5.0 10.0 15.0 19.1 38.1 63.5 88.9 Ethylbenzene (Oh) 12.442 14.855 15.647 18.925 19.304 21.754 22.525 Ethylbenzene (4 h) 3.291 6.871 10.237 18.734 19.297 21.553 22.386 Ethylbenzene (8 h) 0.600 3.576 6.542 18.632 19.117 21.057 22.079 m,p­Xylene (Oh) 9.339 12.225 17.719 18.906 18.902 18.906 18.817 fji,p­Xylene (4h) 4.197 7.813 12.625 17.831 18.321 18.607 18.819 m.p­Xylene (8h) 0.203 3.731 7.492 16.490 17.937 18.467 18.782 203 Sand 70/100, Air Flow Rate 120 mLymin, Temperature 21°C Air Phase Concentrations (mg/L) Time (min) 0 I 3 5 12 21 31 45 60 90 120 180 210 300 360 420 480 Toluene 0.0000 0.1165 0.0795 0.0713 0.0643 0.0625 0.0563 0.0555 0.0548 0.0530 0.0503 0.0468 0.0468 0.0453 0.0448 0.0437 0.0384 Chlorobenzene 0.0000 0.1578 0.1103 0.0953 0.0970 0.0900 0.0880 0.0820 0.0800 0.0735 0.0700 0.0598 0.0598 0.0573 0.0503 0.0456 0.0416 Siyrene 0.0000 0.1690 0.1170 0.1025 0.1005 0.0960 0.0885 0.0855 0.0855 0.0830 0.0875 0.0685 0.0575 0.0405 0.0385 0.0276 0.0274 Propylbenzene 0.0000 0.0965 0.0615 0.0456 0.0423 0.0495 0.0364 0.0260 0.0259 0.0141 0.0135 0.0116 0.0091 0.0088 0.0081 0.0079 0.0079 1.2 DCB 0.0000 0.0000 0.0230 0.0235 0.0213 0.0189 0.0109 0.0115 0.0067 0.0058 0.0038 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 1.2.4 TCB 0.0000 0.0000 0.0100 0.0292 0.0142 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 Liquid Phase Concentrations (mg/L) t = Oh Depth (mm) 5.0 10.0 15.0 19.1 38.1 63.5 88.9 Toluene 32.119 32.424 44.179 47.148 47.612 47.835 47.971 Chlorobenzene 47.662 52.500 70.958 72.333 75.156 75.314 75.574 Styrene 33.878 36.866 38.678 41.253 46.688 47.305 47.482 n­propylbenzene 18.229 23.534 25.665 29.286 30.974 31.457 32.262 1.2 DCB 19.309 25.182 25.665 28.883 30.374 30.457 30.462 1.2,4 DCB 12.452 14.398 14.830 14.700 14.916 15.511 15.560 1.2 DCB 5.671 17.761 22.043 27.567 30.298 30.312 30.315 1.2.4 DCB 6.251 8.721 12.155 14.121 14.909 15.413 15.467 Liquid Phase Concentrations (mg/L) t = 4h Depth (mm) 5.0 10.0 15.0 19.1 38.1 63.5 88.9 Toluene 19.918 28.224 33.719 47.023 47.412 46.917 47.018 Chlorobenzene 29.471 45.014 64.032 70.271 72.693 74.876 75.108 Styrene 18.071 24.997 32.003 32.997 43.627 47.291 47.397 n­propylbenzene 9.012 18.789 20.178 28.347 30.931 31.358 32.213 204 Liquid Phase Concentrations (mg/L) t = 8h Depth (mm) 5.0 10.0 15.0 19.1 38.1 63.5 88.9 Toluene 8.692 15.964 21.593 41.075 44.735 46.053 46.835 Chlorobenzene 8.884 29.546 54.218 63.200 72.444 74.516 74.960 Styrene 5.962 10.091 18.067 24.172 43.589 46.998 47.180 n­propylbenzene 4.776 10.609 18.761 26.408 30.861 31.395 32.199 1.2 DCB 2.896 10.316 20.574 26.688 30.297 30.259 30.455 1.2.4 DCB 2.36! 6.485 11.745 13.360 14.843 15.473 15.459 Mass Balance Calculations Initial mass of VOC X,­£ilVXC:,AZ. Final mass of VOC X, =eLWY,cl,AZ, VOC mass removed i'=e,lc„A», Mass balance difference P= X , ­ X f ­ Y Where Ck„ for (/ = o or/) is the average concentration of VOC in the aqueous phase in the layer of thickness L is the length of the single­air channel semp (17.5 cm), W is the width of the single­air setup (5 cm), £ is the porosity of the porous media, Qa is the air flow rate (L/min), and Atn is the time interval elapsed between two consecutive air phase samples 205 Ottawa Sand, Air Flow Rate 10 mL/min voc Xoimg) Benzene Ethylbenzene ni,p­Xylenes o­Xylene Toluene Chlorobenzene Styrene n­Propylbenzene 1,2­Dichlorobenzene 1.2,4­trichlorobenzene 41.38 13.47 16.05 10.89 24.35 25.12 15.48 8.38 16.52 3.58 Y (mg) P(nifi) 4.99 1.45 1.51 0.92 1.80 1.60 0.98 0.91 0.97 0.31 0.59 ­0.13 ­0.25 0.16 0.24 0.20 ­0.10 ­0.01 ­0.02 ­0.09 35.80 12.15 14.80 9.81 22.31 23.32 14.60 7.48 15.57 3.36 Ottawa Sand, Air Flow Rate 25 mL/min VOC Benzene Ethylbenzene m,p­Xylenes o­Xylene Toluene Chlorobenzene Styrene n­Propylbenzene 1,2­Dichlorobenzene 1.2,4­trichlorobenzene Xo(mg) Xf(m) Y (mg) Peng) 36.39 10.42 10.38 8.04 12.45 12.43 8.34 9.93 3.30 1.64 33.71 9.71 9.43 7.41 10.38 10.53 7.29 8.43 2.98 1.58 2.15 0.61 0.64 0.50 2.36 2.28 1.30 1.42 0.34 0.13 0.53 0.10 0.31 0.13 ­0.29 ­0.38 ­0.25 0.08 0.06 ­0.07 Y(mg) P (nig) 2.87 0.82 0.72 0.49 4.20 5.11 3.44 2.17 1.53 0.92 0.63 0.36 0.45 0.14 ­0.27 ­0.72 ­0.36 ­0.26 0.56 0.10 Ottawa Sand, Air Flow Rate 55 mL/min VOC Benzene Ethylbenzene nn,p­Xy[enes o­Xyiene Toluene Chlorobenzene Styrene n­Propylbenzene 1,2­Dichlorobenzene 1.2,4­trichlorobenzene Xo(mg) 40.75 10.41 10.94 9.21 28.73 32.20 20.48 13.20 20.34 6.55 37.25 9.23 9.77 8.58 24.80 27.81 17.40 11.29 18.25 5.53 206 Ottawa Sand, Air Flow Rate 120 mL/min voc Benzene Ethylbenzene m,p­Xylenes o­Xylene Toluene Chlorobenzene Styrene n­Propylbenzene 1,2­Dichlorobenzene 1.2.4­trichlorobenzene Xo(mg) Xffmg) y <mg) P (mg) 44.67 20.15 24.14 14.29 12.53 13.79 39.91 17.80 21.84 12.80 10.60 11.58 7.10 7.16 7.48 2.42 3.78 2.06 1.95 1.24 2.84 3.11 1.74 1.00 1.04 0.98 0.29 0.35 0.25 ­0.91 ­0.90 ­0.41 ­0.44 ­0.13 ­0.28 8.60 8.46 8.65 3.18 1.91 Sand 30­50, Air Flow Rate 10 mLymin VOC Benzene Ethylbenzene ni,p­Xylenes o­Xylene Toluene Chlorobenzene Styrene n­Propylbenzene 1,2­Dichlorobenzene 1.2.4­trichlorobenzene Xo(mg) Xf(mg) Y(mg) P (mg) 19.95 3.54 3.76 3.24 12.04 43.32 20.85 10.44 16.81 4.16 18.03 3.15 3.40 2.99 11.00 39.12 18.88 9.38 16.17 3.86 1.81 0.34 0.29 0.19 0.99 3.18 1.32 0.71 0.14 0.05 0.11 0.05 0.07 0.06 0.05 1.02 0.65 0.35 0.50 0.25 Sand 30­50, Air Flow Rate 25 mLymin VOC Benzene Ethylbenzene m,p­Xylenes o­Xylene Toluene Chlorobenzene Styrene n­Propylbenzene 1,2­Dichlorobenzene 1.2.4­trichlorobenzene Xo(mg) Xf(mg) Y (ma) P (niRj 46.40 18.39 20.30 11.99 17.93 69.82 30.88 18.26 18.68 3.49 41.58 16.25 18.34 10.58 15.75 61.21 27.36 16.87 18.14 3.27 5.80 2.67 2.37 1.61 2.72 8.37 3.08 0.96 0.25 0.03 ­0.98 ­0.53 ­0.41 ­0.20 ­0.54 0.24 0.44 0.43 0.29 0.19 207 Sand 30­50, Air Flow Rate 55 mL/min voc Benzene Ethylbenzene m.p­Xylenes o­Xylene Toluene Chlorobenzene Styrene n­Propylbenzene 1.2­Dichlorobenzene 1.2,4­trichlorobenzene Xo( m g ) 49.49 23.23 25.05 14.47 15.83 50.54 22.65 7.76 3.02 2.84 Xf(mg) Y(mg) P (nig) 45.77 20.95 23.11 13.13 14.35 45.84 20.56 6.66 2.65 2.58 3.23 1.95 1.64 1.15 1.98 5.71 2.38 1.43 0.00 0.00 0.49 0.33 0.30 0.19 ­0.50 ­1.01 ­0.29 ­0.33 0.37 0.26 Ottawa Sand, Air Flow Rate 120 mL/min VOC Benzene Ethylbenzene m,p­Xylenes o­Xylene Toluene Chlorobenzene Styrene n­Propylbenzene 1.2­Dichlorobenzene 1.2,4­trichlorobenzene Xo(mg) Xf(mg) Y(mg) P(mg) 31.54 13.46 14.00 8.00 19.45 73.46 31.57 15.22 17.21 7.22 28.43 11.82 12.54 7.24 17.53 67.37 28.88 13.26 16.01 7.19 3.35 1.91 1.14 0.88 2.40 6.04 3.14 1.73 0.82 0.23 ­0.24 ­0.27 0.32 ­0.12 ­0.48 0.05 ­0.45 0.23 0.39 ­0.19 Sand 70­100, Air Flow Rate 10 mL/min VOC Benzene Ethylbenzene m.p­Xylenes o­Xylene Toluene Chlorobenzene Styrene n­Propylbenzene 1,2­Dichlorobenzene 1.2.4­trichlorobenzene Xo( m g ) Xf( m g ) Y(mg) P(mg) 37.67 9.63 12.04 6.60 15.77 27.10 18.34 12.23 10.68 8.11 34.67 8.76 11.14 6.16 14.55 25.57 17.24 11.51 10.00 7.83 3.12 0.69 0.76 0.37 1.33 1.68 0.85 0.51 0.14 0.02 ­0.12 0.18 0.14 0.07 ­0.11 ­0.15 0.25 0.22 0.54 0.26 208 Sand 70­100, Air Flow Rate 25 mL/min voc Xo (ma) Xf(mg) Y(mR) P(mg) Benzene Ethylbenzene m,p­Xylenes o­Xylene Toluene Chlorobenzene Styrene n­ftopylbenzene 1,2­Dichlorobenzene 1.2,4­trichlorobenzene 24.67 6.65 8.38 4.37 15.76 26.77 15.75 12.81 11.86 7.21 22.01 5.85 7.61 3.81 14.10 24.05 14.40 11.73 11.00 6.67 3.32 1.15 1.19 0.74 2.06 3.49 1.71 0.96 0.29 0.09 ­0.66 ­0.35 ­0.42 ­0.18 0.40 ­0.77 ­0.36 0.12 0.57 0.44 Sand 70­100, Air Flow Rate 55 mL/min VOC Benzene Ethylbenzene m,p­Xylenes o­Xylene Toluene Chlorobenzene Styrene n­Propylbenzene 1,2­Dichlorobenzene 1,2,4­trichlorobenzene Xo (mg) Xf(mg) Y(nig) P(mg) 30.40. 7.77 9.34 6.20 15.43 23.48 13.15 7.27 4.68 6.00 27.75 7.01 8.44 5.64 13.44 21.33 11.48 6.78 4.36 5.78 2.89 0.90 0.90 0.67 2.69 2.09 2.00 0.25 0.06 0.02 ­0.24 ­0.14 0.00 ­0.11 ­0.70 0.06 ­0.33 0.24 0.26 0.20 Sand 70­100, Air Flow Rate 120 mL/min VOC Benzene Ethylbenzene m,p­Xylenes o­Xylene Toluene Chlorobenzene Styrene n­Propylbenzene 1.2­Dichlorobenzene 1,2,4­trichloroben2ene Xo(mg) Xf(mg) Y(mg) P(mg) 32.05 7.70 6.88 4.69 17.65 27.70 17.27 11.50 11.20 5.79 28.85 6.91 6.12 4.05 15.62 25.30 14.75 10.68 10.31 5.26 2.77 0.77 0.61 0.91 2.49 3.18 3.06 0.71 0.14 0.03 0.43 0.02 0.15 ­0.23 ­0.16 ­0.78 ­0.54 0.11 0.75 0.50 309 APPENDIX C. ONE­D DIFFUSION MODEL. COMPUTER CODES 210 Computer Code for Ethylbenzene, Single Air Channel Setup TITLE THIS PROGRAM COMPUTES THE CONCENTRATION TITLE PROFILE OF VOCs IN THE SOIL MATRIX FOR TITLE A SINGLE AIR CHANNEL SETUP ASSUMING TITLE ID MOVEMENT OF THE CHEMICALS FROM THE SOIL MATRIX TITLE TO THE AIR CHANNEL, WHERE D REPRESENTS THE MOLECULAR TITLE DIFFUSION COEFFICIENT, ALPHA THE TORTUOSITY FACTOR, TITLE K REPRESENTS THE MASS TRANSFER TITLE COEFFICIENT BETWEEN LIQUID AND AIR PHASES, PORO REPRESENTS TITLE THE POROSITY AND KH REPRESENTS THE DIMENSIONLESS TITLE HENRY'S LAW CONSTANT FOR THE VOC TITLE THE SOIL MATRIX HAS BE12N DIVIDED IN 26 LAYERS TITLE OF VARIABLE THICKNESS DZ PARAMETER D=0.0004578,K=0.00514, AREA=87.5, PORO=.4, KH=.294 PARAMETER ALPHA=.47, Q=115.0, V=13.825 STORAGE FLUX(27) , DZ{2S), C(26), Z(26) / DIMENSION FNET(26), MCON(26), IMCON(26) / EQUIVALENCE (FNETl,FNET(1)), (MCONl,MCON(1)), (IMCONl, IMCON(l)) TABLE DZ(1­18)=18*.25, DZ(19­21)=3*.50,DZ(22­26)=5*1. INITIAL NOSORT FUNCTION CURVE1=0, .00001,1, .1143,2, .073 5,5, .068,11, .049, ... 23, .0453,30, .0218,40, .017,60, .0174,90, .0137, ... 120,.0122,180,.0115,240,.0111,300,.0105,360,.01,... 420, .00987,480, .00908 FUNCTION CURVE2 = 0, .009953, .5,0.012442,1.0, .014 855,1.5, .01554 7, ... 1.905, .018925,3.81, .0193 04,6.35, .021754,8.89, .022525,11.0, .0 22 525 R(l)=.5*DZ(1) DO 2 1=1,26 Z(I+1)=2(1)+.5*(DZ(I+1)+DZ(I)) 2 C(I)=AFGEN(CURVE2,R(I)) C(26)=0.022525 DO 5 1=1,26 5 IMCON(I)=PORO*C(I)*D(Z) *AREA DYNAMIC NOSORT CAVE=(C(1)+C(2)+C(3)+C(4)+C(5)+C(6)+C(7)+C(8)+C(9))/9 CAIR=AFGEN(CURVEl,TIME) FLUX(1)=­K*{( (CAVE)*KH)) *AREA CC=((.1143/1000)+FLUX(1)/Q)*(EXP(­Q*(TIME­.5)/V))­FLUX(1)/Q FIXED I NOSORT DO 10 1=2,26 10 FLUX (I) = (C(I­l) ­C(I) ) *D*ALPHA*AREA/ (Z(I)­Z(I­l)) FLUX(27)=0 FIXED I NOSORT DO 20 1=1,26 20 FNET(I)=FLUX(I)­FLUX(I+1) MCONl = INTGRL(IMCONl,FNETl, 26 ) FIXED I NOSORT DO 30 1=1,26 3 0 C(I) =MCON(I) / (PORO*DZ(I)•AREA) FLIX1=FLUX(1) MTRANS=­INTGRL(0.0,FLUXl) AIRPH=Q*INTGRL(0.0,CC) AIRPH1=(Q/1000)*INTGRL(0.0,CAIR) TIMER FINTIM=480, PRDEL=60, DELT=1.0 METHOD SIMP DO 31 1=1,26 211 31 35 40 IF(C(I) ­LT. 0.0) GO TO 35 GO TO 40 C(I)=0.0 CONTINUE C1=C(1)*1000 C2=C(2)*1000 C3=C(3)*1000 C4=C(4)*1000 C5=C(5)*1000 CS=C(6)*1000 C7=C(7)*1000 C8=C(8)*1000 C9=C(9)*1000 C10=C(10)*1000 C11=C(11)*1000 C12=C(12)*1000 C13=C(13)*1000 C14=C(14)*1000 C15=C(15)*1000 C1S=C(16)*1000 C17=C(17)*1000 C13=C(18)*1000 C19=C(19)­1000 C20=C(20)*1000 C21=C(21)*1000 C22=C(22)*1000 C23=C(23)*1000 C24=C(24)*1000 C25=C(25)*1000 C2S=C(26)*1000 FLUX1=FLUX(1) PRINT CI,C2,C3,C4,C5,C6,07,C8,C9 , CIO,Cll, C12,C13,C14,C15,C16, . . . C17, CIS,C19,C20,C21,C22,C23,C24,C25,C2S, CAVE,FLUXl,CC, ... MTRANS,AIRPH,AIRPHl END STOP 212 Computer Code for Ethylbenzene, Single Air Channel Setup, Adsorption TITLE TITLE TITLE TITLE TITLE TITLE TITLE TITLE THIS PROGRAM COMPUTES THE CONCENTRATION PROFILE OF VOCs IN THE SOIL MATRIX FOR A SINGLE AIR CHANNEL SETUP ASSUMING XD MOVEMENT OF THE CHEMICALS FROM THE SOIL MATRIX TO THE AIR CHANNEL, WHERE D REPRESENTS THE MOLECULAR DIFFUSION COEFFICIENT. ALPHA THE TORTUOSITY FACTOR, K REPRESENTS THE MASS TRANSFER COEFFICIENT BETWEEN LIQUID AND AIR PHASES, PORO REPRESENTS TITLE THE POROSITY AND KH REPRESENTS THE DIMENSIONLESS TITLE HENRY'S LAW CONSTANT FOR THE VOC TITLE THE SOIL MATRIX HAS BEEN DIVIDED IN 26 LAYERS TITLE OF VARIABLE THICKNESS DZ PARAMETER D=0.0004578,K=0.00199, AREA=a7.5, PORO=.377, KH=.294 PARAMETER ALPHA=.52, Q=55.0, V=13 .a25, R=4.8405 STORAGE FLUX(27), DZ(26), C(26), Z(26) / DIMENSION FNET(26), MCON(26), IMCON(26) / EQUIVALENCE (FNETl,FNET(1)), (MCONl,MCON(1)), (IMCONl, IMCON(l)) TABLE DZ(1­18)=18*.25, DZ(19­21)=3*.50,DZ{22­26)=5* 1. INITIAL NOSORT FUNCTION CURVE1=0,.00001,1, .2465,5, .0965,10, .00785,15, .057, ... 30, .04455,45, .04165,60,.03725,90,.0293,120,.02085,... 18 0,.0202,24 0,.0169,300,.01545,360,.0145,4 20,.01765,... 480,.01375,540,.01115,600,.01245 FUNCTION CURVE2=0,.012909,.5,0.0161359,1.0,.0200568,... 1.905, .0218664,3.81, .0217156,6.35, .0217156,8.89, .0226 958,11.0, .022696 Z(1)=.5*DZ(1) DO 2 1=1,26 Z(I + 1)=Z(I)+.5*(DZ(I + 1)+DZ(I) ) 2 C(I)=AFGEN(CURVE2,Z{I) ) C(26)=0.0226958 DO 5 1=1,26 5 IMCON(I)=PORO*C(I)*DZ{I)*AREA DYNAMIC NOSORT CAVE=(C(l)+C(2)+C(3)+C(4)+C(5)+C(6)+C(7)+C(8)+C(9)+C(10)+C(11).. . ) /II 10 20 30 CAIR=AFGEN(CURVEl,TIME) FLUX(l)=­K*(((CAVE)*KH))*AREA CC=({.2465/1000)+FLUX(1)/Q)*(EXP(­Q*(TIME­1)/V))­FLUX(l)/Q FIXED I NOSORT DO 10 1=2,26 FLUX(I)=(C(I­1)­C(I))*(D/R)*ALPHA*AREA/(Z(I)­Z(I­l)) FLUX(27)=0 FIXED I NOSORT DO 20 1=1,26 FNET(I)=FLUX(I)­FLUX(I+1) MCONl=INTGRL(IMCONl,FNETl,26) FIXED I NOSORT DO 30 1=1,26 C(I) =MCON(I)/(PORO*DZ(I)*AREA) FLIX1=FLUX(1) MTRANS=­INTGRL(0.0,FLUXl) AIRPK=Q*INTGRL(0.0,CC) AIRPH1=(Q/1000)*INTGRL(0.0,CAIR) TIMER FINTIM=600, PRDEL=120, DELT=1.0 METHOD SIM? 213 31 35 40 DO 31 1=1,26 IF(C{I) ­LT. 0.0) GO TO 35 GO TO 40 C(I)=0.0 CONTINUE C1=C(1)*1000 C2=C(2)*1000 C3=C(3)*1000 C4=C(4)*1000 C5=C(5)*1000 C6=C(6)*1000 C7=C{7)*1000 C8=C(8)*1000 C9=C{9)*1000 C10=C{10)*1000 C11=C(11)*1000 C12=C(12)*1000 C13=C(13)*1000 C14=C(14)*1000 C15=C(15)*1000 C1S=C(16)*1000 C17=C(17)*1000 C18=C(18)*1000 C19=C(19)*1000 C20=C(20)*1000 C21=C(21)*1000 C22=C{22)*1000 C23=C(23)*1000 C24=C(24)*1000 C25=C(25)*1000 C26=C(26)*1000 FLUX1=FLUX(1) PRINT CI,C2,C3,C4,C5,C6,C7,C8,C9,CIO,Cll,C12,C13,C14,C15,C16, C17,CIS,C19,C20,C21,C22,C23,C24,C25,C26,CAVE,FLUXl.CC,... MTRANS,AIRPH,AIRPHl END STOP 214 Computer Code for Styrene, Soil Column TITLE THIS PROGRAM COMPUTES THE FINAL CONCENTRATION TITLE PROFILE OF VOCs IN THE SOIL MATRIX FOR A COLUMN MODEL AS TITLE AN ARRAY OF AIR CHANNELS OF LENGTH L ASSUMING RADIAL TITLE ID MOVEMENT OF THE CHEMICALS FROM THE SOIL MATRIX TITLE TO THE AIR CHANNEL, WHERE D REPRESEOTS THE MOLECULAR TITLE DIFFUSION COEFFICIENT(cm^2/min), ALPHA THE TORTUOSITY FACTOR, TITLE AREA IS THE AREA OF THE AIR CHANNEL AIR­WATER INTEFACE {cm^2) TITLE K REPRESENTS THE MASS TRANSFER (cm/min) TITLE COEFFICIENT BETWEEN LIQUID AND AIR PHASES, PORO REPRESENTS TITLE THE POROSITY, KH REPRESENTS THE DIMENSIONLESS TITLE HENRY'S LAW CONSTANT FOR THE VOC, Q (cm"3/min) IS THE AIR FLOW TITLE RATE. X IS THE NUMBER OF CHANNELS, AND S IS THE EMPTY REACTOR TITLE VOLUMEN (cra^3) TITLE THE SOIL MATRIX SURROUNDING THE AIR CHANNEL TITLE HAS BEEN DIVIDED IN 7 CONCENTRIC LAYERS TITLE OF THICKNESS DR TITLE MASS TRANSFER COEFFICIENTS WERE DETERMINED USING CORRELATIONS TITLE DEVELOPED IN CHAPTER IV AND RCH AND AREA ARE EXPERIMENTAL PARAMETER D=0.0004734,K=0.01173, AREA=33.9, PORO=.379, KH=.0967 PARAMETER AL?HA=.52, Q=1100, L=49, RCH=0.15, X=175, S=1150 STORAGE FLUX(8), DR(7), C(7), R(7), V(7), IA(7) / DIMENSION FNET(7), MC0N(7), IMC0N(7) / EQUIVALENCE (FNETl,FNET(1)), (MCONl,MCON(1)), (IMCONl, IMCON(1)) TABLE DR(l­7)=7*0.05 INITIAL NOSORT FUNCTION CURVE1=0.1,0.0254 98, .2,0.0254 98, .3, .025498, .4, .0254 98 FUNCTION CURVE2=0, .0001,1, .249,3, .69,10, .434,17, .3075,30, ... .214 5,40, .169,50, .0945,60, .1075,75, .1155,90, .0815,120, .1135, ... 165, .071,180, .0615,225, .046,306, .0365,330, .035, ... 360, .032,420,.0325,480,.0295 R(l)=RCH + 0.5*DR(1) DO 2 1=1,7 R (I + 1)=R(I) + .5*(DR(I+1)+DR(I) ) 2 C(I)=AFGEN(CURVE1,R(I))­(.119/1000) C(7)=0.0254 98­ (.119/1000) DO 5 1=1,7 V(I)=3.1416*L*(((0.5*DR(I)+R(I))**2)­((R(I)­0.5*DR(I))**2)) IA(I+1)=3.1416*(L­13)*2*(R(I)+0.5*DR(I)) 5 IMCONd) =PORO*C(I) *V{I) VOL= (V(l)+V{2)+V(3)+V(4)+V(5)+V{6)+V(7)) TVOL=X*VOL*PORO VAIRE= 3.1416*(L­9)*(RCH**2) IMCONA=1150*.249/1000 DYNAMIC NOSORT CAVE=(C(1)*V(1)+C(2)*V(2)+C{3)*V(3)+C(4)*V(4)+C{5)*V(5) +... C(6)*V(6)+C(7)*V(7))/VOL TMASS=X * CAVE*VOL * PORO CAIR=AFGEN{CURVE2,TIME) FLUX(l)=­K*(((CAVE)*KH))*AREA FIXED I NOSORT DO 10 1=2,7 10 FLUX(I)=(C(I­1}­C(I))*D*ALPHA*IA(I)/(DR(I)) FLUX(8)=0 FNETA=(­FLUX(1)­(Q/X)*CC)•X MCONA=INTGRL(IMCONA,FNETA) CC=MCONA/( S + X*VAIRE) FIXED I 215 NOSORT DO 20 1=1,7 FNET(I)=FLUX(I)­FLUX(I+1) MC0N1=INTGRL{IMCONl,FNETl,7) FIXED I NOSORT DO 30 1=1,7 C(I)=MCON(I)/(PORO*V(I)) FLUX1=FLUX(1) MTRANS=­X*INTGRL(0.0,FLUXl)+IMCONA AIRPH=Q*INTGRL(0.0,CC)+IMCONA AIRPH1=(Q/1000)*INTGRL(0.0,CAIR) TIMER FINTIM=48 0, PRDEL=3 0, DELT=1.0 METHOD SIMP FIXED I NOSORT DO 32 1=1,7 IF(C(I) .LT. 0.0) GO TO 3 5 C(I) =0 CONTINUE C1=C(1)*1000 C2=C(2)*1000 C3=C(3)*1000 C4=C(4)*1000 C5=C(5)*1000 C6=C (6)*1000 C7=C(7)*1000 FLUXl=FLUX(1) FLUX2=FLUX(2) FLUX3=FLUX(3) FLUX4=FLUX(4) FLUX5=FLUX(5) FLUX6=FLUX(6) FLUX7=FLUX(7) PRINT CI,C2, C3,C4,C5,06,C7,CAVE,FLUXl,FLUX2,FLUX3,FLUX4,FLUX5, .. . FLUX6,FLUX7 , CC,TVOL,MTRANS,AIRPH,AIRPHl,TMASS END STOP 216 BIOGRAPHICAL SKETCH Washington J. Braida was bora on December 18, 1956 in Montevideo, Uruguay. He received his Bachelor of Science degree in Chemistry and a Master of Science degree in Chemical Engineering from Universidad de la Republica (Uruguay) in 1978 and 1982, respectively. After 10 years of working in the pharmaceutical industries, he was admitted into the graduate program of environmental engineering at Polytechnic University, Brooklyn. New York. He received his Master of Science degree in Environmental Engineering in June 1994. Since August 1994, he has been working towards his Ph.D. degree at Iowa State University. During his studies at Iowa State, he was awarded the Society of Hispanics Professional Engineers (SHPE) Region VI Graduate Scholarship in 1995 and the Merwin Dougal Award/Scholarship (Best Graduate Student in the Environmental Engineering Program) in 1997. He has served as a Research and Teaching Assistant in the Department of Civil and Construction Engineering at Iowa State University. He is married with Gaye Goker and has two daughters, Maria Eugenia and Constanza.