characterizing grain-scale availability of aromatic hydrocarbons in mgp site soils and assessment of the impact on the dynamics of soil and pore water concentration during and after remediation

The following conclusions could be made concerning the site post-remediation: 1 despite significant reduction of the contamination levels, the chemicals remained available at the grain-

CHARACTERIZING GRAIN-SCALE AVAILABILITY OF AROMATIC HYDROCARBONS IN MGP SITE SOILS AND ASSESSMENT OF THE IMPACT ON THE DYNAMICS OF SOIL AND PORE WATER CONCENTRATION DURING

AND AFTER REMEDIATION

A Dissertation Presented to the Faculty of the Interdisciplinary Graduate Program

in Environmental Engineering, University of Houston

UMI Number: 3225012

INFORMATION TO USERS

The quality of this reproduction is dependent upon the quality of the copy submitted Broken or indistinct print, colored or poor quality illustrations and photographs, print bleed-through, substandard margins, and improper alignment can adversely affect reproduction

In the unlikely event that the author did not send 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

® UMI UMI Microform 3225012 Copyright 2006 by ProQuest Information and Learning Company

All rights reserved This microform edition is protected against unauthorized copying under Title 17, United States Code

ProQuest Information and Learning Company

300 North Zeeb Road P.O Box 1346 Ann Arbor, MI 48106-1346

CHARACTERIZING GRAIN-SCALE AVAILABILITY OF

AROMATIC HYDROCARBONS IN MGP SITE SOILS AND

ASSESSMENT OF THE IMPACT ON THE DYNAMICS OF

SOIL AND PORE WATER CONCENTRATION DURING

AND AFTER REMEDIATION

Approved:

Committee Members:

KC (leit

Larry C Witte, Associate Dean,

Cullen College of Engineering

RoberfBryan Luper '

4.24 & ` /Z⁄/„

Chair of the Committee GZ

William G Rixey, Associate’Professor, Civil and Environmental Engineering

LhA 0z

Dennis A Clifford, Prof,

Civil and Environmenta 2 eering dla Hanadi S Rifai, Associate Professor 7 Civil and Environmental Engineering

ACKNOWLEDGEMENTS

As much as these years have been a joy to me, I cannot express the degree of

relief and amount of pride that I feel to finally be completing the PhD program at the

University of Houston As a civil engineer coming back to school after more than seven years out of academics and no experience in environmental engineering, my learning curve has been steep For this reason, I would first like to thank Dr Rixey for his

continued support and patience through these years while I struggled to master the

necessary concepts to perform this work I would also like to thank Dr Rixey for his invaluable guidance It has been a great comfort to be able to depend on the accuracy and soundness of his advice in all aspects of this research I also thank the other committee members Dr Clifford, Dr Capuano, Dr Mohanty, and Dr Rifai whose input and

sacrifice of precious time have also been extremely helpful and have added to the value

of the research and my education

I would especially like to thank both the former and current members of my

research group I have always been able to count on, whenever needed, their advice, help

and assistance I also give thanks my fellow graduate students, both within and outside

of my research group for their daily kindness, cordiality and consideration They have all made coming to school each day a pleasure

I cannot give sufficient thanks to my wife, Belen, for all of the sacrifices she has made to make it possible for me to pursue my goals She has been long suffering from

my absences both during the week and on so many weekends If possible, the completion

of my doctorate is an even greater relief to her than to me For our son Nathan, words cannot express the joy that he has brought and continues to bring into our lives Every

day of delay in graduating to provide for his care has been paid back a hundredfold in the fulfillment of having him in our lives For my parents and siblings, they are and continue

to be my foundation, I cannot express the degree to which they give me the ability to

continue on in tough times

Finally, but very importantly, I would like to acknowledge and give thanks for the

funding and logistical support received from RETEC, URS, the Gas Technology Institute

and the Gulf Coast Hazardous Substance Research Center

CHARACTERLZING GRAIN-SCALE AVAILABILLTY OF AROMATIC HYDROCARBONS IN MGP SITE SOILS AND ASSESSMENT OF THE IMPACT ON THE DYNAMICS OF SOIL AND PORE WATER CONCENTRATION DURING

AND AFTER REMEDIATION

An Abstract of a Dissertation Presented to the Faculty of the Interdisciplinary Graduate Program

in Environmental Engineering, University of Houston

ABSTRACT

To predict groundwater concentration levels which emanate from a contaminated

source, early release models assumed local equilibrium partitioning As the validity of

this assumption has been questioned, models have been developed which incorporate

kinetic release To further this work, two complementary studies were performed The

first involved applying developed laboratory and analytical techniques for measuring availability during the remediation of an existing contaminated site The second involved the application of a two-site equilibrium/rate-limited release model to illustrate the

impact of specific degrees of limited availability

For the first study, as an initial measure of contaminant availability, short-term

batch tests were performed before, during and after the remediation effort Test samples

which indicated limited availability were selected for long-term testing to quantify the

availability parameters The following conclusions could be made concerning the site

post-remediation: (1) despite significant reduction of the contamination levels, the

chemicals remained available (at the grain-scale) for additional treatment; (2) the

groundwater concentrations in monitoring wells near the source could be predicted to rebound significantly; (3) despite the absence of remediation to an endpoint, methods

were developed to determine environmentally acceptable soil concentration endpoints in

the laboratory; (4) confirmation of short-term results with long-term experiments is necessary; and (5) the apparent Ky can used to justify soil concentrations endpoints significantly higher than those predicted using typical soil screening level fo.*K., values

For the second study, simulated and actual availability data were used with

simulated site-conditions (i.e velocity, porosity, etc.) to generate groundwater

concentration vs time curves The curves illustrate a two-step process consisting of

aggressive remediation, followed by a post-remediation step where natural hydrologic

conditions are re-established Release rate constants of about 10° day’ or faster allow

remediation at a practicable, albeit slower, rate Upon the return to natural conditions, the

concentration rebounds to a level determined by the apparent Kg Slower rate constants

significantly reduce the effectiveness of remediation efforts, but result in post-

remediation concentrations lower than predicted by the apparent Ky

TABLE OF CONTENTS

SOF OOERH EHEC TENTHS ERE DERESHEHEMEEHODERDEE EAH EEHEROCTOHEHER EERE ERE ESEREOSEED

PRUHH HEHE HESS OEE EOC ERT CEH TET EC ESE SEHR ESE HO HEE EH OOHER ETE EONS

AMOR C ECCS RS EEEEEET RE EREEHEEEERE THE TEER HEED EEE HH HERS EHOREEDEOTEEHOEE

1.1 Problem Statement PRON ee Ree aE meee ea meena eet a One Đoh 4o BÊ Đo A9 0ọm Đe k6

1.2 Objectives/Significance of Study —“—-_

1.3 Iilustrative Example of the Effect of Slow Release on Pore-water

Ame m ee eee ea tee eee eee eee me eT TEE EE AEH HEHE O eee mE eae eee eee

Seema ba acne eae ea em ere se rere rar as en rae snares eeerreseeseses

POMOC He Reade eE ae ESE REDE SESLESED EERO ERED ERD OREETE

2.2 Nonequilibrium Relationships ¬“.-ˆˆ

2.3 Nonequilibrium Desorption Models ` ` -

2.3.2 Two-site Kinetic Model

2.3.3 Continuum Model DOO e dame eee meta meee eee eer eae e eee eee beeen nee

2.3.4 Two-region Kinetic Model Come meee emer were mene near anes tenant seaneeeseres

2.4 Subsurface Fate & Transport Modeling Demme eee ame renee eee e sean eer eanees

2.4.1.1 Completely Stirred Tank Reactor Models

2.4.1.2 Continuous Flow Stirred Tank Reactor Models ee ee er ey

2.4.2 Nonideal Systems

2.4.2.1 CSTRs-in-series

2.4.2.2 Advection-Dispersion-Reaction Equation (Plug flow with dispersion)

2.5 Laboratory Methods for Measuring Rates of Sorption/Desorption

2.5.1

2.5.2

2.5.3

2.5.4

2.5.5

2.7 Summary

3 MATERIALS AND METHODS

3.2 Sample Processing

3.2.1

3.3.1

3.3.2

3.4 Laboratory Methods

3.4.1

Gas Purge

Ce ee ry

ie ey

Supercritical Fluid Extraction

Fixed-bed Soil Columns

`

CA Soils Somers ence nce n eae eneaeaee Sencar ena reas Analytical Data Significance of Soil Analysis

¬ ÔỒÔÔÔÐ Batch Partitioning Experiments Heese eee mance neas Ammer e meter rete ease eaesenes “`

ec eee ee ee ee ee eee ee ee ee ee ee re sy “` ÖÖŠÖ ` eres waneee CS i rr ee eee ee ee ee re ere ry i ee i ee er re DOOM ase ame ere ence eee nm mentee eran are senees ee ee ee ee ey CaO e areca ener nam ere ese aasesearsraaes SEC eer meee e eter eta emer ates anes SO eR ee eee eeesereerersatesesaserasetenreneas ÔÖÖ

¬ ., , , , ,

ÔÔÔÔÔÔÔÔÔÔÔỒÔỞÓÔ Ô ,.,

—

Am

—._._ `

ee ee ee re ee ee

TOPO R a eee eae e nema reese aren men enaens

19

19

21

21

21

22

22

23

24

25

27

27

27

27

28

28

28

29

29

31

32

33

¬ 34

3.4.1.2 48-hr A queous ConcenfratiOnS chen ec 34 3.4.1.3 Long-term Aqueous Concentrations ccc non coi 35 3.4.1.4 Total Hydrocarbon Measurement .0 cccccecccsececceecececees 35

3.4.2 Depletion of Contaminated Soils Using Laboratory Columns 36

3.4.3.1 Preparation of Standard Solutions ccccccccccecccceseccccecee 38

3.4.3.2 Purge & Trap Gas Chromatography to non 38

3.4.3.3 Calibration CurV€S cu n HH nh nho 39

4 MATHEMATICAL FRAMEWORK ào ccccccs se 40

4.1 Qualitative Data Analysis of Batch Partitioning Experiments 40 4.2 Quantitative Data Analysis of Time-Series Batch Partitioning

Experiments SH ni tk Ki HT hit vn ng kề cà LH TH TH S1 kg xe: 44 4.3 Quantitative Data Analysis of Fixed-Bed Rate of Release Experiments 48 4.4 CSTR Two-site Equilibrium/Rate-Limited Source Zone Model 49

45 SUMMALY 2 ccc cecceccceeccaseseceuustueessucecennesees eee eaeeeteueeuse 55

5 AVAILABILITY EXPERIMENTS AT A FORMER MANUFACTURED

GAS PLANT SITE DURING A REMEDIATION PROJECT 58

he nh 58

5.2 ODJECHVES na 60

3.4 Field Sampling and Processing nhe 62 3.4.1 Field Sampling n2 nho 62

5.4.2 Soil Characterization " ĐD 65

3.4.3 Sample PreparafiOn cu TH Hs Tnhh hiến 65 3.4.4 Spatial Sampling Variabilify con nh nn hà 66 5.5 Results and Discussion :.cccsseeeseeceeceusecsesceuecseneensceeueesen 68 5.5.1 Short-term Partitioning Data cv ryêy 68 5.3.2 Boring BI Trends c.cQnnnnnnnn nh nha nhung 69 5.5.3 Boring B2 Trends ccc ccccccecceccvccuceuscceccucceseesneanseuceens 80 5.6 Long-term Batch Partitioning Experiments 00 ccccececeeceueeceues 82 5.6.1 Case A: Soil Remediated in the Field to an Endpoint: Boring B2-1 85

5.6.2 Case B: Soil with Partial Remediation — Boring B1-1 (Q7 & Q8) 90

5.6.3 Case C: Soil Depleted in the Laboratory — Boring B1-2 & B1-3 93

5.6.3.1 Q8-B1-2 Column Depletion Experimenfs 93

3.6.3.2 Q8-B1-2D Partitioning Experiments on Depleted Soils 95

2.6.3.3 Q8-B1-3 Column Depletion Experiments 97

5.6.3.4 Q8-B1-3D Partiioning Experiments on Depleted Soils 102

3.7 Monitoring Well/Leachate Test Relationship - 106

5.8 Determining Environmental Endpoints using ẾT,/long-term‹‹‹ các cà cà 109 5.9 Conclusions 6 ANALYSIS OF THE IMPACT OF NONEQUILIBRIUM PARAMETERS ON SOURCE ZONE REMEDIATION AND POST-REMEDIATION GROUND WATER CONCENTRATION REBOUND 113 6.1 Effect of the slow release of naphthalene on concentrations during and after

6.4 Effect of slow release on pore water concentrations for naphthalene and benzene during and after remediation of CA soils

6.5 Effect of slow release on pore water concentrations for naphthalene and benzene

during and after remediation of Site A soils 0 0 cccccceseceeeee neve 130

7 CONCLUSION AND FUTURE WORK ccccccecescsecesseeeeeens 139

7.1 Application of availability parameters to the decision-making process of a remediation project and the further development of laboratory release

techniQU€S c ng HS eeeeeuesaeeuseeeeeeaueeaecsaneuaeees 139 7.2 Source zone release analysis ccccce sec cesccesecuseesucseeeuseuecaeans 142 17.3 Recommended Future Research - ch 143

APPENDICES .cccceccuseeenscecseeeeecesessceusscesaseecesctscescceuseavecseness 153

Illustrative Example of the Effect of Slow Release on Pore-water

Concentrations POCA eA STOOD EEE HEEER HEHE EEO ET ER AEE EERE E SOHC ORE REDS

Experimental/Analytical Protocol COD ee eeserereraesesaseoueeeneneoen

Diagram of fixed-bed column depletion experiment eeaassoet°se

Experimental/Analytical Protocol ———_

Site Map Ẻ969920940464A0600040946000049460496606000094600204069600902060090660606606066666s421

Sampling Groups Soil & Aqueous Leachate Concentrations of Benzene

Soil & Aqueous Leachate Concentrations of Naphthalene

TPH Concentrations vs Soil Groups BORO HTC eRe eee ereserenesence

Naphthalene COOP H HOHE HTT THEE SEHHE RE EEET ESE RE DS EHE TEE ON ETE REET OTOL OS

Q8-B1-2D Depletion Column curve COO ema eaensresesnsceeresessneeenes

Case C: Q8-B1-2D: Long-term Leaching Results: Benzene

Q8-B1-3b 2"° Depletion Column curve ——— _

Figure 5.14a - Case C: Q8-B1-3b: Long-term Leaching Results: Benzene 103

Figure 5.14b -Case C: Q8-B1-3b: Long-term Leaching Results: Naphthalene 103 Figure 6.1a -

AD, post-remediation = 0 d' (simulated avaÏlability data) 117

Predicted impact of slow release on soil and groundwater concentrations for naphthalene in a sandy soil Cy,= 100 mg/kg,

Kg = 10 L/kg, F = 0.90, k; = 1.0E-05 a dp,remediation = 1 d™',

AD, post-remediation = 0 d" (simulated availability data) 117

Predicted impact of slow release on soil and groundwater concentrations for naphthalene in a sandy soil Cy,= 100 mg/kg,

Predicted impact of slow release on soil and groundwater

‘concentrations for naphthalene in CA-17 soil Cy,= 1290 mg/kg,

Ka = 330 L/kg, F = 0.88, k, = 4.0E-05 d™ (simulated k, value),

Ap,remediation = 20 d, Ap,post-remediation =O Ï, cccccccsee 122

Predicted impact of slow release on soil and groundwater concentrations for naphthalene in CA-2 soil C7,= 60 mg/kg,

Ky = 5933 L/kg, F = 0.18, kz = 1.7E-03 d™, Ap remediation = 20 d™,

Ap,posteremediation =O ' ccceeeeeees 124

Predicted impact of slow release on soil and groundwater

concentrations for naphthalene in CA-5 soil Cy,= 120 mg/kg,

Ka = 1035 L/kg, F = 0.32, k, = 1.0E-03 d™, Ap remediation = 20 a",

— -1

concentrations for naphthalene in CA-18 soil C7,= 4100 mg/kg,

concentrations for benzene in unremediated Q8-B1-3b soil

Cro= 114 mg/kg, Ka = 17 L/kg, F = 0.97, ky = 2.7E-02 d", Ap,remediation = 1d", Ap,posteremediation =O U7 .ccscssscessssssenseees 132

Predicted impact of slow release on soil and groundwater concentrations for benzene in laboratory remediated Q8-B1-3b soil

Ap,remediation = 1 ', Ap,post-remediation = OO ' ssssesssesssceseseeeees 133

Predicted impact of slow release on soil and groundwater concentrations for naphthalene in unremediated Q8-B1-3b soil

Cro= 2,077 mg/kg, Ka = 157 L/kg, F = 0.50, k = 1.5E-02 d",

= -l = -1

Figure 6.7a -

Figure 6.7b -

Figure 6.8a -

Figure 6.8b -

Predicted impact of slow release on soil and groundwater

concentrations for naphthalene in partially remediated Q8-B1-3b soil

Cy = 238 mg/kg, Ka = 45 L/kg, F = 0.97, k;ạ = 1.0E-02 d1,

Predicted impact of slow release on soil and groundwater

concentrations for naphthalene in laboratory remediated Q8-B1-3b soil Cro= 10 mg/kg, Ka = 104 L/kg, F = 0.23, k) = 6.0E-03 d”,

Ap remediation = 1d”, Ap,post-remediation = Ú Ì, c.cccsscce<< 137 Predicted impact of slow release on soil and groundwater

concentrations for naphthalene in remediated Q7-B1-1 soil

Table 3.1a

Table 3.1b

Table 3.2

Table 4.1

Table 5.1

Table 5.2

Table 5.3

Table 5.4

Table 5.5

Table 5.6

Table 5.7

Table 5.8

Table 5.9

Table 5.10

Table 5.11

Table 5.12

Table 5.13

LIST OF TABLES

Soil Characterization Data for CA soils 29

Soil Characterization Data for Site A soils 30

Q8-B1-3b: Clay Minerology Data 30

Summary of Mathematical Eramework 56

Groundwater Concentrafion Goals 58

B2-2 Short-term batch testing results at discrete vertical Dep(HS on HH HH ng KH ko nu v64 67 B1-1 Short-term batch festfing results 70

B1-2 Short-term batch festing resul(s 71

B1-3 Shorf-term batch testing resul(s 72

B2-1 Short-term batch testing resulfs 73

B2-2 Short-term bafch testing resulfs 74

B2-3 Short-term batch festing resulfs 75

Summary of Availability Values Determined from Long-term Partitioning ExperimenfS c co s<<e 84 Q8-B2-1 Short-term measurements — Pre- and Post-storage 86

Q7-B1-1/Q8-B1-1 Short-term measurements — Pre- and POSt-StOLAGE ccecscscecsccsceececececacseveececcesesecsceeeesseas 87 Q8-B1-2 Short-term measurements — Pre-storage, Post-storage & Post-depletion ccccccecsccscecesetceescsseessessscecessaccacs 88 Q8-B1-3/Q8-B1-3b Short-term measurements — Pre- and l0 0c he 89

Groundwater Sampling Data from Wells near Boring

Locations CORKS HASHES E EEOC CER ERE ETH EOHOER TEES EO HEELS HEH EE ESET S OSE DEDS

Parameters for two-site equilibrium/slow release source

zone model —¬—“ ˆ-.Ô Ắ

eo+adese“e

Appendix A — Predicted K, Values

APPENDICES

SOOO H AH RER EEOC OTe SESETeS OR EDEDOD ORO EERE EE EERE TOS

Appendix B - Coupled Source Zone/Fate & Transport Model Appendix C — Standard Deviafions - co uc 0101111 ve s2

CHAPTER 1 INTRODUCTION

1.1 Problem Statement

Contamination in the form of nonaqueous phase liquids (NAPLs) such as gasoline, crude oil, chlorinated solvents, etc., typically enters the subsurface as the result of spills

or leakage and is both a public health and ecological concern Contamination caused by

these releases and the processes that affect their fate and transport in the subsurface

requires sound understanding in order to make appropriate site-management and

remediation decisions A released volume of NAPL will flow downward by gravity

forces into the subsurface As the NAPL flows downward through the pore structure of the unsaturated zone of the subsurface (the vadose zone), an increasing portion of its total

volume will become trapped in soil pores by capillary forces (Conrad et al., 1992) If the volume of the NAPL spill is larger than what can be entrapped within the vadose zone, then a portion of the NAPL released will reach the water table as a free phase liquid Light NAPLs, such as gasoline, upon reaching the water table, will float on top of the water table and spread laterally as free product As the height of the water table

fluctuates, an additional portion of the volume will become trapped in soil pores below the average mean sea level of the water table (the saturated zone)

Dense NAPLs, such as chlorinated solvents, in lieu of floating on the water table, will displace the water and continue to migrate downwards through the water table, leaving entrapped NAPL within the soil pores of the saturated zone as it does This continues until all the NAPL volume becomes entrapped within pores or a remaining free-phase portion reaches an impermeable barrier such as a dense clay or bedrock and then pools

there as free product These sequestered dense and/or light NAPLs are relatively

immobile under natural hydraulic conditions and act as long-term sources of

contamination

Due to legislation, such as the Resource Conservation and Recovery Act (RCRA) and the Comprehensive Environmental Response, Compensation and Liability Act

(CERCLA), requiring the clean-up of contaminated sites, the remediation of

contamination sources has in many instances been initiated However, attempts to restore

contaminated subsurface sites to their original pristine condition have in most instances

proven to be unattainable by existing technologies (Travis and Doty, 1990) In response

to the inability of existing remediation technologies to completely remove contaminants from the subsurface, more risk-based corrective action (RCBA) approaches based on environmentally acceptable endpoints (EAE) are being implemented An

environmentally acceptable endpoint (EAE) is a site-specific clean-up level that is

protective of human-health and the environment (Loehr, 1999) The RCBA approach establishes a three tiered risk assessment process that can be used for establishing site- specific clean-up levels based on potential risk to human health and the environment (ASTM, 1994) The EAE and RCBA approaches recognize that the risks posed by chemicals in the subsurface soils are dependent not only on their concentration but also

on their mobility and toxicity Therefore, the presence of high concentrations of

chemicals in subsurface soils may not necessarily be associated with adverse effects to the public and environment (Alexander, 1995) As the required level of risk assessment moves from screening to a detailed investigation, the cost, time and data required for conducting a proper analysis increases However, the increased level of investigation

may also result in increased protection of the public and environment and/or in

considerable savings in remediation costs

A common exposure pathway to the public and environment from a contaminated site

is through groundwater transport Figure 1-1 provides a theoretical representation of

such an exposure from a residual source zone located in the saturated zone

Figure 1.1 — Illustration of groundwater pathway exposure

In predicting the concentration of contaminants reaching an exposure point, initial

groundwater models have assumed that, at any particular local point within the

contaminated area, equilibrium partitioning is occurring between the residual NAPL, soil,

air and groundwater phases and that these relative concentrations can be expressed with the use of partition coefficients (Karickhoff et al., 1979) This is commonly called the -local equilibrium assumption (LEA) and is valid when the rate of release of a chemical from.a residually trapped NAPL or a sorbed contaminant in the soil is more rapid than the

rate of aqueous chemical concentration changes by other means (e.g advection,

volatilization, biodegradation, etc.) However, researchers, in the mid to late 1980’s, began to realize that kinetic limitations were often playing a significant role in sorption and desorption and that equilibrium modeling is often not sufficient to accurately

characterize the release of chemicals from soil (Pignatello, 1989; Xing and Pignatello,

1997) The erroneous assumption of local equilibrium can lead to an underestimation

and/or overestimation of the true extent of sorption and desorption over time, false

predictions about the mobility and toxicity of contaminants, and can ultimately lead to incorrect field management and site-cleanup decisions concerning contaminated sites

Therefore, in order to accurately model the true risk posed by contaminants, the

phenomena of kinetically limited release must be considered and its effects quantified

Due to the spatially and temporally complex naturés of contamination and

contaminated soil matrices, the actual phenomena or combination of phenomena causing

kinetically limited release from the soil matrix is not yet well understood and is still a

subject of much research (Cornellissen et al., 2006; Cornellissen et al., 2005; Cornellissen

et al., 1997; Ghosh et al., 2003; Ghosh et al., 2000; Hawthorne et al., 2002; Luthy et al., 1997; Xing and Pignatello, 1997) For this reason, kinetic release parameters have been

incorporated into release models to empirically describe the nonequilibrium chemical desorption from soil matrices to the aqueous phase Various laboratory methods have also been developed to determine these release parameters These models and methods

are discussed in greater detail in the Literature Review

1.2 Objectives/Significance of Study

The purpose of this research has been to develop and use laboratory methods and

mathematical models to quantify the nonequilibrium partitioning characteristics of

benzene, toluene, ethylbenzene, xylene and naphthalene (BTEXN) from weathered

former manufactured gas plant (MGP) soils and then incorporate the rate-limited release

parameters into a two-site equilibrium/rate-limited release model to predict the

groundwater concentrations which would emanate from a source during remediation and

the potential rebound that can be expected after the termination of remediation activities For this study, the kinetic-release parameters are expressed by the following: (1) Kg - the apparent soil-water partitioning coefficient for the contaminant at equilibrium between the soil and water phases, (2) F - the fraction of the contaminant available for equilibrium release, and (3) k, a slow release rate constant for the remaining fraction (1-F) Note that the methods presented in this study address only grain-scale contaminant unavailability Additional field-scale release limitations, due to macro-scale effects (e.g flow

bypassing), are not addressed in this study

That the nonequilibrium release of contaminants from soils occurs and that

nonequilibrium parameters can be quantified using experimental methods has been

established (Kim, 2002; Opdyke, 2000; Opdyke, 2002) That the rate-limited release data can be incorporated into a groundwater fate and transport screening model has also been demonstrated (Opdyke, 2002) This previous work has been extended by:

(1) the further development of the laboratory methods,

(2) the application of the developed laboratory and analytical procedures to the

decision making process of an actual MGP site remediation project, and

(3) the incorporation of measured and simulated nonequilibrium data into an

analytical source zone model to evaluate the relative impact of different

kinetic-release parameters in model predictions

1.3 Illustrative Example of the Effect of Slow Release on Pore Water Concentrations

In order to better explain the potential consequences of limited availability, Figure 1.2 provides an illustrative example of the effects of limited availability on a proposed

aggressive remediation process for the pore water of a contaminated source zone (Ct =

10 mg/kg of naphthalene) For this example, the remediation is evaluated for a

contaminated source for which the contaminant availability parameters of F, Ka, and k>

have already been quantified The aggressive remediation step lasts for a period of one- year As shown in the illustration, during the initial portion of the remediation, the

available fraction (F=0.5) of the contamination is rapidly removed from the soil so that, within only a month or two, the soil concentration is reduced to approximately 5 mg/kg

and to a pore-water concentration which approximately corresponds to the partitioning

coefficient of 10 L/kg (i.e C, ~ 0.5 mg/L when Cy; = 5 mg/kg) However, after the

available fraction has been removed, the soil remediation rate is sharply reduced to the

rate at which naphthalene is released from the soil matrix to the pore water (kạ = 107 day

'), For the next few months, the continued aggressive removal of naphthalene from the

100

Remediation | Post Remediation

Figure 1.2 - Illustrative Example of the Effect of Slow Release on Pore-water

Concentrations — Cy, = 10 mg/kg, F = 0.5, Kg = 10 L/kg, k, = 10“ day"

pore-water results in a continued decrease of the pore-water naphthalene concentration (C,) until a pseudo-steady state value of only a few pg/L is obtained However, in this same time period, no significant additional removal of the remaining soil contamination

has been achieved Unless the remediation process can be modified to enhance the kinetic-release rate of the naphthalene (i.e solvent, surfactant and/or thermal methods), the level of aggressiveness of the continuing remediation effort is irrelevant The soil contamination is not available for rapid remediation Furthermore, when remediation efforts are discontinued and natural hydrologic conditions return, the pore-water

concentration can be expected to rebound However, note in the illustration, that

although the pore-water concentration does rebound by an order of magnitude, it reaches

a pseudo steady-state value which is more than an order of magnitude lower than the

expected equilibrium concentration for the remaining soil concentration (Cy = Cy/Ka) and

is slightly less than the illustrated ground water standard In this particular example, the aggressive remediation effort modeled for this illustration could, in six months, providing other macro-scale remediation limitations do not occur, achieve a remediation level

which is protective of the public health

In conclusion, the quantification of grain-scale availability parameters can

theoretically provide information regarding what reduction in soil concentration levels

can be achieved before kinetic-release limits remediation In other words, a site-manager can, before a remediation process is implemented, obtain information (based on

laboratory measurements) concerning whether a particular soil needs to be or can be treated to meet a given groundwater quality or soil concentration goal

1.4 Outline of the Dissertation

The current laboratory methods in use and the mathematical frameworks used are described in detail in Chapter 3 and Chapter 4, respectively In Chapter 5, the

implementations of the methods described in the previous chapters are applied to an

existing field remediation project with the results analyzed and discussed In Chapter 6, a quantitative source zone model is used to quantitatively assess the impact of rate limited release on soil remediation and the subsequent groundwater rebound using release

parameters obtained from laboratory experiments for several site soils that exhibited

different degrees of contaminant availability Conclusions and recommended future work are given in Chapter 7

CHAPTER 2 LITERATURE REVIEW

As the exact mechanisms of the pore-scale dissolution and desorption of chemical

compounds from the NAPL and soil phases to the aqueous phase are not yet well

understood, dissolution/desorption equations, both equilibrium and kinetic based, have been developed in an attempt to quantitatively describe the release of the compounds In order to model the transport and fate of these contaminants in the subsurface, these

desorption equations have typically been coupled to an advection-dispersion-reaction

(ADR) equation The parameters for use in these models have been determined using various laboratory procedures for measuring the rates of sorption and desorption of

organic contaminants in different oil/soil/water and soil/water systems

2.1 Equilibrium Relationships

Although in recent years, the assumption of equilibrium partitioning at the pore-

scale level has come into question, linear and nonlinear isotherm models were initially used to describe equilibrium sorption to the aqueous phase (Karickhoff et al., 1979)

Initial studies indicated that the local equilibrium assumption was valid for dissolution from the NAPL phase to the aqueous phase (Van der Waarden et al., 1971; Fried et al,

1979) When equilibrium desorption is used, it is common to assume that the equilibrium isotherm is linear (Karickhoff et al., 1979; Karickhoff, 1981; Ball and Roberts, 1991)

Recent research has indicated that nonlinear isotherms can occur for many soil/chemical

systems under various conditions (Kan et al., 1994; Huang and Weber, 1997; Xing and Pignatelio, 1997; Chen et al., 1999; Jepsen and Lick, 1999); however, kinetic models that use Langmuir or Freundlich isotherms to represent the equilibrium condition were not

Judged necessary for the range of values presented in this research and will not be

presented

The linear isotherm is represented by the following equation

where:

q = mass of contaminant sorbed per mass of solid phase (mg/kg)

Ka = sorption partition coefficient (L/kg), and

C,, = concentration of the contaminant in water (mg/L)

It has been shown that the sorption partition coefficient of hydrophobic organic compounds onto many soils is related to the fraction of organic carbon, fo,, within the soil

matrix and their partition coefficient with respect to the organic fraction, Kyo It can be

desorption and that equilibrium modeling is often not sufficient to accurately model the

release of chemicals from the soil matrix (Pignatello, 1989) When local equilibrium

conditions cannot be assumed, a mass transfer expression must be used A typical

expression of the rate of mass transfer of organic chemicals from the NAPL phase to the aqueous phase has been in the form ofa linear-driving-force term (Hunt et al., 1988a and 1988b; Miller et al., 1990; Powers et al., 1991; Powers et al., 1992)

where:

J = net flux of a species between phases (g/sec-cm”)

C, = aqueous concentration in equilibrium with NAPL phase concentration (g/ cm’)

Cy = average concentration dissolved in water (g/cm”)

k; = mass-transfer coefficient (cm/sec), and

a = NAPL-water interfacial area per unit volume of medium (cm?/cm’-medium)

When the flux of species between phases is sufficiently fast in relation to the rate of

aqueous chemical concentration changes by other means, equilibrium between the phases will be obtained and partition coefficients can be used in lieu of Equation 2.2.1 This is known as the local equilibrium assumption (LEA) Some studies have shown that the local equilibrium assumption (LEA) for dissolution from the NAPL phase to the aqueous phase is valid for certain ranges of NAPL saturations and aqueous velocities (Hunt et al., 1988a; Miller et al 1990; Van der Waarden et al., 1971; Fried et al, 1979; Mackay et al.,

1991) However, most of these studies have covered only a limited range of aqueous

velocities and saturation ranges Studies have indicated that, as the interfacial area is reduced, mass transfer rate limitations will eventually occur (Hunt et al., 1988b; Imhoff et

al., 1993; Geller and Hunt, 1993; Rixey et al., 1991; Rixey, 1996) Additionally, even if

the dissolution occurs from a largely insoluble matrix where the available interfacial area

does not reduce (e.g gasoline), a study carried out by Borden and Kao, 1992 indicated

that nonequilibrium conditions are still likely to eventually occur In their study, an effluent column study was carried out over several orders of magnitude of concentration Although equilibrium conditions were observed initially, mass transfer limitations began

to occur as the experiment progressed The authors reasoned that the eventual limitations occurred due to the initial depletion of contamination from smaller NAPL blobs, leaving only the larger blobs, which have smaller interfacial areas per unit volume of NAPL It has also been theorized that the eventual mass-transfer limitations observed in the study may also have been the result of the occurrence of slow desorption from the soil phase rather than slow dissolution from the NAPL blobs (Garg, 1998)

If equilibrium conditions cannot be assumed, a mass transfer rate, k,a, must be determined However, based on several different factors (e.g type of porous medium, organic fluid properties, wetting history, etc.), NAPL volumes from spills are distributed into the subsurface in a currently immeasurable variety of sizes and shapes (Conrad et al 1992), and there are currently no well-developed methods for precisely estimating the available interfacial area per unit volume of medium for any particular system Some studies have attempted to use an idealized NAPL geometry in order to estimate a specific interfacial area for different systems (Hunt et al., 1988a; Powers et al., 1991); however, due to the current lack of understanding about the nature of NAPL volume distributions

in the subsurface, several empirically based models have been developed and are

described in the following section

2.3 Nonequilibrium Desorption Models

There are a number of empirical formulations appropriate to describe the

nonequilibrium rate of sorption and desorption of chemicals from the NAPL and/or soil

matrix to the aqueous phase A few of the more common ones will be discussed in this section All of the formulations discussed are mass-transfer models which assume a linear isotherm to represent the equilibrium condition Assumed is that the phases in question are well mixed and any resistance to mass transfer between the boundaries is due to a boundary layer As indicated in Equation 2.2.1, mass transfer equations take the general form (Cussler, 1997)

Rate of mass transferred = Rate constant * Concentration difference

2.3.1 Bicontinuum Kinetic Model

In order to describe desorption, the bicontinuum model assumes that there are two sites for sorption: q; and qo Site q; is in equilibrium with the aqueous phase, while site q2 is kinetically limited, with a desorption rate of ky The initial fraction of the total

amount sorbed that is in equilibrium with the aqueous phase (the q; fraction) is denoted

by F The remainder of the initial sorbed chemical (the q; fraction) is given by (1-F) Therefore

2.3.2 Two-site Kinetic Model

The two-site model is very similar to the bicontinuum model, except that the rate

of sorption to and from site S; is not instantaneous, but is kinetically limited

(Cornellissen et al 1997; Williamson et al 1998; Ghosh et al 1999) The (1-F) fraction

of the qp site is identical to the q; site of the bicontinuum model The F fraction of the q; site is kinetically limited but at a significantly higher mass transfer rate (k,) than q’s ky

rate Therefore

Sometimes the k; value better represents the initial period of rapid release However, due

to the adequacy of the bicontinuum model in most situations, this model will not be used

in this research

2.3.3 Continuum Model

Although adequate to describe rate-limited release, the two-site and bicontinuum

models describe only two sites of what is likely a continuum of N release sites (Fi, Fa, F3,

; Fy), each with a successively lower release rate Often this distribution of release

rates is described using a gamma (T) or log normal probability distribution function

(Deitsch and Smith, 1999; Culver et al., 1997; Deitsch et al., 1998) The I’-probability distribution function (PDF) can be expressed as follows

where:

n = the shape parameter

B =the scale parameter, and

x = dummy variable of integration

The shape and magnitude of the [-PDF are uniquely determined by altering the shape and scale parameters The mass of soil contamination is divided equally among the N

sites Soil desorption can be expressed as follows

qi = mass sorbed in compartment i with respect to qr (g/g)

a; = first-order mass-transfer coefficient for compartment i (1/sec), and

F= we = mass fraction of solute sorbed in each site at equilibrium (assumed equal for all compartments)

This model can more closely describe the rate of release of contaminants from a soil Additionally, confidence intervals can be developed for the model parameters A recent article extends this model to include the determination of the kinetic release parameters

using the temporal moments from the experimental data, in lieu of best-fitting the

experimental data (Cunningham et al., 2005) However, due to the adequacy of the bicontinuum model in most situations, this model will not be used in this research

2.3.4 Two-region Kinetic Model

In lieu of describing kinetic sorption, the two-region model (mobile and immobile regions) was developed to model large-scale physical non-equilibrium conditions (e.g

diffusion into layers of low permeability) In both mobile and immobile regions, sorption between the aqueous and solid phases is assumed to follow a linear equilibrium isotherm and to be instantaneous Mass transfer limitations occur due to slow diffusive mass

transfer from the immobile regions Although mathematically consistent with the

bicontinuum and two-site models, the two-region model is not conceptually equivalent (Valocchi, 1988) and will not be used in this research

2.4 Subsurface Fate and Transport Modeling

Subsurface models consist of mathematical formulas that can be used to represent actual subsurface processes, particularly the fate and transport of chemicals in the vadose and saturated zones A subsurface system can be represented by using any one of a

number of ideal or non-ideal reactor equations The system equation coupled with one of the sorption/desorption equations described in the previous section can be used to create a subsurface source zone model A brief description of some of the systems used in the

literature is presented below

2.4.1 Ideal Systems

2.4.1.1 Completely Stirred Tank Reactor Models

There is no flow into or out of a complete stirred tank reactor models (CSTR) Since a CSTR is considered to be perfectly mixed, it is homogeneous with respect to chemical/physical properties and concentrations Assuming no degradation, a mass

balance on a CSTR used to study desorption kinetics leads to the following ordinary

Equation 2.4.1 describes the decrease of concentration in the sorbed phase and the

corresponding increase in the aqueous phase as time progresses To convert this

expression into algebraic expressions for C and q as a function of time, a desorption equation for dq/dt must be substituted This type of model has been used to describe the results of batch studies in several articles (see laboratory methods concerning batch testing in a later section) and will be used in this dissertation

2.4.1.2 Continuous Flow Stirred Tank Reactor Models

The continuous flow stirred tank reactor model (CFSTR) is identical to a CSTR except that there is flow into and out of the tank As with the CSTR, the system is

considered to be homogeneous with respect to chemical/physical properties and

concentrations Therefore, the effluent concentration is equal to the tank’s concentration