USGS
science for a changing world
United States
Environmental
Protection Agency
Air and Radiation
(6608J) Emergency and
Remedial Response
EPA402-C-00-001
November 2000
www.epa.gov
Field Demonstration Of Permeable
Reactive Barriers To Remove
Dissolved Uranium From
Groundwater, Fry Canyon. Utah
+ +
September 1997 through September 1998
Interim Report
Contaminant
plume
Trench (filled)
Chemical
barrier
Aquifer
-------
EPA 402-C-00-001
November 2000
FIELD DEMONSTRATION OF PERMEABLE
REACTIVE BARRIERS TO REMOVE
DISSOLVED URANIUM FROM
GROUNDWATER, FRY CANYON, UTAH
September 1997 through September 1998
Interim Report
Prepared in cooperation with the
U.S. Geological Survey
U.S. Bureau of Land Management
U.S. Department of Energy
U.S. Environmental Protection Agency
Office of Air and Radiation
Washington, DC
2000
-------
CONTRIBUTORS
This report was prepared under the supervision of the Fry Canyon Study Team.
PRINCIPAL AUTHORS
David L. Naftz, Ph.D.
U.S. Geological Survey
Edward M. Feltcorn
U.S. Environmental Protection Agency
Christopher C. Fuller
U.S. Geological Survey
Ronald G. Wilhelm
U.S. Environmental Protection Agency
James A. Davis, Ph.D.
U.S. Geological Survey
Stan J. Morrison, Ph.D.
*Environmental Sciences Laboratory
Geoffrey W. Freethey
U.S. Geological Survey
Michael J. Piana
U.S. Geological Survey
Ryan C. Rowland
U.S. Geological Survey
Julie E. Blue**
U.S. Environmental Protection Agency
* Operated by MACTEC-ERS for the U. S. Department of Energy Grand Junction Office
** Present affiliation: Oberlin College, Oberlin, Ohio
DISCLAIMER
The following is intended solely as guidance to Environmental Protection Agency
(EPA) and other environmental professionals. This document does not constitute
rulemaking by the Agency, and cannot be relied on to create a substantive or
procedural right enforceable by any party in litigation with United States. EPA may
take action that is at variance with the information, policies, and procedures in this
document and may change them at any time without public notice.
Reference herein to any specific trade names, products, process, or services does not
convey, and should not be interpreted as conveying, official EPA approval,
endorsement, or recommendation.
-------
ACKNOWLEDGEMENTS
The Project Officer of the Fry Canyon Permeable Reactive Barrier Demonstration Project, Ed
Feltcorn of EPA's Office of Radiation and Indoor Air's Center for Radiation Site Cleanup, would
like to acknowledge the support provided by the following individuals and organizations Without
their continued support, this project would not have been possible.
EPA Office of Emergency and Remedial Response, 3/8 Center.
EPA Office of Radiation and Indoor Air, Radiation Protection Division
Mr. Ronald Wilhelm of the Office of Radiation and Indoor Air and Mr. Randy
Breeden, at the time a member of EPA's Office of Emergency and Remedial Response.
From 1993 through 1995 they strongly advocated the use of PRBs for inorganic
contaminants. Their initial efforts created the basis for development of this
project.
EPA Office of Emergency and Remedial Response Project Coordinators: Randy
Breeden, Paula Estornell, Dr. Julie Blue and Kelly Madalinski.
Peer reviewers for this report: Dianne Marozas, DOE Albuquerque; Richard
Steimle, USEPA, Technology Innovation Office; and Dr. Ralph Ludwig, USEPA, ORD
Ada Lab.
US Geological Survey, Utah District Office lead by Dr. Dave Naftz who did everything he could
to ensure a successful project. His leadership and tireless efforts at resolving issues motivated the
entire project team to excellence.
The Project Officer would also like to note the achievements of the University of Waterloo,
Canada, in their work with Permeable Reactive Barriers and thank them for their cooperation.
in
-------
TABLE OF CONTENTS
ii CONTRIBUTORS
iii ACKNOWLEDGEMENTS
v LIST OF TABLES
vi LIST OF FIGURES
1.0 INTRODUCTION 1
1.1. Abstract 1
1.2. Background 1
1.3. Purpose and Scope 3
1.4. Project History 4
1.5. Technology Description 4
2.0 SITE CHARACTERIZATION AND REACTIVE MATERIAL SELECTION 4
2.1. Site Characterization 4
2.1.1.Hydrologic Characterization Results 4
2.1.2. Water Quality Characterization Results 7
2.2. Reactive Material Selection 15
2.2.1.Characteristics of Reactive Material 15
2.2.1.1. Evaluation of Phosphate Material for Use in Permeable Reactive Barrier
Demonstration 15
2.2.1.2. Evaluation of Zero-Valent Iron Material for Use in Permeable Reactive
Barrier Demonstration 16
2.2.1.3. Evaluation of Amorphous Ferric Oxyhydroxide Material for Use in Permeable
Reactive Barrier Demonstration 17
2.2.2. Laboratory Evaluation of Phosphate Materials 17
2.2.3.Laboratory Evaluation of Zero Valent Iron and Amorphous Ferric Oxyhydroxide Materials...27
2.2.4.Selection of Materials for Demonstration 31
2.3. Further Development of Barrier Materials 32
3.0 PERMEABLE REACTIVE BARRIER DESIGN 32
4.0 MONITORING NETWORK DESIGN 34
5.0 BARRIER AND MONITORING NETWORK INSTALLATION 41
6.0 POST-INSTALLATION SAMPLING AND ANALYSIS 46
7.0 YEAR ONE RESULTS OF PERMEABLE REACTIVE BARRIER DEMONSTRATION 48
7.1. Volume of Groundwater Treated 48
7.2. Changes in Uranium Concentration 51
7.3. Water-Quality Effects of Barrier Materials 57
8.0 COST ANALYSIS OF SITE CHARACTERIZATION, PRB DESIGN, AND PRB INSTALLATION 64
9.0 REMAINING QUESTIONS FOR RPMS DURING THE RI/FS 65
10.0 RECOMMENDATIONS FOR PRB IMPLEMENTATION: LESSONS LEARNED 66
APPENDICES
A. Summary of Activities at Other Sites Using Permeable Reactive Barriers to Remove Uranium 68
B. Summary of Deep Emplacement Methods for Permeable Reactive Barriers 72
C. Site Evaluation and Selection Process 75
D. Health and Safety Issues Associated with Installation of Permeable Reactive Barriers 78
E. Sample Collection, Analysis, Quality Assurance, and Field Measurement Calibration 79
F. Glossary of Terms 84
G. List of References 86
IV
-------
LIST OF TABLES
Table Title Paj
2.1 Physical properties, trace-element concentration, and measured discharge
at Fry Creek surface-water sites, Fry Canyon, Utah 14
2.2 Products tested during the laboratory investigation 16
2.3 Physical characteristics of phosphate rock and bone meal materials
evaluated 19
2.4 Physical properties of bone-char phosphate pellet formulations CP3 and
CP5 20
2.5 Summary of column results for uranium breakthrough and uptake by
different phosphate materials for 12 milligrams per liter dissolved uranium
in pH 7 artificial groundwater 25
2.6 Hydraulic conductivity values measured by constant head method reported
in feet per day 30
6.1 Sampling period, chemical constituents, and number of samples taken
during the first year of permeable reactive barrier operation at Fry Canyon,
Utah 47
7.1 Percentage of input uranium concentration removed after traveling
approximately 1.5-feet into each of the permeable reactive barriers during
September 1997 through September 1998, Fry Canyon, Utah 55
8.1 Actual cost and duration of project planning through installation of three
permeable reactive barriers at the Fry Canyon site, Utah 64
A.1 Other permeable reactive barrierfield projects 69
D.1 Permeable reactive barrier chemicals used at the Fry Canyon site and the
associated toxicity characteristics 78
E.1 Uranium concentration and pH, specific conductance, and oxidation-
reduction potential value changes during pumping of well ZVIFS1 during
September 1998, Fry Canyon, Utah 79
E.2 Chemical analysis of selected major-, minor-, and trace-element
constituents from blank samples processed during pre-installation and
year 1 barrier monitoring activities, Fry Canyon, Utah, September 1996 to
September 1998 81
E.3 Chemical analysis of selected major-, minor-, and trace-element
constituents from duplicate samples collected during pre-installation
characterization and year 1 barrier monitoring activities, Fry Canyon, Utah,
September 1996 to September 1998 82
-------
LIST OF FIGURES
:igure Title PC
1.1 Schematic diagram of permeable reactive barrier 2
1.2 Location of the Fry Canyon demonstration site in southeastern Utah 3
2.1 Generalized profile of geologic units exposed in the Fry Canyon area 5
2.2 Potentiometric surface of the colluvial aquifer during October 1996, Fry
Canyon, Utah 6
2.3 Conceptualization of ground-water movement in the colluvial aquifer at the
Fry Canyon study site 7
2.4 Pre-installation ground- and surface-water sampling sites and
potentiometric surface of the colluvial aquifer during October 1996, Fry
Canyon, Utah 8
2.5 Pie charts comparing the pre-installation major-ion and uranium
concentration in milliequivalents per liter from 7 groundwater and 2
surface-water sites, Fry Canyon, Utah 9
2.6 Pre-installation trace-element concentrations measured in groundwater
samples collected during December 1996, February 1997, and April 1997.. 10
2.7 Comparison of uranium concentration in filtered and unfiltered water
samples collected during site characterization, December 1996, Fry
Canyon, Utah 11
2.8 Uranium concentrations (unfiltered) during a 1-hour pumping cycle at well
FC3, September 1996 11
2.9 Total uranium concentrations in subsurface sediment samples collected
during September 1996, Fry Canyon,Utah 12
2.10 Uranium desorption results using samples collected from the colluvial
aquifer at Fry Canyon during September 1996 13
2.11 Electron micrograph of bone-char phosphate pellets (CP5) illustrating
morphology and porosity of the material 21
2.12 Batch uranium uptake on various phosphate materials at pH 7 in Fry
Canyon artificial groundwater 23
2.13 Breakthrough of uranium in bone-char column plotted as the ratio of
column effluent dissolved uranium to influent dissolved uranium versus
number of column pore volumes passed 24
2.14 Release of uranium from bone char column plotted as the ratio of column
effluent dissolved uranium to initial influent dissolved uranium versus
number of column pore volumes passed 26
2.15 Uranium concentrations in 3-day batch tests on a variety of reactive
materials 28
2.16 Uranium concentrations in effluents from column experiments using
amorphous ferric oxyhydroxide/sand (0.2 weight percent iron as
amorphous ferric oxyhydroxide) and zero valent iron foam pellets/sand (55
weight percent zero valent iron) 29
2.17 Testing hydraulic conductivity of amorphous ferric oxyhydroxide /gravel
mixture 30
VI
-------
3.1 Schematic diagram showing the funnel and gate design used for the
installation of PRBs at Fry Canyon, Utah 33
3.2 Three-dimensional views of the sacrificial frame design used for
installation of PRBs at Fry Canyon, Utah 34
4.1 Schematic diagram showing monitoring well placement and sample site
identification for the bone-char PRB 35
4.2 Schematic diagram showing monitoring well placement and sample site
identification for the zero-valent iron PRB 36
4.3 Schematic diagram showing monitoring well placement and sample site
identification for the amorphous ferric oxyhydroxide PRB 37
4.4 Two-inch monitoring well, Fry Canyon, Utah 38
4.5 Pressure transducer deployed during PRB demonstration, Fry Canyon,
Utah 39
4.6 Schematic diagram of the automatic data recording system within and
adjacent to the permeable reactive barriers, Fry Canyon, Utah 40
4.7 Automated data logging equipment used during the Fry Canyon barrier
demonstration project 41
5.1 Trench box used to protect workers during installation of the PRBs at Fry
Canyon, Utah 43
5.2 Placement of AFO barrier material into the gate structure of the permeable
reactive barrier, Fry Canyon, Utah 44
5.3 Location and dimensions of permeable reactive barriers after construction,
Fry Canyon, Utah 46
7.1 Configuration and altitude of the potentiometric surface of the colluvial
aquifer at Fry Canyon, Utah prior to the permeable reactive walls being
installed in (a) July 1997, and (b) January 1997 49
7.2 Configuration and altitude of the potentiometric surface in the colluvial
aquifer at Fry Canyon, Utah, December 1998, and the approximate area of
aquifer influenced by the permeable reactive barriers 51
7.3 Changes in dissolved uranium concentrations in the bone char phosphate
permeable reactive barrier from September 1997 through September
1998, Fry Canyon, Utah 52
7.4 Changes in dissolved uranium concentrations in the zero valent iron
permeable reactive barrier from September 1997 through September
1998, Fry Canyon, Utah 53
7.5 Changes in dissolved uranium concentrations in the amorphous ferric
oxyhydroxide permeable reactive barrier from September 1997 through
September 1998, Fry Canyon, Utah 54
7.6 Variation in pH and percent uranium removal from September 1997
through Septeber 1998 in two monitoring points completed in the
amorphous ferric oxyhydroxide barrier, Fry Canyon, Utah 56
7.7 Comparison of uranium concentration in filtered and unfiltered water
samples collected after installation of permeable reactive barriers, October
1997, Fry Canyon, Utah 57
Vll
-------
7.8 Changes in water temperature within the bone char phosphate, zero valent
iron, and amorphous iron oxyhydroxide permeable reactive barriers from
September 1997 through September 1998, Fry Canyon, Utah 58
7.9 Changes in pH values within the bone char phosphate, zero valent iron
and amorphous iron oxyhydroxide permeable reactive barriers and
background well FC3 from September 1997 through September 1998, Fry
Canyon, Utah 59
7.10 Changes in dissolved oxygen concentration from within the bone char
phosphate, zero valent iron, and amorphous iron oxyhydroxide permeable
reactive barriers from September 1997 through September 1998, Fry
Canyon, Utah 60
7.11 Changes in oxidation reduction potential from within the bone char
phosphate, zero valent iron, and amorphous iron oxyhydroxide permeable
reactive barriers from September 1997 through September 1998, Fry
Canyon, Utah 60
7.12 Changes in ferrous iron concentration in water samples from upgradient,
within-barrier, and downgradient wells in the zero-valent iron permeable
reactive barrier from September 1997 through September 1998, Fry
Canyon, Utah 61
7.13 Changes in ferrous iron concentration in water samples from upgradient,
within-barrier, and downgradient wells in the amorphous ferric
oxyhydroxide permeable reactive barrier from September 1997 through
September 1998, Fry Canyon, Utah 62
7.14 Changes in phosphate concentration in water samples from upgradient,
within-barrier, and downgradient wells in the bone char permeable reactive
barrier from September 1997 through September 1998, Fry Canyon,
Utah 63
C.1 Location and uranium concentration in water samples from abandoned
mine sites considered for field demonstration of permeable reactive
barriers 76
Vlll
-------
1.0 INTRODUCTION
1.1 Abstract
The Fry Canyon site in southeastern Utah was selected in 1996 as a long-term field
demonstration site to assess the performance of selected permeable reactive barriers for the
removal of uranium (U) from groundwater. Permeable reactive barriers (PRBs) are permeable
walls that are installed across the flow path of a contaminant plume. The wall is designed to be at
least as permeable as the surrounding aquifer material. The PRBs contain a zone of reactive
material that is designed to act as a passive in-situ treatment zone for specific contaminants as
groundwater flows through it.
The use of PRBs for remediating organic-contaminated groundwater is fairly well documented.
This project demonstrates one of the first uses of PRBs for remediating uranium contaminated
groundwater. The U concentrations measured in groundwater at the Fry Canyon site prior to
PRB installation were as high as 16,300 micrograms per liter (|lg/L) with a median concentration
of840|ig/L.
A series of laboratory experiments were conducted on three classes of potential PRB materials
(phosphate, zero valent iron, and ferric iron) to determine uranium removal efficiencies and
hydrologic properties. A PRB material from each class was selected for field demonstration. The
selected materials had suitable hydraulic conductivity, high U removal efficiency, and high
compaction strengths.
A funnel and gate design was used with wing walls on each end of the structure to channel the
groundwater into the PRBs. Each gate structure was 3 feet (ft) thick and 7 ft wide. Depths of
barrier materials varied from 3.2 to 3.7 ft. Sixteen monitoring wells were located along two
parallel flow paths in each PRB to evaluate short-term changes in water quality.
During the first year of operation (September 1997 through September 1998), the PRBs removed
most of the incoming U. The zero-valent iron (ZVI) PRB has consistently lowered the input U
concentration by more than 99.9 percent after the contaminated groundwater had traveled 1.5 ft
into the PRB. The percentage of U removed in the bone-char phosphate (PO4) and amorphous
ferric oxyhydroxide (AFO) PRBs exceeded 70 percent for most measurements made during the
first year of operation. The U concentrations in monitoring wells downgradient of the PRBs are
at or near background concentrations. This project has demonstrated that PRBs are an efficient
and financially viable means of remediating uranium contaminated groundwater. Because
mechanisms similar to those which remove uranium in PRBs are also responsible for the removal
of other inorganic contaminants, the results of this project have wide applicability.
1.2 Background
Potable groundwater supplies worldwide are contaminated or threatened by advancing plumes
containing radionuclides and metals. Surface drainage from abandoned and inactive mines has
percolated into underlying aquifers and contaminated groundwater with uranium (U), radium
(Ra), molybdenum (Mo), arsenic (As), selenium (Se), chromium (Cr), and vanadium (V), as well
as other radionuclides and metals.
The problem of the U migration from inactive and abandoned mines and tailings piles is not
limited to U ores. Because of the enrichment of U in ores of other metals and phosphate
-------
deposits, high U concentrations have been found in mine drainage from hardrock base- and
precious-metal mines, as well as industrial mineral mines. In the case of metal mines, U
concentrations on the order of several thousand |lg/L have been found, generally with higher U
concentrations in more acidic drainages. In addition to uranium, concentrations of associated
metals have also been found at levels exceeding drinking-water standards. In addition to mining
and milling operations (at Department of Energy (DOE) and Department of Defense (DOD)
facilities and elsewhere), industrial activities such as machining, plating, and manufacturing have
resulted in groundwater contamination. Also, some waste repositories are currently leaking
contaminants into the underlying groundwater. Repository liners and caps designed to prevent
infiltration of precipitation are failing due to a variety of perturbations, including differential
settling and bio-intrusion. Therefore, viable approaches to dealing with such problems may have
widespread applicability.
Currently, the most widely used method of groundwater remediation is the combination of
extraction, ex-situ treatment, and discharge of the treated water known as pump and treat.
However, pump-and-treat methods are costly and often ineffective in meeting long-term
protection standards (Travis and Doty, 1990; Gillham and Burris, 1992; National Research
Council, 1994). Permeable reactive barriers (PRBs) offer a low cost alternative to these methods.
PRBs are permanent, semi-permanent, or replaceable units that are installed across the flow path
of a contaminant plume (Feltcorn and Breeden, 1997). The PRBs contain a zone of reactive
material that acts as passive in-situ treatment zones that degrade or immobilize contaminants,
such as radionuclides, as groundwater flows through them (fig. 1.1).
Contaminant
plume
Trench (filled)
Chemical
barrier
Aquifer
Figure 1.1. Schematic diagram of permeable reactive barrier.
The impetus for the development of innovative treatment technologies is based on federal law
and policy. Under the Superfund Amendments and Reauthorization Act of 1986 (SARA),
USEPA is required to select remedial actions involving treatment that "permanently and
significantly reduces the volume, toxicity, or mobility of the hazardous substances, pollutants,
and contaminants" [Comprehensive Environmental Response, Compensation, and Liability Act
(CERCLA), Section 121(b)]. Furthermore, "EPA expects to consider using innovative technology
-------
when such technology offers the potential for comparable or superior treatment performance or
implementability, fewer or lesser adverse impacts than other available approaches, or lower costs
for similar levels of performance than demonstrated technologies" [National Oil and Hazardous
Substances Pollution Contingency Plan (NCP), 40 CFR Part 300.430 (a) (1) (ii) (E)]. This field
demonstration project develops and tests an innovative use of an existing treatment technology.
The results provide valuable information to decision makers regarding the use of this technology
at existing Superfund sites.
1.3 Purpose and Scope
The overall objective of this project is to demonstrate the use of PRBs to control the migration of
radionuclides and other metals in groundwater. Three PRBs were installed in September 1997 at
Fry Canyon (fig. 1.2). The purpose of this report is to summarize the experimental work leading
to the design of the Fry Canyon experimental installation of three PRBs and the first year of
treatability study results. This report summarizes preliminary laboratory work and first year
field results. Based on these results, Chapter 10 provides a summary of recommendations
for PRB implementation at sites contaminated with radionuclide and trace-metal
contamination in groundwater. This report details information on the Fry Canyon site
characterization, reactive material selection, PRB design and construction, operation and
maintenance, and technology performance. This report is intended for use by Remedial Project
Managers (RPMs), EPA regional technical support staff, contractors, stakeholders, technology
vendors, and others tasked with remediation sites contaminated with radionuclides.
114°
42--.I
t
I
1
! UTAH
!
I / I/Installing monitoring wells at the Fry Canyon
j <#f Y demonstration site.
f
i
|
..J 1 1—.-^isf.—..^..—
Lake Powell
0 20 40 60 KILOMETERS
20 40 60 MILES
Figure 1.2. Location of the Fry Canyon demonstration site in southeastern Utah.
-------
1.4 Project History
Laboratory research conducted prior to the Fry Canyon proposal indicated the likelihood that
PRBs would be effective for treating groundwater contaminated by uranium. This research
included: (1) laboratory batch testing comparing the effectiveness of a wide variety of reactive
materials to remove U from groundwater (Morrison and Spangler, 1992); (2) batch and column
testing confirming the efficiency of amorphous ferric oxyhydroxide (AFO) to sorb U (Morrison
and Spangler, 1993; Morrison and others, 1995).
A reconnaissance stage investigation to evaluate potential sites for treatability study of PRBs
was conducted in August 1995. The four candidate sites were located in southeastern Utah
(Appendix C, fig. C.I).
1.5 Technology Description
PRBs show promise as an inexpensive and effective remediation technique for cleaning up
radionuclide-contaminated groundwater that commonly exists near numerous abandoned mill
tailing piles throughout the Western United States. Operational and maintenance costs are
significantly lower than pump-and-treat methods. Reactions within the wall either degrade
contaminants to non-toxic forms or transfer the contaminants to an immobile phase.
The use of reactive chemical walls for inorganic groundwater contaminants including metals and
radionuclides has received less attention. The development of new forms of reactive materials
and rapid increases in the number of treatability studies and field experiments performed are
demonstrating the viability of this technology for inorganics.
PRBs are best suited for sites that have well defined flow paths. It is preferable to have an
impermeable layer to key the wall into, but hanging walls can be designed that capture plumes
without a bottom layer. PRBs have previously been installed using conventional trenching
technologies at depths of no more than 45 feet (ft) below ground surface (BGS). New
emplacement methods (grouting, fracing, driven mandrels, injection, etc...) are being investigated
that can extend this depth range. Appendix B contains a summary of these other methods.
2.0 SITE CHARACTERIZATION AND REACTIVE MATERIAL SELECTION
2.1 Site Characterization
Candidate sites for PRB remediation must undergo a thorough geochemical and hydrologic
characterization to ensure proper system design. Based on the site selection procedure for this
demonstration project (Appendix C), Fry Canyon was chosen in June 1996 for initial site
characterization activities. The objective of the initial site characterization at Fry Canyon was to:
(1) determine the type and amount of groundwater contamination in the shallow colluvial aquifer
beneath the tailings and (2) ensure that hydrologic and geochemical conditions were conducive for
the installation and demonstration of PRBs. Site characterization activities were conducted at the
site from September 1996 through June 1997.
2.1.1 Hydrologic Characterization Results
Characterization of the hydrogeologic conditions at Fry Canyon was based on previous geologic
studies, site reconnaissance, drilling, field measurements and testing, and laboratory analyses.
Near the study site, which lies in a sedimentary stream valley, colluvial deposits remain as
paleochannel deposits beneath and west of the existing stream channel. Results from drilling 9
test holes showed that these deposits are as thick as 18 ft with as much as the lower 5 ft
saturated. The elevation of the sandstone bedrock surface under these deposits varies, but in
-------
places it is lower than the adjacent rock stream channel (fig. 2.1). Analyses of cores collected
during drilling indicate the deposits consist of silt to gravel-size particles derived from the
sandstone and shale formations that exist upgradient and upslope from the site.
o
o
Moss Back Mr
Fry Canyon
,Shinarump Mr
Moenkopi Fm
Hoskinnini Mr
Geology adapted from Thaden, Trites, and Finnell, 1964
Figure 2.1. Generalized profile of geologic units exposed in the Fry Canyon area.
Hydrologic properties were estimated from field and laboratory measurements. Specific-capacity
measurements and slug tests on wells indicate that hydraulic conductivity values for the aquifer
are probably in the range of 5 to 50 feet per day (ft/d). Hydraulic-conductivity values measured
in the laboratory on disturbed samples ranged from 55 to 85 ft/d. Saturated thickness in the
vicinity of the PRBs ranges from 2 to 4 ft; thus, transmissivity values for the aquifer probably
range from 10 to 200 ft /d. The porosity of a repacked drilling sample, as measured in the
laboratory, is 12.6%. Porosity values from the literature indicate the in-situ effective porosity is
probably greater—20 to 25% for mixed sand, gravel, and silt (Freethey, Spangler, and Monheiser,
1994).
The stream channel deposits being used for the demonstration are limited in extent vertically and
laterally by the Permian Cedar Mesa Sandstone. The hydraulic conductivity of the unfractured
sandstone measured in the laboratory by Jobin (1962) was about 0.003 ft/d or about 1,000 times
smaller than the hydraulic conductivity of the channel deposits. Wells tapping the deep saturated
zone of the Cedar Mesa Sandstone yield small to moderate amounts of water, implying that the
hydraulic conductivity of the saturated sandstone is 1 or 2 orders of magnitude larger, probably
because the calcareous cement has been partially dissolved by groundwater. This indicates that
-------
the contact between the sandstone and the colluvial deposits is probably an impediment to flow
in the shallow groundwater system. The lateral edge of the aquifer is its deposit!onal limit where
it contacts the Cedar Mesa Sandstone. The existence of Fry Springs indicates that the sandstone
more readily transmits groundwater through bedding-plane fractures, and it is thus possible that
some groundwater could move between the sandstone and the channel deposits.
Water-level elevations in 6 wells and the elevations of Fry Creek adjacent to the aquifer provided
data to construct an initial potentiometric contour map of the study site (fig. 2.2). The contour
map shows that the aquifer is recharged by subsurface inflow from Fry Creek upstream of the
site, by precipitation directly on the site, and by runoff from the sandstone upslope from the
site. Additional recharge could be coming from lateral subsurface inflow at the contact between
the sandstone and the channel deposits. Groundwater discharges from the aquifer by seeping
back into Fry Creek, by evaporation where the saturated sediments are near land surface, by
riparian vegetation (Tamarisk) transpiration and possibly downward leakage into the sandstone.
Discharge measurements in Fry Creek in November 1997, indicate that about 10 to 15 acre-feet
per year (ac-ft/yr) of groundwater seep into the stream along this 300-ft reach. Figure 2.3
illustrates how water moves into and out of this shallow aquifer system.
Explanation
— 5,358— Potentiometric contour—Shows altitude at which
water level would have stood in tightly cased wells, October 1996.
Interval is 0.2 foot
5,359.56
• Control point—Number is altitude of water level, in feet
Potentiometric contours
October 1996
Figure 2.2. Potentiometric surface of the colluvial aquifer during October 1996, Fry Canyon, Utah.
-------
N
Figure 2.3. Conceptualization of ground-water movement in the colluvial aquifer at the Fry Canyon
study site.
2.1.2 Water Quality Characterization Results
Pre-installation ground- and surface-water quality were determined to establish baseline values
that could be compared to post-installation water-quality and geochemical data. The pre-
installation data were also used during design of the PRBs. Water quality and geochemical data
were collected from 7 wells and 2 surface-water sites located on Fry Creek (fig. 2.4). Samples
were collected during September 1996, December 1996, and April 1997 to determine seasonal
variability prior to installation of the PRBs. A description of sample processing, analytical
methods, and quality assurance results are presented and discussed in Appendix E.
-------
FRYCRK3
(located 350 feet downstream from
symbol)
N
5,356
EXPLANATION
. Potentiometric contour—Shows altitude at which
water level would have stood in tightly cased wells in
he colluvial aquifer.
Contour!nterval 0.2 feet. Datum is sea level.
Ground-water site and designation.
Surface-water site and designation
FRYCRK2
FC1 (located 1,200 feet southest from symbol)
Figure 2.4. Pre-installation ground- and surface-water sampling sites and potentiometric surface of
the colluvial aquifer during October 1996, Fry Canyon, Utah.
Pie charts were used to compare the major-ion chemistry, in milliequivalents per liter (meq/L),
between the sample sites (fig. 2.5). The background well (FC1, U concentration = 60 |ig/L) and
the surface-water sites (FRYCRK2 and FRYCRK3, U concentration = 60 and 140 |ig/L) have
similar major-ion chemistry. The dominant cation is sodium. Bicarbonate and sulfate are the
dominant anions (fig. 2.5). The similarity between the major ion chemistry at the background and
surface-water sites verifies the strong surface-water/groundwater interaction in the colluvial
aquifer at the site. Non-background wells in the colluvial aquifer that contain U concentrations of
less than 600 micrograms per liter (|lg/L) have major-ion chemistry similar to the background and
surface-water sites (fig. 2.5). As the U concentration increases above 800 milligrams per liter
(mg/L), the proportion of calcium and sulfate increases and the proportion of bicarbonate and
sodium decreases. Water from wells FC3 and FC7 contained the largest U concentrations
measured during the pre-installation characterization and are calcium sulfate water types.
-------
Uranium = 60
FC1
April 1997 so
HC03
^^Mg
Uranium = 1,020|j,g/L
Uranium = 2,800
FC3
April 1997
HC03
Uranium = 840 |j,g/L
FC4
December 199 7
-— ^^^Mg
/Ca
Uranium =110 jig/L
FC5
April 1997
^^Mg
Uranium = 16,300
FC7
April 1997
Uranium = 550 jig/L
April 1997 ^^^^ so
Uranium = 60
FRYCRK2
April 1997
/
HC03 /
Mg
Uranium = 140
FRYCRK3
April 1997
HCO3
Figure 2.5. Pie charts comparing the pre-installation major-ion and uranium concentration in
milliequivalents per liter from 7 groundwater and 2 surface-water sites, Fry Canyon, Utah.
Box plots were used to display the concentration ranges of U and selected trace elements in
groundwater samples during the pre-installation sampling periods (fig. 2.6). Elevated iron and
manganese were detected in site groundwater samples. The median iron concentration was 90
|ig/L and the median manganese concentration was 180 |ig/L (fig. 2.6). The measurable iron and
manganese concentrations indicate slightly reducing conditions in the colluvial aquifer. For
example, the oxidation reduction potential measured in well FC3 during September 1996 was
-12 millivolts (relative to the silver-silver chloride, platinum electrode system) with a
corresponding dissolved oxygen concentration of 0.6 mg/L.
-------
cc
LU
10,000
o E 1'°°°
z < 100
LLJ o:
o o
o cc
o o 10
DT = Lower analytical
reporting limit
EXPLANATION
Maximum
99th precentile
95th precentile
75th precentile
Median
-I - 25th precentile
— 5th precentile
O - 1 st precentile
A - Minimum
CHEMICAL CONSTITUENT
Figure 2.6. Pre-installation trace-element concentrations measured in groundwater samples
collected during December 1996, February 1997, and April 1997.
The median copper concentration was below the analytical reporting limit of 4 |ig/L and the
median zinc concentration was at the reporting limit of 10 |ig/L. The U concentrations were
elevated at the site, ranging from 60 to 16,300 |ig/L with a median concentration of 840 |ig/L.
The U concentrations were determined in filtered, 0.45 micrometer (|im) and unfiltered water
samples collected during December 1996 to document concentration differences. Because drinking
water supplies are not typically filtered, it is important to document that sample filtration is not
biasing the U concentration data. It is also important to document the filtered and unfiltered U
concentration prior to PRB emplacement to determine if PRBs change the U distribution between
filtered and unfiltered fractions.
To prevent clogging of analytical instruments the unfiltered samples were acidified with nitric
acid to a pH of less than 2.0 units. After a 24-hour (hr) period, the acidified water sample was
filtered and analyzed for U concentration. It is probable that this procedure will mobilize the U
associated with particulates in the sample prior to the filtration step.
Comparison of the filtered (0.45 |im) and unfiltered samples (fig. 2.7) indicates no difference in
the U concentration between the filtered and unfiltered samples. The slight differences that are
observed are all within plus or minus 10%, which is the analytical uncertainty of the analytical
method.
10
-------
10,000 -
DC
LU
Oj
5;
1,000 -
o
- 100 -
^H Filtered (0.45 micrometers)
I I Unfiltered
SAMPLE SITE
Figure 2.7. Comparison of uranium concentration in filtered and unfiltered water samples
collected during site characterization, December 1996, Fry Canyon, Utah.
Short-term fluctuations in U concentration in groundwater were measured during September 1996
to document the existence of a stable contaminant source. Well FC3 was pumped for about a 1-
hour (hr) sampling period while unfiltered water samples were collected at one-gallon intervals.
Analysis of these samples for U concentration indicated a stable U concentration, consistently
above 3,000 |ig/L (fig. 2.8).
z" 3,700
o t
I- -i 3,600
DC
3,500
3,400
3,300
3,200
3,100
jjjE
il
0<
0
_ 12345
VOLUME PUMPED, IN GALLONS
Figure 2.8. Uranium concentrations (unfiltered) during a 1-hour pumping cycle at well FC3,
September 1996.
The measured pH of the groundwater at Fry Canyon was considered an important attribute in
PRB design because changes in pH can affect numerous geochemical reactions. For example,
under certain geochemical conditions, increases in pH can cause desorption of U from
11
-------
contaminated sediments or precipitation of carbonate mineral phases. The pH values in 20
groundwater samples were near neutral during the pre-installation monitoring period. The pH
values ranged from 6.9 to 7.7 and the median value was 7.3 units.
The U concentrations in sediment samples from the Fry Canyon site were determined prior to
barrier installation to evaluate the potential for re-release of U from sediments downgradient of
the PRBs. The PRBs were installed within the contaminant plume at this site for several reasons
including efficiency of the installation process at this particular site. This project was not aimed
at remediating the site but at assessing the effectiveness of PRBs in removing U. Obviously, at a
remedial site it would be preferable to locate the PRBs downgradient of any contaminant plume.
At this site, however, U desorbing from the aquifer material downgradient of the PRBs and
upgradient of the monitoring wells had to be accounted for. Total U in subsurface samples of
saturated colluvial material ranged from 2.95 parts per million (ppm) at well FC1 (background
site) to 21.2 ppm at well FC3 (fig. 2.9). Total U concentrations were higher in the unsaturated
colluvium samples than in the saturated samples collected from FC1 (6.6 ppm) and FC5 (59
ppm), indicating that rainwater has percolated through the tailings on the surface and
concentrated significant quantities of U within the unsaturated colluvium.
5,356-
2.95
[6.60]
EXPLANATION
Potent!ometric contour—Shows altitude at which water level would
have stood in tightly cased wells in the colluvial aquifer
Contour interval 0.2 feet Datum is sea level
Ground-water site and designation—Number indicates uranium
concentration in sediment sample collected from saturated zone.
Number in brackets indicates uranium concentration in sediment
sample collected from unsaturated zone
70 feet
FC1
2.95
[6.60]
Figure 2.9. Total uranium concentrations in subsurface sediment samples collected during
September 1996, Fry Canyon, Utah. Concentration expressed in parts per million. Site FC1 is
located offsite, approximately 0.2 miles to the southeast.
Desorption experiments indicate that sediments from the contaminated part of the colluvial
aquifer at the Fry Canyon site contain a large amount of U that can be readily desorbed (fig.
2.10). Using sample FC3 as an example, 2.8 ppm out of the 21.2 ppm of the total U is readily
desorbed. Making the following assumptions: (1) porosity equals 40%; (2) density of the solid
12
-------
phase equals approximately 2.7 grams per cubic centimeter (g/cm ); and (3) density of
groundwater equals 1.00 g/cm ; there would be about 4,000 grams (g) of sediment per liter of
water resulting in 11.3 milligrams (mg) of desorbable U compared with the measured U
concentration of 3.8 mg per liter of groundwater collected from well FC3. The desorption results
indicate that water exiting the PRBs will initially desorb significant amounts of U from the
contaminated sediments. This would not occur in a remediation application because the PRB
would likely be placed in non-contaminated sediments downgradient of the contaminant source
term.
DC
LU
CO
LU
FC5
FC4
FC3
FC2
FC1
| | Potential uranium concentration
from sediment desorption
Uranium concentration in
ground water
8
10
12
URANIUM CONCENTRATION,
IN PARTS PER MILLION
Figure 2.10. Uranium desorption results using samples collected from the colluvial aquifer
at Fry Canyon during September 1996.
Two surface-water sampling sites were established on Fry Creek (fig. 2.4) to define pre-
installation water quality and quantity on Fry Creek. Site FRYCRK2 was located upstream of
the potential site for PRB installation. Site FRYCRK3 was located approximately 350 ft
downstream of the upgrader tailings. The pH of the water samples from Fry Creek range from
8.4 to 8.6 units (table 2.1), which is more than 1 pH unit higher than the median pH of the
shallow groundwater at the site. Concentrations of most trace elements in water samples from
Fry Creek are less than the analytical reporting limit. The U concentrations in Fry Creek were
significantly lower than in the groundwater samples and ranged from 60 to 140 |lg/L during the
pre-installation monitoring phase. Although Fry Creek is perennial throughout the study area, the
discharge was very low, ranging from 0.025 to 0.043 cubic feet per second (cfs).
13
-------
Table 2.1. Physical properties, trace-element concentration, and measured discharge at Fry Creek
surface-water sites, Fry Canyon, Utah.
[pH, in units; uS/cm, microsiemens per centimeter at 25 degrees Celsius; mg/L, milligrams per liter;
ug/L,micrograms per liter; <, less than reported value; cfs, cubic feet per second; Al, aluminum; Cu, copper;
Fe, iron; Li, lithium; Mn, manganese; P, phosphorus; Si, silicon; Zn, zinc; U, uranium]
Surface-
water site Date Time
FRYCRK2 12/18/96 1520
FRYCRK3 12/18/96 1605
FRYCRK2 04/09/971330
FRYCRK3 04/09/97 1245
PH,
field,
in
units
8.6
8.4
8.6
8.6
Specific
conduc-
tance, in
uS/cm
2,440
2,110
1,850
1,865
Al,
in
ug/L
<50
<50
<50
<50
Cu,
in
ug/L
<4
<4
<4
<4
Fe,
in
ug/L
<20
<20
<20
<20
Li,
in
ug/L
2.1
1.8
1.5
1.5
Mn,
in
ug/L
20
20
<10
<10
P, Si,
in in
ug/L ug/L
<0.1 8.5
<0.1 6.9
<0.1 6.2
<0.1 5.9
Sr,
in
ug/L
1.8
1.5
1.4
1.4
U,
in
ug/L
90
140
60
140
Zn,
in
ug/L
<10
<10
<1 0
15
Discharge
in cfs
0.036
0.043
0.025
0.040
-------
2.2 Reactive Material Selection
Prior to work on the Fry Canyon Project, numerous materials had been tested in laboratory
experiments for their ability to remove U and other inorganic contaminants from groundwater.
Some of the results of these investigations are presented in Spangler and Morrison (1991),
Blowes and Ptacek (1992), Morrison and Spangler (1992), Morrison and Spangler (1993),
Kaplan et al. (1994), Morrison et al. (1995), Bostick et al. (1996), and Morrison et al. (1996).
Based on these studies, three groups of reactive materials were selected for consideration: (1)
phosphate, (2) zero-valent iron (ZVI), and (3) ferric iron (AFO). These materials are believed to
remove U by the following mechanisms: (1) phosphate - precipitation of an insoluble uranyl
phosphate phase, (2) ZVI - reduction of U to +4 oxidation state and subsequent precipitation,
(3) AFO - by adsorption to the iron oxyhydroxide surface. Additional laboratory investigations
were conducted to select a specific reactive material from each of these three groups to be used at
Fry Canyon. Factors considered in selecting the materials included: (1) availability; (2) cost; (3)
more permeable than surrounding aquifer material; (4) structural strength (resistance to
compactive crushing when placed in the ground); (5) extent, rate, and duration of U removal; (6)
mobility (i.e. the tendency for the material to move with the groundwater; e.g. the tendency for
AFO to form mobile colloids); (7) potential for re-release of uranium; and (8) possible
detrimental effects on groundwater quality such as pH change or release of iron or phosphate.
2.2.1 Characteristics of Reactive Material
2.2.1.1 Evaluation of Phosphate Material for Use in Permeable Reactive Barrier
Demonstration
Hydroxyapatite and other apatite minerals have been found effective in immobilizing lead and
other metals through the formation of metal phosphates that are insoluble over a range of
chemical conditions (Ma and others, 1994a, b; Zhang and others, 1997). Recently, additions of
hydroxyapatite to U contaminated sediments were shown to decrease U solubility (Arey and
others, 1999). The hypothesis for the use of natural apatites to remove dissolved U from
groundwater is that they provide a source of phosphate with which aqueous U(VI) should react
to form insoluble uranyl phosphates, such as hydrogen, calcium, magnesium, potassium or
sodium autunite (e.g. [Ca(UO2)2(PO4)2*10H2O]) (Sowder et al, 1996; Sandino and Bruno, 1992;
Arey and other, 1999). The effectiveness of removal of aqueous U by commercially available
natural apatite materials (phosphate rock, bone meal, and bone meal charcoal) was determined in
laboratory batch uptake and column experiments. The results of the laboratory evaluation were
used for choosing specific phosphate material for the field demonstration. Other criteria for
choice of material for use in PRBs include the extent of reversibility of U removal, permeability
of the reactive material, release of solutes detrimental to water quality.
Batch uptake experiments with synthetic, reagent grade hydroxyapatite indicated that 15 mg/L of
U(VI) was completely removed by 6.7 gram per liter (g/L) of solid. Because the cost of this
material was prohibitive for use in field applications, lower cost, commercially available natural
apatite materials were evaluated. Phosphate rock samples were obtained from mining companies
in Florida, North Carolina, and Utah. The mined rock had been separated from accessory
minerals and crushed to fine-sand or silt grain size prior to receipt. The phosphate bearing
minerals in phosphate rock are fluoroapatite and carbonated fluoroapatite. Fertilizer-grade bone
meal and bone charcoal also were obtained for testing because the inorganic component in bone is
primarily hydroxyapatite.
15
-------
The results for both the batch and column experiments described below indicated that bone-meal
phosphate materials were much more effective for U removal from groundwater. However, use
of these materials in field-scale demonstration would require dilution with an inert coarse
material to obtain adequate barrier permeability. Instead, a pelletized bone charcoal (2-mm
diameter) was also evaluated for use in the reactive barrier. Pelletized bone charcoal was
produced by firing fertilize-grade bone meal mixed with aluminum and phosphate binders at
1100° C in the absence of air. Several formulations of bone-char pellets were evaluated which
included pellets fired in air to remove carbon and pellets with an iron binder. Cercona of America
manufactured all the formulations of bone-char pellets evaluated. The cost of the Cercona of
American bone-char pellets was about $65 per cubic foot (table 2.2).
2.2.1.2 Evaluation of Zero Valent Iron Material for Use in Permeable Reactive
Barrier Demonstration
ZVI is a scrap product available principally from the automotive industry. Several companies
broker the scrap ZVI and sieve it to customers' specifications. Usually, the brokers also heat
treat the ZVI by roasting it in an oven at temperatures of about 1,200 degrees C. Heating
removes cutting oils that may have been present. ZVI products tested in this investigation are
listed in Table 2.2.
Table 2.2. Products tested during the laboratory investigation. Products 2, 12, and 15 were
selected for the Fry Canyon Demonstration project. Bulk prices may be less.
[ZVI, zero valent iron; AFO, amorphous ferric oxyhydroxide; vol, volume; —, information not
available]
1
2
3
4
5
6
7
8
9
10
11
12
15
Vendor
Cercona
Cercona
Cercona
Cercona
Cercona
Cercona
Cercona
Connelly GPM
Connelly GPM
Connelly GPM
Master Builder
Cercona
Noah Ind.
Product
Name
Cast iron
Foam Pellets
Foam Aggregate
MV Pellets
MV Foam Block
ZVI/Zeolite
ZVI/Magnetite
CC-987
CC-1004
CC-1010
GX027
Bone char pellets
AFO slurry
Sieve
Size
-8+50
-3+50
-3+50
--
--
--
--
-18+40
-8+50
-18+60
-8+50
-3+20
-5+18
Cost per
cubic foot
$30
$40
$35
--
--
--
--
$30
$30
$30
$30
$55
$40
Description
ZVI filings
Foamed aluminosilicate-bound pellets
Foamed aluminosilicate aggregate
Magnetite/ZVI (mixed valent) pellets
Magnetite/ZVI (mixed valent) foam block
ZVI + 25% zeolite in pellets
ZVI + 20% Fe3Od in pellets
ZVI filings
ZVI filings
ZVI filings
ZVI filings
Charred bone meal with aluminophosphate binder
AFO slurry mixed with gravel (1 :2 vol/vol)
Products 1, 8, 9, 10, and 11 are ZVI filings that have been sieved and heat treated by the brokers.
The other products have been custom processed by Cercona. To produce product 2 (Foam
Pellets), Cercona pelletizes fine-grained ZVT with an aluminosilicate binder. A foaming agent
(aluminum powder which produces hydrogen gas) is added during the pelletizing process. The
pellets are then fired which increases their strength. The process results in low-density, high
strength, porous pellets with greater than 90% ZVI. The manufacturing of product 3 (Foam
Aggregate) is similar to the Foam Pellets except that the foamed aluminosilicate bonded ZVI is
not rolled into pellets and is not fired. This product is less costly than the Foam Pellets but does
not have as high structural strength and is more likely to be crushed when buried. The
aluminosilicate binder can be used to produce other shapes such as Foam Blocks (product 5).
16
-------
Products 4, 5, 6, and 7 were manufactured using the aluminosilicate binder but used a variety of
raw materials.
The sieve size that has been used most often in PRBs is -8 +50. That is, the material passes
through an 8-mesh sieve (2 mm) but does not pass through a 50-mesh sieve (0.3 mm). The cost
of the raw material screened to -8 +50 is about $350 per ton. Equivalent weights of Cercona
customized products are more expensive. However, because their density is much less, the cost
per unit volume is similar to unprocessed ZVI filings. The cost of ZVI foam pellets is about $40
per cubic foot which is only slightly higher than the $30 per cubic foot cost of unprocessed ZVI
filings (Morrison and others, 1998).
2.2.1.3 Evaluation of Amorphous Ferric Oxyhydroxide Material for Use in
Permeable Reactive Barrier Demonstration
AFO is prepared by rapid hydrolysis of a ferric salt solution. Ferric chloride (FeCl3) is the least
expensive form of ferric salt. Hydrolysis is accomplished by adding a base such as sodium
hydroxide (NaOH) to a solution of ferric chloride. The ferric chloride solution is acidic and the
hydrolysis reaction is exothermic. The temperature of the solution increases as NaOH is added.
A temperature increase to about 50°C causes substantial amounts of goethite (FeOOH) to form.
Goethite is more crystalline than AFO and less effective in stabilizing uranium.
AFO is not available commercially in a form that can be used in a PRB. The raw materials (ferric
chloride and sodium hydroxide) can be purchased commercially at a cost of about $625 per ton of
AFO [as Fe(OH)3]. Material costs could be reduced by using a locally available base such as
limestone in lieu of sodium hydroxide. For a large job it may be cost-effective to mix the ferric
chloride and base at the site. AFO is also available as a food-grade slurry at a cost of about
$23,000 per ton of AFO. Although a lower grade of AFO would be acceptable for use in PRBs,
no vendors produce and market a lower grade. The slurry has to be mixed with sand or gravel to
use in a PRB. At the Fry Canyon demonstration site, food-grade AFO slurry (2% Fe) was mixed
with pea gravel. The cost of this mixture was about $40 per cubic foot.
2.2.2 Laboratory Evaluation of Phosphate Materials
Physical characterization. Phosphate materials were characterized for release of phosphate
as a function of time in batch experiments identical to U uptake experiments except without the
addition of dissolved uranium. Grain-size distribution of phosphate materials was determined by
dry sieving if not provided by the vendor. Specific surface areas of phosphate materials were
determined by single-point nitrogen BET measurement. Bone-char pellet solid density was
determined by water displacement. Internal porosity of bone-char pellets was determined by
weight loss during drying of water-saturated pellets. The permeability of bone-char pellet
formulation chosen for the PRB demonstration was determined from the flow rate of water
through packed columns as a function of hydraulic head using a modified Marriot bottle
apparatus. Permeability, expressed as hydraulic conductivity (K ft/day), was calculated from the
dependence of flow rate on hydraulic head using Darcy's law and corrected for conductivity of
the apparatus.
Laboratory batch uptake experiments. Batch uptake experiments provide a test of the relative
efficiency of each material for U removal under similar solution conditions. In addition, the time
dependence of the removal process can be evaluated in the absence of transport limitations.
Uptake of U was measured as a function of time and total U concentration. The phosphate
materials were used as received. Solid concentrations of 10 and 100 grams solid per liter of
17
-------
artificial groundwater (AGW) were used for fine-grain and coarse-grain phosphate materials,
respectively. Batch U(VI) uptake experiments were conducted in pH 7 AGW of similar major
ion composition to Fry Canyon groundwater (alkalinity 4.8 meq/L, ionic strength 0.04 molar
(M)). In this groundwater, dissolved U(VI) speciation is dominated by uranyl carbonate species.
The solid was equilibrated with the AGW in either 50- or 250-mL centrifuge bottles prior to
addition of dissolved uranium. Dissolved U (U(VI)) stock solutions were prepared from reagent
uranyl nitrate and added to batch experiments to yield total U concentrations of 2.4 to 48 mg/L.
Sample pH was measured after addition of U and adjusted to pH 7 with dilute acid or base as
needed. Samples were then equilibrated from 1 to 96 hrs. The pH was measured and samples
centrifuged at 16,000 Gravitational Units (G) for 10 minutes. The supernatant was sampled for
dissolved uranium, cations and phosphorous (P) analyses. Initial experiments used a U tracer
added to the U(VI) stock solution and U concentration was determined by liquid scintillation
OQQ
counting (LSC) of the U alpha decay. Dissolved U was determined in subsequent experiments
by kinetic phosphorescence analysis (KPA). LSC had a precision of ±2% and 5 microgram per
OQQ
liter detection limit for U labeled 2.4 mg/L initial total U. The extent of U uptake was
determined by the change in dissolved U concentration after equilibration with the phosphate
solid divided by the solid concentration.
Column experiments. Column experiments were used to determine the volume of uranium-
contaminated groundwater that can be passed prior to U(VI) breakthrough. In addition, column
tests provide a means to determine the total capacity of phosphate materials for U removal under
concentrations and flow-rates expected in the field application.
Glass columns of 1-centimeter (cm) inside diameter fitted with 20-|im end-cap screens were used
for bone meal and phosphate rock experiments. Bone-char pellets were packed dry into 2.5-cm
diameter columns. Column outlets were fitted with 0.45-mm syringe filters. Initial tests of
columns packed with bone meal failed because of clogging either from bacterial growth or loss of
integrity. The small grain size of some phosphate materials resulted in too low a column
permeability for experimental tests. Because of the small grain size and/or clogging, both the
phosphate rock and bone meal were diluted ten-fold by weight with coarse sand (20 to 30 mesh)
to achieve adequate permeability. The sand had negligible U uptake in a batch experiment.
Mixtures of phosphate solid and sand were prepared by weight, wetted with AGW, and packed
into the column in ten to fifteen increments. Packed columns were saturated with AGW by slow
upward flow. Column test with bone charcoal was not feasible due to mobility of this very fine-
grained (3 to 15 |im diameter) material out of the column.
The larger particle size of the bone-char pellets enabled packing into column without dilution
with sand. Packed columns were flushed with carbon dioxide (CO2). Deionized water was then
passed to dissolve the CO2. This method was effective in eliminating gas pockets.
Column pore-volume was calculated from the weight of the packing material and total column
volume. The internal porosity was assumed negligible for the phosphate rock, bone meal and
Ottawa sand. Internal porosity of bone-char pellets (see below) was included in calculation of
column pore volume.
Artificial groundwater with 12 mg U/L was passed through the columns at 10 mL/hour using
gravity feed and upward flow. The AGW feed reservoir was continuously equilibrated with a 2%
CO2 in air to maintain constant pH of 7±. 1 pH units. Volume and flow rate of AGW passed was
18
-------
determined from the mass of each effluent sample. Column-effluent dissolved U was measured
by either LSC or KPA. Cation and dissolved P in column effluents were measured by inductively
coupled plasma optical emission spectrometry (ICP/OES). Effluent pH was measured
periodically using a flow-cell electrode. Total U uptake per gram of phosphate material was
calculated from the difference between total U input and outflow, divided by mass of phosphate
material in the column. The influent was changed to U-free artificial groundwater after U-
breakthrough to test the reversibility of U uptake.
Physical and chemical characteristics. Grain size and N2-BET surface area of the various
phosphate materials tested are given in Table 2.3 along with abbreviation for each material. The
grain size was used in determining the solid concentration used in initial batch experiments. A
solid concentration of 10 g/L was used for the fine-grained materials and 100 g/L for the coarser
materials. Surface area for the B-2 bone meal was very low in comparison to the other materials
perhaps because of its large particle size. XRD patterns of bone charcoal (BK1) and bone char
pellets (CP3) are dominated by hydroxyapatite peaks.
Table 2.3. Physical characteristics of phosphate rock and bone meal materials evaluated. Grain
size expressed as percentages by weight of diameter in microns (|im).
[<, less than; m2/g, square meters per gram; mg P/L, milligrams phosphorous per liter; |im,
microns;%, percent]
Material
Phosphate
Rock
Phosphate
Rock
Phosphate
Rock
Bone meal,
cooked
Bone meal,
steamed
Bone Charcoal
ID
SF
PCS
CF
B1
B2
BK
Source
SF Phosphates, LTD
Vernal, Utah mine
PCS Phosphates
Raleigh, North
Carolina
CF Industries
Plant City, Florida
Fertilizer company
Dale Alley Company
St Joseph, Missouri
Fertilizer company
Dale Alley Company
St Joseph, Missouri
EM Scientific
Specific
surface
area
(m2/g)
4.2
14.4
12.2
6.6
0.3
64
Grain-size
75% 125 to 1000 urn
25% < 125 urn
95% 125 to 1000 urn
5%<125 urn
51% 125 to 1000 urn
47% < 125 urn
55% >500 urn
45% <500 urn
1 00% < 63 urn
3-15 urn
Phosphate
concentration
(mg P/L)*
0.9
14
0.8
1.0
17
5.3
# Phosphate concentration in batch experiment after 24-hour equilibration with artificial ground
water in the absence of uranium.
Bone-char pellets are 80% by weight or greater 2-4 mm diameter with larger particles comprising
most of the remainder (table 2.4). The internal volume or porosity of the pellets was 0.6 to 0.8
cm3 per gram of pellets. Scanning electron-microscope (SEM) images illustrate the highly porous
structure of the pellets (fig. 2.11) which likely account for the high surface area of this material
19
-------
(table 2.4) compared to expected surface area of spheres of the size range of these pellets. Pores
range from a few |im in diameter to upwards of 100 |im. SEM images also indicate presence of a
fine-grained material of 0.5 and 5 |im in diameter. Only Ca and P were detected in these fine-
grained particles using energy dispersive X-ray analysis. These elements were also detected on
pellet surface in areas devoid of these fine-grained particles. Smaller amounts of aluminum from
the binder were also detected.
Table 2.4. Physical properties of bone-char phosphate pellet formulations CP3 and CP5. CP5
used in PRB field demonstration.
[%, percent; mm, millimeters; ft/day, feet per day; m2/g, square meters per gram; cm3/g, cubic
centimeters per gram; mg P/L, milligrams phosphorous per liter; cm/sec, centimeters per second]
Property
Grain-size
% >4 mm
% 2-4 mm
% 1 -2 mm
%<1 mm
Hydraulic Conductivity
Hydraulic Conductivity
Specific Surface Area
Inter-particle Porosity
Intra-particle Volume
Intra-particle Porosity
Phosphate Concentration *
CP-3
7.5
65.3
23.7
3.5
0.85
2400
33
20
0.83
55
0.9
CP-5
13.8
42.9
39.4
3.8
0.55
1560
44
19
0.59
51
7.8
UNITS
percent
percent
percent
percent
cm/sec
ft/day
m2/g
percent
cm3/g pellets
percent
mgP/L
# Phosphate concentration in batch experiment after 24-hour equilibration with artificial ground
water in the absence of uranium.
20
-------
Figure 2.11. Electron micrograph of bone-char phosphate pellets (CP5) illustrating morphology
and porosity of the material.
The effective porosity determined from H breakthrough agreed within 5% of the calculated
column pore volume when the internal volume of the pellets was included. This comparison
indicates that there is rapid exchange of groundwater between the inter-particle porosity and the
21
-------
bone-char internal porosity. The measured hydraulic conductivity of the CP3 and CP5 bone-char
pellet formulations greatly exceeded estimates of aquifer hydraulic conductivity (table 2.4).
Dissolved phosphate ([PO4]) released into AGW from apatite solids was measured in batch
experiments in the absence of dissolved uranium. After 24-hr reaction period, the bone meals and
bone charcoal had greater [PO4] concentrations than phosphate rock with the exception of the
PCS phosphate rock (table 2.4). The CP3 bone-char pellets phosphate release was about a factor
often lower than the CP5 formulation of pellets (table 2.4).
Results of batch and column U uptake experiments. The U uptake reached steady state
within 24 hrs for all solids tested. A 48-hr equilibration time was used to compare uptake among
the different phosphate materials. On a per mass basis, the bone meal and bone charcoal removed
1.5 to 2 orders of magnitude more U than the crushed phosphate rock (fig. 2.12). The U uptake
by bone-char pellets (CP3 and CP5) was about a factor ten lower than the bone meal and bone
charcoal. The bone-char pellets with an iron binder had similar U uptake as the CP3 and CP5
pellets but caused the pH to increase to 9.4 or higher. The bone meal pellets fired in air had
about 40% lower U uptake compared to pellets fired in the absence of air (CP3 and 5). The
uptake of U in batch experiments was independent of the amount of [PO4] released from the
solid prior to addition of uranium.
22
-------
1.5
1.0
0.5
0.0
-0.5
!=! -1.0
<3o
_!
lug
li
^UJ
-1.5
^ Bone Meal B1
T Bone Meal B2
O Bone-Black Charcoal BK
O Bone-char Pellets CP3
D Bone-char Pellets CP5
A Phosphate Rock SF
• Phosphate Rock PCS
A Phosphate Rock CF
0
0
0
-3-2-10 1 2
LOG DISSOLVED URANIUM, MILLIGRAMS PER LITER
-1
-3
0 Phosphate Rock SF
A Phosphate RockPCS
| Phosphate Rock CF
^ Bonemeal B1
^ Bone meal B2
O Bone Charcoal BK
A Bone-char Pellets CP5
O Bone-char Pellets CP3
O
O
V
-3-2-1012
LOG DISSOLVED URANIUM, MILLIGRAMS PER LITER
Figure 2.12. Batch uranium (U) uptake on various phosphate materials at pH 7 in Fry Canyon
artificial groundwater. The U uptake (log milligrams (mg) U per gram (g) phosphate solid) versus
equilibrium dissolved U concentration (log dissolved U in mg/L).
The absolute reactivity of solid phases for removal of solutes is best compared on a site density
basis if the reaction process occurs at the solid surface, instead of on a per mass basis. The N2-
BET surface area was used as an indicator of surface site density by normalizing U uptake to a
23
-------
rj
per surface area basis (mg U/m ). Normalizing uptake per unit area resulted in much smaller
differences among all the materials except the bone meals, which had significantly greater U
uptake per m (fig. 2.12). It is unclear why the fertilizer-grade bone meals had greater U uptake
9 9
per m . With the exception of the bone meal data, the smaller range in U uptake per m is
consistent with a removal process occurring on the surface of the apatite solids. However, for
evaluation of potential PRB materials uptake per mass is a better indicator since the size
(thickness) of the PRB limits the mass of reactive material that can be used in the barrier.
In column tests at 12 mg/L dissolved U influent, the phosphate rock materials reached 50%
breakthrough rapidly (within 7 to 18 pore volumes). Maximum uptake of U ranged from 0.16 to
0.4 mg U per gram of phosphate rock. In contrast, 216 pore volumes were required to reach 50%
breakthrough with bone meal. A maximum U uptake of 7.7-mg U per gram of bone meal Bl was
observed. The column with B2 bone meal diluted with sand was terminated due to clogging after
155 pore volumes. No significant U breakthrough was measured to this point. Clogging was
likely due to bacterial growth.
In a column packed with the CP3 bone-char phosphate pellets (undiluted), 50% breakthrough
occurred after 100 pore volumes (fig. 2.13). Complete breakthrough (100%) occurred at 190 pore
volumes of 12 mg U/L in AGW. A maximum uptake of 1.4-mg U per gram pellets was observed.
The column test with bone char pellets with iron binder clogged within ten pore volumes due to
oxidation and cementing of pellets in the upgradient end of the column. The bone meal pellets
fired in air (CP2) had complete U breakthrough within 20 pore volumes. Table 2.5 summarizes
the breakthrough and maximum uptake for the different phosphate materials tested.
<
cc
LU LLJ
li
S o
Li. S
O ^
o -z.
P" <
< DC
<
a
_i
OQ
o<
HOC
o<
ODC
0.0
50
100 150
COLUMN PORE VOLUMES
200
Figure 2.13. Breakthrough of uranium in bone-char column plotted as the ratio of column
effluent dissolved uranium to influent dissolved uranium versus number of column pore volumes
passed. The cumulative uptake of uranium (mg uranium/g solid) versus column pore volumes is
also calculated and assumes uranium uptake is uniform by the entire mass of solid within the
column.
24
-------
Table 2.5. Summary of column results for uranium breakthrough and uptake by different
phosphate materials for 12 milligrams per liter dissolved uranium in pH 7 artificial groundwater.
All phosphate materials were diluted ten-fold with sand except bone-char pellets.
Material
Pore volumes to
50% breakthrough
Uranium uptake at
100% breakthrough
mg U per gram solid
Phosphate Rock
SF
PCS
CF
7
11
18
0.16
0.22
0.41
Bone meal
B1
B2*
216
155
7.7
4.6
Bone-char pellets®
CP3
CP5#
100
250
1.4
0.7
* column terminated at 20% breakthrough due to clogging
® bone-char pellets packed undiluted
# 2.4 mg/L dissolved uranium influent
There was little change in the pH of groundwater passing through any of the columns packed
with any of the phosphate materials tested except the bone char pellets with ZVI binder (CP1)
which had an effluent pH of 9 or higher. Little change in major ion concentrations was observed
in column effluents.
Phosphate release from the CP3 bone-char pellet column was initially 3.5 mg P/L and decreased
to about 0.7 mg/L after 37 pore volumes. Phosphate concentrations remained relatively constant
at 0.68 ± 0.11 mg P/L throughout the uptake to 100% U breakthrough and subsequently during
the U-release test using U-free AGW. Similar levels of [PO4] release were observed in the first
100 pore volumes of U-free AGW passed through a column of CP3 pellets. These results suggest
that U removal is not dependent on release of phosphate to solution, that is U removal is not
occurring by precipitation directly from groundwater, but instead is removed by some process
occurring at the solid surface. In addition, the release of P from the PRB would be sustained at
levels that might be detrimental to aquatic systems downgradient if no subsequent processes
removed P from groundwater. At Fry Canyon, removal of [PO4] by sorption to iron coatings on
aquifer sediments is expected to remove most [PO4] prior to discharge into Fry Creek.
Characterization of potential for release of U from barrier materials using column
experiments. Release of U from phosphate materials was initiated after 100% breakthrough by
passing U-free artificial groundwater through the column. Greater than 65% of the U uptake was
released from phosphate rock columns within 20 to 30 pore volumes of U-free AGW passed. In
contrast, a slower release of U was observed from bone meal columns with less than 30% of total
U uptake released after 350 pore volumes. An initial release of about 40% of the total U uptake
(1.4 mg U/g) from the CP3 bone-char pellet column occurred over the first 75 pore volumes of
elution (fig. 2.14). At that point the effluent U concentration had decreased to 2.4 mg U/L. A
25
-------
slower release of U continued with 1,000 pore volumes required to reach U effluent concentration
equal to 1% (0.12 mg/L) of the initial influent concentration. At this point, 80% release of the U
from the solid had occurred. The concentrations of U in column effluents in release experiments
were in excess of the proposed drinking limit of 20 |ig/L. The extent of release measured
indicates that in the field, U may be released to groundwater after complete breakthrough if
uranium-free groundwater subsequently entered the barrier. However, about half of the U release
would occur over a long period of time.
150
0.0
400
650 900
COLUMN PORE VOLUMES
1,150
Figure 2.14. Release of uranium from bone char column plotted as the ratio of column effluent
dissolved uranium to initial influent dissolved uranium versus number of column pore volumes
passed.
Possible U removal processes. The removal of dissolved U is postulated to occur by formation
of a uranyl phosphate phase, in part because of the low solubilities of uranyl phosphates such as
various forms of autunite (Ca, Mg, K, Na, or H uranyl phosphate). However, because there are
many uranyl phosphate phases and the solubility constants for some of these phases are not well
defined (Grenthe and others, 1992), calculation of the degree of saturation for these phases can
not be used to determine the precipitation of a specific phase. Instead, spectroscopic and X-ray
diffraction techniques were used to characterize the process of U uptake. In addition, it is
unclear whether U removal occurs by precipitation directly from solution or by reaction at the
apatite surface. Knowledge of the process of U uptake by phosphate materials is needed for
modeling U transport in the PRB.
Extended X-ray absorption fine-structure spectroscopy (EXAFS) was used to characterize the
process(es) of U uptake by apatite on a molecular scale. The U-edge EXAFS spectra are
characteristic of the local bonding environment of uranium. The spectra are used for phase
identification by comparing with spectra of phases of known structure. In addition, the distance
and coordination number of nearest and next-nearest neighboring atoms to U can be derived from
26
-------
sample spectra for comparison to distance and coordination of likely bonding environments.
Synchrotron-source X-ray diffraction (XRD) also was used to identify the presence of
crystalline uranyl phases in the reacted apatite materials. This technique provides significantly
greater angular resolution and sensitivity than standard laboratory XRD instruments.
The U-LIII X-ray absorption edge positions of reacted bone meal, bone charcoal, and bone-char
pellets indicate that U remains in the +6 oxidation state instead of being reduced. The U-edge
EXAFS of U reacted with bone-char pellets, bone charcoal, and bone meal at total U
concentrations to 5,500 ppm indicate a different bonding environment than observed for uranyl
phosphates autunite [Ca(UO2)2(PO4)2*10H2O], meta-ankoleite [K2(UO2)2(PO4)2*6H2O],
saleeite [Mg(UO2)2(PO4)2*10H2O] or for schoepite [UO2(OH)2*2H2O] and several uranyl
carbonate mineral specimens. However, the EXAFS spectra of hydroxyapatite with U uptake of
greater than 7,000 ppm were similar to autunite.
Synchrotron-source XRD patterns of these samples also did not indicate presence of any known
uranyl phosphate phase. Crystalline U(VI)-phosphate solids, autunite and chernikovite
[(H3O)2(UO2)2(PO4)2*6H2O] were present only in XRD patterns of reagent-grade
hydroxyapatite with U uptake concentrations of greater than 7,000 ppm. Detection limits of
about 350 ppm and 2,000 ppm U, were determined for autunite and chernikovite, respectively.
No evidence for these crystalline precipitates were observed in any of the samples prepared from
bone-char materials, which had U(VI) solid concentrations ranging from 800 to 5,500 ppm.
These results, in combination with EXAFS measurements, suggest that the predominant U(VI)-
removal process is complexation by phosphate in the apatite surface (e.g. adsorption) at uptake
levels of column experiments or expected in the field demonstration.
2.2.3 Laboratory Evaluation of Zero Valent Iron and Amorphous Ferric
Oxyhydroxide Materials
Issues addressed in the laboratory investigations include (Spangler, 1997): 1) development of a
suitable form of AFO for use in a PRB, 2) efficiency of U removal by Z VI and AFO from Fry
Canyon groundwater, 3) hydraulic conductivity, and 4) mobility of the reactive materials in
groundwater. Four types of experiments were performed: (1) tests to determine suitable
mixtures of AFO slurry and gravel, (2) batch tests of contaminant uptake, (3) column tests of
contaminant uptake, and (4) hydraulic conductivity measurements. Batch and column
experiments used groundwater (FRGW) from well FR3 at the Fry Canyon site. FRGW had the
following composition: Ca 270 mg/L, Na 286 mg/L, Mg 84 mg/L, Mn 0.292 mg/L, K 5.12 mg/L,
Fe 0.0104 mg/L, Sr 1.75 mg/L, Cl 108 mg/L, PO4 <0.086 mg/L, SO4 987 mg/L, U 2.09 mg/L,
alkalinity 411.0 mg/L as CaCO3, and pH 7.2 to 7.6.
Laboratory batch uptake experiments. Batch tests were performed by agitating the reactive
material with 35 mL of FRGW in an end-over-end rotator. After agitation for 3 days, the
mixtures were centrifuged to remove particles with greater than 2 |im diameters. The U
concentrations were measured on the decanted fluids.
Three-day batch tests were conducted on a variety of ZVI-based materials and two AFO
mixtures. The results indicated that with the exception of 3 materials (Cercona ZVI/zeolite,
Cercona mixed valent pellets, and Cercona mixed valent foam block), all reactive materials had
potential for significant U removal (fig. 2.15).
27
-------
2,000 —
LU
(f>
CC CC
H UJ
Z Q.
m tn
O £
Figure 2.15. Uranium concentrations in 3-day batch tests on a variety of reactive materials.
Experiments with amorphous ferric oxyhydroxide contained the indicated amount amorphous
ferric oxyhydroxide with 35 milliliters of Fry Canyon groundwater. All other experiments
contained 0.5 g of reactive material and 35 milliliters of Fry Canyon groundwater. Initial uranium
concentration was 2,090 |ig/L.
Selection of materials for additional investigation in column experiments was based on: (1) results
presented in Figure 2.15, (2) data from published literature, (3) cost, (4) hydraulic conductivity,
and (5) the objectives of the Fry Canyon project. Two materials were selected: AFO-coated
gravel and ZVI foam pellets. The ZVI foam pellets have excellent structural strength, high
hydraulic conductivity, and removed significant amounts of U from solution. AFO was capable
of removing a significant amount of U from solution. Because it is very fine grained and would
have low hydraulic conductivity in a PRB, it had to be mixed with gravel to achieve a suitable
hydraulic conductivity. AFO slurry-to-gravel ratios were determined by mixing various
proportions in a 3-gallon bowl and observing the consistency of the mixture. The maximum
amount of AFO slurry that could be used without the mixture becoming too "soupy" was about
2 weight% Fe as AFO.
Column experiments. Column tests were conducted in 10-cm inside diameter (ID) by 26.5 cm
long Plexiglas columns. The ZVI foam pellets were mixed with sand in an attempt to achieve
greater hydraulic conductivity. The column contained 2,152.9 g of Cercona ZVI pellets and
1,775.7 g of sand. The AFO/sand column contained 7.70 g of AFO and 3,048 g of sand. (This
column contained only about 0.2 weight% Fe as AFO. Tests performed later showed that a
mixture containing up to 2 weight% Fe as AFO with gravel was suitably permeable.) FRGW
was pumped through the columns from bottom to top at a rate of 0.5 mL/min for the first 5
days. A pump rate of 0.5 mL/min produced a linear flow rate of 1.5 ft/day, which is
approximately the groundwater flow rate in the colluvial aquifer at the Fry Canyon site. After 5
28
-------
days, the flow rate was increased to 3.0 mL/min to simulate the increased flow rate that would
occur if in the PRBs emplaced at Fry Canyon using a funnel and gate system.
The U concentrations in the effluents from the AFO/sand column remained below detection
(0.5 |ig/L) for about 5 L of effluent and then increased steadily until the experiment was
terminated at 7.2 L (fig. 2.16). The U concentrations in the effluents from the ZVI foam
pellets/sand column remained below detection throughout the experiment.
DC DC
Q I- UJ
Uj Z 0-
o
-------
Table 2.6. Hydraulic conductivity values measured by constant head method reported in feet
per day.
[AFO, amorphous ferric oxyhydroxide; Fe, iron; %, percent; ZVI, zero valent iron]
Material
Sand
AFO/Sand (0.40% Fe as AFO)
AFO/Sand (2.27% Fe as AFO)
AFO/Sand (0.25% Fe as AFO)
ZVI/Sand (42% Fe)
ZVI (Master Builder -8 +50)
ZVI Foam Pellets/Sand (55% Fe)
Gravel (3/8 inch)
AFO/Gravel (0.4% Fe as AFO)
Fry Canyon Colluvium
Fry Canyon Alluvial Aquifer
Hydraulic Conductivity
1,417.5
680.4
7.088
241.0
822.2
198.5 to 283.5
396.9 to 822.2
4,167
3,204
56.7 to 85.0
39.69 to 65.20
Hydraulic conductivity of AFO/gravel mixture with 2 weight Fe% (as AFO) was determined in a
Plexiglas tank (fig. 2.17). Hydraulic conductivity was calculated from the head difference across
the tank produced by a given flow rate. The head difference was too small to measure even for
the maximum pump rate, thus the hydraulic conductivity is greater than 2,835 ft/d. The
hydraulic conductivity measured in the tank experiment was significantly higher than that
measured for an AFO/gravel mixture by the constant head column method. The difference may
result from the different flow orientations in the two tests. The column test measures vertical
hydraulic conductivity whereas the tank test measures horizontal conductivity. Horizontal flow
is more analogous to field conditions.
Figure 2.17. Testing hydraulic conductivity of amorphous ferric oxyhydroxide/gravel mixture.
30
-------
Possible U removal processes. Release of U back to solution was not evaluated during this
study for ZVI or AFO materials. The following discussion of potential re-release of U is based
on theoretical considerations. U(VI) adsorbed to the outer surfaces of AFO particles will
eventually desorb and reenter the groundwater. If the upgradient groundwater becomes clean
(such as could occur if the contaminant source is removed), the U concentration front will
continue to migrate through the reactive medium until it reaches the downgradient edge. Once the
front is at that position, U will be released to the environment at concentrations controlled by the
chemistry of the incoming groundwater. If incoming clean groundwater has the same pH and
major-ion chemistry, then the concentration of U in the outflow will be the same as it was in the
contaminated groundwater.
If the PRB containing ZVI were to become depleted in Fe or if the reactive surfaces became
coated with mineral precipitates, oxidized groundwater could re-mobilize the U minerals.
Uraninite has not yet been identified in any studies of U uptake by ZVI although the redox state
of the groundwater chemistry is conducive to its formation. Uraninite dissolution has been
observed to occur rapidly when laboratory experiments become oxic; however, no quantitative
studies are known (Rai, 1999). Laboratory experiments conducted by Abdelouas and others
(2000) indicate that reoxidation of uraninite is significantly reduced if the uraninite is precipitated
together with mackinawite. This detailed mineralogical information is currently (1999) not
available. If the rate of uranium-bearing mineral(s) dissolution is high, then PRBs containing ZVT
will need to be removed or sealed off to prevent future remobilization of U into groundwater.
2.2.4 Selection of Materials for Demonstration
Phosphate. Both the batch and column uptake tests indicate that the bone meal and bone
charcoal apatites would be more effective on a per gram basis in removing dissolved U from Fry
Canyon groundwater than phosphate rock in PRB application. However, the potential clogging
of the PRB constructed from fertilizer-grade bone meal diminishes its effectiveness since dilution
on the order of 10-fold or higher with coarse grain non-reactive materials would be required to
maintain permeability. Because of its small particle size, the bone charcoal powder also suffers
from the need to dilute with coarse material and may be transported out of the PRB. The large
hydraulic conductivity of the pelletized bone-char phosphate allows its use in a PRB without
dilution with coarse-grained non-reactive material. The 5 to 10-fold lower U uptake capacity of
bone-char pellets compared to the fertilizer grade bone meal and bone charcoal powder is offset
by the large hydraulic conductivity of the pellets. Based on these results, the CP3 pellets were
chosen as the best phosphate material for the PRB. The U uptake in batch experiments and
hydraulic conductivity of the formulation produced for the PRB (CP5) was measured prior to
PRB construction. Column experiments with the CP5 pellets were conducted after barrier
emplacement and will be more fully described in a subsequent report.
Zero valent iron and amorphous ferric oxyhydroxide. Based on the laboratory test results,
cost, and availability, one AFO-based material and one ZVI-based material were selected for use
in the Fry Canyon PRB: AFO mixed with 3/8 inch gravel (2% Fe), and Cercona ZVI foam
pellets. By selecting both AFO and ZVI, the efficiency of 2 different chemical mechanisms could
be compared. The selected materials had suitable hydraulic conductivity, high U removal
efficiency, and high compaction strengths.
31
-------
2.3 Further Development of Barrier Materials
The Fry Canyon Demonstration project is the one of the first to consider the treatment of
groundwater contaminated by metals or radionuclides in a PRB. At the time of the installation at
Fry Canyon (September 1997), PRB technology was not well developed. The most suitable
reactive materials were selected for the demonstration, however, it was believed that better
materials would likely be developed in the future. As an example, although AFO was shown to
be an effective adsorbent in laboratory studies, its fine grain size prohibits concentrations greater
than about 2 weight% to be used in a PRB.
A project, funded by the DOE and undertaken in collaboration with the EPA project manager for
the Fry Canyon Demonstration project, conducted a laboratory investigation to evaluate
improved reactive materials (Morrison and others, 1998). Five categories of reactive materials
were investigated: (1) ZVI-based materials, (2) phosphate-based materials, (3) AFO-based
materials, (4) peat and humic acid based materials, and (5) materials containing mixtures. During
this study, a relatively inexpensive form of ZVI (dubbed HSA or "high surface area" metal) was
located and tested in addition to the forms of ZVI typically used in PRBs. HSA has metal
uptake characteristics comparable to other forms of ZVI. Another result of this study was a new
process to coat materials with AFO. The process was used to coat grains with up to 6% by
weight of AFO. While this is a significant improvement, it is still probably not sufficient for
most PRB applications. Cost comparisons indicated that ZVI-based compounds were more
efficient than the other forms of reactive materials. Disadvantages of ZVI-based compounds are
the potential for releasing Fe and Mn (a contaminant present in ZVI) to the groundwater, and the
potential for clogging of the barrier or passivation of the reactive surfaces due to mineral
precipitation.
3.0 PERMEABLE REACTIVE BARRIER DESIGN
Results from the laboratory testing of the reactive barrier materials indicated that each of the
three materials could successfully remove U during the field demonstration. In order to
accommodate all three reactive materials, three PRBs were designed for installation and
concurrent "side by side" operation at the site. A funnel and gate design was chosen, consisting of
three "permeable windows" where each of the PRBs would be placed, separated by "no-flow
walls" and wing walls on each end to channel the groundwater flow into the PRBs (fig. 3.1). Each
PRB and no-flow boundary was keyed into the bedrock (Cedar Mesa Sandstone) underlying the
colluvial aquifer. Heavy equipment consisting of a trac-mounted backhoe and a bulldozer were
chosen to install the PRBs. This design and installation technique was chosen for the following
reasons: (1) ameanable for multiple PRBs placed side by side; (2) low construction cost; (3)
shallow groundwater system; and (4) transferability to other remote, abandoned mine sites with
contaminated groundwater.
32
-------
Impermeable Wall Ground-Water
Flow Direction
Well FC3
U = 3 ppm
Figure 3.1. Schematic diagram showing the funnel and gate design used for the
installation of PRBs at Fry Canyon, Utah.
Planned dimensions of each PRB was 7-ft long by 3-ft wide by 5-ft deep. A 1.5-ft wide layer of
pea gravel was placed on the upgradient side of PRBs to facilitate uniform flow of contaminated
groundwater into each PRB (fig. 3.1). A sacrificial steel frame was designed to act as a template
during barrier construction (fig. 3.2). Each PRB template consisted of plywood end panels that
remained in place during backfilling and two plywood panels with lifting slots on the upradient
and downgradient ends (fig. 3.2). The lifting slots were designed to facilitate panel removal during
backfilling with the pea gravel and reactive barrier material.
33
-------
- REACTIVE MATERIAL FILL
PLYWOOD END PANELS
TO REMAIN IN PLACE
(BOTH ENDS)
PLYWOOD PANELS WITH LIFTING SLOTS
" (REMOVED DURING INSTALLATION)
DETAIL A-\
MATERIAL LIST
3/4" CDX PLYWOOD
2
2
2
4' x 8' SHEET
2'-10ix5'-Oi SHEET
V-e'xS'-O1 SHEET
DETAIL A
NO SCALE
DETAIL B
NO SCALE
DETAIL C
NO SCALE
Figure 3.2. Three-dimensional views of the sacrificial frame design used for installation of PRBs
at Fry Canyon, Utah.
Numerous factors should be considered when designing the thickness of a PRB and include
ground water velocity, rate of contaminant removal by the reactive material, contaminant removal
capacity of the barrier material, estimated mass of the contaminant in the ground-water plume,
and physical constraints of the trenching equipment. In an actual remediation project, these
factors and possibly others would be used to determine the thickness of the PRB. Addressing all
of these design criteria was not practical for the field demonstration of PRBs conducted during
this study. The three-foot thickness of each PRB at the Fry Canyon site was based on the
following criteria: (1) large enough thickness to sample and map gradients in dissolved chemical
constituents across the PRB; (2) sufficient residence time for ground water to chemically interact
with the PRB material during changing hydrologic conditions; and (3) physical limitations of the
trac hoe and trench box used during PRB construction. Results from the PRB demonstration
project at Fry Canyon can be used to estimate proper PRB thickness during future applications
of PRBs for removing uranium from groundwater.
4.0 MONITORING NETWORK DESIGN
Because the objective of this project is to determine if PRBs present a feasible and long-term
groundwater technology, an extensive monitoring network was required and designed. Monitoring
networks used in future applications of PRB technology at sites contaminated by trace-elements
and radionuclides will require less monitoring wells and equipment compared to the Fry Canyon
demonstration project site. However, the monitoring network design and equipment used at the
34
-------
Fry Canyon site will still have direct application to future PRBs installed for remediation
purposes. A description of each monitoring component and subsequent use of the data is
presented in the following sections.
A large number of monitoring/sampling points were installed in each of the PRBs to gain a full
understanding of the potential geochemical reactions and changes in water-quality and
groundwater flow that occurs as groundwater is passively treated. Each PRB contains
16, 1/4-inch schedule 40 PVC wells located along two parallel flow paths in each PRB (figs. 4.1,
4.2, and 4.3). The 0.5 ft spacing of the 1/4-inch monitoring wells is required to evaluate short-
term changes in water quality within the PRBs. A 1/4-inch well casing was selected to minimize
the effects of pumping to adjacent wells during PRB sampling activities. Multi-level 1/4-inch
wells were installed at three sites along each of the two flow paths (figs. 4.1, 4.2, and 4.3). The
purpose of the multi-level wells is to provide sampling ports to monitor changes in water-quality
with depth in the reactive material contained in each PRB. A hose adapter was attached to each
of the 1/4-inch wells to facilitate sampling with a peristaltic pump.
Plan View
Not to scale
1.75 ft
3.5ft
1.75 ft
-M-
TRANSDUCER PORT
DG3
M-
o
TRANSDUCER PORT
DG4
3ft
1.5ft
©
(5
•
PO4R1S-7 /^
P04R1-8 Barrier material \2/
PO4R1-6
FLOW-SENSOR PORT
PO4R1S-4 f\ /JN
P04R1-5 11 P04FS1 V^J
PO4R1-3 •
TRANSDUCER AND MINIMONITOR PORT
P04R1S-1 f\ p04T3 £N
PO4R1-2 y_y V^
TRANSDUCER PORT
P04T1 S^\
PO4R2S-7
PO4R2-8
PO4R2-6
PO4R2S-4
PO4R2-5
PO4R2-3
PO4R2S-1
PO4R2-2
TRANSDUCER PORT
Pea gravel ^""\ P04T2
EXPLANATION
© 1/4-inch well, two sampling depths
• 1/4-inch well, one sampling depth
O 2-inch monitoring well
Figure 4.1. Schematic diagram showing monitoring well placement and sample site identification
for the bone-char PRB.
35
-------
2.0ft
Plan View
Not to scale
3.0ft
2.0ft
3ft
®
•
@
0
ZVIR1S-7
ZVIR1-8
ZVIR1-6
ZVIR1S-4
ZVIR1-5
ZVIR1-3
Barrier material
•
FLOW- SENSOR PORT
fj ZVIFS1 VjJ
*
ZVIR2S-7
ZVIR2-8
ZVIR2-6
ZVIR2S-4
ZVIR2-5
ZVIR2-3
TRANSDUCER AND MINIMONITOR PORT
®
ZVIR1S-1
ZVIR1-2
TRANSDUCER PORT
ZV,T, Q
fj ZVIT3
ZVIR2S-1
ZVIR2-2
TRANSDUCER PORT
Pea gravel ^"\ zviT2
1.5ft
EXPLANATION
© 1/4-inch well, two sampling depths
• 1/4-inch well, one sampling depth
O 2-inch monitoring well
Figure 4.2. Schematic diagram showing monitoring well placement and sample site identification
for the zero-valent iron PRB.
36
-------
2.3ft
Plan View
Not to scale
2.4ft
2.3ft
-M-
-M-
DG1
o
TRANSDUCER PORT
DG2
o
3ft
®
•
®
•
AFOR1S-7
AFOR1-8
AFOR1-6
AFOR1S-4
AFOR1-5
AFOR1-3
Barrier material
FLOW-SENSOR PORT
CJ AFOFS1
TRANSDUCER AND MINIMONITOR PORT
>2/ AFORI^1 Cj AFOT3
TRANSDUCER PORT
AFOT1 S^\
®AFOR2S-7
AFOR2-8
• AFOR2-6
®AFOR2S-4
AFOR2-5
£ AFOR2-3
/£l AFOR2S-1
\^/ AFOR2-2
TRANSDUCER PORT
Pea gravel f\ AFOT2
1.5ft
EXPLANATION
© 1/4-inch well, two sampling depths
• 1/4-inch well, one sampling depth
O 2-inch monitoring well
Figure 4.3. Schematic diagram showing monitoring well placement and sample site identification
for the amorphous ferric oxyhydroxide PRB.
Each PRB was designed to contain 6, 2-inch inside diameter PVC monitoring wells for measuring
water-levels upgradient, within, and downgradient of the PRBs. The water-level data will be used
to construct potentiometric surface maps of the PRBs, monitor potential plugging tendencies as
the barriers age, and construct a groundwater flow model of the PRBs. Each of the 2-inch wells
also contains a sampling port consisting of dedicated flexible tubing to collect water samples
using a peristaltic pump (fig. 4.4).
37
-------
Figure 4.4. Two-inch monitoring well, Fry Canyon, Utah.
Short-term trends in groundwater levels, selected water-quality constituents, and flow directions
and velocities in each PRB are needed to better evaluate barrier performance. For example, hourly
pH data can be used to determine if PRB aging may result in pH changes that may decrease U
removal efficiency. The 2-inch wells are used as access points for the deployment of pressure
transducers, water-quality minimonitors, and flow sensors. The Waterlog H-310, 15-psi pressure
transducers are used to measure hourly water levels and temperature in four wells within each
PRB (fig. 4.5). Each PRB also contains a Yellow Springs Instrument 600XL water quality
minimonitor (figs. 4.1, 4.2, and 4.3) that measures hourly values of pH, specific conductance,
water temperature, dissolved oxygen, and oxidation reduction potential. A 2-inch well in each
PRB (figs. 4.1, 4.2, and 4.3) is used as an access point to measure flow direction and velocity
with a K-V Associates Model 40 portable groundwater flow meter.
38
-------
Figure 4.5. Pressure transducer deployed during PRB demonstration, Fry Canyon, Utah.
Hourly data from the 12 pressure transducers and 3 water-quality minimonitors are
automatically recorded using a Campbell CR10 data logger and SMI92 data storage module (fig.
4.6) located inside a mobile laboratory parked adjacent to the PRBs (fig. 4.7). The automated data
collection equipment was powered with batteries that were recharged with solar panels installed
on the roof of the mobile laboratory. To prevent data loss, the data collection system was backed
up with a duplicate data logger, data storage module, and power supply.
39
-------
Mob
ile laboratory
Data logger #1
and power
supply
Solar panels
mounted on
roof
Junction box
Data logger #2
and power
supply
Amorphous ferric oxide
permeable reactive barrier
(4 transducers and 1
minimonitor)
Zero-valent iron permeable
reactive barrier
(4transducers and
1 minimonitor)
Bone-char permeable
reactive barrier
(4transducters and
1 minimonitor)
Figure 4.6. Schematic diagram of the automatic data recording system within and adjacent to the
permeable reactive barriers, Fry Canyon, Utah.
40
-------
Figure 4.7. Automated data logging equipment used during the Fry Canyon barrier
demonstration project.
The 12 pressure transducers and 3 water quality minimonitors were serviced every 1 to 1.5
months. During service visits, the measured water level was compared to the water level recorded
by each transducer. If the transducer water level was +/- 0.03 ft different than the measured water
level, the y offset was adjusted until the transducer water level matched the actual value. Water
quality sensors in each of the three minimonitors were cleaned and recalibrated with the
appropriate standards during service visits.
5.0 BARRIER AND MONITORING NETWORK INSTALLATION
Prior to PRB installation, a health and safety plan (HASP) was developed to address the issues
associated with trench construction and installation of the chemical material to form the PRBs
(Appendix D). A copy of this HASP can be obtained by contacting the USGS District Chief
located in Salt Lake City, Utah.
The design for the Fry Canyon test site included 3 permeable reactive gates separated by
impermeable sections of cement-bentonite. The design also included short sections of cement-
bentonite walls at the ends to help funnel groundwater into the reactive gates. Previous water
level data indicated that groundwater flowed subparallel to Fry Creek and the gates were placed
perpendicular to the flow. The angled spur of impermeable wall on the east end was intended to
capture additional contaminated groundwater that would otherwise flow into Fry Creek.
41
-------
The design was modified in the field for several reasons. A number of problems were
encountered during installation of the PRBs. The first significant problem was the unanticipated
occurrence of a large bedrock nose of the Ceder Mesa Sandstone during initial trench excavation.
The bedrock nose caused the trench orientation to be rotated approximately 35 degrees in a
clockwise direction. This trench rotation did not allow the gate structures to intercept the
groundwater flow at right angles as planned. Instead, the gate structure of each PRB intercepted
the anticipated groundwater flow direction about 35 degrees from perpendicular.
After lowering the ground surface by about 4 ft with a dozer, the wall was installed using a
trackhoe. A trench box was used to hold back caving soils and to protect workers (fig. 5.1). The
trench was scraped down to bedrock on which the sacrificial boxes were placed with the
trackhoe. The sacrificial boxes (fig. 3.2) were constructed from angle iron and plywood and
delivered to the field site in several pieces that were assembled on site. Reactive material was
placed in the sacrificial box and then the trackhoe was used to place native fill around it. Reactive
material and native soil were filled gradually to avoid overpressuring one side of the box. After
the box was partially filled, the two longest plywood sheets were raised part way up and after
complete filling, were removed. The frame and the other two plywood sheets remain in the
ground.
42
-------
Figure 5.1. Trench box used to protect workers during installation of the PRBs at Fry Canyon,
Utah.
The second problem was encountered during the construction of the no-flow barrier or "wing
wall," designed to funnel contaminated groundwater to the gate structures that contained
permeable reactive material. Bentonite slurry was dumped into the wing wall trench from an
adjacent cement mixing truck. The slurry could not be contained in the wing wall area while the
gate structures were constructed. The bentonite slurry was removed and the wing wall was
constructed with plywood and plastic sheeting. The bentonite was reused later on the west end
of the wall (fig. 3.1) after placing sufficient native soil to prevent any chance of the slurry
invading the AFO reactive gate. Bentonite was grouted into the junctions using a half-round of 8-
inch PVC pipe to hold it in place. Associated problems with the bentonite slurry included the
use of cement trucks at remote sites with limited water. If the bentonite slurry is not dumped
from the cement truck within 4 hrs, the slurry begins to congeal and must be removed from the
cement mixer using a specialized and costly procedure. The congealing can be delayed by adding
water to the mixture; however, limited water supplies at the Fry Canyon site did not allow for
43
-------
this option. Another problem with the bentonite slurry is the potential for leakage to other areas
of the trench prior to backfilling. For example, a small amount of slurry leakage to the gate areas
could prevent or reduce groundwater flow, thereby decreasing PRB treatment efficiency.
The assurance of a consistent seal between the underlying bedrock and the no-flow barriers was
problematic because the bedrock surface was not flat. The presence of groundwater in the trench
made it difficult to visually inspect the seal between the no-flow barrier and the underlying
bedrock.
Bone char and foamed ZVI pellets were delivered to the site in 8-cubic-foot cloth bags. The bags
were lifted with the trackhoe forks and suspended over the trench while the bottom was cut open
allowing the materials to flow into the sacrificial box. The ZVI foam pellets and bone char pellets
were used as received without mixing with sand or gravel . AFO was delivered as a slurry in 55-
gallon drum. The slurry contained 13% AFO [as Fe(OH)3]. The slurry was stirred by hand in
the drums and then mixed with gravel (3/8 inch) in a cement mixer at a proportion of 1 part slurry
to 2 parts gravel (by volume). The mixture was placed in the trackhoe scoop and then dumped
into the sacrificial box (fig. 5.2).
Figure 5.2. Placement of AFO barrier material into the gate structure of the permeable reactive
barrier, Fry Canyon, Utah.
44
-------
Monitoring wells were placed using a template secured to the top of the sacrificial box. For the
first gate, the sacrificial box was placed in the trench and then the monitoring wells were placed.
This proved to be cumbersome while working in the trench. For the other 2 gates, the monitor
wells were placed in the sacrificial box first and then the entire unit was lowered into the trench.
After construction of the PRB, the ground surface was brought back up to grade with the dozer.
Care was taken not to damage the monitoring wells. The monitoring devices were wired to data
collectors in a mobile trailer. Plastic pump boxes were installed to protect the well heads. The
slope was stabilized with straw mats and drainage upslope was slightly modified to prevent flash
floods from eroding the project area.
The final "as built" dimensions of each PRB are shown in Figure 5.3. Depths of barrier materials
were less than the designed depth of 5 ft because of materials settling after removal of the trench
box (fig. 5.1). This resulted in the water table approximately 1.9 ft above the top of the PO4
PRB. The saturated zone above the PRB was backfilled with native colluvium that is less
permeable than the reactive materials. In a remediation application, the highest projected
elevation of the water table must always be below the top of the barrier material to ensure that all
groundwater is treated. This was not a stringent requirement in this demonstration project
because monitoring wells were located within each PRB to document U removal efficiencies. The
"as built" volume of reactive material in each PRB were calculated from the field survey data as:
(1) PO4 = 67.2 ft3; (2) ZVI = 77.7 ft3; and (3) AFO = 67.2 ft3. Discrepancies between as-built
volume and delivered volumes are due to reactive material spreading into the void left as the
trench box was removed.
45
-------
EXPLANATION
1 Amorphous ferric hydroxide
1 Zero-valent iron
I Bone char phosphate
J Pea gravel
] Native fill and tailings
_ Water table measured
January, 1999
-A1 Line of reactive wa
cross-section
Amorphous ferric hydroxide cross-section
VERTICAL EXAGGERATION
Zero-valent iron cross-section
Bone char phosphate cross-section
A J
L
U-1.5ft.*U 3.0 feet d
U-1.5 ft*U 3.0 feet d
B1
U-1.5ft.-*U 3.0 feet d
Figure 5.3. Location and dimensions of permeable reactive barriers after construction, Fry
Canyon, Utah.
After backfilling of the PRBs was completed, each of the wells was pumped to remove the fine
particulates introduced during the barrier construction activities. The 2-inch diameter wells were
developed with a Brainard Killman ball and piston hand pump and the 1/4-inch diameter wells
were developed with a peristaltic pump. The 2-inch diameter wells completed in the pea gravel
had very few fine particulates and the water cleared up after approximately 2 casing volumes
were removed. Wells completed in the undisturbed colluvial aquifer downgradient of the PRBs
typically required more than 2 casing volumes to adequately remove the fine particulates
introduced during well installation. With the exception of the AFO barrier, both the 1/4-inch and
2-inch diameter wells completed in the barrier material required pumping less than 2 casing
volumes to adequately remove the majority of the fine particulates. Wells completed in the AFO
barrier material contained significant quantities of ferric oxide particles that did not adsorb to the
surface of the pea gravel during barrier installation. Wells within the AFO barrier still contain
significant iron particulates 12-months after pumping the wells periodically for water samples.
6.0 POST-INSTALLATION SAMPLING AND ANALYSIS
Monitoring of the three PRBs was designed to meet technology demonstration, quality
assurance/quality control, and state compliance goals. During the first year of PRB operation,
groundwater was sampled 7 times (Sep 97, Oct 97, Nov 97, Jan 98, Apr 98, Jun 98, and Sep 98).
Table 6.1 list the number of water samples and the chemical constituents that were analyzed
during each of the time periods. More than 370 water samples were collected and analyzed in the
first year of PRB operational and monitoring activities.
46
-------
Table 6.1. Sampling period, chemical constituents, and number of samples taken during the first
year of permeable reactive barrier operation at Fry Canyon, Utah.
Sampling period
September 1997
October 1997
November 1997
January 1998
April 1998
June 1998
September 1998
Chemical constituents
analyzed
pH, specific conductance, total
alkalinity, temperature, iron,
phosphate, uranium
pH, specific conductance, total
alkalinity, temperature, iron,
phosphate, uranium
aluminum, calcium, copper,
iron, potassium, lithium,
magnesium, manganese,
phosphorus, sodium, silicon,
strontium, zinc, sulfate,
chloride
pH, specific conductance, total
alkalinity, temperature, iron,
phosphate, uranium
pH, specific conductance, total
alkalinity, temperature, iron,
phosphate, uranium
aluminum, calcium, copper,
iron, potassium, lithium,
magnesium, manganese,
phosphorus, sodium, silicon,
strontium, zinc, sulfate,
chloride
pH, specific conductance, total
alkalinity, temperature, iron,
phosphate, uranium
pH, specific conductance, total
alkalinity, temperature, iron,
phosphate, uranium
aluminum, calcium, copper,
iron, potassium, lithium,
magnesium, manganese,
phosphorus, sodium, silicon,
strontium, zinc, sulfate,
chloride
pH, specific conductance, total
alkalinity, temperature, iron,
phosphate, uranium
Number of samples taken
47
58
47
58
47
58
*58
*Additional samples
were installed during
in September 1998 reflect additional 2-inch diameter monitoring wells that
August 1998
Downgradient wells (DG1, DG2, DG3, DG4, FC2, and DP4) were monitored to ensure that
groundwater was not further degraded after installation of the PRBs. Iron and pH were monitored
in downgradient wells because iron corrosion reactions within the ZVIPRB could cause increases
in pH and iron concentrations in the treated groundwater. The U was monitored in downgradient
wells because of the potential for increased pH values causing desorption of U from the
previously contaminated colluvial sediments. The U concentrations from the downgradient wells
were not used to evaluate PRB performance because these wells are completed in U contaminated
47
-------
colluvium. Phosphate was monitored in downgradient wells to address the possibility of
phosphate transport downgradient from the PO4 PRB.
A description of sample processing, analytical methods, and quality assurance results are
presented and discussed in Appendix E.
7.0 YEAR ONE RESULTS OF PERMEABLE REACTIVE BARRIER
DEMONSTRATION
7.1 Volume of Groundwater Treated
On the basis of hydrologic and physical properties of the aquifer and the hydraulic gradient
derived from potentiometric contour maps (fig. 7.1) prior to wall installation, the average linear
velocity of groundwater in the aquifer ranges seasonally from 0.2 to 2.5 ft/d. The velocity
changes in response to changes in the amount of recharge occurring (hydraulic gradient) and as a
result of the lateral variability in hydraulic conductivity and effective porosity of the sediments.
Average linear velocity during the summer when phreatophytes are using groundwater is about
70% of winter velocities.
48
-------
Explanation
| | Evaporation basins and tailings
| | Phreatophytes
_5 358_ Potentiometric contour—Shows altitude at which
water level would have stood in tightly cased wells,
July 1997. Interval is 0.2 foot
Control point—Number is altitude of water level, in feet
Potentiometric contours
July 1997 (low)
'a/
Explanation
| | Evaporation basins and tailings
| | Phreatophytes
—5,358— Potentiometric contour—Shows altitude at which
water level would have stood in tightly cased wells,
January 1997. Interval is 0.2 foot
5,359.56
• Control point—Number is altitude of water level, in feet
Potentiometric contours
January 1997 (high)
Figure 7.1. Configuration and altitude of the potentiometric surface of the colluvial aquifer at
Fry Canyon, Utah prior to the permeable reactive walls being installed in (a) July 1997, and (b)
January 1997.
The flow system was altered after PRB construction was completed. Hydraulic gradients were
increased because the bentonite and plywood wing walls concentrate more groundwater from a
wider area of the aquifer than was common in the natural state. Based on the change in hydraulic
gradient, the average hydraulic conductivity, and porosity of the aquifer between wells FC-4 and
FC-3, average-linear velocity in the aquifer between the two wells upgradient of the walls
49
-------
increased from about 0.75 ft/d before wall installation to about 1.25 ft/d after wall installation.
When groundwater moves into the influence of the reactive-wall construction zone the hydraulic
gradient becomes nearly flat because the hydraulic conductivity of the pea-gravel buffer zone and
the wall materials is nearly 10 times larger than in the aquifer. However, average linear velocity of
groundwater probably remains about the same as in the aquifer because as the large hydraulic
conductivity values of the PRBs tend to increase velocity, the lower hydraulic gradient tends to
decrease velocity, thus having the opposite effect.
The estimated capture zone during December 1998 for the installed PRBs is shown in Figure 7.2.
Even though this zone includes only about 14,500 ft , it represents the terminus of a drainage
area nearly 5 times that size. The quantity of water that enters and exits this capture zone was
estimated to be no greater than 90 ft3/d (about 10% of the precipitation falling on the drainage
area annually). Thus, the amount of groundwater being treated would be no greater 90 ft /d or
approximately 33,000 ft during the first year of operation.
50
-------
^ XX
^ //
1 "
200 FEET
_|
Explanation
| | Evaporation basins and tailings
| | Approximate capture zone for the chemical reactive walls in
December 1998
—5,358— Potentiometric contour—Shows altitude at which
water level would have stood in tightly cased wells,
December 1998. Interval is 0.2 feet
5,359.56
• Control point—Number is altitude of water level, in feet
Figure 7.2. Configuration and altitude of the potentiometric surface in the colluvial aquifer at
Fry Canyon, Utah, December 1998, and the approximate area of aquifer influenced by the
permeable reactive barriers.
The amount of water passing through each individual PRB is some portion of the total, and is
dependent on the orientation of each wall to the direction of groundwater flow and the hydrologic
properties of each wall. Computer simulations coupled with tracer tests have been initiated and
may be used to refine the estimates of treated water and the mass of U removed during the life of
the project.
7.2 Changes in Uranium Concentration
As of September 1998, 1 year of uranium-concentration data had been collected since operation
of the PRBs began in September 1997. The input U concentrations are significantly different for
51
-------
each PRB, ranging from less than 1,000 |ig/L in the PO4 PRB to more than 20,000 |ig/L in the
AFO PRB (figs 7.3, 7.4, and 7.5). The input U concentrations to each of the PRBs also vary
seasonally by approximately 3,000 in the PO4 PRB to greater than 9,000 |ig/L in the AFO PRB.
PO4 BARRIER-ROW 1
DC
111
K
_l
DC
111
Q_
1
DC
O
DC
O
S
4,000
3,000
2,000
1,000
0
\ SEPTEMBER 1997
-0- OCTOBER 1997
'\
-A- NOVEMBER 1997
—V— JANUARY 1998
x.
--*-- APRIL 1998
\
JUNE 1998
0 \ SEPTEMBER 1998
- V '••- V-
^^^-^ / " . ' ' ' ^'~-~^
i . i . i . i . i . i . i
O 5,000
PO4 BARRIER-ROW 2
<
DC
111
O
z
O
O
4,000
3,000
2,000
1,000
-Q- SEPTEMBER 1997
-o- OCTOBER 1997
-A- NOVEMBER 1997
-v- JANUARY 1998
-*- APRIL 1998
-°- JUNE 1998
SEPTEMBER 1998
-1.0 -0.5 0.0
0.5 1.0 1.5 2.0
2.5 3.0
CROSS SECTION DISTANCE FROM GRAVEL/
BARRIER MATERIAL INTERFACE, IN FEET
Figure 7.3. Changes in dissolved uranium concentrations in the bone char phosphate permeable
reactive barrier from September 1997 through September 1998, Fry Canyon, Utah.
52
-------
DC
LU
DC
LU
Q.
DC
O
O
DC
O
<
DC
111
O
•z.
O
O
<
DC
15,000, , 1 r-
ZVI BARRIER-ROW 1
10,000
5,000
6,000
5,000
4,000
3,000
2,000
1,000
0
O
SEPTEMBER 1997
-O- OCTOBER 1997
-A- NOVEMBER 1997
--V-- JANUARY 1998
•••*-•- APRIL 1998
-Q- JUNE 1998
—O— SEPTEMBER 1998
ZVI BARRIER-ROW 2
-Q- SEPTEMBER 1997
-o- OCTOBER 1997
A- NOVEMBER 1997
v- JANUARY 1998
-*- APRIL 1998
-°- JUNE 1998
o- SEPTEMBER 1998
-1.0 -0.5 0.0
0.5 1.0 1.5 2.0
2.5 3.0
CROSS SECTION DISTANCE FROM GRAVEL/
BARRIER MATERIAL INTERFACE, IN FEET
Figure 7.4. Changes in dissolved uranium concentrations in the zero valent iron permeable
reactive barrier from September 1997 through September 1998, Fry Canyon, Utah.
53
-------
AFO BARRIER-ROW 1
20,000
DC
111
j 15,000
DC
111
°- 10,000
DC 5,000
DC
0 o
s
§* SEPTEMBER 1997
ff;'*--,^ --0- OCTOBER 1997
'•;-:-:>:> -A- NOVEMBER 1997 '
- O.., '"'"C'X *; --V-- JANUARY 1998 -
'"•-•,.v "::.:. V \ - X -APRIL 1998
""~'-';'g\ \ -D- JUNE 1998
•\OA \--O S EPT EM B ER 1 998~
V. '.\ ~x
A g— —--ft- -.--„- m
,1,1,1,1,1,1,1,
<
DC
111
O
z
O
O
DC
25,000
20,000
15,000
10,000
5,000
AFO BARRIER-ROW 2
-B- SEPTEMBER 1997
-o- OCTOBER 1997
-A- NOVEMBER 1997 -\
-v- JANUARY 1998
-*- APRIL 1998
-°- JUNE 1998
-o- SEPTEMBER 1998
-1.0 -0.5
0.0
0.5
1.0
1.5
2.0
2.5
3.0
CROSS SECTION DISTANCE FROM GRAVEL/
BARRIER MATERIAL INTERFACE, IN FEET
Figure 7.5. Changes in dissolved uranium concentrations in the amorphous ferric oxyhydroxide
permeable reactive barrier from September 1997 through September 1998, Fry Canyon, Utah.
During the first year of operation, the PRBs removed most of the incoming U (figs. 7.3, 7.4, and
7.5); however, the percentage of U removal varies with time and barrier material (table 7.1).
Percent uranium removal was calculated using the following formula:
Uremoved = 1 00 - (Ubarr/Umput) (7.1)
Where
Uremoved
is the % of U removed from either row 1 or row 2 flow paths
is the concentration of U in groundwater 1 .5 ft from the pea gravel/PRB interface in
either row 1 or row 2 monitoring points
is the concentration of U in groundwater prior to entering the PRB in either row 1
or 2 monitoring points
54
-------
Table 7.1. Percentage of input U concentration removed after traveling approximately 1.5-feet
into each of the permeable reactive barriers during September 1997 through September 1998, Fry
Canyon, Utah.
Date
SEP
1997
OCT
1997
NOV
1997
JAN 1998
APR
1998
JUN 1998
SEP
1998
PO4
barrier,
row 1
99.7
94.8
89.4
79.2
96.7
98.3
> 99.9
PO4
barrier,
row 2
94.4
71.9
71.6
61.8
77.4
88.6
92.0
ZVI
barrier,
row 1
> 99.9
> 99.9
> 99.9
> 99.9
> 99.9
> 99.9
> 99.9
ZVI
barrier,
row 2
> 99.9
> 99.9
> 99.9
> 99.9
> 99.9
> 99.9
> 99.9
AFO
barrier,
row 1
95.3
94.9
93.6
85.9
77.8
81.9
87.4
AFO
barrier,
row 2
87.4
81.4
65.1
60.1
47.5
66.7
37.4
The ZVI PRB has consistently lowered the input U concentration by more than 99.9% after the
contaminated groundwater had traveled 1.5 ft into the PRB. This trend is true for both row 1 and
row 2 monitoring points (table 7.1). Figure 7.4 indicates that the majority of U is removed from
the groundwater after traveling 0.5- to 1.0-ft into the ZVI PRB. During the first year of
operation, there is no indication that U removal efficiencies in the ZVI PRB have decreased.
The percentage of U removed in the PO4 PRB is less than the ZVI PRB (table 7.1). The U
removal efficiency consistently decreased through January 1998 to below 70% U removal in row
2 monitoring points. After January 1998, the percentage of U removal has consistently increased
on both sides of the PO4 PRB. The observed increase in U removal may be due to a transition to
more chemically reducing conditions over time in the PO4PRB. This transition to more reducing
conditions is indicated by a consistent increase of dissolved iron in the PO4 PRB. For example, in
well PO4R1-2, in the PO4 PRB, the dissolved iron concentration increased from 60 |lg/L in
October 1997 to 1,090 |ig/L in January 1998 and 4,240 |ig/L in June 1998. The mechanism
causing the transition to reducing conditions in the PO4 PRB is unknown at this time.
Results from row 1 monitoring points in the AFO PRB indicate that more than 90% of the input
U concentration was removed through November 1997 (table 7.1). From January 1998 through
September 1998 the U removal percentage monitored by row 1 wells was reduced to less than
90%. The U removal in the other half of the AFO PRB sampled by the row 2 monitoring points
was significantly less (table 7.1) and was never higher than 88%.
Variation in the % of U removal in the AFO PRB appears to be related to large pH changes in
selected row 1 monitoring points (fig. 7.6). For example, the pH in wells AFOR1-5 and AFOR1-
55
-------
8 varied from about 7.5 to greater than 8.5 pH units (fig. 7.6). The mechanism causing this change
in pH is unknown; however, the U removal by the AFO barrier appears to be inversely related to
the increases in pH. Elevated pH values during the November 1997 and February and April 1998
monitoring periods resulted in decreasing percentages of U removal. Decreases in pH during the
June and September 1998 monitoring periods resulted in an increased amount of U removal (fig.
7.6).
o.u
8.4
8.2
8.0
7.8
7.6
7.4
....... WELLAFOR1-5 '
»
/ * —
\ Percent U removal
* '• V \» "
/ •
J '•• "•.
• • " x
\0 VN
•
'•
.
^' pH value
i i i i
96
94
92
90
88
86
84
82
80
78
7fi
3D
m
I
Q.
o.u
8.4
8.2
8.0
7.8
7.6
7 4
• •. WELLAFOR1-8 -
/v
/ vx
'/ X Percent U removal
N"-- \
/ \ ""X\-» "
' '. *> •'
•\
\ x
' \ x
' \ • x
' \ '• x
\ '' ...» "*m
• pH value -..-•' -
i i i i
98
96
94
92
90
88
86
Q/l
7/21/97 10/29/97 2/6/98 5/17/98 8/25/98
m
3D
o
m
SAMPLE DATE
Figure 7.6. Variation in pH and percent uranium removal from September 1997 through
September 1998 in two monitoring points completed in the amorphous ferric oxyhydroxide
barrier, Fry Canyon, Utah.
The U concentrations were determined in filtered (0.45 |im) and unfiltered water samples
collected in October 1997 after installation of the PRBs. The purpose of this was to determine if
unfiltered U concentrations were higher relative to pre-installation data. It is also important to
document if PRB installation has increased the relative amount of unfiltered U concentration
above that expected by analytical variation. For example, an increase in unfiltered U
56
-------
concentration relative to filtered U concentration may indicate that U is associated with colloids
and subject to migration from the PRB to the downgradient aquifer.
Similar to the pre-installation samples (fig. 2.7), comparison of the filtered and unfiltered samples
indicates no increase (above that expected by analytical variation) in unfiltered relative to filtered
U concentrations for most samples (fig. 7.7). The majority of samples taken from within the
AFO PRB have significantly higher unfiltered relative to filtered U concentrations (fig. 7.7). The
AFO material was in a slurry that was mixed with pea gravel during PRB installation. After AFO
PRB installation, it is likely the residual AFO slurry that did not attach to the pea gravel
adsorbed U resulting in an elevated U concentration in the unfiltered samples. The unfiltered U
concentration in the non-barrier wells was not elevated relative to the filtered concentration
indicating that the paniculate U was not transported more than a few ft beyond the barrier.
pffi
100,000
10,000
o-I 1,000
DC
8*
100
10
DC
LU
0.1
Error bar +/- 10 percent
O
O
D Bone char sam pies
O Zero valent iron samples
;'...';. Amorphous ferric oxyhydroxide samples
Non-barrier samples
0.1
1
10
100
1,000
10,000 100,000
FILTERED URANIUM CONCENTRATION,
IN MICROGRAMS PER LITER
Figure 7.7. Comparison of uranium concentration in filtered and unfiltered water samples
collected after installation of permeable reactive barriers, October 1997, Fry Canyon, Utah.
Additional geochemical and hydrological factors that affect U removal efficiencies
and processes in each of the PRBs are currently (1999) being evaluated. These factors include but
are not limited to changes in the amount and velocity of water flowing through the PRBs, type
and quantities of minerals forming within the PRBs, and small-scale groundwater flow paths
through the PRBs.
7.3 Water-Quality Effects of Barrier Materials
A major concern of using PRBs for long-term remediation of contaminated groundwater is
changes in the chemical quality of water caused by the reactive material. For example, high iron or
manganese concentrations from iron filing barriers could have negative impacts to downgradient
57
-------
water quality (Grand Junction Office, 1998). Selected data from the water quality minimonitor in
each PRB were plotted during the first year of operation to determine changes in water quality as
a result of barrier aging.
Water temperature followed a seasonal cycle in each PRB (fig. 7.8). The minimum water
temperature occurred in late February and the maximum occurs in early- to mid-September. An
annual temperature difference of approximately 10 degrees Celsius was observed during the first
year of PRB operation. Each of the 3 PRBs reacted similarly to the changes in seasonal
temperature. The effects of these temperature changes PRB performance are unknown.
LLJ
LiJ
LiJ
LiJ LLJ
I
18
16
14
12
10
--O--PO, BARRIER
—D-ZVI BARRIER
-A-AFO BARRIER
i
_L
7/11/97
9/29/97
12/18/97
3/8/98
DATE
5/27/98
8/15/98
11/3/98
Figure 7.8. Changes in water temperature within the bone char phosphate, zero valent iron, and
amorphous iron oxyhydroxide permeable reactive barriers from September 1997 through
September 1998, Fry Canyon, Utah.
The pH values in water samples from the PO4 PRB are similar to pre-installation pH values;
however, water samples from the ZVI and AFO PRBs contain higher pH values relative to
samples from the upgradient well, FC3 (fig. 7.9). The elevated pH in the ZVI PRB is probably
the result of iron corrosion in either aerobic or anaerobic conditions as shown in equations 7.2 and
7.3,
2Fe° + O2 + 2H2O <—> 2Fe2+ + 4OH" (7.2)
2+
Fe + 2H2O <— > Fe + H2 + 2OH
(7.3)
where:
Fe° is metallic iron,
O2is dissolved oxygen,
H2O is water,
OH" is hydroxide ion,
H2 is hydrogen, and
Fe +is ferrous iron.
58
-------
Both reactions can result in an increase in pH. Upgradient wells contain measurable dissolved
oxygen concentrations implying that reaction 7.2 probably occurs until all the oxygen is
consumed. Dissolved oxygen in the groundwater within the ZVIPRB is entirely consumed after
traveling 0.5 ft thus allowing reaction 7.3 to occur. Gas generation is occurring in the center part
of the ZVI barrier and it is likely hydrogen gas; however, this has not been confirmed
analytically.
I
Q.
.D--D--D
10.0
9.5
9.0
8.5
8.0
7.5
7.0
6.5
6.0
7/11/97 9/29/97 12/18/97 3/8/98 5/27/98 8/15/98 11/3/98
DATE
EXPLANATION
--O-- PO4 BARRIER
---D--- ZVI BARRIER
-A-AFO BARRIER
O WELL FC3
Figure 7.9. Changes in pH values within the bone char phosphate, zero valent iron, and
amorphous iron oxyhydroxide permeable reactive barriers and background well FC3 from
September 1997 through September 1998, Fry Canyon, Utah.
The dissolved oxygen concentration and oxidation-reduction potential in water samples from each
PRB are distinctly different (figs. 7.10 and 7.11). The dissolved oxygen concentrations in the ZVI
PRB were below the lower reporting limit of 0.10 mg/L during the first 7 months of barrier
operation and increased to slightly above the lower reporting limit since mid-March 1998. These
data support the consumption of dissolved oxygen by iron oxidation as shown in equation 7.2. In
addition, the lack of dissolved oxygen during the first 7 months of barrier operation also support
the presence of anaerobic corrosion (equation 7.3) and the generation of hydrogen. This slight
increase in dissolved oxygen concentrations after mid-March 1998 could indicate a decrease in the
efficiency of oxygen consumption by the barrier material. Sivavec and others (1997) postulate
that anaerobic corrosion (equation 7.3) may lead to the formation of iron hydroxide precipitates
that may coat the iron surface and affect its redox properties. The oxidation reduction potential
values in the ZVI PRB have been less than -370 millivolts during the first year of operation;
however, there was a slight upward trend during the first 6-months of barrier operation
(fig. 7.11).
59
-------
)ISSOLVED OXYGEN, IN
MILLIGRAMS PER LITER
1— 1 ^
3.5
3.0
2.5
2.0
1.5
1.0
0.5
0.0
i i i i i i
O __O--PO4 BARRIER
P~Q ;\ -D-ZVI BARRIER .
/ V> -A- AFO BARRIER -
? 9 / \ A
/ V ' v / \
' \ .'
' V ,'
\AL / > \
/6^A \ \ o
: ° A / q\ /\A :
/\ \ \ ' x
A \ / V° '
\ / ?^ /
n X -Q n -n
^^-^ LOWER REPORTING LIMIT ~
i i i i i i
7/11/97
9/29/97 12/18/97
3/8/98 5/27/98
DATE
8/1 5/98
11/3/98
Figure 7.10. Changes in dissolved oxygen concentration from within the bone char phosphate,
zero valent iron, and amorphous iron oxyhydroxide permeable reactive barriers from September
1997 through September 1998, Fry Canyon, Utah.
LiJ
B
0.
O
O
^
o
(A
O
400
200
8
-200
-400
-600
.--O--PO4 BARRIER
-D-ZVI BARRIER
-A-AFO BARRIER
j
7/11/97
9/29/97 12/18/97
3/8/98 5/27/98 8/15/98 11/3/98
DATE
Figure 7.11. Changes in oxidation reduction potential from within the bone char phosphate,
zero valent iron, and amorphous iron oxyhydroxide permeable reactive barriers from September
1997 through September 1998, Fry Canyon, Utah.
Dissolved oxygen concentrations in the PO4 and AFO PRBs were similar to the concentrations
measured in pre-installation groundwater samples (fig. 7.10). No oxygen consuming reactions are
expected to occur in either the PO4 or AFO PRBs. Monthly variations in the dissolved oxygen
concentrations are probably a function of interaction between surface-water and groundwater at
60
-------
the site. During rainfall events, the groundwater gradient in the PRBs can reverse causing water
from the ZVI barrier to enter the gravel pack and enter the PO4 or AFO barriers for short time
periods.
Oxidation reduction potential values in the PO4 and AFO barriers are generally positive (7.1 1)
and reflect the measurable dissolved oxygen in these barriers (fig. 7.10). The mechanism causing
the consistent decrease in oxidation reduction values in the PO4 PRB since March 1998 is not
known; however, this decrease has consistently increased the U removal efficiencies in the PO4
PRB.
Ferrous iron concentrations during the first year of ZVI and AFO PRB operation were
summarized for upgradient, within-barrier, and downgradient monitoring wells (figs. 7.12 and
7.13). Excessive amounts of iron in groundwater are considered undesirable because it can form
red oxyhydroxide precipitates that stain laundry and plumbing fixtures. Ferrous iron
concentrations in the upgradient ZVI wells were less than 3 mg/L during the first year of barrier
operation (fig. 7.12). Numerous wells within the ZVI PRB contained ferrous iron concentrations
that exceeded the upper reporting limit of 12 mg/L. Well DP4, a downgradient monitoring well,
contained ferrous iron concentrations consistently higher than 5 mg/L. These concentrations are
larger than the ferrous iron concentrations in groundwater that entered the ZVI PRB during the
first year of operation.
o
FCC
o
O
cc
HI
16
14
1°
10
8
6
4
2
0
i i i i i i i i i i i i i
-
-
-
-
-
J,,J
.
L
1 i
1
i
i
Upper reporting limit -
of field determination -
_^
L i i i
-
-
-
-
-
& & 3>
,•& ,•& ,•&
EXPLANATION
I I SEP 97
I I OCT97
I I NOV97
EH JAN 98
I I APR 98
V77\ JUN 98
I I SEP 98
WELLS WITHIN THE ZVI PRB
MONITOR ING WELLS
5 <
ooc
Q o
Figure 7.12. Changes in ferrous iron concentration in water samples from upgradient, within-
barrier, and downgradient wells in the zero-valent iron permeable reactive barrier from September
1997 through September 1998, Fry Canyon, Utah. Samples with a ferrous iron concentration
above the upper reporting limit were assigned a value of 0.3 times the upper reporting limit of
12 mg/L.
61
-------
o
!
-------
dissolved oxygen concentrations and positive oxidation-reduction potentials measured within this
barrier.
Phosphate concentrations in samples from well FC3, upgradient of the PO4 PRB were lower
than the lower reporting limit of 0.2 mg/L during the first year of barrier operation (fig. 7.14).
Monitoring points in the gravel pack in front of the P04 PRB contained elevated phosphate
concentrations that periodically exceeded 10 mg/L. The elevated phosphate concentration in the
upgradient wells probably resulted from the low ground-water gradient in the barrier allowing
water to move from the PO4 PRB into the pea gravel during selected time periods. The
persistence of high phosphate concentrations in the pea gravel is probably the result of the
limited amount of metal oxides expected to be contained in the pea gravel. Metal oxides,
especially iron and manganese oxy-hydroxides usually limit phosphate concentrations to
significantly less than 1 mg/L (Hem, 1989). Water samples collected from within the PO4 PRB
contained elevated phosphate concentrations as high as 81 mg/L (fig. 7.14). Downgradient
monitoring points contained low phosphate concentrations that were less than 0.2 mg/L during
the first year of PRB operation. These low phosphate concentrations are probably due to the
large amounts of naturally occurring iron oxyhydroxides in the colluvial sediments downgradient
from the PO4 PRB.
z
IT RATION,!
ER AS P04
o
o-1
ZCC 1
I DC n 1
Q- (3
W —
^^ ^J
^L ^£
n m
-
~
1
1
1
1
1
1
1
1
1
1 r
1
I
\
I
•
1
[
\
1
1
1
ll
; EXPLANATION
- r^\ SEP 97
dl OCT 97
EZ1 NOV 97
! CZl JAN 98
: CHAPR98
- E53JUN98
EZ1SEP98
g WELLSWITHINTHEPO4PRB z
5 w z 5 ^
%g MONITORING WELLS Q°5
Figure 7.14. Changes in phosphate concentration in water samples from upgradient, within-
barrier, and downgradient wells in the bone char permeable reactive barrier from September 1997
through September 1998, Fry Canyon, Utah. Samples with a phosphate concentration below the
lower reporting limits were assigned a value of 0.7 times the two lower reporting limits of 0.05
and 0.2 mg/L.
63
-------
8.0 COST ANALYSIS OF SITE CHARACTERIZATION, PRB DESIGN, AND
PRB INSTALLATION
Actual costs of the demonstration project activities are summarized (table 8.1) to document the
potential costs that could be associated with a full-scale PRB deployment at another site. Cost
figures are provided for the following three project phases:
• site selection, characterization, and PRB material testing
• PRB design
• PRB installation
Phase I activities include site selection and site characterization, as well as laboratory testing and
development of potential reactive materials for placement in the PRBs. The cost of Phase I tasks
at Fry Canyon was $280,000 and took place over a 1-year period. The time period and cost of
Phase I activities could be reduced significantly in a remediation application for the following
reasons: (1) the site will already be selected; (2) existing information on type and performance of
reactive material will decrease the laboratory testing time and subsequent cost; (3) site
characterization activities will likely be complete.
Table 8.1. Actual cost and duration of project planning through installation of three permeable
reactive barriers at the Fry Canyon site, Utah. Cost figures include indirect costs incurred by the
U.S. Department of Energy and U.S. Geological Survey. Indirect cost rates will probably be
higher for private organizations.
Project phase and
approximate duration
Tasks
Cost, in
U.S. dollars
Phase I
March 1,1996 to March
15,1997
1. Project planning
2. Site selection
3. Laboratory testing of reactive materials
4. Selection of reactive materials
5. Regulatory permitting
6. Site Health and Safety Plan for site characterization
7. Site characterization
$280,000
Phase II
March 16,1997 to
August 15, 1997
1. Design of permeable reactive barrier structures
2. Design of monitoring network
3. Logistical planning
4. Analysis and awarding of subcontracts
5. Development of Health and Safety Plan for
construction phase
$148,000
Phase III
August 20 to
September 4, 1997
1. Purchasing and shipment of material, supplies, and
equipment
2. Excavation of trench
3. Installation of monitoring network
4. Placement of reactive material and
backfilling/recontouring operations
$246,000
TOTAL
$674,000
64
-------
Phase II tasks include design of the PRBs and associated monitoring networks. The cost of Phase
II tasks was $148,000 and occurred within a 5-month interval. In an actual remediation
deployment, the cost and timeframe of these activities could be significantly reduced for the
following reasons: (1) the demonstration of three PRBs in one structure involved additional
design considerations and costs; (2) the costs needed to design a monitoring network for three
PRBs was much more extensive than designing a monitoring network for a single PRB in a
remediation application that would probably have less stringent monitoring needs; and (3)
permitting issues for the demonstration project delayed the actual start date by about 2.5
months.
Phase III tasks included installation of three PRBs and the associated monitoring network. The
cost of Phase III tasks was $246,000 and involved approximately 16 days of effort. A detailed
breakdown of the installation cost is listed in table 8.1. The time frame and cost of Phase III
activities could be higher or lower during the installation of a single PRB for groundwater
remediation. PRB installation at the Fry Canyon demonstration site may be more cost effective
than at other sites due to the following factors: (1) less than 15 ft to groundwater; (2) average
saturated thickness less than 5 ft; (3) tailings and overburden removed during trench excavation
did not require special handling; and (4) only 21 linear ft of PRB was emplaced.
PRB installation at the Fry Canyon demonstration site may be less cost effective than at other
sites due to the following factors: (1) only small quantities of reactive materials were purchased,
negating any bulk discounts that may have been available; (2) a large number of monitoring points
were required to meet the objectives of a demonstration project using three PRBs; (3) twelve
transducers, three water-quality minimonitors, and two automated data loggers were required to
monitor PRB performance; and (4) the demonstration site is in a relatively remote site causing
increased shipping costs and travel costs for construction, scientific, and management personnel.
9.0 REMAINING QUESTIONS FOR RPMs DURING THE RI/FS
This study has addressed a number of issues associated with the use of PRBs for the removal of
U from groundwater; however, important questions that a site RPM might have about using this
technology for ground water remediation remain unanswered. Some of these questions can be
addressed by continued work at the Fry Canyon site. Questions are listed below:
1. Is the long-term performance of PRBs for removal of inorganic contaminants, including
uranium, cost effective compared to other contaminant removal technologies currently
available?
2. What methods can be used to determine if PRBs will be an effective remediation technology
at a particular site for the removal of inorganic contaminants? Specific issues that need to be
addressed include: (1) correlation of laboratory column and batch tests to actual field
performance of PRBs; (2) specific inorganic contaminants that can be removed from
groundwater by PRBs; and (3) type of ground water systems amenable for PRB deployment.
3. How long will a PRB remove inorganic contaminants to concentrations at or below the clean-
up goals for a particular site? Specific issues that need to be addressed include: (1) effective
lifetime of the PRB; (2) important geochemical processes negatively affecting PRB
performance (clogging, passivation, dissolution of barrier material, and re-mobilization of
contaminants); and (3) understanding the geochemical and hydrologic processes that may
65
-------
reduce long-term contaminant removal efficiencies of PRBs. For example: 1. The surface of
the ZVI barrier material may become coated by either carbonate or iron hydroxide mineral
phases.; 2. Chemical precipitation or microbial growth may reduce PRB permeability.; 3. The
mechanism(s) of contaminant removal by the reactive material will influence the potential for
re-release of contaminant. For example, a removal reaction resulting in the uranium being
precipitated in a mineral structure may potentially be more resistant to re-release than an
adsorption removal process.
4. Can PRBs remain in place or will removal and proper disposal of barrier material become
necessary for inorganic contaminants to remain unavailable to the post-remediation
groundwater?
5. What are the regulatory issues associated with long-term disposal of material from PRBs?
6. What are the positive and negative effects of microbes on PRB performance?
7. What are the potential detrimental effects of the barrier material itself to downgradient water
quality?
8. How will the quantity and major ion chemical composition of groundwater flowing through a
PRB affect expected PRB longevity and inorganic contaminant removal efficiency?
10.0 RECOMMENDATIONS FOR PRB IMPLEMENTATION: LESSONS
LEARNED
Based on the installation of the three PRBs at the Fry Canyon demonstration site, a number of
improvements and issues should be considered during future, full-scale installations of PRB for
site remediation. The following list of considerations is not exhaustive, rather it reflects project-
specific observations based on experiences at Fry Canyon:
1. The uneven surface of the underlying confining unit made it difficult to ensure that each PRB
gate structure or no-flow barrier was in direct contact with the underlying confining unit. If
either structure was placed on small lenses of the residual colluvial aquifer, this may have
provided a pathway for contaminated groundwater to bypass the reactive material.
A possible solution to prevent this would be the use of a more powerful track hoe that would
be able to excavate into the underlying confining unit. This equipment would allow for a
smooth surface and a gradient could be established that would drain the groundwater away
during excavation, allowing the observation of the underlying confining unit. In addition, the
use of pumps with a capacity exceeding the groundwater inflow will allow for visual
inspections of the seal between the PRB and the confining layer.
2. The use of pre-mixed bentonite slurry for construction of the no-flow barriers was
problematic. It was difficult to control the movement of slurry from the wing wall to the gate
structure of each PRB. It is critical to know exactly where the bentonite slurry is. If the
slurry flows into a gate structure it could impact the flow and treatment of contaminated
groundwater in the finished PRB. In a worst case scenario, the gate structure of the PRB
could be sealed off, preventing the treatment of contaminated groundwater.
A possible solution would be the use of non-hydrated bentonite chips for the construction of
PRB wing walls and associated no-flow barriers. After placement, the chips would hydrate
with the natural groundwater, ensuring correct placement during PRB construction. In
addition, the use of bentonite chips would not require the added expense of cement mixers for
slurry transport.
66
-------
3. The placement of monitoring wells within the wing walls and other no-flow areas between
PRB gate structures is important to ensure proper operation. Including these wells in the
routine monitoring network can provide critical water-level and water-quality data useful in
assessing PRB operation. For example, water levels measured in wing wall monitoring wells
would be expected to respond more slowly to naturally occurring water-level increases
observed within the PRB gate structures.
4. A large bedrock nose was encountered during PRB installation that resulted in a re-orientation
of the PRBs. This re-orientation resulted in the entry of groundwater at an oblique angle into
the PRB gate structures, rather than the perpendicular angle that was anticipated.
In order to prevent this problem in future PRB installations, a more detailed view of the
bedrock topography is needed during site characterization activities. Additional data on
bedrock topography could be obtained by increased drilling density during pre-installation
characterization activities or possibly by subsurface geophysical methods such as seismic or
ground penetrating radar techniques.
5. During the pre-installation characterization, it is important to determine the amount of readily
desorbable U contained in the contaminated aquifer sediments. In a remediation scenario
where the source term is either removed or stabilized in place, the total mass of readily
desorbed U will eventually pass through the PRB. Quantification of this mass is needed to
properly design the contaminant removal capacity of the PRB prior to emplacement.
6. Numerous hydrologic and water-quality characteristics should be considered prior to the
selection of an appropriate point of compliance (POC) well when installing a PRB for
ground-water remediation. When a PRB is placed within the contaminant plume, the POC
well should probably be placed within the PRB. If the POC well is placed downgradient of
the PRB it is likely that contaminant free water exiting the PRB could become re-
contaminated with readily desorbable contaminants from the aquifer sediments.
For some PRB materials, such as ZVI, the placement of a POC well within the PRB could be
problematic. For example, the high iron concentrations and high pH values of water within
the Fry Canyon ZVI PRB may not meet water-quality compliance standards. However, if the
POC wells were placed downgradient of the ZVI PRB, the high iron concentration and high
pH values would be significantly reduced. In many situations the location of POC wells may
have to be parameter specific depending on the barrier material and on-site hydrologic and
geochemical conditions.
7. Quantification of groundwater flow during pre-installation characterization and after PRB
emplacement is critical. This information is needed for PRB design and to monitor changes in
PRB hydraulic conductivity after emplacement. Groundwater models provide adequate
information to address barrier design issues during the pre-installation characterization phase.
After PRB emplacement, tracer injection methods appear to be best suited for monitoring
PRB performance over time. Results to date (1999) at the Fry Canyon site have indicated
that in-situ flow sensors are of limited use in monitoring changes in either ground-water flow
or direction within PRBs; however, these data are still being evaluated.
67
-------
Appendix A
Summary of Activities at Other Sites Using Permeable Reactive Barriers to
Remove Uranium
Currently (1999) there are 45 field projects involving the use of the PRBs to treat contaminated
groundwater (Appendix A, table A.I). Including the Fry Canyon demonstration site, 6 of the 44
PRB field projects are treating water containing uranium. A PRB was recently installed at the
Rocky Flats DOE facility in Colorado; however, no results are currently (1999) available. The
Rocky Flats PRB is mainly intended to treat cVOCs and only trace concentrations of U are
present. At Oak Ridge, TN the DOE has installed 2 PRBs to treat uranium; Channel and South
Plume. No information is currently (1999) available about the South Plume site.
68
-------
Table A.l. Other permeable reactive barrier field projects.
compound (magnesium peroxide); F&G, funnel and gate; IO, iron oxide;
No. Countrv/State/Prov Site Contaminants Install Date Status
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Alabama
British Columbia
California
California
California
California
California
California
California
California
Colorado
Colorado
Colorado
Colorado
Delaware
Florida
Florida
Florida
Ireland
Kansas
Kentucky
New Hampshire
New Jersey
New Jersey
Maxwell AFB
Vancouver
Chico
Alameda
Ft. Bragg
Moffat Field
Mountain View
Newbury Park
Sunnyvale
Valley Wood
Federal Center
Lowry AFB
Durango
Rocky Flats
Dover AFB
Cape Canaveral
Cape Canaveral
Cape Kennedy
Belfast
Coffeyville
Paducah
Summersworth
Caldwell Trucking
Fairfield
cVOC
Pb, Zn, Cu, Cd
TCE
cVOC
cVOC
cVOC
cVOC
cVOC
cVOC
Cr
cVOC
cVOC
U, Mo, As, Se
cVOC, U
cVOC
cVOC
cVOC
cVOC
cVOC
cVOC
cVOC
cVOC
cVOC
cVOC
9
3/97
1995
12/96
9
4/96
9/95
7
1/95
1997?
10/96
7
10/95
9/98
7
10/97
11/97
7
12/95
1/96
1995
1997
3/98
9/98
Installing now?
Operating
Pilot Complete
Operating
Operating
Pilot operating
Full Scale
Pilot operating
Operating
Operating?
Operating
Operating
Operating
Operating
Operating?
Operating
Operating
Operating?
Operating
Operating
Operating
Operating
Operating
Operating
P, phosphates; GAC, granular activate charo
Depth Reactant Tvrje Designation
75ft
24ft
100ft
9
7
20ft?
20ft
87ft
30ft
9
25ft
25ft
5ft
5ft
7
45ft
45ft
7
40ft
28ft
9
9
7
25ft
ZVI
Compost
Microbes
ZVI
GAC
ZVI
ZVI
ZVI foam
ZVI
Reductant
ZVI
ZVI
ZVI
ZVI
ZVI
ZVI
ZVI
ZVI/Sonic
ZVI
ZVI
ZVI/+
ZVI
ZVI
ZVI/Sand
Frac/Jet
Trench
Injection
F&G
Cannister
F&G
Trench
Fracing
F&G
Injection
F&G
F&G
Cannister
Cannister
F&G?
Mandrel
Jetting
Deep Mixing
Cannister
F&G
Mandrel
F&G
Fracing
Trench
Commercial
Pilot
Pilot
Pilot
Commercial
Pilot
Commercial
Pilot
Commercial
Commercial
Commercial
Pilot
Pilot
Commercial
Pilot?
Pilot
Pilot
Pilot
Commercial
Commercial
Pilot
Pilot
Commercial
Commercial
ON
-------
Table A.l. (Continued)
No.
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
Countrv/State/Prov
New Jersey
New Mexico
New Mexico
New York
New York
North Carolina
Ohio
Ohio
Ontario
Ontario
Oregon
South Carolina
South Carolina
South Carolina
Tennessess
Tennessee
Tennesse
Utah
Washington
Washington
Wyoming
Site
Parlin
Belen
Sandia
West Valley
Sherburne
Elizabeth City
Portsmouth
Portsmouth
Borden CFB
Nickel Rim
Unnamed
SRS-D Area
SRS-Siphon
Manning
Y12 South Plume
Y12 Channel
WAGS
Fry Canyon
Hanford 100D
Hanford 100H
Christensen Ranch
Contaminants
none
BTEX
Cr/TCE/CC14
90 c,
Sr
cVOC
cVOC, Cr
cVOC
cVOC
cVOC
Metals/S04
cVOC
metals/TCE/SO4
TCE
cVOC
U/Tc/PCE/N03
U/Tc/N03
90Sr
U
Cr
Cr
U
Install Date
1997
9
1997
7
12/97
6/96
1998
1996
1993
8/95
1998
9
7/97
1998?
12/97
12/97
11/94
9/97
9/97
9/95
10/97
Status
No contaminant
Operating
Completed
7
Operating
Operating
Operating
Operating?
Completed
Operating
Operating
9
Operating
Operating
7
Operating
Operating
Operating
Operating
Operating
Operating
Depth
15ft
9
8ft
9
15ft
26ft
9
piped
8ft
10ft
30ft
9
15ft
29ft
25ft?
30ft
12ft
15ft
100ft
100ft
440ft
Reactant
ZVI
ORC
ZVI/GAC/+
Zeolite
ZVI
ZVI
MnO4
ZVI/+
ZVI/Sand
Wood chips/+
ZVI
9
ZVI
ZVI
9
ZVI
Zeolite
ZVI/IO/P
Dithionite
Dithionite
P/IO
Type
Jetting demo
Socks
Jetting
9
F&G
Cont. trencher
Injection
Cannisters
F&G
Trenching
Cont. Trencher
Environwall
Geosiphon
Cont Trencher
Concrete Vault
Trench
Concrete Vault
F&G
Injection
Injection
Barrier package
Designation
Demonstration
Commercial
Demonstration
Commercial
Commercial
Commercial
Pilot
Pilot
Pilot
Pilot
Commercial
Commercial
Pilot
Commercial
Commercial
Commercial
Commercial
Pilot
Pilot
Pilot
Pilot
-------
Oak Ridge National Laboratory Y-12 Plant, Channel Site, TN. A PRB was installed at the
DOE Oak Ridge National Laboratory in December 1997. Eighty tons of ZVI was emplaced in a
trench using guar gum to keep the trench open (Gu and others, 1998). An enzyme breaker was
added that causes the guar gum to dissipate into the groundwater to minimize permeability loss.
The trench penetrates a shallow groundwater aquifer contaminated by uranium, technetium-99,
and nitrate. Preliminary results indicate that the PRB is effectively removing all the
contaminants. Ferrous iron concentrations in the groundwater were high shortly after
emplacement but have since decreased and the pH has increased to greater than 9. The guar gum
caused an increase in microbial activity and some problems were encountered by the guar failing
to breakdown properly and flow out of the system.
Uranium Mill Tailings Repository Site at Durango, CO. A seep coming from a U mill
tailings repository is currently (1999) collected in a gravel trench and then piped to an
underground steel tank containing foam blocks of ZVI (Morrison, 1998). This PRB has been
operating intermittently since May 1996. The ZVI has consistently reduced uranium,
molybdenum, and nitrate to concentrations below water quality standards. Concentrations of U
contained in the ZVI were as high as 2% by weight; and concentrations of vanadium as high as
10% by weight. After about 2 years of operation, the tank became clogged and was no longer
allowing the passage of water. Gas bubbles were observed during excavation of the tank. The gas
composition included hydrogen and methane. The gases were a likely cause of the blockage.
71
-------
Appendix B
Summary of Deep Emplacement Methods for Permeable Reactive Barriers
New techniques are being developed to increase the depth range for emplacement of PRBs. Six
new methods have been field tested: (1) vertical hydraulic fracturing, (2) jet grouting, (3) chemical
injection, (4) deep soil mixing, (5) driven mandrels, and (6) arrays of non-pumping wells.
Following are summaries of recent developments in these deep emplacement technologies. In
addition to increasing the emplacement depths, these methods have two other advantages over
trench and fill installation: (1) ZVI can be emplaced more efficiently around utility lines, and (2)
workers are afforded more protection because less contaminated material is exposed at the
surface.
Vertical hydraulic fracturing. Vertical hydraulic fracturing is initiated by driving a spade-
shaped tool into a borehole in the desired location. Water is not viscous enough to suspend and
carry the ZVI to the fracture zone. To increase the viscosity, guar gum is mixed with the water
and ZVI. An enzyme is added to the mixture that is designed to cause the guar to degrade after
several days. Pumping the grout (ZVI with water and guar) into the incipient fracture causes the
fractures to expand and lengthen. The PRB wall is emplaced by fracturing in a series of adjacent
borings.
Vertical hydraulic fracturing has been used at only one site to emplace reactive materials; the
Cal dwell Trucking facility in New Jersey (Appendix A, table A.I). About 110 tons of ZVI were
injected to form two parallel walls about 3 inches thick each (Hocking and others, 1998). The
PRB is about 65 ft deep and 150 ft long and is being used to degrade cVOCs in groundwater.
This project demonstrated that ZVI could be emplaced into unconsolidated rocks by creating
vertical fracture zones. ZVI has also been emplaced as a proponent during hydraulic fracturing of
jointed bedrock at the Newbury Park site in California (Marcus and Farrell, 1998). This
demonstration project was also aimed at controlling cVOC contamination in groundwater.
Jet grouting. Jet grouting has been used for some time to emplace cement or cement/bentonite
impermeable barriers but only recently has been used to emplace PRBs. Grouting is performed
by lowering a jetting tool into a borehole. As the tool is withdrawn upwards, grout is pumped
through a small orifice at pressures often exceeding 5000 psi. As with hydraulic fracturing, the
grout consists of ZVI suspended in a guar gum and water mixture containing an enzymatic
breaker. The grout mixes turbulently with the sediments and some of the sediment is brought to
the surface through the annulus around the drill pipe. The PRB is formed by successively jetting
a line of boreholes. The jetting tool is rotated during extraction to form columns of ZVI with
diameters up to 8 ft. Alternatively, "thin diaphragm" configurations with a thickness of about 3
to 6 inches are formed if the tool is not rotated.
Jet grouting has been used to emplace ZVI at three sites (Appendix A, table A.I): (1) Cape
Canaveral Air Force Station, Florida, (2) a Du Pont facility at Parlin, New Jersey, and (3) Sandia
National Laboratory, New Mexico. At Cape Canaveral, high-pressure water jetting was used to
cut a slot in coastal-plain sands contaminated with cVOCs (Marchand and others, 1998). The
slot was then expanded and filled by injecting ZVI mixed with guar gum, cross linking compound,
enzyme breaker, and water. The resulting PRB was 70 ft in length, about 4 inches thick, and 45
ft deep. This project demonstrated that ZVI could be emplaced using jetting and the resulting
PRB would degrade cVOCs. At Parlin, ZVI suspended in water with guar and enyzme breaker
72
-------
was injected under 5000 psi into the upper 15 ft of unsaturated sands and gravels (Landis, 1997).
The project demonstrated the viability of the jet grouting method to emplace columns and thin
diaphragms of ZVI; however, no contaminant was present. At Sandia, jet grouting was used to
emplace grouts in 6-foot diameter, 8-foot deep cylindrical culverts (Dwyer, 1998). Eight types
of grout were used, each in a separate installation: (1) ZVI with guar, (2) cement slag, (3)
activated carbon, (4) colloidal ZVI, (5) mordenite (zeolite), (6) clinoptilolite (zeolite), (7) ZVI
with zeolite and guar, and (8) ZVI with cement slag and guar. Contaminants were not present at
the site but chromium (VI), TCE, and CC14 were introduced after the jetting to determine the
effectiveness of the reaction zones. This demonstration indicated that these materials could be
emplaced by jet grouting and that the desired reactions were occurring.
Chemical injection. Chemical injection involves the introduction of reactive materials through
wells. The materials are dissolved in water and the solution is passed into the formation by
exerting a hydrostatic pressure (either via gravity or pumping) at the well head. Injections are
made in a line of adjacent wells to form the PRB. Reactants that have been injected include
microbes, microbial nutrients, chemical reductants, and oxidants.
Four field projects have been conducted to test chemical injection emplacement methods
(Appendix A, table A.I): (1) Chico Municipal Airport, California, (2) DOE Facility at
Portsmouth, Ohio, (3) Hanford 100D Area, Washington, and (4) Hanford 100H Area,
Washington. At Chico, resting-state microbial cells were injected into a groundwater plume
containing TCE (Duba and others, 1996). About 50% of the cells attached to the subsurface
sediments, forming a reaction zone. In the reaction zone, TCE was degraded by enzymes present
in the microbes. This project demonstrated that resting-state microbes could be emplaced by
injection through wells and that once injected they would degrade TCE. At the Portsmouth site,
potassium permanganate was injected into the subsurface through wells. Few results are
available but apparently the oxidation was able to cause the degradation of TCE in the
groundwater. This project demonstrated an alternative to using reductive dechlorination by ZVI
for TCE degradation, and that oxidation of the subsurface through injection was possible. At the
Hanford 100D and 100H areas, sodium dithionite solution was injected into the subsurface
through wells (Fruchter and others, 1998). Dithionite is a reduced sulfur compound that is
capable of reducing ferric iron to ferrous iron. The ferric iron is contained in silicates and oxides
of the sediment. The ferrous iron formed by the dithionite injection is capable of reducing Cr(VI)
to Cr(III) which precipitates out of the groundwater as a hydroxide mineral. The ferrous iron is
also capable of degrading TCE by reductive dechlorination and precipitation of U as uraninite. A
PRB was formed by multiple injections of dithionite in adjacent wells. This project showed that
dithionite could be injected to form a chemically reducing zone capable of remediating
groundwater.
Deep soil mixing. Deep soil mixing involves the use of large (up to 12-foot diameter) augers that
mix the subsurface soils as they are rotated into it (Shoemaker and others, 1995). To form a
PRB, reactive materials are augured into the soil. Deep soil mixing can emplace reactive materials
up to 150 ft deep in unconsolidated sediments.
Deep soil mixing has been used to emplace reactive material at only one site: Cape Kennedy
Space Center, Florida (Appendix A, table A.I). Details of this demonstration are not available
but apparently the project demonstrated that emplacing ZVI using deep soil mixing is feasible.
73
-------
Driven mandrels. A hollow tube (the mandrel) can be driven into the subsurface and filled with
reactive material (Shoemaker et al.,1995). When the mandrel is pulled back out, the reactive
material remains in the ground. A "shoe" is placed on the lower end of the mandrel to prevent it
from filling with sediment during the insertion. The shoe remains in the ground during
withdrawal. A mandrel with a rectangular cross section of up to 30 inches by 6 inches can be
driven up to 120 ft in unconsolidated sediments. A PRB is formed by repeatedly inserting the
mandrel at adjacent locations.
Reactive materials have been emplaced at two sites using a mandrel: (1) Paducah, Kentucky and
(2) Cape Canaveral Air Force Station, Florida. At Paducah, a mandrel was used to install ZVI
and granular activated carbon in a tight clay-rich formation. Groundwater containing TCE was
moved through the reaction zone using electrokinetics and in the reaction zone it was either
adsorbed or degraded. This project demonstrated the feasibility of mandrel emplacements to
about 25-foot depths. At Cape Canaveral, a mandrel was used to emplace ZVI to a depth of 45
ft (Marchand and others, 1998). A 70-foot long by 4-inch thick zone of ZVI was constructed to
degrade cVOCs. This project demonstrated that a large-scale PRB could be installed at a
reasonable cost using a mandrel and that the PRB is capable of degrading cVOCs.
Arrays of non-pumping wells. Use of arrays of unpumped wells has been proposed by Wilson
and Mackay (1997) as a method to remediate contaminant plumes when the installation of
treatment walls is not possible because of technical or financial constraints. This type of
deployment technology is useful for treatment of deeper contaminant plumes. Barrier
deployment tubes (Naftz and others, 1999) are used to place reactive material into the arrays of
non-pumping wells. The combination of barrier deployment tubes with arrays of non-pumping
wells allows for the cost-effective retrieval and replacement of reactive material, which would not
be possible with other deployment technologies.
Under natural flow conditions, groundwater converges to non-pumping well arrays and the
associated barrier deployment tubes in response to the difference in hydraulic conductivity
between the well and aquifer. Numerical simulations of ground-water movement through a non-
pumping well array indicate that each well intercepts groundwater in a portion of the upgradient
aquifer approximately twice the inside diameter of the well (Naftz and others, 1999).
Reactive materials have been emplaced in arrays of non-pumping wells using barrier deployment
tubes at two sites: (1) Christensen Ranch In-Situ U Mine, Wyoming and (2) Fry Canyon, Utah
(Appendix A, table A.I). Barrier packages containing mixtures of bone-char phosphate and iron
oxide were deployed into groundwater at the Christensen Ranch site with a U concentration of
20,000 |ig/L to depths exceeding 430 ft below land surface. Initial U removal efficiencies exceeded
99.9% during a 7-month deployment period at the Christensen Ranch site. Both projects are in
progress.
74
-------
Appendix C
Site Evaluation and Selection Process
During a reconnaissance-phase investigation, four sites were considered for the demonstration of
PRBs to remove U from groundwater (Appendix C, fig. C. 1). Based on the results of the
reconnaissance-phase investigation, three of the four candidate sites had sufficient groundwater
for a viable demonstration project. The Tony M Mine U tailings were dry and eliminated from
further consideration. The Blue Cap and Firefly-Pygmy mine sites contained steep tailings slopes
that made them unsuitable as potential demonstration sites. The steep and potentially unstable
tailings slopes would incur high costs for the drilling and installation of monitoring wells.
Conducting a long-term demonstration project on steep slopes would result in unnecessary
safety issues. In addition, the measured U concentration in a water sample from the Firefly-
Pygmy mine site was low (Appendix C, fig. C.I).
75
-------
Fry Canyon site. Measured
uranium concentration in
ground water was 3,160
M9/L.
Tony M Mine
uranium tailings.
No ground water
was present.
Blue Cap mine. Uranium
concentration ranged from
575 to 1,040 ug/L in
samples of water
discharging from the audit
and below the tailings.
Firefly-Pygmy Mine.
Uranium concentration
measured at the base of
the tailings was 115 ug/L.
Figure C.I. Location and uranium concentration in water samples from abandoned mine
sites considered for field demonstration of permeable reactive barriers.
-------
The following characteristics were considered during the reconnaissance-stage investigation:
(1) extent and type of groundwater contamination;
(2) depth to contaminated groundwater;
(3) complexity of groundwater flow system;
(4) past, present, and future status of site clean up;
(5) ownership;
(6) topography;
(7) access;
(8) historical and background information;
(9) climate;
(10) transferability;
(11) health and safety issues.
During the reconnaissance-phase investigation the Fry Canyon site appeared to contained
numerous favorable characteristics for the long-term field demonstration of PRBs. These
favorable characteristics included: (1) groundwater containing U concentrations exceeding 3,000
|ig/L; (2) riparian vegetation indicating the presence of shallow groundwater; (3) a No Further
Action Planned rating for the site in 1990, thus preventing the potential for clean-up activities at
the site during field demonstration; (4) the site is managed by BLM; (5) access roads to the
tailings exist; (6) flat surface topography that is conducive to drilling, monitoring, and
construction activities; (7) a moderate climate at the site that is conducive to year-round site
access and monitoring; (8) the site is typical of other abandoned mine sites in the arid- and semi-
arid Western United States.
77
-------
Appendix D
Health and Safety Issues Associated with Installation of Permeable Reactive
Barriers
Health and safety issues specifically associated with PRB installation were separated into three
distinct activities. Activity 1 consisted of excavating the tailings and colluvium in order to expose
the saturated zone of the aquifer system. Activity 2 consisted of excavating in the saturated zone
to the confining unit beneath the colluvial aquifer. Activity 3 was filling the trench with the
appropriate chemicals for each of the PRBs.
Each of the construction activities had different hazards associated with them. Primary hazards
associated with Activity 1 was worker exposure to ionizing radiation, radioactive materials, and
silica dust. Trench collapse was the main hazard associated with activity 2. Exposure to PRB
chemicals was the primary hazard associated with activity 3. Table D.I details the toxicity
characteristics of the PRB chemicals used at the Fry Canyon demonstration site.
Table D.I. Permeable reactive barrier chemicals used at the Fry Canyon site and the associated
toxicity characteristics.
[PEL/TLV-TWA, permissible exposure limit/threshold limit value-time weighted average]
Chemical
Amorphous ferric
oxide
Foamed zero-
valent iron
Bone char
(phosphate)
Routes of
exposure
Inhalation, skin
and eye contact,
ingestion
Inhalation, skin
and eye contact,
ingestion
Inhalation, skin
and eye contact,
ingestion
Target organs
Eyes,
respiratory
system
Eyes,
respiratory
system
Eyes and
respiratory
system
Effects of exposure
Acute: mild skin and
eye irritation.
Chronic: fibrosis and
liver cirrhosis if large
quantities are inhaled.
Acute: bronchitis, eye
irritation.
Chronic: siderosis.
Acute: irritation of
nasal and respiratory
passage, eye irritation.
Chronic: none.
PEL/TLV-TWA
Not available
Not available
Not available
Safety procedures were instituted for each of the three construction activities to minimize the
associated hazards. During activity 1 (excavation to the saturated zone) the following safety
procedures were followed: (1) a qualified Radiological Control Technician monitored radiation
hazards and (2) respirable dust levels were monitored by the USGS Health and Safety Officer to
minimize exposure to silica dust. During activity 2 (excavation below the saturated zone) a trench
box was placed on the bedrock surface within the trench (fig. 5.1) to allow workers to enter the
trench prior to backfilling operations. During activity 3 (filling the trench with PRB chemicals)
workers were protected from contact with the chemicals by wearing personal protective
equipment as deemed appropriate. In addition, all workers practiced as low as reasonably
achievable (ALRA) techniques to minimize exposure to the chemical materials during the
backfilling operations.
78
-------
Appendix E
Sample Collection, Analysis, Quality Assurance, and Field Measurement
Calibration
Sample Collection. During the pre-installation monitoring phase, water was purged from each
well until the field parameters (pH, dissolved oxygen, oxidation-reduction potential, specific
conductance, and water temperature) stabilized. The shallow water table allowed for the
collection of groundwater samples with a peristaltic pump. Stabilization of field parameters
usually occurred after four to five gallons of water was removed from the well. At a minimum,
this volume represented the removal of greater than 3 casing volumes from the 2-inch diameter
monitoring wells.
The volume of water extracted from each well was modified during the PRB monitoring phase
because of the close proximity of monitoring points within and adjacent to the PRBs. Less water
was purged from the 2-inch wells to minimize the creation of pumping induced gradients within
and adjacent to the PRBs during monitoring activities. 1 gallon of water was removed from the
2-inch diameter monitoring wells and 1 liter of water was removed from the 0.25-inch
monitoring wells prior to sample collection. With a maximum saturated thickness of five ft in the
PO4 barrier (fig. 5.3), the extraction of 1 gallon of water from the 2-inch monitoring wells
represents approximately 1.2 casing volumes. Extraction of 1 liter of water from the 0.25-inch
monitoring wells represents approximately 20.8 casing volumes.
During September 1998 a total of 4 gallons of water was removed from the 2-inch diameter well
ZVIFSI in the ZVI PRB to address changes in water chemistry as a function of pumpage
volumes. Water samples were collected and analyzed after every gallon that was pumped
(Appendix E, table E. 1). Results from the chemical analysis indicate that U concentration and
field parameters are stable sometime after 1 gallon and sometime before 2 gallons of water have
been removed from the well casing. These results indicate that removal of 1 gallon of water from
each 2-inch diameter well completed within and adjacent to the barrier is probably the best
compromise to obtain a representative groundwater sample while not imposing pumping induced
gradients within the PRBs. The higher hydraulic conductivity of the barrier materials and pea
gravel relative to the native aquifer material is the probable reason that smaller purge volumes are
needed from the wells completed in the PRB and pea gravel.
Table E.I. Uranium concentration and pH, specific conductance, and oxidation-reduction
potential value changes during pumping of well ZVIFSI during September 1998, Fry Canyon,
Utah.
[|lg/L, milligrams per liter; |iS/cm, microsiemens per centimeter; mV, millivolts]
Volume of water
removed during
pumping of well
ZVIFS1 in gallons
1
2
3
4
Uranium
concentration
in |^g/L
< 0.06
< 0.06
< 0.06
< 0.06
pH, in
units
8.88
9.05
9.04
9.03
Specific conductance
in |^S/cm
1,760
1,760
1,760
1,770
Oxidation-
reduction
potential in
mV
-272
-235
-222
-222
79
-------
After purging, water samples were filtered on site using a 0.45 |im capsule filter and collected in
field-rinsed polyethylene bottles. Samples for analysis of uranium, aluminum, calcium, copper,
iron, potassium, lithium, magnesium, manganese, phosphorus, sodium, silicon, strontium, and
zinc were acidified on site with ultra-pure concentrated nitric acid to a pH<2. Approximately 100
ml of deionized water was then pumped through the tubing between samples. Each monitoring
point contains a dedicated sampling tube to minimize cross contamination. Surface-water samples
from Fry Creek were collected in field-rinsed, one-gallon containers and processed according to
the procedures used for groundwater samples.
Sample Analysis. A variety of analytical methods were used to determine the concentrations of
the chemical constituents that were monitored during the laboratory simulation, pre-installation,
and the first year of PRB operation. Water analyses were conducted at the USGS Research
Laboratories in Menlo Park, California. Dissolved U was determined by kinetic phosphorescence
analysis (KPA). The KPA-11A has a ±3% precision and 0.06 microgram/L detection limit.
Samples were diluted 1:10 in 0.1 M HNO3 to minimize potential chloride interference.
(Chemchek Instruments, Richland WA). Aluminum, calcium, copper, iron, lithium, magnesium,
phosphorus, manganese, sodium, silicon, strontium, and zinc concentrations were determined by
ICP/OES using a Thermo Jarrel Ash ICAP 61 (Standard Methods, 1992). Potassium was
determined by direct air-acetylene flame atomic absorption spectrometry (AA) using a Perkin
Elmer AA 603. Sulfate and chloride concentrations were measured by ion chromotography using
a Dionex Chromatograph CHB (Standard Methods, 1992).
Selected chemical and physical constituents in groundwater samples were determined in the field.
The pH, specific conductance, and temperature of each water sample was determined in a flow-
thru chamber using a Yellow Springs Instrument 600XL minimonitor that was calibrated daily
with respect to pH and specific conductance. Total alkalinity (as CaCO3) of filtered (0.45 |im)
water samples was determined on site using a HACK digital titrator and 1.6 normal sulfuric acid.
Ferrous iron and phosphate (as PO4) were determined on site with a colorimetric method using
CHEMetrics self-filling vials and a portable photometer.
Quality Assurance. A series of quality assurance (QA) samples were collected and processed
with the routine water-quality samples collected from within and adjacent to the PRBs during
September 1997 through September 1998. The QA samples collected during pre-installation and
year 1 barrier monitoring activities consisted of process blanks and field duplicates.
Deionized, distilled water of known major-, minor, and trace-element composition from the
USGS Water-Quality Service Unit, Ocala, Florida, was used for the process blanks during each
sampling trip. The process blanks were used to assess contamination during field processing of
water samples. Ten process-blank samples were analyzed (Appendix E, table E.2). Median
concentrations for the process blanks were all below the analytical detection limits (Appendix E,
table E.2). Uranium concentrations in the process blank samples never exceeded the analytical
detection limit of 0.06 |lg/L.
80
-------
Table E.2. Chemical analysis of selected major-, minor-, and trace-element constituents from
blank samples processed during pre-installation characterization and year 1 barrier monitoring
activities, Fry Canyon, Utah, September 1996 to September 1998.
[mg/L, milligrams per liter; <, less than reported value; |lg/L, micrograms per liter]
Chemical
constituent
BUranium
Aluminum
Calcium,
Copper
Iron
Potassium
Lithium
Magnesium
Manganese
Phosphorus
Sodium
Silicon
Strontium
Zinc
Sulfate
Chloride
Unit
V&L
V&L
mg/L
V&L
V&L
mg/L
mg/L
mg/L
V&L
mg/L
mg/L
mg/L
mg/L
Vg/L
mg/L
mg/L
Number of
samples
10
6
6
6
6
6
6
6
6
6
6
6
6
5
6
6
Median
concentration
<0.06
<50
<0.01
<5
<20
<0.05
<0.04
<0.04
< 10
<0.1
<30
<0.05
< 0.015
< 10
< 1
<0.25
Low
concentration
<0.06
<50
<0.01
<5
<20
<0.05
<0.04
<0.04
< 10
<0.1
<30
<0.05
< 0.015
< 10
< 1
<0.25
High
concentration
<0.06
50
0.2
<5
30
<0.04
<0.04
<0.04
< 10
<0.1
<30
0.13
< 0.015
< 10
< 1
10
A total of 11 field duplicates were analyzed to ensure consistency in the methods used to collect
the water samples during pre-installation and year 1 barrier monitoring activities. At least one
duplicate sample from a randomly selected well was collected, processed, and submitted for
chemical analyses during each sampling trip from September 1996 through September 1998
(Appendix E, table E.3). The majority of field duplicate results are within plus or minus 10% of
one another for all constituents. With respect to uranium, the duplicate sample collected during
January 1998 was 11.5% higher in U than the routine sample (Appendix E, table E.3). The
January 1998 sample was the only duplicate that exceeded the plus or minus 10% margin of
error.
81
-------
Table E.3. Chemical analysis of selected major-, minor-, and trace-element constituents from duplicate samples collected during pre-
installation characterization and year 1 barrier monitoring activities, Fry Canyon, Utah, September 1996 to September 1998.
[Al, aluminum; Ca, calcium; Cu, copper; Fe, iron; K, potassium; Li, lithium; Mg, magnesium; Mn, manganese; P, phosphorus; Na,
sodium; Si, silicon; Sr, strontium; Zn, zinc; U, uranium; SO4, sulfate; Cl, chloride; ND, not determined; mg/L, milligrams per liter; <,
less than reported value; |ig/L, micrograms per liter]
Well
FC5
DUPLICATE
FC1
DUPLICATE
FC2
DUPLICATE
ZVIR2-2
DUPLICATE
DG2
DUPLICATE
ZVIR2S-1
DUPLICATE
DG1
DUPLICATE
AFOR1S-1
DUPLICATE
ZVIR1S-7
DUPLICATE
DG4
DUPLICATE
TI1
DUPLICATE
Date
9/13/96
9/13/96
12/19/96
12/19/96
4/10/97
4/10/97
9/24/97
9/24/97
9/25/97
9/25/97
10/29/97
10/29/97
11/19/97
11/19/97
1/29/98
1/29/98
4/22/98
4/22/98
6/23/98
6/23/98
9/10/98
9/10/98
Al,
|ig/L
<50
<50
<50
<50
50
<50
ND
ND
ND
ND
<20
<20
ND
ND
40
60
ND
ND
<20
<20
ND
ND
Ca,
mg/L
59
60
48
50
120
120
ND
ND
ND
ND
130
130
ND
ND
170
180
ND
ND
100
100
ND
ND
Cu, |ig/L
<4
<4
<4
<4
10
<5
ND
ND
ND
ND
<5
<5
ND
ND
20
30
ND
ND
<2
<2
ND
ND
Fe,
l-ig/L
30
40
<20
<20
<20
<20
ND
ND
ND
ND
24.7
23
ND
ND
10
50
ND
ND
450
430
ND
ND
K, mg/L
5.4
5.4
4.0
4.5
4.05
4.05
ND
ND
ND
ND
6.6
6.7
ND
ND
7.2
7.6
ND
ND
0.1
0.1
ND
ND
Li,
mg/L
1.6
1.6
1.4
1.3
1.9
1.8
ND
ND
ND
ND
0.2
0.2
ND
ND
0.2
0.2
ND
ND
60
60
ND
ND
Mg,
mg/L
48
48
48
51
71
71
ND
ND
ND
ND
70
70
ND
ND
70
70.0
ND
ND
0.3
0.2
ND
ND
Mn,
!^g/L
160
160
<10
<10
180
170
ND
ND
ND
ND
500
500
ND
ND
0.2
0.2
ND
ND
0.3
<0.1
ND
ND
P,mg/L
<0.1
<0.1
<0.1
<0.1
<0.1
<0.1
ND
ND
ND
ND
7.6
10
ND
ND
0.4
0.4
ND
ND
3.7
3.9
ND
ND
Na,
mg/L
280
290
310
310
270
280
ND
ND
ND
ND
300
300
ND
ND
240
240
ND
ND
300
300
ND
ND
Si, mg/L
6.2
6.0
7.7
7.8
4.9
4.9
ND
ND
ND
ND
5.2
5.3
ND
ND
2.4
2.7
ND
ND
5
5
ND
ND
Sr,
mg/L
1.2
1.2
1.3
1.3
1.6
1.6
ND
ND
ND
ND
0.4
0.4
ND
ND
1.7
1.7
ND
ND
1.4
1.4
ND
ND
Zn,
|ig/L
20
70
<10
<10
15
10
ND
ND
ND
ND
<4
<4
ND
ND
<2
<2
ND
ND
<2
<2
ND
ND
U, |ig/L
260
260
60
60
1,020
1,030
370
340
1,770
1,670
50
50
1,250
1,230
6,150
6,950
<0.06
<0.06
720
720
7,520
7,430
SO4,
mg/L
310
310
330
330
630
630
ND
ND
ND
ND
640
740
ND
ND
810
880
ND
ND
490
490
ND
ND
Cl,
mg/L
110
110
110
120
110
110
ND
ND
ND
ND
110
110
ND
ND
110
100
ND
ND
120
120
ND
ND
-------
Field Measurement Calibration. The Yellow Springs Instrument 600XL water-quality
minimonitors and Waterlog H-310 pressure transducers were re-calibrated at approximately 5-
week intervals during the first year of barrier operation. The specific conductance, dissolved
oxygen, oxidation-reduction potential, and pH probes were calibrated on the water-quality
minimonitors. Specific conductance was calibrated using two standards with a known specific
conductance (0 and 2,500 mS/cm). Calibration of the dissolved oxygen probe was done in an air-
saturated chamber using a barometric pressure and temperature corrected value. A Zobell solution
relative to the silver-silver chloride/platinum electrode (corrected for temperature) was used to
calibrate the oxidation-reduction potential probe. The pH electrode was calibrated using both the
pH 7 and 10 standards.
The pressure transducers were calibrated by comparing the measured water level, using a
electronic measuring tape, with the water level recorded by the pressure transducer. If the water
level recorded by the transducer differed by more than +/- 0.03 feet from the actual water level,
the transducer was adjusted until the two measurements agreed.
83
-------
Appendix F
Glossary of Terms
Aerobic Presence of oxygen.
Anaerobic Absence of oxygen.
Aquifer An underground geologic formation that is filled with water and which is permeable
enough to transmit water to wells and springs.
Confining unit A body of material distinctly less permeable than the aquifer adjacent to it.
Darcy's Law Expressed by an equation that can be used to compute the quantity of water
flowing through an aquifer.
Effective porosity The volume of interconnected void spaces through which water or other fluids
can travel in a rock or sediment divided by the total volume of the rock or sediment.
Ex-situ The execution of an environmental cleanup by removing the contaminants from the
existing location to another matrix.
Groundwater Subsurface water found in the saturated zone below the water table in formations
known as aquifers.
Hydraulic conductivity A coefficient of proportionality describing the rate at which water can
move through a permeable medium.
Hydraulic gradient The rate of change of pressure head per unit of distance of flow at a given
point and in a given direction.
In-situ In place.
Milliequivalents The formula weight, in milligrams, of a dissolved ionic species divided by the
electrical charge.
Oxidation The loss of one or more electrons by a substance during a chemical reaction.
Permeable Pertains to the relative ease with which a porous medium can transmit a liquid under
a hydraulic or potential gradient.
Permeable reactive barrier Permanent, semi-permanent, or replaceable bodies of chemically
reactive material that are installed across the flow path of a contaminant plume.
pH The negative logarithm of the hydrogen ion activity in solution.
Phreatophyte A plant that obtains its main water supply from the saturated zone or through the
capillary fringe above the saturated zone.
84
-------
Total porosity Ratio of the volume of void spaces, interconnected and unconnected, in a rock or
sediment to the total volume.
Potentiometric surface The level to which water rises in a well. This level, generally called the
hydraulic head, is the sum of the elevation head and the pressure head. Elevation head is a result
of the elevation of the point in question above a datum, and pressure head is the height of the
column of water rises above the point in question.
Pump and treat Contaminant removal process where water is pumped from the aquifer into a
treatment cell, treated, and then pumped back into the aquifer.
Recharge The entry of water into the saturated zone.
Redox A chemical process where the loss (oxidation) and gain (reduction) of electrons among
reactants affect the charge of the medium and can be expressed as an oxidation reduction potential
(ORP).
Reduction The gain of one or more electrons by a substance during a chemical reaction.
Saturated zone Zone of porous medium in which all interconnected voids are filled with water.
Specific conductance An approximation of the salinity in a water sample. The reciprocal of
electrical resistivity.
Transmissivity The rate at which water at the prevailing density and viscosity is transmitted
through a unit width of the aquifer under a unit hydraulic gradient.
Transpiration The process by which water adsorbed by plants is discharged into the
atmosphere from the plant surface.
Water table That surface of a body of unconfined ground-water at which the pressure is equal to
that of the atmosphere.
85
-------
Appendix G
List of References
Abdelouas, A., Lutze, W., andNuttall, H.E. 1999. Oxidative dissolution of uraninite precipitated
onNavajo sandstone. Journal of Contaminant Hydrology, 36:353-375.
Arey, IS., Seaman, J.C., and Bertsch, P.M. 1999. Immobilization of uranium in contaminated
sediments by hydroxyapatite addition. Environmental Science and Technology. 33:337-342.
American Society for Testing and Materials. 1994. Standard Test Method for Permeability of
Granular Soils (Constant Head), D 2434-68 (Reapproved 1974). Reprint form Annual Book of
ASTM Standards, Philadelphia, PA.
Blowes, D. W. and Ptacek, C. J. 1992. Geochemical Remediation of Groundwater by Permeable
Reactive Walls: Removal of Chromate by Reaction With Iron-Bearing Solids. Proceedings of the
Subsurface Restoration Conference, Dallas, Texas, 1992. pp. 214-216.
Bostick, W. D., Jarabek, R. J., Slover, W. A., Fiedor, J. N., Farrell, J., and Helferich, R. 1996.
Zero-valent iron and metal oxides for the removal of soluble regulated metals in contaminated
groundwater at a DOE site. U. S. Department of Energy Report Number K/TSO-35P, Oak Ridge
National Laboratory, Tennessee. 64 pp.
Duba, A. G., Jackson, K. J., Jovanovich, M. C., Knapp, R. B., and Taylor, R. T. 1996. TCE
remediation using in-situ resting-state bioaugmentation. Environmental Science and Technology.
30:1982-1989.
Dwyer, B. 1998. Remediation of deep soil and groundwater contamination using jet grouting and
innovative materials. In Subsurface Barrier Technologies, Engineering Advancements and
Application Considerations for Innovative Barrier Technologies, compilation of papers from the
International Business Communications Conference, Tucson, Arizona, January, 1998.
Fruchter, J. S., Cole, C. R., Williams, M. D., Vermeul, V. R., Szecsody, J. E., and Evans, J. C.
1998. Recent progress on in-situ redox manipulation barriers for chromate and trichloroethylene.
In RTDF Permeable Reactive Barriers Action Team Meeting Presentation Materials, Beaverton,
Oregon, April 1998. pp. 117-130.
Freethey, G.W., Spangler, L.E., and Monheiser, W.J. 1994. Determination of hydrologic
properties needed to calculate average linear velocity and travel time of groundwater in the
principal aquifer underlying the southeastern part of Salt Lake Valley, Utah. U.S. Geological
Survey Water-Resources Investigations Report 92-4085. 30 p.
Gillham, R.W., and Burris, D.R.. 1992. Recent developments in permeable in situ treatment walls
for remediation of contaminated groundwater. Proceedings, Subsurface Restoration Conference,
June 21-24, Dallas, Texas.
Grand Junction Office. 1998. Results of field treatability studies of the Monticello, Utah, PeRT
wall. U. S. Department of Energy, Grand Junction Office Report GJO-98-69-TAR, 32 p.
86
-------
Grenthe, I, Fuger, J., Konings, R.J.M., Lemire, R.J., Muller, A.B., Nguyen-Trung, C., and
Wanner, H. 1992. Chemical thermodynamics of uranium. North-Holland, Amsterdam, 715 p.
Gu, B., Watson, D., Bogle, M. A., and Liang, L. 1998. Treatment of radionuclides and inorganic
contaminants in groundwater by permeable reactive barriers. Remediation Technology
Development Forum Permeable reactive Barriers Action Team Meeting Presentation Materials,
Oak Ridge, Tennessee, pp. 182-204.
Hocking, G., Wells, S. L., and Ospina, R. I. 1998. Performance of the iron reactive permeable
barrier at Calwell Superfund site. In Remediation Technology Development Forum Permeable
Reactive Barriers Action Team Meeting Presentation Materials, Oak Ridge, Tennessee,
November 1998, pp. 229-242.
Jobin, D.A. 1962. Relation of the transmissive character of the sedimentary rocks of the
Colorado Plateau to the distribution of uranium deposits. U.S. Geological Survey Bulletin 1124,
151 p.
Kaplan, D. I, Cantrell, K. J., and Wietsma, T.W. 1994. Formation of a barrier to groundwater
contaminants by the injection of zero-valent iron colloids: Suspension properties. 33rd Hanford
Symposium on Health and the Environment; In-Situ Remediation: Scientific Basis for Current
and Future Technologies (eds. G. W. Gee and N. R. Wing), Battelle Press, Columbus, Ohio, p.
821-837.
Landis, R.. 1997. Potential use of jetting to emplace PRBs. In Summary of the Remediation
Technologies Development Forum Permeable Reactive Barriers Action Team Meeting, Virginia
Beach, Virginia, September, 1997, p. 11-12.
Ma, Q.Y., Logan, T.J., Traina, S.J., and Ryan, J.A. 1994. The effects of NO3", Cl", SO42" and
9 94-
CO3 " on Pb immobilization by hydroxyapatite. Environmental Science and Technology.
28:408-418.
Marchand, E. G., Shirley, P. A., McNeils, K. A., Fiorillo, T. L. 1998. New installation
techniques in side-by-side demonstrations at Cape Canaveral Air Station. In RTDF Permeable
Reactive Barriers Action Team Meeting Presentation Materials, Beaverton, Oregon, April 1998,
p. 21-28.
Marcus, D. L. and Farrell, J. 1998. Reactant sand-fracking pilot test results. In Remediation
Technology Development Forum Permeable Reactive Barriers Action Team Meeting
Presentation Materials, Beaverton, Oregon, April 1998, p. 29-36.
Morrison, S. J. and Spangler, R. R. 1992. Extraction of uranium and molybdenum from aqueous
solutions: A survey of industrial materials for use in chemical barriers for uranium mill tailings
remediation. Environmental Science and Technology. 26:1922-1931.
Morrison, S. J., and Spangler, R. R. 1993. Chemical barriers for controlling groundwater
contamination. Environmental Progress. 12:175-181.
87
-------
Morrison, S. J. 1998. Status of the UMTRA Durango permeable reactive treatment walls. In
Remediation Technology Development Forum Permeable Reactive Barriers Action Team
Meeting Presentation Materials, Oak Ridge, Tennessee, November 1998, p. 205-216.
Morrison, S. J., Helferich, R., Thomson, B., and Emilia, D. 1998. Improvements to reactive
materials for use in permeable reactive barriers to remediate groundwater contaminated by metals
and radionuclides. U. S. Department of Energy report prepared for MSE Technologies, Inc., 49
pp.
Morrison, S. J., Spangler, R. R., and Morris, S. A. 1996. Subsurface injection of dissolved ferric
chloride to form a chemical barrier: Laboratory investigations. Groundwater. 34:75-83.
Morrison, S. J., Spangler, R. R., and Tripathi, V. S. 1995. Adsorption of uranium (VI) on
amorphous ferric oxyhydroxide at high concentrations of dissolved carbon (IV) and sulfur(VI).
Journal of Contaminant Hydrology. 17:333-346.
Morrison, S.J., and Spangler, R.R. 1992. Chemical barriers for controlling groundwater
contamination. Environmental Progress. 12:175-181.
Naftz, D.L., Davis, J.A., Fuller, C.C., Morrison, S.J., Freethey, G.W., Feltcorn, E.M., Wilhelm,
R.G., Piana, M.J., Joye, J., and Rowland, R.C. 1999. Field demonstration of permeable reactive
barriers to control radionuclide and trace-element contamination in ground water from abandoned
mine lands. U.S. Geological Survey Toxic Substances Hydrology Program—Proceedings of the
Technical Meeting, Charleston, South Carolina, March 8-12, 1999 (eds. D.W. Morganwalp and
H.T. Buxton): U.S. Geological Survey Water-Resources Investigations Report 99-4018A, p. 281-
288.
National Research Council. 1994. Alternatives for ground water cleanup. National Academy
Press, Washington, D.C., 315 p.
Rai, D. 1999. Personal communication.
Feltcorn, E., and Breeden, R., 1997. Reactive barriers for uranium removal. U.S. Environmental
Protection Agency Ground Water Currents. Issue No. 26:1-2.
Sandino, A., and Bruno, J. 1992. The solubility of (UO2)3(PO4)2*4H2O(s) and the formation of
U(VI) phosphate complexes: Their influence in uranium speciation in natural waters. Geochimica
et Cosmochimica acta. 56:4135-4145.
Shoemaker, S. H., Greiner, J. F., and Gillham, R. W., 1995. Permeable reactive barriers. In R. R.
Rumer and J. K. Mitchell (eds.), Assessment of Barrier Containment Technologies, a product of
the International Containment Technology Workshop, Baltimore, Maryland, August, 1995, p.
310-348.
Sivavec, T.M., Mackenzie, P.D., Horney, D.P., and Baghel, S.S., 1997. Redox-active media for
permeable reactive barriers: International Contaminant Technology Conference Proceedings,
February 9-12, 1997, St. Petersburg, Florida, p. 753-759.
-------
Sowder, A.G., Clark, S.B., and Fjeld, R.A. 1996. The effect of silica and phosphate on
thetransformation of schoepite to becquerelite and other uranyl phases. Radiochimica Acta.
74:45-49.
Spangler, R. R. and Morrison, S. J. 1991. Laboratory-scale tests of a chemical barrier for use at
uranium mill tailings disposal sites. ER'91 Conference Proceedings, Pasco, Washington,
U. S. Department of Energy, 1991, p. 739-744.
Spangler, R. R. 1997. Laboratory investigation of FeO and amorphous ferric oxyhydroxide as
permeable barrier materials at Fry Canyon. Department of Energy Grand Junction Office Internal
Report, 9 pp.
Standard Methods. 1992. Standard methods for the examination of water and wastewater, 18th
Edition. Greenberg, A.E., Clesceri, L.S., and Eaton, A.D., Editors.
Thaden, R.E., Trites, A.F. Jr., and Finnell, T.L. 1964. Geology and ore deposits of the White
Canyon area, San Juan and Garfield Counties, Utah. U.S. Geological Survey Bulletin 1125, 166 p.
Travis, C.C., and Doty, C.B. 1990. Can contaminated aquifers at Superfund sites be remediated?.
Environmental Science and Technology. 24:1464-1466.
Wilson, R.D., and Mackay, D.M. 1997. Arrays of unpumped wells: An alternative to permeable
walls for in situ treatment. International Contaminant Technology Conference Proceedings,
February 9-12, 1997, St. Petersburg, Flor., p. 888-894.
Zhang, P, Ryan, J.A., Bryndzia, L.T. 1997. Pyromorphite formation from goethite adsorbed lead.
Environmental Science and Technology. 31:2673-2678.
89
------- |