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.