BARRIER SYSTEMS for ENVIRONMENTAL CONTAMINANT CONTAINMENT and TREATMENT - PART 3 pptx

2000, Naftz et al.2002, and Reddi and Inyang 2000 for detailed information on the generalcharacteristics of barrier materials mix design approaches and performance issues.In this chapter

referred to the preceding book in the containment series, Assessment of Barrier Containment Technologies (Rumer and Mitchell, 1995), as well as Daniel (1993),Gavaskar et al (1998), LaGrega et al (2000), Blowes et al (2000), Naftz et al.(2002), and Reddi and Inyang (2000) for detailed information on the generalcharacteristics of barrier materials mix design approaches and performance issues.

In this chapter, the emphasis is on fundamental factors and laboratory and fieldobservations that relate to the long-term performance of materials used in con-structing various types of containment systems The overall performance of thesesystems has been analyzed holistically using the systems approach in Chapter 1.Chapter 2 dealt with models of water and contaminant fate and transport throughcomponents of containment systems It is herein recognized that material properties

* With contributions by David W Blowes, University of Waterloo, Waterloo, Ontario, Canada; David

A Carson, U.S Environmental Protection Agency, Nashville, Tennessee; Peter W Deming, Mueser Rutledge Consulting Engineers, New York, New York; Jeffrey C Evans, Bucknell University, Lewis- burg, Pennsylvania; Glendon W Gee, Battelle Pacific Northwest National Laboratory, Richland, Washington; Hilary I Inyang, University of North Carolina at Charlotte, Charlotte, North Carolina; Stephan A Jefferis, University of Surrey, Surrey, United Kingdom; Mark R Matsumoto, University

of California at Riverside, California; Gustavo Borel Menezes, University of North Carolina at Charlotte, Charlotte, North Carolina; Stanley J Morrison, Environmental Services Laboratory, Grand Junction, Colorado; Scott D Warner, Geomatrix Consultants, Oakland, California; John A Wilkens, DuPont, Wilmington, Delaware

144 Barrier Systems for Environmental Contaminant Containment & Treatmentplay a significant role in overall system performance This chapter is divided intothree primary subsections, each of which addresses materials performance for aspecific type of containment structure.

3.1.1 T HE R OLE OF B ARRIER M ATERIAL M INERALOGY AND M IX

to the barrier mix in cases where the natural clay content of the barrier material

is insufficient to provide the required mix characteristics In other cases, barriermaterials are fabricated and used to provide specific functions An example is ageomembrane that can be incorporated as a component into a containment structurefor fluid retention, separation of clay to minimize the chance of attack by aggres-sive permeants, and diversion of gas flow to desirable control points Table 3.1provides a general listing of various characteristics of barriers that affect classes

of phenomena that relate to the most significant barrier design objectives Some of

TABLE 3.1 Containment System Design Considerations and Material Characteristics that are Usually Evaluated in Bench-Scale Tests

Physico-Chemical

Significant Barrier Material Properties

Reduction of contaminant release and transport

Advection Hydraulic conductivity

Density Moisture content Gradation Porosity Crack density

Leachability Crack density Chemical compatibility Inadequate retardation Density Physical durability Chemical attack Mineralogy relative to

contaminant chemistry Radiation transport Density, mass attenuation

coefficient

Material Stability and Applications 145

the barrier parameters such as hydraulic conductivity, porosity, and crack densityapply to compacted, cemented, and fabricated materials

For granular barrier materials that may be compacted or cemented into barrierlayers, the component material mineralogy and specific surface area are keymaterial factors that, in combination with the emplacement density, control theinitial and long-term barrier material textures when exposed to physical stressesand chemical contact Mineralogy controls the physico-chemical interactions(including the reactivity) of a barrier component with permeating fluids under agiven environmental condition Under the most frequently encountered temper-ature, pressure, and pH–Eh conditions in the field, clays (comprising mostlyaluminosilicates) react with permeants much more aggressively than sands (com-prising mostly silica) Because of their mineralogy, the charged clay surfacespresent opportunities for the chemisorption of charged contaminants such asheavy metals as summarized by Inyang (1996) in Table 3.2

For a barrier material that has favorable mineralogy (i.e., a mineralogy thatfavors its interaction with permeating fluids in reactions that remove soluteswithout degrading the barrier), the opportunity for its interaction with the per-meant is enhanced if its specific surface is high The specific surface is the ratio

of surface area to weight of a material, and it is inversely proportional to thegrain size of the material For surface reactions like cation exchange and adsorp-tion that are prevalent in barriers, their role in increasing the contaminant distri-bution coefficients (i.e., cleaning the permeating fluid in terms of its entry vs.exit chemistries) increases as the specific surface of the component materialincreases, as reflected in results plotted by Milne-Home and Schwartz (1989)presented in Figure 3.1 Often, even when a specific barrier component exhibits

a desirable material characteristic, it may not be adequate with respect to anothercharacteristic For example, a clay mineral such as sodium montmorillonite may

be sorptive enough for heavy metals but inadequate in terms of providing strengthagainst desiccation Yet still, cost considerations usually preclude the use ofsingle-component barrier systems in waste containment Essentially, most barriermaterials are composites, the proportions of which are designed to optimizeperformance characteristics at minimal cost In the case illustrated in Figure 3.2,D’Appolonia (1980) evaluated the effects of fines (% minus #200 sieve) on thepermeability of soil-bentonite (SB) backfill candidate materials and found thatfor both plastic fines and nonplastic/low-plasticity fines, the permeabilitydecreased as the fines content increased Permeability values for the plastic fineswere generally lower than those of the nonplastic/low-plasticity fines Presumably,the plastic fines comprise more moisture-sensitive or expansive minerals than thenonplastic/low-plasticity fines Figure 3.3 shows the effects of bentonite (mont-morillonite) content on the permeability of the SB backfill candidate materialmixes A bentonite content of 3% (by dry weight) was adequate to reduce thepermeability values from 5 × 10–5 to 5 × 10–3 centimeters per second (cm/s) toabout 10–7 cm/s for well-graded coarse materials

In another investigation that illustrates the optimization of mix composition

to obtain a favorable material characteristic, Ryan and Day (1986) evaluated the

146 Barrier Systems for Environmental Contaminant Containment & Treatment

TABLE 3.2

Sorption Characteristics of Soil Minerals and Chemical Additives

for Hazardous and Radioactive Metals

Single

Component

Material

General Properties

Metals Tested

Test Type

particle size

<2 µ m

Cd 2+ Batch Initial pH

range = 4.6–7.3

95%, 95%, and 90%

of Cd 2+ sorbed by Na-, Ca-, and K-montmorillonite, respectively Montmorillonite

from Texas

(Puls and

Bohn, 1988)

Ca — saturated

(Ziper et al.,

1988)

K — fixed, 500–1000

µ m particle size, SSA = 22.5 m 2 /g

Zn > Ni 50% of metals were adsorbed within pH range 4.49–5.80 Kaolinite (Yong

and

Galvez-Cloutier, 1993)

LI = 61%, SSA =

40 mL of lead solutions

Maximum Pb 2+

adsorption decreased at high

pH due to precipitation Goethite

(iron oxide)

(Coughlin and

Stone, 1995)

SSA = 47.5 m 2 /g

Coordination chemistry of oxides affects adsorption 50% of Cu 2+ , Pb 2+ ,

Co 2+ , Ni 2+ removed

at pH 4.5, 4.8, 6.3, 6.8, respectively Goethite

Selectivity order:

Zn 2+ > Cd 2+ > Ca 2+

Material Stability and Applications 147

permeability ranges of three mix compositions for a fly ash cement-slurry wall,the results of which are presented in Figure 3.4 Test results developed (Flemingand Inyang, 1995) for fly ash amended materials, which may, in some cases,exhibit cementation if the ash mineralogy is favorable or some cementing agentsare added, show that initial and longer term permeabilities of cemented barriermaterials can be significantly influenced by reactions among the mix components.Figure 3.5 shows the conceptual textural patterns proposed by Fleming and Inyang(1995) in a comparative study of the effects of class F (nonreactive) fly ash andclass C (reactive) fly ash amendment of barrier clay on changes in permeabilityunder freeze-thaw action The patterns are similar, but the reactive fly ash exhibitsinitial and final permeabilities that are lower than those of the nonreactive ash

3.1.2 A PPROACHES TO M ATERIAL E VALUATION AND S ELECTION

Bench-scale tests provide the best opportunity to evaluate the fundamental acteristics of barrier materials However, holistic assessments of a barrier systemperformance are most meaningfully performed through a combination of bench-scale testing and field quality assurance and monitoring tests The bench-scaleapproach has been widely used to evaluate barrier material parameters in batch

char-TABLE 3.2 (continued)

Sorption Characteristics of Soil Minerals and Chemical Additives

for Hazardous and Radioactive Metals

Single

Component

Material

General Properties

Metals Tested

Test Type

of competing ions = 0.1 N

Selectivity order:

Pb 2+ > Sr 2+ > Cu 2+

> Cd 2+ > Zn 2+ >

Cs 2+ at 25 mg/L lead concentration, absorbed Pb =

35 µ g/g Pyrolusite

(MnO2) (Ajmal

et al., 1995)

Crushed samples

Pb 2+ , Cd 2+ ,

Zn 2+ , Mg 2+

Batch Washed and

dried at 40° C; pH range of about 2–8

At pH = 6.5, 100%

of initial 22.7 mg/L

of Pb 2+ was sorbed; other results show high sorption for

Zn 2+ and Cd 2+ but low sorption for

Mg 2+

Source: Inyang, H.I (1996) Sorption of inorganic chemical substances by geomaterials and additives, Report CEEST/001R-96, University of Massachusetts, Lowell, MA.

148 Barrier Systems for Environmental Contaminant Containment & Treatment

systems, monoliths of scaled down dimensions, or columns of media The lattercan be densely compacted, as in the case of earthen materials considered forfluid/contaminant transport barriers or loosely emplaced as in reactive columns.Most of the granular barrier material characteristics that are usually targeted aresummarized in Table 3.3 Not all of these tests need to be performed for all barriermaterials Some tests, exemplified by porosimetry, are not usually performedbecause the influence of the pore size distribution measured is represented alongwith barrier material density and reactivity with specific contaminants in dataobtained from column tests for contaminant retardation coefficient estimation.The tests listed in Table 3.3 have designations that vary from one country toanother, although they are most standardized under the American Society for

FIGURE 3.1 Specific surface vs bulk cation exchange capacity for various sediments

and minerals (From Milne-Home, W.A and Schwartz, F.W., 1989 Proceedings of the Conference on New Field Techniques for Quantifying the Physical and Chemical Proper- ties of Heterogeneous Aquifers, Dallas, Texas, pp 77–98 With permission.)

Discrete particle clays Pore ilning clays Pore bridging clays

1

1 2 3

Material Stability and Applications 149

rials such as geomembranes are tested under protocols that are different from

those of granular barrier materials Fundamental tests are important because they

can provide data that are helpful in performing a general durability evaluation of

barrier materials and understanding mechanisms that are determinants of their

durability

3.1.3 G EOSYNTHETICS AND THEIR D URABILITY IN B ARRIER S YSTEMS

In general, the ability of barrier materials to retard fluid transport, resist chemical

and biological attack, and maintain structural integrity under externally imposed

stresses depends on their composition, emplaced thickness, and the quality

assur-ance practices implemented during construction Early in the development of

containment system design configurations, earthen and cementitious barrier

mate-rials were used almost exclusively A more recent development, particularly

within the past two decades, is an increase in the use of geosynthetic materials

to enhance containment system barrier layer performance Both earthen and

geosynthetic barrier materials have advantages and disadvantages Earthen

bar-riers are most commonly clayey soils that are either compacted into layers as in

landfills and surface impoundments or emplaced as slurry backfill as in slurry

cutoff walls While they can retard contaminant transit through a variety of

processes (e.g., sorption, induced precipitation of dissolved substances within

inter-particle pore spaces), significant variability and uncertainty can exist in the

FIGURE 3.2 Effects of fines content on the permeability of soil-bentonite backfill (From

D’Appolonia, 1982 Proceedings of the 13th Annual Geotechnical Lecture Series,

Phila-delphia Section, American Society of Civil Engineers, Philadelphia, PA With permission.)

SB Backfill permeability, cm/sec

Nonplastic or low plasticity fines

Testing Materials (ASTM) protocols As evident in Section 3.2, fabricated

mate-150 Barrier Systems for Environmental Contaminant Containment & Treatment

spatial distribution of barrier transport parameters such as hydraulic conductivity

and diffusion coefficient Furthermore, under aggressive chemical environments

and sustained desiccation processes, earthen barriers can develop enlarged flow

channels that allow contaminants in both the gaseous and liquid phases to travel

through the barrier easily Geosynthetic materials such as geomembranes have less

FIGURE 3.3 Effects of bentonite content on the permeability of SB backfill (From

D’Appolonia, 1982 Proceedings of the 13th Annual Geotechnical Lecture Series,

Phila-delphia Section, American Society of Civil Engineers, PhilaPhila-delphia, PA With permission.)

FIGURE 3.4 The effects of cement/water ratio and fly ash/cement ratio on the

perme-abilities of slurry wall mixtures (From Ryan, C.R and Day, S.R., 1986 Proceedings of

the 7 th National Conference on Management of Uncontrolled Hazardous Waste Sites,

Washington, DC With permission.)

Poorly graded silty sand w/30 to 50% nonplastic fines

Clayey silty sand w/30 to 50% fines

10 –5 Average

(typ.)

Material Stability and Applications 151

variability in the spatial distribution of transport parameter magnitudes because

they are manufactured in tightly controlled processes Furthermore, they are less

permeable to fluids and offer the opportunity to minimize the overall design

thickness of a barrier layer On the other hand, punctures, poor joints, and internal

degradation can diminish their effectiveness as barrier layers Giroud et al (1992,

1997) have developed quantitative methods for estimating liquid transport through

geomembrane defects

Geosynthetic barrier materials have been used as barrier layers that

comple-ment the functions of earthen barrier layers Many composite cover designs such

FIGURE 3.5 Effects of reactions among barrier constituents on the permeability of

ash-modified clayey barrier soil subjected to freeze-thaw cycling (From Fleming, L.N and

Inyang, H.I., 1995 ASCE Journal of Materials in Civil Engineering, 7(3), 178–182 With

d Class C fly ash-modified clay soil

Longitudinal fracture

Reactive ash particle Clay platelet Reacted rim

Nonreactive ash particle

152 Barrier Systems for Environmental Contaminant Containment & Treatment

as those consistent with the minimum design standards developed for the

Resource Conservation and Recovery Act (RCRA), comprise both soil barrier

layers and geosynthetic materials Othman et al (1997) have performed studies

of the performance of such barrier configurations in the field The results indicate

that with adequate quality control, such systems can perform effectively, at least

within the few decades that they have been in service Another composite barrier

system that typically produces desirably low hydraulic conductivities in barrier

systems is the geosynthetic clay liner (GCL) that has been studied by many

researchers (Estornell and Daniel, 1992; Rad et al., 1994; and Petrov et al., 1997)

The GCL is gaining wider acceptance in the containment industry because of its

cost effectiveness, relatively easy installation, and low barrier thickness

Instal-Although test protocols, design methods, and quality assurance methods have

been developed [Koerner and Daniel, 1997; Haxo, 1987; United States

Environ-mental Protection Agency (USEPA), 1985], concerns about the long-term

dura-bility of geosynthetic materials in barrier systems remain This concern is driven

by the knowledge that all materials that are exposed to stressors degrade with

time Such degradation in the long term is not limited to geosynthetic materials,

but extends to emplaced earthen barrier materials as well For geosynthetic

TABLE 3.3

General Testing Approaches and Methods for Significant

Characteristics of Batch and Compacted Barrier Materials

Soil Texture

Dispersivity Indeterminate; evaluate experimentally

Hydraulic conductivity a,b Permeameter tests

Moisture content Drying tests

Path length/tortuosity a Indeterminate; evaluate experimentally

Pore size distribution Porosimetry

Porosity (effective) a Empirical methods, porosimetry

Soil Composition

Chemical (elemental) composition Chemical tests (e.g., x-ray fluorescence)

Mineralogy (crystallinity) Mineralogy tests (e.g., x-ray diffraction)

a Denotes a property dependent on compaction.

b Denotes a property dependent on mineralogy.

Source: Adapted from Inyang, H.I et al (1998) Physico-Chemical Interactions

in Waste Containment Barriers, Encyclopedia of Environmental Analysis and

Remediation, Vol 2, Wiley, New York, pp 1158–1165.

lation methods are summarized in Section 3.4.3

Material Stability and Applications 153

materials that have been effectively installed, degradation mechanisms include

aging, chemical attack, and photo-oxidation To assess the potential effectiveness

of geosynthetic barriers in containment systems over 500 service time frames,

Badu-Tweneboah et al (1999) analyzed prospective effects of various degradation

processes of a 1.5-millimeter (mm)-thick high-density polyethylene (HDPE)

geomembrane that was installed within a landfill cover They used data from

studies performed on geomembranes and other polymeric materials to evaluate

the damage potential under sustained contact with aging agents such as oxygen,

microorganisms, heat, ultraviolet radiation, and radioactivity, as well as flaw

development due to abrasion, thermal stresses, animal burrowing, and plant root

penetration The analysis led to the conclusion that up to 5% reduction in yield

strain can occur per 25 years of service, resulting in an estimated yield strain of

zero if a liner deterioration pattern is assumed or 36% of the original yield strain

in 500 years if a logarithmic deterioration pattern is assumed

On the basis of their analysis, Badu-Tweneboah et al (1999) estimated that

the progressive stiffening of the geomembrane due to molecular rearrangement

under induced stresses in common containment system configurations would

likely result in stress cracking after 300 years of service The challenge is to

relate the damage potential to flaw sizes and numbers — a necessary step for

estimating potential fluid transport rates through geosynthetic materials

3.2 MATERIAL PERFORMANCE FACTORS IN CAPS

Caps or surface barriers in general are used to isolate buried wastes or

contam-inated soils from the atmosphere and biota on the earth’s surface To design an

effective cap, it is necessary to consider multiple objectives, including biota

intrusion (i.e., intrusion of plants, animals, and humans into the underlying waste

or contaminated soils), wind and water erosion, gas control, and percolation of

water into underlying waste The material performance criteria established for

each of these objectives depend on the type of waste to be contained and the

risks imposed by the waste on the nearby environment For example, stringent

mix design criteria may need to be used for facilities containing long-lived and

toxic radioactive wastes, whereas less stringent criteria can be applied to facilities

containing largely inert construction and demolition wastes The life span over

which the cap must function is generally associated with the type of waste as

well (e.g., 1,000 years for radioactive wastes or 30 years for solid wastes) In

most containment applications, however, there is no intent of ever exhuming the

waste Thus, a cap must meet the performance criteria as long as the material

being contained poses a risk to the surrounding environment In most cases, this

means that caps need to be designed for perpetuity and that a plan be in place to

monitor and maintain the cap as needed

Percolation from the base of the cap is the primary design criterion in most

cases A capping approach that will meet a percolation criterion (e.g., <1 mm/year)

is usually selected Then, the materials and geometry (e.g., layering) are selected

and configured to meet the percolation criterion, as well as the other criteria

(e.g., erosion, biota intrusion, gas control) Two general cap designs are used:resistive designs and water balance designs Examples of resistive designs areshown in Figure 3.6; examples of water balance designs are shown in Figure 3.7.Resistive designs employ a barrier system with high hydraulic impedance to limitpercolation (Benson, 2001) The barrier system can consist of geomembranes,fine-grained earthen materials, asphalt layers, or combinations of these or similarmaterials A drainage system is often used to limit the driving head on the barrierand ensure physical stability The water balance approach employs the store andrelease principle to limit percolation to an acceptable amount (Benson, 2001).Materials are selected that have adequate capacity to store infiltrating water duringwet periods without appreciable percolation Vegetation is used to remove thestored water and return it to the atmosphere so that the cover has the capacity tostore water during subsequent infiltration events.

The resistive and water balance design approaches are fundamentally different.The resistive design approach is predicated on constructing and maintaining asystem that blocks natural water movement In contrast, the water balance approach

FIGURE 3.6 Profiles of caps relying on a resistive barrier: (a) Compacted clay barrier

and (b) composite barrier.

Vegetated surface layer (150 mm)

Clay liner (>600 mm)

Clay liner (>600 mm) (a)

(b)

Vegetated surface layer (150 mm)

uses natural processes to limit natural water movement The natural approach usedfor water balance covers is considered by some to be superior The logic is that

a system that works with nature (i.e., water balance cap) is believed to be lesslikely to fail over the long term than a system that works against nature(i.e., resistive cap) However, currently there is no direct evidence demonstratingthat one approach is superior, provided that the cap is designed and constructedproperly

3.2.1 M ATERIAL P ERFORMANCE F ACTORS IN C OMPOSITE B ARRIERS

Resistive designs generally employ engineered materials to provide the hydraulicimpedance needed to meet a percolation criterion These materials include com-pacted natural clays, bentonites used alone in layers (e.g., as in a geosyntheticsclay liner) or mixed with other earthen materials (e.g., a compacted mixture ofsand and bentonite), polymeric sheets known as geomembranes, and asphalt andasphalt concrete layers (Koerner and Daniel, 1997) During the last decade, awealth of experience has accrued regarding the characteristics of these materialsand the elements that are required to reduce percolation to small amounts Expe-rience has shown that systems that rely solely on an earthen barrier (i.e., compacted

FIGURE 3.7 Schematic water balance caps: (a) Monolithic cap design and (b) two-layer

clay barrier or GCL) are prone to failure, even after short service lives, whereascomposite designs that combine a geomembrane underlain by an earthen barrierappear to function extremely well, at least for the relatively short experience record(<10 years) that currently exists (Benson, 2001, 2002) The performance of capsthat rely solely on a geomembrane or asphalt layer is largely unknown.

The following two examples illustrate how resistive designs that rely solely

on an earthen barrier can fail soon into their service lives One is a cap employing

a compacted clay barrier consisting of 460 mm of compacted clay placed oncompacted subgrade and overlain with 150 mm of topsoil vegetated with Bermudaand rye grasses This type of cap is often the presumptive remedy (i.e., the defaultdesign) for sites in the United States Superfund program, as was the case for thecap described here The other is a similar design, except a GCL was used instead

of a compacted clay barrier, and 600 mm of “protective cover soil” was placedbetween the GCL and the topsoil layer The topsoil layer was vegetated withcrown vetch to minimize erosion

The clay barrier was compacted in a manner that yielded a field hydraulicconductivity of 5 x 10–8 cm/s at the time of construction (the design criterion was

10–7 cm/s) The cap was intended to transmit less than 30 mm/year of percolation.Concerns about long-term cap performance led to installation of a system formonitoring all components of the water balance (Benson, 2002; Roesler et al.,2002) and, most importantly, the percolation rate Water balance data collectedfrom the cap since the time of construction are shown in Figure 3.8

FIGURE 3.8 Water balance data for the clay cap.

Soil water storage

No rain

Drying soil

4/1/00 8/1/00 12/1/00 4/1/01 8/1/01 12/1/01 4/1/02

Approximately 10 months after construction (September/October 2000), aperiod with little precipitation persisted for approximately six weeks During thisperiod, the cap desiccated as evidenced by the monotonic decrease in soil-waterstorage during this period Prior to this period, the cap transmitted percolation atrate of approximately 30 mm/year, which is consistent with the design criterion.Afterward, the percolation rate was approximately 500 mm/year (approximatelyone half of annual precipitation) Inspection of the clay barrier after it desiccatedshowed that the barrier contained desiccation cracks (Albright and Benson, 2002;Roesler et al., 2002) that served as preferential flow paths, causing the largepercolation rate increases that were measured and the stair-step character of thecumulative percolation record.

Concerns about the field performance of a cap that relies solely on a GCLalso led to percolation rate monitoring using two 10 m by 10 m lysimeters(Thorstad, 2002) The cumulative percolation recorded by the lysimeters is shown

in Figure 3.9 Excessive percolation was first noticed during the spring thaws of

1997 The GCL was exhumed in June 1997 and inspected to determine the cause

of the excessive leakage rates GCL thinning due to pressure applied by gravel

in the lysimeter was the suspected cause of the high percolation rate, but noquantitative assessment of the failure mechanisms was made A layer of sand wasadded to the lysimeter above the gravel as a cushion, a new GCL was installed,and the over-lying soil layers were replaced

Percolation monitoring continued after the lysimeters were rebuilt in 1997.Approximately 15 months after reconstruction, the percolation rate became exces-sive again Monitoring continued until October 1999, when one of the lysimeters(BL2) was exhumed to inspect the GCL Monitoring of the other lysimeter (BL1)continued Percolation recorded by lysimeter BL1 continued relatively steadilyand averaged 211 mm/year

Inspection of the GCL exhumed from directly over lysimeter BL2 revealedthat the bentonite was dry and cracked No thinning due to uneven pressure applied

by the underlying soil was observed Hydraulic conductivity tests on samples ofthe GCL exhumed from inside and outside the lysimeter showed a saturatedhydraulic conductivity ranging between 1.4 × 10–6 cm/s and 1.0 × 10–4 cm/s or

as much as 50,000 times the as-built hydraulic conductivity (2 × 10–9 cm/s).Chemical analysis showed that the exchange complex of the bentonite wasdominated by calcium and magnesium, whereas sodium was originally the pre-dominant cation (Thorstad, 2002) The exchange of calcium and magnesium forsodium reduced the swell potential of the bentonite sufficiently so that cracksthat formed during drier periods could not swell shut during wetter periods As

a result, the hydraulic conductivity of the GCL became unacceptably high.When lysimeter BL2 was exhumed in October 1999, it was rebuilt using acomposite barrier consisting of a thin (0.5 mm) polyethylene geomembrane heatbonded to one side of the GCL This barrier was installed with the geomembranedown, as recommended by the manufacturer The overburden soils removed duringexhumation were replaced after the new GCL was installed Very little percolationfrom the new GCL has been recorded during the two years of monitoring since

installation (2.4 mm/year on average), suggesting that the composite barrier isfar superior to the GCL alone.

Positive field performance of caps that employ a resistive design with acomposite barrier has been reported by others as well (Melchior, 1997; Dwyer,

FIGURE 3.9 Profile (a) and cumulative percolation record (b) for GCL cap.

Vegetated surface layer (150 mm)

Silty base layer (600 mm)

Elapsed time (days)

1996

1997 1998

1st rebuild

2nd rebuild BL2

2001; Albright and Benson, 2002) Melchior (1997) reported percolation ratesbetween 0.8 and 3.0 mm/year for a cap in Germany employing composite barrierdesign The barrier consisted of 600 mm of clay (saturated hydraulic conductivityless than 10–7 cm/s) overlain by a 1.5-mm-thick HDPE geomembrane, a sanddrainage layer 250 mm thick, and a vegetated topsoil layer 750 mm thick Dwyer(2001) reported an annual percolation rate of 0.1 mm/year for a cap in semi-aridAlbuquerque, New Mexico, having a design similar to Melchior’s cap Dwyer(2001) also reported a percolation rate of 1.8 mm/year for a similar cap inAlbuquerque employing a composite barrier with a GCL as the earthen component.The USEPA’s Alternative Cover Assessment Program (ACAP) is also moni-toring the percolation rate from seven caps employing composite barrier layersconsisting of a geomembrane underlain by a GCL or compacted clay barrier(Albright and Benson, 2002; Roesler et al., 2002) Percolation rates from thesecaps are summarized in Table 3.4 The percolation rates generally are near zero

in semi-arid and arid climates, and less than 4 mm/year in humid climates Thus,the composite barrier generally seems to be effective, largely because thegeomembrane is nearly impervious and the fine-grained soil beneath the geomem-brane provides impedance to flow at points where the geomembrane may containdefects

The exception is the cap located in Monterey, California This cap is located in

a semi-arid environment, but is transmitting 18 mm/year of percolation (Table 3.4).The cover soil placed on the geomembrane for this cap consisted of soil from

TABLE 3.4

Summary of Precipitation and Percolation Rates from Caps

with Composite Barriers Monitored by ACAP

Site

Duration

Total Precipitation (mm)

Percolation (mm/year)

Percentage of precipitation in parentheses.

Source: Data from Albright, W and Benson, C (2002) Alternative Cover Assessment Program

2002 Annual Report, Publication No 41182, Desert Research Institute, Reno, NV; Roesler, A.

et al (2002) Field Hydrology and Model Predictions for Final Covers in the Alternative Cover Assessment Program — 2002, Geo-Engineering Report No 02-08, University of Wisconsin, Madison, WI.

demolition projects and contained a variety of debris, including reinforcing barsand angular chunks of concrete These materials may have caused puncturing ofthe geomembrane, which may be responsible for the higher percolation rates(Roesler et al., 2002) This example illustrates an important point: caps con-structed with suitable barrier materials can function poorly if other aspects of thedesign are not properly implemented.

Although the performance record for caps with composite resistive barriers

is good, the record is short relative to the life span over which the caps areintended to function Melchior’s study has the longest record (eight years).Dwyer’s record is four years, and the monitoring is continuing at ACAP sites Ingeneral, composite barriers that have been exhumed appear to be in excellentcondition even after several years of service, including those barriers located inthe arid desert in southern California (Corser and Cranston, 1991; Melchior,1997) Additionally, several studies suggest that geomembranes should performadequately for hundreds of years, if not longer (Hsuan and Koerner, 1998; Clarke,2002; Rowe and Sargam, 2002) However, these predictions are primarily heu-ristic or based on ancillary measurements (e.g., depletion rate of anti-oxidants).The reality is that little hard data exist that can be used to make reliable predictionsregarding the life span of geomembranes in composite covers Given the dearth

of information on life expectancy, this is an area in need of research given thatcaps employing composite barriers are ubiquitous

3.2.2 M ATERIAL P ERFORMANCE F ACTORS IN W ATER

B ALANCE D ESIGNS

Water balance designs generally employ broadly graded finer-textured soilsbecause of their capacity to store significant amounts of water with little drainageand their ability to deform without cracking Coarse-grained materials are alsoused to form capillary breaks that enhance storage in the finer layer or divertwater under unsaturated conditions The coarse material can also be used toremove water from the barrier through advective drying (Albrecht and Benson,2002; Stormont et al., 1994) Caps that employ a single layer of fine-texturedsoil are generally referred to as monolithic barriers, whereas those with two ormore layers with contrasting particle size are referred to as capillary barriers(Figure 3.7)

The performance record for water balance designs generally is shorter thanthat associated with resistive designs, although a large effort has been underway

in North America during the last decade to collect field data on water balancecaps (Khire et al., 1997; Ward and Gee, 2000; Dwyer, 2001; Albright and Benson,2002) Perhaps the most notable monitoring program has been conducted at thesemi-arid Hanford site (south-central Washington) for a cap designed to limitpercolation to <0.5 mm/year The cap is intended to have service life of 1000 yearswithout maintenance [United States Department of Energy (USDOE), 1999; Wardand Gee, 2000] A full-scale test section of the cap was constructed in 1994 and

has been monitored under natural conditions and conditions that are extremelywet for the region.

Because a 1000-year life without maintenance was required, natural tion materials that are known to have existed in place for thousands of years wereselected The top-to-bottom profile consists of a 2-m-thick layer of vegetated silt-

construc-loam overlying layers of sand, gravel, basalt rock (riprap), and asphalt (Figure

3.10) Each layer serves a distinct purpose The silt-loam is for storing infiltration(600 mm of water can be stored in the silt loam before it will drain) and providesthe medium for establishing plants that are necessary for transpiration The coarsermaterials placed directly below the fine soil layer create a capillary break thatenhances the storage capacity of the silt-loam Placement of the silt-loam directlyover coarser materials also creates an environment that encourages plants andanimals to limit their natural biological activities to the near surface, therebyreducing biointrusion into the lower layers The coarser materials also help deterinadvertent human intruders The asphalt layer (asphalt concrete overlain by layer

of fluid-applied asphalt) acts as a secondary barrier that employs a resistiveapproach to impede and divert water passing through the capillary break A shruband grass cap was established on the cap in November 1994 Two sideslope

configurations, a clean fill gravel on a 10:1 slope and a basalt riprap on a 2:1

slope, were also part of the overall design and testing

FIGURE 3.10 Hanford cross section of Hanford cap showing (a) interactive water balance

processes, (b) gravel sideslope, and (c) basalt riprap sideslope.

Lateral drainge

Upper neutron probe access tube

Erosion- resistant gravel admix

Clean fill side slope

(pit run gravel)

(c) Basalt side slope

Vertical drainage

Waste crib

Precipitation (P)

Evapo-transpiration

Neutron probe access tube

Upper silt w/admix 1.0 m Lower silt 1.0 m Sand filter 0.15 m Gravel filter 0.3 m Basalt rock Riprap 1.5 m Drainage gravel 0.3 m min

Composite asphalt (asphaltic concrete coated w/fluid applied asphalt 0.15 m min.)

Top course 0.1 m min

Sandy soil (structural) fill

In situ soil

From November 1994 through October 1997, sections of the cap were jected to an irrigation regime of three times the long-term average annual pre-cipitation, which included a simulated 1,000-year storm event (70 mm of water)during the last week of March for three years (1995 through 1997) Percolationdid not occur from the cap until the third year, and then only a small amount(less than 0.2 mm) was transmitted from one section subjected to the enhancedirrigation treatment No drainage has occurred since then from this section orfrom any other portion of the cap In fact, the percolation that was recorded hasbeen attributed to lateral flow from water diverted off an adjacent roadway ratherthan flow through the cap (USDOE, 1999).

sub-Despite the large amount of water that was applied, all available stored soilwater was removed from the entire soil profile by late summer each year byevapo-transpiration (Figure 3.11), which maintained the water storage in the silt-loam layer well below the estimated drainage limit of 600 mm If the silt-loamthickness was reduced from 2 m to 1.5 m, the storage data indicate that little or

no percolation would be expected However, if the silt-loam thickness was

FIGURE 3.11 Temporal variation in mean soil water storage in the silt-loam in the

Hanford cap Monitoring was interrupted 1998–2000 Horizontal dashed lines represent estimated storage limits for caps with silt-loam layers 2 m, 1.5 m, and 1.0 m thick (From USDOE, 1999 200-BP-1 Prototype Barrier Treatability Test Report DOE/RL-99-11, U.S Department of Energy, Richland, WA; Ward, A and Gee, G., 2000 In Looney, B and

Falta, R (Eds.), Vadose Zone Science and Technology Solutions, Battelle Press, Columbus,

OH, pp 1415–1423 With permission.)

2.0 m silt loam

1.5 m silt loam

1.0 m silt loam

Drainage under natural conditions

9/30/1998 9/29/1996

reduced to 1 m, it appears that the cap would not perform well under extremelywet conditions.

The cap tested at Hanford represents perhaps the most sophisticated andredundant type of water balance design ever considered The level of complexityassociated with the cap is needed for the radioactive wastes that it is designed toisolate For many sites (e.g., municipal solid wastes, demolition debris, contam-inated soils), however, less sophisticated water balance caps are needed Anassessment of more typical water balance caps is being conducted by ACAP undernatural climatic conditions (Bolen et al., 2001; Albright and Benson, 2002) Thecaps tested by ACAP are intended to meet a target percolation rate that rangesbetween 3 and 30 mm/year depending on the type of waste, the regulations inplace at each site, and the climate (semi-arid or arid vs humid) Laboratorymeasurements of unsaturated and saturated soil properties were used in conjunc-tion with common methods accepted in practice to design each cap (Bolen et al.,2001) Typically, an unusually wet condition was used for the design calculations.Percolation rates measured for the ACAP water balance caps as of April 2002are summarized in Table 3.5, along with the design percolation rates Ninemonolithic barriers and five capillary barriers are being evaluated The designcriterion is being achieved at eight of the 10 semi-arid sites, but at none of thehumid sites The factors contributing to the higher than anticipated percolationrates are currently under evaluation, but the data do illustrate that water balancecaps do not necessarily perform as intended

One key factor contributing to the higher than anticipated percolation ratesappears to be the influence of pedogenesis on hydraulic properties near thesurface Samples are currently being collected from the surface of each test section

as large undisturbed blocks to characterize the hydraulic property changes thathave occurred A summary of the saturated hydraulic conductivity measurementsobtained to date is provided in Table 3.6 The saturated hydraulic conductivityhas increased due to factors such as desiccation and root penetration at three ofthe four sites for which tests have been conducted At the fourth site, the hydraulicconductivity has remained about the same Understanding how the hydraulicproperties change over time is critical to predicting how water balance caps willperform over the long term Long-term performance prediction is an issue in need

of research before water balance caps can be considered a long-term solution forcontainment Another important issue probably contributing to higher than antici-pated percolation rates is scaling between hydraulic properties measured in thelaboratory and those operative in the field Additional study of scaling issues andhow they can be incorporated in design is needed to understand long-term capperformance

3.2.3 C OUPLING OF V EGETATION AND M ATERIAL

P ERFORMANCE F ACTORS

Vegetation is not a cap material per se like soils and geosynthetics, but it is critical

to the long-term behavior of most caps, as discussed in detail in Chapter 1

or Environmental Contaminant Containment & T

Duration

Total Precipitation (mm)

Percolation (mm/year)

Percentage of precipitation in parentheses.

© 2006 by Taylor & Francis Group, LLC

Vegetation reduces erosion and, for water balance caps, is mostly responsible forremoving water stored in the cap There are three important factors that affectthe success associated with establishing vegetation: proper preparation of the capsurface (e.g., not over-compacted), provision of nutrients, and selection of veg-etation that is consistent with the surrounding environment (e.g., a heavy grasscover should not be used for a water balance cap in the desert of Las Vegas,Nevada).

When these issues are considered during design and construction, vegetationhas largely been successful For example, at the Hanford site, the survival rate

of transplanted shrubs has been remarkably high (97% for sagebrush and 57%for rabbitbrush) Heavy invasions of tumbleweed have occurred (e.g., in 1995),but have not persisted Grass cover consisting of 12 varieties of annuals andperennials, including cheatgrass, several bluegrasses, and bunch grasses, currentlydominates the surface Approximately 75% of the surface remains covered byvegetation requiring no maintenance, which is a value typical of shrub-steppeplant communities (Gee et al., 1996) A similar example is shown in Figure 3.12for the water balance caps at the ACAP site in Sacramento, California Withinone year of construction, a healthy cover of grasses and forbs was establishedwith a leaf area index on the order of 1.4 (Roesler et al., 2002)

Characterizing the transpiration that can be expected from vegetation is amore challenging issue (Figure 3.13) Figure 3.13 shows water balance quantitiesfor the thinner (1,080 mm) monolithic water balance cap in Sacramento beingmonitored by ACAP (test section on right-hand side of photographs shown inFigure 3.12) During the first growing season after construction (2000), thevegetation was able to extract the water and deplete the soil-water storage to thewilting point (approximately 180 mm), thereby providing an adequate soil res-ervoir for storing water during the subsequent winter However, the vegetationwas far less effective in extracting the water in Spring 2001, even though theprecipitation record was similar in both years, the water stored at the end of bothwet seasons was comparable (approximately 400 mm), and the vegetation

TABLE 3.6 Summary of Saturated Hydraulic Conductivities

of Samples Retrieved from the Surface of Covers being Monitored by ACAP

Site

Geometric Mean Hydraulic Conductivity (cm/s)

Cedar Rapids, IA 1.5 × 10 –5 4.6 × 10 –4

appeared no different during either growing season Despite these similar tions, the vegetation removed approximately 140 mm less water during the 2001growing season Inadequate water removal resulted in inadequate storage capacitythe following wet season As a result, the storage capacity (approximately 430 mm)was quickly exceeded during the wet period, and most of the water that infiltratedthe cap surface became percolation.

condi-The inadequate transpiration observed during the 2000 growing season didnot persist During the 2001 growing season, the vegetation removed all of theavailable stored water However, the reason for these differences remains a mys-tery, and efforts are currently underway to better understand why transpirationwas greatly lower in 2001 This example illustrates, however, that characterizingand understanding the characteristics of vegetation is as important as understand-ing other materials used for caps, particularly for water balance caps that rely ontranspiration as a critical barrier system process

FIGURE 3.12 ACAP test sections in Sacramento, CA, at the end of construction (a) and

one year after construction (b).

(a)

(b)

3.3 MATERIAL PERFORMANCE FACTORS IN PRBS

In contrast to most containment systems, which are usually designed to impedethe flow of water, PRBs provide containment by treating contaminated water thatpasses through them PRBs rely on a reactive material placed in the subsurface(or manipulation of the physico-chemical properties of the subsurface environ-ment) to treat contaminated groundwater (Figure 3.14) As contaminated waterpasses though the PRB, reactions occur between the contaminants and the reactivemedium, resulting in effluent that meets a target concentration, such as a maximumcontaminant level (MCL) (depicted as “remediated water” in Figure 3.14)

A variety of reactive media are used for PRBs, including granular iron metal,granular activated carbon, zeolitic minerals, compost, limestone, and other “solid”materials placed in the subsurface to promote the physical, chemical, and bio-logical conditions necessary for contaminated groundwater treatment A summary

of many of the materials being used is provided in Table 3.7 A photograph ofgranular iron and clinoptilolite is shown in Figure 3.15

The most commonly used treatment material is granular iron metal, which iseffective for treating groundwater affected by both organic and inorganic constit-uents (Gillham and O’Hannesin, 1994) Although the proportion of all PRBapplications using granular iron has not been computed, a reliable estimate isthat 70% to 90% of PRBs installed as tests or full-scale applications have used

FIGURE 3.13 Water balance quantities for thin cover (1080 mm thick) monolithic water

balance covers being monitored by ACAP in Sacramento, CA (Data from Roesler et al.,

2002 Field Hydrology and Model Predictions for Final Covers in the Alternative Cover Assessment Program — 2002, Geo-Engineering Report No 02-08, University of Wisconsin, Madison, WI; Albright, W and Benson, C., 2002 Alternative Cover Assessment Program

2002 Annual Report, Publication No 41182, Desert Research Institute, Reno, NV.)

12/30/00 3/31/00

granular iron as the reactive medium Other materials, such as granular activatedcarbon (GAC), compost, crushed limestone, alumino-silicates such as zeoliticminerals, and other materials are less used thus far, but are being tested in avariety of diverse applications.

3.3.1 A PPROACH TO S ELECTION OF PRB M ATERIALS

The criteria for selecting a reactive material are described by Blowes et al (2000)and include an assessment of the range of materials that can be used to removecontaminants and an assessment of the duration of material reactivity Thesecriteria, coupled with an assessment of the potential for the release of hazardous

FIGURE 3.14 Schematic of a PRB used to intercept and treat a plume of contaminated

groundwater.

TABLE 3.7

List of Reactive Materials that have been Used in PRBs

Zero-valent metals (including

iron) (may or may not include

metal couples)

Methanes, ethanes, ethenes, propanes, chlorinated pesticides, freons, nitrobenzene, certain metals (Cr, U, As, Tc, Pb, Cd) Ferric oxides Mo, U, Hg, As, P, Se

Zeolites Sr, Pb, Al, Ba, Cd, Mn, Ni, Hg, certain organics Activated carbon Mo, U, Tc, chlorinated VOCs, BTEX

Peat, humate Mo, U, Cr, As, Pb

Sawdust, compost Nitrate

Oxygen Aromatic hydrocarbons, MTBE, vinyl chloride

Flow direction Aquifer

materials or contaminant by-products (e.g., release of vinyl chloride due to thereductive dechlorination of dichloroethylene), can be used to assess the potential

of the barrier material to provide adequate groundwater treatment Interactionsbetween natural groundwater constituents can result in extensive formation ofsecondary mineral precipitates within the barrier These precipitates can hinderbarrier performance by clogging the pore space and reducing barrier permeability,

or by obscuring reactive particle surfaces The assessment can be combined with

an understanding of contaminant concentrations, groundwater geochemistry, andsite hydrogeology to determine whether a practical remedial system can beconstructed Then, a preliminary cost estimate can be developed and compared

to remedial alternative estimates

Implementation of a remedial system employing a PRB can proceed through

a series of steps, with accompanying decision points leading to the installation

of an optimized system These steps start with a theoretical assessment of thepotential for treatment using existing PRB materials State and federal guidancemanuals have documented the PRB materials that were employed at existing PRBinstallations, the contaminants that were treated, and the contaminant removalthat was attained [e.g., Interstate Technology Regulatory Council (ITRC), 1999a,b].This information can be used in conjunction with theoretical calculations, such

as the use of geochemical speciation/mass transfer computer codes or the use of

FIGURE 3.15 Examples of reactive media used in PRB applications: granular iron metal

(left) and the zeolite clinoptilolite (right) U.S quarter shown for scale.

pH–Eh diagrams to assess the potential for contaminant removal If contaminantremoval is possible, then laboratory treatability testing is considered.

Laboratory treatability tests can be used to assess the potential for nant removal and develop reaction parameters to assist in barrier design Batchexperiments can be conducted to determine contaminant reactivity and measurereaction rates under static conditions Column experiments can be used to measurerates of contaminant removal under dynamic flow conditions and assess the poten-tial for the precipitation of secondary minerals and barrier clogging Where pos-sible, mineralogical examination of column materials following the testing programcan be used to verify the presence and structure of secondary precipitates to assessthe stability of these precipitates within the barrier and evaluate the potential forbarrier clogging Complementary geochemical modeling, including reactivetransport modeling, can be used to develop design parameters at this stage Thegeochemical modeling, coupled with groundwater flow and transport modeling,can be used to provide preliminary estimates of barrier performance and longevityand to design parameters for pilot- or full-scale installations

contami-3.3.2 E VALUATION OF F IELD P ERFORMANCE U SING

P ILOT T ESTING

The decision whether to conduct a pilot-scale test or move directly to full-scaleimplementation depends on the history of the technology and the confidence ofthe client and regulators Many PRB technologies have been demonstrated suf-ficiently to satisfy regulators that the treatment processes are well understood andthe installation success depends on site-specific processes Pilot-scale installationsvary in scale and degree of monitoring, from small-scale column experiments

conducted ex situ at a field site to large-scale installations that ultimately form a

portion of a full-scale PRB

The key objective of the pilot-scale installation is to simulate conditions in

a full-scale system as closely as possible Using the candidate reactive materialsand natural aquifer materials in contact with site groundwater and typical con-taminant concentrations provides a close approximation to the characteristics offull-scale systems The small size of pilot installations provides an opportunityfor monitoring at a level of detail that is sufficient to provide design parametersfor the full-scale installation Pilot-scale installations should be sufficiently ver-satile so that variability in treatment media and groundwater flow rates can beassessed The results of the pilot-scale installation can be used to confirm con-taminant reactivity and assess the potential for negative secondary reactions such

as scaling or clogging The pilot-scale system should also have well-defineddimensions and performance characteristics to simplify scaling up to the finalremedial system

One type of cost-effective in situ pilot test is conducted with a reactive test

well (RTW) consisting of reactive material placed in a 300-mm-diameter borehole(Figure 3.16) One or more 25-mm-diameter polyvinyl chloride (PVC) casings

are placed along the central axis of the borehole for groundwater sampling.Several well casings with slots at different depths can be used to obtain multiplesamples at different depths A peristaltic pump is used to collect low-flow samplesfrom the slotted section of each casing for analysis.

RTWs were first used to test the efficacy of different reactive media forremoving arsenic from groundwater at a DuPont site in East Chicago, Indiana.Data collected from the RTWs over a nine-month period were used to select PRBmaterial Basic oxygen furnace (BOF) slag was selected for use in the PRB based

on data collected from RTWs, whereas laboratory studies indicated that anothermaterial was more appropriate

The coaxial configuration of the RTW ensures that groundwater passedthrough approximately 75 mm of reactive material before sampling regardless ofthe local groundwater flow direction For accelerated tests, groundwater can becontinually extracted through the casing Because RTWs are simpler and lesscostly than a full-scale pilot wall, multiple RTWs can be installed at a given site

to test different materials or act as controls RTWs also have several advantages

over ex situ field demonstrations (Table 3.8)

Installation quality is important in a RTW providing reliable data The drillingprocess should not create a smear zone at the well interface that might impedeflow Centrality of the well casing is also important so that the flow path throughthe reactive medium is the same at all points in the well A centralizer consisting

of a plastic disk with threads that match the well string is generally placed atthe bottom of a RTW, along with conventional stainless-steel centralizers along

FIGURE 3.16 RTW using passive groundwater flow.

Groundwater samples

Bentonite seal

Reactive material in 300-mm borehole

Slotted well casing Ground-

water

the length of the well casing (Figure 3.17) The stainless-steel centralizers areinstalled above the slotted section so that the water being sampled is not exposed

to any extraneous reactivity

Data from a RTW can be interpreted at several levels, from strict tion of contaminant removal to development of break-through and capacity cor-relations and projections of service life By manipulating flow rates, kineticexpressions can also be developed A passive RTW (i.e., operating under naturalgroundwater flow conditions) can provide an assessment of effectiveness in nearlyreal time, i.e., one month of field data is equivalent to one month of ultimatePRB exposure To project the ultimate life of a PRB, an extractive RTW can beemployed In this technique, groundwater is pumped out of the central casing at

demonstra-an accelerated rate, demonstra-analogous to using higher throughputs in a laboratory column.Avoiding kinetic limitations (i.e., from an extraction rate that is too high) with

an extractive RTW is important unless a kinetic study is intended An appropriaterate should be determined in the laboratory and then translated to the field test

3.3.3 E FFECTS OF H YDRAULIC C ONSIDERATIONS ON R EACTIVE

M ATERIAL P ERFORMANCE

To date, PRB research has focused mostly on the reaction mechanisms, kinetics,and conversion efficiency associated with the reactive materials (Tratnyek et al.,2003) Much less effort has focused on factors that affect PRB hydraulics, even

TABLE 3.8

Comparison of Reactive Test Wells vs Ex Situ Field Tests

Key Technical Parameters

Contaminant losses in system Essentially none Can be significant

Flow rate through bed Natural (uncontrolled, cross-flow); or

enhanced (pumped), radial flow

Controlled, precise, axial

Logistical Parameters

Duration limit Unlimited, at low cost Limited by cost, etc.

Overall Assessment

Final wall approximation Very close; a mini-wall Approximation

Special potential Long-term performance evaluation Precise flow control

though hydraulic factors can have as large an impact on PRB effectiveness(Eykholt et al., 1999; Elder et al., 2002).

As more PRB systems are implemented and monitored, performance datasuggest that hydraulic characteristics of PRB materials need to receive greaterattention during design Recent reviews of PRB applications have suggested thatmost cases of unintended performance are due largely to inadequate hydraulicperformance Few cases are related to inadequacies in the chemical treatmentmethodology (Warner and Sorel, 2001; Battelle, 2002) These findings indicatethat designers need to consider hydraulics as a critical factor affecting successfulPRB deployment, and approach hydraulic design with the same level of care asreaction effectiveness Hydraulic aspects that can have a large impact on PRBeffectiveness are aquifer material heterogeneity and spatial variability of thegroundwater flow field The importance of geological heterogeneities and theneed for characterization was illustrated in a recent case study of a PRB con-structed near Kansas City, Missouri (Laase et al., 2000) The PRB was installed

in an alluvial aquifer to intercept a plume containing trichloroethylene (TCE).Data from a hydrogeological study were used as input to a groundwater modelused to select PRB orientation and breadth The breadth was to be sufficiently

FIGURE 3.17 Centralizers for maintaining casing position in a RTW: Base centralizer

(left) and stainless steel centralizer (right).

large to capture the entire width of the plume An extensive set of monitoringwells (12 upgradient, 16 downgradient, and 10 adjacent to the ends of the PRB)was installed to monitor influent and effluent conditions and check for bypassing.Data from the monitoring program showed that the wall was not functioning

as intended While the reaction mechanisms appeared to have been accounted forproperly, a sandy gravel region toward the southern end of wall was not detectedduring hydrogeological characterization and caused a portion of the plume tobypass the PRB, as shown in Figure 3.18 In addition, reversals in the hydraulicgradient during recharge events caused the southerly extent of the plume to curlnorthward and, at times, flow backward through the PRB Bypassing was occur-ring along the northern end of the PRB as well

Few PRBs are monitored as closely as the PRB in Kansas City Thus, thefrequency of problems caused by heterogeneity is unknown However, a recentmodeling study by Elder et al (2001, 2002) suggests that geological heterogeneitymay be having a much larger impact on PRB effectiveness than previouslythought Elder et al (2001, 2002) constructed a series of heterogeneous aquiferscontaining PRBs and simulated flow and transport through the aquifer and PRB.Because a model was used, effluent concentrations were characterized in fargreater detail than is possible in the field, even with a dense network of monitoringwells

Typical results reported by Elder et al (2001, 2002) are shown in Figure 3.19.The simulation consisted of a TCE source with a uniform concentration of 1000

FIGURE 3.18 Schematic of plume bypassing southern end of PRB installed near Kansas

City (Adapted from Laase, A et al., 2000 In Wickramanayake, G et al (Eds.), Chemical Oxidation and Reactive Barriers, Remediation of Chlorinated and Recalcitrant Com- pounds, Battelle Press, Columbus, OH, pp 417–424).

N Plume

Flow

PRB