.
u; I

 

: .rdnvvmmau...

MIHM a..."

 

 

 

 

 

 

1
'3»: 9‘

=5. 5.3.!
.3? 51‘. 3..

I .

I ‘ 'n n L ‘9'
‘.£....n.._..>u........m§

I. .151
a: 4.4.;

v3.

 

‘\.:V.r . .zl‘n.

. l... :5 .v. , .e
(.1... flow.)

gag: “menu”... V ' infil..xmwwmr.@ r
i . Swim-alt!“ VT... . .v.! u .A

 

RIBS

Wt

MICHIGAN STATE UNIVERSITY B\

\ Ni\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\‘i\

3 1293 01402 788

 

 

 

 

 

 

i

 

This is to certify that the

thesis entitled

Rates of Removal of Dissolved Gasoline Components
from Laboratory Soil Columns with and without
Microbiological Activity

presented by

Mara E. Hollinbeck

has been accepted towards fulfillment
of the requirements for

M.S. Environmental Engineering

 

 

degree in

in M

Major professor

Date April, 1995

 

0-7639 MS U is an Affirmative Action/Equal Opportunity Institution

 

 

LIBRARY
Michigan State
University

 

 

 

PLACE IN RETURN BOX to roman this checkout from your mood.
TO AVOID FINES Mum on or bdoro duo duo.

DATEIDUE DATE DUE DATE DUE
MAGIC 2 ‘

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

L___L_J
-| [:3

||___J[::l
ll I |

MSU ioAn Namath. Action/Equal Opportunity IMIII‘IOI'I
W

 

 

 

 

 

 

 

 

 

 

 

 

 

m1

RATES OF REMOVAL OF DISSOLVED GASOLINE COMPONENTS
FROM LABORATORY SOIL COLUMNS WITH AND WITHOUT
MICROB IOLOGICAL ACTIVITY

By

Mara E. Hollinbeck

A THESIS
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

MASTER OF SCIENCE

Department of Civil and Environmental Engineering

1995

ABSTRACT
RATES OF REMOVAL OF DISSOLVED GASOLINE COMPONENTS

FROM LABORATORY SOIL COLUMNS WITH AND WITHOUT
MICROBIOLOGICAL ACTIVITY

By

Mara E. Hollinbeck

A preliminary study was performed to identify the factors affecting remediation of
gasoline contaminated soil under saturated flow conditions. After contamination with
groundwater containing benzene, toluene, ethylbenzene, o-xylene, and naphthalene, a
series of “active” columns were seeded with an acclimated mixed culture of
microorganisms and flushed with highly oxygenated nutrient-enhanced groundwater.
“Inactive” columns were flushed using a deoxygenated non-nutrient-enhanced solution.
Effluent contaminant concentrations, dissolved oxygen levels, microbiological activity
levels, and solid phase contaminant concentrations from sacrificed columns were

measured as flushing proceeded.

Microbiological conditions within each column system were significantly different based
on dissolved oxygen data and microbial plating results. Flow interruption showed non-
equilibrium partitioning was occurring. Both column systems approached similar liquid
remediation levels rapidly indicating that microorganisms did not offer a significant
advantage in the extent of contaminant depletion. Mass balance calculations indicate

substantial error associated with solid phase analysis.

PLEASE NOTE

Pa9e(s) not included with original material
and unavailable from author or university.
Filmed as received.

iii-vi

University Microfilms International

TABLE OF CONTENTS

List of Tables ...................................................... ix
List of Figures ..................................................... x
CHAPTER

I. INTRODUCTION AND RESEARCH OBJECTIVES

1.1. Introduction ............................................. 1
1.2. Research Objectives ....................................... 3
1.3. Research Approach ....................................... 4
H. LITERATURE REVIEW ....................................... 6

III. MATERIALS AND METHODS

3.1. Soil Characteristics ....................................... 12
3.2. Contamination .......................................... 13
3.3. Apparatus/Operating Conditions
3.3.1. Packed Soil Columns ............................... 14
3.3.2. Carbon Dioxide Flush and Water Saturation ............. 14
3.3.3. Contaminant Saturation ............................. 16
3.3.4. Seed Solution and Seeding Procedure .................. 17
3.3.5. Flushing Solutions
3.3.5.1. Inactive Columns .......................... 18
3.3.5.2. Active Columns ........................... 19
3.4. Operating Procedure ..................................... 19
3.5. Sampling Procedures
3.5.1. Liquid Samples ................................... 21
3.5.2. Solid Samples .................................... 22
3.6. Analysis Procedures .
3.6.1. Gas Chromatography ............................... 22
3.6.2. Soil Biomass Plating ............................... 23

vii

viii

IV. RESULTS
4.1. Microbiological Data .................................... 27
4.2. Oxygen Data .......................................... 28
4.3. Liquid Phase Data ...................................... 30
4.4. Solid Phase Data ....................................... 37

V. DISCUSSION, CONCLUSIONS, AND RECOMMENDATIONS

5.1. Discussion ............................................ 51

5.2. Conclusions ........................................... 56

5.3. Recommendations ...................................... 57
APPENDIX

A. CONTAMINANT RECOVERY FROM SOIL

A. 1. Contaminant Recovery .................................. 59
A2. Effect of Soil Mass ..................................... 60
B. EFFECT OF CONT AMIN ANT AGE ON RECOVERY ............. 62
C. BATCH SORPTION EXPERIMENT ........................... 64

LIST OF REFERENCES .......................................... 69

TABLE

4.4.1.

A.1.1.

A.2.1.

8.1.

CI.

C2.

C3.

C4.

LIST OF TABLES

Comparison of Expected and Actual Results .................... 50
Contaminant Recovery from Spiked Soil ....................... 60
Effect of Soil Mass on GC Response .......................... 61
Estimated Kd Values Based on Contaminant Age ................. 63
Parameters Used to Design Batch Sorption System ............... 65
Thimerosal Effect on K, ..................................... 67
Sodium Sulfite Effect on K, .................................. 67
Partition Coefficients Based on Batch Sorption Experiment ........ 68

ix

FIGURE

3.1.

4.2.1.

4.3.1.

4.3.2.

4.3.3.

4.3.4.

4.3.5.

4.3.6.

4.4.1.

4.4.2.

4.4.3.

4.4.4.

4.4.5.

4.4.6.

4.4.7.

4.4.8.

4.4.9.

LIST OF FIGURES

Schematic of Operating System .............................. 15
Influent and Effluent Dissolved Oxygen ....................... 29
Effluent Benzene Concentrations ............................. 31
Effluent Toluene Concentrations ............................. 32
Effluent Ethylbenzene Concentrations ........................ 33
Effluent o-Xylene Concentrations ............................ 34
Effluent Naphthalene Concentrations ......................... 35
Effluent COD ............................................ 36
Solid Phase Benzene Concentrations ......................... 38
Solid Phase Toluene Concentrations .......................... 39
Solid Phase Ethylbenzene Concentrations ...................... 40
Solid Phase o-Xylene Concentrations ......................... 41
Solid Phase Naphthalene Concentrations ...................... 42
Solid Phase COD ......................................... 43
Cumulative Mass of Benzene Removed ....................... 44
Cumulative Mass of Toluene Removed ....................... 45
Cumulative Mass of Ethylbenzene Removed ................... 46

xi
4.4.10. Cumulative Mass of o-Xylene Removed ..................... 47

4.4.11. Cumulative Mass of Naphthalene Removed .................. 48

4.4.12. Cumulative COD Removed ............................... 49

CHAPTER 1

INTRODUCTION AND RESEARCH OBJECTIVES

1.1. Introduction.

Significant attention has been focused on the issue of groundwater remediation specifically
related to petroleum hydrocarbons due to contamination from leaking underground storage
tanks (U STs). From a regulatory standpoint, the alkylbenzene components (or light aromatic
compounds), which include benzene, toluene, ethylbenzene, and xylene (BTEX), are of
particular concern due to their potential for rapid migration relative to other contaminants
(Voice, 1993). Drinking water contamination is the primary route of human exposure to
these chronically toxic compounds (Voice, 1993). Heavier components of gasoline,
including naphthalene (a polyarornatic hydrocarbon, PAH), are also of interest for

comparative purposes due to their increased persistence.

Pumping contaminated groundwater to above ground treatment facilities (pump and treat)
has been, and continues to be, an accepted and common method of treatment, however,
sorption of contaminants to soil and organic matter tends to complicate removal rates.

Pumping a system until the removed water meets a desired level of treatment does not ensure
complete aquifer restoration (Piwoni and Keeley, 1990). Desorption, driven by the

concentration gradient between the sorbed and liquid phases, slowly releases the

2
contaminants, resulting in long treatment times. In addition, after removal from the ground,

the contaminated water still requires treatment using above ground bioreactors or activated

carbon adsorption systems.

In comparison, in-situ biodegradation has the potential to convert the contamination to
innocuous byproducts without removal from the ground. In addition to this advantage, in-
situ bioremediation is receiving more attention as a result of the problems associated with
pump and treat technologies. Sufficient information on successful remediations, however,

has not been provided which prevents assessment of the degree to which each process and
parameter influences the performance of biorernediation (Abriola and Chen, 1993). In either

case, there remain unanswered questions regarding the relationship between sorption and

biodegradation.

The purpose of this study was to make a direct comparison between two systems which
loosely represent the two technologies commonly used to remediate contaminated
groundwater, namely pump and treat and in-situ biodegradation. Of primary interest was the
rate of remediation observed using either technology to decrease the level of soil
contamination in two one-dimensional saturated soil column systems contaminated with
dissolved gasoline components. The two systems differed only in tint one was biologically
active (to represent in-situ biodegradation) while the second was inactive (to represent
removal by pumping alone). This direct, quantitative comparison was intended to allow

conclusions to be drawn regarding the relative rate of clean-up allowed in either system

3
along with possible factors affecting this rate. The target compounds include benzene,

toluene, ethylbenzene, and o-xylene (BTEX) along with naphthalene.

1.2. Research Objectives.

The hypothesis underlying this research is that the degree to which mass transfer limits solid
phase removal rates of dissolved gasoline components from saturated soil columns will be
decreased by the presence of degradative microorganisms. The underlying assumption is
that solid phase removal by degradation is unlikely, therefore, under desorption limiting
conditions (maintained by optimizing degradation conditions) the effect of microorganisms
can be assessed by comparing the two systems under similar conditions. The
microorganisms may increase solid phase removal rates by decreasing liquid phase
concentrations which results in a higher driving force for desorption. Altemately, or in
addition to this effect, the microorganisms may have surfactant-like properties which

improve desorption.

Based on the above paragraph the three-fold objective can be explicitly stated m follows. (1)
To compare two saturated soil systems one of which is microbiologically active while the
second is inactive; (2) To quantify the solid phase removal rates of dissolved gasoline
components sorbed to saturated soil through either flushing alone or in the presence of
degradative microorganisms; and (3) To assess the effect microorganisms have on the
desorption limitation. The results of this study were intended to provide a quantitative

comparison between the two primary technologies used to remediate contaminated

groundwater.

1.3. Research Approach.

The first step in meeting the specified research objectives was to design a system with the
features necessary to isolate the variables of interest. This process of broadly addressing
each issue constitutes the research approach. The importance of this preliminary stage is
justified when considering it the foundation for the entire project. In addition, the
fundamental reasoning behind use of the experimental system is beneficial when attempting

to apply conceptual ideas to further research or research in related areas.

There are essentially four factors upon which the core system features were based. The first
factor was the interest in groundwater contamination which specifies that packed soil column
systerm be used as representative of that dynamic environment. The second factor is the
measurement of solid phase contaminant levels which requires the use of multiple columns
due to the destructive nature of this procedure. The third factor is related to the comparison

aspect which requires two series of columns to be used which differ only in tint one

contains microbiological activity while the other is sterile. This assists in isolating the effect
microorganism have on removal rates. Finally, the interest in desorption limitation effects
dictates a system in which measurements are taken during column flushing after

contaminants have been previously sorbed to the soil.

5
The result when combining the mentioned factors is a system which includes two series of

multiple packed soil columns (active and inactive) flushed to remove previously sorbed
contaminants. The remaining components which make up the system are important yet don’t

change the overall operating system.

CHAPTER 2

LITERATURE REVIEW

Significant attention has been focused on the issue of groundwater remediation specifically
related to petroleum hydrocarbons due to contamination from leaking underground storage
tanks (U STs). From a regulatory standpoint, the alkylbenzene components (or light aromatic
compounds), which include benzene, toluene, ethylbenzene, and xylene (BTEX), are of
particular concern due to their potential for rapid migration relative to other contaminants
(Voice, 1993). Drinking water contamination is the primary route of human exposure to
these chronically toxic compounds (Voice, 1993). Heavier components of gasoline,
including naphthalene (a polyaromatic hydrocarbon, PAH), are also of interest for

comparative purposes due to their increased persistence.

Pumping contaminated groundwater to above ground treatment facilities (pump and treat)
has been and continues to be an accepted and common method of treatnrent, however,
sorption of contaminants to soil and organic matter tends to complicate removal rates.

Pumping a system until the removed water meets a desired level of treatment does not ensure
complete aquifer restoration (Piwoni and Keeley, 1990). Desorption, driven by the
concentration gradient between the sorbed and liquid phases, slowly releases the

contaminants, resulting in long treatment times. In addition, after removal from the ground,

7
the contaminated water still requires treatment using above ground bioreactors or activated

carbon adsorption systems.

In comparison, in-situ biodegradation has the potential to convert the contamination to
innocuous byproducts without removal from the ground. In-situ bioremediation is also
receiving more attention as a result of the problems associated with pump and treat
technologies, however, there remains unanswered questions regarding the relationship
between sorption and biodegradation. In addition, sufficient information on successful
rernediations has not been provided which prevents assessment of the degree to which each
process and parameter influences the performance of bioremediation (Abriola and Chen,

1993).

Coupled-process (biodegradation and sorption) research was initiated by soil scientists
studying the fate of pesticides in soils (Angley et al., 1992). More current research has
focused on the bioavailability of sorbed contaminants to microbial species. Sorption,
theoretically can cause either an increase or a decrease in bioconversion rate (Rijnaarts et al.,
1990). An increase would result if the contaminant is toxic to the microbes since sorption
will reduce the level of solution contaminant thereby lowering the impact on the microbes
using only dissolved substrates. More often a decrease in rate is seen for one of two reasons.
The first is due to a decrease in aqueous phase concentration below threshold levels while a
second effect is realized when desorption controls bioconversion which limits high potential

rates due to slow desorption velocities.

8
Numerous studies have been performed with the specific intent of assessing the relationship

between sorption and biodegradation. Results vary depending on the contaminant and
contaminated matrix (sediments or soil), organic matter content, moisture content,
microorganism(s) involved, age of contamination, and type of system (batch or colurrm).

Generally the studies can be divided into three interrelated areas of interest: bioavailability,

modeling issues, and rate limitations.

Modeling is of value not only as a method for quantifying the transport and degradation of
contaminants but for use as a predictive tool and way to understand the mechanisms involved
in both processes. Sorption was typically found to follow a two-phase process whereby a
fraction sorbs instantaneously while the remainder sorbs in a time-dependent or rate-limited
rmnner (van Genuchten et al., 1989; Angley et al., 1992). Rate limited sorption models were
accordingly found to fit data better than models which assume local equilibrium exists
(Angley er al., 1992) except for compounds which sorb to a lesser degree (i.e. benzene).

Models also exist which include the use of the local equilibrium assumption (LEA) (Scow
and Hutson, 1992). The validity of this assumption can be tested using the flow interruption

technique as described by Pennell et al. (1993).

Non-linear sorption has been attributed to structural heterogeneity at the intraparticle scale in
terms of microporosity while hysteresis was found to contribute due to irreversible sorption
to organic matter (Farrell and Reinhard, 1994). Linear sorption is also a common assumption

(Scow and Hutson, 1992). Slow release (desorption) of compounds has been attributed to

9
restricted diffusion through both soil organic matter and intraaggregate pores (Farrell and

Reinhard, 1994). Residual contaminants also resist mobilization and microbial degradation
due to entrapment in intraparticle micropores (Steinberg et al., 1987). Biodegradation is
often assumed to follow first-order kinetics (Angley et al., 1992) however, exceptions do
exist For example, Scow and Hutson ( 1992) found that biodegradation was initially fast due
to solution phase degradation followed by a slower rate, limited by desorption/diffusion. It is
also interesting to note that while some researchers have identified the possibility of
enhanced desorption due to microorganisms (W szolek and Alexander, 1979; Gordon and

Millero, 1985; Rijnaarts et al., 1990), no models can accommodate this effect.

Bioavailability of sorbed compounds to degradation has also been addressed in some detail.
The most cited of these studies found that sorbed 2,4-dichlorophenoxyacetic acid (2,4—D)
was completely protected from microbial degradation but solution 2,4~D was degraded with
equal efficiency by both solution and sorbed microorganisms (Ogram et al., 1985). On the
other hand, adsorption of phenol, an antimicrobial agent at high concentrations, has been
found to increase degradation rates (Scott et al., 1983; van Loosdrecht et al., 1990; Shimp
and Young, 1988). Chlorophenols fall somewhere between the two extremes since a
"surface" fraction of bound substance was found to be releasable while the remainder is

bound to a "core" which is entirely inaccessible to microorganism (Dec and Bollag, 1988).

Gordon and Millero (1985) tested various substrates that sorb to different degrees on

hydroxyapatite and found decreased availability to be a function of the degree of adsorption.

10
Organism-specific properties were found to dictate whether and to what extent sorbed

naphthalene is biologically available by Guerin and Boyd (1992). Finally, quaternary
ammonium compounds (QACs - surfactants) were found to be available in the sorbed phase
with the explanation being that they were utilized as sole carbon and energy source and have

specific enzymes to bind and transport the QACs (Larson and Vashon, 1983).

Rate limitations are primarily thought to be due to desorption (van Loosdrecht er al., 1990;
Rijnaarts et al., 1990), however, this may not initially be limiting (Robinson et. al., 1990).
Alvarez-Cohen et al. (1993) found the transformation reaction to be rate limiting while
Weissenfels er a1. (1992) associated the type of binding with degradation limitations.
Finally, diffusion/desorption was found to be faster in non-sterile systems due to a steeper
concentration gradient resulting from microbial removal and decreased diffusion distances

(Rijnaarts et al., 1990; van Loosdrecht et al., 1990).

Of the research related to this proposal, two studies are of particular interest. One of these
performed a direct comparison experiment to test the effect sorption has on biodegradation
with the use of three systems (Voice et al., 1992). One system allowed contaminant removal
using both sorption and biodegradation (biologically active carbon fluidized bed) while the
other two considered removal by only one of the mechanisms, either sorption (granular
activated carbon fluidized bed) or biodegradation alone (non-activated carbon with biofilm).
This work concluded that substrate sorbed to the outer portions of the particles is readily

available to microorganisms. More importantly, the study provided incentive regarding the

1 1
type of system necessary to isolate the information of interest to this work.

A second study used computer modeling to compare pump and treat, biorestoration, or a
combination of the two technologies under varying conditions (Marquis and Dineen, 1994).

Limited published documentation involving direct comparisons between the systems was the
impetus for the study. Either in-situ biorestoration alone or in combination with pump and
treat were found to be remedial options for sand and gravel aquifers contaminated with

biodegradable organic compounds.

The overwhelming interest in comparing rermdiation alternatives based on efficiency and
expediency of site clean-up combined with the scarcity of direct quantitative comparisons
validates the need for this work. In particular, the relative merits of pump and treat versus in-
situ bioremediation technologies are debated with each side convinced their choice is the
"best" often times based only on qualitative comparisons. This study intends to quantify this

on-going debate and provide additional “food for thought”.

CHAPTER 3

MATERIALS AND METHODS

The two series of packed soil columns which were operated simultaneously are referred
to as active and inactive systems based on the presence or absence of microbiological

activity, respectively.

3.]. Soil Characteristics.

The soil used for this work was a sandy soil collected from Clare County, Michigan in the
spring of 1989. The organic matter content of the soil was 0.5% as determined using a
soil adapted CHN analyzer. Since a direct comparison was made between the two sets of
columns, it was not necessary to quantify soil parameters such as porosity and bulk
density except to verify consistency from column to column. The porosity was 0.35 :l:
0.01 while the bulk density was found to be 1.73 :I: 0.04 g/cm3. The pore volume was
thus calculated to be 4.22 and used as a basis for graphical comparisons. Equilibrium
sorption/desorption partitioning coefficients for this soil were determined to assist in
estimation of solid phase contaminant levels using effluent concentrations. The results of

this sorption study are presented in Appendix C.

12

1 3
3.2. Contamination.

Simulated gasoline-saturated water reservoirs were generated and used as a uniform
contaminant source during contaminant saturation. The target compounds, benzene,
toluene, ethylbenzene, o-xylene, and naphthalene (BTEXN), were dissolved in each other
in the ratios found in gasoline and spiked into water to create levels similar to those found
in gasoline-saturated water. The concentrated BTEXN stock solution was prepared by
dissolving 0.9 g naphthalene in 40 mL toluene and adding 7.4 mL of this solution to 5.6
mL benzene, 2 mL ethylbenzene, and 1.1 ml. o-xylene. Multiple reservoirs were then
generated by spiking the concentrated BTEXN stock solution into 160 mL serum vials
containing deionized water passed through a 0.22 micron filter, two drops of 6 M HCl,
two (2) 4 mm glass beads, and approximately 10 mL of headspace. Each reservoir was
then sealed, tumbled for two (2) days to promote dissolution and uniform distribution,
and refrigerated until use. The initial concentrations of each component were determined
and verified periodically throughout the experiment prior to filling the feed syringes.
Using the five (5) components mentioned represents the various fractions of contaminants

present in gasoline.

Tim hydrochloric acid was used to prevent microbiological degradation of the substrates
prior to or during contaminant saturation. A result of its use was a decrease in column pH
which may have created conditions unfavorable for microbiological growth. For this
reason, sodium hydroxide (NaOH) was added to the final contaminant solution to

produce neutral pH conditions within the columns prior to inoculation.

1 4
3.3. Apparatus/Operating Conditions.

3.3.1. Packed Soil Columns.

Two series of 10 borosilicate glass columns (Kontes, 1.0 cm in diameter x 15 cm long)
were used with identical fittings and physical constraints. All soil was sterilized initially
by gamma-irradiation (2 Mrad). This was necessary to prevent microbiological activity
in the inactive columns and was subsequently necessary for the soil in the active columns
and for that used in standard curve generation. This precaution addresses the possible
effect gamma-irradiation has on the structure of the soil organic matter which could

change the sorption characteristics of the soil.

Aseptic column packing included use of sterile gloves and autoclaved equipment in a
laminar flow hood. Soil was packed into the columns incrementally using a glass rod to
mix after each addition to create uniform packing and prevent layering effects. The bulk
density was calculated using the soil mass (found by weighing the columns before and
after soil packing) in conjunction with the column volume occupied by soil. The column
pore volume and porosity were also determined using these parameters to verify

consistency among columns.

3.3.2. Carbon Dioxide Flush and Water Saturation.
The dry soil columns were initially flushed with 240 mL of carbon dioxide (C02), at a
rate of 10 mL/hr, to replace the air in the pore space. This prevents entrapment of gas

during water saturation since the C02 will subsequently dissolve. A feed rate of 10

15

10 mL plastic D.O.
collection syringe

-------
>.-\.-.--. .-\.;
---------

     
  

1 mL glass eflluent
collection syringe

3-way valve

packed soil column

 

 

10 mL gas tight

teflon-lined stainless steel .
influent tubing . glass feed synnge

multi-head

 

 

 

 

syringe pump

 

Figure 3.1. Schematic of Operating System

 

16
mleay was used to water saturate the columns in an upflow mode. The water was

obtained from the aged (to ensure volatilization of any chlorine residual) tap water
reservoir maintained throughout the experiment for consistency. Prior to use, this water
was autoclaved to assist in maintaining sterile conditions during contaminant saturation.
Plastic syringes and teflon tubing were used during this portion of the experiment to
connect influent feed syringes to the base of the columns. Feeding was consistently

performed using Harvard Apparatus (Model 22) syringe pumps with multi-head adapters.

3.3.3. Contaminant Saturation.

Maintenance of sterile conditions prevented the presence of microorganisms from
impeding contaminant saturation which could cause a concentration gradient to develop
instead of the desired uniform contamination. This was accomplished by filtering the
water used to prepare contaminant solutions, autoclaving the feed syringes, and soaking
the valves, tubing, and connectors in a 70% ethyl alcohol solution. The columns were
saturated with contaminant solution at an increased rate until the effluent concentrations
stabilized. After this point, the rate was decreased to 10 mleay to verify that saturation
levels were reached. Contaminant saturation continued for a total of approximately 3.5

months.

Four (4) syringe pumps with multi-head adapters were used to accommodate five (5)
syringes apiece to ensure the columns received identical feed rates. It was inevitable that

losses still occurred due to the nature of the columns however keeping uniform influent

17
concentrations and monitoring effluent concentrations allowed the degree of consistency

among columns to be assessed. Use of gas-tight glass syringes (10.0 mL) during
contaminant saturation helped to ensure uniform influent conditions and prevented
significant losses due to sorption or volatilization. Gas-tight glass syringes (sterilized by
autoclaving for 20 minutes at 121°C) were also used during flushing to prevent oxygen
loss or gain (depending on the system and corresponding concentration gradients) due to
diffusion. Teflon lined stainless steel tubing (1/8") was used to prevent changes in
influent oxygen levels during flushing due to diffusion through teflon or other types of

tubing.

3.3.4. Seed Solution and Seeding Procedure.

A mixed culture obtained from an acclimated fluidized bed reactor used to treat BTX
contaminated groundwater for three (3) years was used to seed the active columns.
Growth of this biomass seed solution included addition of 250 mL of reactor seed to tap
water to create a total volume of 1000 mL. Trace nutrients were present in the tap water
and bulk nutrients were provided by addition of 3 mL of concentrated solution containing
33 g/L NH4C1 and 7.8 g/L KH2P04. Substrate consisted of the concentrated BTEXN
stock solution prepared in conjunction with the contaminant solutions described
previously. The stock substrate solution was added to the seed solution to create a
toluene concentration of approximately 100 ppm (45 uL BTEXN stock). The seed
solution was then sealed with a rubber stopper and mixed continuously using a stir bar

and plate. Periodic addition of additional substrate and nutrient solutions ensured

18
continued bioactivity prior to column seeding. The approximate organism density was

determined by plate counts.

The seeding procedure consisted of adding 15 mL of seed solution at the high rate of 100
mL/day using plastic syringes, teflon tubing, and syringe pumps. The inactive columns
were also flushed at the same rate but using 15 mL of the appropriate flushing solution to
maintain comparable conditions. Influent and effluent samples were plated using a series

dilution procedure to estimate the percentage of biomass retained in the columns.

3.3.5. Flushing Solutions.

3.3.5.1. Inactive Columns.

The use of thirnerosal, a mercury derived biocide, to maintain sterile conditions within
the inactive columns was tested and found to affect sorption of the compounds
significantly. For this reason, a nitrogen flushed reservoir was used to displace the
oxygen in the water and maintain low oxygen levels in the influent flushing solution. The
reservoir consisted of a water filled 1.0 L graduated cylinder flushed with nitrogen gas by
use of a flow meter, tubing, and a diffuser. Maintaining low influent oxygen levels was
intended to prevent significant microbiological activity within the columns. Addition of
microorganisms from the flushing solution was prevented by filtering the water through a
0.22 um filtering unit (Millipore Sterivex-GS) when filling the feed syringes. Sodium
sulfite was also tested in conjunction with nitrogen flushing to serve as an oxygen

scavenger and remove the remaining trace levels of oxygen, however, it was also found

19
to have an effect on sorption therefore could not be used.

3.3.5.2. Active Columns.

The seeded organisms were enhanced using an oxygenated nutrient flushing solution
generated by addition of nitrogen and phosphorus to aged tap water (to ensure
volatilization of any chlorine residual). Sufficient nutrient levels were determined based
on the maximum decrease in chemical oxygen demand under maximum oxygen provision
to prevent nutrient limitations. A flowmeter was used to mix gaseous nitrogen and pure
oxygen (~50% of each) prior to addition to a 1.0 L graduated cylinder using a diffuser to
create a consistent reservoir containing between 13 and 28 mg M as needed. The
effluent oxygen levels were ideally maintained above 4 mg OzlL by adjusting the influent

oxygen levels to ensure an oxygen limitation wasn’t encountered.

3.4. Operating Procedure.
Operation of the aseptically packed soil column systems consisted essentially of five
stages. Each is described in detail elsewhere however the following is a summary to

integrate each procedure in terms of stages.

Sgge 1: Carbon dioxide flush. The intention here was to prevent air entrapment by
replacing the gas in the pore space with carbon dioxide which subsequently dissolves

upon water saturation.

20
Stage 2: Water saturation. Performed to create the saturated soil conditions of interest to

this study.

S_tagfi; Contaminant saturation. This stage continued for some time after effluent levels
stabilized to allow time for intraparticle sorption to occur. These conditions more
closely resemble the contamination encountered in the field although it is acknowledged
that the time frame involved does not adequately address the issue of long term

contamination.

Stage 4: Inoculation of active columns. Performed to provide an acclimated
microbiological culture to “activate” the microbiologically active system. The inactive

system was also flushed during this stage to keep the systems similar for comparison

purposes.

Stage; Flushing. The previous stages were in preparation for this final stage which was
intended to address the objectives of this research by looking at removal of sorbed
contaminants from soil by flushing with non-contaminated water. The active columns
were flushed with oxygenated nutrient solution while the inactive columns were flushed
with deoxygenated nutrient-free water. Sampling was performed during this desorption
mode of operation to allow the effect of desorption on removal rates in both cases to be
investigated and offer insight into the effect that microorganism have on desorption.

Partitioning coefficients (determined in the batch sorption study) were used to

21
back-calculate the solid phase contaminant levels in each column based on effluent

concentrations. This would ideally allow the columns to be sacrificed at intervals which

appropriately generate the removal curves in their entirety.

3.5. Sampling Procedures.

3.5.1. Liquid Samples.

Both influent and effluent contaminant concentrations were determined during initial
column contamination to determine the time it takes to reach saturation ("complete"
adsorption). During flushing, 0.2 mL effluent samples were collected in triplicate using
1.0 mL glass syringes connected to valves on the effluent end of each column. Influent
and effluent dissolved oxygen (D.O.) levels were monitored (using an oxygen probe) in
the active columns to determine the oxygen uptake. These measurements were intended
to provide an indication of conditions in the columns and assist in attributing losses to the
appropriate cause while also verifying that the mixed culture seed had acclimated.
Dissolved oxygen levels in the inactive columns were also monitored to verify that

oxygen in the system remained low as intended.

All liquid samples were collected by increasing the feed rate to 5 mL/day to allow
collection in a short, discrete time period. The importance of this method of sample
collection is two-fold: (1) to avoid loss of contaminants due to volatilization during
collection and (2) to prevent biased oxygen results from changes in effluent oxygen

levels with time due to diffusion through the plastic collection syringes. Additionally, in

22
both systems, 10 11L effluent samples were periodically plated on agar pour plates to

qualitatively assess that bioactivity was present (in the active system) or absent (in the

inactive system) as desired.

3.5.2. Solid Samples

Initial solid phase contaminant levels were assessed by sacrificing and analyzing the soil
in the first column from each series prior to flushing but after the microbiological
inoculation procedure. The consistency of contaminant conditions among columns was
verified by monitoring effluent concentrations which stabilized to similar levels after

sufficient contamination occurred.

After a column was sacrificed (removed from operation), it was sealed and frozen (to
prevent volatilization losses) and removed as a slug which was physically partitioned into
uniformly sized pieces (=1 g/sample) to allow a comparison of conditions at multiple
locations within each column in addition to the total contaminant mass. Headspace vials
were weighed prior to use to allow soil mass per sample to be determined after analysis

and drying to remove moisture.

3. 6. Analysis Procedures.
3.6.1. Gas Chromatography.
Liquid samples were analyzed using a combination Perkin-Elmer headspace sampler

(model 11840) and gas chromatograph (GC Autosystem) with a megabore column (PE

23
624: 30 m x 0.53 mm) fitted with a flame ionization detector (FID) using helium as the

carrier gas. Solid samples were analyzed using the static headspace analysis method
described by Voice and Kclb (1994). Briefly, this entails adding 5 mL chilled deionized
water to each frozen soil sample contained in a headspace vial. The samples were sealed
using teflon-coated septa and aluminum crimp tops. Soil calibration standards were
chilled prior to contaminant spiking to minimize volatilization losses followed by rapid
shaking for 1 hour and equilibration overnight prior to preparation and analysis.
Sonication was tested and not found to improve recovery therefore was not used as part

of the analysis procedure.

The operating conditions of the GC were as follows. Each liquid sample was equilibrated
at 90°C for 1 hour while the solid samples had a therrnostating time and temperature of 1
hour at 95°C. The GC oven temperature started at 60°C and ramped at 15°C/min to a
final temperature of 200°C which was held constant for 1 minute. Both high and low
sensitivity methods were used to span the full range of each contaminant encountered. A
minimum of three check standards and blanks were analyzed in conjunction with each

sample analysis period. The detection limits for all compounds was 1 jig/L or less.

3.6.2. Soil Biomass Plating.
To obtain a quantitative estimate of the microbial numbers on the column soil, the
biomass was removed from one column fiom each system, plated and incubated in

growth chambers. In this manner, a comparison between the levels of biological activity

24
specifically due to growth on the substrates of interest in either the active or inactive

columns was afforded. This comparison is important to attribute any difference in solid
phase contaminant concentration to the presence of microbial activity. The specific

procedure and equipment used are described in the following text.

Agar pour plates were prepared in advance by adding the following to each liter of
deionized water: 5 grams of bacto-agar (no carbon source), 1.0 mL trace element
solution, and 100 mL major ion solution. The trace element and major ion solutions were
prepared as described by Ridgway et al. (1990). The entire solution was autoclaved at
121°C for 20 minutes, tempered, and poured into sterile plates (60x15 mm Corning
polystyrene tissue culture dishes). The pH 7 phosphate buffer solution contained per liter
of deionized water 2 g monobasic phosphate (KH2P04) and 3.5 g dibasic phosphate
(K2H2P04). Disposable culture tubes (Baxter 13x100 mm, T1290-4) were filled with 1.8
mL of this buffer solution, capped (Kimble glass natural kim-kaps, 13 mm, T1291-13A),

and autoclaved prior to use.

Triplicate 1 g soil samples were aseptically removed from each sacrificed column,
weighed, and added to the first dilution tubes (containing phosphate buffer) in the
triplicate dilution series. Each was mixed for 3 minutes (using a touch mode Therrnolyne
mixer) followed by a 2 minute wait in a 3-cycle period for a total of 15 minutes. This
first stage was intended to remove the biomass from the soil. A dilution series was then
performed by adding 200 uL from the first tube to the second and 200 pl. from the

second to the third and so on to give successive 1:10 dilutions. Ten (10) pl. from each

25
tube ,was then plated to give 104, 103, 10“, 10's, 10‘, 10'7 , and 104‘ dilutions. The

dilutions and plating were all done using new sterile automatic pipette tips prior to each
step to prevent cross contamination. Plates were spread using a sterilized glass rod and

turntable.

After plating, seven (7) bacto-agar plates were placed in a growth chamber containing a
headspace vial saturated with concentrated BTEXN substrate solution and cotton to
prevent leakage. The concentrated BTEXN solution was prepared as described for
contaminant solution preparation. Vapors then serve as the carbon source for growth of
organisms. Growth will not occur if the concentration is too high (inhibitory to the
microbes) or too low (carbon limiting) however, it was found previously that adding 0.5
ml. of BTEXN solution was appropriate to allow for growth if organisms with the ability

to degrade these compounds are present.

Fourteen (14) additional plates were prepared to serve as controls. .Seven (7) of these
consisted of 10 uL from each dilution tube spread on bacto-agar plates incubated outside
of the growth chambers. This provided an indication of the number of organisms which
may grow on trace substrates found in the agar due to impurities. The remaining 7 plates
contained 10 pL from each dilution tube on nutrient agar (23 g/L H20) also grown
outside of the growth chambers. These plates allow the total number of organisms

present to be quantified and provide a comparison to the number which grow on BTEXN.

The appropriate incubation period was determined by plating 10 pl. of the seed solution

26
on replicate bacto-agar plates, placing them in a growth chamber with the appropriate

vapor carbon source, and monitoring growth. When no new colonies appeared from one
day to the next, this time period was selected as the appropriate incubation period and all

plates were counted using this guideline of 5 days.

CHAPTER 4

RESULTS

4.1. Microbiological Data.

The density of the microbiological seed used to inoculate the active columns was
determined by plating on nutrient agar pour plates and on bacto-agar pour plates
incubated for five (5) days in growth chambers which contained contaminant vapors. In
addition, the effluent from three columns (to generate triplicate samples) was collected
during inoculation and plated on nutrient agar pour plates to estimate the percent of

organisms which were retained in the columns.

It was found that 2.66 i 0.33 x 104 colony forming units per milliliter (CFU/mL) of
acclimated target substrate degraders were present in the seed solution based on the
growth chamber incubated bacto-agar plates. Growth of this culture on nutrient plates
resulted in 1.24 :t 3.6 x 104 CPU/ml. which is statistically similar to the growth chamber
results indicating that the vast majority of organisms had the ability initially to degrade
benzene, toluene, ethylbenzene, o-xylene, and/or naphthalene. Effluent plating produced
variable results of 2.10 :i: 1.7 x 102 CFU/mL which indicates a maximum retainage of

approximately 98%.

27

28
Triplicate soil biomass plating was performed on one column from each system after

flushing for 17.5 days (41.5 pore volumes). Samples were collected from three (3)
locations within the columns to try to assess the distribution of organisms. These
locations included the column inlet and two points in the middle of the soil columns.
Triplicate plates consisting of bacto-agar in growth chambers and controls of bacto-agar

with no substrate plus nutrient agar plates were counted.

The non-incubated bacto-agar controls contained more colonies after 5 days than the
incubated bacto-agar plates. This indicates that significant growth occurred on trace
substrates present in the agar with possible inhibition occurring in the presence of the
target substrates. Nutrient agar plates were therefore used as an indication of the soil
microbial numbers assuming the majority of growth is comprised of BTEXN degraders.
No trend was seen in terms of location within the columns, however, the active column
contained an average of 4.28 x 108 CFU/g soil compared to 3.20 x 10‘5 CFU/g soil in the

inactive column.

4.2. Oxygen Data.

lnfluent oxygen levels were measured daily and maintained consistently. Effluent
oxygen samples were collected periodically by increasing the flushing rate to 5 mL/hr to
allow collection in a shortened period of time in an attempt to minimize bias due to
atmospheric diffusion. Figure 4.2.1. provides a summary of influent and effluent

dissolved oxygen concentrations. Effluent levels were averaged over the columns in each

29

 

 

 

 

 

 

 

 

 

35
30 4»
I
25 i-
l
0 2° ‘
on
E. .
O. 15 .L t +Inactrvc Influent
d , -)(- Active Influcnt
+ Active Effluent
10 i * -l— Inactive Effluent
ll
_ t-
0 r ‘r i 4 i i
0.00 20.00 40.00 60.00 80.00 100.00

# Pore Volumes

Figure 4.2.1. Influenth Effluent Dissolved Oxygen.

120.00

30
series with typical deviations of less than 15% except for the later active column values

which increased in variability to a maximum of 35%. An increase in oxygen
concentration by approximately 2 mg/L was associated with the influent fed to the
inactive columns while approximately 20 mg/L of oxygen was consistently consumed in

the active columns.

4.3. Liquid Phase Data.

Liquid phase data consists primarily of effluent contaminant concentrations analyzed
using headspace gas chromatography. Between one and three 0.2 mL samples were
analyzed from each column at regular intervals and average values for each series of
columns were used for interpretation. Deviations from column to column within the
column series were typically less than 10% however this variability increased to a
maximum average of 25% as the sample size decreased due to column sacrifice and as
concentrations dropped into the lower ranges. Figures 4.3.1 - 6 provide a comparison
between active and inactive concentrations for each target compound in addition to a

combined removal represented by chemical oxygen demand.

To test whether equilbrium partitioning was occurring in the column systems, flow was
interrupted for a period of four days starting on day 40. Effluent samples were collected
prior to interuption and after flow was resumed. Comparison of the contaminant
concentrations in the samples showed a two-fold increase in the effluent levels after

resumption (see Figures 4.3.1 - 6). This indicates non-equilibrium partitioning was

31

 

 

 

 

 

 

 

 

 

7
-l- Inactive Columns

6 T + Active Columns

5 ...
E

4 -_
.95
g 3 __ flow interuption
U

0

2 ._

l -_

0 . i : : i

0 20 40 60 80 100 120

# Pore Volumes

Figure 4.3.1. Effluent Benzene Concentrations

32

 

 

 

 

 

20

13 if -I- Inactive Columns
16 l -0- Active Columns
14 ..

 

Conc. (ppm)
8

 

 

 

 

 

 

8 I flow interuption
5 -.
4 a
2 «L 2
0 i i 4 i i
O 20 40 6O 80 100 120

# Pore Volumes

Figure 4.3.2. Effluent Toluene Concentrations

33

 

 

 

 

 

5
4, 5 ._ ~I- Inactive Columns
4 ,_ -0- Active Columns
3.5 4r

Conc. (ppm)
N
U!

flow interuption

 

 

 

 

0 20 40 60 80 100 120
# Pore Volumes

Figure 4.3.3. Effluent Ethylbenzene Concentrations

34

 

 

 

 

 

 

 

 

 

 

3.5 1
+ Inactive Columns
3 I +Active Columns
2.5 ~-
2, 2 i
3‘
o“
g 1.5 -—
U flow interuption
1 .4-
0.5 ~
0 t .L t i t
0 20 40 60 80 100 120

# Pore Volumes

Figure 4.3.4. Effluent o -Xylene Concentrations

Conc. (ppm)

35

 

0.6

0.5 -

0.4 ~—

0.3 ‘-

0.2 ~~

0.1 1-

 

    
 

 

~I~ Inactive Columns
+ Active Columns

 

 

 

flow interuption

 

I l I 1 l

 

 

I I I l l

20 40 60 80 100
# Pore Volumes

Figure 4.3.5. Effluent Naphthalene Concentrations

120

Conc. (ppm)

36

 

120

 

--I- Inactive Columns
'00 " + Active Columns

 

 

 

flow interuption

 

 

 

 

 

 

60 80 100
# Pore Volumes
Figure 4.3.6. Effluent COD

120

37

occurring.

4.4. Solid Phase Data.

Sacrificed soil columns provided the primary data of interest to this study. The “true”
starting contaminant levels were those determined after inoculation and initial flushing of
the inactive system. Sacrifice of one column from each system followed by solid phase
analysis provided the initial concentrations. In the active system, these levels were 1.44
mg/kg benzene, 10.77 mg/kg toluene, 10.88 mg/kg ethylbenzene, 6.83 mg/kg o-xylene,
and 3.25 mg/kg naphthalene. Conversion to chemical oxgyen demand (COD) resulted in
an initial concentration of 84.85 mg/kg. The inactive system levels were 0.95 mg/kg for
benzene, 9.08 mg/kg toluene, 11.87 mg/kg ethylbenzene, 7.26 mg/kg o-xylene, 3.2 mg/kg

naphthalene, and 97.11 mg/kg COD.

No trend was seen when incremental concentrations were plotted versus depth up through
the columns, therefore, interpretation of changes due to location could not be assessed.
For this reason, the total contaminant mass on each column served as single data points
during plotting of solid phase removal rates. Figures 4.4.1 - 6 are presented to compare
solid phase removal in either system and for each contaminant. Figures 4.4.7-12

represent a mass balance by comparing the liquid to solid removal levels.

Mass (mg/kg)

38

 

 

1 40 q -0- Inactive Columns
' -|- Active Columns

 

 

 

 

 

+

 

*0
in

.L l
l

 

 

 

0.00 i i i .
0.0 20.0 40.0 60.0 80.0 100.0
# Pore Volumes

Figure 4.4.1. Solid Phase Benzene Concentrations

120.0

12.00 ’

Mass (mg/kg)
N .e O~ 9° .5
8 8 8 8 8

8

39

 

 

-O- Inactive Columns
~l- Active Columns

 

 

 

1*
I

 

 

 

f

4
4

 

 

0.0 20.0 40.0 60.0 80.0 100.0
# Pore Volumes

Figure 4.4.2. Solid Phase Toluene Concentrations

120.0

12.00

Mass (mg/kg) __
a 9° .0
8 8 8

P
8

0.00

40

 

 

-O- Inactive Columns
—I-- Active Columns

 

 

 

 

 

 

 

0.0 20.0 40.0 60.0 80.0 100.0
# Pore Volumes

Figure 4.4.3. Solid Phase Ethylbenzene Concentrations

 

1

1 20.0

Mass (mg/kg)

 

8.00

7.00

6.00~ .

5.00 .

4.00 1

3.00 4i

2.00 -

1.00 «

 

 

+ Inactive Columns
-I- Active Columns

 

 

 

 

0.00

0.0

 

80.0 100.0

# Pore Volumes

Figure 4.4.4. Solid Phase 0 -Xylene Concentrations

 

 

42

 

3.50
3.00 ._
2.50 T
2.00 l.

1.50 at

Mass (mg/kg)

1.00 *—

0.50 L

 

J

 

+ Inactive Columns
~§~ Active Columns

 

 

 

 

 

l
H

 

 

0.00
0.0

20.0

40.0

60.0 80.0 100.0
# Pore Volumes

Figure 4.4.5. Solid Phase Naphthalene Concentrations

120.0

Conc. (mg/kg)

43

 

 

 

+ Inactive Columns
-l-- Active Columns

 

 

 

 

 

 

 

 

0.0 20.0 40.0 60.0
# Pore Volumes

Figure 4.4.6. Solid Phase con

120.0

 

 

 

 

 

 

 

 

 

 

400.00
Note: Liquid phase removals based on effluent concentrations.
350.00 - ~
300.00 -~
’63
:3) 250.00 +
“E 200 00 I +Inactive Liquid
m --I- Active Liquid
§ 150.00 ,_ +Active Solid
-*- Inactive Solid
100.00 ._
50.00 « -
0.00 j I i 3 i i i Q
0 20 40 60 80 100 120

# Pore Volumes

Figure 4.4.7 . Cumulative Mass of Benzene Removed

Mass (microg)

45

 

 

 

  

 

 

 

 

 

 

 

 

 

1200.00 .
Note: Liquid phase removals based on effluent concentrations.
1000.00 ~
800.00 -
+ Inactive Liquid
600.00 ’ ~a~ Active Liquid
+Active Solid
400.00 A +Inactive Solid
200‘00 ‘ b k—a—g—t—a i
V— x n I‘ n n
0.00 i a‘ fi‘ # i
O 20 40 60 80 100 120

# Pore Volumes

Figure 4.4.8. Cumulative Mass of Toluene Removed

 

 

 

600.00
+ Inactive Liquid
500.00 a r + Active Liquid
+ Active Solid
_ ~X~ Inactive Solid

 

 

 

 

Mass (rnrcrog)
.8 .“8’ .8 .8
8 8 8 8

 

1 1 I 1 1

Note: Liquid phase removals based on effluent concentrations.

 

 

j T l I I

8

0 20 40 60 80 100

# Pore Volumes

Figure 4.4.9. Cumulative Mass of Ethylbenzene Removed

120

47

 

 

 

 

 

 

 

 

 

350.00
+Inactive Liquid
300‘” ‘i +Activc Liquid
+Activc Solid
250.00 «L ~x~lnaerive Solid
A
or)
§ 200.00 ~
E5 150.00 -—
2 ' / ___ - -——_=‘
100.00 -- M
50.00 +~
Note: Liquid phase removals based on effluent concentrations.
0.00 fi‘ % i i i
0 20 40 60 80 100 120

Figure 4.4.10. Cumulative Mass of o -Xylene Removed

# Pore Volumes

48

 

 

 

 

 

 

 

 

 

90.00

80.00 ~~

70.00 ~-
83 60.00 «-
8
'5 50.00 J- [/1
E: 40.00 T +Inactive Liquid
3 ~I- Active Liquid
2 30'00 ‘" +Active Solid

20.00 “i- -*-Inactive Solid

10.00 ‘-

0 00 N ore: Liquid phase removals based on effluent concentrations.
0 20 40 60 80 100 120

# Pore Volumes

Figure 4.4.11. Cumulative Mass of Naphthalene Removed

49

 

7000.00

6000.00 +~

5000.00 4 -

8
3

3000.00 - ~

Mass (microg)

2000.00 . -

1000.00 fl

 

  
  

 

+ Inactive Liquid
"Ir- Active Liquid
+ Active Solid
—*- Inactive Solid

W! in

Note: Liquid phase removals based on effluent concentrations.

1 I I

 

 

 

 

 

 

 

0.00

T I I I I

20 40 60 80 100 120
# Pore Volumes

Figure 4.4.12. Cumulative Mass of COD Removed

50
Table 4.4.1 contains data generated to assess the validity of sample data. It essentially is

a comparison of estimated data and measured data to determine whether the results are
similar to those expected. This is useful to locate potential errors while also identifying
unexpected results which may be of particular interest. The table contains the same

information for each target compound including converted COD data.

The contaminant saturation levels are the concentrations fed during contaminant
saturation. These numbers represent the maximum possible concentration of each
compound loaded prior to flushing. The distribution coefficients (Kd) were estimated
during the batch sorption experiment as discussed in Appendix C. The maximum
possible contaminant mass on the soil was then found by multiplying the liquid levels by
the distribution coefficients and is represented as estimated mass on soil. Cumulative
mass removals for both active and inactive columns were calculated based on the effluent

concentrations measured during flushing. The measured mass on soil were determined

using headspace gas chromatography to analyze the column soil.

Table 4.4.1. Comparison of Expected and Actual Results

 

 

 

 

 

 

 

 

 

Contaminant Estimated Cumulative Mass Measured
Saturation Kd Mass on Removed in Effluent Mass on
Compound Level (mg/L) [(mg/kgyppm] Soil (ug*) Inactive Active Soil (ug*)
Benzene 32 1.11 639.36 376.39 320.96 17.10
Toluene 41 1.21 892.98 1 108.94 882.08 163.44
Ethylbenzene 10 2.06 370.80 538.41 389.94 213.66
o-Xylene 6 1.90 205.20 319.86 252.93 130.68
Naphthalene 1 3.53 63.54 93.22 69.31 57.60
COD 270 -- 65 15.64 7309.82 5744.15 1747.98

 

 

 

 

 

 

 

* ug = micrograms

 

CHAPTER 5

DISCUSSION, CONCLUSIONS, AND RECOMMENDATIONS

5.1. Discussion.

A number of interrelated issues arose during experimentation which merit discussion in
relation to the conclusions and understanding which can be gained from this research
effort. There are essentially six (6) areas into which these “issues” can be grouped. The
first is related to the solid phase data, while the others include generation and use of
oxygen data, microbial quantification, similarity of conditions among columns, soil
organic matter content, and the possible effect of oxygen on the sorption characteristics

of the soil.

The large difference between cumulative mass removal of contaminants when comparing
solid phase and liquid phase data indicates a loss in contaminant mass which has not been
accounted for (refer to Figures 4.4.7 - 12). One possibility could be low recovery of
contaminants from soil due to the type and extent of sorption. The effect of contaminant
age on recovery was, however, investigated as described in Appendix B. The results did
not indicate a strong contaminant age correlation with decreased recovery for the length
of this study. For this reason, loss during column sacrifice, physical partitioning of the

soil columns, and/or analysis must be considered.

51

52
One method used to identify the source for the error detected during performance of a

mass balance was described in the results section with the accompanying data shown in
Table 4.4.1. Comparing the mass removed in either the active or inactive columns (based
on effluent data) to the total contaminant mass on the soil (estimated from liquid data and
distribution coefficients) provides an indication of order of magnitude similarities. Since
there is some error associated both with the estimated distribution coefficients and
effluent concentrations, in addition to the conservative estimate of saturation levels
(probably lower due to losses), the values are fairly similar. On the other hand, looking at
the initial soil levels determined from analysis, the numbers are much less than the same
predicted values. Coupling this with the fact that the effluent data (Figures 4.3.1 - 6) is
similar for both systems but somewhat lower for the active columns as expected indicates

that the error is likely due to analysis of soil samples.

A comparison of liquid and solid removals for the increasingly volatile compounds
(benzene > toluene > o-xylene > ethylbenzene > naphthalene) show larger errors in mass
balance terms indicating possible loss due to volatilization. In addition, prior to analysis
of the final pair of columns, half of the samples in sealed headspace vials from each

column were sonicated for 1 hour at 50°C in an attempt to improve contaminant recovery

during analysis. The use of sonication had been previously tested and found to
insignificantly effect the results however, the possible effect may have become more
significant during analysis of lower concentrations which was the reason for a second

assessment. The results were somewhat varied but there was no trend indicating

53
improved recovery associated with sonication.

The oxygen issue can be sub-divided into (1) diffusion effects in the column system and
during sampling and (2) abiotic oxygen uptake. Figure 4.2.1 shows the influent and
effluent oxygen data for both series of columns. The effluent oxygen from the inactive
columns was somewhat higher than the influent due to the first issue of diffusion and
sampling effects. The effluent oxygen from the active columns was maintained at very
low levels which indicates a constant oxygen uptake. This could be due to abiotic effects
or microbiological activity however both would decrease over time. The average oxygen
uptake of 20 mg/L corresponds to a COD of approximately 4,400 mg/L when 2 moles of
oxygen is consumed per mole of COD degraded. This level of removal should be
represented by the difference in initial contaminant mass on the soil (6,515.64 mg/L COD
from Table 4.4.1) and cumulative mass removed from the active columns (5,744.15 mg/L
COD) however, it’s significantly higher. These comparisons indicate an error is
associated with using the collected oxygen data to estimate removal due to degradation as

intended.

The change in dissolved oxygen levels during column flushing due to atmospheric
diffusion was minimized through the use of gas-tight glass syringes and teflon-lined
stainless steel tubing. Some diffusion remained unavoidable due to the columns and the
connections, however, this level was estimated by flushing a sterilized ottawa sand

packed column with high levels of oxygen. The resulting effluent dissolved oxygen

54
levels were approximately 2 - 3 mg/L lower than the average influent levels of 15 mg

OzlL.

Biased results during sampling and analysis were addressed by using an increased
sampling rate to minimize the time for diffusion into the effluent collection syringes.
After collection, a needle was used to rapidly introduce the 2 mL sample into the bottom
of a customized oxygen vial designed to allow for reduced volume analysis. This
diffusion effect was also better understood by operating the last three inactive columns in
an anaerobic glove box. Influent and effluent oxygen levels were the same and constant
at approximately 0.8 mg OzlL. It is believed that there was in fact “no” oxygen in the
nitrogen flushed solution therefore the 0.8 was introduced during transfer from the
columns to the vial for measurement. The effluent levels of 2 - 3 mg OzlL minus 0.8 mg

02/L would therefore be due to diffusion into the columns during operation.

Abiotic oxygen uptake was quantified by flushing a non-contaminated sterilized packed
soil column with highly oxygenated water and measuring the effluent dissolved oxygen
levels. Significant oxygen uptake was observed, however, the study did not continue
long enough to understand the maximum abiotic uptake levels. It is understood that these
losses will reduce the consumption by microorganisms, however, it would be suprising if
this abiotic consumption is a continuous oxygen sink therefore it does not serve as an

explanation for the large and continuous oxygen uptake.

55
Quantification of microbial numbers was problematic in some respects. Initially a protein

assay was performed and intended to allow for conversion to microbial concentration at
different locations within the columns. Humic acids present in the soil clouded the
results of this procedure. The soil biomass plating procedure was then considered,
however, due to freezing of sacrificed columns (to minimize volatilization losses of
contaminants) prior to headspace analysis, the organisms would suffer. Effluent plating

was not found to be representative of soil microbial levels, therefore, was also ruled out.

Quantification of soil microorganism numbers occurred at a single point in time through
the sacrifice of two (2) columns solely for the purpose of microbial analysis. While this
provided some indication of the comparative concentrations, the point of measurement
may not correspond to the steady-state microbial levels due to decreasing substrate
availability. In addition, use of bacto-agar plates in growth chambers did not identify the
number of organisms present which were capable of degrading the target compounds.
For this reason, a most probable number (MPN) procedure is recommended using column
effluent to quantify the microbial numbers capable of degrading BTEXN. Altemately,
soil biomass plating using high purity agarose media in growth chambers could have

provided improved results.

Similarity in conditions from column to column is an important factor since each column
represents a point in the treatment series. Numerous sources of variability were

introduced due to non-uniformity of column packing, contamination, microbial

56
distribution, and flow. Considering these effects, the deviation in results is not suprising

or extensive.

Use of a higher organic matter content soil may have been useful to increase the
sorption/desorption effect. Desorption from the soil used in this study appeared to be
rapid resulting in a relatively short experimental run time. While being representative of

groundwater environments, the time frame may have affected the results.

The finding that sodium sulfite effects soil sorption characteristics was suprising and may
indicate that the change in oxygen level actually produced the effect. If this is true, the
results would be highly dependent on the vast difference in influent oxygen levels in

either column system which warrants further investigation.

5.2. Conclusions.

Using the results summarized in the previous chapter, the following conclusions can be

drawn.

1. The microbial conditions within each series of columns were
significantly different based on the dissolved oxygen data and
microbial plating results.

2. Non-equilibrium partitioning occurred and indicates the desorption
limitation was achieved.

3. Both column systems approached similar liquid remediation levels
rapidly, indicating that microorganisms did not offer a significant
advantage. The somewhat higher removal in the active columns is
potentially due to an increased concentration gradient.

57

4. Mass balance calculations indicate substantial loss of contaminants
prior to or during solid phase analysis.

5. There was a poor correlation between oxygen data and substrate
removal due to biodegradation.

6. Methods used to quantify solid phase microbiological activity levels
were problematic.

5.3. Recommendations.

Due to the preliminary nature of this study, there are a broad range of areas into which
further effort could be invested. The comparative nature offers decided advantages when
attempting to isolate system differences. For this reason, addressing some of the focus
issues introduced here could maximize the effectiveness of this type of experiment in

future study.

The recommendations can be divided into four (4) areas. The first is further study to
isolate the potential reasons for soil analysis error. This can be subdivided into three
studies to understand whether contaminant age effects, contaminant recovery, or losses

during column sacrifice and preparation were responsible for low headspace responses.

The second recommendation is related to the type of column system used. A series of
small packed soil columns connected by valves to allow influent and effluent sampling to
be directly associated with the solid slug is recommended. In addition, glass columns

with stainless steel fittings would minimize losses due to diffusion and sorption.

58
The “inactive” system could also be operated and sampled in a nitrogen atmosphere to

minimize atmospheric oxygen bias. Gamma-irradiation after the columns have been
packed could also decrease the possibility of microbial contamination. Oxygen levels in
the inactive column flushing solution could also be decreased closer to zero through the

use of zero-grade nitrogen.

The third recommendation is to develop a more appropriate soil biomass quantification
procedure or method. This is necessary to allow a better correlation between removal rate
and relationship to microbiological activity level to be developed. The final
recommendation is for additional study to address the effect of oxygen level on sorption,

abiotic soil uptake, and contaminant loss during soil analysis.

APPENDICES

APPENDIX A

CONTAMINANT RECOVERY FROM SOIL

In order to determine what percentage of each contaminant was recoverable from soil
samples during analysis by headspace gas chromatography a recovery study was
performed. This study consisted of analyzing replicate samples with and without soil
spiked with varying levels of low concentration standard solution. This standard solution
contained 226.24 mg/L benzene, 248.48 mg/L toluene, 138.4 mg/L ethylbenzene, 93.04
mg/L o-xylene, and 24.088 mg/I. naphthalene. Additionally, the effect of variation in soil

mass on analysis results was investigated.

11.]. Contaminant Recovery.

Soil samples were prepared using the standard procedure (as described elsewhere) to
contaminate five (5) replicate 1.0 g soil samples to the same level using the BTEXN
standard solution. Samples containing this same level of contamination spiked directly
into empty headspace vials were also prepared. Both soil containing and soil free
samples were then shaken, equilibrated, and chilled prior to addition of 5 mL chilled
deionized water. Preparation of these samples in a similar manner was necessary to
maintain similar loss conditions and prevent biased results. Check standards were also

prepared by spiking contaminant solution directly into 5 mL deionized water. All of the

59

60
above samples were then analyzed by headspace gas chromatography.

The difference between the samples with and without soil allowed contaminant recovery
to be determined. Naphthalene recovery was based on a comparison between the check
standard response and soil response due to loss of naphthalene in the soil free systems
due to solidification onto the headspace vials. Table A.1.1. provides a summary of these
results which indicate that a significant portion of the contamination present is available
for analysis. Recovery based on varying levels of contamination was also tested and

found to have an insignificant effect.

Table A.1.l. Contaminant Recovery from Spiked Soil.

 

 

 

Contaminant Recovery (%)
._B.enzene Toluene___Ethxlhenzene___a_Xrlan_- _blanhthalsme_
96.0 96.7 96.3 95.3 68.4

 

 

 

 

 

 

 

A.2. Efiect of Soil Mass.

The effect of soil mass on analysis results was investigated by spiking similar levels of
contaminant solution onto soil samples which varied in mass by 0.2 g in the range from
0.4 g to 1.6 g. The results of this study are presented in Table A.2.1. The deviations in
GC response over this entire range (60% variation in mass) for each component were not
significant, therefore, soil mass was not incorporated into the analysis procedure except

for use during calculation of concentration per mass of soil analyzed.

61

Table A.2.1. Effect of Soil Mass on GC Response.

 

fizz), " I‘ll-5.:u”z'...;\~'\ ,ggyfif‘» . \-
-‘.-1.., 51.1.: \:

 

 

 

 

 

 

 

 

 

 

 

 

\ GC Response

Soil Mass, g Benzene Toluene Ethylbenzene o-Xylene Naphthalene
0.4 398128.64 469834.54 252575.16 172305.93 27055.44
0.6 433103.40 516573.74 286552.79 195810.98 27285.79
0.8 434138.04 520757.97 279001.10 193205.33 27419.72
1.0 450783.56 554753.85 311977.41 213143.30 25725.27
1.2 443538.94 546382.41 307157.29 210835.05 25025.01
1.4 447359.65 556838.77 315531.35 213661.53 23682.55
1.6 412723.38 521173.29 304083.42 208636.37 25122.53

Average 431397 526616 293840 201085 25902
Deviation 19339 30235 22575 15132 1407
% Deviation 4.5 5.7 7.7 7.5 5.4

 

 

 

 

 

 

 

APPENDIX B

EFFECT OF CONTAMIN ANT AGE ON RECOVERY

Research has shown that the recovery of contaminants from soil decreases with time due
to interparticle and irreversible sorption (Farrell and Reinhard, 1994; Steinberg et al.,
1987). To quantify this effect for the contaminants and soil used in this research a

recovery study was designed and implemented.

Neutral, clear-glass ampoules (OIC No. 130021) were used to create a sealed
environment to allow the effect of decreased recovery to be assessed. Six (6) series of six
(6) ampoules were used including triplicate samples and triplicate controls for each
series. All were prepared at the same time and analyzed to coordinate with column

sacrifice times to allow a direct comparison to be made.

The samples consisted of 1 g of gamma-irradiated soil with the remaining volume
containing contaminant reservoir solution. The approximate concentrations of each target
compound were 28 mg/L (benzene), 34 mg/L (toluene), 9 mg/L (ethylbenzene), 5.5 mg/L
(o-xylene), and 0.9 mg/L (naphthalene). The controls contained this same solution
without soil. All ampoules were sealed using a purging and sealing unit (01 Corporation

Model 524PS) with minimal headspace followed by periodic shaking to ensure

62

63
equilibrium sorption occurred. At the time of sacrifice the soil containing samples were

frozen to allow the soil to be removed with minimal volatilization losses prior to
preparation and analysis. The controls were sampled using a glass syringe to remove

triplicate 1.0 mL samples after the glass top was cut away from the ampoule body.

The data generated was used to calculate the solid phase contaminant mass of each
component and the accompanying change over time. The change with time will
presumably be due to decreased recovery from the soil which allows a recovery with time
curve to be generated. This information is useful in and of itself to quantify an effect
which previously hasn’t been quantified. In addition, associating the ampoule recovery
study with the analysis of the packed soil columns may allow the effect of decreased

recovery to be taken into consideration upon data reduction and use.

Table B.1. Estimated Kd Values Based on Contaminant Age.

 

 

 

 

 

 

 

Kd [(mg/kngpm]
# Days Benzene Toluene Ethylbenzene o-Xylenc Naphthalene
5 .46 .6 .86 .85 1.81
10 1.06 1.32 1.8 1.73 2.93
12 1.29 1.52 1.97 1.85 2.79
18 .78 .99 1.47 1.39 2.41
27 1.47 1.76 2.40 2.28 3.55
27 1.21 1.48 1.99 1.91 3.25

 

 

 

 

 

 

 

 

 

The results presented in Table 8.1. contain significant variation, however, the values
estimated from the last two days were significantly similar and high indicating no

decreased recovery.

APPENDIX C

BATCH SORPT ION EXPERIMENT

Estimation of solid phase contaminant levels based on effluent analysis is important to
determine column sacrifice intervals. To obtain a reasonable estimate, it is necessary to
quantify the partitioning coefficient (Kd) of each contaminant. For this reason, a batch
sorption experiment was performed to obtain these values while also assessing whether
different concentrations affect the results. In addition, thimerosal and sodium sulfite use
was investigated to observe whether changes in equilibrium sorption characteristics of the

soil occurred.

Gamma-irradiated soil was used due to the potential change this procedure may have on
the structure of the soil organic matter which in turn effects the sorption characteristics.
The small quantity of this soil available dictated the use of a small system. Disposable 5
mL (volume is actually greater than this) glass centrifuge tubes with screw cap finish
(Baxter C3795-5) in conjunction with screw thread vial caps (Baxter C4810-51) and
silicone rubber septa (Baxter C4810-58) constituted the batch system. Approximately
five (5) mL of liquid and 2 g of soil were used with thirty, 2 mm glass beads used to take
up the excess volume to prevent the presence of headspace thereby avoiding gas-phase
partitioning. These values (volume = V, mass of soil = M,) were used to estimate the

initial contaminant levels based on chosen equilibrium concentrations (C.) and

64

65
partitioning coefficients (Kd) roughly estimated previously.

Three initial levels of contaminant (Co) were used to assess whether sorption is linear. A
mid-range concentration was representative of most values while a value 20% lower
provided an idea of K, changes at lower levels. High concentrations representative of
influent values allowed K, in this range to be found and compared. The following table

(C. 1) identifies the specific concentrations that were used.

Table C.l. Parameters used to Design Batch Sorption System.

 

 

 

 

 

 

 

 

Benzene Toluene Ethylbenzene o-Xylene Naphthalene
NM [(mg/kg)/ppm] 0.996 1.129 1.53 1 1.4065 1.726
Low Co (ppm) 0.257 0.794 0.431 0.28 0.149
Concentration Ce (ppm) 0.2 0.6 0.3 0.2 0.1
Mid- C0 (ppm) 1.285 3.968 2.156 1.402 0.747
Concentration Ce (ppm) 1 3 1.5 1 0.5
High Co (ppm) 41.106 55.548 15.812 8.411 1.493
Concentration Ce (ppm) 32 42 1 1 6 1

 

 

 

 

 

 

 

 

 

The initial concentration levels (Co) were prepared in 160 mL serum vials by spiking
concentrated BTEXN solution (0.3, 1.5, and 20.5 11L) into 150 mL filtered DI. water plus
approximately 580 mg/L HCl to prevent microbial degradation of the constituents. Glass
beads were added and the solutions mixed overnight on a rotary shaker, to insure

dissolution and uniform distribution, prior to transferring 5 mL into the respective vials.

Three time periods (7, l4, and 21 days) were used to ensure equilibrium was reached. A

66
total of 33 vials were used as follows: three sets of three controls containing no soil

tumbled for 7, 14, and 21 days, six high and mid-range vials in each period, three low
concentration vials in 7-day period, and three vials containing sodium sulfite (Na2803) in
the 7-day period. The N a2803 was added to the reservoir in sufficient quantity (0.03 g) to
consume the oxygen present in the water and leave a residual level of approximately 100

mg/L. Thimerosal was also added to three vials to create a concentration of 200 mg/L.

The initial concentrations were measured in triplicate by sampling the serum vial
reservoirs using glass syringes (triplicate 1 mL samples). Equilibrium concentrations
were measured after centrifuging the batch vials and sampling the supernatant (triplicate
0.2 mL samples). Centrifugation occurred for 10 minutes at 3200 rpm using a six-place
angle rotor Adams Compact 11 Centrifuge (Fisher, 05-100-90). Soil samples were also
analyzed by freezing and removing the soil at the end of the experiment followed by

preparation and analysis as described elsewhere.

Thimng Effects

The effect of thimerosal on sorption characteristics was significant for each compound
excluding naphthalene. Table C.2. provides a comparison of K, values for each

contaminant in systems which either contain or don’t contain thimerosal.

67
Table C.2. Thimerosal Effect on K...

Benzene Toluene lbenzene o-X

Without Thimerosal 0.9960 1. 1290 1.53 10 1.4065 1.7260
% Difference 36.4 31.3 23.8 23.1 2.3

 

Sodium Sulfite Effects
Sodium sulfite was also found to have an effect on sorption of the target compounds

therefore was not used. The following table (C.3.) summarizes the data to support this

 

 

 

 

 

 

 

 

 

 

 

 

 

conclusion.
Table C.3. Sodium Sulfite Effect on Kd.
Contaminant Concentration on Soil (mg/kg)

42“.”, Benzene Toluene Ethylbenzene o-Xylene Naphthalene

With Na2801 1.84 1.81 0.71 0.31 0.16

Without Na2503 2.16 2.83 1.14 0.66 0.16
% Difference 14.8 36.0 37.7 53.0 0.0
Estimati n

Low, mid and high concentration vials produced similar results such that a single value of
K. for each contaminant was determined and used regardless of contaminant
concentration. Due to the quantity of mid-concentration sorption vials these were then
used to calculate Kd values. These values were estimated in two ways; using only liquid
phase data to compare the difference between soil containing and soil free systems or

using a combination of liquid phase and solid phase data collected from the soil

68
containing systems. The variability in results using only liquid phase data was significant

and may be due to losses within the system. For this reason, the partition coefficients
were determined based on the combination of liquid and solid phase data. Table A.3.4.

below summarizes these values and provides a comparison for each equilibration period.

Table C.4. Partition Coefficients Based on Batch Sorption Experiment.

 

 

 

 

 

 

 

 

Equilibration Kd [(mfigyppm]
Period Benzene Toluene Ethylbenzene o-Xylene Naphthalene
7-Day 1.1 1 1.09 2.07 1.89 4.00
14-Day 1.09 1.19 2.02 1.91 4.00
21-Day 1.14 1.34 2.09 1.90 2.60
Average 1.11 :1: 0.03 1.21 d: 0.13 2.06 :1: 0.04 1.90 :1: 0.01 3.53 :i: 0.81
% Deviation 2.26 10.43 1.75 0.53 22.88

 

 

 

 

 

 

 

 

LIST OF REFERENCES

LIST OF REFERENCES

Abriola, L.M. and Chen, Y.M. (1993) A Mathematical Model of BTX Biodegradation and
Transport: Comparisons with Experimental Measurements. CoBioRem Conference. March
1-2. Lansing, MI.

Alvarez-Cohen, L., McCarty, P.L., and Roberts, P.V. (1993) Sorption of Trichloroethylene
onto a Zeolite Accompanied by Methanotrophic Biotransforrnation. Environ. Sci. Technol.
27(10), 2141-2148.

Angley, J .T., Brusseau, ML, Miller, WL., and Delfmo, J .J . (1992) Nonequilibrium Sorption
and Aerobic Biodegradation of Dissolved Alkylbenzenes During Transport in Aquifer
Material: Column Experiments and Evaluation of a Coupled-Process Model. Environ. Sci.
Technol. 26(7), 1404-1410.

Boyd, SA. and Guerin, W.F. (1992) Differential Bioavailability of Soil-Sorbed Naphthalene
to Two Different Bacterial Species. Appl. Environ. Microbiol. 58(4), 1142-1152.

Criddle, C.S., Alvarez, L.M., and McCarty, PL. (1991) Microbial Processes in Porous
Media, In Transport Processes in Porous Media, Ed. J. Bear and Y. Corapcioglu, NATO
Advanced Studies Institute Series B: Applied Sciences Vol. 202, Kluwer Academic
Publishing Co., Dordrecht, The Netherlands, pp. 641-691.

 

Dec, J. and Bollag, J-M. (1988) Microbial Release and Degradation of Catechol and
Chlorophenols Bound to Synthetic Humic Acid. Soil Sci. Soc. Am. J. 52, 1366-1371.

Fanell, J. and Reinhard, M. (1994) Desorption of Halogenated Organics from Model Solids,
Sediments, and Soil Under Unsaturated Conditions. 1. Isotherms. Environ. Sci. Technol.
28(1), 53-62.

Fanell, J. and Reinhard, M. (1994) Desorption of Halogenated Organics from Model Solids,
Sediments, and Soil under Unsaturated Conditions. 2. Kinetics. Environ. Sci. Technol. 28(1),
63-72.

Gordon, AS. and Millero, R]. (1985) Adsorption Mediated Decrease in the Biodegradation
Rate of Organic Compounds. Microbial Ecology. 11, 289-298.

Hayden, N .J . (1992) Retention, Enhanced Volatilization, and leaching of Gasoline in
Unsaturated Soils. PhD Dissertation. Michigan State University.

69

70

Hu, C-Z, Korus, R.A., Levinson, WE, and Crawford, R.L. (1994) Adsorption and
Biodegradation of Pentachlorophenol by Polyurethane-Immobilized Flavobacterium
Environ. Sci. Technol. 28(3), 491-496.

Larson, R]. and Vashon, RD. (1983) Adsorption and Biodegradation of Cationic
Surfactants in Laboratory and Environmental Systems. Dev. Ind. Microbiol. 24, 425-434.

Marquis, S.A. Jr. and Dineen, D. (1994) Comparison Between Pump and Treat,
Biorestoration, and Biorestoration/Pump and Treat Combined: lessons from Computer
Modeling. Ground Water Monitoring Review. 105-119.

Oglam, A.V., Jessup, R.E., Ou, L.T., and Rao, P.S.C. (1985) Effects of Sorption on
Biological Degradation Rates of 2,4-Dichlorophenoxyacetic Acid in Soils. Appl. Environ.
Microbiol 49(3), 582-587.

Ridgway, HF, Safarik, J., Phipps, D., Carl, P. and Clark, D. (1990) Identification and
Catabolic Activity of Well-Derived Gasoline-Degrading Bacteria From a Contaminated
Aquifer. Appl. Environ. Microbiol. 56(11), 3565-3575.

Rijnaarts, H.H.M., Bachman, A., Jumelet, J.C., and Zehnder, AJB. (1990) Effect of
Desorption and Interparticle Mass Transfer on the Aerobic Biomineralization of a-
Hexachlorocyclohexane in a Contaminated Calcareous Soil. Environ. Sci. Technol. 24(9),
1349-1354.

Robinson, K.G., Farmer, W.S., and Novak, J.T. (1990) Availability of Sorbed Toluene in
Soils for Biodegradation by Acclimated Bacteria. Wat. Res. 24(3), 345-350.

Scott, H.D., Wolf, DC, and Lavy, TL. (1983) Adsorption and Degradation of Phenol at
Low Concentrations in Soil. J. Environ. Qual. 12(1), 91-95.

Scow, KM. and Hutson, J. (1992) Effect of Diffusion and Sorption on the Kinetics of
Biodegradation: Theoretical Considerations. Soil Sci. Soc. Am. J. 56(1), 119-127.

Scow, KM. (1993) Effect of Sorption-Desorption and Diffusion Processes on the Kinetics of

Biodegradation of Organics Chemicals in Soil. Soil Science Socieg of America Spggial
Publication No. 32. Madison, WI.

Shimp, RJ. and Young, R.L. (1988) Availability of Organic Chemicals for Biodegradation
in Settled Bottom Sediments. Ecotoxicol. Environ. Safety. 15, 31-45.

Smith, S.C., Ainsworth, C.C., Traina, S.J., and Hicks, RJ. (1992) Effect of Sorption on the
Biodegradation of Quinoline. Soil Sci. Soc. Am. J. 56(3), 737-746.

71

Steinberg, S.M., Pignatello, J.J., and Sawhney, BL. (1987) Persistence of 1,2-
Dibromoethane in Soils: Entrapment in Intraparticle Micropores. Environ. Sci. Technol.
21(12), 1201-1207.

van Genuchten, M. Th. and Wagenet, R]. (1989) Two-Site/Two-Region Models for
Pesticide Transport and Degradation: Theoretical Development and Analytical Solutions.
Soil Sci. Soc. Am. J. 53(5). 1303-1310.

van Loosdrecht, M.C.M., Lyklema, J ., Norde, W., and Zehnder, AJ.B. (1990) Influences of
Interfaces on Microbial Activity. Microbial. Rev. 54(1), 75-87.

Voice, T.C. and Kolb, B. (1994) Comparison of European and American Techniques for the
Analysis of Volatile Aromatic Compounds in Environmental Matrices. 1994 Pittsburgh
Conference, February 27 - March 4, Chicago, IL.

Voice, T.C. and Wallace, RB. (1994) Biodegradation Kinetics, Oxygen Transfer, and
Bioavailability in Bioventing and Related Technologies (portion of proposal: Objective 3).

Voice, T.C. (1993) Fundamentals of Petroleum Spill Chemistry. CoBioRem Conference.
March 1-2. Lansing, MI.

Voice, T.C., Pak, D., Zhao, X., Shi, J ., and Hickey, RF. (1992) Biological Activated Carbon
in Fluidized bed Reactors for the Treatment of Groundwater Contaminated with Volatile
Aromatic Hydrocarbons. Wat. Res. 26(10), 1389-1401.

Weissenfels, W.D., Klewer, H-J, and Langhoff, J. (1992) Adsorption of Polycyclic Aromatic
Hydrocarbons (PAHs) by Soil Particles: Influence on Biodegradability and Biotoxicity.
Appl. Microbial Biatechml 36, 689-696.

Wszolek, PC. and Alexander, M. (1979) Effect of Desorption Rate on the Biodegradation of
n-Alkylamines Bound to Clay. J. Agric. F aad Chem. 27(2), 410-414.

"lullinn“