The BAT groundwater monitoring system in contaminant studies

MISSING IMAGE

Material Information

Title:
The BAT groundwater monitoring system in contaminant studies
Physical Description:
xviii, 256 leaves : ill., photos. ; 29 cm.
Language:
English
Creator:
Mines, Barry Shaun, 1962-
Publication Date:

Subjects

Subjects / Keywords:
Civil Engineering thesis Ph. D
Dissertations, Academic -- Civil Engineering -- UF
Genre:
bibliography   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1992.
Bibliography:
Includes bibliographical references (leaves 244-254)
Statement of Responsibility:
by Barry Shaun Mines.
General Note:
Typescript.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 001800847
oclc - 27669796
notis - AJM4605
System ID:
AA00004744:00001

Full Text












THE BAT GROUNDWATER MONITORING
SYSTEM IN CONTAMINANT STUDIES



















By

BARRY SHAUN MINES


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA


1992


































Copyright 1992

by

Barry Shaun Mines



































This dissertation is dedicated to my loving wife, Wendy,

who put many of her dreams on hold so that I might achieve one

of mine.















ACKNOWLEDGEMENTS


There are numerous individuals to whom I must express my

gratitude for their help and guidance in my academic

progression. It is definitely true that no man is an island

as I have been significantly influenced by those around me.

Dr. Bloomquist has served as a technical consultant. By

introducing high technology materials he has taught me to be

innovative and to think like an inventor. He has also given

me insight into many scientific principles.

Dr. Davidson has stressed to me the fundamentals of soil

behavior. He has listened with keen interest to my progress

reports after each field and laboratory test and has advised

me well. My gratitude to him for reading and editing my

research proposals and dissertation drafts are immeasurable.

Dr. Townsend has guided my professional development in

several ways--first, by giving me numerous scientific articles

dealing with my research which always seemed to come at an

appropriate time. His teaching methods instilled in his

students' confidence in their ability to design. Of all the

teachers I have had he has stressed the practical aspects of

design engineering, ensuring that students evaluate all

possible concerns to come up with the most effective and

economical designs. He has won me over to the teaching

iv









profession through his sincere involvement in the students'

development. He took groups to numerous construction sites to

see engineering works firsthand and to seminars and

conferences on new and innovative engineering techniques. As

part of his involvement with the Engineers' Fair, he had

students build reinforced earth walls with cardboard and paper

strips, giving them the opportunity to work as a group on a

engineering project as done in industry.

Charlie Schmidt enabled me to become a pseudo-chemist.

He taught me the theory of mass spectrometry and allowed me

hands-on use of the Gas Chromatograph/Mass Spectrometer for

chemical analyses. He honed my skills as a chemist and gave

me insight into the properties of hydrocarbons and chemical

solvents.

Dr. Tom Stauffer provided technical assistance and

arranged the financial support for this study. I am very

appreciative of his promptness in reading my drafts and

supplying helpful testing suggestions along with his hands on

approach of going to the site to view the field testing.

I am deeply appreciative of Dr. Don Myhre's sincere

interest in my project. He has always taken the time to

listen to the progress of my research and to make helpful

suggestions. From his soil chemistry class, he has stimulated

my interest in the sorption capacities of soils and the

partitioning of hydrophobic organic compounds.









Dr. Thompson and his staff have been most helpful in all

administrative affairs, from working with the Air Force to

obtain my funding, to ordering required equipment for my

research. Dr. Thompson has always had time to discuss my

research progress and provide liaison help with the Air Force.

His seminar course has been on target with various speakers on

current engineering issues. He has helped me to be a better

engineer/manager by showing me that it is not enough to be a

sound technical engineer; one must also be an informed citizen

and be aware of changing current events and how they affect

civil engineering.

My sincere appreciation is due to Ed Dobbson. Ed

accompanied me on many BAT and electric cone penetration

tests. He instructed me well on the use of the penetrometer

rig to the point where I could manage on my own. Ed would

always check to see when I would like to do more testing and

was always most helpful. I could always count on Ed to help

me fix any mechanical equipment with which I had problems.

Ed, along with Dr. Bloomquist, was great at thinking of

alternative ways to obtain groundwater samples.

I am grateful to the Handex Company, Geosolutions Inc.,

and the Alachua County Department of Environmental Control for

letting me use their contaminated sites for this study.

To my brother, Dr. Richard Mines, I am grateful for his

interest in my research. As a civil-environmental engineer,









he was responsible for introducing me to this area of study

and for helping me develop into an engineer.

Richard and Dreama Mines, my parents, were responsible

for instilling a high regard for education in my life. From

my mother I have inherited an unquenchable desire for reading,

which, along with my father's high level of motivation and

determination, has allowed me to succeed in life. They taught

me that if you don't push yourself in life, you will never

know your limits.

I am most grateful to my wife, Wendy. She made several

edits to my dissertation and allowed me to work numerous

nights and weekends, sacrificing much of our time together.

She was always there to help me when testing did not go well

or equipment did not arrive.


vii
















TABLE OF CONTENTS
ACKNOWLEDGEMENTS . . .. iv

LIST OF TABLES . .. xii

LIST OF FIGURES .. . xiv

ABSTRACT . . ... xvi

CHAPTERS

1 INTRODUCTION. . 1

1.1 Problem Statement . 1
1.2 Objectives .. . 3

2 REVIEW OF THE LITERATURE . 5

2.1 Introduction . 5
2.2 Direct Push Technology . 6
2.2.1 Hydropunch . 6
2.2.2 BAT Probe . 9
2.3 DPT Field Studies .. . .. 10
2.4 Sample Preservation. . .. 16
2.5 Necessity for Current Study .. 17

3 LAB PERMEABILITY TESTING . .. 20

3.1 Permeability Limitations of the BAT System 20
3.2 Lab Permeability of Soils .. 22

4 BAT GROUNDWATER MONITORING SYSTEM LAB STUDIES 30

4.1 Introduction . .. 30
4.2 Monitoring Well Model . .. 31
4.3 Experiment 1 Inside Model Monitoring Well 33
4.4 Experiment 2 Inside the Model Monitoring Well
. . 38
4.5 Experiment 3 Inside the Model Monitoring Well
. . 41
4.6 Experiment 4--Sampling Within Tank Spiked
Water . . .. 46
4.7 Experiment 5--Sampling Within Tank Spiked
Water . ... 51
4.8 Experiment 6--Sampling Within Tank Spiked
Water . ... 52

viii










4.9 Teflon Ball and Tube Sampling Apparatus 54
4.10 BAT Vadose Zone Probe Testing 56
4.11 Summary. . . 58

5 ANALYSIS, TRANSPORT, AND PROPERTIES OF VOLATILE ORGANIC
COMPOUNDS .. 61

5.1 Introduction . ... 61
5.2 Organic Compounds . 61
5.3 Chemical Analysis . .... 64
5.4 Regulatory Contaminant Levels .. 70
5.5 Solute Transport . 71

6 FIELD STUDIES--CAVALIER PRODUCTS BUILDING SITE 76

6.1 Introduction . .. 76
6.2 Field Test 1 . 77
6.3 Field Test 2 . ... 79
6.4 BAT Sampling Within MW-17 . 84
6.5 CPT Testing . .. 86
6.6 BAT Sampling Within MW-15 .. 87
6.7 BAT Field Test 5 . ... 90
6.8 Summary . . 93

7 FIELD STUDIES--TEXTILE TOWN . .. 94

7.1 Introduction . .. 94
7.2 BAT Test 1 . . 95
7.3 BAT Test 2 . 98
7.4 BAT Testing at MW-7 . .. 100
7.5 Vertical Contamination Profile at MW-7 103
7.6 Vertical Profile Testing at MW-11 .. 104
7.7 BAT Probe testing in MW-11 107
7.8 Plume Chasing . ... 109
7.9 CPT Testing . ... .111
7.10 Summary . . 112

8 BAT MODIFICATION TESTING .. ... 114

8.1 Vacuum Pump Test 1 . ... 114
8.2 Vacuum Pump Test 2 . .. 118
8.3 Field Testing of Drill Rods as a Monitoring
Well . . 122
8.4 Summary . . 124

9 CONCLUSIONS AND RECOMMENDATIONS .. 127

9.1 Conclusions . 127
9.2 Recommendations for Future Testing 130
9.3 Advantages of the BAT Groundwater Monitoring
System . . 132









9.4 Disadvantages of the BAT Groundwater Monitoring
System . . 133

APPENDICES

A PERMEABILITY DATA . ... 135

A.1 Needle Permeability Without Filter 136
A.2 Permeability of Needle and HDPE Filter 141
A.3 Permeability of Needle and Steel Filter 146
A.4 Permeability of Kaolin-Sand Mixture 158
A.5 Permeability of Yellow Fine Mortar Sand 163
A.6 Permeability of a Uniform Sand ... 169
A.7 Permeability Data from Lynch Park .. 175
A.8 Sieve Analysis Data . ... 177
A.9 Constant Head Permeability Data ..... 178
A.10 Falling Head Permeability Test ... 179
A.11 Atterberg Limits . .. 179
A.12 Derivation of Formulas for BAT Permeability
Calculations . .. 180

B ORGANIC CHEMICAL DATA . ... .187

C BAT FIELD SAMPLING LOG . ... .190

C.1 Cavalier Site . .. 190
C.2 Textile Town Site . .. 192

D CHEMICAL ANALYSES DATA . ... .198

D.1 GC/MS Data for Cavalier Site .. 198
D.2 GC/MS Data for Textile Town Site .. 200

E DECONTAMINATION PROCEDURES .. ... 205

E.1 Bailer Decontamination ... 205
E.2 Decontamination of BAT Glass Sample Vials 205
E.3 BAT Probe Decontamination . 206
E.4 Decontamination of Enviro probe ... 207

F CONE PENETRATION DATA . ... .208

F.1 CPT Sounding at Lynch Park Adjacent to
MW-17 . . 208
F.2 CPT Sounding at Lynch Park Adjacent to
MW-15 . . 208
F.3 CPT Sounding at Textile Town Around MW-11 209
F.4 CPT Sounding at Textile Town Around MW-7 209

G HEADSPACE CORRECTIONS . .. 210









H OVERVIEW OF GROUNDWATER STRATEGIES


H.1 Groundwater Studies . .. 215
H.1.1 Planning . 215
H.1.2 Conventional Sampling Mechanisms .216
H.2 Monitoring Well Design Considerations 224
H.2.1 General ............. 224
H.2.2 Monitoring Well Size .. .224
H.2.3 Expected Contaminants 225
H.2.4 Water Table Depth .. 225
H.2.5 Screen and Casing Materials 226
H.3 Well Development . .. .229
H.4 Purging of Wells . 230
H.5 Sampling Studies . .. ..231
H.5.1 Lab Studies . .. 231
H.5.2 Field Studies . .. .235
H.6 Underground Storage Tank Regulatory Programs 238
H.6.1 Congressional Acts .. 238
H.6.2 EPA's Underground Storage Tank
Program .. . 239
H.6.3 Florida DER Programs .. .242

REFERENCE LIST. . .... 244

SUPPLEMENTAL BIBLIOGRAPHY . .. 251

BIOGRAPHICAL SKETCH . ... 255


215
















LIST OF TABLES


Table 3.1 Permeability Limitation Values of BAT
System . .. 23

Table 3.2 Permeability of Three Soils ... 29

Table 4.1 Chemical Analyses of Sampling Within Model
Monitoring Well . .. 39

Table 4.2 Chemical Analyses of Experiment Two Inside
the Model Monitoring Well ... 40

Table 4.3 Chemical Analyses of BAT Probe Sampling
Inside the Model Monitoring Well 44

Table 4.4 Chemical Analyses of Sampling Within Tank
Spiked Water . ... 50

Table 4.5 Chemical Analyses from Sampling Within Tank
Spiked Water . . 53

Table 4.6 Chemical Analyses of Experiment 6 .. 55

Table 4.7 Vadose Probe Testing . ... 59

Table 5.1 Henry's Constant for Selected VOCs ... 63

Table 5.2 Primary Drinking Water Standards (MCLs) 69

Table 5.3 Florida Ground Water Target Levels ... 71

Table 6.1 Chemical Analyses from MW-17 at Cavalier
Site . .... 81

Table 6.2 Chemical Analyses from MW-17 at Cavalier
Site . .... 83

Table 6.3 Chemical Analyses from BAT Sampling Within
MW-17 . . 86

Table 6.4 Chemical Analyses from BAT Sampling Within
MW-15 . . 89

Table 6.5 Chemical Analyses from Lab Insitu Class 92

xii










Table 7.1 Chemical Analyses from MW-17 at Textile Town 103

Table 7.2 Chemical Analyses of Vertical Sampling at
MW-7 . . 105

Table 7.3 Analyses from BAT Testing in MW-11 .. 110

Table 7.4 Chemical Analyses from Plume Chasing 112

Table 8.1 Chemical Analyses from MW-7 .. 119

Table 8.2 Total Times for Truck Set Up and Sampling 120

Table 8.3 Chemical Analyses from Vacuum Pump Apparatus 121

Table 8.4 Chemical Analyses from BAT Testing around
MW-7 . . 125

Table 9.1 Summary Comparison of BAT Versus BAiler
Recovery of VOCs ... 128

Table G.1 Henry's Law Constants for Selected Organic
Compounds ..... . 212


xiii


















LIST OF FIGURES


Figure 2.1 Hydropunch in Closed and Open Positions

Figure 2.2 BAT Enviroprobe in Closed and Open
Positions . .

Figure 3.1 Gradation Curve of Uniform White Fine Sand


S8


S 211

. 25


Figure 3.2 Gradation Curve of Fine Mortar Sand

Figure 4.1 Teflon Bailer . .

Figure 4.2 Model Monitoring Well Set-Up .

Figure 4.3 BAT MK2 Probe . .

Figure 4.4 Cascaded Sampling for Zero Head Space
Sample . .

Figure 4.5 BAT Probe with Reaction Beam .

Figure 4.6 Balloon and Test Tube Apparatus .

Figure 4.7 BAT Sampling in Nalgene Container .

Figure 4.8 Typical Soil Moisture Curve .

Figure 5.1 Transport of a Typical LNAPL .

Figure 5.2 Transport of a DNAPL . .

Figure 5.3 Typical Ion Chromatograph .

Figure 5.4 Typical Mass Spectra . .

Figure 6.1 Cavalier Site Plan . .


. 32

. 34

. 36


. 37

. 43

* 47

* 48

. 60

. 62

. 65

. 67

* 68

* 77


Figure 6.2 View of Cavalier Site--Lynch Park on Left

Figure 7.1 Textile Town Site Plan . .

Figure 7.2 View of Textile Town--Penetrometer Rig
near MW-7 . .


xiv









Figure 7.3 Cone over MW-11 . .

Figure 7.4 View of Stripping Tower for Remediation

Figure 7.5 Penetrometer Rig around MW-7 .

Figure 7.6 Vertical Contamination Profile at MW-7

Figure 7.7 Vertical Contamination Profile at MW-11

Figure 7.8 Lateral Plume Delineation .

Figure 8.1 Brass Adaptor for Vacuum Pump Testing

Figure G.1 Fraction Remaining C/Co versus V,/V. for
Headspace Related Errors for Selected
Aromatics. Compounds Apply to 20 C .

Figure G.2 Fraction Remaining C/Co versus V,/Vs for
Headspace Related Errors for Selected
Chlorinated Compounds. Compounds Apply
to 20 C . .

Figure H.1 Conventional Groundwater Sampling
Mechanisms . .


. 97

S 101

S 102

S 106

S 108

S 113

S 115



S 213




S 214


218















Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy

THE BAT GROUNDWATER MONITORING SYSTEM IN
CONTAMINANT STUDIES

By

Barry Shaun Mines

August 1992

Chairperson: John L. Davidson
Major Department: Civil Engineering

Groundwater contamination continues to be a serious threat to

our environment. One prevalent cause of such contamination is

leakage from underground storage tanks. Detection and

assessment of contamination must be made before remediation of

any site can begin. This is traditionally accomplished by

sampling from monitoring wells. The BAT Groundwater

Monitoring System is a recently developed device which can

collect samples of pore fluid without the need for a drilled

well.

An experimental study was conducted on the BAT System,

with the major objective of evaluating its effectiveness in

sampling volatile organic compounds (VOCs). Both large-scale

laboratory and field investigations were carried out. At many

locations BAT testing was compared to adjacent bailer sampling

from monitoring wells.


xvi









Neither the well bailer nor the BAT system consistently

recovered more VOCs, though concentrations recovered in most

cases were comparable. BAT samples recovered using a

stainless steel filter element consistently exhibited higher

concentrations of VOCs than did samples from probes with the

HDPE filter. Concentrations of volatile constituents in BAT

samples displayed a lower standard deviation than did samples

obtained using the bailer.

The effect of headspace in the BAT's sampling tubes was

investigated. VOCs measured in single tubes with small

amounts of headspace compared favorably with those in test

tubes obtained in a cascaded set-up which had no headspace.

The use of an inert material balloon within the sampling tube

to eliminate headspace and vacuum effects showed promise.

Modifications to the equipment were made to investigate the

possibility of drawing the much larger fluid samples required

for some chemical tests.

The BAT System can be used to estimate a formation's

coefficient of permeability. It was found that this

capability is limited, by the pore size of the filter element,

to determinations in very fine material, specifically silts

and clays.

The BAT test is a relatively rapid, inexpensive

penetration test which provides high quality and reproducible

pore fluid samples. It has the potential for use in the

assessment of contaminated sites, especially in the


xvii









delineation of leakage plumes and in siting of permanent

wells.


xviii















CHAPTER 1
INTRODUCTION



1.1 Problem Statement


Groundwater contamination continues to be a serious

threat to our environment at a time when the demand for water

resources has never been greater. Many states, including

Florida, obtain the majority of their water from groundwater

(aquifers). Contamination of this water can occur from

numerous sources. Some of the more prevalent are the

application of pesticides for agricultural uses, leakage of

fuel from storage tanks (above or below ground), leakage of

leachate from sanitary landfills or hazardous waste sites, and

spillage/leakage from industrial plants. For, example the

Florida Department of Environmental Regulation (Stuart, 1986)

estimates that 6000 of the state's 60,000 petroleum storage

tanks are leaking. Currently there are over 1.4 million

underground storage tanks in the United States that are

regulated by the Resource Conservation and Recovery ACT

(RCRA). Of these it is estimated that approximately 80% are

constructed of bare steel, which is easily corroded.

Determination of the presence of groundwater

contamination is typically performed by installing monitoring









2

wells for groundwater sampling. Alternative methods include

soil sampling for analysis, extraction of vapors from the

vadose zone, magnetic/resistivity surveys, and direct push

(or penetration) technology (DPT).

DPT makes use of a cone device which is pushed or

hammered into the ground. Some tips have sensors for

measuring the changes in the resistivity/conductivity of the

soil which can provide a measure of the total dissolved solids

through correlations. Others tips use fluorescence sensors to

detect contamination (Cooper and Malone, 1991). Fluorescence

sensors radiate ultraviolet light out into the soil where some

contaminants absorb the radiation and then fluoresce (emit

light). The emitted light is received by a fiber-optic link

which sends the information to an optical analzyer for

spectral analysis. Cooper and Malone (1991) state that these

sensors are only good for contamination concentrations greater

than 200 ppm. Another problem with this system is that not

all organic compounds fluoresce. Some that do are polynuclear

aromatic hydrocarbons (PAHs), phenol, toluene, and the

xylenes. Naturally occurring materials also fluoresce causing

interference.

A relatively new penetrometer, the BAT probe

(Torstensson, 1984), allows groundwater samples to be taken

without installation of costly monitoring wells. A porous

element in the probe's tip connects via a hypodermic needle

and septum with an evacuated sample tube, which is lowered











down the center of the push rods from the surface. However,

considerable doubt has existed in the regulatory arena as to

whether such a sampling device would cause a loss of VOCs due

to the initial vacuum in the sample tube. Some environmental

regulatory agencies are not willing to use the BAT until a

larger data base has been established on its ability to obtain

a representative sample. Peristaltic pumps used to obtain

groundwater samples from monitoring wells have generally been

ineffective for sampling VOCs because the reduced pressure

causes the VOCs to volatilze away.

The question to be answered is whether the BAT probe can

be penetrated into the ground and recover groundwater samples

which contain contamination levels comparable as that obtained

using a monitoring well and bailer.



1.2 Objectives

The purpose of this study is to provide an in-depth,

comprehensive study to compare results from the BAT probe and

and Teflon bailers from nearby monitoring wells. VOCs are

typically the most difficult contaminants to sample. The

research was performed by taking samples within a small radius

around monitoring wells at two leaking underground storage

tank sites and taking bailer samples from the monitoring

wells. BAT sampling will also be performed inside the

monitoring wells to ensure basically the same water is being

sampled.









4

A secondary purpose of the study is to evaluate the

repeatability of BAT probe testing, i.e., what is the

precision of the device? Does it have a high or a low

standard deviation compared to the Teflon bailer? Other

objectives of the research are to:



1. Evaluate the significance of headspace in BAT sample

vials and look at possible modifications to eliminate this

headspace.

2. Perform laboratory testing to determine the limitations

of the BAT probe for the determination of the coefficient of

permeability. Does the BAT probe give reasonable values of

the coefficient of permeability?

3. Examine modifications which could be made to the BAT

groundwater monitoring system to allow collection of larger

sample sizes.

4. Perform BAT testing to determine the extent of vertical

and horizontal contamination at a site.

5. Determine typical sampling times for the BAT probe along

with time estimates for setting up a penetrometer rig and

taking groundwater samples with the BAT probe.

6. Provide an overall evaluation of the BAT system, its

advantages, disadvantages, and uses in groundwater studies

(specifically contamination assessments).

7. Reommend future possible research with the BAT system.















CHAPTER 2
REVIEW OF THE LITERATURE


2.1 Introduction


The ability to characterize the quality of groundwater

adequately and economically is a major concern in all

contamination studies. In order to determine whether a site's

groundwater is contaminated, samples must be obtained for

chemical analysis to verify the presence of contaminants and

their concentrations. Engineers, owners, contractors, and

regulatory officials are concerned with obtaining groundwater

samples that are "representative" of actual insitu conditions.

Obtaining "representative" groundwater samples for chemical

analysis is extremely difficult, if not impossible. It

parallels the problem geotechnical engineers have in obtaining

"undisturbed" soil samples.

The literature review that follows first gives

descriptions of two DPT systems (BAT System and the

Hydropunch) used for performing groundwater contamination

assessment and then presents a critical review of previous

testing of these devices. Lastly, a discussion of headspace

in sample vials is presented.











2.2 Direct Push Technology

Development of DPT within the past ten years has made it

possible to take groundwater samples without having to drill

and install a monitoring well. This is a significant

breakthrough since Pettyjohn et al. (1981) have shown that

drilling monitoring wells can alter the chemical and microbial

environment in the vicinity of the well. Drilling fluids can

especially alter the chemistry of the groundwater. Direct

push technology has been around for decades in the form of

cone penetration testing. Cone penetration testing is used to

determine soil stratigraphy by measuring the end bearing and

frictional resistance on a tip of standard dimensions, which

is pushed into the ground at a standard rate. ASTM D3441

governs the performance of the test. In DPT a sampling

device is attached to a string of drill rods and either

hammered or hydraulically pushed into the ground to the

required sampling depth. Two such commercial devices which

will be discussed below are the Hydropunch and the BAT

Groundwater Monitoring System.



2.2.1 Hydropunch

The Hydropunch device was introduced in March 1985. Edge

and Cordry (1989) give an excellent overview of the system.

The device has a stainless steel drive cone, a stainless steel

perforated intake for sampling, and a stainless steel sample

chamber. To obtain a sample, the device is attached by means











of an adapter to either cone penetration rods or drill rods.

The device is pushed hydraulically to the required sampling

depth. The push rods are then pulled up 1.5 feet to expose

the stainless steel intake. The drive cone is held in place

by the friction of the soil. Once the intake is exposed,

water flows through the intake into the sample chamber due to

hydrostatic pressure. Once the chamber is filled (5 mL) the

device is pulled to the surface. A ball valve, similar to

that in a bailer, closes during extraction ensuring no loss of

sample. At the surface the device is disassembled and a

discharge tube inserted into the unit to allow transfer of the

fluid sample to a container for storage and transport to a

lab. Figure 2.1 is a diagram showing the Hydropunch

operation.

The device minimizes cross contamination. Since the

exterior of the device is smooth, contaminated soil is not

transported down as the sampler is advanced. It also has the

advantage of not exposing the sample to negative pressures

which could cause degassing and loss of VOCs. It has the

disadvantage of only obtaining a very small sample (5 mL). In

sandy soils, samples can be obtained in as little as 5

minutes. In clayey soils it may take 45 minutes or longer.

Like the bailer, the sample must be transferred to a sample

container, a process which could cause loss of VOCs. The

hydropunch requires no purging of water as is necessary in the

use of monitoring wells.




















































Figure 2.1 Hydropunch in Closed and Open Positions











2.2.2 BAT Probe

Torstensson (1984) describes another probe that was

developed for groundwater sampling using DPT. The standard

device, known as a BAT probe, consists of a tip with a porous

filter. The probe is attached to special drill rods (1 inch

minimum inner diameter) and then pushed hydraulically into the

ground to the required sampling depth. At the ground surface,

a vacuum pump is used to evacuate a test tube (35 mL) which is

placed within a housing. A chain of weights is attached to

the housing and to a steel cable, which is used to lower the

apparatus through the center of the drill rods. The housing

also contains a double-ended needle, which is installed within

a spring. Once the housing reaches the bottom of the drill

rod, the hypodermic needle first makes contact with a septum

in the BAT probe and is then pushed up through a septum in

the test tube. Water is thus drawn through the porous filter

of the BAT probe and into the evacuated test tube. When

equilibrium is reached the apparatus is manually pulled to the

surface and the test tube removed from the housing, labeled,

and placed in a cooler for transport to a laboratory. When

pulling up on the steel cable, the spring in the housing

causes the double-ended needle to simultaneously lose contact

with the test tube and the septum of the filter, ensuring no

sample loss. The BAT system has other attachments which can

be lowered down the drill rods to measure pore water pressure

and to perform hydraulic conductivity tests. Torstensson









10

(1984) describes the use of the system in monitoring Cl-

concentrations with depth and with time at a Stockholm,

Sweden, site in 1981. Figure 2.2 shows a diagram of the BAT

Enviroprobe.



2.3 DPT Field Studies

Edge and Cordry (1989) discuss several case histories in

which the Hydropunch was used in groundwater contamination

studies. It was used in 1985, 1986, and 1987 at a California

landfill for the detection of leaking, low level, VOCs. At a

schoolyard in the spring of 1986 in Los Angeles, California,

it was used on a weekend to confirm the presence of benzene,

toluene, and xylene contamination. The study was carried out

over the weekend to minimize disruption to the school and was

completed at a third of the cost of monitoring wells. In

1988, it was used at a petrochemical facility in Louisiana to

detect low levels of chlorinated organic. The concentration

and extent of contamination was determined which allowed

planning of remedial measures. Klopp et al. (1989) discuss

several case histories of the use of the BAT system in ground

water contamination studies. It was used to study the

stratification of arsenic near San Francisco in 1986. It was

also used for several projects in Texas under the review of

the USEPA and the Texas Water Commission. These included

sampling at a leaking storage tank, delineation of a plume,
















SEPTUM


=II- 4 RETAINING NUT


PROBE BODY

THREADED CONNECTION


POROUS FILTER m



TIP


Figure 2.2 BAT Enviroprobe in Closed and Open Positions


7-7









12

detection of buried hazardous waste, and sampling at an

abandoned land fill. This paper notes that the BAT system can

take a sample at one elevation, be decontaminated at that

elevation, then pushed to another depth for additional

sampling in order to vertically delineate the ground water

quality.

Smolley and Kappmeyer (1991) used the Hydropunch at

Silicon Valley, California for the plume delineation of VOCs

that had leaked from underground storage tanks. The

Hydropunch was first validated by pushing it beside an

existing monitoring well and taking ground samples from each

for comparison. The Hydropunch gave higher concentrations of

VOCs than samples from the monitoring well. The operators of

the Hydropunch only had a 70% sample collection success rate

at this site. They found that the check valve does not always

close properly, causing a loss of sample. The study concluded

that by using the Hydropunch samples could be collected at 20

to 40 percent of the projected cost of monitoring wells. The

work was also completed in a third of the time that would have

been required for monitoring wells. This paper states that

samples up to 500 mL can be taken with the sampler, which is

significantly larger than that stated by Edge and Cordry

(1989). This must be due to some modification to the system.

Strutynsky and Sainey (1990) discuss the use of both the

Hydropunch and the BAT system at an industrial site in

southern Ohio to delineate a trichloroethene plume. These









13

systems were chosen because the plume had moved off the

manufacturer's site onto the property of a farmer who did not

want monitoring wells installed on his property. At this site

the operators had an 81% sampling success rate with the

Hydropunch (22 samples out of 27 attempts). It was noted that

the tip had to be placed at least 4 feet below the water table

for sample collection. The BAT system was tried at two

locations at this site. The first was not successful as the

thin walled casing used to push the BAT probe buckled.

Normally, heavy duty steel drill rods are used to push the BAT

probe. At the second location the BAT Enviroprobe was

successful in obtaining multiple samples.

Lammons et al. (1991) discuss the use of the Hydropunch

at an industrial site in South Carolina to delineate the

vertical and horizontal extent of ground water contamination.

The Hydropunch took samples at several locations which were

then made into permanent monitoring wells. The Hydropunch was

able to take a significantly greater number of samples than

conventional methods for the same cost, which made it possible

to more effectively delineate the contaminant plume.

Torstensson and Petsonk (1988) discuss the work done by

the Earth Technology Corporation at two contaminated sites in

California. At one site samples were taken from a monitoring

well with a conventional teflon bailer. The BAT probe was

then placed inside the well and additional samples taken. The

samples were analyzed by EPA methods 601 and 602. The BAT









14

samples showed consistently higher concentrations of VOCs. For

one contaminant, tetrachloroethene, the BAT probe recovered

77% more than the bailer did. The actual concentrations of

contaminants were probably even greater than obtained with the

BAT probe, as the 35 mL sample vials had a slight headspace in

them. Torstensson suggested that two sample vials could be

used in conjunction to avoid having any headspace. Two test

tubes would be placed in a sample housing with a hypodermic

needle between them. During sampling, the bottom test tube

would fill completely with the overflow going up into the

second test tube. Therefore the second test tube would

contain water and headspace while the bottom one would be

completely filled.

Geomatrix Consultants (1986) used the BAT probe at East

Palo Alto, California, to delineate an arsenic plume.

Fourteen BAT soundings were made to determine the

concentration of arsenic with depth. Samples obtained from

the BAT system were in general agreement with those obtained

from the few monitoring wells that were already in place.

The BAT probe can also be used for vadose zone monitoring

as demonstrated by Haldorsen et al. (1985). A sintered

ceramic filter with an average pore size of 2 microns is used.

Since initially a vacuum is placed on a test tube, this device

can only obtain a ground water sample when the soil tension is

less than about 10 meters of water. By measuring the pore

pressure changes with time as the water flows into the









15

sampler, the unsaturated hydraulic conductivity can be

calculated. The saturated hydraulic conductivity in an

unsaturated zone can be determined by performing an outflow

test while measuring pore pressure changes with time. This is

done by partially filling a test tube with water and then

pressurizing it. The water is forced into the unsaturated

zone and gradually saturates the soil. In a fairly

homogeneous soil the probe can be used at different depths,

each having a different soil tension, to obtain the hydraulic

conductivity which can be used to plot the K(h) curve.

Petsonk (1985) discusses the theory for performing hydraulic

conductivity tests which is based upon work by Hvorslev,

Dachler, and from Boyle's Law.

A large scale field comparison of several ground water

sampling devices was performed at the Desert Research

Institute in Las Vegas, Nevada, in conjunction with the USEPA

Environmental Monitoring Systems Laboratory (Blegen et al.,

1988). The sampling devices used in the comparison included

a teflon bailer, a bladder pump (Well Wizard), the West Bay MP

System, two in situ BAT probes, and a BAT well probe. Six

"monitoring" wells were installed at a site contaminated with

benzene and chlorobenzene. Three wells were of the

conventional type. Two others were constructed by drilling a

borehole, inserting a BAT probe into it, placing a gravel pack

around the probe, and backfilling the borehole. The last well

was a borehole in which the Westbay MP system was installed.










16
The wells were set up in a rectangular grid with a 20 foot

spacing between them. Samples were taken with each device

over an eight week period. The bladder pump and the BAT probe

generally obtained the highest recovery of organic; the

Westbay system obtained the lowest, with the teflon bailer in

between.

2.4 Sample Preservation

Johnson et al. (1987) looked at the effect of headspace

in sample bottles on the loss of VOCs. They found that if the

volume of the headspace is greater than 5% of the total volume

of the container, significant loss of VOCs could occur. In

one instance, a sample bottle only half filled with a sample

had a 50% reduction in l,l,l-trichloroethane.

Pankow (1986) also studied the effect of headspace on the

loss of VOCs. He provides an excellent table and figure

relating the loss of VOCs to different headspace volumes for

numerous organic chemical compounds. Appendix G provides

expanded figures of the concentration remaining in solution

versus different headspace volumes for selected aromatic and

chlorinated organic compounds. He found that samples

containing benzene and toluene and with a ratio of volume of

headspace to volume of sample as great as .1 only suffered a

compound loss of 1%. However, with some other organic, such

as vinyl chloride and chloroethane, a .1 ratio of headspace to









17

sample would result in a loss of 25% or more. The effect of

headspace is thus very dependent upon the contaminant in

question.

Clesceri et al. (1989) recommend that sample vials for

volatile analysis have no headspace to avoid loss of volatiles

by volatization. Samples once taken in the field should be

placed in a cooler supplied with ice packs to maintain the

temperature at 4C. By keeping the samples cool and dark

there is less chance of loss of volatiles and less chance of

growth of microorganisms.



2.5 Necessity for Current Study

Blegen et al. (1988) performed a study comparing seven

different samplers, including the BAT groundwater monitoring

system and a Teflon bailer, at a single site in Nevada.

Several issues were not addressed in this study. At no time

were BAT and Teflon bailer samples obtained from the same

monitoring well to be sure that basically the same water was

being sampled. In this study, a hole was predrilled, the BAT

probe installed, and the hole then backfilled with a gravel

pack, fine silica sand, and a cement-bentonite slurry. The

BAT probe in this case was basically an installed monitoring

well. This defeats the major purpose of the BAT, which is to

eliminate drilling and installation of a monitoring well. The

question is whether the BAT can be penetrated into the ground

and recover representative samples which contain contamination









18

comparable to that obtained using a monitoring well and

bailer.

In Blegen's study the BAT probe was installed 20 feet

from the monitoring well where Teflon bailer samples were

obtained. Such a significant distance in itself may cause a

discrepancy in the concentrations of contaminants measured.

Variations can also be caused by the sampling depth interval.

The BAT groundwater monitoring system samples over a length of

2 inches while the wells installed were screened over a 12

inch interval. No information in their study is given as to

whether or not BAT samples were obtained using cascaded

sampling techniques to eliminate headspace. No data was

presented where the BAT probe was used with a stainless steel

filter in comparison to a HDPE filter.

The current study is needed to develop a larger database

with the BAT groundwater monitoring system. Numerous

regulatory agencies are skeptical of new and innovative

systems until they are thoroughly tested to validate their

use. This study will provide sampling with the BAT probe

using both steel and HDPE filters to evaluate the better of

the two. BAT sampling will be performed inside monitoring

wells to allow comparison directly with Teflon bailer samples

to validate its ability to recover VOCs. BAT sampling will

also be performed adjacent to monitoring wells to show that it

can recover higher concentrations of VOCs than the teflon

bailer in monitoring wells due to the dilution effect which









19

can occur in monitoring wells. Statistics will be given to

show that the BAT system is a more precise device by showing

a lower standard deviation and relative standard deviation

(coefficient of variation) than the bailer.

BAT samples will be analyzed at different amounts of

headspace to see if significant losses of VOCs occur in the

BAT sample tubes due to headspace. Pankow (1986) has

previously shown that losses of BTEX in sample vials with

small amounts of headspace was minimal.















CHAPTER 3
LAB PERMEABILITY TESTING



3.1 Permeability Limitations of the BAT System


The insitu coefficient of permeability can be determined

using the BAT groundwater monitoring system. Both inflow and

outflow testing is possible. For inflow testing the BAT MK2

probe is hydraulically pushed to the desired depth. The pore

water pressure is determined by using the BAT pore water

pressure adaptor. The adaptor uses a single ended hypodermic

needle to make hydraulic connection between the BAT probe and

a pore pressure transducer which is connected to a digital

display unit. The insitu pore pressure is recorded from the

hand held display unit. A double ended test tube is then

placed in a housing and is connected to the pressure

transducer by means of a single ended hypodermic needle which

is connected to an extension cylinder. The extension cylinder

has a small port which contains a rubber septum. For inflow

testing a needle is used to pierce the septum and either a

syringe or vacuum pump is used to evacuate the test tube. For

outflow testing, water is placed into the test tube and then

pressurized using a syringe through the septum port. The

pressure in the test tube can be read on the display unit.









21

Once the desired pressure is attained the needle is removed

from the septum and the test tube remains sealed at that

pressure. The housing is then lowered down the drill rods

until connection is made with the probe. Upon connection a

stopwatch is started. Pressure readings should be taken from

the hand held digital read-out at regular intervals. In the

inflow test, as water is drawn into the test tube the pressure

becomes more positive. The initial pore pressure reading

along with the pressure readings with time are input into a

computer program "Perm" Version 13 developed by the BAT

company which calculates the coefficient of permeability.

Additional information on the program "Perm" is provided in

section A.12 of Appendix A.

In both the inflow and outflow tests water must pass

through the probe's porous filter and through the hypodermic

needle which provides connection between the sample tube and

the probe. It is apparent that there will be a maximum soil

permeability which can be correctly measured. In a more

permeable soil the flow of water will be governed not by the

soil but by the filter and/or the needle. The determined

"permeability" will be of the device and not of the penetrated

soil. This limiting permeability value was determined by

laboratory testing.

The BAT probe was placed in a bucket of water and inflow

permeability testing performed. An initial vacuum (negative

pressure) was applied to the test tube using a vacuum hand









22

pump. The test tube was then lowered down the drill rods

until needle contact was made with the probe's septum.

Because the test tube pressure was lower than the insitu

(bucket) water pressure, water was drawn into the tube. This

continued until equilibrium was reached, i.e., the pressure in

the test tube was equal to the external water pressure.

Twelve tests were performed using the stainless steel

porous filter and five using the high density polyethylene

(HDPE) porous filter. To determine if the porous filter or

the needle actually provided the limiting permeability, five

additional tests were performed with no porous filter. The

results are shown in Table 3.1. The average permeability of

the set-up using the steel filter was 7.8E-04 cm/sec, using

the HDPE filter 1.7E-04 and with no filter 6.1E-03. This

demonstrates that it is the pore size of the probe filter

which is the limiting component.

The BAT groundwater monitoring system with the porous

filters tested is not suitable for permeability testing in

soils with a coefficient greater than approximately 1.OE-04

cm/sec, i.e., clean sands and sand-gravel mixes. The methods

may be suitable in such soils as clays, silts, and clay or

silt-sand mixes (Cedergren, 1977).



3.2 Lab Permeability of Soils

The coefficients of permeability of three soils were

determined in the laboratory using the BAT groundwater












Table 3.1 Permeability Limitation Values of BAT
System


TRIAL

1
2
3
4
5
6
7
8
9
10
11
12

AVERAGE
STD DEV


PERM
W/STEEL
FILTER
(cm/sec)

9.6E-05
4.6E-04
4.1E-03
2.4E-03
6.2E-04
1.1E-04
1.1E-03
9.6E-05
8.3E-05
8.5E-05
9.0E-05
7.2E-05

7.8E-04
1.2E-03


PERM
W/HDPE
FILTER
(cm/sec)

1.4E-04
1.6E-04
1.4E-04
1.5E-04
2.4E-04








1.7E-04
3.4E-05


PERM
NO
FILTER
(cm/sec)

5.0E-04
8.2E-03
2.1E-02
6.2E-05
7.2E-04








6.1E-03
7.3E-03


monitoring system and then compared with values obtained from

constant/falling head tests. The soils were a uniform white

silica sand, a fine yellow mortar sand, and a 50% silica sand-

50% kaolinite clay mixture. The uniform white silica sand was

obtained from the Feldspar Corporation of Edgar, Florida (EPK

Sand, CAS NO. 14808-60-7). It had an effective size (D1,) of

.16 mm, a uniformity coefficient (Cu) of 1.7 and a

coefficient of curvature (C,) of .93. Its grain size

distribution curve is shown in Figure 3.1. This soil also

classified as an A-3 according to the AASHTO system. The fine

yellow mortar sand (Figure 3.2) had a Do, of .19 mm, a Cu of

1.7 and a C, of 1.08. This soil classified as an A-3 in the









24

AASHTO system. Both soils classified as a SP (poorly graded

sand) according to the Unified Soil Classification System

(USCS). The sand-clay mixture was obtained by blending the

EPK sand with pulverized kaolin that was also obtained from

the Feldspar Corporation (CAS No. 1332-58-7). The mixture had

a plastic limit of 20.0 and a liquid limit of 31.8.

Coefficients of permeabilities of the sands were also

estimated using Hazen's equation:

k-C* (Do1) 2



where k = permeability in cm/sec

C = empirical factor with an average value of 1

Do = effective diameter in mm



Hazen estimates are included in the table of results, Table

3.2



Before performing any BAT permeability testing in the lab

it was necessary to determine the size of container (bucket)

which would avoid any boundary effects that could influence

the results. Drawdown was estimated by assuming a porosity

of the sand, knowing the radius of the available containers

and knowing the volume of water which would be removed per

test (35 mL).












35
H-hw-
7Er2*n


Figure 3.1 Gradation Curve of Uniform White Fine Sand



where H h, = drawdown in cm

n = porosity or saturated volumetric water content

r = radius of the container in cm

To be conservative a low porosity (n) of .3 was selected. A

low porosity would cause a greater drawdown. A bucket of

radius 5.5 inches (14 cm) was selected. For this bucket the

drawdown was calculated as:


WHITE SILICA SAND


LOG DIAMET1 (MM)














































Figure 3.2 Gradation Curve of Fine Mortar Sand


H-hw- 35 -.19cm-.00623ft
,*142*.3



To determine the upper limit of the coefficient of

permeability that could be used with this bucket the following

equation was used (Joint Technical Manual Departments of the

Army, Air Force, and Navy, 1983):


R-C* (H-h,) *v,


YEUIW/ORANGE SAND
100

9 -o -I- --- ---

80 --f-~- ___ ---t __----

70.1----

S60
I. -------_-- -
250- -- ----_ ----- -- -- V.
40- ____
aI I I II

a-
20-



0.01 0.1 1 10
ILD DUAIIE (MM)












where R = radius of influence in feet

H = height of water level beyond the zone of influence

h, = height of water at the probe

H h, = the drawdown in feet

C = empirical factor normally equal to 2 or 3

k = coefficient of permeability in units of 10'- cm/sec


5.5-3*.00623*/T
12
K-6.01*10-2 cm
sec



This coefficient of permeability is significantly greater

than the upper limit of the BAT system (1.OE-04 cm/sec). The

11 inch diameter bucket is therefore satisfactory as regards

to boundary affects.

The experiments with the sand consisted of placing a

known volume of water into the bucket, locating the BAT probe

in the center of the bucket and raining a known amount of dry

sand (by weight) in around it. The bucket was shaken to

vibrate and settle the soil thereby eliminating any large

voids. This procedure was followed until the bucket was

filled with sand (an approximate height of 13 inches) and the

water level was at the surface of the sand. This allowed the

exact water pressure and the unit weight of the material to be

known. The pore pressure was also checked with the BAT pore









28

pressure device. This gave water heights within 2 cm of the

known heights in the bucket. Since the steel filter had the

highest permeability it was used for all testing.

For the BAT permeability testing the white silica sand

had a dry unit weight of 80.6 pcf and a moist (saturated) unit

weight of 109 pcf. The yellow mortar sand had a dry unit

weight of 90.6 pcf and a moist unit weight of 114.4 pcf. The

same respective dry unit weights were used in the constant

head tests.

The sand-kaolin mixture was prepared by placing a known

amount of water into the bucket and adding a known dry amount

of kaolin. The water and kaolin were mechanically mixed.

Sand was mixed in gradually until a 50-50 mixture of sand-

kaolin had been made. The mixture had a dry unit weight of

67.5 pcf and a moist unit weight of 99.3 pcf. The moist unit

weight in the falling head test was 118 pcf.

The results of the permeability testing are shown in

Table 3.2. The kaolin-water mixture had the lowest

coefficient of permeability of the three soils as expected.

The BAT underestimated the coefficient of permeability by a

factor of 100 when compared with the falling head test even

though the material was in a denser state in the falling head

apparatus. The permeabilities' calculated using Hazen's

equation for the sands compared rather well with the values

obtained from the constant head permeability tests. Values

obtained with the BAT probe did not compare well with the










29

constant head tests. It was evident that the fine sands were

more permeable than the steel filter used on the BAT probe.

The BAT testing in the fine sands was still a measure of the

permeability of the steel filter and not of the fine sand. It

is evident that the BAT is severely limited for permeability

testing. Further research could include comparison of the

coefficient of permeability from the BAT probe and from field

pumping tests since both measure predominantly the horizontal

coefficient of permeability.

Table 3.2 Permeability of Three Soils


TRIAL

1
2
3
4
5
6

AVERAGE
STD DEV

HAZEN'S EQ

CONSTANT HEA
FALLING HEAE


PERM
SAND-KAOLIN

(cm/sec)

2.8E-07
2.1E-07
2.0E-07
2.0E-07
2.0E-07


2.2E-07
2.9E-08



D6 -
S6.3E-05


PERM
FINE MORTAR
SAND
(cm/sec)

4.2E-04
6.5E-04
5.9E-05
9.6E-05
5.3E-04
2.6E-05

3.0E-04
2.3E-04

3.6E-02

2.3E-02
-


PERM
UNIFORM
SAND
(cm/sec)

1.8E-05
2.0E-05
1.3E-05
7.4E-05
5.3E-05
5.8E-05

3.9E-05
2.2E-05

2.6E-02

1.6E-02


I















CHAPTER 4
BAT GROUNDWATER MONITORING SYSTEM LAB STUDIES



4.1 Introduction


This chapter discusses laboratory studies which were

performed to evaluate the BAT probe's ability to recover VOCs

as compared to that of the Teflon bailer in a controlled

environment. A model monitoring well was built to sample with

the BAT probe and with the Teflon bailer. Additional testing

with the BAT probe and the bailer was performed in a large

nalgene tank filled with water spiked with gasoline

constituents. This testing was performed to directly compare

the amount of VOCs recovered from the BAT probe and the Teflon

bailer to determine if the BAT performed comparable to the

bailer.

Each section describes in detail all set-up procedures

and testing sequences. All chemical analyses for this study

(both laboratory and field) were performed on a Gas

Chromatograph/Mass Spectrometer (GC/MS) in accordance with EPA

Methods 524/624 which allows water sample storage at 4 C

without preservatives up to seven days. The equipment used

was a Hewlett-Packard 5985 GC/MS "benchtop" system with an HP

5840A gas chromatograph. The GC had a 30 meter capillary









31

column with a 0.32 mm inner diameter. Samples were purged for

11 minutes, desorbed for four minutes, and baked for 6

minutes.

Other topics include modifications of the BAT system to

provide samples with no headspace including the use of

balloons inside test tubes and Teflon tubes with Teflon balls.

A BAT probe with a ceramic filter is used in the laboratory to

simulate vadose zone testing.



4.2 Monitoring Well Model


A ground water monitoring well was constructed in the lab

to model a typical field installation. The purpose of the

model was to allow an evaluation of two types of groundwater

sampling mechanisms, a Teflon bailer (350 mL--Norwell

Company) and the BAT probe. The Teflon bailer is shown in

Figure 4.1.

The model was constructed within a metal 55-gallon drum

which was lined with a plastic nalgene container. The nalgene

container was used to decrease the chemical interaction

sorptionn or leaching) that could occur between the

contaminants and either the metal drum or regular plastic

garbage cans. Because of the flexibility of the nalgene

container the metal drum was needed to provide rigid

containment. Since the nalgene tank had a flow valve on the

bottom, a small slit was made down the side of the metal drum

with a welding torch to allow its insertion. The monitoring










32

well was constructed of a 2.5 foot section of "triloc" slotted

pvc well screen with a #10 slot (.01 inch opening). This was

threaded onto a 2.5 foot section of pvc casing on one end and

a pointed tip on the other. The monitoring well had an inner

diameter of 2 inches. A uniform sand with an effective

diameter of .012 inches was rained in around the monitoring


Figure 4.1 Teflon Bailer









33

well. Figure 4.2 is a photograph of the model monitoring well

set-up.

4.3 Experiment 1 Inside Model Monitoring Well

In the first experiment a solution was prepared by mixing

50 liters of water with benzene, toluene, and o-xylene, each

at a concentration of 20 pg/l (20 ppb). This contaminated

water was then siphoned through a flexible tube into the

bottom of the monitoring well, from which it spread radially

into the sand-filled container. Filling took approximately

two and a half hours.

Several attempts to mix the chemicals directly with water

were unsuccessful due to the relative insolubility of these

volatile aromatics. Each attempt only provided a non-aqueous

phase liquid (NAPL) above the water due to its lower specific

gravity and hydrophobic characteristics. To overcome this

problem 0.5 grams of benzene, toluene, and o-xylene were added

to 50 ml of methanol (CH3OH) to dissolve the aromatics. This

provided a solution with a concentration of 10,000 mg/l (ppm).


.5g 103mL 10mg10000mg
50mL L g L



To achieve a desired concentration of 20 ppb, 100 p1 of

the above solution was injected into a tank containing 50

liters of water. The water and chemicals were gently

mechanically mixed with a wooden rod.












1 X1,00 g x103R
1x1O,OOO0x lQCji
100LLx 106 pL L mg =20tig
50L L


Figure 4.2 Model Monitoring Well Set-Up


When the filling of the tank with water was completed, a

Teflon bailer, which had been cleaned and stored in aluminum

foil, was lowered down the monitoring well to obtain a water

sample. This first bailer sample was discarded. The bailer









35

was again lowered down the well to obtain a sample. Upon

retrieval, two 40 mL glass vials (teflon sealed) were filled

using the bottom control flow valve. While inserting the flow

control valve into the bottom of the bailer, it appeared that

a small air bubble was introduced into the water and traveled

up through the bailer. Two 40 mL glass vials were also filled

by decanting the water through the top of the bailer. The

vials were refrigerated for sample preservation. A BAT probe

(Figure 4.3) was then lowered down through the center of the

well to obtain samples. The first sample was discarded

because 8 mL of the water in the sample was from the water

that was used to saturate the porous filter. Three attempts

were made to obtain samples using the cascaded type system of

two test tubes in series (Figure 4.4). This procedure was

used to collect a bottom tube sample with no headspace while

the upper sample will contain some headspace. Cascaded BAT

samples were numbered with odd integers for the lower (zero

headspace) tube samples, e.g BAT3, and with the next (even)

integer for the upper (with headspace) tube sample for the

same test. Only one of the three attempts yielded a sample

with no headspace. There appeared to be a bad connection

between the two test tubes. In the two unsuccessful attempts,

the bottom test tubes were partially filled while the upper

tubes were empty. The upper test tubes when opened still had

vacuums. Samples were stored for less than one day before

performing the chemical analyses. Results are shown in Table
















EVACUATED TEST
TUBE


SEPTLI -
DOUBLE ENDED
NEEDLE
SEPTUM





POROUS FILTER H


Figure 4.3 BAT MK2 Probe













-,U UPPER SAMPLE TUBE
WITH SOME HEADSPACE


- HYPODERMIC NEEDLE




-_LOWER SAMPLE TUBE
NO HEADSPACE


---- HYPODERMIC NEEDLE

S--- TIP SEPTUM


Figure 4.4 Cascaded Sampling for Zero Head Space Sample









38

4.1. The bailer samples recovered more VOCs than did the BAT

samples. The single BAT sample with no headspace still

recovered 3% less benzene, 15% less toluene, and 24% less

xylene than did the average of all the bailer samples.



4.4 Experiment 2 Inside the Model Monitoring Well

After the chemicals had remained in the drum for three

days, additional samples were taken. Two full samples without

headspace were obtained using the cascaded technique (BAT1 and

BAT3) and one sample was obtained using the single vial

technique (BAT5). These samples were also taken by placing

the BAT probe down the center of the monitoring well. The

results are shown in Table 4.2.

These two rounds of sampling did show the effect that

headspace has on the loss of VOCs. Generally, the larger the

headspace the smaller the amount of VOCs observed. The sample

that was obtained during experiment 1 which was stored for

five days did not show any additional loss of VOCs as compared

to those stored for only one day.

A very important factor discovered was that the method of

extracting the water sample from the BAT test tubes played a

large role in the levels of observed contaminants. It was

determined that the best method to obtain the water from a

double-ended test tube was to hold the tube vertically, remove

the top end of the test tube, insert the needle of the syringe

through the bottom end and draw the water out of the test












Table 4.1 Chemical Analyses of Sampling Within Model
Monitoring Well


SAMPLE HEAD BENZENE TOLUENE XYLENE COMMENTS
SPACE ppb ppb ppb
%

TANK 14.6 11.8 12.1
WATER

BAT 15 13.7 8.8 7.8 SINGLE-
ENDED TEST
TUBE

BAILER 16.1 11.2 12.3 BOTTOM
FLOW
CONTROL
VALVE

BAT 57 10.3 8.0 7.3 DOUBLE-
ENDED TEST
TUBE

BAILER 16.0 11.8 12.4 DECANTED
TOP

BAILER 16.8 12.2 12.9 DECANTED
TOP

BAT 0 16.0 10.0 9.6 DOUBLE-
ENDED TEST
TUBE

BAT 50 11.9 6.7 6.4 DECANTED
THROUGH
NECK

BAILER 17.0 12.0 13.0 BOTTOM
FLOW
CONTROL
VALVE
AVG BAILER 16.5 11.8 12.7

True concentration of water in tank was to be 20 ppb (20
1g/l).
BAT samples obtained by lowering probe down the center of
the pvc monitoring well.












Table 4.2 Chemical Analyses
Monitoring Well


of Experiment Two Inside the Model


HEAD BENZENE
SPACE ppb
%

0 4.7




9 3.5




0 2.6




43 2.8




5 2.3


BAT 39
STORED
FIVE
DAYS
(FROM EXPERIMENT


11.8


TOLUENE
ppb


2.7




2.7




1.5




1.8




1.3




7.2


XYLENE
ppb


4.7




4.4




3.7




3.8




2.9




5.0


COMMENTS


DOUBLE-ENDED
TEST TUBE
NEEDLE
EXTRACTED

SINGLE-ENDED
TEST TUBE
NEEDLE
EXTRACTED

DOUBLE-ENDED
TEST TUBE
NEEDLE
EXTRACTED

SINGLE-ENDED
TEST TUBE
NEEDLE
EXTRACTED

SINGLE-ENDED
TEST TUBE
NEEDLE
EXTRACTED

DOUBLE-ENDED
TEST TUBE
NEEDLE
EXTRACTED


tube. When extracting the sample with both ends sealed

(also occurs when using a sealed single-ended test tube), it

becomes quite difficult to remove the sample and if the


SAMPLE


BAT1




BAT2




BAT3




BAT4




BAT5









41

analyst is not careful the sample can be pulled back into the

test tube. If an additional needle is placed in the bottom

test tube to relieve the vacuum while the sample is being

extracted with the syringe, little air bubbles move up through

the sample. This may cause a loss of VOCs. Water should be

slowly drawn out of the test tubes to allow the syringe to

gradually fill without any headspace. If the water is removed

too quickly, bubbling can occur as the water enters the

syringe. The most successful method for extracting the water

from a single-ended test tube is to remove the end of the test

tube and to hold the test tube in a near horizontal position.

The needle of the syringe is then placed in the test tube and

kept under the water level. As the water level is lowered,

the test tube is inverted slightly to keep the needle under

the water. Another method not attempted would be to remove

the seal and, holding the tube vertically, use a syringe with

a long enough needle to reach the bottom of the test tube. If

an adequate needle is not available, it may be possible to

place a length of thin tubing over the needle which can be

lowered down the test tube.



4.5 Experiment 3 Inside the Model Monitoring Well

A third round of testing was performed after draining the

water from the tank the previous day. A new contaminated

solution of 20 ppb each of benzene, toluene, and o-xylene was

mixed and slowly siphoned by gravity down the monitoring well









42
to fill the model. A 40 mL vial was filled with the

contaminated water. A Teflon bailer was used to obtain two

samples from the well (BAILER1 and BAILER2). This sampling

was performed within 5 minutes of the filling. Since there

was little time for interaction between the contaminants and

the pvc well, no purging was performed.

BAT probe sampling was performed for the first time in

the soil, adjacent to the monitoring well. The probe was

pushed to the bottom of the tank using a hydraulic jack. A

load frame constructed of four inch steel channel, and shown

in Figure 4.5, provided the reaction for the penetration push.

Two samples without headspace (BAT1 and BAT3) were obtained

after drawing the water out of the filter. The BAT probe was

then removed from the tank and a second probe inserted at a

different location and to a shallower depth. Two samples with

no headspace (BAT5 and BAT7) were obtained from this depth.

The results of the chemical analyses are shown in Table 4.3.

The bailer samples contained higher concentrations of

contaminants than the BAT samples. The BAT samples taken from

the very bottom of the tank were quite low. This was probably

due to incomplete draining of the tank which allowed the old

contaminated water to be sampled. The longer the water remains

in the tank, the greater the chance of sorption of the

contaminants onto the soil and loss of VOCs. The

concentrations of contaminants in the upper BAT samples were

closer to those obtained from the bailer samples, but were












































Figure 4.5 BAT Probe with Reaction Frame


still generally around 35% lower.

At this time, it was believed that there were two

principal reasons why the BAT system was not recovering

similar levels of VOCs as the bailer. One reason was that the

bailer samples were taken within a few minutes of filling the

well, with little time for the VOCs to volatilize. The water

obtained from the BAT samples taken a couple of hours later,










44

Table 4.3 Chemical Analyses of BAT Probe Sampling Inside the
Model Monitoring Well


SAMPLE HEAD BENZENE TOLUENE XYLENE COMMENTS
SPACE ppb ppb ppb
%

TANK 18.1 21.3 24.3
WATER

BAILER 17.7 19.0 21.5

BAILER 15.5 16.4 18.6

BAT1 0 4.4 3.4 2.1 BOTTOM OF
TANK

BAT3 0 5.2 4.2 2.7 BOTTOM OF
TANK

BAT5 0 10.6 10.6 9.4 UPPER PART
OF TANK

BAT7 0 12.4 12.0 13.9 UPPER PART
OF TANK

BAT pushed into the soil with the use of Hydraulic Jack.

Water spiked to provide concentrations of 20 pg/1 (ppb)
for each contaminant.







however,had plenty of time to interact sorbb) with the soil

perhaps resulting in a lower recovery of VOCs. The second

possible explanation for the lower recovery of VOCs was that

the BAT's use of a vacuum causes a loss of VOCs. It was

thought that the water entering the BAT test tube would bubble

due to the vacuum which had been placed on the test tube. The

bubbling would cause a loss of volatiles as they would enter









45

the gaseous phase. A lab experiment was performed to see if

the water did bubble when entering the test tube. A vacuum

was placed on two test tubes which were then connected with a

double-ended needle. Another double-ended needle was placed

into the septum of the BAT probe, which had been placed in a

bucket of water. The bottom test tube was then placed in

contact with the exposed needle from the BAT probe. At the

instant contact was made, water was pulled into the test tube

and bubbling did occur. Bubbling occurred but it became less

dramatic as the test tube filled.

The test was repeated with the probe's porous filter

removed to see if it could have been only partially saturated.

In which case, the bubbles that formed would be due to air

entrapped in the porous filter and pulled into the test tube.

The test showed considerable bubbling, which eliminated the

filter as the responsible party.

Another test was performed using degassed water (boiled

water) to see if the bubbling effect was due to dissolved gas

being pulled out of solution by the vacuum. Less bubbling

occurred. Bubbling will probably always occur as long as

there is a head space when the water enters the test tube.

When a syringe without any headspace is used to slowly

withdraw water from a test tube or vial almost no bubbling

occurs. If the syringe is pulled strongly back and a

headspace is formed, the water will bubble when entering the

syringe due to the reduced pressure.









46

In order to eliminate headspace it was decided to make

use of a membrane inside a double-ended test tube. A balloon

was used as the membrane. The balloon was placed in the test

tube with its opening stretched over the neck of the test

tube. The top was then screwed on over the balloon (Figure

4.6). A syringe was inserted through this top to evacuate the

air from the balloon. This caused the balloon to collapse.

The top was then screwed onto the other end of the test tube

and the air evacuated with a syringe. This membrane test tube

was used with the probe in the lab with tap water with

virtually a 100% success rate. When sampling, the water would

enter the balloon and fill it. Once the unstretched length of

the balloon filled, it would continue to fill as the water

stretched the balloon until it came in contact with the walls

of the glass tube. The water would then continue to expand

the balloon upward. When viewing the test tube after

sampling, a small bubble was observed in the water filled

balloon.



4.6 Experiment 4--Sampling Within Tank Spiked Water

This experiment consisted of filling a nalgene container

with 200 liters of distilled water. The container had less

than 10% headspace. Benzene, toluene, and o-xylene were

injected into the tank to give it a concentration of 10 pg/l

of each contaminant. The tank was mechanically mixed with a

pvc slotted well screen. The objective here was to directly












































Figure 4.6 Balloon and Test Tube Apparatus



compare the bailer and the BAT without the presence of the

sand. The sand was thought to sorb some of the contaminants,

resulting in the lower recovery of volatiles by the BAT system

in earlier experiments.

The BAT probe was lowered into the tank and suspended

slightly below the water level using two "C" clamps on the

drill rod, Figure 4.7. Tape was placed around the drill rod

to seal the hole and the escape of any gaseous fumes. The












































Figure 4.7 BAT Sampling in Nalgene Container


first BAT sample was as usual discarded, as it contained at

least 8 mL of the distilled water which had been used to

saturate the probe. Three BAT samples (BAT1, BAT3 and BAT5)

were collected without headspace by the cascaded technique.

A minimum of 10 minutes was required to completely fill the

bottom test tube before any filling of the upper test tube

occurred. Two BAT samples (BAT7 and BAT8) obtained using a

single-ended test tube, filled approximately 90% within 7









49

minutes. Two BAT samples, BAT9 and BAT10, were obtained using

the balloon technique. After sampling with the BAT system,

two samples of the tank water were obtained with the Teflon

bailer, BAILER1 and BAILER2.

Results of the chemical analyses are shown in Table 4.4.

The bailer samples again recovered the highest percentages of

VOCs. Samples using the balloon technique recovered the

lowest percentage of VOCs. This was undoubtedly due to

sorption of the contaminants onto the rubber balloon. There

was an extreme variation in the results obtained from the BAT

samples with and without headspace.

Statistical data such as the standard deviation (STD) and

the relative standard deviation (RSD) were calculated from the

equations given below. The relative standard deviation is

also known as the coefficient of variation. The standard

deviation and relative standard deviation are both measures of

skewness. They give us an idea on the precision of our data.

The smaller the skewness in the data the higher the precision

in the sampling procedure and device. This infers that the

sampling procedure is also highly reproducible and gives us a

high level of confidence.












Chemical Analyses of Sampling Within Tank Spiked


Water


HEAD BENZENE
SPACE ppb
%


TOLUENE XYLENE
ppb ppb


COMMENTS


BAT9
BAT10


BAT1
BAT3
BAT4


BAT5
BAT7
BAT8


AVG BAT

BAT-NO HEAD
AVG
STD
RSD


BAILER1
BAILER2

AVG
BAILER
STD
RSD


2.49
1.72

5.84
9.02
5.40


7.68
8.20
8.61

7.46


SPACE (#1
7.51
1.3
17.3


9.08
9.60

9.34

0.3
3.2


1.29
1.01

7.01
9.10
6.30


8.65
7.69
8.30

7.84


,3,5)
8.25
0.9
10.8


9.80
9.99

9.90

0.1
1.0


0.56
0.44

7.81
9.20
6.96


9.13
8.21
9.50

8.47


8.71
0.6
6.9


11.20
11.90

11.55

0.3
2.6


BALLOON
BALLOON



PRESSURIZED
BEFORE
EXTRACTING


EXCLUDING
BALLOON
SAMPLES


Tank Spiked to give actual concentrations of 10 14g/1
(ppb) for each contaminant.


SAMPLE


Table 4.4













STD= (-) 2
N
RSD- STx100
X-actual concentration
X-mean concentration
N-numberofsamples



4.7 Experiment 5--Sampling Within Tank Spiked Water


This experiment consisted of spiking 225 liters of water

with benzene, toluene, and xylene to achieve a concentration

of 8.9 ppb for each contaminant. The chemicals were mixed

mechanically as in Experiment 4. The container had less than

2% headspace. Five samples (BAT1, BAT3... BAT9) without

headspace were collected with the BAT probe using the cascaded

technique. Each test tube did contain a small bubble. Five

samples were obtained with the teflon bailer by decanting from

the top into 40 mL vials. Results of the chemical analyses

are shown in Table 4.5.

Since the results again showed that the BAT recovered

lower BTX concentrations, an attempt was made to obtain

samples using hydrostatic pressure rather than with reduced

pressure. This method is used by the Hydropunch system. A

cascaded type system was used with two modifications. Two

double-ended test tubes were used. A cap was not placed on

the top of the upper test tube. This allowed air to vent from

the test tubes as they filled. The second modification

consisted of grilling a hole into the top of the metal plug









52

which is screwed down on to the test tube container housing.

This allowed the air to vent from the test tube and from the

container housing.

The BAT probe was placed approximately three feet below

the container free water surface. After sixteen hours the

lower test tube was about 80% (28 mL) full. Such a length of

time would, in most situations be impractical. Also such a

long period of time would allow a significant amount of the

VOCs to vaporize.

Another attempt at collecting a sample hydrostatically

was performed. The threaded glass ends were removed from the

tube and fused onto a smaller diameter tube of approximately

the same length. The modified tube held approximately 12.5 mL

of water, about 1/3 the standard tube's volume. It was hoped

this would significantly reduce the time required for

sampling. The modified tube filled completely, with no

headspace, and approximately 10 mL entered the upper tube

within seventeen hours. This however was also considered

inadequate.

4.8 Experiment 6--Sampling Within Tank Spiked Water

This experiment again used the balloon technique.

Testing was as previously tried with the exception that the












Table 4.5 Chemical Analyses from
Water


Sampling Within Tank Spiked


SAMPLE



BAILER1
BAILER2
BAILER3
BAILER4
BAILER5


HEAD
SPACE


0
0
0


AVG BAILER

STANDARD DEVIATION

RELATIVE STANDARD
DEVIATION


BAT1
BAT3
BAT5
BAT7
BAT9


AVG BAT


STANDARD DEVIATION

RELATIVE STANDARD
DEVIATION

BAT % LOWER


Tank spiked to provide actual concentrations of 10 pg/l
(10 ppb) for each of the contaminants.


BENZENE
ppb


9.9
9.5
7.4
9.9
8.3

9.0

1.0

11.1



6.9
7.6
8.4
6.7
6.7

7.3


0.7

9.6


18.9


TOLUENE
ppb


9.7
9.1
9.7
9.9
10.2

9.7

0.4

4.1



9.1
9.1
8.7
9.0
8.4

8.9


0.3

3.4


8.2


XYLENE
ppb


10.1
10.0
10.7
10.6
11.0

10.5

0.4

3.8



9.1
9.4
8.7
9.3
8.4

9.0


0.4

4.4


14.3









54

inside of the balloon was sprayed with a dry film lubricant

and mold release agent, manufactured by Crown Industrial

Products of Hebron, Illinois (#6075). The product label

states that it is chemically similar to TFE (Teflon) as

manufactured by Dupont.

A solution of 10 ppb (10 gg/l) of benzene, toluene, and

xylene was made. Three BAT samples (BAT1, BAT3 and BAT5) with

no headspace were taken using the cascade technique. Two BAT

samples were collected using the balloon which had been coated

with the teflon spray. Two bailer samples (BAILER1 and

BAILER2) were obtained for comparison. The results are shown

in Table 4.6.

After performing the chemical analysis on the two bailer

samples and two of the no headspace BAT samples, analysis was

performed on the balloon sample. This sample overloaded the

GC/MS system due to the freon propellant that is used in the

teflon spray coating. The third BAT sample with no head space

was run directly after the balloon sample but could not be

properly interpreted. Analysis on the second balloon sample

was not performed. In this test the BAT samples, without any

headspace recovered more VOCs than did the Teflon bailer.



4.9 Teflon Ball and Tube Sampling Apparatus

Another modification to the BAT system was tried to

obtain zero headspace samples. Teflon tubes were manufactured









55

with a constant inner diameter of 1/4 inch and threads

machined on either end to fit the BAT test tube caps.

A Teflon ball, 1/4 inch in diameter, was placed inside

the Teflon tube. The ball was to be in contact with the walls

of the tube with no air space around the wall. After the ball

was placed in the end of the tube, one cap was screwed on and

a vacuum pulled on that side of the tube. The second cap was

then screwed onto the other end and a vacuum placed on that

side. When the hypodermic needle made contact with the BAT

septum and the septum of the Teflon tube, it was hoped that

the higher pressure of the water would push the teflon ball up

the tube and yield a sample without any head space.



Table 4.6 Chemical Analyses of Experiment 6


SAMPLE HEAD BENZENE TOLUENE XYLENE
SPACE ppb ppb ppb

BAILER1 0 11.4 10.0 8.9
BAILER2 0 10.7 8.8 8.2

AVG BAILER 11.1 9.4 8.6


BAT1 0 12.0 10.3 9.5
BAT7 0 11.9 10.0 8.4

AVG BAT 0 12.0 10.2 9.0


BAILER % LOWER 7.5 7.8 4.4
THAN BAT

Water spiked to provide actual concentration levels of 10
Mg/l (10 ppb) for each contaminant.









56

Several trials with this method met little success. Even

though Teflon has a very low coefficient of friction, the

water pressure was not sufficient to push the ball up the

tube. The diameter of the tube was slightly enlarged to see

if this would help. The ball did move slightly better, but

would typically become stuck somewhere in the middle of the

tube. This was probably due to the flexibility of the teflon

tube. If the tube became the slightest bit distorted in any

direction, the inner diameter would change and cause the ball

to become stuck.



4.10 BAT Vadose Zone Probe Testing

Laboratory testing was conducted with a BAT probe using

a ceramic filter. Such a filter, with its small pore size of

2 microns, is necessary if sampling is to be attempted in the

unsaturated zone. Standard BAT filters made of steel or HDPE

have larger pore sizes. These filters allow air to be pulled

into the filter which inhibits the flow of water because a

full vacuum cannot be maintained. This means sampling is

greatly hindered.

A uniform fine silica sand was placed in a plastic

concrete cylinder casing (12" high by 6" diameter) to a height

of 8 inches. Before placing the sand in the container a small

hole was made in the bottom of the casing to allow water to

drain out. A piece of white cotton sheet was taped over the

hole on the inside of the casing to serve as a filter. A









57

piece of strapping tape was placed over the hole on the

outside of the casing to inhibit flow.

The dry soil had a mass of 6012 grams. Water, with a

mass of 1555 grams was then poured into the soil. The

strapping tape was removed and the water allowed to drain into

a pan. When drainage was complete, 161 grams of water had

been collected, which left 1394 grams in the soil. The

initial water content (by weight) was then 23.2% and the moist

unit weight was 124.5 pcf. The BAT probe with ceramic filter

was then saturated in a bucket of water and pushed by hand

into the soil filled cylinder. Pore pressure readings were

taken and sampling performed with the BAT groundwater

monitoring system. Results are shown in Table 4.7.

Some problems were evident with the pore pressure

readings. By removing the water from the soil the pressure

should have become more negative as the test progressed. The

pore pressure reading problems could have been caused from all

the water in the ceramic filter being pulled out.

By taking a soil sample from the field and performing a

test like that above, the soil moisture curve could be

developed. A typical soil moisture curve is shown in Figure

4.8. With this information pore pressures could be read in

the field and correlated to the actual water content from this

graph. The soil water content profile is needed for all

unsaturated flow problems.











4.11 Summary

Testing within a tank filled with water spiked with

gasoline constituents proved to be a better method than

modeling a well inside a 55-gallon drum filled with sand. For

the two experiments in the tank filled water there was

conflicting data. In one test the bailer recovered more VOCs

than did the BAT probe and in the other test the BAT probe

recovered more than the bailer. The data from the BAT probe

showed it to be a more precise device than the bailer by

having a lower relative standard deviation.





















Table 4.7 Vadose Probe Testing


Sampling
Time



20 min


40 min


2 hrs
15 min

24 hrs
15 min

43 hrs
30 min


Water
Recovered
mL
(Water content)

4
(23.1%)

1.5
(23.1%)

4
(23.0%)

23.5
(22.6%)

0.0


Pore
Pressure
cm of
Water

-.42


-.43


-.42


0.0


-.03


Evaporation was not considered in water calculations.

Water = 1 g/cm3 1 ml = 1 cm3
























SANDY SOIL
0.35


0.3-


^0.25 ---------- -
0.25--


8 0.2





0.1


0.05 --- _
1 10 100
NEGATIVE MATRIC PI1ENTIAL cm


Figure 4.8 Typical Soil Moisture Curve















CHAPTER 5
ANALYSIS, TRANSPORT, AND PROPERTIES OF VOLATILE ORGANIC
COMPOUNDS


5.1 Introduction

This chapter is provided to give a overview of some basic

principles of geo-environmental engineering. Discussions are

provided on organic compounds, chemical analysis of water

samples, regulatory contaminant levels, and solute transport.

This background information is necessary before looking at the

field contaminant studies that were performed.

5.2 Organic Compounds

Organic compounds are defined as compounds which contain

some amount of carbon. Hydrocarbons are compounds which

contain only hydrogen and carbon. The most familiar

hydrocarbons are benzene, toluene, ethylbenzene, and the

xylenes. These four compounds are typically known as BTEX.

All four are constituents in petroleum products such as

gasoline. Petroleum hydrocarbons typically have specific

gravities less than one making them float on top of the

groundwater in a separate phase. These compounds are

sometimes called LNAPLs (Light non aqueous phase liquids) or

floaters. Hydrocarbons such as BTEX which typically have low

solubilities in water and volatilize easily are known as VOCs.

Figure 5.1 illustrates the typical transport of a LNAPL.

61









62

Henry's Constant, H, is a coefficient which describes a

compound's partitioning between the liquid and vapor phases.

The higher the Henry's Constant the more likely the compound

is to come out of water and go into a vapor phase. Table 5.1

lists the Henry's Constant of several compounds, from Pankow

(1986).


* DISSOLVED PHASE

I NONAQUEUOUS PHASE LIQUID


GROUNDWATER
FLOW
No-


Figure 5.1 Transport of a Typical LNAPL


MsNN MIM\M\N: n\11M&M,1












Table 5.1 Henry's Constant for Selected VOCs


COMPOUND

BENZENE

CHLOROBENZENE

TOLUENE

ETHYLBENZENE

O-XYLENE

M-XYLENE

P-XYLENE


H

0.0055

0.0036

0.0067

0.0066

0.0050

0.0070

0.0071


COMPOUND

METHYLENE CHLORIDE

CHLOROFORM

CHLOROETHANE

VINYL CHLORIDE

TRICHLOROETHENE

TETRACHLOROETHENE

ETHYLENE DIBROMIDE


H

0.0020

0.0029

0.15

0.081

0.0091

0.0153

0.00082


H is in atm.m'/mol


If VOCs are present in groundwater they vaporize and

migrate vertically and horizontally in the gas phase through

the soil pores until they reach the atmosphere. This is a

natural remediation process. Ballestro et al. (1991) state

that nonhalogenated compounds such as BTEX, when present in

low concentrations in groundwater, readily degrade in

oxygenated soil. This does not occur, however, when large

concentrations are present.

Halogenated organic compounds contain hydrogen, carbon,

and one or more of the halogens, fluorine, chlorine, bromine,

or iodine.

Chlorinated compounds are typically denser than water and

are known as sinkers or dense non-aqueous phase liquids










64
(DNAPLs). These compounds will sink through groundwater until

they reach a confining layer and will then move laterally with

gravity. If the confining layer is angled, the contaminant

can even move upgradient. Figure 5.2 illustrates a simulated

transport of a DNAPL. Chlorinated DNAPLs include chloroform,

tetrachloroethene or perchloroethene (PCE), trichloroethene

(TCE), methylene chloride, l,l,l-trichloroethane (TCA) and,

1,1,2-trichlorotrifluorethane (freon). Trichloroethene is a

degreasing solvent and is the most common contaminant found in

groundwater. Tetrachloroethene is used in dry cleaning fluid.



5.3 Chemical Analysis

Once groundwater samples are taken they must be analyzed

to determine the presence and concentration of contaminants.

For gasoline spills or leaking underground storage tanks, EPA

Method 602 titled Purgeable Aromatics is run to determine the

presence and concentration of the aromatic chemicals. These

include benzene, chlorobenzene, the three dichlorobenzenes,

ethylbenzene, and toluene. This analysis consists of

injecting a sample into a purging device where an inert gas

such as helium is bubbled through the water sample to

volatilize the contaminants. These are then trapped on a

sorbant material. The trap is then heated and backflushed

with helium to desorb the contaminants which are then sent to

a gas chromatograph (GC) for separation and detection. Before












































Figure 5.2 Transport of DNAPL

running any samples the gas chromatograph must be calibrated

for the contaminants of concern.

These contaminants are individually run through the gas

chromatograph since it cannot absolutely distinguish between

compounds. By running each compound separately the retention

time is determined for each compound. Compounds come off the

GC column in order of their boiling points. With this method

there can still be some error because several compounds may









66

elute (come off the column) at the same retention time.

Figure 5.3 shows a typical total ion chromatograph.

The state of the art for groundwater analysis makes use

of a gas chromatograph in conjunction with a mass spectrometer

(GC/MS). EPA Method 624 titled Purgeables makes use of the

GC/MS for the detection and quantitation of not only the seven

contaminants found in Method 602 but 24 other compounds. This

analysis is run in a similar manner to that of Method 602 with

the exception that after the sample leaves the GC it is sent

to the mass spectrometer. The mass spectrometer bombards the

compounds with electrons to try to ionize them by knocking off

electrons and some of the atoms. As the compound is bombarded

with electrons it is scanned several times a second to

determine the atomic mass units (AMUs) that are present and

their relative intensities. This allows better determination

of compounds. Each compound has a mass spectrum which is its

own unique fingerprint under the given conditions. The mass

spectrum of a compound shows the fragmentation ions that are

present and their relative amounts. A typical mass spectrum

is shown in Figure 5.4. The compound shown is o-xylene (1,2

dimethylbenzene). Its chemical formula is C,Ho, resulting in

a molecular weight of 106. This is one of the peaks shown.

Another peak shown is 91 which comes when a methyl group

(CH3), having a weight of 15, is knocked off the compound.

The mass spectrum for each peak can be viewed to determine














AREA TABLE ENTRIES: FRN 17067

Entry Time Ma. Ar-ea

1 92.9 94.7 4680. 104.0
2 11.3 95.7 24860. 53:1.2
3 11.5 113.7 18472. 394.7
4 10.0 127.7 2058. 44.8
5 2.,8 54.7 2168. 46.3
6 14.9 76.7 5121. 109.4
7 10.9 77.7 56148. 1199.7
6 14.3 91.7 31937. 692.5
9 19.5 185.7 16821. 359.4
CALCIRLfTE %. OH ENTRY #:
1Ba25 ML BAILER SAM + 180/25 INT9 18 SEPT 91 EM I 17067
IiJLDB-624 311 / .32 11M / 1.8 UI1 FILM







i--l' I I --I I I I I -l












36814
8671


















I I I
7 3 9 10 11 12 13, 14 15 16 17 18 19 20 21 22 23 24




Figure 5.3 Typical Ion Chromatograph



what compound or compounds are present at that particular


retention time.























Entry Statistics Hass/int Mass/Int Iass/Int Hiass/Int Hass/Int Hass/Int
Fil: LIN(N71B)
Entry; 1724 ? 76 92 75
Peaks: 16 51 167 183 50
Meo Ut: 16 52 70 185 173
63A. if 187 400
C8 H10. 65 7 107 33
m m
mini l I I ..... ._ ^ __ _


Figure 5.4 Typical Mass Spectra


Before running any samples the GC/MS must be calibrated.
This is done by running known compounds at known
concentrations through the system to obtain response factors.
Response factors are the actual known concentrations divided
by the peak area for the compound in question. These factors
are obtained by running a wide range of concentrations of the
particular compounds such as 2 ppb, 4ppb, 10 ppb, and 20 ppb.
These data are then averaged to give a response factor for
each compound to be analyzed.


J.-OZENE, 1,2-DIMETHn-,












Table 5.2 Primary Drinking Water Standards (MCLs)


Metals Ag/1

Arsenic 50
Barium 1000
Cadmium 10
Chromium 50
Lead 50
Mercury 2
Selenium 10

Volatile Organics

Vinyl Chloride 2
Trichloroethene 5
Benzene 5
Carbon Tetrachloride 5
1,2-Dichloroethene 5
1,l-Dichloroethene 7
1,1,l-Trichloroethane 200

Semivolatiles

1,4-Dichlorobenzene 75
2,4,5-Trichlorophenol 10

Pesticides/Herbicides

2,4-Dichlorophenoxyacetic acid 100
gamma-BHC 4
Methoxychlor 100
Toxaphene 5

Additional Parameters

Nitrate 10,000
Fluoride 4,000








Method 524 titled Measurement of Purgeable Organic

Compounds in Water by Capillary Column Gas Chromatography/Mass

Spectrometry covers a total of sixty compounds. The method of









70

detection limit (MDL) depending upon the type of column ranges

from .02 to .35 ppb. The method of detection limit is the

minimum concentration above zero that is detected 99% of the

time.

5.4 Regulatory Contaminant Levels

The Safe Drinking Water Act (SDWA) which was passed in

1974 sets the maximum contaminant levels (MCLs) for drinking

water. These levels are listed in Table 5.2. States may,

however, implement even more stringent requirements. The

Florida Department of Environmental Regulation (DER) has set

its own state ground water target levels. For closing an

underground storage facility in Florida the contaminant levels

must not exceed those listed in Table 5.3. Methyl tert-butyl

ether (MTBE) and ethylene dibromide (EDB) are fuel additives.

Ethylene dibromide is also used in soil fumigants. Neither of

these two compounds is listed in EPA Method 602.

EPA Method 610 is titled Polynuclear Aromatic

Hydrocarbons. This method covers sixteen organic compounds

that are associated with fuels other than gasoline such as

diesel, kerosene, jet fuel A, JP-4 (jet fuel), and No. 6

heating oil. This method requires a minimum sample size of

250 ml. EPA Method 625 titled Base/Neutrals and Acids covers

61 compounds. It includes all the compounds from method 610

plus several polychlorinated biphenyls (PCBs) and several

pesticides including DDT, aldrin, chlordane, toxaphene, and

dieldrin.











Table 5.3 Florida Ground Water Target Levels


tg/l
Gasoline (EPA Method 602)

Benzene 1

Total VOA 50
-Benzene
-Toluene
-Total Xylenes
-Ethylbenzene

Methyl Tert-Butyl 50
Ether (MTBE)


Kerosene/Diesel (EPA Method 610)

Polynuclear Aromatic Hydrocarbons (PAHS) 10





5.5 Solute Transport

Solutes (contaminants) migrate through soil due to three

processes: advection, diffusion, and dispersion. Advection

is contaminant flowing with the groundwater. Diffusion is the

process of spreading due to chemical gradients, i.e., moving

from a high concentration to a lower concentration. It can

take place when there is no flow of groundwater. Dispersion

is the spreading out of the contaminant longitudinally and

laterally due to velocity effects as it moves through the

tortuous paths through the soil pores. Water moves through

the center of pores faster than at the edges where it drags on

the soil particles.











The solute transport equation can be derived from the

continuity equation:



aM__ aJ
at ax



Where J, = Total solute flux

M-total mass-OC+pS
J,- -D C+qC
ax




and q = Darcy flux
D = Hydrodynamic Dispersion Coefficient (lumps
dispersion and diffusion together)
8 = Porosity (Volumetric Water Content when saturated)
p = Bulk Density of the Soil (ML-3)
S = Mass Adsorbed Solute/Mass of the Soil
C = Solution Concentration (ML-3)
Kd= Partition or Sorption Coefficient (LM-1)


+C _+p +s -p- [-OD +qC]
at 9t at 9t 9x ax




Assume: steady q
constant 8 (no change in water content)
constant p (no change in soil density)


ac 9s a C BC
8 +p =OD _-qa
at at a82 ax











For linear sorption S=KC

as_ ac
8t =dat



Substituting for S gives:


6t +pK- -D- -_q-
at atd X2 a x



Dividing through by 8 gives:


P+Kd aC -C q a
6 at ax2 E ax
pKJ
R-Retardation factor-l+ d

Vo-pore water velocity--




Substitution gives:

ac a2c ac
R -D ---V
at ax2 ax
V,-velocity of the solute- -
R



This shows that the contaminant will travel at a

velocity, Vs, which is slower than the velocity of water by a

factor R, the retardation factor. For any computer model

using the solute transport equation it is necessary to first

determine the partition coefficient, K,. This can be

determined by taking a soil sample from the field and










74
performing a laboratory test in which known contaminants at

known concentrations are passed through a column of the soil,

and measurements made of the concentrations in the effluent,

to determine how much was sorbed by the soil.

It is very important to realize the effect of diffusion

in this equation. For years landfills were designed just

considering advection. Contaminants were assumed to move just

with the water. Clay liners were built with a minimum

thickness of three feet and with a permeability of less than

10-7 cm/sec. The water velocity becomes negligible when the

permeability is small (v=ki). Eliminating this from the

transport equation shows that the contaminants will still move

through the clay barrier before the water will and can

contaminate the groundwater. Shackelford and Daniel (1991)

have found that in fine grained soils diffusion may be the

primary transport mechanism in solute transport. Contaminants

may show up below clay liners years earlier than predicted

from advection alone.

The retardation factor is a function of the sorption

between the contaminants and the soil particles. Sorption can

be due to ion exchange where higher valence cations replace

lower ones. It can also be due to the hydrophobic nature of

some contaminants which easily go out of solution. Clays can

retard contaminant migration due not only to their lower

permeability but also due to their negative charges which

allows for ion exchange unlike sand particles which have small









75

surfaces areas and no charge. Anions, negatively charged

ions, such as Cl-, NO,-, SO,-, can be repelled from clays and

will move with the water. Soils with a high percentage of

organic also retard many contaminants such as pesticides.

Acar and Haider (1990) give the partition coefficients

for several contaminants in some particular soils. Generally

the order of retardation for some contaminants from lowest to

highest is as follows: benzene, toluene, ethylbenzene, and

o-xylene.















CHAPTER 6
FIELD STUDIES--CAVALIER PRODUCTS BUILDING SITE



6.1 Introduction


Field work with the BAT groundwater monitoring system was

performed at the Cavalier Products Building (previously a

Shell gasoline station). The site is located at the

intersections of SW 4th Avenue and S. Main Street in

Gainesville, Florida. Figure 6.1 shows the site plan. The

site is under the jurisdiction of the Alachua County Office of

Environmental Protection which contracted with the Handex

Company to complete a contamination assessment study. The

Handex company installed several 2 inch monitoring wells on

the site and also downgradient of the site in Lynch Park.

Two types of tests were performed at this site: BAT probe

sampling inside existing monitoring wells and BAT probe

sampling within the soil adjacent to the monitoring wells.

The purpose of the testing was to show that the BAT

groundwater monitoring system could recover VOCs. By

collecting samples with the BAT probe inside the monitoring

well they could be compared directly to samples obtained with

the Teflon bailer from the same monitoring well. The BAT MK2


















N



LYNCH PARK


DIRECTION OF FLOW


S. MAIN STREET


Figure 6.1 Cavalier Site Plan

probe with steel and HDPE filters were used to evaluate which

filter type sorbed lower amounts of VOCs.

6.2 Field Test 1

Monitoring well MW-17 was purged by removing 3 well

volumes of water with a three foot long teflon bailer. A one

foot Teflon bailer was then used to obtain groundwater

samples. The water was decanted from the bailer into 40 mL

septum vials and stored in a cooler with ice packs. Four

samples were obtained (BAILER1....BAILER4). The depth to the


MW-16


3
N

4 TH




MW-14


C
A
V
A
L

E
R


* *


MW-15
*


MW-17
0




































,. 1: ::l i~


Figure 6.2 View of Cavalier Site--Lynch Park on Left



water table in the well was approximately 9.5 feet. The smell

of petroleum was prevalent when sampling.

BAT sampling was performed with the University of

Florida's 20 ton electric cone penetration test truck

positioned as closely as possible to the well. Since several

obstacles (trees, shrubs, etc.) were present the proximity was

severely limited. A BAT MK2 probe with a steel filter was

pushed to a depth of 4 meters (13.1 feet) to obtain









79

groundwater samples. The probe was located 11.5 feet

horizontally away from the well. After purging the probe,

three 35 mL test tubes were filled using the cascaded sampling

technique (BAT1, BAT3 and BAT5) which required more than

thirty minutes per sample. The rods were then pulled, the

truck moved slightly and a second penetration performed. This

consisted of pushing a BAT MK2 probe with a HDPE filter to a

depth of 3.5 meters (11.5 feet). This penetration was located

12.0 feet horizontally away from the well. One BAT sample

(BAT7) without headspace was obtained at this location.

Chemical analyses were performed with a GC/MS in

accordance with EPA Methods 524/624. Compounds specifically

analyzed for were BTEX, trichloroethene (TCE), 1,2

dibromoethane (EDB), methyl tert-butyl ether (MTBE), and

tetrachlorothene (PERC). Only BTEX was detected. Results are

presented in Table 6.1.



6.3 Field Test 2

Additional testing at Lynch Park (Cavalier Site) was

performed on 23 October 1991. The depth to the water table in

monitoring well MW-17 was determined by two methods. Using a

bailer, the depth to the water table was determined from

hearing it touch the water level and measuring this distance.

This gave a value of 9 feet and 6 inches. Using an electronic

device (Soiltest, Inc., Model DR-760A Water level indicator)

gave a water table depth of 9 feet and 11 inches.









80

The monitoring well was purged by removing three well

volumes with a large Teflon bailer (1 L). While purging the

well a strong hydrocarbon odor was quite prevalent. A small

Teflon bailer (350 mL) was used for sampling. Two samples

were pulled and two 40 mL vials were filled from each. These

samples were designated BAILER1, BAILER2, BAILER3 and BAILER4

(which was not analyzed). BAILER1 and BAILER2 were obtained

from the same bailer, as were BAILER3 and BAILER4. The

samples were quite cloudy.

Distilled water was pulled through the BAT probe and

placed in a 40 mL vial to serve as an equipment blank to

ensure that no contamination remained from previous testing.

The probe had been decontaminated with boiled distilled water

after its previous use. The BAT probe was then penetrated 10

feet and 4 inches horizontally away from MW-17 and to a depth

of 10.5 feet. This depth was chosen to ensure that the probe

was extremely close to the actual water level where most of

the BTEX compounds should be, since they are lighter than

water. Before sampling, the pore pressure device was used to

measure the pore water pressure at the BAT probe as another

check on the water level. The digital readout gave a value of

0.1 meter. This meant that the probe was basically .1 m below

the water table.

The BAT probe was purged twice before sampling. The

first BAT cascaded sampling (BAT1) took 60 minutes and did not

yield a full sample. A second cascaded sampling (BAT3), took



















Table 6.1 Chemical Analyses from MW-17 at Cavalier Site


Benzene Toluene
ppb ppb


Ethylbenzene Xylenes
ppb ppb


Water Table--9'-6"


BAILER1
BAILER2
BAILER3

AVG BAILER


7198
4410
6724

6110


3120
2606
3301

3009


7880
6555
8048

7494


BAT Depth--13'-2"
BAT MK2 probe with
from MW-17.


steel filter was located 11'-6"


BAT1
BAT3
BAT5


AVG BAT


BAT Depth--11'-6"
BAT MK2 probe with
from MW-17.


HDPE filter was located 12'


BAT 7


All BAT samples contained no headspace









82

40 minutes and also did not yield a full sample. A third

sample was collected at this depth using a single test tube

(BATS) and had approximately 80% headspace.

The probe was then pushed down another meter to a depth

of 13 feet and 1 inch. Water was purged from the probe before

sampling. Two samples were obtained at this depth (BAT7 and

BAT8). Each of these samples yielded only 1 mL of fluid.

Time for each sample was 25 minutes.

All samples were iced at the site and then transferred to

a refrigerator. No preservatives were used since the analysis

would be performed within 7 days.

Chemical analyses were performed the following day with

the GC/MS. In order to run these samples on the GC/MS, they

had to be highly diluted since normally 20-30 ppb is the

maximum level that should be run on this equipment. Some

dilutions were made by placing 0.5 mL of the actual sample

into 100 mL of deionized water. Other samples were more

highly diluted by placing 0.5 mL of sample into 200 mL of

deionized water. The concentrations of BTEX are given in

Table 6.2. Ethylene dibromide (EDB) and methyl tert-butyl

ether (MTBE) were not present in the groundwater samples.

Trimethylbenzene was found but not quantitated. Naphthalene

was present in both BAT and bailer samples. In the bailer

samples the concentrations of naphthalene for the four samples

were 479 ppb, 525 ppb, 599 ppb, and 672 ppb.