March 1996
     Assessing  UST Corrective Action
                  Technologies:
Diagnostic Evaluation of  In Situ SVE-Based
               System Performance
                            by

                      Richard L. Johnson
                    Oregon Graduate Institute
                  Portland, Oregon 97291-1000
                       R. Ryan Dupont
                     Utah State University
                      Logan, Utah 84322
                           and
                       Duane A. Graves
                        IT Corporation
                   Knoxville, Tennessee 37923
                    Contract No. 68-C2-0108
                        Project Officer

                        Chi-Yuan Fan
              Water Supply and Water Resources Division
            National Risk Management Research Laboratory
                   Edison, New Jersey 08837
            National Risk Management Research Laboratory
                Office of Research and Development
                U.S. Environmental Protection Agency
                    Cincinnati, Ohio 45268

-------
                                   Notice
The U.S. Environmental Protection Agency through its Office of Research and
Development partially funded the research described here under Contract No. 68-C2-
0108, Work Assignment No. 2, to IT Corporation . It has been subjected to the
Agency's peer and administrative review and has been approved for publication as an
EPA document.

-------
                                   Foreword
The U.S. Environmental Protection Agency (EPA) is charged by Congress with pro-
tecting the Nation's land, air, and water resources. Under a mandate of national envi-
ronmental laws, the Agency strives to formulate and implement actions leading to a
compatible balance between human activities and the ability of natural systems to sup-
port and nurture life. To, meet this mandate, EPA's research program is providing data
and technical support for solving environmental problems today and building a scientific
knowledge base necessary to manage our ecological resources wisely, understand how
pollutants affect our health, and prevent or reduce environmental risks in the future.

The National Risk Management Research Laboratory is the Agency's center for inves-
tigation of technological and management approaches for reducing risks from threats to
human health and the environment.  The Laboratory research  program focuses on
methods for the prevention and control of pollution to air, land, water, and subsurface
resources; protection of water quality in public water systems;  remediation of contami-
nated sites and 'groundwater; and prevention and control of indoor air pollution. The
goal of this research effort is to catalyze development and implementation of innovative,
cost-effective environmental technologies, develop scientific and engineering informa-
tion needed by EPA to support regulatory and policy decisions, and provide technical
support and information transfer to ensure effective implementation of environmental
regulations and strategies.                                          :

This publication has been produced as part of the Laboratory's strategic long-term re-
search plan.  It is published and made available by EPA's Office of Research and De-
velopment to assist the user community and to link researchers with their'clients.
                                          E. Timothy Oppelt, Director
                                          National Risk Management Research
                                          Laboratory

-------
                                    Abstract
In situ corrective action technologies are being proposed and installed at an increasing
number of underground storage tank (LIST) sites contaminated with petroleum products
in saturated and unsaturated zones. It is often difficult to accurately assess the per-
formance of these systems for remediating soils and groundwater.  Because of the
complex subsurface characteristics encountered at leaking LIST sites, a series'of
tests/tools are needed for evaluating the appropriate application and remediation per-
formance of these corrective action technologies.

In response to this need, the U.S. Environmental Protection Agency (EPA) Office of
Research and Development (ORD) National Risk Management Research Laboratory
(NRMRL) has provided technical support to EPA Regions for evaluating in situ cor-
rective action technologies.  This report presents a series of test procedures that were
developed for evaluating intrinsic biodegradation, bioventing, soil vapor extraction
(SVE), and in situ air sparging (IAS), These procedures present diagnostic tools that
are designed to assist in assessing remediation performance. Most of the discussions
focus on evaluating systems designed for petroleum  hydrocarbon contamination; how-
ever, these procedures can also be used to evaluate remediation perforrriance of in situ
technologies used to address a wide range of contaminants.
                                       IV

-------
                                   Contents
Foreword	iii
Abstract	iv
Figures . .	  vii
Tables . .	xi
Example Calculations	  xii
Acknowledgments	'. .  . xiii

  1.   Introduction	'	1-1
          Purpose	1-1
          Background	1-1

  2.   Procedures for Conducting Tracer Tests to Evaluate Air Flow During
       Soil Vapor Extraction	 2-1
          Introduction	2-1
          Test Objectives	2-7
          Theory  	,	 2-7
          Test Equipment	2-13
          Test Procedures	.[	2-20
          Data Analysis	2-32
          Field Examples	2-34
          References for Section 2	  2-36

  3.   Procedures for Conducting Tracer Tests to Evaluate Recovery of Injected
       Air During In Situ Air Sparging	3-1
          Introduction	3-1
          Test Objectives	,	3-3
          Theory  .	  . .	3-3
          Test Equipment	 . .	3-5
          Test Procedures	3-13
          Data Analysis	,	|	   3-16
          Field Examples	   3-18
          References for Section 3	 . . .;	   3-18

  4.   Procedures for Conducting Tracer Tests to Evaluate the Distribution of
       Injected Air  During In Situ Air Sparging	4-1
          Introduction	4-1
          Test Objectives	 .	4-3
          Theory  	'.	4-3

                                      v«

-------
                           Contents (continued)

        Test Equipment	 4-4
        Test Procedures	4-6
        Groundwater Sample Collection	4-9
        Data Analysis	  4-12
        Field Example	  4-12

5.  Procedures for Bioventing Field System Design and Evaluation—Air Flow,
     Tracer, and In Situ Respiration Tests	 .	5-1
        Introduction/Purpose 	5-1
        Background/Theory	5-1
        Bioventing System Test Procedure for Treatability Assessment, Design,
         Process Monitoring and Performance Evaluation	5-8
        Summary and Conclusions	(	5-27
        Calibration of Field Instrumentation for Hydrocarbon and Oxygen/
         Carbon Dioxide Determinations -  Electronic Detection Instruments  . . . 5-28
        Field Sampling and Analysis for Hydrocarbon and O^CO2
         Determinations - Electronic Detection Instruments	 5-31
        References for Section 5	5-33

6.  Procedures for Evaluating Natural Attenuation in Groundwater	6-1
        Introduction	6-1
        Site Characterization and Selection of Natural Attenuation for Site
         Remediation	6-5
        Biological Monitoring Parameters	 6-13
        Physical/Chemical Monitoring Parameters	6-24
        Defining an Efficient and Cost-Effective Monitoring Plan for Natural
         Attenuation		6-29
        Modeling Natural Attenuation	6-30
        Data Presentation	 . . .	6-32
        Applying the Principles of Natural Attenuation for Environmental
         Management, Risk Reduction, and Remediation		6-33
        References for Section 6	6-38
                                    VI

-------
                                   Figures


2-1    Schematic drawing of petroleum hydrocarbon distribution in the subsurface. . 2-2

2-2   Relationship between compound vapor pressure and number of carbon
      atoms	2-4

2-3   Schematic drawing of a typical SVE remediation system in plan view (a)
      and section view (b)	2-5

2-4   Flow diagram showing the major steps involved in an airflow tracer test	2-9

2-5   Schematic drawings of possible SVE flow configurations	2-11

2-6   Schematic drawing of tracer "short-circuiting" through a higher-perrneability
      zone	2-12

2-7   Schematic drawing of the flow meter calibration system	2-16

2-8   Schematic drawing for flow meter calibration under positive pressure	2-18

2-9   Schematic drawing showing setup for measurement of sample purhp flow
      rate vs. vacuum and leakage vs. vacuum	2-19

2-10  Schematic drawing of the flow measurement system for the SVE/BV
      system	2-21

2-11  Schematic drawing showing the system for measuring SVE/BV flow
      using a tracer gas	2-22

2-12  Example showing estimation of radius of influence of the SVE/BV system
      from steady-state pressure data	2-25

2-13  Geometry for calculation of tracer dilution	2-28

2-14  Schematic drawing of the sample collection apparatus	2-31
                                      VII

-------
                              Figures (continued)


2-15  a) Tracer breakthrough curve; b) Fraction of volume recovered during
      the tracer test	2-33

2-16  Travel times from various locations within the OGI large physical model  . . . 2-35

3-1    Schematic configuration of an IAS/SVE remediation system	3-2

3-2   Steps in conducting a helium recovery tracer test	 . . .	3-4

3-3   Schematic drawing of the equipment used to calibrate the helium detector. . . 3-6

3-4   Schematic drawing of the equipment used to prepare helium standards	3-8

3-5   Schematic drawing of the flow meter calibration system	3-9

3-6   Schematic drawing for flow meter calibration under positive pressure	3-11

3-7   Schematic drawing showing setup for measurement of sample pump flow
      rate vs. vacuum and leakage vs. vacuum	3-12

3-8   Schematic drawing showing the system for measuring SVE flow using a
      tracer gas.	3-14

3-9   Schematic drawing showing the setup for sample collection during the
      tracer recovery test	3-17

3-10  Recovery of helium during an air recovery test in an OGI large experi-
      mental aquifer.		 3-19

4-1    Schematic configuration of an IAS/SVE remediation system	4-2

4-2   Steps involved in an lAS air flow tracer test		4-5

4-3   Schematic drawing for flow meter calibration under positive pressure	4-7

4-4   Schematic drawing of the tracer gas injection system.	'	 . 4-8

4-5   Plan view of sparging test site showing the locations of the monitoring wells
      (B-series) and the vertical profile locations (P-series)	4-13


                              Figures (continued)

                                      viii

-------
4-6    Site plan view showing SF6 vertical profiles	4-15

4-7    SF6 and vertical distribution of contaminants near monitoring well B6	4-16    '•

4-8    SF6 and vertical distribution of contaminants near monitoring well B7	4-17

5-1    Typical SVE system with labeled system components	5-5

5-2    Typical bioventing system schematic	5-5

5-3    Field bioventing system treatability assessment, design, and performance
       evaluation procedure.  :	5-9

5-4    Typical soil respiration gas data collected during field in situ respiration test. 5-13

5-5    Typical zero order soil respiration gas data collected during field in;situ
       respiration test. Zero order respiration rate = -0.145 vol %/hour	5-13  '  \

5-6    Typical first order soil respiration gas data collected during field in situ
       respiration test. First order respiration rate = -0.015/hour.  . .	5-14

5-7    Typical regression results for linear regression analysis of field respiration           '<
       data for bioventing system	5-15

5-8    Zero order linear regression results for a set of hypothetical in situ '
       respiration rate  data	5-19    ;
                                                                     -                  !

5-9    Residuals plot for zero order linear regression results for the set of                  ;
       hypothetical in situ respiration  rate data	5-20

5-10   First order linear regression results for a set of hypothetical in situ
       respiration rate  data	5-20

5-11   Residuals plot for first order linear regression results for the set of
       hypothetical in situ respiration  rate data.	5-21

5-12   Respiration rate/contaminant concentration relationships generated  in               |
       laboratory-scale microcosm studies conducted with JP-4 contaminated
       soil from Hill AFB, Utah	 . . :	 5-25


                                Figures (continued)
                                         IX

-------
5-13  Sample initial and continuing calibration data record sheets for hydrocarbon
      and respiration gas measurement instruments	 5-30

5-14  Sample format for soil gas monitoring data record sheet	5-32

5-15  Sample format for in situ respiration test data record sheet	5-33

6-1    Analysis of soil core segments accurately defines location of separate
      phase and dissolved phase contaminants	6-11

6-2   Contaminant mass estimation using the Thiessen Method for area;
      assignments.	6-12

6-3   Electron acceptors for common microbial respiration pathways and
      approximate oxidation/reduction potential (Redox or Eh) where each
      pathway occurs	 6-23
                                                                 i
6-4   Simplified groundwater elevation and contaminant content and composition
      figure showing relative contaminant composition and content in three
      monitoring wells	6-34

-------
                                  Tables


2-1  Recommended Spacings for Tracer Injection Points	2-14

3-1  Typical Pressure Values Used in Preparing Helium Standards.	3-7

5-1  Carrier Fluid Oxygen Supply Requirements	5-3

5-2  General Design and Application Considerations Appropriate for Conventional
     Versus Bioventing SVE Systems	5-7

5-3  Potential Oxygen Transfer Rates in Various Soils	,	5-11

5-4  Example Treatability study and Full-Scale Bioventing System Respiration
     Rates Reported From Various Sources	5-17

6-1  Analytical Methods for Natural Attenuation	:	6-3

6-2  Simple Microcosm Study to Define Aerobic and Anaerobic Biodegradation. . 6-19
                                     XI

-------
                              Example Calculations
2-1    Calculation of actual SVE/BV flow using a tracer	2-23
2-2    Estimation of soil permeability from steady-state pressure measurements.  . 2-26
2-3    Calculation of the volume of SF6 to be injected	2-27
2-4    Calculation of the volume of air+tracer to be injected.	2-29
2-5    Calculation of actual SF6 injection volume	2-29
2-6    Calculation of fraction of injected mass recovered	.....;	2-32
2-7    Calculation of effective porosity using a simple geometry	2-36
2-8    Estimation of velocity at the tracer injection point	2-36
3-1    Calculation of actual SVE flow using a tracer	'....•	3-15
3-2    Calculation of expected concentration and recovery efficiency.  . .  :	3-16
4-1    Calculation of headspace concentration	4-11
4-2    Calculation of percent saturation	 4-11
6-1    Groundwater flow rate	......  4	6-6
6-2    Application of a conserved chemical's concentration to calculate the rate
       and magnitude of abiotic changes in concentration.  	6-8
6-3    Contaminant dissolution from  nonaqueous phase liquids (NAPL)	6-14
6-4    Mass flux of contaminant in NAPL to water	6-16
6-5    Retardation coefficient	6-18
6-6    Biodegradation potential of an aquifer	6-20
6-7    Biodegradation rate and target compound half-life	6-25
                                       XII

-------
                             Acknowledgments
This document was prepared for the U.S. Environmental Protection Agency (EPA), Of-
fice of Research and Development, National Risk Management Research; Laboratory
(NRMRL) under Contract No. 68-C2-0108 by IT Corporation.

IT acknowledges the guidance and assistance provided by Anthony Tafuri, UST Re-
search Program Manager, Michael Grunfeld, NRMRL's Project Officer, and Chi-Yuan
Fan, NRMRL's Work Assignment Manager for this Work Assignment.

This document was produced under the direction of Robert Amick, IT's Program Direc-
tor. Floy Chaudet served as the initial Work Assignment Leader and produced the first
draft.  Thomas Clark, the current Work Assignment Leader, oversaw development of
the final document. Dr. Richard Johnson of the Oregon Graduate Institute of Science
and Technology, Dr.  R. Ryan Dupont of Utah State University, and Duane Graves of IT
are the principal authors. Technical review of the tracer tests was provided by Dr. Paul
Johnson of Arizona State University.  Jerry Day provided the editorial support, and
Sherri O'Regan prepared the manuscript.
                                     XIII

-------
                                   Section 1

                                 Introduction
Purpose
It is often difficult to accurately assess the performance of systems for remediating soils
and groundwater. This is due in large part to the complexity and heterogeneous nature
which exists in the subsurface at a given site. Because of the complex subsurface
characteristics encountered at leaking underground storage tank (UST) sites, a series
of tests/tools are needed for evaluating the appropriate application and remediation per-
formance of selected corrective action technologies.

In response to this need, the U.S.  Environmental Protection Agency (EPA) Office of
Research and Development (ORD) National Risk Management Research Laboratory
(NRMRL), UST Research Program has provided technical support to EPA Regions for
evaluating selected corrective action technologies.  Under this program, NRMRL devel-
oped a series of test procedures for evaluating intrinsic biodegradation, bioventing, soil
vapor extraction (SVE), and in situ air sparging (IAS). The test procedures described
here are diagnostic tools designed to aid in assessing remediation performance. While
most of the discussions focus on the widespread problem of petroleum hydrocarbon
contamination, these procedures can also be used  to evaluate remediation perform-
ance of technologies used to address a wide range of contaminants.
Background
Releases of petroleum hydrocarbons fuels (e.g., gasoline, diesel, aviation fuel) fre-
quently contaminate soil and shallow groundwater, and are widely recognized as an
important environmental problem. Fuel spills are different than many other groundwater
contamination problems, in that much of the fuel often exists in the subsurface as a
nonaqueous-phase liquid (NAPL). NAPLs can be distributed in the subsurface as iso-
lated "blobs" within pore spaces, or as larger continuous zones or "pools'1 near the wa-
ter table or associated with heterogeneous zones in the medium. Vadose zone soils in
the source area often contain "residual" hydrocarbons in isolated pore spaces. This
residual may account for 10 to 60 liters of hydrocarbons per cubic meter of soil. Be-
cause petroleum NAPLs are less dense than water (LNAPLs), they often accumulate at
the water table. They can occupy a significant fraction of the pore space in this region,
and can often move  up  and down as the groundwater level changes. As'the result of

                                      1-1

-------
fluctuating water tables, the NAPL can become entrapped below the water table in what
has become known as a "smear zone."

At many sites, LNAPLs both above and below the water table represent long-term
sources of contaminants. (Dense  NAPLs also cause significant subsurface contamina-
tion and can be remediated by SVE and IAS; however, the focus of this discussion will
be on LNAPLs.) As groundwater  flows through NAPL zones, dissolution forms
aqueous-phase plumes.  These plumes move under the influence of groundwater flow
and can migrate for significant distances.  Because of limited solubilities of the fuels,
the sources can be long-lived. As a consequence, treating the  groundwater plume
without removing the source can result in regrowth of the groundwater plume. This
means that it is often necessary to treat both the plumes and the sources, although the
approaches for each may be significantly different.

Remediation of NAPL-contaminated zones has generally been  accomplished by flush-
ing the appropriate fluid through the medium (i.e., air in the vadose zone, water in the
groundwater zone). Under those  "pump-and-treat" scenarios, mass is removed from
the vadose zone by volatilization and from the groundwater zone by dissolution.   Both
of these technologies have demonstrated successes and failures. Failures can often
be traced to heterogeneities in the medium and/or to the inability to effectively remove
NAPL.

IAS is a relatively new remedial technology. During IAS, air is injected into the satu-
rated zone to volatilize contaminants in much the same way that contaminants are
volatilized in the unsaturated zone.  In addition to volatilization/dissolution, remediation
occurs by biodegradation. The operational conditions under which biodegradation is
optimized are somewhat different  than for volatilization/dissolution. For biodegradation,
the primary criterion will be how well oxygen can be delivered to the contamination
zone. In shallow, well-drained soils, molecular diffusion may be adequate to maintain
aerobic conditions. In many cases, however, it will be necessary to actively flush air
through the system to maximize biodegradation. Similarly, in some cases in the
groundwater zone, there may be an adequate supply of compounds which are useful to
degrade petroleum hydrocarbons  (i.e., electron acceptors).  In other cases it may be
necessary to enhance the delivery of electron acceptors to achieve significant biodeg-
radation in the groundwater zone.

There are a number of criteria which are important to the success of remediation  by
flushing. These include:
                                                                 i
      •   Volatility/solubility of contaminant of interest
      •   Distribution of contaminants
      •   Current and potential biodegradation activity
      •   Soil permeability                                         ;

                                      1-2

-------
       •    Heterogeneity of the medium             ;                  i
       •    Flow pathways of the flushing fluid

Of these six criteria, the physical properties of the contaminants and the bulk soil per-
meability are relatively easy to determine.  For NAPL contamination sites,  however, the
heterogeneity of the medium (which is extremely important in controlling the distribution
of contaminants) is very difficult to determine using conventional site characterization
techniques, and significantly impacts how the flushing fluid flows through;the medium,
and therefore how well it contacts the NAPL.  This difficulty is the reason the tracer test
procedures presented here were developed.

Finally, the role of biodegradation, both in the vadose and groundwater zones, not only
depends on the physical characteristics of the medium, but also on the chemical and
biological characteristics of the system.  These characteristics are often difficult to mea-
sure, and even if they are determined it is difficult to predict the role of biodegradation.
Therefore,  practical field procedures are presented here which directly measure biodeg-
radation activity in the field. These measurements form the basis for determining the
potential for successful bioremediation.

The five test procedures described  here are diagnostic tools which can be used to eval-
uate remediation performance. Three of the procedures are tracer tests which can be
used to evaluate air flow in the subsurface (SVE air flow, IAS air recovery, and IAS air
distribution).  The tracer tests are new procedures which have been tested at a small
number of sites and can be expected to undergo significant changes in the future. The
other two procedures are used to evaluate biodegradation in the subsurface (bioventing
and natural attenuation). The biodegradation procedures have been demonstrated at a
much larger number of sites and as a consequence are likely to undergo fewer chang-
es. The five procedures have been designed as "stand-alone" sections. Section 2 pro-
vides procedures for SVE air flow tracer tests. Section 3 discusses IAS air recovery
tests. Section 4 provides a procedure for determining  air distribution in the saturated
zone during air sparging.  Section 5 presents  procedures for evaluating bioventing.
Section 6 summarizes test procedures for evaluating natural attenuation.  Depending
on the remedial objectives, these five procedures can be used either independently for
a separate  corrective action approach or, where appropriate, can be used  together in
an integrated approach to evaluate remediation.
                                       1-3

-------
                                   Section 2

                   Procedures for Conducting Tracer Tests to
                 Evaluate Air Flow During Soil Vapor Extraction
Introduction
Introduction to Soil Vapor Extraction (SVE)
SVE is a remediation technique in which air is drawn through the subsurface to remove
volatile contaminants and to enhance aerobic biodegradation. It is generally designed
to optimize mass removal through volatilization, although some biodegradation occurs
in virtually all hydrocarbon remediation settings.  In that context, conventional SVE and
bioventing can. perhaps be thought of as end members of a continuum of air flushing
technologies. SVE is generally used to remediate "source  zones" where contaminants
may be present as nonaqueous-phase liquids (NAPLs). These NAPLs areas can be
distributed within the unsaturated zone as a "residual" NAPL, perched above lower-
permeability/higher-water-content zones, or in the vicinity of the water table (Fig-
ure 2-1).  SVE is generally not applicable to the remediation of groundwater zone con-
tamination. However, if water tables fluctuate significantly, seasonally the SVE system
may be able to access portions of the subsurface zone which are sometimes saturated.

The success of SVE generally depends upon both the chemical/physical properties of
the contaminants of concern and the physical properties of the soil to be remediated.
For the purposes of this discussion it will be assumed that the contaminants of interest
are petroleum hydrocarbons.  To be effective, SVE must establish good contact be-
tween the flowing air and the contaminant (NAPL).  Important factors include the per-
meability of the soil, the scale and severity of the heterogeneity, and soil water content.
These factors will determine the distribution of the NAPL, the  rate at which  it will be pos-
sible to draw air through the system, and the pathways the air will follow.

To be effective, the contaminant of interest must also be sufficiently volatile to  be effec-
tively removed from the soil. Since the contaminants may be associated with the air,
water, soil,  or NAPL phases, all of these must be considered in the context of volatiliza-
tion. Volatile gasoline-related compounds such as butane, pentane and hexane, for
example, will be primarily associated with the air, water and NAPL phases. As a conse-
quence, if air can be induced to move within a contaminated soil, they will be quickly
removed.  Less-volatile contaminants, such as those found in diesel or weathered
                                      2-1

-------
                                     ^RESIDUAL

                                "SMEAR ZONE"
 WATER TABLE
. FLUCTUATIONS
Figure 2-1.  Schematic drawing of petroleum hydrocarbon distribution in the
subsurface.
                                    2-2

-------
gasoline, do not readily volatilize and must therefore be treated over a longer timeframe
via biodegradation or by some more-aggressive remediation technique (e.g., heating,
chemical flooding). The relationship between vapor pressure and molecular weight is
shown in Figure 2-2.  The figure implies that the volume of air required to remove a
given compound increases exponentially with increasing molecular size.

SVE System Design
The typical components of an SVE system are shown in Figure 2-3. One or more air
extraction wells are generally placed in the zone of highest contamination. This is done
in order to maximize the recovery of volatile contaminants. One or more injection wells
(either actively or passively injected) may also be  placed in the zone of remediation.
The patterns of these wells may follow some simple geometry (e.g., concentric circles)
or may be optimized for site stratigraphy. The construction of both the extraction and
injection wells is often similar to conventional monitoring wells/points. Typically they will
consist of 2-inch or 4-inch PVC well screen with a length of 5 to 20 feet connected to a
PVC stand pipe.  The wells are often installed using a hollow-stem drill rig; however,
extraction points  may also be implaced by direct pushing or other means. The extrac-
tion wells are often connected to a common manifold, which is in turn connected to the
air handling and off-gas treatment system.  These systems are made up of the following
components:

       1.   A vacuum blower which is capable of 10 to 200 standard cubic feet per min-
          ute (scfm) at a 0.5- to 10-psi pressure  drop. (The specific performance of
          the blower will depend upon site conditions, number of wells, off-gas treat-
          ment stream, etc.)
                                                                 '   v,
      2.   A "knockout drum" for removing liquids from the air stream will generally pre-
          cede the vacuum blower.  It must be designed to handle the range of sub-
          ambient pressures produced by the blower.

      3.   At most sites, some type of off-gas treatment of the blower exhaust will be
          required. This may include combustion, carbon filtration, biodegradation  or
          other approaches.

SVE Operating Conditions
SVE has  been  applied over a wide range of field operating conditions. These are sum-
marized briefly below:

      Depth - SVE systems can be installed over a range of depths from 10 to 500+
          feet. With shallow systems, one of the most difficult obstacles is minimizing
          surface leakage. With deeper systems, heterogeneity of the medium be-
          comes a potentially important limitation.
                                      2-3

-------
I
3
          0.01  -
         0.001  -
0.0001  -
       0.00001  -
      0.000001  -
      0.0000001
     0.00000001
                            5               10
                               NUMBER OF CARBON ATOMS
                                                   15
                                                                   20
   Figure 2-2.  Relationship between compound vapor pressure and number of
   carbon atoms.
                                      2-4

-------
     a)
                                      GROUNDWATER
                                      PLUME

b)





!> f, f
?< i-w*-.. J
•Afcte*-. . ^


t^t/irsat
= Tw?t
fct-i - r»
NAPL. at.-'BP

X
^\
^\^


                                                      x  ;
                                                Off-gas treatment
                                                Extraction blower
                                                Knockout drum
                                         Extraction well  i
Figure 2-3. Schematic drawing of a typical SVE remediation system in plan view
(a) and section view (b).
                                 2-5

-------
       Flow rate - Flow rate is generally in the range of 10 to 100 scfm per well. This
           depends upon a number of factors, including the number of wells and the
           permeability, heterogeneity, and water content of the soil.

       Soil permeability - SVE systems have been shown to be effective jn soils ranging
           from gravel to clay. Once again, soil heterogeneity and water content are as
           important as bulk permeability.

       Water content - Water content affects not only the air permeability of the soil, but
           also the pathways in which the air moves through the soil and the mass
           transfer of contaminants from the soil to  the air.  For example, in lower per-
           meability soils, which often have higher water content, air may flow exclu-
           sively through channels separated by tens of centimeters. In these cases,
           mass transfer from the water-saturated zones between the channels may be
           quite slow.

Summary
SVE is a remediation technique which has been demonstrated to effectively remove
volatile contaminants from a wide variety of soil types. In many cases, SVE has suf-
ficiently remediated sites to allow their closure. In other cases, however,;remediation
has proved difficult. The reason for failure in these cases can often be traced to nori-
uniform air flow due to soil characteristics (heterogeneity, high water content, etc.). The
procedures described in the following section provide a means of assessing air flow
pathways and, as a consequence, evaluating the remediation performance using SVE.

Introduction to Air Tracer Tests
At most sites where SVE and/or bioventing using vapor extraction (BV) is used, it is dif-
ficult to relate measured soil vacuum data to the air  flow field.  Vacuum data are fre-
quently used to define the radius of influence; however, the  vacuum data do not provide
much insight into the structure of the soil or the airflow pathways through the soil.
Vacuum data tend to present a picture of the flow field which is much more uniform
than is generally the case.  Small strata of lower or higher permeability can have
profound effects on flow patterns, and these effects  may not be reflected in the vacuum
data.  At many sites, there is more flow from the surface than is commonly assumed,
and at many sites there is less flow near the water table than is commonly assumed.
As a consequence, at many sites the time required for soil cleanup using SVE/BV is
much longer than  predicted, based on simple calculations or analytical models.

Tracer tests to directly measure the air flow field are easy to perform, and have the
potential to significantly improve the conceptual model of how air is actually flowing at a
site.  Both naturally occurring and introduced compounds can be used as' tracers.
Oxygen and carbon dioxide concentrations can be used to assess where air is flowing
                                      2-6

-------
in the subsurface.  Inert gases such as helium or sulphur hexafluoride can be injected
into the subsurface and tracked in situ and in SVE/BV off-gas.

Test; Objectives
The primary objectives of SVE/BV tracer tests are to assess air velocities within the
subsurface, and to assess the remediation performance of SVE/BV. The tests can be
used to answer a number of critical questions, including:

      •    Do short-circuit pathways for the air exist?
      •    Are there significant stagnant zones?
      •    Is the zone of contamination being effectively flushed?
      •    Are current operating conditions optimizing vapor flow rates through the
           zone of contamination?
      •    Is SVE/BV delivering oxygen for biodegradation?

Theory
The procedures for conducting air flow tracer tests are straightforward. However, the
interpretation of the observations may be more complicated, and numerical modeling
may be required to fully understand airflow patterns.  In addition, the number, types,
and locations of the tracer injection points will have a significant impact on the extent to
which the tests can be interpreted.

The basic procedure is to make an instantaneous injection of tracer at a point at some
distance from an SVE/BV well and to watch the arrival of the tracer at the SVE/BV well
or at intermediate points.  The injection point can be a monitoring well, a soil gas probe,
or virtually any other preexisting point in the unsaturated zone.  However, locations with
small diameters and short screened intervals will generally work best for these tests. In
this context, driven soil gas probes are ideal  as tracer sources.  As soon ;as the tracer is
released  into the injection point, it will begin to move towards the SVE/BV well. Its
transport velocity will depend upon the air permeability of the medium and the pressure
gradient at  that point.

If a single SVE/BV well is being used, the vacuum gradient and flow field strength will
generally increase  in a nonlinear manner as the tracer approaches the well. If high-per-
meability (i.e., high velocity) pathways exist, they may be at lower pressure than other
portions of  the system. As a consequence, the tracer will  move towards a high-velocity
zone. If the tracer is injected into a high-permeability zone, arrival at the SVE/BV well
will be rapid. If it is injected into a lower-permeability zone or at a point which is sepa-
rated from the extraction well by low-permeability media, travel time will be longer. In
addition,  if the tracer is injected into a lower-permeability zone, the  resulting "pulse" of
tracer arriving at the well may be spread out over a much  longer time interval than the
high-permeability case.
                                      2-7

-------
The tests should be conducted at a variety of distances from the SVE/BV well and at a
variety of depths. The latter is particularly important because of the importance of the
ground surface boundary and because unconsolidated soils often have a dominant hor-
izontal structure. Again, it is very important to keep the vertical extent of the screened
interval to a minimum. In this context, driven soil gas probes are ideal. It is also impor-
tant to place injection points in close proximity to the zones of contamination. In this
context, it is important to define the zone(s) where contamination is present. For pe-
troleum hydrocarbons, for example, these zones will often be associated,with lower-
permeability materials in the unsaturated zone, or near the water table. As such, the
effectiveness of mass removal will depend upon the soils in the immediate vicinity of the
contamination. If the tracer tests are to be effective in evaluating air flow;patterns, it is
important to understand flow from these contaminated zones.

Steps in Conducting an Air Flow Tracer Test
There are eight major steps in the execution of an SF6 tracer test (Figure 2-4):

      1.    Injection of the tracer - Under most circumstances an "instantaneous" injec-
          tion of a known volume of SF6 will be used.

      2.   Monitoring of the SVE/BV offgas - The SF6 detector will be configured to col-
          lect a sample from the extraction manifold at regular time intervals (e.g., 2 to
          20 minutes), depending upon the duration of the test.

      3.   Plotting of the SF6 breakthrough - SF6 concentrations in the offgas will be
          automatically determined using the SF6 detector.  These data can then  be
          plotted as a function of time.

      4.   Calculation of the volume of SF6 recovered - The concentration data can be
          multiplied by the SVE flow rate and the time between sample intervals to de-
          termine the volume of SF6 recovered during each  interval. These data can
          be summed to determine the total volume of SF6 recovered.

      5.   Estimation of the fraction of SF6 recovered - The total volume estimated in
          the previous step is compared to the injected volume of SF6 to calculate the
          fraction  of SF6 recovered.  Recovery should be between 50 and 150 per-
          cent.

      6.   Estimation of the breakthrough time - The volume data can be used to de-
          termine the breakthrough time of the tracer. The breakthrough time is the
          time at which 50 percent of the recovered volume  is reached.
                                      2-8

-------
                                           fn/ection of tha tracer

                                       Instantaneous injection of a known
                                               volume of SF6
                                     Monitoring of the SVE/BVoff gas

                                      Sample collected from the extraction
                                         manifold at regular intervals
                                        Plotting SF6 concentrations

                                        SF6 concentrations plotted as a
                                               function of time
                                  Calculation of volume of SF6 recovered

                                   Concentration data multiplied times the flow
                                   rate and interval between sampling; products
                                   summed to get cumulative volume recovered
                                      estimation of percent recovery

                                      Recovered volume compared to the
                                            injected SF6 volume
                                     Estimation of breakthrough time

                                    Breakthrough time estimated as point at
                                    which 50% of recovered mass is reached
                                  Calculation of effective air-filled porosity

                                   Breakthrough time, distance from extraction
                                   and injection points, and extraction rate are
                                      used to calculate air-filled porosity
                                  Calculation of velocity atin/ecfion point

                                  Assuming cylindrical flow the velocity at any
                                     point can be calculated using a simple
                                            geometric equation
Figure 2-4.  Flow diagram showing the major steps involved in in airflow tracer
test.
                                                2-9

-------
      7.   Calculation of effective air-filled porosity using a simple geometry - If it is ini-
           tially assumed that flow to the extraction well is radial, the velocity increases
           with decreasing distance to the well.
           •   Calculate the volume of the cylinder containing the test (nr2h). Where r
              is the distance between the injection and extraction points and h is the
              thickness over which air is assumed to be flowing.
           •   Calculate the extraction volume (flow rate x breakthrough time).
           •   Calculate the effective porosity (n, breakthrough volume/cylinder vol-
              ume).
           •   In an ideal case, the value of n determined from each tracer test would
              be the same. In practice, n will vary depending upon the conditions en-
              countered in the subsurface.
           •   If the calculated value of effective porosity is not reasonable (e.g., great-
              er than 1 or less than 0.1), any of a number of conditions shown in
              Figure 2-5 might exist (e.g.,  surface leaks, high-permeability zones, low-
              permeability zones).
           •   If high-permeability zones exist, travel times may vary in a manner which
              does not relate well to distance. This is shown schematically in Fig-
              ure 2-6.                                             ;

      8.   Calculation of the velocity at the injection point - For the case of simple,
           homogeneous media with no surface leakage (i.e., cylindrical flow), velocity
           at any point can be described by the equation:
                                        cr
                                V =
                                     2T\Rhn                        \

          where   F  =  the extraction rate
                   R  =  the distance from the point of interest to the extraction well
                   h  =  the height of the cylinder
                   n  =  the effective air-filled porosity.

Interpretation of the Air Flow Tracer Tests
Once several tests have been conducted, it is possible to compare travel time and ve-
locity data and develop a travel-time pattern for the site.  For example, tracer injected at
points near the water table at a certain distance from the extraction well may have
much longer breakthrough times than points nearer the surface at the same distance.
Those data can be used to identify additional locations within the remediation zone to
conduct tracer tests.

If the site is a complicated one, or if there are several extraction wells, it may be neces-
sary to evaluate the test data using an appropriate numerical model. The numerical
model should be sophisticated enough to handle the complexities found at the site.  In

                                      2-10

-------
 a)
 b)
•f

                                     ^
                                    Lower Permeability
 d)
Figure 2-5.  Schematic drawings of possible SVE flow configurations.
                                2-11

-------
 Tracer
 Injection 1
 Tracer
 Injection
 Tracer
 Injection
                                                     High-k Zone
    0.5


>    0

1
e-  °-5
CO
""    0


    0.5


      0
         0
                                Tracer Injection 1
               Tracer Injection 2
                                            Tracer Injection 3
                 200      400       600
                         Time (Minutes)
800
1000
Figure 2-6. Schematic drawing of tracer "short-circuiting" through a higher-
permeatibility zone.
                              2-12

-------
general terms the model should include both flow and transport.  It should be able to
run either 3-D or 2-D axis-symmetric problems, and should be able to incorporate het-
erogeneities (e.g., layers) into the flow field.

Test Equipment
Overview of Experimental Setup
In order to simplify interpretation, the tests should be conducted using a single SVE/BV
well. Tracer injection points should be placed  in at least three depths and at a variety of
locations of interest.  If contamination is present near the water table, it is important to
place tracer injection points there. Locations of the other points should be determined
by stratigraphy and soil/vapor concentrations.  They can be placed in both high- and
low-permeability zones.  In that context, soil boring logs may be useful in identifying site
stratigraphy.  Chemical analyses of cores or cuttings may be useful in identifying zones
of contamination. Soil gas surveys may help identify those regions which are not being
effectively remediated.

Design and Installation of Extraction Wells
In nearly all cases, the air flow tracer tests will be conducted in conjunction with vapor
extraction or injection operations. In that context, the design and installation of the ex-
traction/injection wells will be dictated by the remediation design. As a consequence,
design and installation of the extraction/injection wells will not be discussed  here. How-
ever, the design of the extraction/injection wells will have an impact on the design and
installation  of the tracer injection points.

Design and Installation of Tracer Injection Points
Wherever possible, tracer tests should be conducted using existing monitoring wells
and vapor monitoring points. However, monitoring wells with screened lengths of great-
er than 5 feet should not be used. Frequently, test points will be part of other testing
(e.g., in situ respirometry). In general, the locations are selected on  the same basis as
for in situ respirometry and soil gas measurements. The primary criterion for placement
is that they be within contaminated zones so that air flow within those zones will  be di-
rectly measured. Vapor monitoring points, especially driven vapor probes, are the best
choice for tracer injection points. Again, they are best if the screened interval is  small.

Most regulatory agencies do not regulate unsaturated-zone monitoring point construc-
tion. Nevertheless, prior to construction it is necessary to check with regulators to as-
sure compliance with any regulations that may exist. Monitoring  point construction will
vary depending on the depth and the installation techniques. The following  approaches
are commonly used for installing points:

      1.   Hand-augured holes
      2.   Drill-rig augured holes
      3.   Pushed or driven probes

                                      2-13                         '

-------
 Tracer injections should be made at a minimum of three locations, and at each location
 to at least three depths. Whenever possible, the monitoring points should be located in
 contaminated soils.  A sufficient number of points should be emplaced to characterize
 flow through the contaminated soil. If contamination has reached the water table, it is
 important to place points near the water table. Points should be located in both high
 and low permeability zones so that spatial variability in velocity can be examined.  This
 typically involves a minimum of 3 or 4 locations and 3 to 4 depths each.  Recommended
 spacings for tracer injection points are listed in Table 2-1. In general, the minimum
 number of points will increase if the site is heterogeneous.
Table 2-1. Recommended Spacings for Tracer Injection Points
(adapted from Hinchee et al., 1992).ab
Soil Type
Coarse sand
Medium sand
Fine sand
Silt
Clay
Depth to Top of Vent Well Screen
(feet)
1.5
3
>5
1.5
3
>5
1.5
3
>5
1.5
3
>5
1.5
3
>5
Radial Distances for Placement of
Monitoring Points (feet)
1,3,6
3, 6, 12
6,9,18
3,6, 9
4, 8, 12
6, 12, 18
3, 6, 12
4,9, 18
6, 12, 24
3, 6, 12
4, 9, 18
6, 12, 24
3,6,9
3, 6, 12
4,9, 18





     a Assuming 10 ft of vent well screen.  If more screen is used, the >15-ft spacing will be used.
     k Note that monitoring point intervals are based on a venting flow rate of 1 cfm/ft screened
       interval for clays to 3 cfm/ft screened interval for coarse sands.

In general each areal location should have a minimum of three tracer injection points at
various depths.  If contamination exists at the water table, one or more points should be
located at that depth. The shallowest screen will normally be 1 to 2 m below ground
surface.  One or more points should be installed between these points.  In general, one
point can be located near the depth of the top of the extraction well screen.  If the soil in
the contaminated zone is heterogeneous, additional points should be installed at appro-
priate depths.  In this context it is important to have enough site characterization data to
allow good placement of the points.
                                       2-14

-------
The ideal design of a tracer injection point, in most cases, will consist of a short (10 to
30 cm) well screen connected to ground surface via a small-diameter tube.  Depending
upon the emplacement technique, the screen may be positioned within a sand pack or
in direct contact with the soil.  In general, single completion wells are preferred over
multiple completion wells. This is because of possible problems in isolating the various
layers from one another.  If points are installed with a drill rig, it may be necessary to
use multiple completions to minimize cost.

In general, materials used for monitoring well construction will be appropriate for tracer
points. The specific materials chosen will depend upon the types of samples to be col-
lected from the point. (For SF6, Teflon tubing should be avoided.  The preferred choic-
es are high-density polyethylene or stainless steel.)

Each point should be clearly labeled with a unique identification code, which includes,
at a minimum, designators for the areal location of the point and for the vertical position
of the point. In most cases, the vertical position should  be identified using the depth of
the well screen below ground surface.

Calibration of Analytical Equipment
SF6 [Detector                                                     ;
There are several types of detectors for SF6 analysis. A gas chromatograph with an
electron capture detector is the basic instrument for detecting SF6 at low concentrations
in vapor samples. There are a wide variety of commercially available gas chromato--
graphs ranging from sophisticated research instruments to more "user friendly" instru-
ments. An SF6-specific gas chromatograph is available from Lagus Applied Technol-
ogies (LAT) in San Diego, CA.  It is automated and has a detection limit of ~10 parts
per trillion by volume.  In the following discussion, it will  be assumed that a LAT
Autotrac is being used.

One of the features of the LAT Autotrac is that it comes with an internal cylinder of cali-
bration gas. As a consequence, single-point calibrations in the field can be easily ac-
complished. However, it is desirable to conduct multipoint calibrations to verify linear
instrument response. Standards can be obtained commercially, prepared by dilution, or
by using a dynamic calibration instrument.

Air Flow Meters
Flow rates for the SVE system will generally be in the 10- to 200-scfm range.  Vacuum
levels will  be at 10 to 200 inches of water below ambient pressure. The flow meter cal-
ibration system is shown in Figure 2-7. At the high flow rates, a large dry gas meter will
be required. An alternate approach will be to use another calibrated flow meter to
                                      2-15

-------
                       VACUUM
                        GAUGE
VACUUM
PUMP
                       VALVE
                                     FLOW
                                     METER
    DRY GAS
    METER
Figure 2-7.  Schematic drawing of the flow meter calibration system.
                                   2-16

-------
calibrate the one to be used at the site. Actual versus observed flow rates should be
determined over the range of the flow meter at several vacuums between 0 and 0.9 at-
mosphere. Those data should be plotted as a family of curves with each line corre-
sponding to a different vacuum value.

Flow rates for the IAS system will generally be less than those used for the extraction
system. However, pressures will be above atmospheric, rather than below.  The appa-
ratus used to calibrate flow meters at above atmospheric pressure is shown in Fig-
ure 2-8.

The tracer gas can be  used not only for the tracer tests, but also to determine the
extraction air flow rate.  To measure extraction air flow rate, a continuous injection of
tracer into the extraction manifold can be made, and the observed dilution of the tracer
can be used to determine the extraction air flow rate.  In general, flow rates for
continuous tracer injection will be in the range of 0.1 to 1 L/min. The apparatus for
calibrating the flow meter for continuous tracer injection is the same as for air injection
(Figure 2-8).

Evaluation of the Sampling Pump
Under most operating conditions, the SVE/BV manifold will be under sufficient vacuum
that most automated analytical equipment will not be able to draw a sample from the
manifold.  In these cases, it will be necessary to use a good quality vacuum pump to
draw samples from the manifold and deliver them to the automated analytical equip-
ment.. In the context of tracer tests, two potential problems arise with respect to the
pump.  First, the pump must be able to move sufficient volumes of air to meet the
needs of the analytical equipment, and second the pump should not leak air, which can
provide additional dilution of the sample stream.  The procedures  below describe how a
pump performance can be measured.

Prior to obtaining a sampling pump, check the specifications of any automated sam-
pling equipment to be used to determine the volumes of air required by  each. (The
Autotrac detector requires ~100 mL/min.) Next, connect the sampling pump to be
tested to the apparatus shown in Figure 2-9.  If two dry test meters are not available,
two calibrated flow meters of the appropriate ranges can be used. Turn on the pump
and open the valve so that no vacuum is observed on the gauge.  Determine the flow
into and out of the pump by recording the volume of flow that occurs in one minute on
each of the dry test meters. Partially close the valve to produce a vacuum of 5 inches
of mercury, and determine the flow rates into and out of the pump.

Repeat the previous procedure with vacuums of 5, 10, 15, 20, and 25 inches of mercu-
ry.  Prepare a plot of flow rate in and out versus vacuum.  Determine if the pump flow
rate is adequate for the analytical equipment and vacuum to be used. Determine if the
                                     2-17

-------
                    VACUUM
                     GAUGE
                     -$3-
                    VALVE
   DRY GAS
    METER
                                FLOW
                                METER
                                            COMPRESSOR
Figure 2-8.  Schematic drawing for flow meter calibration under positive
pressure.
                             2-18

-------
              VACUUM
               GAUGE
                         SAMPLE
                          PUMP
  DRY GAS
   METER
DRY GAS
 METER
Figure 2-9. Schematic drawing showing setup for measurement of
sample pump flow rate vs.vacuum and leakage vs. vacuum.   ,
                          2-19

-------
 leakage of the sampling pump is acceptable (e.g., inflow rate is within a factor of two of
 outflow rate).

 Test Procedures
 Qveiview of Procedures
 Test activities can be divided into the following seven components. Each of them are
 described briefly in the following sections.

       1.  Installation of SVE/BV extraction wells
          - Usually part of some remediation strategy.
       2.  Installation/identification of tracer injection locations.
       3.  Calibration of analytical instrument.
       4.  Calibration of the air flow system/determination of SVE/BV flow rate.
       5.  Make preliminary measurements.
       6.  "Instantaneous" injection of SF6 into soil gas point.
       7.  Collect samples from SVE/BV off-gas.
       8.  Determine effluent concentrations.

 Measurement of SVE/BV Flow
 It is necessary to determine the actual flow of the SVE/BV system if mass removal rates
 are to be calculated. .Measurement of SVE/BV air flow can be made either by using a
 direct-reading flow meter or a tracer. Both procedures are described below.

 Flow Measurement Using a Flow Meter
 If a flow meter is used to measure SVE/BV flow,  it is essential that the flow meter be
 calibrated by the procedure described above.  This needs to be done at the appropriate
 pressure/vacuum. The apparatus is shown in Figure 2-10. Prior to measuring flow, the
 extraction flowrate should be adjusted to the desired flow/vacuum using the ball valve
 between the flow meter and the pump.  If more than one extraction well is being used, it
 is desirable,  although not necessary, to measure the flow at each well.  This involves
 placing a ball valve, vacuum gauge, and a flow meter in line with each well.  Read the
vacuum and flow rate from the gauge and meter and use the calibration table for that
flow meter to determine actual flow.

 Flow Measurement Using the Tracer Gas to Determine SVE/BV Flow
 If an independent calibration of the flow meter is  not available, the tracer gas can be
used to calibrate the SVE/BV flow system.  In some cases (especially when the extrac-
tion vacuum  is large), water in the system makes it difficult to get good flow meter read-
ings, and the tracer procedure is preferred.                            ;

The apparatus for tracer calibration is shown in Figure 2-11. Prior to beginning the
procedure, it is necessary to estimate the SVE/BV flow.  The specifications for the
extraction blower are adequate for this.  Next, calculate SF6 inflow rate to produce a

                                     2-20                         ;

-------
                                                              TO
                                                              OFF GAS
                                                              TREATMENT
                       KNOCK-OUT
                       DRUM
          s "T
                                                        EXTRACTION
                                                        BLOWER
           VACUUM GAUGE (OPTIONAL)


     BALL VALVE

           VACUUM GAUGE
     FLOW METER
   FROM
   EXTRACTION
   WELL
Figure 2-10. Schematic drawing of the flow measurement system for the
SVEfBV system
                                2-21

-------
                                                                    TO
                                                                    OFF GAS
                                                                    TREATMENT
                           KNOCK-OUT
                           DRUM
   FROM
   EXTRACTION
   WELL
EXTRACTION
BLOWER
                                             EXHAUST LINE
Figure 2-11. Schematic drawing showing the system for measuring SVE/BV
flow using a tracer gas.
                                    2-22

-------
 concentration of 1 ppbV.  An air flow rate in the 0.1- to 1-L/min range is desired. (Use
 an SF6/air mixture with a concentration of ~1000 ppmV for this.)  Example Calcula-
 tion 2-1 shows how these calculations are to be done. Next, install a go9d vacuum
 pump (metal bellows or diaphragm) to the manifold-this will be the same setup as for
 the tracer tests themselves.  Make sure the pump does not leak at the system pressure.
 Connect the SF6 detector to the sampling pump.
  Determination of tracer injection rate
        Approximate SVE Flow rate = 35 scfm = 1000 L/min
        tracer mix concentration = 100 ppmV = 105 ppbV
        Desired final concentration = 10 ppbV
        need a dilution of 104
        tracer flow rate = 1000 L/min / 104 = 0.1 L/min

  Calculation of actual SVE/BV flow

  Using a flow rate of 0.1 L/min and an input concentration of 105 ppbV the actual
  SVE/BV flow rate is determined by dividing the input concentration by the effluent
  concentration  and multiplying by the tracer injection flow rate.
  If the observed effluent concentration were 20 ppbV, the actual flow would be:

        105 / 20 * 0.1 L/min = 500 L/min ~ 17.7 scfm

    Example Calculation 2-1. Calculation of actual SVE/BV flow using a tracer.
To begin the flow measurement, connect the SF6 source to the manifold near the ex-
traction point and add SF6 at the prescribed rate. Monitor tracer concentration in the
extraction system until it stabilizes.  (This should take only a few minutes.)  Calculate
the observed dilution factor (influent tracer concentration/effluent tracer concentration).
The SVE flow rate is calculated by the dilution factor times the SF6 inflow rate.

Vacuum Survey
The procedure described  below assumes that the remediation system is already in op-
eration. It is important to collect subsurface vacuum data prior to initiation of the tracer
tests. These data provide insight into the general nature of the flow system. For exam-
ple, if little or no vacuum is recorded at a monitoring point, it can be expected that there
is little flow at that point. Large vacuums may indicate areas of active flow; however,
these values can also occur within low-flow regions adjacent to higher flow regions.
Nevertheless, these data can frequently be helpful in understanding the general nature
of air flow at the site.
                                      2-23

-------
The general approach will be to measure soil vacuum with differential pressure gauges
(e.g., Magnehelic™ gauges).  However, the same measurements could also be made
with a manometer or other calibrated gauge.  For most sites, it will be necessary to
have Magnehelics in the following ranges (in inches of water): 0 to 1 inches, 0 to
10 inches, and 0 to 100 inches.  Vacuums of less than 1 inch of water are subject to
error from a variety of sources including barometric pressure effects and should be in-
terpreted with caution.  After the remediation system has been operating iseveral hours,
determine the soil vacuum at each point in the system by connecting the -appropriate
gauge to the  point.  After connection to the monitoring point, sufficient time should be
allowed for the vacuum to stabilize (commonly 1 minute).

Soil Gas Permeability Test Procedures - Steady-State Method
Prior to  beginning the soil gas permeability test, examine the site for any structures that
could serve as vertical conduits for gas flow. These must be sealed to prevent short-
circuiting and to ensure the validity of the soil gas permeability test.  Next, values of the
permeability and the radius of influence can be estimated using the equations de-
scribed  in the following example calculations. For the purposes of the tracer tests,
estimation of permeability using the steady-state method is acceptable.

Estimation of the "Radius of Influence"
The term radius of influence should be used with caution. The procedure described
below is often used to estimate the region over which the SVE system is effective. As
described in this procedure, airflow can be spatially quite variable and as a conse-
quence  a pressure-based radius of influence is subject to considerable uncertainty.

The example in Figure 2-12 is taken from Hinchee et al., 1992, p. 77.  The first step is to
conduct a soil vacuum survey such as the one described in the previous section. Next,
plot the  final pressures of each point versus the log of its radial distance from the vent
well. Then, extrapolate the best-fit straight line through the data to zero vacuum. That
distance is the estimated radius of influence. The radius of influence in Figure 2-12 is
-30 rn.

Estimation of Permeability
The example  is once again taken from Hinchee et al., 1992, p. 77. The equation for
permeability to vapor flow from steady-state pressure data is:
                                       1 ~(PatmlPw)2]

where    k   = permeability
          Q  = flow rate (cm3/s)
          u   = viscosity of air (1 .8x1 0"4 g/cm-s)
          Rw -radius of well (cm)

                                     2-24
                             k = - —• -                 (2-2)
                                                 2

-------
J.U
x—s 9
CD 0
•& 8
* 7
q-H
0 ^
CO 6
CD
3!
GX 4
1 3
I2
> 1
O










i 	
\
\
\
\
\
\
\
\
\
\
\
\
\
\
\
\









\
    0.3                3                 30                300

         Distance from the extraction well (m) \
Figure 2-12.  Example showing estimation of radius of influence of the
SVE/BV system from steady-state pressure data.
                              2-25

-------
           RI = radius of influence (cm)
           H  = depth of well screen (cm)
           Pw = absolute pressure at well (g/cm-s2)
           patm  = air pressure (1.013x105 g/cm-s2)

 The calculation of permeability for this example is shown in Example Calculation 2-2.
        In this example the following values are used:
                 Q  =1.4x104cm3/s
                 H  =2.7 ft. = 81 cm
                 Pw = -80 in.H2O = 0.816 x 106 g/cm-s2
                 Rw = 2.54 cm
                 RI  =457 cm                                     ;
        The value of the air permeability calculated from those data is 1.2 x 10~7 cm2.

    Example Calculation 2-2.  Estimation of soil permeability from steady-state
                            pressure measurements.
Measurement of Background SF6 Concentrations
In most cases, background concentrations of SF6 will be essentially zero.,  However, it is
important to make that determination prior to starting any test.  These measurements
can be made while the extraction system is in continuous operation. If previous tracer
tests have been conducted at the site, initial concentrations will probably be non-zero.
If concentrations are decreasing with time (i.e., on the tail of the previous test), then if
possible, conditions should be allowed to stabilize prior to initiation of the next test. If it
is not practical to wait additional time prior to initiating the test, the volume of injected
tracer can be increased. However, this may necessitate diluting samples: prior to
analysis.

Measurement of Oxygen and Carbon Dioxide Concentrations
Measurements of oxygen and carbon dioxide are very important in that they provide
insight to areas where air is or is not flowing.  Since SVE/BV tracer tests will often be
conducted in conjunction with in  situ respirometry tests, information on oxygen demand
and carbon dioxide production may also be available.  The combination of concentra-
tion and respirometry measurements provides a good measure of the extent to which
flushing is occurring. If biodegradation is the primary remediation pathway, these tests
may be sufficient and the tracer tests described here may not be necessary. If addition-
al information on flushing is desired, tracer tests in conjunction with the oxygen and car-
bon dioxide data can provide a good picture of air flow.
                                     2-26

-------
Preparation of SF6 for Injection
The first step in the preparation of SF6 for injection is the estimation of the volume of
pure SF6 to be injected. The Autotrac detector has a working range of 0.01 ppbV to
50 ppbV.  A good target concentration is 1 ppbV.  It is desirable to predict the extent to
which dilution will occur so that the correct volume of SF6 will be injected. This will be
most easily accomplished once several tracer tests have been conducted at the site.
Initially a simple airflow geometry can be used. An example of this is shown in Exam-
ple Calculation 2-3 and the geometry is shown in Figure 2-13.
  Distance between injection and extraction points                 5 m
  Thickness of the unsaturated zone                             5 m
  Estimated air-filled porosity                                    0.20
  SVE flow rate                                           2.8 me/min (~100 scfm)
  Target SF6 concentration in offgas                        10 ppbV (10"8 vol/vol)

  Air volume in cylinder of soil surrounding                        ~100 me
       the extraction well.  The radius of the
       cylinder is equal to the distance between
       the injection point and the extraction
       well (n*5*5*5* 0.25)

  Desired volume of pure SF6  (100 me * 10'8)                10'6 me = 1 mL

    Example Calculation 2-3.  Calculation of the volume of SF6 to be injected.
The second step is to determine the total volume of gas (air+SF6) to be injected. This is
necessary because when tracer tests are conducted it is important to introduce all of
the SF6 into the medium quickly. To do this, the volume of gas injected must be suffi-
ciently large to flush the bulk of the SF6 out of the well into the formation. At the same
time it is desirable to keep the injection volume as small as possible in order to estimate
the air flow rate in the stratum of interest. In general, the volume of SF6 necessary for
the test will be very small. In this case, the SF6 must be diluted with air to produce the
final injection volume.

The injection volume should be ~3 times the volume of the well. This insures that a
significant fraction of the SF6 gets out of the well point into which it is injected and di-
rectly to the subsurface. Injection volumes for conventional wells can become substan-
tial. Therefore, it is desirable to use smaller volume wells (e.g., vapor monitoring
points). An alternate approach for conventional wells is to pack off the well just above
the well screen and inject below the packing.  In this case, the injection volume is calcu-
lated based on the length of the well screen. Injection  of the tracer should be followed
                                      2-27

-------
                            n
             R
                                                          t
                                                         /i
                                                               h
Figure 2-13. Geometry for calculation of tracer dilution.
                             2-28

-------
 immediately by the injection of flushing air to move the remainder of the tracer out of
 the well.

 As discussed above, it is desirable to move all of the SF6 out of the injection point and
 into the formation. To accomplish this, a volume of air should be injected into the injec-
 tion point following the introduction of the tracer.  In general terms, this flushing volume
 should be ~3 times the volume of the well. Example Calculation 2-4 shows the volume
 of air+tracer to be injected.
  Total length of the well (I)        = 6 m
  Inside radius of the well (r)       = 0.019 m (0.75 inch pipe)
  Volume of the well (nr2!)         = 0.014 me
                                = 1.4L                            ;

  Injection volume (3X well volume) ~ 4.2 L

  Example Calculation 2-4.  Calculation of the volume of air+tracer to be injected.
In most cases, the injection volume will be on the order of 1 to 10 liters. In general, the
best way to introduce the tracer into the subsurface will be delivery from a pressurized
canister.  Dilutions can be prepared on a volume or pressure basis. For example, SF6
stock can be introduced into an evacuated canister to bring the pressure up to a prede-
termined value which represents the volume of SF6 to be injected. The canister can
then be pressurized up to some final positive pressure which reflects the total injection
volume.  It is important to remember that the injection volume will be determined by the
gauge pressure, but that some of the SF6 will remain in the canister at the conclusion of
the injection.  Therefore, the actual volume of SF6 injected must be adjusted according-
ly. An example of this is shown in Example Calculation 2-5.  Once the injection volume
has been prepared, prepare a second pressurized canister filled with air at approxi-
mately the same final pressure as the tracer injection sample for use in flushing the SF6
out of the well.
 Volume of SF6 placed in the canister                1.25 ml_
 Total injection volume                             4 L
 Volume of canister                                1 L
 Gauge pressure to produce the injection volume      4 atm
 Total air volume in the canister                     5 L [1 L * (4+1) atm]

 Volume of SF6 injected into the subsurface           1 ml_ [1.25 ml_ * 4 L / 5 L]

      Example Calculation 2-5. Calculation of actual SF6 injection volume.

                                     2-29

-------
 Introduction ofSF6 Into the Subsurface
 Once the preliminary data has been collected and the analytical instrument is calibrat-
 ed, the tracer test can be initiated. The analytical instrument should be set on automat-
 ic operation, and the initial SF6 concentration in the subsurface should be determined.
 Next, the pressurized canister should be attached to the injection point (The manner in
 which this connection is made will depend upon the construction of the injection point.)
 The valve on the injection volume canister should be opened and the time noted.  The
 pressure at the injection point should be monitored to insure that the canister has
 dropped to atmospheric pressure or below so that no tracer is lost to the atmosphere
 when the canister is removed. Immediately following tracer injection, the flushing vol-
 ume should be injected.

 Sample Collection (Figure 2-14)
 Samples should be collected prior to the extraction  pump to avoid dilution and other
 errors which may occur in the extraction pump. (Samples can be collected after the
 extraction pump if the system is correctly calibrated; however, that procedure will not be
 discussed here.)  The pressure at this point will be below atmospheric, so care must be
 taken to insure that a good sample is collected.  In general, samples can  be taken using
 a good quality diaphragm pump or metal bellows pump, or manually by syringe. (Care
 should be taken to insure that the pump does not leak and introduce dilution air.) Pres-
 sures below 0.5 atm require extreme care to insure that a good sample is collected.

 In high-vacuum situations, the capacity of the pump in the SF6 detector may exceed the
 capacity of the sample pump. This problem can be addressed either by getting a larger
 sample  pump or by adding an exhaust line (e.g., a 30 m length of 0.06 m  tubing onto
 the flow system; see Figure 2-14). Because the SF6 detector samples intermittently, the
 exhaust line will be continuously flushed with extraction air, and therefore a good sam-
 ple can  be drawn from the line when needed by the detector. If necessary, samples
 can be injected manually into the Autotrac via a syringe port on the front of the instru-
 ment. A 10-mL gas sample is required. Manual sampling is useful if ana|yzing samples
 from vapor monitoring points, but is not recommended for monitoring breakthrough un-
 less sample dilution is required.

 Sample Dilution
As described above, the LAT SF6 detector has a useful  range of 0.01 to 50 ppbV. If
 breakthrough occurs more rapidly than expected, SF6 concentrations may exceed
 50 ppbV.  In that case it will be necessary to dilute the sample into the useful range of
the instrument.  A number of procedures are available for this, and  depend upon the
extent of the dilution.  However, it  is important to realize that any dilution process in-
creases the chances for errors in the measurements. Therefore, whenever possible the
tests should be conducted such that dilutions are not necessary.
                                     2-30

-------
FROM
EXTRACTION
WELL
                           KNOCK-OUT
                           DRUM
                                                       VACUUM
                                                       RELIEF
                                                       VALVE
AIR
FILTER
                                              EXHAUST LINE
                           TO
                           OFF GAS
                           TREATMENT
EXTRACTION
BLOWER
    Figure 2-14.  Schematic drawing of the sample collection apparatus.
                                    2-31

-------
The procedures for sample dilution depend upon the magnitude of the dilution. Dilu-
tions of 2x to 10x can often be accomplished within the syringe in which the sample is
collected. For example, a 5x dilution can be prepared by drawing 2 ml_ of sample into a
10-miL syringe, and the remainder of the syringe filled with SF6-free air.  For dilutions
greater than 10x, it is generally necessary to inject a known volume of sample into a
larger vessel filled with SF6-free air. The vessels can be canisters, Tedlar bags/or
other containers. Once the dilution is complete, a 10-mL sample can be analyzed by
injection into the septum port on the SF6 detector.

Determination of Concentration
The Autotrac automatically reports concentrations in ppbV (or pptV, parts per trillion by
volume).  The concentration data can be written to a DOS-compatible disk for direct
transfer to a spreadsheet or other data management system.           i

Data Analysis
Calculation of Recovery Efficiency
The SF6 volume removal rate can be calculated from the concentration data by multiply-
ing the concentration by the  air flow rate for the SVE/BV system and by the sampling
interval. The individual measurements can be plotted as a function of time to obtain the
breakthrough curve, and the measurements can be integrated to determine the cumula-
tive recovery of SF6 (Figure 2-15). If sampling times are  regular, integration  of the
breakthrough curve can be made by simply summing the volumes collected  during each
sampling interval.

The SF6 recovery efficiency can be calculated by dividing the total SF6 recovery by the
volume of SF6 injected (Example Calculation 2-6).
 Injection:                2 L of 1000 ppmV SF6
 Equivalent volume of SF6: 2 x 10~3 L
 Extraction rate:           2830 L/min
 Sample interval:          10min                                  -
 To calculate the volume of SF6 removed at each interval multiply:

       [ppbV] x 10'9 x 2830 x 10 = [ppbV] x 2.83 x 10"5

 To calculate the total volume of SF6 removed, sum the volumes for each interval.
 The fraction recovered is the sum divided by 2 x 10~3 L
 For the example in Figure 2-11, the total volume was 1.7 x 10"3 L, and the fraction
       recovered was 0.85.

   Example Calculation 2-6. Calculation of fraction of injected mass recovered.
                                     2-32

-------
        100
         80
      I
         60
      .9  40
      a  20

      § 0.8
I  °-4
;p

%  0.2
a


1   °o
               50   100   150   200   250   300


                    Time (minutes)
Figure 2-15.  a) Tracer breakthrough curve; b) Fraction

of volume recovered during the tracer test.
                       2-33

-------
 Calculation of Breakthrough Time
 Breakthrough curves may be highly nonsymmetrical. Significant tailing may result due
 to mass transfer limitations during the movement of the tracer towards the extraction
 well.  An example of this was shown in Figure 2-6.  Tracer injection 1 shows consider-
 able tailing due to mass transfer limitations out of the lower permeability zone into which
 it was injected.  For this reason, the breakthrough time should be chosen as the time at
 which 50 percent of the recovered mass has been removed from the subsurface.

 Calculation of Effective Porosity
 Effective porosity can be estimated from the tracer test using a simple geometry. As
 discussed in Section 3, this estimate provides a useful evaluation of air flow. The steps
 in calculating effective porosity are shown in Example Calculation 2-7.  For most soils,
 the air-filled porosity will be between 0.05 and 0.25.  If the calculated value is outside of
 that range, it is likely that significant heterogeneity exists in the soil or that leakage from
 the ground surface is significant.

 Estimation of Velocity Near the Injection Point
 As described in Example Calculation 2-8, if the flow field is assumed to be radial,
 velocity will increase moving  towards the extraction well. The flow  rate, the distance
 between the injection and extraction wells, and the effective porosity can be used to
 approximate the velocity near the injection point (see Theory):
                                 V  =
                                      2nRhn
(2-3)
Comparison of Velocity Distributions of Several Tests
Comparison of velocity profiles of several tracer tests can provide insight into spatial
variations in air velocity in the subsurface. In most cases this analysis will be qualita-
tive, in part because the number and locations of the tests will be limited.  However, It is
certainly possible to use the tracer data, combined with other site information, to eval-
uate air flows using numerical models.

Field Examples
Tracer Tests in the OGI Large Layered Experimental Aquifer
Air flow tracer tests were conducted in a large (9mx9mx3m deep) physical model at
the Oregon Graduate Institute.  The model consists of layers of sand, gravel, and clay
as shown  in Figure 2-16. Air was extracted only from the sand at the bottom of the
model. Travel times in minutes are listed in Figure 2-16b. As can be seen from the
data, the highest velocities were in the gravel layer above the screened sand interval.
The data in Figure 2-16b also indicate that there is significant leakage through ground
surface. This occurred despite the presence of a PVC barrier. Numerical modeling
indicated that approximately half of the total flow through the model was leaking
through ground surface.

                                      2-34                         !

-------
a)
                         D
                                                 B
12
11
10
9
8
/
b
5
4
3
2
1

•vC ; v" v*
«. ' x ,-
x"-1 s ''•''•-f.
x '^ s *"^
x
x
x
V ~ X *
''. x . '


X
X
"x?^
'X
X
X
X

x -,
"x*

"-% ,, Sx''5
X
x

X
X
X
•^ x
^ ^y ,/ 1»X_


f^

N





r--B
/.- x
~^ ; 	 ~ — x—
LICA FLOOR >x
-<> x
^J x
X<- ' fXa
X
X
X
/ "* , x
El^lfolsilTE^ 5
-*• "^ x <•
	 ^ 	 ^ — 3c 	
' '^ V!T'
" SAND x "**--*
x
^ -X ^ ,^-
X
GRAVEL x
x
' «. X
VSA'ND^ '*x-

	 TL 	 \
^ x ' ••
' ' X ','},

X ' ^
~ >4
x
x
x
x
x -


                                                                2.2m
                              •8.4m
b)
x 85
y
x 62
x
X
x22
x
x 138, 216
x
X 67
x
X
X
X
x 15
x
X
X
x
X
x22
X
X
x
X
X
x
X
X
x
x 22
X
X
x 9
x 7
x
X
x 112
x 282, 372
x
x 23
x
X
x 6
x 3
x
x
x
x 210
X
x17
' x13
x1O
x 7
x 4
x
x
x
x
	 	 	 	 	 — 	 , — \ 	 1
                       Q = 50 scfm
 Figure 2-16. Travel times from various locations within the OGI large
 physical model.
                             2-35

-------
  Assume, as in Figure 2-11, that the breakthrough time was 105 min and that the fol-
  lowing conditions were present at the site:

        Distance from injection to extraction point:     5 m
        Thickness of unsaturated zone:               5 m
        Depth of injection point:                      4m

  If a cylindrical geometry is assumed, the breakthrough time should be the time re-
  quired to sweep the air from a 5  m radius around the extraction well.

  Volume of a cylinder with a 5-m radius:
        5*5*5*n* ~ 400 me

  The volume of air extracted in 105 minutes was:

        105 min * 2830 L/min /1000 L/me - 300 me

  The effective porosity is then estimated to be 300/400 or 0.75. This value is unrealis-
  tically high. The most likely explanation for the high effective porosity isithat the
  injection point lies within a region where air flow is low. Since the injection point is
  near the bottom of the unsaturated zone,  it is possible that air is leaking from the sur-
  face and short-circuiting to the extraction well.
        Example Calculation 2-7.  Calculation of effective porosity using a
                                simple geometry.
 Flow rate (me/min)            = 2.8
 Radius of cylinder (R) (m)     = 5
 Height of cylinder (h) (m)      =5
 Air-filled  porosity             = 0.25

 V = 2.8/[2*n*5*5*0.25]
   = 0.071 m/min
   Example Calculation 2-8.  Estimation of velocity at the tracer injection point.
References for Section 2

Hinchee, R. E., S. K. Ong, R. N. Miller, D. C. Downey, and R. Frandt. 1992. Test Plan
and Technical Protocol for a Field Treatability Test for Bioventing. Prepared for the U.S.
Air Force Center for Environmental Excellence. Revision 2, May 1992.

                                      2-36

-------
                                   Section 3

          Procedures for Conducting Tracer Tests to Evaluate Recovery
                    of Injected Air During In Situ Air Sparging
Introduction
Introduction to In Situ Air Sparging (IAS)
IAS is a groundwater remediation technique in which air is injected directly into a water-
saturated medium to remove contaminants by volatilization and to enhance aerobic
degradation.  IAS is used both to remediate aqueous groundwater plumes and to treat
sources which contain nonaqueous-phase liquids (NAPLs).

The setup of an IAS remediation system is shown schematically in Figure 3-1.  It gener-
ally consists of one or more air injection wells and one or more SVE wells. As  men-
tioned above, the primary purposes of the injection well(s) are to volatilize contaminants
and to increase aerobic biodegradation by introducing additional oxygen into the
groundwater.  These wells are usually designed in a manner similar to groundwater
monitoring wells, except that they generally have short screens (i.e., 1 to 2 ft) and are
screened entirely below the water table.

The principal purpose of the extraction wells is to prevent the off-site migration of va-
pors volatilized by the IAS system. Generally, the setup of the extraction wells is similar
to conventional soil vapor extraction (SVE) systems. This often will include an  air blow-
er, a "knockout" drum for removing liquids, and an off-gas treatment system (Fig-
ure 3-1).

The equipment required for the IAS portion of the system is minimal.  In addition to the
injection well, all that is generally required is a compressor capable of delivering air at
the desired flow rate at a pressure governed by the depth of injection. It is also de-
sirable to be able to measure and control air flow and pressure at the injection well.

Introduction to Air Recovery Tests
Air recovery tests are an important means of evaluating the performance of SVE sys-
tems for capturing air injected below the water table as part of an IAS remediation
system.  The recovery tests are important because they provide direct evidence of the
extent to which injected air may be moving off site. Off-site migration is potentially
                                      3-1

-------
                 Air
                 Compressor
                         Air
                         In
       VADOSEZONE
              NAPLr
.Air  //    X
f Out/ / Off-gas treatment

   ^"^Extraction blower
       ^Knockout drum

    Extraction well
     GROUNDWATER ZONE
Figure 3-1. Schematic configuration of an IAS/SVE remediation system.
                                3-2

-------
important because it is a means by which potentially hazardous concentrations of
contaminants can be carried to adjacent properties.

Test Objectives
General Comments
In order to prevent off-site migration of vapors during IAS, combined IAS/SVE systems
are often designed in such a way that extracted air flow exceeds air injection by some
multiplicative factor (e.g., 5X).  In addition, to demonstrate that the design is working,
soil gas vacuum surveys in the vicinity of the IAS/SVE system are usually conducted.  It
is generally concluded that if no pressures greater than ambient are observed, all of the
IAS air is being captured by the SVE system.  However, it is generally difficult to relate
vacuum data to recovery of IAS air.  This is the case because numerous potential air
flow patterns in the groundwater zone can exist.  For example, if IAS air is injected into
sand below a continuous clay layer, the air may move laterally beyond the radius of
influence of the SVE well before it has the opportunity to reach the water table. In this
case, the sparge air might not be captured by the SVE system.

The previous example implies that under some circumstances pressure measurements
alone will not conclusively demonstrate that IAS air is being captured.  As a conse-
quence, it is important to conduct tests which can unambiguously determine if all of the
IAS air is being captured by the SVE system.

Primary Objective
The primary objective of helium recovery tracer tests described here is to unambiguous-
ly determine the recovery efficiency of air injected during IAS.

Theory
Underlying Principal
The principal underlying the helium recovery tests is simple. Helium is injected into the
subsurface at a known rate and the rate of helium recovery at the SVE is calculated
from the observed helium concentration in the SVE effluent and the SVE flow rate.

Practical Considerations
In order to successfully conduct a  helium tracer test, it is necessary to accurately mea-
sure flow rates and helium concentrations. As a result, calibration of the analytical
equipment (both flow meters and the helium detector) is extremely important.  It is also
very important to have a system which is free of leaks. This means not only the injec-
tion and extraction systems, but also the sampling and analysis systems.

Steps in Conducting a Helium Recovery Test
There are four steps in conducting the helium recovery test (Figure 3-2). They are:

     1.     Determination of the "100 percent recovery" concentration

                                      3-3

-------
                    Determination of the 100% recovery
                             concentration
                       Injection of the helium tracer
                   Measurement of the helium concentration
                           in the SVE off-gas
                   Plotting of percent helium recovered as
                           a function of time
Figure 3-2. Steps in conducting helium recovery tracer test.
                                3-4

-------
             Helium is injected at a known rate (the same rate used in the tracer test)
             directly into the extraction manifold prior to the helium detector.  The con-
             centration measured at the helium detector is the 100 percent recovery
             concentration.
      2.     Injection of the helium tracer
             Once the 100 percent recovery concentration is determined, helium injec-
             tion into the sparge air can be initiated. This injection rate must be the
             same as the rate used to determine the 100 percent recovery concentra-
             tion.                                                  ;
      3.     Measurement of the helium concentration in the SVE off-gas.
             Once helium injection in the sparge air has been initiated, air samples are
             collected from the extraction manifold at regular intervals until the helium
             concentration in the effluent stabilizes.                   ;
      4.     Plotting of percent helium recovered as a function of time.

             Observed helium concentrations divided by the 100 percent recovery con-
             centration times 100 are plotted as a function of time  since the initiation of
             helium injection. The final values represent the fraction of the injected
             helium which is recovered by the SVE system.

Test equipment
Overview of Experimental Setup
In order to simplify interpretation, the tests should be conducted by injection of helium
into a single IAS well and recovery from a single SVE well.  In nearly all cases, tracer
tests will be conducted in conjunction with vapor extraction and injection operations.  In
that context, the design and installation of the extraction/injection wells will be dictated
by the remediation design. As a consequence, design and installation of the extrac-
tion/injection wells are not discussed here.

Calibration of Analytical Equipment
Calibration of the Helium Detector
Helium in the extracted air will be measured with a Mark Products helium detector Mod-
el 9822 or equivalent with a minimum sensitivity of 100 ppm (0.01 percent).  Calibration
of the helium detector is made using the experimental setup shown in Figure 3-3.

The helium detector should be turned on and equilibrated for at least 10 minutes prior
to conducting a calibration or obtaining measurements. As part of the calibration
process, the internal sampling pump of the helium detector should be checked prior to
operation to ensure that it is functioning.
                                      3-5

-------
        VENT
   HELIUM
   DETECTOR
                       PRESSURIZED
                       SOURCE OF
                       HELIUM AT
                       A KNOWN
                       CONCENTRATION
Figure 3-3. Schematic drawing of the equipment
used to calibrate the helium detector.
                      3-6

-------
 The helium detector should be calibrated each day using helium calibration standards
 in air.  These standards should be pressurized cylinders of 10, 1, 0.1, and 0.01 percent
 helium in air. The instrument is calibrated by connecting it to one of the pressurized
 standards and adjusting the flow from the cylinder such that some flow comes out of the
 vent line. Flow should continue until a stable reading is achieved on the meter
 (-30 seconds).

 Once measurements have been made for each concentration, a calibration curve can
 be constructed.  If any measured value differs from the reported  standard value by
 greater than 20 percent, that standard should be reanalyzed. If the value fails to agree
 upon reanalysis, the source of the problem should be identified.

 Helium standards can be purchased from a specialty gas supplier or they can be pre-
 pared on a pressure or volume basis. The pressure-based approach will be discussed
 here. In general, standards should be prepared in canisters which can withstand 10 at-
 mospheres of pressure and which do not affect the quality of the standard. The final
 pressure of the standards described here will be 9 atm gauge pressure, which corre-
 sponds to a 10-fold dilution of the concentration of helium added to the canister. Prep-
 aration of standards can begin with canisters filled with helium-free air at a pressure
 equal to  atmospheric. Standards should be prepared using good-quality pressure
 gauges which are calibrated against a reference. Water or mercury manometers are
 excellent references. The canister should be connected to a helium source-either
 100 percent helium or a certified mixture (e.g., 1 percent He in air) and a Magnehelic™
 gauge as shown in Figure 3-4. Helium is allowed to flow into the canister until the pres-
 sure rises to a predetermined gauge pressure. Typical values are listed in Table 3-1.
 The canister can then be brought to a final pressure of 9 atmospheres.

 Table? 3-1.  Typical Pressure Values Used in Preparing Helium Standards.	
                                                       Final Pressure of Stock in
  Final  Concentration (%)     Stock Concentration (%)      Standard Canister (atm)

         90                          100                         9
         10                          100                         1
          0.9                         1                          9
          0.1                         1                          1
          0.01                        1                          0.1


 Calibration of the Air Flow Meters
 Flow rates for the SVE system will generally be in the 10- to 200-scfm range. Vacuum
 levels will be at 10 to 200 inches of water below ambient pressure.  The flow meter cal-
 ibration system is shown in Figure 3-5.  At these high flow rates, a large dry gas meter
will'be required.  If a large dry gas meter is not available, an alternate approach is to

                                      3-7

-------
                          PRESSURE
                            GAUGE
          STANDARD
           CANISTER
 a)
                     PRESSURIZED
                     SOURCE OF
                     HELIUM AT
                     A KNOWN
                     CONCENTRATION
                          PRESSURE
                            GAUGE
          STANDARD
           CANISTER
 b)
AIR
Figure 3-4. Schematic drawing of the equipment
used to prepare helium standards.
                         3-8

-------
                       VACUUM
                        GAUGE
VACUUM
PUMP
                       VALVE
   DRY GAS
    METER
                                     FLOW
                                     METER
Figure 3-5.  Schematic drawing of the flow meter calibration system.
                                   3-9

-------
use another calibrated flow meter to calibrate the one to be used at the site. Actual ver-
sus observed flow rates should be determined over the range of the flow meter at sev-
eral vacuums between 0 and 0.9 atmosphere. Those data should be plotted as a family
of curves with each line corresponding to a different vacuum value.

Flow rates for the IAS system will generally be less than those used for the extraction
system. However, pressures will be above atmospheric,  rather than below. The
experimental apparatus used to calibrate flow meters at above atmospheric pressure is
shown in Figure 3-6.

Flow rates for tracer injection will be in the range of 0.1 to 1 L/min. The experimental
apparatus for calibrating the flow meter for tracer injection is the same as for air
injection (Figure 3-6).

Calibration of the Sampling Pump
Under many operating conditions the SVE manifold will be under sufficient vacuum that
automated analytical equipment will not be able to draw an adequate sample from the
manifold. In these cases it will be necessary to use a good quality vacuum pump to
draw samples from the manifold and deliver them to the automated analytical equip-
ment. In the context of tracer tests, two potential problems arise with respect to the
vacuum pump. First, the pump must be able to move sufficient volumes of air to meet
the needs of the analytical equipment,  and second the pump should not leak air which
can provide additional dilution of the sample stream. The procedures below describe
how pump performance can be measured.

Prior to selecting a sampling pump, check the specifications of any automated sampling
equipment to be used to determine the volumes of air required by each. (The Mark
Products helium detector requires ~100 mL/min.) The first step is to connect the pump
to be tested to the apparatus shown in Figure 3-7.  If two  dry test meters  are not avail-
able, two calibrated flow meters of the appropriate ranges can be used. Then turn on
the pump and open the valve so that no vacuum is observed on the gauge. Determine
the flow into and out of the pump by recording the volume of flow that occurs in 1 min-
ute on each of the dry test meters or the flow rates on the flow meters. Partially close
the valve to produce a vacuum of 5 inches of  mercury,  and determine the flow rates into
and out of the pump.

Repeat the previous procedure with vacuums of 10, 15, 20 and 25 inches of mercury.
Prepare a plot of flow rate in and out vs. vacuum. Based  on those data, determine the
maximum vacuum that provides sufficient flow for the helium detector. Next, determine
the sampling pump leak ratio as a function of vacuum.  Determine if the leakage of the
sampling pump is acceptable (e.g., inflow rate is within a factor of two of outflow rate).
                                     3-10

-------
                    VACUUM
                     GAUGE
   DRY GAS
    METER
                                            COMPRESSOR
                                FLOW
                                METER
Figure 3-6. Schematic drawing for flow meter calibration under positive
pressure.
                              3-11

-------
              VACUUM
               GAUGE
                         SAMPLE
                          PUMP
  DRY GAS
   METER
DRY GAS
 METER
Figure 3-7. Schematic drawing showing setup for measurement of
sample pump flow rate vs.vacuum and leakage vs. vacuum.
                          3-12

-------
Test Procedures
Qvewiew of Experimental Procedures
Experimental activities can be divided into the following components. Each is described
briefly in the following sections.

            1.     Determination of the "100 percent recovery" concentration
            2.     Injection of helium into an IAS point
            3.     Collection of samples from SVE off-gas
            4.     Determination of recovery rate (percent) of helium

Determination of "100 Percent Recovery" Concentration
It is necessary to determine the concentration of helium in the off-gas which represents
the concentration at "100 percent recovery" of helium.  To do this, helium is injected into
the extraction manifold prior to the sample pump at a rate which is the same as will be
used for the recovery test. The apparatus for this is shown in Figure 3-8. The steps
involved in determining the "100 percent recovery" concentration are:

      1.    Estimate the SVE flow for preliminary calculations (e.g., use the  flow
            meter reading).

      2.    Calculate the inflow rate of 100 percent helium to produce 1 percent con-
            centration in the effluent (See Example Calculation 3-1).    ;

      3.    Install a good vacuum pump (metal bellows or diaphragm) to the manifold.
            (This will be the same setup as for the tracer tests.)   Make sure the
            pump has adequate flow and does not leak at the system pressure.

      4.    Connect the helium source to the manifold near the extraction point and
            add  helium at the prescribed rate using a calibrated flow meter.

      5.    Monitor tracer concentration  in the extraction system until it stabilizes.
            (This should take only a few minutes.)  This value represents the 100 per-
            cent recovery" concentration.                           ;

Vacuum Survey
It is important to collect subsurface vacuum data prior to initiation of the tracer tests.
These data provide insight into the general nature of the flow system. For example, if
little or no vacuum is recorded at a monitoring point,  it can be expected that there is
little flow at that point. Large vacuums may indicate areas of active flow; however,
these values can also occur within low-flow regions adjacent to  higher flow regions.
Nevertheless, these data can frequently be helpful in understanding the general nature
of air flow at the site..
                                     3-13

-------
                                                                TO
                                                                OFF GAS
                                                                TREATMENT
                        KNOCK-OUT
                        DRUM
 FROM
 EXTRACTION
 WELL
EXTRACTION
BLOWER
Figure 3-8. Schematic drawing showing the system for measuring SVE flow
using a tracer gas.
                                3-14

-------
 Determination of tracer injection rate
       Approximate SVE Flow rate       = 35 scfm = 1000 L/min
       Tracer concentration             = 100%
       Desired final concentration        = 1.0%
       need a dilution of 102
       tracer flow rate = 1000 L/min /102  =10 L/min
 Calculation of "100 percent recovery" concentration
 To determine the concentration which corresponds to 100 percent recovery, pure
 helium is injected into the extraction manifold at the same rate (e.g., 10 L/min) that
 will be used during the tracer test.  The helium concentration observed under these
 conditions is considered to be the value which corresponds to 100 percent recover/.

     Example Calculation 3-1. Calculation of actual SVE flow using a tracer.
The general approach will be to measure soil vacuum with Magnehelic™ gauges.  The
same measurements can be made with a manometer or other calibrated vacuum
gauge. For most sites it will be necessary to have gauges in the following ranges (in
inches of water): 0 to 1 inch, 0 to 10 inches, and 0 to 100 inches.

When the remediation system has been  operating for more than one day, determine the
soil vacuum at each point in the system by connecting the appropriate gauge to the
point. After connection to the monitoring point, sufficient time should be allowed for the
vacuum to stabilize (commonly 1 minute).

Measurement of Background Helium Concentrations
In most cases, background concentrations of helium will be essentially zero. However,
it is important to make that determination prior to starting any test. These measure-
ments can be made while the extraction  system is in continuous operation. If previous
tracer tests have been conducted at the  site, initial concentrations may be non-zero. If
concentrations are decreasing with time  (i.e., on the tail of the previous test), then if
possible, conditions should be allowed to stabilize prior to initiation of the next test. If it
is not practical to wait for stabilization prior to initiating the test, the volume of injected
tracer can be increased. However, helium concentrations in  the influent air should be
kept below 5 percent.

Estimation of the Rate of Pure Helium to be Injected
A volume fraction of helium in the effluent stream in the range of 0.002 to 0.01  (0.2 to
1 percent) is desired. To estimate the rate of helium injection necessary,to produce this
concentration, some initial estimate of SVE air flow must be made.  The input rate for
helium is simply the approximate SVE air flow rate times the  target volume fraction.  If
the IAS rate is low (e.g., <20 percent of the SVE rate), the target effluent volume

                                     3-15

-------
fraction should be kept at the bottom end of the range to avoid buoyancy effects in the
injection air (i.e., helium concentrations in the influent air should be kept below ~5 per-
cent).

Introduction of Helium into the Subsurface
Once the preliminary data have been collected and the analytical instrument is calibrat-
ed, the tracer test can  be initiated.  The IAS/SVE system should have been in operation
for a period of several  days prior to initiation of the tracer test. The first step is to start
the analytical instrument and determine the initial  helium concentration in the subsur-
face.  If these concentrations are adequately low,  the helium source can be connected
to the I AS well and the test initiated.

Sample Collection
Samples should be collected prior to the extraction pump to avoid dilution and other
errors which may occur in the extraction pump (Figure 3-9). (Samples can be collected
after the extraction pump if the system is correctly calibrated; however, that procedure
will not be discussed here.)  The pressure at this point will be below atmospheric, so
care must be taken to  insure that a good sample is collected.  In general, samples can
be taken from the extraction manifold using a good quality diaphragm pump or metal
bellows pump, or manually by syringe.  (Once again, care should be taken to insure
that the pump does not leak and introduce  dilution air.) Pressures below 0.5 atm re-
quire extreme care to insure that a good sample is collected. In high-vacuum situa-
tions, the capacity of the pump on the helium detector may exceed the capacity of the
sample pump. This problem must be addressed by using a sampling pump with ade-
quate capacity.

Data Analysis
Calculation of Recovery Efficiency at a Particular Point in Time
Recovery efficiency is  simply calculated  as the ratio of the observed concentrations to
the "100 percent recovery" concentration determined at the beginning of the test (Ex-
ample Calculation 3-2)
 SVE flow rate                   = 1000 L/min (-35 scfm)
 Injection rate of pure helium      =10 L/min
 Expected concentration          =1% by volume
 Observed concentration          = 0.65%
 Recovery efficiency             = .65/1*100 = 65%

         Example Calculation 3-2. Calculation of expected concentration
                  	and recovery efficiency.	;
                                     3-16

-------
                                                                   TO
                                                                   OFF GAS
                                                                   TREATMENT
                          KNOCK-OUT
                          DRUM
 FROM
 EXTRACTION
 WELL
    EXTRACTION
    BLOWER
Figure 3-9. Schematic drawing showing the setup for sampl
during the tracer recovery test.
e collection
                                    3-17

-------
Time-Series Analysis of Recovery Data
In most cases helium will begin to be recovered within an hour of the initiation of tracer
injection. Helium concentrations can be expected to rise rapidly initially and then to
asymptotically approach some final value. It may be necessary to continue the test for
a period of 24 hours or more to establish the final value of recovery efficiency.

Interpretation of the Recovery Efficiency Data
Recovery efficiencies of less than 100 percent imply that some of the IAS air is escap-
ing the SVE system.  The significance of the lost air will depend upon the potential risks
posed by off-site migration  of the sparge air. There is, of course, some uncertainty in
the measurement of recovery efficiency. That uncertainty stems from uncertainty in
flow measurements (injected helium, extracted air) and measured  helium concentra-
tions. In this context, recoveries of greater than 80 percent probably indicate adequate
recovery, and efficiencies of less than 50 percent generally indicate incomplete recov-
ery.

Field Examples
IAS In an OGI LEAP Tank
An air recovery test was conducted at one of the OGI large experimental aquifers.
During the test the  air injection rate was 3.5 scfm and the SVE extraction rate was 30
scfm, The depth to water was ~1 m and the surface was covered with  a PVC barrier.
Helium was injected into the sparge air at ~1.5  L/min. The data in Figure! 3-10  indicate
that the helium recovery rate climbed to 100 percent of the injection rate in a period of
~2 hours.

References for Section 3

Hinchee, R. E., S. K. Ong,  R. N. Miller, D. C. Downey, and R. Frandt. 1992. Test Plan
and Technical Protocol for  a Field Treatability Test for Bioventing.  Prepared for the
U.S.  Air Force Center for Environmental Excellence. Revision 2, May 1992.
                                     3-18

-------
 O
    120
    100
     80
    60
 o

I  40
    20
      -50
0              50
     Time (minutes)
100
150
 Figure 3-10.  Recovery of helium during an air recovery test in an OGI large
 experimental aquifer.
                                  3-19

-------
                                    Section 4

               Procedures for Conducting Tracer Tests to Evaluate
            the Distribution of Injected Air During In Situ Air Sparging
Introduction
Introduction to In Situ Air Sparging (IAS)
IAS is a groundwater remediation technique in which air is injected directly into a water-
saturated medium to remove contaminants by volatilization and to enhance aerobic
degradation.  IAS is used both to remediate aqueous groundwater plumes and to treat
sources that contain NAPLs.

The setup of an IAS remediation system is shown schematically in Figure 4-1. It gener-
ally consists of one or more air injection wells and one or more soil vapor extraction
wells. As mentioned above, the primary purposes of the injection well(s) are to volatil-
ize contaminants and to increase aerobic biodegradation by introducing additional oxy-
gen into the groundwater.  IAS wells are usually designed in a manner similar to
groundwater monitoring wells, except that they generally have short screens (i.e., 1 to
2 ft) and are screened entirely below the water table.

The principal purpose  of the extraction wells is to prevent the off-site migration of va-
pors volatilized by the  IAS system.  Generally, the setup of the extraction wells is similar
to conventional SVE systems. This often will include an air blower, a "knockout" drum
for removing liquids, and an off-gas treatment system (Figure 4-1).

The equipment required for the IAS portion of the system is minimal. In addition to the
injection well, all that is generally required is a compressor capable of delivering oil-free
air at the desired flow rate at a pressure governed by the depth of injection. It is also
desirable to be able to measure and control air flow and pressure at the injection well.

Introduction to IAS Air Flow Tracer Tests
IAS air flow tests are conducted by injecting a gas-phase tracer, such  as SF6, along
with the IAS air and determining the distribution of tracer in the subsurface by collecting
water samples from discrete locations and depths and  determining the concentration of
the tracer in the water.  In the approach described below, tracer can be injected for a
period of 1  week,  followed by groundwater sampling in the vicinity of the IAS well.
                                      4-1

-------
                 Air
                 Compressor
                        Airi
                        In  »
       VADOSEZONE
              NAPL-
 , ~ir  / /
 f Out/ / Off-gas treatment

>   \. ^^Extraction blower
>\     ^-^
                                         Knockout drum
     Extraction well
     GROUNDWATER ZONE
Figure 4-1. Schematic configuration of an IAS/SVE remediation system.
                                4-2

-------
Introduction to Vertical Groundwater Profiling
Vertical groundwater profiling (VGP) is a technique that allows water samples to be col-
lected at a number of discrete depths  in the subsurface.  It is generally accomplished by
driving a small (e.g., 1-inch) -diameter pipe into the ground. The leading edge of the
pipe usually consists of a drive point followed by a screened interval through which
water can be drawn. The pipe assembly can be advanced by hammering, vibrating, or
pushing.

Water samples can be drawn to the surface using  a variety of devices.  If the water ta-
ble is within the suction limit, water can be drawn to the surface through a tube connect-
ed to a peristaltic pump. If the water table is deeper, a small-diameter bailer or bladder
pump may be used. Vertical profiles are generally made at a number of locations and
distances around the IAS well to create a three-dimensional picture of the air distribu-
tion.

Test Objectives
General Comments
Air pathways produced by IAS are highly erratic. As a consequence, it  is difficult to de-
fine the "radius of influence" using conventionally measured parameters (e.g., dissolved
oxygen in wells, water level changes). Tracer tests and vertical profiling during IAS pro-
vide a means of not only characterizing the radius over which the air is moving, but also
the vertical distribution of the air.  The latter is important because for the IAS process to
be effective at remediating zones of residual NAPL contamination,  there must be good
contact between the contaminated zones and the  sparge air.

The IAS air distribution test described below should be applicable to porous media sites
where the permeability is greater than 0.001 cm/s  (e.g., fine sand or coarser). At per-
meabilities below this range it will be difficult to withdraw water from the subsurface
using the small-diameter driven sampler. In this case, core samples may be appro-
priate for characterizing the air distribution.

Primary Objective
The primary objective of tracer tests described here is to characterize the distribution of
air pathways below the water table at  IAS sites.

Theory
Underlying Principal
The principal underlying the IAS air distribution tests is that as the air moves through
the groundwater zone, some of the tracer introduced with the sparge air will partition
from the air to the groundwater during the sparging process.  For the water in imme-
diate contact with the sparge air, tracer concentrations will rise to or near saturation
values with respect to the tracer input concentration.  An injection period of 1 week is
adequate to give a representative picture of air flow patterns, but short enough to

                                      4-3

-------
minimize advective transport of the tracer in the groundwater.  In areas not in direct
contact with the sparge air, tracers can arrive by diffusion or groundwater. advection and
concentrations will generally be significantly lower.

Practical Considerations
In order to successfully conduct an IAS air distribution test it is necessary to be able to
collect groundwater samples at discrete depths below the water table. The ground-
water samples must be collected without headspace or volatilization losses during sam-
pling. This can generally be accomplished  by vertical profiling and careful groundwater
extraction.

A variety of tracers can be used to conduct this test. SF6 is the tracer used in the
procedure described below.  SF6 is a gas at room temperature and pressure; it has a
modest solubility (40 mg/L) and a high dimensionless Henry's gas constant (-150). Its
primary advantage is that it can be detected at very low concentrations in air and water.
These properties are similar to those of oxygen.  Consequently, the distribution of SF:6
tracer can be seen as an analog for oxygen distribution.

Steps in Conducting an IAS air distribution test
The five steps in conducting the IAS air distribution test (Figure 4-2) are as follows:

       1.     Injection of the tracer
       2.     Determination of the tracer injection concentration
       3.     Groundwater sample collection
       4.     Analysis of SF6 in groundwater samples
       5.     Mapping of the percent saturation of SF6 in plan and profile views

Test Equipment
Overview of Experimental Setup
In order to simplify interpretation, the tests should be conducted by injection of SF6 into
a single IAS well and recovery from a single SVE well. In nearly all  cases, tracer tests
will be conducted in conjunction with vapor extraction and injection operations. The
design and installation of the extraction/injection wells will be dictated by  the remedia-
tion design. Therefore, design and installation of the extraction/injection wells will not
be discussed here.

The tracer is added to the IAS air from a pressurized cylinder attached to.the injection
manifold. The rates of air and SF6 injection determine the concentration  in the IAS air
and the concentrations which will be observed at 100 percent saturation in the ground-
water.

The SF6 concentrations in the groundwater are determined by gas chromatography us-
ing an electron capture detector.  A wide variety of gas chromatographs,  ranging from

                                      4-4                         ;

-------
                        Injection of the SF6 tracer
                    Determination of the tracer injection
                             concentration
                     Groundwater sample collection
                  Analysis of SF6 in groundwater samples
                   Determination of the tracer injection
                             concentration
Figure 4-2.  Steps involved in am IAS air flow tracer test.
                               4-5

-------
sophisticated research instruments to more "user friendly" instruments, are commercial-
ly available. For this application an instrument that is robust enough and portable
enough for field use is desirable. In addition, it should have a very low detection limit
and an automatic data acquisition system. An  SF6-specific gas chromatograph, avail-
able from Lagus Applied Technologies (LAT) in San Diego, California, is one instrument
that satisfies all of the above criteria.  It is automated and has a detection limit of ~10
parts per trillion (0.01 ppbV) by volume. In the following discussion it will be assumed
that a LAT Autotrac is being used.

Calibration of Analytical Equipment
SF6 detector
Because the LAT Autotrac is equipped with an internal cylinder of calibration gas,
single-point calibrations can be easily accomplished in the field.  However, it is desira-
ble to conduct multipoint calibrations to verify linear instrument response.  Standards
can be obtained commercially, prepared by dilution, or prepared using a dynamic cali-
bration instrument.  For field applications such as the one described here, pressurized
canisters over a range of concentrations provide a convenient means of multipoint
calibration.

Air flow meters
All flow meters to be used in the tests should be calibrated with the appropriate gases
over the appropriate pressure ranges.  For the  IAS system, pressures will be above
atmospheric.  The flow meter calibration system for systems above atmospheric pres-
sure Is shown  in Figure 4-3. Actual versus observed flow rates should be determined
over the range of the flow meter at several pressures between 1 and 5 atmospheres.
Those data should be plotted as a family of curves with each line corresponding to a
different pressure value.

Test Procedures
Overview of Experimental Procedures
Experimental activities can be divided into the following components:

      1.    Injection of the tracer
      2.    Determination of the tracer injection concentration
      3.    Groundwater sample collection
      4.    Analysis of SF6 in groundwater samples

Each is described briefly in the following subsections.

Injection of the Tracer
The setup for injection of the tracer is shown schematically in Figure 4-4.  The tracer is
added to the sparge air between the  compressor and the point at which the air enters
the subsurface. Because the air injection line is at a positive pressure relative to the

                                      4-6

-------
                    VACUUM
                     GAUGE
   DRY GAS
   METER
                                            COMPRESSOR
                                FLOW
                                METER
Figure 4-3. Schematic drawing for flow meter calibration under positive
pressure.
                              4-7

-------
    AIR
    COMPRESSOR
  JV
      TRACER
      SOURCE
                       PRESSURE
                       GAUGE
                                              BACK PRESSURE
                                                    VALVE
FLOW
METER
                                                          SPARGE
                                                          WELL
Figure 4-4. Schematic drawing of the tracer gas injection system.
                                   4-8

-------
atmosphere, the tracer must be injected at a greater pressure. To ensure a stable flow
of tracer, it is recommended that the cylinder valve and backpressure valve be adjusted
such that the desired flow is achieved at a pressure of ~50 psi.  In general, this will be
sufficiently above the air injection pressure that the tracer flow will remain constant,
despite changes in air flow conditions.

To produce the desired SF6 concentration in the sparge air, the following calculation
can be used: if the IAS air flow is 150 L/min (~5 scfm), an SF6 flow rate of 45 mL/min
will produce an input concentration of 3 x 105 ppbV.

Determination of the Tracer Injection Concentration
As described in the previous subsection, SF6 is injected at a known rate directly into the
IAS manifold.  To determine the SF6 input concentration, an air sample is collected from
the manifold after the SF6 injection point and the tracer concentration in the air sample
is determined. For the example described here, an injection concentration of ~3 x
105 ppbV was used. Therefore, the samples must be diluted approximately 10,000-fold
to get them in the range of the LAT detector. This can be easily accomplished, for ex-
ample, by filling a Tedlar bag with 10 liters of SF6-free air and injecting 1 ml_ of the IAS
air into the bag (~2 minutes should be allowed  for the air to mix in the bag).

Grotindwater Sample Collection
The collection of good-quality groundwater samples is key to the success of this tracer
test.  The sample collection technique must be capable of collecting samples from dis-
crete depths and delivering those samples to a storage vessel (e.g., a 40-mL vial) with-
out volatilization loss.  As described previously in this section ("Introduction to  IAS Air
Flow Tracer Tests"), a variety of methods can be used to accomplish this. Most involve
advancing a pipe with  a screened tip using a percussion hammer or vibration.

If the water table is less than ~25 feet, it is generally possible to use suction to draw wa-
ter through the pipe to the surface using a peristaltic pump.  If a steady stream of water
can be produced (e.g., no air bubbles in the sampling line), the water flowing from the
pump can be delivered to a sample bottle for storage.  To ensure that a good sample is
collected, the tube from the peristaltic pump should be placed in the bottom of the sam-
ple bottle and the water delivered to the vial in a "gentle" manner (i.e., no splashing,
etc.).  If samples are collected in 40-mL vials, a rate that fills the bottle in ~10 seconds
is appropriate. If 40-mL vials are used, they should be overfilled with ~100 mL of water,
and then the tube should be slowly removed from the vial.  The cap should be placed
on the vial such that no headspace remains in the vial. The samples should then be
stored inverted in the dark until the time of analysis, which should be within 2 weeks.

If a steady supply of water cannot be drawn to the surface with vacuum,a variety of
alternative approaches are available to accomplish this. Small-diameter bailers are the
simplest means of accomplishing this, but small-diameter bladder pumps;are the best

                                      4-9

-------
means of delivering a high-quality water sample. Another approach is to-use evacuated
vials that can be lowered down the pipe and opened at depth (e.g., BAT™ samplers,
Hogentogler and Co., Inc, Columbia, Maryland).

Analysis ofSF6 in Groundwater Samples
In order for profiling to provide an accurate picture of air distribution in the groundwater
zone, an accurate measure of concentration is required. This can be accomplished us-
ing a variety of analytical approaches on a gas chromatograph with an electron capture
detector (e.g., headspace, direct aqueous injection). An LAT detector will be used for
discussion purposes in the following example. Although it has excellent sensitivity for
SF6, the LAT detector requires a 10-mL air injection, which somewhat complicates
sample preparation.

Because SF6 is analogous to oxygen, it is useful to report concentrations as a percent
of saturation. In that context it is useful to report values that range from 100  percent of
saturation with respect to the input concentration down to ~1 percent.  The first step in
calculating the percent saturation is to measure the  aqueous concentration of the tracer
in the water sample, and then convert that value to a percent of saturation based on the
input air concentration.

The easiest way to measure aqueous concentrations of SF6 using the LAT detector Is
by headspace analysis. This requires that conditions be adjusted to provide  a head-
space concentration within the range of the LAT. The following example outlines this
method.

If an SF6 input concentration of 3 x 105 ppbV is used, based on a solubility of 40 mg/L
(a dirnensionless Henry's constant of 150), concentrations in the groundwater could
reach ~1.2 x 10~5 g/L. If a headspace of equal volume to the water is created by
removing half of the water, essentially all of the SF6  (>99 percent) will partition to the
headspace.  As shown in Example Calculation 4-1, this will produce a headspace con-
centration of ~2000 ppbV, which is greater than the  maximum concentration for the LAT
detector. A maximum  headspace concentration on the order of 20 ppbV is desired. To
accomplish this, a headspace to water ratio of ~100 should be used.  This is  achieved
by injecting 0.4  mL of water sample into a 40-mL vial that had been previously flushed
with SF6-free air. The water and air should be allowed to equilibrate for 1 to 2 minutes
before an air sample is withdrawn.

As mentioned above, the LAT detector requires that -10 mL of air be injected into the
sample loop.  This is accomplished by withdrawing the air through the septum cap us-
ing a 10-mL syringe. However, using a syringe to withdraw that volume from a 40-mL
vial will cause a significant reduction in the internal pressure of the vial. When the sy-
ringe is exposed to the atmosphere, ambient air will be drawn into the syringe.  If the
ambient air contains SF6, this will lead to errors in the analysis.  To prevent this, 10 nnL

                                     4-10

-------
of clean air should be injected into the vial as the sample is withdrawn. If this is done
carefully (e.g., one needle tip at the top of the vial, one at the bottom), dilution of the
sample by the injected air can be avoided.
  SF6 input concentration                  = 3 x 10s ppbV
                                        = 3x 10'4 mole fraction
  SF6 concentration in water
        at equilibrium with tracer           = 0.04 g/L* 3x10"4
                                        = 1.2x10-5g/L
  Concentration in air at equilibrium
        with an equal volume of water
        saturated with respect to the
        injection air                      =1.2x10"5 g/L / 6 g/L
                                        = 2 x 10'6 mole fraction
                                        = 2000 ppbV
  Concentration in air at equilibrium
        with water saturated with
        respect to the injection air
        at an airwater ratio of 100:1        = 1.5 x 10"5 g/L / 6 g/L /100
                                        = 2 x 10~8 mole fraction
                                        = 20ppbV

           Example Calculation 4-1. Calculation of headspace concentration.
Once the concentration of SF6 in the headspace is determined, the concentration in the
aqueous phase can be determined by calculating the total mass in the headspace and
dividing that number by the volume of water in the vial.  The aqueous concentration can
then be expressed as a percent of the saturation value (Example Calculation 4-2).
 Measured headspace concentration        = 5 ppbV
                                        = 6g/L*5x10-9
                                        = 3x10-8g/L
 Volume of air in the vial                  = 40 mL
 Total mass in headspace                 = 0.04 L * 3 x 10"8 g/L           :
                                        = 1.2x10-9g
 Volume of water in the vial                = 0.0004 L
 Aqueous concentration                   = 3 x 10"6 g/L
 Saturation concentration
        (fraction SF6 in IAS air * solubility)   = 3 x 10'4 * 0.04 g/L
                                        = 1.2x10'5g/L
 Percent saturation                       = 3 x 10'6 g/L /1.2 x 10'5 g/L * 100
                                        = 25%                        ;

              Example Calculation 4-2. Calculation of percent saturation.


                                       4-11

-------
 Data Analysis
 Once percent saturation in groundwater numbers have been calculated, it is useful to
 construct plots of concentration versus distance below the water table for the profile
 locations. If, for example, several profiles were made in a vertical plane through the
 sparge well, these data could be plotted together to show the distribution of air as a
 function of distance from the sparge point.

 Alternatively, profile data from the site can be plotted in plan view with individual plots
 located approximately where the profiles were taken. [An example of this is shown later
 in this section (Field Example).] If data on hydrocarbon concentrations as a function of
 depth are available, it is useful to plot the SF6 and hydrocarbon data together to assess
 the extent of contact between the two.

 The data collected and presented as described above can be used to examine the
 areal and vertical distribution of the air from an IAS well.  The test can also be used to
 assess the contact between the IAS air and the zone of contamination. It is assumed in
 this discussion that the zone of contamination includes residual NAPL at or below the
 water table.   In this case, good contact between the air and the NAPL is important for
 cleanup to be accomplished within a reasonable timeframe.  If the vertical distribution of
 contaminants is known, the test described here can provide a good measure of the con-
 tact between the NAPL and the air. If vertical profiles have been made at a number of
 locations at the site, the test can also provide a good indication of the area over which
 the IAS well is effective.  If the vertical distribution of the contaminants is not known, the
 test can still  provide useful information about the vertical and areal distribution of the
 sparge air; however, it may be difficult to assess the effectiveness of the IAS well for
 remediating  the site.

 One of the most commonly measured parameters with  regard to an IAS well is the
 "radius of influence" (ROI) of that well. Most measures of ROI (e.g., groundwater
 mounding, vadose zone pressure) produce a picture that is areally much more uniform
 than field data (e.g., Field Example) suggest is the case at many sites.  The test
 described here provides a much more accurate picture of the zone over which IAS is
 active, in both the vertical and areal directions.                        ;

 The example described above uses a single IAS well. The test can also be applied at
 sites where multiple sparge wells are in operation.  In the latter case,  the same profiling
 and analysis procedures can be used. However, it  may be desirable to increase the
 number of profile locations in order to adequately describe the distribution of air at the
 site.

 Field Example
 Figure 4-5 shows a plan view of the sparging test site at which an IAS air flow tracer
test was conducted. The subsurface at the site consisted of fine to coarse sands over

                                     4-12

-------
       xP8

   xB8                 xPB



                xB7
                                                 xB9
                                                     xP9
xB6

   xP6
                                xB5
                                _           xB15
                                SPARGE
                                 POINT
                            xB3     xBl
                                                            xBlO


                                                         xPIO


               xB14





                                   5 meters
Figure 4-5.  Plan view of sparging test site showing the
locations of the monitoring wells (B-series) and the vertical
profile locations (P-series).
                               4-13

-------
 the entire treatment zone. The water table was at a depth of ~6 meters and the sparge
 air was injected at a depth of -9.5 m.

 In Figure 4-6, the SF6 data are superimposed on a plan view map of the site. As the
 figure shows, the distribution of SF6 around the sparge well is not uniform, either areally
 or vertically. These data, combined with observations during sampling, have significant
 implications for IAS at the site.

 During sampling it was  observed that in the vicinity of Wells B-6, B-9, and B-10, there
 appeared to be a lower permeability zone from the water table to a depth of approxi-
 mately 7 m. The samples immediately below that depth produced a significant amount
 of air with the water.  These samples also showed significant SF6  concentrations. This
 suggests that an areally extensive air pocket had formed beneath the lower-permeabili-
 ty layer and that  upward air flow was low.

 In Figure 4-7, SF6 data  from B-6 are plotted along with previous soil sample data. The
 data indicate that air flow  occurs below the depth of highest soil contamination. Based
 on the field observations,  it would appear that the high soil concentrations correspond
 with a lower-permeability zone above 7 m and that the high air flow region corresponds
 to a higher-permeability zone below that depth.  This suggests that air may be bypass-
 ing the most contaminated zones at the site.

 In the vicinity of Wells B-7 and B-8, little SF6 was observed in the water samples and
 IAS appears to be ineffective. It has been speculated that this is caused by anisotropic
 medium. Our observations suggest that the lower-permeability zone above 7 m was
 not as pronounced in this  area; as a consequence, lateral air movement under the layer
was Jess significant. SF6 and soils data obtained near Well B-7 (Figure 4-8) show a
 pattern similar to those  obtained near Well B-6.  In the case of Well B-7, the SF6
concentration is quite low, indicating that air flow is limited. However, the data do
suggest that air flow once again appears to bypass the zone of highest contamination.

Throughout the site there  appeared to be limited air distribution below a depth of 8 m.
Once again, our interpretation of these data is that the distribution of air was controlled
primarily by lower-permeability zones above ~7 to 8 meters which  covered portions of
the site.
                                     4-14

-------
OJ
-6
-6.5
-7
•7.5
-8
-15
-9
-95








P8





-5.5
-6
•6.5
-7
-7.5
•8
-8.5
•9
- - n c
" 0 20 40 60 80 100 -«
-6
45
O.J
-6
-6.5
-7
-7.5
-8
-8.5
.9
-9.5






P7





0 20 40 60 80 11
-V
-7.5
-8
•8.5

-9.5





PB




-5.5
*,
' -V
.3
,,
0 20 40 60 80 100 .9









P9
___
••^^™
L


'J 0 20 40 60 80 101







P6
^^^^^
^MM








0 20 40 60 80 100
0 :
                                A SPARGE
                                w POINT
                                                  -5.5
                                                   -6
                                                  -6.5
                                                   -7
                                                  -7.5
                                                   -8
                                                  -8.5
                                                   -9
        0 20 40 60 80 100

- PIO
0 20 40 60 80 101
Figure 4-6.  Site plan view showing SF6vertical profiles.
                              4-15

-------
    -1
     2
  I"
   § -4
     -6
   §

     -8
     -9
    -10
              5       10      15

                   Thousands

                 [C5-CH] (mg/kg)
20        0     20     40     60     80     100

                 SF6 (% Saturation)
Figure 4-7.  SF6 and vertical distribution of contaminants near monitoring well B6.
                                      4-16

-------
    -1
ts>
urface (m
d»
  1
m gro
O>
  8
  s
i-7


 -8


 -9


-10

            5         10
                Thousands
              [C5-CU] (mg/kg)
                                 15
20  0      5      10     ;15
           SF6 (% Saturation)
                                                                          20
Figure 4-8.  SF6 and vertical distribution of contaminants near monitoring well B7.
                                      4-17

-------
                                   Section 5

         Procedures for Bioventing Field System Design and Evaluation
                 Air Flow, Tracer, and In Situ Respiration Tests
Introduction/Purpose
Conventional soil vapor extraction (SVE) systems are designed to optimize system
performance by maximizing airflow rates and air/contaminant contact to yield a maxi-
mum recovery rate of volatiles from contaminated soil. Performance may deteriorate
over time, however, due to occluded residual saturation and enrichment of residual
contamination in the less volatile waste components.

Bioventing is a modification of the conventional, gas based soil remediation technology
which has been successfully applied and documented for the remediation of hydrocar-
bon contaminated soils either used alone or for the "polishing" of residual, semi-volatile
contaminants remaining in soil following high rate SVE. Bioventing entails the use of
SVE systems for the transport of oxygen to the subsurface, where indigenous orga-
nisms are stimulated to aerobically metabolize contaminants located there. Bioventing
systems are designed and configured to optimize oxygen transfer and oxygen utilization
efficiency, and are operated at much lower flow rates and with significantly different
configurations from than those of conventional SVE systems.           '

This procedure has been developed to provide an  integrated approach for the eval-
uation of air flow/air permeability (characteristics of gas phase remediation systems
common to both SVE and bioventing systems), along with biodegradation rates (char-
acteristic of the biologically based bioventing technology) quantified from respiration
measurements collected under field conditions for  use in the design and evaluation of
field-scale in situ bioventing systems. Both air flow data, relating to oxygen supply, and
biodegradation rate data, relating to oxygen utilization, are required for trie rational de-
sign of bioventing systems, and both types of data can be collected in the field proce-
dure described below.
Background/Theory
Biological Remediation of Contaminated Soils
The biodegradation of organic compounds in soil environments has been extensively
described in the technical literature, and details of metabolic pathways and microbial

                                      5-1

-------
populations responsible for compound biotransformation have been summarized in a
large number of textbooks and reviews on soil microbial ecology (Alexander, 1977;
Atlas, 1981; Dragun, 1988).  For direct biodegradation of hazardous organics to be
successful, four conditions must be satisfied. First, the contaminants of interest must
serve as a carbon and energy source for the indigenous microbial population, i.e., they
must be able to serve as an electron donor.  Secondly, an appropriate electron accep-
tor must be available so that energy can be extracted from the electron donors at envi-
ronmentally significant rates. Thirdly, macro- and micronutrients essentia) for the
production of cellular material must be available in the appropriate ratio for microbial
growth to proceed unhindered (C:N:P mass  ratios typically recommended for soil bio-
remediation applications are 100:10:1).  Finally, environmental conditions within the
contaminated soil environment must not be inhibitory to the indigenous microflora soil
environmental conditions to ensure effective bioremediation include:  soil water at 50 to
80 percent of soil field capacity ~ 1/3  bar; soil pH from 5.5 to 8.5; soil temperature in
the rnesophilic range from 15 to 45°C; and an absence of organic or inorganic toxicants
that can inhibit microbial activity.

The most critical limitation to successful bioremediation is generally the lack of appro-
priate electron acceptors.  A variety of electron acceptors can be used by soil micro-
organisms to carry out the oxidation of organic contaminants.  These include oxygen,
nitrate, sulfate,  iron, manganese, carbon dioxide and organic carbon.  Of these, oxygen
provides the organism with the highest energy yield, providing nearly twice that of ni-
trate, and an order of magnitude higher energy release to the microorganism when it is
utilized as an electron acceptor as compared to sulfate, carbon dioxide and organic
carbon. Oxygen metabolism is therefore energetically selected for, and subsequently,
oxygen utilizing microorganisms are ubiquitous in soil environments. Oxygen is also
the preferred electron acceptor from an engineering standpoint, as accelerated degra-
dation rates generally occur under aerobic (oxygen rich) conditions as compared to
anoxic or anaerobic (oxygen deficient) conditions.

These principles of biodegradation  have historically been applied to the in situ aerobic
bioremediation of contaminated soils and ground water using  water to carry oxygen to
subsurface contamination.  Efforts have been made to increase the level of oxygen in
this water by saturating the water with pure oxygen or through the addition of hydrogen
peroxide. These efforts have generally met with limited success, however, because of
the inability to transfer adequate oxygen to areas of subsurface contamination.  This
oxygen transfer is caused by the physical limitations to the transfer of this oxygen
saturated liquid through contaminated soils  (Downey et al., 1988;  Hinchee and
Downey, 1988; Hinchee et al., 1989; Lee et al., 1988;  Wetzel et al., 1987).

The inherent disadvantage of utilizing water as the carrier medium for the transfer of
oxygen to the subsurface can be graphically illustrated by determining the mass of wa-
ter required to transport a unit mass of oxygen when the carrier medium is saturated

                                       5-2

-------
with oxygen.  These values are summarized in Table 5-1 and show that due to the low
solubility of oxygen in water, prohibitively large amounts of oxygen-saturated water are
required, even when using pure oxygen or hydrogen peroxide saturated solutions. This
oxygen supply limitation is exacerbated by the high oxygen demand of hydrocarbon
contaminants, as indicated by the simple stoichiometric reactions for hexane oxidation
shown below, assuming no substrate incorporation into cell material:

                        C6H14 + 9.5 O2 - 6CO2 + 7H2O          \         (5-1)
                                3.5 g O2/g C6H14                   ;

Table 5-1.  Carrier Fluid Oxygen Supply Requirements.	

                                                        Ib carrier/lb oxygen

 Water
                                      Air Saturated            125,000
                             Pure Oxygen Saturated            25,000
             Hydrogen Peroxide Saturated (500 mg/L)            10,000

 Air (20.9% Oxygen)	4.8
Assuming an oxygen requirement of only 3 g O2/g hydrocarbon for hydrocarbon miner-
alization, a 3,785 L (1.,000 gal) fuel spill weighing approximately 3,175 kg (7,000 Ib)
would require 9,525 kg (21,000 Ib) of oxygen. This equates to an air-saturated water
volume of approximately 1,191,000,000 L (315,000,000 gal), a pure O2 saturated vol-
ume of 238,000,000 L (63,000,000 gal), or a saturated peroxide solution Volume of
95,300,000 L (25,200,000 gal) to provide the required oxygen for fuel bioremediation.  It
becomes apparent from these calculations that hydraulic limitations would be severe for
the remediation of a spill even as small as 3,785 L (1,000 gal) due to massive water
volumes required when using liquid carrier bioremediation approaches in either the
saturated or the unsaturated zone.

Bioventing  Technology Description
Bioventing describes the process in which the air medium is utilized to deliver oxygen to
the subsurface to stimulate the in situ biodegradation of organic contaminants. As indi-
cated in Table 5-1, air is an extremely efficient oxygen transfer medium due to its high
oxygen content (20.9 vol. percent, i.e., 209,000 ppmV) and low viscosity as compared
to that  of saturated water. Bioventing represents a hybrid physical/biological process
utilizing SVE systems for oxygen transfer, while focusing not on  contaminant stripping,
but rather on in situ aerobic contaminant biodegradation for the remediation of a con-
taminated site.
                                      5-3

-------
 Consideration of SVE for oxygen transfer to the subsurface was proposed in 1988 by
 Wilson and Ward (1988), who noted that systems designed for the removal of volatiles
 from soil could also be used to transport oxygen. A number of other authors have
 discussed the potential improvement of in situ, aerobic, subsurface bioremediation
 using SVE for oxygen transfer (Bennedsen, 1987;  Riser, 1988; Ely, and Heffner, 1988;
 Ostendorfand Kampbell, 1989; Stapps, 1989), and there has been ample recent
 evidence demonstrating field-scale bioventing system effectiveness for fuel con-
 taminated site remediation (Dupont etal., 1991; Miller etal., 1991; Hincheeetal., 1991;
 Ong et. al., 1994; van Eyk,  1994).

 Bioventing systems are composed of hardware identical to that of conventional SVE
 systems, with vertical wells and/or lateral trenches, piping networks, and a blower or
 vacuum pump for gas extraction. They differ significantly from conventional systems,
 however, in their configuration and philosophy of design and operation.  As indicated
 above, the primary purpose of a bioventing system is to use moving soil gas to transfer
 oxygen to the subsurface where indigenous organisms can  utilize it as an electron
 acceptor to carry out aerobic metabolism of soil contaminants. As such, bioventing
 system extraction wells are not placed in the center of the contamination as in
 conventional SVE systems (Figure 5-1), but on the periphery of the site (Figure 5-2),
 where low flow rates [4.6 to 23 actual L/s (10 to 50 acfm) versus 46 to 700+ actual L/s
 (100 to 1,500+ acfm)] for conventional SVE systems) maximize the residence time of
 vent gas in the soil to enhance in situ biodegradation and minimize contaminant
 volatilization.

 Because it is a biological treatment approach, however, bioventing does require the
 management of environmental conditions to ensure maintenance of bioactivity at the
 site. Management of soil moisture and soil nutrient levels to avoid inhibition of microbial
 respiration within the vadose zone can be accomplished fairly easily,  and have been
 used to optimize contaminant biodegradation at field sites when other variables, e.g.,
toxicity, do not limit microbial activity (Dupont et al., 1991; Miller et al., 1991).

 Oxygen transfer to the subsurface via SVE systems is generally more rapid than oxy-
gen uptake rates observed under field conditions (Dupont et al., 1991; Hinchee et al.,
 1991). This results in the oxygenation of soil gas to near ambient levels if vent system
blowers are operated on a continuous basis.  To minimize system operating costs, and
more importantly to reduce or even perhaps eliminate off-gas treatment requirements
entirely, cyclic or "surge" pumping of vent systems in bioventing operations is recom-
mended.  Surge pumping in a bioventing mode entails operating the blower system until
soil gas oxygen levels reach near ambient conditions throughout the site being reme-
diated. The system is then shut off for some period of time,  during which soil gas
                                     5-4

-------
  By-Pass
                                    By-Pass
                                       11
Knock-Out
 Drum   Blower
                 Off-Gas
                Treatment
 Figure 5-1. Typical SVE system with labeled system components.
 By-Pass  t
                                        Krock.-Out
                                                         Off-Gas
                                                       Treatment
Figure 5-2. Typical bioventing system schematic. Note extraction and
injection from the periphery of contamination.
                                 5-5

-------
 oxygen concentrations would be routinely monitored until they reach a level which inhib-
 its aerobic microbial activity. Once this limiting soil gas concentration is reached, the
 vent system would be restarted, and the on-off cycle would continue once again.
 Based on a Henry's Law constant for oxygen,  a limitation would be expected to occur at
 a soil gas concentration of approximately 2.0 vol. percent, corresponding to soil water
 oxygen concentrations of approximately 1 mg/L. An inhibition of soil respiration has
 been reported at the 2.0 vol. percent soil oxygen level in venting systems treating JP-4
 contaminated soils (Dupont et al., 1991) and in vented soil piles contaminated with PCP
 waste (McGinnis et al., 1994), suggesting that this value represents a good operating
 number for field scale applications.

 Based on observed field respiration data from  various JP-4 jet fuel contaminated sites
 (Dupont et al., 1991; Hinchee and Ong, 1992;  Ong et al., 1994) and bioventing of PCP
 contaminated soil piles (McGinnis et al., 1994), field oxygen uptake rates of 0.03 to
 1.4 vol. percent /hour (0.8 to 39.7 g O2/m3 soil-d @ air filled porosity = 40 vol. percent)
 can be expected. These rates can be nearly an order of magnitude lower as reme-
 diation progresses to near de minimis soil hydrocarbon levels (Dupont et al., 1991),
 allowing typical bioventing systems to be operated on schedules of 8 hours on,
 16 hours off at the initiation of remediation, to 8 hours on, 7 days off near the end of the
 field effort, while still maintaining aerobic conditions within the contaminated soil during
 non-venting periods. Table 5-2 presents a summary of general design, operational and
 application considerations appropriate for conventional  SVE systems versus those uti-
 lized in a bioventing operating mode.

 Bioventing System Design
 Despite the large number of bioventing systems being implemented in both the public
 and private sectors for hydrocarbon contaminated soil remediation, there is currently a
 lack of quantitative design recommendations for their application and performance eval-
 uation. The U.S. Air Force has been a leader in the development and implementation
 of bioventing  systems for remediation  of many of their fuel release sites, and they have
 developed, through the Air Force Center for  Environmental Excellence (AFCEE), a field
treatability procedure for bioventing system design (Hinchee et al.,  1992).  In addition,
 an addendum to this procedure document detailing the integration of soil gas survey
 results into bioventing system evaluation was published by AFCEE in 1994 (Downey
and Hall, 1994). These AFCEE field bioventing procedure documents were written as a
guide for the field scale evaluation of the potential application of bioventing for remedia-
tion of Air Force sites, and focus heavily on field methods for in situ respiration/deg-
radation rate determinations using procedures  adapted from Hinchee and Ong (1992).
While the importance of air flow and permeability evaluation is clearly stated, little detail
is provided regarding field-scale permeability determinations nor is information provided
on the use  of tracer tests for improved site assessment and system design.
                                      5-6

-------
Table 5-2. General Design and Application Considerations Appropriate for
Conventional Versus Biovention SVE Systems.	
           Parameter
          Conventional SVE
       Bioventing
 Compound type
 Vapor pressure
 HC (dimensionless)
 Aqueous solubility
 Soil concentration
 Depth to groundwater
 Air-phase permeability
 Subsurface conditions
 NAPL phase
 General vent well placement
 General injection well placement
 Operating mode

 Flow  rate

 Pore  volumes/d
 Optimal soil moisture
 Nutrient requirement
 Soil gas O2 levels
 Toxicants
Volatile at room temperature
>100mmHg
>0.01
<100 mg/L
>1 mg/kg
>20 ft
>1 xKT'cm/s
Little or no stratification
Little or none
Within contamination
Outside contamination
Maximum soil gas exchange rate

46 to 700+ actual L/s
(100to1500+acfm)
1 to 15
«25% field capacity
NA
NA
NA
 Biodegradable
 NAa
 NA
 NA
NA
Biodegradable
Outside contamination
Within contamination
Maximum retention time
and aerobic conditions
4.6 to 23 actual L/s
(10to50acfm)
0.1 to 0.5
=75% field capacity
C:N:P= 100:10:1b
>2 vol.%
Little or none
 NA = Not available.
 Caution should be used in considering a nutrient requirement as field
 shown mixed results in performance with nutrient addition. This ratio
 requirement that may or may not be needed at a given site.
                            -scale; bioventing'research has
                             represents a maximum theoretical
                                          5-7

-------
The procedure described in this document relies heavily on the methodology as de-
scribed in the AFCEE procedure, but attempts to improve on it by more completely
integrating field bioventing  system evaluation and design from the initial site char-
acterization activities through the system design, process monitoring and performance
evaluation steps.  Schematically this improved bioventing system test procedure is
summarized in Figure 5-3,  describing the objectives, activities and outcome/interpre-
tation of each of its five phases.

Bioventing System Test Procedure for Treatability Assessment, Design, Process
Monitoring and Performance Evaluation

Phase I - Assessment of the Potential for Contaminant Biodegradation Under
Field Conditions                                                 ',
This phase is the first step  in the evaluation of the potential application of bioventing, or
more  generally, any biologically based remediation system at a given field site under
existing site conditions. To determine the potential for in situ biodegradation of vadose
zone contaminants via bioventing, existing soil microbial activity should be quantified
during site assessment investigations. This can be readily accomplished through the
analysis of soil gas O2 and CO2 composition, in addition to the more routinely measured
total hydrocarbon  concentrations, prior to venting activity at the site. Total petroleum
hydrocarbon as well as O2  and CO2 concentrations can be measured during standard
soil gas surveys using a variety of measurement techniques.  While both respiration
gases can be easily measured, oxygen concentrations are considered a better indicator
of microbial activity in soil systems because there are rarely abiotic sinks for oxygen in
these environments.  CO2 is produced through anaerobic as well as aerobic microbial
activity and can also be affected by  assimilation or dissolution of carbonate rock.

The key to the evaluation of soil bioactivity using these methods is the determination of
the extent of O2 depletion and CO2 enrichment in soil gas at a site with respect to back-
ground, uncontaminated soil levels.  It cannot be overemphasized that these determina-
tions must be based on a comparison to uncontaminated soil conditions, as only levels
of O2 depletion and CO2 enrichment in excess of background are indicative of increased
microbial activity compared to normal, basal respiration levels seen in uncontaminated
soils at the site. It is also important  to note that despite the collection of hydrocarbon
concentration data during initial site investigations conducted  using soil gas surveys,
respiration gas (O2/CO2) measurements are rarely made, even though they can be
made using the same soil gas probes, and virtually at the same time as hydrocarbon
measurements are taken.  These respiration gas readings are unequivocal indicators of
microbial activity at the site under actual field conditions,  and as indicated in Figure 5-3,
are critical in evaluating the next step in bioventing feasibility assessment at a  given
site.
                                      5-8

-------
         OBJECTIVE
                                        ACTIVITY
                                          OUTCOME
                                         Interpretation
 Phase 1


    Assessment of the
      Potential for the
     Biodegradation of
    Site Contaminants
        Under Field
        Conditions
 Phase II
     Assessment of Air
      Fiov* Rates and
      Oxygen Transfer
      Possible Under
      Field  Conditions
  Soil Gas Survey
     In Situ Air
Permeability/Tracer
       Test
                                                                 Depressed Oxygen/Elovated
                                                                 Carbon Dioxide Compared to
                                                                      Background Soils
                                                               Biological Activity Evident. Further
                                                               Evaluation of Bioventing System
                                                                  Warranted.  Go To Phase II

                                                                   Oxygen/Elevated Carbon
                                                                     Dioxide Compared to
                                                                 Background Soils Not Evident
                                                               Biological Activity Inhibited Due to
                                                                Site Limitations/Toxicity, Further
                                                               Evaluation of Bioventing System
                                                                        Not Warranted
                                                                Adequate Flow Rates. Little to
                                                                 No Short-Clrcuiting Evident
                                                                 Acceptable Air Flow/Oxygen
                                                                Transfer Potential. Bioventing
                                                                System Warranted.  Go To Full
                                                                        Scale Design

                                                                    Lov* Flow Rates. Short-
                                                                       circuiting Evident
                                                                 Low Air FlovwfOxygen Transfer
                                                                  Potential. Bioventing System
                                                                         Not Warranted
     Assessment of the
       Biodegradation
         Rate of Site
       Contaminants
         Under Field
         Conditions
 In Situ Respiration
        Test
                                                               Respiration Rates Corrected for
                                                                Background Rates 2 Published
                                                                            Results
                                                                   Biological Activity Evident,
                                                                 Bioventing System vVarranted.
                                                                   Go To Full Scale Design
                                                                Respiration Rates Corrected for
                                                                 Background Rates <: Published
                                                                             Results
                                                                   Biologicai Activity Minimal,
                                                                    Inhibition or Recalcitrant
                                                                 Contaminant Residual Evident.
                                                                     Bioventing System-Not
                                                                           Warranted
     Phase III
         Full Scale
    Bioventing System
          Design
    Assessment of
    Oxygen Uptake
       Rate Law
                                                                             Zero Order Rate
                                                                          Assess Desirability of
                                                                    Continuous versus Pulse Venting
                                                                      Operation in Full Scale Design
                                                                             First Order Rate
                                                                      Use Continuous Pufse Venting
                                                                      Operation in Full Scale Design
  Phase IV
       Full Scale
  Bioventing Syster
    •  Monitoring &
     Performance
      Evaluation
  Quarterly Oxygen
    Uptake Rate
  Determinations
                                                               Respiration Rates Declining. But
                                                                     =» Background Rates
                                                                Contaminant Mass Removal is
                                                              - Proceeding but Contamination
                                                                  Remains. Assess Operating
                                                                Rate/Mode Changes. Continue
                                                                           Bioventing

                                                                Respiration Rates £ Background
                                                                              Rates
                                                                  Site Remediation  is Indicated.
                                                                Conduct Confirmatory Soil Boring
                                                                   for Performance  Verification
  Phase V
  Verification of
Performance via
Confirmatory Soil
     Borings
                                    Ouantify Residual
                                       Contaminant
                                      Levels via Soil
                                    	Borings	
                                                                    Soil Concentrations s: Clean-Up
                                                                                 Goals  '•
                                                                              Site Closure
                                Soil Concentrations => Clean-Up
                                             Goals
                                  Assess Operating Rate/Mode
                                 Changes. Continue Bioventing
Rgure 5-3.  Field bioventing system treatability assessment, design, and
performance evaluation procedure.
                                                 5-9

-------
 If soil gas organic vapor and soil core data show contamination, but microbial respira-
 tion has not yielded O2 uptake and CO2 production concentrations above background
 levels, conditions within the contaminated soil have resulted in soil microbial toxicity
 and/or severe inhibition, or significant nutrient or moisture limitations exist at the site.
 Unless soil moisture is the cause of this limitation, bioremediation has limited applica-
 tion, and alternative remediation schemes should be considered.

 If soil contamination exists and microbial activity above background levels is evident
 from soil gas measurements, quantification of maximum respiration rates under field
 conditions can be carried out utilizing in situ respiration measurement techniques
 described below. The reader should  refer to the addendum to the AFCEE bioventing
 test procedure by Downey and Hall (1994) for additional examples  of soil gas data inter-
 pretation related to the feasibility of the application of bioventing for site remediation.

 Phase II - Assessment of Air Flow and In Situ Respiration Rates Under Field
 Conditions
 Once biological activity has been verified at the site, quantitation of the rate of air/O2
 supply, as well as the rate of in situ O2 utilization must be determined. As described
 above, the remediation of most contaminated sites is limited by the supply of electron
 acceptor, namely O2, and rational engineering design of bioventing systems requires a
 focus on supplying the oxygen needed to meet the in situ O2 demand.

 Air Flow and Tracer Tests
 Airflow and tracer tests  are first conducted as described in the procedure section en-
 titled "Tracer Tests to Evaluate Air Flow During Soil Vapor Extraction and Bioventing"
 to provide data regarding the existence of short-circuit pathways, stagnant zones, arid
 general conditions of vapor flow and oxygen transport in the subsurface throughout the
 site. These data are essential as efficient oxygen supply to the subsurface is key to
 optimal bioventing system design. Once subsurface air velocities are estimated from
these air flow/tracer tests,  O2 transfer rates and transfer efficiencies can be estimated
for various points throughout the area of contamination.

As stressed in the air flow/tracer test procedure, data from multiple lateral and vertical
points throughout the contaminated soil should be collected to provide information
regarding the spatial distribution and heterogeneity of air flow and oxygen transfer
throughout a site. A minimum of three radial distances and three vertical locations  (a
minimum of nine total sampling points) should  be used to provide the air flow and per-
meability data necessary to assess gas transport conditions at a typical site.  Potential
O2 transfer rates can be  estimated knowing that 1 standard cubic foot/minute (scfm) of
air equals 0.21 scfm of O2 which, from the ideal gas law is equivalent to 7,700 mg
O2/min at 1 atm and 25°C. Table 5-3  provides an estimate of O2 transfer rates for vari-
ous soil types under a uniform SVE system operating condition of 30 in H2O for a 4-in.-
diameter extraction well, a radius of influence of 30 ft, and a well slotting of 10 ft,

                                      5-10

-------
 assuming simple one dimensional, radial flow into the well.  Table 5-3 indicates that
 even in clayey soils where operating flow rates are low, significant O2 transfer rates
 (275 mg Cymin = 396,000 mg O2/day) are possible in bioventing systems.

   Table 5-3. Potential Oxygen Transfer Rates in Various Soils.     	

       Soil Type      Air Flow Rate (scfm)     Oxygen Transfer Rate (mg O,/min)

     Medium sand          35.6                         275,000
     F:inesand              3.56                         27,500
     Siltysand              0.36                          2,750
     Clayey sand	    0.04                            275
 These potential O2 transfer rates as calculated from the air flow/tracer data and subse-
 quent air velocity/flow rate determinations, are representative of actual field conditions,
 and are directly indicative of system performance that can be.expected under full-scale
 conditions.  The final data necessary in this phase of the bioventing treatability assess-
 ment is for the in situ oxygen demand, or in situ oxygen respiration rate produced by
 the site microbial population in the degradation of contaminants found there.

 In Situ Respiration Tests
 In situ respiration tests are conducted following the air flow/tracer tests to quantify the
 rate of oxygen demand expressed under actual field limiting conditions.  In order to ex-
 pedite and minimize efforts and cost of bioventing treatability assessment and design,
 the in situ respiration tests are recommended to be coupled to the air/flow tracer test
 efforts.  Using this approach, in situ respiration tests would be initiated following com-
 pletion of the air permeability/tracer tests when the entire area of influence of the per-
 meability/tracer test extraction well is oxygenated.  In this way, all tracer injection and
 soil gas monitoring points installed as part of the air permeability/tracer test can be uti-
 lized for soil respiration  rate determinations.  Use of these identical monitoring points
 provides air flow and  respiration data that correspond directly to one another and lat-
 erally and vertically distributed data necessary to ascertain the spatial variability and
 distribution of microbial  activity throughout the site. Finally, the background, uncontam-
 inated site location used as a baseline for the soil gas survey should also be incorpo-
 rated into the in situ respiration test effort. This allows for a quantitative determination
 of the significance of measured respiration rates within the contaminated area with
 respect to background oxygen uptake rates in uncontaminated soil.  It also requires that
the background point be oxygenated separately via air injection for a 16- to 24-hour
 period if it does not fall within the area of influence of the air permeability/tracer test
extraction well.

With the entire  flow field oxygenated, the in situ respiration test is initiated by first stop-
ping air flow to  the contaminated soil (as would be done at the completion of the air

                                      5-11

-------
flow/tracer test), followed by the measurement of O2 uptake and CO2 production at the
soil gas probes over time. Typical soil gas respiration data are shown in Figure 5-4.
Selection of an appropriate sampling interval should be flexible based on actual site
conditions, and should be adjusted based on initial readings collected at ,3- to 4-hour
intervals following blower shutdown, to an interval of 6 to 48 hours from Day 2 to the
end of the test. With this as a guide, a typical respiration gas sample collection sched-
ule would be as follows: Day 1 - 0, 3, 6, 12,  18, and 24 hours; Day 2-30 and
42 hours; Day 3-56 and 68 hours; Day 4-80 hours; Day 5-104 hours;  Day 6-128
hours; Day 8-176 hours; and Day 10 - 224 hours.

Respiration Data Reduction. Respiration data reduction is carried out using either a
zero or first order reaction rate model to generate either zero or first order respiration
rate values (vol. percent/hour or 1/hour, respectively) from the slope of these linear
regression relationships.  A zero order reaction is described as one in which the change
in the dependent variable (in our case respiration gases) over time is independent of
the variable's concentration. This independence of the  removal rate is mathematically
described as follows:

                                   dC/dt = -k0                               (5-2)

yielding the integrated form:

                                   C-C0 = -kot                              (5-3)

where  C  =  concentration of the dependent variable  a time t, mg/L or vol. percent
       C0  =  initial concentration of the dependent variable a time t = 0, mg/L or vol.
              percent
       k0  =  zero order reaction rate constant, mg/L or vol. percent/time

This relationship is linear when plotted as measured respiration gas concentration
versus time as indicated in Figure 5-5, the slope of which is k0.

A first-order reaction is described as one in which the change in the dependent variable
(in our case respiration gases) over time is directly related to the variable's concentra-
tion.  This dependence of the removal rate is mathematically described as follows:

                                  dC/dt = -k1C                              (5-4)

yielding the integrated form:

                                ln(C) - ln(C0) = -M -                         (5-5)
                                     5-12

-------
     15
                                          X Oxygen
                                          O Catfcon DIoadde
                    1000
 2000        3000
Time (minutes)
4000  4500
  Figure 5-4. Typical soil respiration gas data collected during field in situ
  respiration test.
                       y = -0.145x + 16.72
                       r2 = 0.9575
                 0
                   0  10  20 30 40 50 60  70  80  90
                                  Time(hr)
Figure 5-5. Typical zero order soil respiration gas data collected during field
in situ respiration test. Zero order respiration rate = -0.145 vol%/hour.
                                5-13

-------
where k1 = the first order reaction rate constant, 1/time. This relationship is non-linear
when the measured respiration gas concentration is plotted versus time as indicated in
Figure 5-4, but can be linearized by plotting the natural log transformed gas concentra-
tion versus time as indicated in Figure 5-6.
                               y =-0.015x +2.89
                               r2 = 0.9701
                       -10 0   10 20  30 40 50 60 70 80  90
                                       Time(hr)
                Figure 5-6. Typical first order soil respiration gas
                data collected during field in situ respiration test.
                First order respiration rate = -0.015/hour.

These regression relationships are generated from linear least-squares analysis of the
field respiration data. The least-squares regression calculations can be carried out us-
ing standard statistical packages available on microcomputers and  many hand held
calculators.  The regression analyses and plots presented in Figures 5-5 and 5-6 were
generated using StatViewll on a Macintosh computer.  These figures show the mea-
sured data, the regression line of best fit, the 95 percent confidence bands (dotted
curves) of the slope of the best fit regression line for the data, the resultant linear re-
gression equation, and the r2 value for the relationship.  Quantification of the observed
respiration reactions using this statistical approach provides a quantitative description of
microbial activity observed at each monitoring point, and allows the quantitative com-
parison of respiration rates spatially at a given point in time, and temporally at a given
location in the contaminated site. These comparisons can  be made using data shown
in Figure 5-7, generated from regression analyses, that include:  the F-test and t-test
statistics, the p  value or probability of a  significant regression, and the confidence inter-
val of the slope of the regression relationship.

The first question that must be answered regarding the regression data should be
whether the relationship is significant, i.e., whether the measured oxygen uptake rate is
significantly greater than zero using a zero or first order regression model. If the slope

                                      5-14

-------
                    Simple Regression X-j: Time(hir)  Y1:02 Concentration (vol%)
Count: R:
5 .9785
Source DF:
R-squared: Adj. R-squared:
.9575 .9434
Analysis of Variance Table
Sum Squares: Mean Square:
RMS Residual:
1 1.2443
F-test:
REGRESSION
RESIDUAL
TOTAL
1
3
4
104.7431
4.6449
109.388
104.7431
1.5483

67.6497
p = .0038

                                No Residual Statistics Computed
 Variable:
                   Simple Regression X-j: Time(hr)  Y1:02 Concentration 
-------
of the regression line is statistically greater than zero, the p value of the regression will
be less than 0.05, and the 95 percent confidence intervals will not include 0.  As shown
in Figure 5-7, with a p value of 0.0038, the slope of this regression is significantly differ-
ent from zero, and is represented by a mean zero order oxygen uptake rate of-0.145
vol. percent/hour with a 95 percent confidence interval of the slope (the range of the
zero order oxygen uptake rate described by the data accurate within 5 percent of the
true value) of-0.089 to -0.201 vol. percent/hour.

Once the respiration rates are evaluated for statistical significance, background soil
respiration rate values should be used to correct contaminated soil values for basal soil
respiration taking place at the site. An inert gas tracer may be injected during soil
aeration so that respiration rate measurements can also be corrected for diffusion of O2
away from and CO2 diffusion to the sampling probe during respiration rate determina-
tions (Hinchee and Ong,  1992). However, the determination of background respiration
rates in uncontaminated soils accounts for both physical diffusion, as well as  biological
reaction mechanisms, and tracer use during respiration rate determinations at any time
other than immediately following air flow/tracer tests is not necessary. If background
respiration rates are significantly greater than zero, they should be subtracted from
respiration rates determined at locations throughout the contaminated to yield
background-corrected respiration  rates.  If background rates are not significantly
different from zero, no correction to rates measured in the contaminated soil is neces-
sary.

Finally, background-corrected respiration rates can be compared to rates published in
the literature for field scale bioventing systems to assess the relative biological activity
measured at the field site. Table 5-4 summarizes reported treatability test and field
demonstration respiration rate data from various sources that can be used for this com-
parison. If background-corrected field respiration rates  compare favorably with these*
reported data, significant biological activity is evident, and full-scale bioventing system
design and implementation is warranted. If background-corrected field respiration rates
are significantly lower (based on 95 percent confidence interval values) than these res-
ported data, less than optimal conditions are evident due to moisture or nutrient
limitations, and/or the presence of inhibitory materials, and the application of a
bioventing system may not be practical at this field site.

Consideration should then be given for the use of a laboratory treatability study to at-
tempt to identify the cause of this limited microbial activity,  or an alternative, nonbiologi-
cal remediation scheme should be evaluated for use at  the site.

Determination of the Governing Rate Law. Determination of the governing rate law can
be made by investigating the nature of the regression residuals generated during the
regression analysis. A residual is defined as the difference between the actual data
point and the value of the dependent variable on the regression line, and can have

                                      5-16

-------
Table 5-4.  Example Treatability Study and Full-Scale Bioventing System
Respiration Rates Reported From Various Sources.
Site Location
Alaska



Florida
Man/land
Nevada
Oklahoma
Utah"



Oxygen
Utilization Rate
(vol. %/day)
13.2
6.9 ±0.0
4.2± 2.6
7.7
10 ±0.5
3.0 ±0.2
6.0 ±0.2
4.0 ±0.5
0.19 to 7.7
0.10 to 3.6
0.06 to 1.3
0.0 to 0.02
Tempera-
ture (°C)
4 to 5
16
8
4 to 5
25
21
21
17
15
15
15
15
Source of
Data3
Treat.
Treat.
Treat.
Field
Treat.
Treat.
Treat.
Treat.
Field
Field (+H2O)
Field (+Nutr.)
Field (Back.)
Reference
Ongetal. (1994)
Ongetal. (1994)
Ongetal. (1994)
Ongetal. (1994)
Hincheeetal. (1992)
Hincheeetal. (1992)
Hincheeetal. (1992)
Hincheeetal. (1992)
Dupontetal. (1991)
Dupontetal. (1991)
Dupontetal. (1991)
Dupontetal. (1991)
 ' Treat. = field in situ respiration treatability test results; Field = field bioventing system
 performance data; +H2O = with moisture addition; +Nutr.) = with inorganic nutrient
 addition, and Back. = background soil respiration rates.
 These data were described by first-order relationships, with values in the table
 representing maximum rate values calculated from K., x 21 vol. percent.
                                     5-17

-------
either a positive or negative value. A standardized residual is the residual divided by
the value of the dependent variable at the point where the residual is calculated.  If a
given rate expression describes a data set, not only should the p value be less than
0.05, and the 95 percent confidence of the regression slope not include 0, but the stan-
dardized residuals should also be randomly distributed over the range of the indepen-
dent variable used in the regression.  If a pattern is observed in the standardized
residuals plot, the assumption that a particular linear model fits the data is not valid, and
an alternative model should be selected for use to describe the data. If the residuals'
plots for a number of models are similar, showing no particular pattern, as a matter of
practice, the simpler model form is selected.

Figure 5-8 shows the zero order linear regression for a set of hypothetical in situ respi-
ration rate data, indicating that the regression coefficient is high (0.8964), and the p
value is well below the 0.05 criteria point. The residual plot for this data set, shown in
Figure 5-9, clearly indicates that the linear model is not an adequate descriptor of thesse
data, due to the obvious pattern of the residual values. The first order model, shown in
Figure 5-10, is much improved, particularly when one inspects the residuals plot  for this
set of data as shown in Figure 5-11. Based on these results, a first order oxygen
uptake rate of -0.01/hour would be reported for these data.

Hydrocarbon Rate Determinations. Hydrocarbon rate determinations can be made us-
ing field determined in situ respiration data assuming the 3.5:1  O2:hydrocarbon mass
stoichiometry presented in Equation 5-1, and from known or estimated properties of the
site soil using the following  expression:


       Hydrocarbon degradation  =  10 k0  ®A  	P (32)	          ,_  „,
           rate (mglkg soil-d)        35   grj  (0.08205)  (273  +  7")

where    qa   = air filled porosity, unit less
        BD   = soil bulk density, kg/L soil
         P   = pressure, atm                                       '
         T   = temperature, °C

Assuming average values for these parameters of: qa = 0.3, BD = 1.4 kg/L soil,  P =  1
atm, and for a temperature = 25°C, the following relationship between measured zero
order degradation rates and equivalent hydrocarbon removal rates can be developed:

     Hydrocarbon  degradation   _ 10 k0 (0.3)       1  (32)      _ Q „. j        ,g -,.
        rate (mglkg soil-d)        35   ^  4)  (0.08205) (298)       °
                                      5-18

-------
Count:
            16
2 14

l<2
c 12
o

"T5 10
         g  8



         3  6

         c
         
-------
    CO
   'to
   
-------
                tn
                CO
 2

1.5

 1
                CD
               *  .5
               "S
               .y  o
               •o
                CO
               •g -.5'
                (0
               co  .^
                 -1.
                    -25  0   25  50  75  100125150175200225250
                                    Time(hr)
              Figure 5-1 1 .  Residuals plot for first order linear regression
              results for the set of hypothetical in situ respiration rate data.

As indicated earlier in Table 5-4, these calculations can be made for sites where first
order oxygen utilization rates are observed by multiplying the first order rate (1/time) by
the concentration of oxygen occurring within the soil at the site (typically 21 vol. per-
cent) to yield an equivalent oxygen utilization rate with units identical to that of k0. The
oxygen utilization rate, however, changes directly with oxygen concentration if removal
follows a first order relationship, so predicted hydrocarbon degradation rates will be
dependent on  oxygen concentrations maintained within the contaminated site under
these conditions.                                                    .

Remediation Time. Remediation time can be estimated from these hydrocarbon
degradation rates knowing the initial concentration of contaminant existing at the site.
Respiration rates decrease linearly with decreasing contaminant concentrations below
approximately 3,000 mg/kg total petroleum hydrocarbons (TPH) (Ravipaty,  1994), so
the time for site remediation provided by Equation 5-8 is the minimum time to site
clean-up.  The actual time to reach soil closure levels will  likely be two to three times
longer.  Equation 5-8 provides an initial estimate of "best-case" time for preliminary
evaluation of the feasibility of a bioventing system applied at a given site.
     Minimal Time to
      Remediation
            (Soil Contamination, mg/kg TPH)	
     (Hydrocarbon Degradation Rate,  mg/kg soil-d)
(5-8)
       Expected Time to Remediation » 2 to 3 x (Minimal Time to Remediation)  (5-9)

                                      5-21

-------
Phase III - Bioventing System Design - Interpretation of Utilization Rate
Once in situ oxygen respiration rate values have been estimated, corrected for back-
ground respiration at the site, and determined to reflect biological activity that warrants
application of a bioventing system at the site, full-scale system design should be carried
out.  Review of detailed bioventing system design procedures is outside the scope of
this document;  however, the application of system design to air flow/tracer test and in
situ respiration test results presented above will be highlighted.

Air Flow Considerations
Air flow considerations are critical to optimal bioventing system design, as it is impera-
tive that the full-scale system effectively deliver the required oxygen to the locations
and at a rate needed to maintain optimal aerobic activity, while minimizing air flow to
reduce or even eliminate volatile emissions from the site.  The air flow rates that must
be maintained can be determined directly from oxygen uptake rate measurements de-
scribed in "In Situ Respiration Tests." This requires careful consideration of tracer
test/air permeability test results, so that design components can be incorporated into
the full-scale system  and overcome air flow limitations due to dead zones, low permea-
bility lenses, short-circuit pathways, etc., which limit vapor flow rates through the zone
of contamination. These design components may include: passive/ active injection
wells strategically placed to minimize dead zones, multiple air injection/ extraction wells
used to treat distinct soil layers existing at the site, etc. The reader is referred to the
section of this document entitled  "Procedures for Conducting Tracer Tests to Evaluate
Air Flow During Soil Vapor Extraction and Bioventing" for further details regarding the
conduct of tracer tests and test data interpretation.

System Operating Conditions for Full-Scale Design
System operating conditions for full-scale design are determined in large part by the
governing rate law for oxygen utilization observed throughout the field site. As indicat-
ed above, if the oxygen uptake data are governed by a zero order rate law, oxygen
utilization, and concomitant contaminant degradation rates are independent of soil gas
oxygen concentrations until oxygen limitations (» 2 vol. percent oxygen) occur. From a
practical standpoint this means that a constant inflow of oxygen to the subsurface is not
required to maintain optimal oxygen uptake rates. Under these conditions, a pulse
pumping system is possible where a blower would be operated for short periods of time
for soil oxygenation, followed by longer periods with no air flow during which time soil
"incubation" and contaminant removal takes place without air movement and possible
air emissions.  This pulsed system could take advantage of existing schedules of facility
operations and maintenance personnel or existing blower equipment, or might provide
a cost effective approach for satisfying stringent requirements on mass emission rates,
limitations on operating hours due to noise considerations, etc. Regardless, zero order
uptake rates allow a much wider range of operating modes than does a system ex-
hibiting first order in situ oxygen uptake rates.
                                      5-22

-------
 Field systems governed by first order oxygen uptake relationships offer no flexibility in
 their operating mode. Performance of these systems is enhanced with increased soil
 oxygen concentrations, and they must be operated in a continuous mode to maximize
 removal of contaminants.  One option that is currently being assessed by the U.S. Air
 Force to maximize contaminant removal, while still minimizing vapor emissions in first
 order rate dominated sites, is the use of pure oxygen in place of atmospheric air as the
 gaseous oxygen source.  While this option has increased costs associated with pure
 oxygen generation, a reduction in airflows by a factor of approximately five significantly
 reduces air emissions, and reduces many overall system costs associated with these
 reduced flow rates. The overall cost effectiveness of these pure oxygen systems has
 yet to be proven, but may be the only way to significantly enhance bioventing systems
 whose performance is found to be sensitive to soil gas oxygen levels.   ;

 Phase IV- System Monitoring and Performance Evaluation
 Routine system monitoring is essential to the optimal operation and control of a field-
 scale bioventing system.  The SVE well and vapor monitoring probes installed for
 conducting the initial air flow/permeability and in situ respiration testing should be incor-
 porated into the full-scale field system as much as is practical. Additional wells should
 be placed as described above to overcome air flow limitations evident from tracer test
 results. In addition, multi-level vapor probes should be added as necessary to provide
 a representative, three dimensional picture of contamination existing at the site prior to
 initiation of the full-scale system.  The logical places for these monitoring points are at
 the locations from which soil  core samples are collected for initial contaminant quantita-
 tion. Multi-level, nested probes, such as those described in the AFCEE procedure by
 Hinchee et al. (1992), minimize the effort and expense of probe placement, as well as
 field sample collection and analyses, by utilizing common bore holes for multiple vapor
 probe depths, and  should be considered strongly for use at new bioventing field sites.
As a rule of thumb, a minimum of nine vapor monitoring points (three spatial, radial
 locations at three depths each) should be installed per SVE well to provide the data
 necessary to adequately monitor vapor flow,  respiration and contaminant removal within
 a field site. The reader is referred to the "Procedure for Conducting Tracer Tests to
 Evaluate Air Flow During Soil Vapor Extraction and Bioventing" for additional discussion
 related to vapor probe design and placement.

 Following complete bioventing field system installation, the soil vapor probes should be
 monitored daily to verify that the specified system design and operating mode is provid-
 ing  air flow to the site that was anticipated. System operating flow configuration and/or
flow rate changes may be necessary to adapt the bioventing system to  actual full-scale
conditions encountered at the field site. This "shake-down" period is expected to last
 one to two weeks, with operation after that time being fairly stable, requiring only
 minimal adjustment and maintenance. It  is recommended that routine system monitor-
 ing  be conducted monthly for the first six months of operation, and then quarterly
                                     5-23

-------
thereafter, to verify proper system operation and allow system "fine-tuning" as
remediation takes place throughout the site.

Routine system monitoring should include, at a minimum, the following parameters:
system flow rate (preferably flow rate to each injection/extraction well in the system);
extraction system gas characteristics (O2/CO2, TPH, temperature, relative humidity,
vacuum) if an air extraction system is utilized; soil gas monitoring point characteristics
(O2/CO2, TPH, temperature, relative humidity, vacuum/pressure); and blower vacu-
um/pressure.

In addition to the collection of routine system monitoring data to ensure system opera-
ting effectiveness, quarterly to semi-annual system shut-down tests should be con-
ducted to assess the progress of remediation taking place throughout the site. These
shutdown tests are conducted in a manner identical to those described in "In Situ Res-
piration Tests" in which the air injection/extraction system is shut off and oxygen  uptake
is allowed to proceed without oxygen replacement. As was done with the Phase II in
situ respiration data, data collected from these routine shut-down tests  are statistically
evaluated for their significance. Comparisons of overlapping confidence  intervals of the
slope of the oxygen uptake relationships, measured at given sampling locations but at
previous time periods, are made to evaluate whether a significant reduction in  respira-
tion rates is occurring (inference that a significant reduction in contaminant levels is
occurring as well). In addition, contaminated site respiration rate confidence intervals
are compared to background respiration rates to determine if microbial  activity at the
site is reaching background activity, suggesting that clean-up levels are being  reached
at the site.

Respiration  rates can decrease due to limited nutrient availability, and/or  low soil water
contents (< 25 percent field capacity) in addition to reduced soil hydrocarbon levels.
With on-going soil vapor relative humidity measurement,  soil drying should be evident
over time. Drying is minimized in bioventing systems due to low air flow movement
through the  soil,  but if it becomes an obvious limitation to system performance, control-
led surface irrigation in coarse grained soils, or injection of water saturated vapor into
fine grained soils, can aid in modifying soil water content to acceptable levels for
improved microbial activity. Nutrient limitation is a more difficult matter, as nutrient
supply and profusion into soils is limited by the high sorptive capacity of soils for typical
inorganic nutrients used in remediation systems. Evidence from field scale bioventing
systems treating fuel contaminated soils is available which  indicates that  nutrient
addition in the field does little to nothing to improve the performance of these systems
(Dupont et al., 1991; Miller 1991), and nutrient limitation should not be considered a
major cause of significant respiration rate reductions observed at a field site.

The primary cause of significant decreases found in in situ  respiration rates, if soil water
is not limiting, is  a significant reduction in degradable contaminant concentrations in the

                                      5-24

-------
soil. This respiration rate/contaminant concentration relationship has been suggested
from field data, and is well documented in the wastewater literature, but has only been
verified recently in laboratory-scale microcosm studies conducted at Utah State Uni-
versity (Ravipaty, 1994). Some of the results of this work for JP-4 contaminated soil
from Hill AFB, Utah is presented in Figure 5-12 for 10 JP-4 concentrations levels from 0
to 10,000 ppm, and at two soil water contents of 50 and 75 percent field capacity.
          §•   6
          CM
          O
          £   5
           o

           Q>
           c
           g   o
           
           N
o Respiration Rate (50% FC)
« Respiration Rate (75% FC)
                        2000
4000   6000    8000   10000
 TPH Concentration (ppm)
                                                                 12000
           Figure 5-12.  Respiration rate/contaminant concentration rela-
           tionships generated in laboratory-scale microcosm studies
           conducted with JP-4 contaminated soil from Hill AFB, Utah
           (from Ravipaty, 1994).
As indicated in Figure 5-12, respiration rate appears to vary linearly with contaminant
concentration to a soil level of approximately 1,000 ppm TPH, beyond which respiration
rate reaches a pseudo steady-state value. The actual steady-state value reached
varies with soil moisture content. This variance with soil water content is postulated to
be related to an increased mass of contaminant available to the soil organisms due to
an increase in the volume of the soil water compartment as soil water content increas-
es.  However, more critical to the use of field determined respiration rates for the
evaluation of the progress of remediation is the fact that respiration rate approaches
zero as a function of soil contaminant level independent of soil water conditions at
which this respiration is taking place. This makes the use of declining respiration rates,
                                     5-25

-------
particularly as they approach zero, or background respiration rate levels, a powerful
indicator of the extent to which contaminant mass removal is taking place at the site.

With this being the case, the possible outcomes of this phase of the bioventing proce-
dure are as follows: 1) respiration rates are shown to remain statistically greater than
background respiration rate levels and within the range reported in the literature (Ta-
ble 5-4), contaminant concentrations remain high (> 2,000 ppm TPH) and bioventing
should continue; 2) respiration rates are shown to be greater than background levels
but are decreasing overtime,  contaminant mass removal is continuing, contaminant
concentrations in the soil are approaching « 1,000 ppm TPH, and bioventing should
continue; and 3) respiration rates are shown to be equal to background levels, con-
taminant removal and site remediation are indicated, and confirmatory soil borings
(Phase V) should be collected to verify that remedial soil concentration goals have been
met.                                                          .

Phase V- Verification of System Performance
The final step in the bioventing test  procedure, that of verification of bioventing system
performance using confirmatory soil core results, is reached through a positive outcome
from Phase IV described above. Again, if respiration throughout the site approaches or
is statistically equivalent to background levels, low residual contaminant levels are indi-
cated and verification of this result should be provided from soil concentration values.
Soil core samples should be collected in a manner identical to that used in preliminary
site assessment activities to allow direct comparison of results between sampling time
intervals.  In addition, due to the large variability inherent in soil sampling,and contami-
nant distribution, confirmatory samples should be collected as close as possible to the
locations of the original soil cores, if valid comparisons of contaminant levels are to be
made over time.

If contaminant mass removal has occurred, and low respiration rate results are indica-
tive of low residual contaminant concentrations, these confirmatory soil core results
should show low levels of both volatile and semivolatile constituents remaining at the
site.  If measured soil concentrations are below regulated  site soil clean-up levels, the
site would be considered for a closure action,  and the bioventing system would no long-
er have to operate at this site. If soil concentrations remain above regulatory action
levels, the rate and mode of operation of the bioventing system should be evaluated,
and system modifications should be made to enhance the removal of remaining con-
taminant so that closure can be accomplished in the future.  In the latter case a
modified bioventing system would go back into operation,  and respiration rates during
shutdown  periods would continue to be monitored on a quarterly basis until once  again
background oxygen uptake rates are observed,  initiating the collection of a new round
of confirmatory soil core samples.
                                      5-26

-------
Summary and Conclusions
The bioventing test procedure presented in this document details an approach to the
site specific determination of the feasibility of bioventing technology that is integrated
with system monitoring and performance evaluation from initial site assessment activ-
ities through final confirmatory soil core analyses. The procedure is composed of five
phases of activity which include the following:

Assessment of the Potential for Contaminant Biodegradation Under Actual Field
Conditions
In this first phase of the procedure respiration gas (O2/CO2) characterization is incorpo-
rated into conventional soil gas survey activities to detect the magnitude and extent of
biological activity, and consequently, oxygen depletion/carbon dioxide enrichment of the
soil gas at the site. If bioactivity is evident from soil gas survey results, the next phase
of the procedure is carried out.

Assessment of Air Flow and In Situ Respiration Rates Under Actual Field Condi-
tions
With biodegradation evident at the field site, air flow/air permeability distribution  and
actual oxygen uptake rates  must then be determined. The test procedure describes a
combined air flow/tracer-in situ  respiration test procedure that takes advantage of moni-
toring probes and subsurface oxygenation provided during the air flow test for collecting
site wide respiration data. Procedures are described to reduce the respiration data to
generate respiration rates and to assess their statistical significance relative to site
background respiration rate levels.  Finally, procedures for converting respiration rates
into equivalent hydrocarbon degradation rates are provided, along with estimation
procedures for the time to site remediation.

Bioventing System Design
Based on air flow and in situ respiration  rate results from Phase II, the potential oxygen
supply rate (airflow) is matched with the oxygen  demand rate (in situ respiration rates)
in rational bioventing  system design. The nature of the respiration rate law observed
from Phase II results  are used to recommend either a pulse operating mode system
(zero order reactions) or continuous mode system (first order reactions) to optimize
overall system performance.                                        '.

Full-scale System Monitoring and Performance Evaluation
The procedure describes  the use of routine shutdown tests to monitor the changes
taking place in respiration rates over time as contaminants are removed  from the site.
These respiration rates are statistically compared to background respiration levels so
that when only background  activity is detected throughout the site, soil core samples
may be taken to confirm system performance.
                                      5-27

-------
System Performance Verification
The final phase of the procedure describes the use of soil core samples, collected from
locations near those used for initial site characterization based on quarterly in situ
respiration results, as the ultimate proof that soil remediation has proceeded to the
point where site closure is possible.

With the use of this integrated approach to site assessment, bioventing system design,
and system performance monitoring and evaluation, an optimal design for site remedia-
tion using this innovative bioremediation technology can be possible. The key to this
optimal design, however, is the collection and interpretation of system performance and
bioactivity data from the initial site assessment stages to the completion of a project so
that the flexibility and economy of this air based bioremediation system can be taken
advantage of on an on-going basis as the "bioreactor," represented as the contami-
nated site soil, evolves and system operating demands change during the course of site
remediation.

Calibration of Field Instrumentation for Hydrocarbon and Oxygen/Carbon Dioxide
Determinations - Electronic  Detection Instruments
Field Preparation Activities

      1.    Visually inspect meters for damage before field sampling event.  Check air
            or liquid filters (clean or replace if necessary) as appropriate and check
            battery condition.

      2.    Calibrate meters in laboratory before field sampling event using proce-
            dures below.  Replace O2 sensor or return instruments to manufacturer for
            repair if instruments will not calibrate. Collect equipment and recording
            sheets as required.

Hydrocarbon Meter
(While the specific meter utilized at a given site will vary, the procedures outlined below
are representative of those necessary to ensure that accurate, representative and
reproducible data are collected from a field effort. Refer to the Owner's Manual for
details of operation and calibration specific for the instrument being used.)

      1.    Turn the meter on and allow to warm up for 5 minutes.

      2.    Adjust the meter to read 0 while sampling atmospheric air using the
            external zero knob.

      3.    Calibrate the meter using hexane calibration gas (a concentration above
            the maximum expected field concentration should be used).
                                      5-28

-------
            a.     Fill a Tedlar bag with calibration gas. Purge and refill.
            b.     Attach the Tedlar bag to the meter and adjust the meter to read the
                   calibration gas concentration using the internal span adjustment.

      4.    Check the zero while sampling atmospheric air.  Adjust if necessary.

      5.    Check the calibration following Step 3 above.  Adjust if necessary.

      6.    A calibration check should be done a minimum of three times daily.
                                                                   /
      7.    Record initial and continuing calibration readings on a calibration data
            sheet (see Figure 5-13 for sample format).

O/CO2 Meter
(While the specific meter utilized at a given site will vary, the procedures outlined below
are representative of those necessary to ensure accurate, representative and repro-
ducible data are collected from a field effort. Refer to the Owner's Manual for details of
operation and calibration specific for the instrument being used.)

CO2 Calibration

      1.    Turn the meter on and allow to warm  up for 5 minutes.

      2.    Adjust the meter to read 0.05 percent while sampling atmospheric air.
            Use the CO2 zero knob to adjust.

      3.    Calibrate the meter using calibration gas (15.0 percent CO2 typically
            used).

            a.     Fill a Tedlar bag with calibration gas. Purge and refill.
            b.     Attach the Tedlar bag to the meter and adjust the meter to read the
                   calibration gas concentration using the appropriate span adjust-
                   ment.

      4.    Check the atmospheric air reading. Adjust if necessary.   ;

      5.    Recheck the calibration following Step 3 above.  Adjust if necessary.

      6.    Check the calibration using a calibration check gas of a different  concen-
            tration (5.00 percent CO2 typically used).  Recalibrate using the calibration
            gas  if the calibration  check is off by more than ±15 percent of the known
            concentration.
                                      5-29

-------
Initial Calibration Date:
Calibration Measured
Time Vapor/Gas Concentration Concentration





































Instrument
Number









Calibration Check
Calibration Measured Instrument
Time Vapor/Gas Concentration Concentration Number













































Figure 5-13.  Sample initial and continuing calibration data record sheets for
hydrocarbon and respiration gas measurement instruments.
                                  5-30

-------
       7.     A calibration check should be done a minimum of three times daily.

       8.     Record initial and continuing calibration readings on a calibration data
             sheet (see Figure 5-13 for sample format).

O2 Calibration

       1.     Turn the meter on and allow to warm up for 5 minutes.

       2.     Adjust the meter to read 21 percent while sampling atmospheric air.  Use
             the appropriate oxygen calibration knob to adjust the meter to the proper
             setting.

       3.     Zero the meter using calibration gas (5.00 percent CO2 in N2 typically
             used).

             a.     Fill a Tedlar bag with calibration gas. Purge and refill.
             b.     Attach the Tedlar bag to the meter and adjust the meter to read
                   zero using the appropriate oxygen zero potentiometer.

       4.     Check the atmospheric air reading. Adjust if necessary.

       5.     Recheck the calibration following Step 3 above. Adjust if necessary.

       6.     Check the calibration using a calibration check gas of a different concen-
             tration (7.0 percent O2 typically used).  Recalibrate atmospheric and zero
             readings if calibration check is off by more than ±15 percent of the known
             concentration.

       7.     A calibration check should be done a minimum of three times daily.

       8.     Record initial and continuing calibration readings on a calibration data
             sheet (see Figure 5-13 for sample format).

Field Sampling and Analysis for Hydrocarbon and O2/CO2 Determinations -
Electronic Detection Instruments
Routine Soil Gas Monitoring

       1.     Calibrate instruments following the directions specified in "Calibration of
             Field Instruments for Hydrocarbon and Oxygen/Carbon Dioxide Deter-
             minations - Electronic Detection Instruments."
                                     5-31

-------
 2.     Fill a Tedlar bag with the gas sample using the vacuum pump. Purge and
        refill. The total sampling time should not exceed 1 minute.  Record the
        vacuum gauge reading while sampling on a soil gas monitoring data rec-
        ord sheet (see Figure 5-14 for sample format). If vacuum reading is less
        than 6 in Hg, check pump for broken connections.
Monitoring
Point









02
(vol. %)









C02
(vol. %)









TPH
(ppm)









Vacuum
(in.H2O)









Comments





(



Figure 5-14. Sample format for soil gas monitoring data record sheet.

 3.    Analyze the gas samples by connecting the hydrocarbon and O2/CO2
       meter to the Tedlar bags.  The meter can be connected in series when
       TPH readings are expected to be below 200 ppm. Record the O2, CO2,
       and TPH readings on a soil gas monitoring data record sheet (see Figure
       5-14 for sample format).

       a.     If the soil gas O2 concentration is less than 12 percent, a dilution
              fitting may be necessary as per manufacturer's requirements for
              the specific instrument being used,  and should be noted on data
              sheet.  Hydrocarbon readings are recorded directly and should be
              multiplied by a factor of 2 prior to final data analysis when a dilution
              fitting is used. (Some manufacturers recommend the use of dilu-
              tion fitting when less than 10 vol. percent O2 occurs in the sample,
              and require its use when the O2 content is less than 8 vol. percent
              02.)

       b.     Recheck atmospheric readings for O2, CO2 and TPH periodically
              and recalibrate meters as necessary.

       c.     Analyze samples indoors during cold weather (below 20°F).
                                5-32

-------
      4.    Measure the soil gas temperature following procedures specified by the
            manufacturer. Record the temperature on a soil gas monitoring data
            record sheet (see Figure 5-14 for sample format).

Respiration Shutdown Test Monitoring

      1.    Shut blowers off to begin respiration test.

      2.    Analyze up to four soil gas samples from each probe per day, following
            the procedures specified in the "Soil Gas Monitoring" section above at
            various time intervals.  Record data, including the date and time, on an in
            situ respiration test data record sheet (see Figure 5-15 for sample format).
Date








Time








02
(vol. %)








CO2
(vol. %)








TPB
(ppm)








Pump
Vacuum
(in.H2O)








Soil
Temp.
(°F)








Comments








Figure 5-15.  Sample format for in situ respiration test data record sheet.

      3.    Continue for up to a 10-day period, or until the soil gas O2 level reaches
            2 vol. percent.

      4.    Turn blowers on at the end of the respiration test.

      5.    Analyze respiration data according to methods presented in "Respiration
            Data Reduction."

References for Section 5

Alexander, M.  1977. Introduction to Soil Microbiology. John Wiley and Sons, Inc.,
New York, NY.

Atlas, R. M.  1981.   Microbial Degradation of Petroleum Hydrocarbons: An Environ-
mental Perspective. Micro. Rev.45(1): 185-209.

Bennedsen, M. B.  1987. Vacuum VOCs From Soil, Poll. Eng. 19(2):66-68.
                                     5-33

-------
Downey, D. C., R. E. Hinchee, M. S. Westray, and J. K. Slaughter. 1988. Combined
Biological and Physical Treatment of a Jet Fuel-Contaminated Aquifer, Proceedings of
the Petroleum Hydrocarbons and Organic Chemicals in Ground Water:  Prevention De-
tection and Restoration, Dublin, OH. pp 627-645.

Downey, D. C., and J. F. Hall.  1994. Addendum One to Test Plan and Technical
Protocol for Field Treatability Test for Bioventing - Using Soil Gas Surveys to Determine
Bioventing Feasibility and Natural Attenuation Potential. Report to U.S. Air Force Cen-
ter for Environmental Excellence, Brooks AFB, TX.

Dragun, J. 1988.  Microbial  Degradation of Petroleum  Products in Soils, Soils Contami-
nated by Petroleum - Environmental and Public Health Effects, E.J. Calabrese and P.T.
Kostecki, Ed., John Wiley and Sons, Inc., New York, NY. pp 289-300.

Dupont, R. R., W. J. Doucette, and R. E. Hinchee.  1991. Assessment of In Situ Biore-
mediation Potential and the Application of Bioventing at a Fuels-Contaminated Site, In
Situ Bioreclamation: Applications and Investigations for Hydrocarbon and Contaminat-
ed Site Remediation, R. E. Hinchee and R.  F. Olfenbuttel, Ed., Butterworth-
Heinemann,  Boston, MA. pp 262-282

Ely, D. L. and D. A. Heffner.  1988. Process for In Situ Biodegradation of Hydrocarbon
Contaminated Soil, Patent No. 4,765,902, U.S. Patent  Office.

Hinchee, R. E. and D. C. Downey.  1988. The Role of Hydrogen Peroxide in Enhanced
Bioreclamation,  Proceedings of the Petroleum Hydrocarbons and Organic Chemicals in
Ground Water: Prevention Detection and Restoration, Dublin, OH.  pp 715-722.

Hinchee, R. E., and S. K. Ong.  1992. A Rapid In Situ  Respiration Test for Measuring
Aerobic Biodegradation Rates of Hydrocarbons in Soils. J. Air Waste Manage. Assoc.
42(10): 1305-1312.

Hinchee, R. E., D. C. Downey, J. K. Slaughter, D. A. Selby, M. S. Westray, and  G. M.
Long. 1989.  Enhanced Bioreclamation of Jet Fuels - A Full-Scale Test at Eglin AFB,
FL, Final Report ESL-TR-88-78, Headquarters Air Force Engineering Services Center,
Tyndall Air Force Base, FL.

Hinchee, R. E., D. C. Downey, R. R. Dupont, P. Aggarwal, and R. N. Miller.  1991.
Enhancing Biodegradation of Petroleum Hydrocarbons Through Soil Venting, J. Haz.
Mat. 27:315-325.

Hinchee, R. E., S. K. Ong, R. N. Miller, D. C. Downey, and R. Frandt. 1992. Test Plan
and Technical Protocol for a Field Treatability Test for Bioventing, Revision 2, Report to
U.S. Air Force Center for Environmental Excellence, Brooks AFB, TX.

                                     5-34

-------
Lee, M. D., J. M. Thomas, R. C. Borden, P. B. Bedient, J. T. Wilson, and C. H. Ward.
1988.  Biorestoration of Aquifers Contaminated with Organic Compounds. CRC Crit.
Rev. Env. Control 18(1):29-89.                                     '

McGinnis, D., R. R. Dupont, K. Everhart, and G. St. Laurent. 1994.  Evaluation and
Management of Field Soil Pile Bioventing Systems for the Remediation of PCP Con-
taminated Surface Soils. Environmental Technology (Letters) 7:729-739.

Miller, R. N., R. E. Hinchee, and C. C. Vogel. 1991. A Field-Scale Investigation of
Petroleum Biodegradation in the Vadose Zone Enhanced by Soil Venting at Tyndall
AFB, Florida, In Situ Bioreclamation: Applications and Investigations for Hydrocarbon
and Contaminated Site Remediation, R. E. Hinchee and  R. F. Olfenbuttel, Ed.,
Butterworth-Heinemann, Boston, MA. pp 283-302.

Ong, S. K., A. Leeson, R. E. Hinchee, J. Kittle, C. M. Vogel, G. D. Sayles, and R. E.
Miller.  1994. Cold Climate Applications of Bioventing, Hydrocarbon Bioremediation,
R. E. Hinchee, B. C. Alleman, R. E. Hoeppel and R. E. Miller, Ed., Lewis Publishers,
Ann Arbor, MI. pp 444-453.

Ostendorf. D. W. and D. H. Kampbell. 1989. Vertical Profiles and Near Surface Traps
for Field Measurement of Volatile Pollution in the Subsurface Environment, Proceedings
of the New Field Techniques for Quantifying the Physical and Chemical Properties of
Heterogeneous Aquifers, Dallas, TX.

Ravipaty, A. 1994.  Determination of Respiration Rate/Soil Concentration Correlations
of Fuel Contaminants in Three Field Soils.  Technical paper presented at Scholar's Day,
Eccles Conference Center, Utah State University, Logan, Utah, May 17. 12 pp.

Riser, E. 1988. Technology Review - In Situ/On-Site Biodegradation of Refined Oils
and Fuels, N68305-6317-7115, Naval Civil Engineering Laboratory.

Stapps, J. J. M. 1989. International evaluation of In Situ Biorestoration of Contami-
nated Soil and Groundwater, 738708006, National Institute of Public Health and En-
vironmental Protection (RIVM).

Wetzel, R. S., C. M.  Darst, D. H. Davidson, and D. J. Sarno. 1987. In Situ Biological
Treatment Test at Kelly Air Force Base, Volume II - Field Test Results and  Cost Model,
Final Report TR-85-52, Headquarters Air Force Engineering Services Center, Tyndall
Air Force Base, FL.

van Eyk, J. 1994. Venting and Bioventing for the In Situ Removal of Petroleum From
Soil, Hydrocarbon Bioremediation, R.E. Hinchee, B.C. Alleman, R.E. Hoeppel  and R.E.
Miller, Ed., Lewis Publishers, Ann Arbor, Ml.  pp 243-251.

                                    5-35

-------
                                   Section 6

          Procedures for Evaluating Natural Attenuation in Groundwater
Introduction
Natural attenuation is a risk management strategy that invokes intrinsic bioremediation,
dilution, dispersion, sorption, and other physical loss mechanisms to control exposure
to contaminants and restore the environment. Several criteria should be considered
before choosing natural attenuation as the principle remedial strategy (Wiedemeier,
1994, Wisconsin Department of Natural Resources, 1993; Underground Tank Technol-
ogy Update, 1993). These include:

      •     Risk of further environmental damage.

            Risk of human endangerment [Defining the risk associated with a hazard-
            ous waste site is beyond the scope of this chapter; however, the natural
            attenuation rate can greatly influence the calculated  risk associated with a
            contaminated site. The Emergency Standard Guide for Risk-Based
            Corrective Action Applied at Petroleum Release Sites (ASTM ES 38-94)
            provides a practical approach for conducting risk assessments].

      •     Detrimental consequences to  local flora and fauna.

            Technical feasibility, practicality, and effectiveness of other technologies.

      •     Site-specific evidence for successful application of intrinsic bioremedia-
            tion.

      •     The cost of natural attenuation compared to other options

When these issues can be addressed in favor of natural attenuation, the technology is a
cost effective and practical remedial alternative for soil and groundwater.

Intrinsic bioremediation is the preferred term to describe the natural biological process-
es that lead to contaminant biodegradation  (Wiedemeier et al., 1994).  Intrinsic biore-
mediation can  occur in any environment that supports microbiological activity; however,
the rate of biodegradation may be slow due to the lack of a suitable respiratory

                                      6-1

-------
substrate (such as oxygen) or inorganic nutrients (such as fixed nitrogen), an extreme
pH, low soil moisture, or limited contaminant bioavailability.  Elimination of the contam-
inant source is essential for the successful application of intrinsic remediation.  Accurate
delineation of contamination, understanding subsurface conditions and characteristics,
and contaminant migration rates and direction are critical for evaluating the success of
natural attenuation and for establishing regulatory support for its use at a site (Davis,
Klier, and Carpenter, 1994; Davis etal., 1994; Wiedemeier et al., 1994, Wilson, 1993,
Wisconsin Department of Natural Resources, 1993).

Major obstacles that complicate the application of natural attenuation at appropriate
sites include defining the potential for natural attenuation and developing a practical,
cost effective monitoring plan that provides conclusive evidence of contaminant mass
reduction.  Identification and  selection of useful monitoring parameters require  a thor-
ough knowledge of site characteristics. The site characterization methods and monitor-
ing parameters used to evaluate natural attenuation are presented here. Table 6-1
lists some of the commonly used analytical methods for these parameters. The men-
tion of a site characterization or monitoring parameter does not, however, mean that the
parameter must be analyzed at each site undergoing natural attenuation.: Inclusion
here is to point out potential parameters, each of which may be useful at a particular
site.  Characterization and monitoring parameters are discussed separately, with
monitoring  parameters segregated based on the contaminant loss mechanism  they are
intended to illuminate.

Site characterization data should include:

      •     Site topography and geology

      •     Hydraulic conductivity of the aquifer

      •     Depth to groundwater, groundwater gradient, direction of flow, flow rate,
            and recharge rate

      •     Contaminant composition, location of non-aqueous phase liquids  (NAPL),
            location of dissolved-phase contaminants, and the mass of contaminant
            present

      •     Physical and chemical properties of the contaminant

      •     Groundwater quality parameters
                   Temperature
                   pH
                   Dissolved oxygen
                                      6-2

-------
Table 6-1. Analytical Methods for Natural Attenuation.
Parameter
Method
Lab Measurements
VOCs
TPH
BTEX
MTBE
Nitrate
Nitrite
Sulfate
Methane
Ammonium
Total Kjeldahl N
Phosphate
Alkalinity
PH
Chloride
SW-846, Method 8260
SW-846, Method 8015
SW-846, Method 8202
SW-846, Method 8020
SW-846, Method 9056, SW-846, Method 300
SW-846, Method 8260, SW-846, Method 300
SW-846, Method 8260, SW-846, Method 300
Headspace analysis by IT Biotechnology Laboratory
Standard Methods for Analysis of Water and Wastewater,
Method 350.2
Standard Methods for Analysis of Water and Wastewater,
Method 35 1.2
Standard Methods for Analysis of Water and Wastewater,
Method 365.3
Standard Methods for Analysis of Water and Wastewater,
Method 31 0.1
Standard Methods for Analysis of Water and Wastewater,
Method 150.1
Standard Methods for Analysis of Water and Wastewater,
Method 300
4500-NH3, SW-846
4500-Norg, SW-846
4500-P, SW-846
2320 B, SW-846
4500-H+ or SW-846
4500-CI, SW-846
Field Measurements
. Iron (II)
Temperature
Conductivity
Redox
pH
Dissolved oxygen
Standard
Standard
Standard
Standard
Methods
Methods
Methods
Methods
Standard Methods
Method 150.1
Standard
Methods
for Analysis
for Analysis
for Analysis
for Analysis
for Analysis
for Analysis
of Water
of Water
of Water
of Water
of Water
of Water
and
and
and
and
and
and
Wastewater,
Wastewater,
Wastewater,
Wastewater,
Wastewater,
Wastewater,
3500-Fe D
2550
2510
2580
B'
B
B
4500-H+ or SW-846
4500-OG
                                    6-3

-------
                   Methane
                   Total Organic Carbon
                   Alkalinity
                   Oxidation/reduction potential (Redox)
                   Nitrate
                   Iron(ll)
                   Sulfate
                   Sulfide
                   Ammonium
                   Phosphate
                   Total Kjeldahl Nitrogen
            -      Chloride
                   Conductivity

A major division of monitoring parameters is biological and physical/chemical.  Poten-
tially useful biological monitoring parameters include the following:

      •     Microbial respiration
      •     Conversion/consumption of respiratory substrates
      •     Biologically induced changes in geochemistry
      •     Biodegradation rates

Potentially useful physical/chemical monitoring parameters include:

      •     Dilution and dispersion
      •     Volatilization
      •     Sorption
      •     Contaminant mass loss
      •     Contaminant concentration reduction
      •     Appearance of degradation products

The conventional monitoring parameter under current regulatory legislation is reduction
in contaminant concentration, and this may be limited to groundwater only or include
both soil and groundwater.

The final section of this chapter presents a stepwise approach for developing a natural
attenuation site management plan. The approach provides a skeleton from which natu-
ral attenuation can be defined and a responsible site management plan implemented.
Explicit details are  intentionally omitted to avoid casting an approach that may not make
sense in total or in  part for a given site. Site specific characteristics, the current level of
site knowledge, previous remedial activities, and public,  regulatory, and environmental
sensitivities must be merged to yield an acceptable  natural attenuation model and
verification or monitoring plan.

                                      6-4

-------
 Site Characterization and Selection of Natural Attenuation for Site Remediation
 Defining site conditions is the first action required to develop a site management plan
 employing natural attenuation. Because each aspect of the site characterization re-
 quires technical proficiency in specialized  areas, discussion will be limited to high-
 lighting particularly usefully or common techniques and approaches. These will be
 generally appropriate for most sites; however, site specific conditions will often mandate
 modifications to assure collection of valid data.                       !

 Site Topography, Geology, and Hydrology
 The surface and subsurface characteristics impact the application of natural attenua-
 tion. Surface contours can influence water infiltration, plume migration, the ability to
 install monitoring wells or treatment systems. The soil type and nature of the aquifer
 material influences the practicality of treatment alternatives, since contaminant distribu-
 tion is directly related to the subsurface geology. For example, fractured rock forma-
 tions provide significant flow pathways for NAPL and dissolved contaminants. Sandy
 soil provides for a higher level of rainwater infiltration which may leach contaminant to
 the groundwater.  Because of differences  in pore space geometry, less residual con-
 tamination is held in sandy soil compared  to silty or clay soils (Gharbeneau et al., 1992).
 Therefore, local geology and topography can provide important insights into contami-
 nant and groundwater behavior.                                     :

 Local! aquifer use, the location of drinking water,  industrial, or agricultural wells, and po-
 tential exposure routes and receptors should  be  identified as part of the overall site
 characterization. These parameters are instrumental for defining the relative health and
 environmental risk associated with the contaminated groundwater.

Aquifer Characteristics
Thorough characterization of the aquifer is essential for predicting the long-term impact
 of the free NAPL,  residual NAPL, and dissolved contaminant.  Several parameters
should be determined using standard techniques. Wells will be required for some tests,
whereas simple monitoring points will be adequate for others.

Groundwater Movement
The movement of groundwater directly affects the progress of natural attenuation;
therefore, defining local aquifer characteristics contributes to an accurate prediction of
the performance of the natural attenuation process.  Hydraulic conductivity of the aqui-
fer is determined using falling or rising head slug tests or actual pump tests. Pump
tests usually give more accurate results, but they also generate relatively large quanti-
ties of water that must be contained and disposed.

The depth to groundwater and the groundwater gradient are also  determined in the
contaminated area.  If a significant change in  gradient or depth is expected based on
local topography and geology, these parameters should be determined in the

                                      6-5

-------
downgradient direction. The groundwater flow rate and direction can be calculated
based on hydraulic conductivity and groundwater gradient data (Example Calcula-
tion 6-1). The risk associated with the contaminated groundwater is evaluated using
these and other parameters.
 Site-Specific Data:

 Hydraulic gradient-a unitless measure of change in depth per change in length. For
 example, an aquifer whose groundwater level changes from 750 feet above sea level
 to 740 feet above sea level over the distance of 100 feet has a hydraulic gradient of:

                           dH  = (750-740) feef =
                           dL       100 feet

 Hydraulic conductivity-a measure of an aquifer's ability to transmit water. Hydraulic
 conductivity is determined from pumping or slug test data. Hydraulic conductivity is
 expressed as length per time, typically as cm/sec.
                                                                  i
 Effective porosity—the percentage of a volume of undisturbed aquifer solids that is
 occupied by water which is available to flow along the hydraulic gradient.  Total
 porosity is the total void volume in a volume of aquifer solids. Depending on the type
 of solids, the effective porosity may be significantly less than the total porosity.

 Calculation of Groundwater Flow Rate
                                       -K dH
                                  v = 	
                                       n. dL
                                        G

 Where:

 v" = average groundwater velocity, units are length per time
 K = hydraulic conductivity
 ne = effective porosity
 dH/dL = hydraulic gradient

                Example Calculation 6-1. Groundwater flow rate.
                                      6-6

-------
Recharge and Discharge
The rate of aquifer recharge and the areas where recharge from the surface and dis-
charge to the surface is likely to occur is determined. These characteristics facilitate a
more accurate determination of plume dilution. Discharge points are considered during
risk evaluations since they represent areas for potential contact with contaminated wa-
ter.

Contaminant Composition, Concentration, Mass, and Properties
Contaminant composition, location of non-aqueous phase liquids (NAPL), location of
dissolved-phase contaminants, and the mass of contaminant present are critical
aspects for understanding the natural attenuation process. Isopach maps showing the
location and thickness of NAPL and isopleth maps showing the location and concentra-
tion of dissolved contaminants are useful visual approximations of subsurface contami-
nation derived from contaminant concentration data.                  ;

The physical and chemical properties of the contaminant also affect the behavior of the
contaminant in the subsurface. Once the composition of the contaminant is known,
properties of individual components are usually available in the published resources.
Particularly useful properties include aqueous solubility, the octanol-water partition co-
efficient, the soil adsorption coefficient, and the Henry's Law  Constant. The application
of these characteristic properties is shown in  various example calculations which follow.

Contaminant Composition
The chemical composition of the contaminant can be determined using several stan-
dard analytical methods. Knowing the chemical composition of the  contaminant permits
changes in composition to be evaluated against time and distance migrated. This infor-
mation is then used to make predictions regarding the attenuation rate, the biodegra-
dation rate, and the change in concentration due to physical  mechanisms.

During site characterization, contaminant chemical composition should be evaluated
with two objectives.  First, the concentration change in chemicals that are likely to bio-
degrade anaerobically should be determined. Second, the change  in chemicals that
are not likely to biodegrade anaerobically should be determined. Concentration chang-
es in the latter can be used to correct for physical losses in the biodegradable com-
pounds. Key target compounds for petroleum fuels that are very slow to biodegrade
anaerobically include 2,3-dimethyl pentane, trimethyl benzenes, and tetramethylben-
zenes ( Cozzarelli et al., 1990, Cozzarelli et al., 1994, Wilson et al.,  1993, Wilson et al.
1994, Wiedemeier).  These compounds have solubility and sorption characteristics
similar to benzene, toluene, ethyl benzene, and xylenes (BTEX) and, thus, they can
serve as practical markers of contaminant movement and BTEX biodegradation. This
approach is the simplest method for calculating dilution, dispersion, and sorption effects
on contaminant concentration (Example Calculation 6-2). When site analytical data do
not contain a suitable conserved marker compound, benzene can be used to normalize

                                      6-7

-------
Site-Specific Data:

One or more site-specific conserved chemicals must be identified. Key characteris-
tics include the consistent presence of the conserved chemical(s) in all or most
groundwater monitoring wells. The conserved chemical(s) must be much less likely
to biodegrade anaerobically than other target chemicals.  The conserved chemical(s)
must have chemical properties similar to target compounds.

Contaminant concentration data from monitoring wells located along downgradient
vectors.

The concentration of a target contaminant is normalized or corrected for abiotic
losses based on the change in concentration of the conserved compound(s):
                                   Mt
                                   —t
                                   M.
Where:                                                          ;
M0 := conserved chemical concentration in the upgradient monitoring well
Mt == conserved chemical concentration in the downgradient monitoring well
R = ratio of change in conserved chemical between the upgradient and the downgra-
dient wells. This value is taken to represent the relative amount of abiotic loss ex-
pected in all chemicals with similar properties.

(If data for more than one conserved chemical are present, an average abiotic loss
ratio "R" can be calculated.)

The abiotic loss factor "R" is used to normalize or correct biodegradable target com-
pound concentrations for the amount of compound removed abiotically.  If the ratio
for the target compound in the downgradient well  is greater than or equal to "R", no
biodegration is occurring in the aquifer. If the target compound ratio is less than "R",
the difference represents the contribution of biodegradation to the overall reduction in
contaminant concentration.  Normalization for abiotic loss can be conducted using
either the upgradient or downgradient contaminant data; however, using the down-
gradient data results in a slightly simpler equation:
Example Calculation 6-2. Application of a conserved chemical's concentration
    to calculate the rate and magnitude of abiotic changes in concentration.


                                    6-8

-------
                      Example Calculation 6-2 (continued)
                                  	
                                   r-.     tnorm
                                   r\
Where:

Ct == measured contaminant concentration in the downgradient monitoring well
Ctnorm = normalized contaminant concentration in the downgradient monitoring well.
This value represents the concentration that would be expected in the well if no
abiotic loss occurred.  (The validity of this approach can be seen by dividing Mt, the
conserved chemical concentration, by "R".)

The biodegradation rate is calculated using the first-order decay equation:
                                  tnorm     o<=
NOTE:  The problem of limited site data often makes the initial evaluation of natural
attenuation challenging, because insufficient data are available to adequately define
intrinsic processes.  When BTEX data are available, examination of site maps and
the benzene-to-toluene (B/T) ratio in each monitoring well often indicates a signifi-
cant shift in the B/T ratio as distance from NAPL or the source area increases. This
is because fresh fuel oils usually have more toluene than benzene, but since toluene
is more susceptible to anaerobic biodegradation than benzene, the ratio shifts to
more benzene than toluene when anaerobic biodegradation is occurring. This shift is
a qualitative indicator of intrinsic biodegradation.

In some cases benzene can be used as a conserved compound to calculate the
abiotic loss of toluene and xylene within the impacted aquifer. Although this ap-
proach is not as rigorous as that described above, it can be used to make a first ap-
proximation of natural attenuation until better data are collected.
                                     6-9

-------
 anaerobic toluene biodegradation and attenuation in BTEX contaminated aquifers.
 However, this approach tends to under estimate biodegradation.

 Contaminant Concentration
 The standard procedure for quantitating a remedial process is reduction in contaminant
 concentration measured with U.S. EPA approved methods. The accuracy of this ap-
 proach for describing the effectiveness of an in situ remedial activity is questionable.
 Contaminant concentration is sensitive to dilution, sorption, and chemical transforma-
 tions of the primary contaminant to other compounds that usually are not detected
 using standard methods.

 Contaminant Mass
 The mass of contaminant in soil and dissolved in groundwater is an important parame-
 ter for quantitating natural attenuation.  Unfortunately, accurate measurement of con-
 taminant mass is difficult and is usually not done with acceptable accuracy. Methods to
 determine contaminant mass involve estimation methods to extrapolate the amount of
 contaminant between monitoring points. Defining the mass of contaminant in a soil
 sample can be accomplished using several analytical methods. Defining the actual
 dimensions and concentrations of contaminant in three dimensions is challenging and
 expensive.

 Soil Core Analysis. One approach is to collect intact soil cores and analyze sections of
 the core to determine the depth distribution of contaminant and the mass; of contami-
 nant in each section (Figure 6-1) (Wilson, 1993). Lateral contaminant distribution is
 estimated based on soil sampling at multiple locations.

 Contaminant Mass and Center of Mass Estimation. Estimation of the amount of con-
taminant present at a site can be estimated using a relatively simple technique known
as the Thiessen Method (Dupont, 1995). Its original use was to calculate the amount of
 rain that fell over an area  using data collected from rain gauges.  The approach is out-
 lined in Figure 6-2.  Using monitoring well data, the method will yield an arbitrary and
 unbiased estimate of contaminant mass as well as the center of mass.  As indicated in
 Figure 6-2, results collected overtime can provide useful insights into the natural atten-
 uation process and the effect of a contaminant source such as NAPL on contaminant
migration and dissolved concentration.

The accuracy of the Thiessen Method contaminant mass estimate is questionable, but
the results are generally precise and can be compared at multiple time intervals. By
combining the contaminant concentration and location data generated from soil core
analysis with the Thiessen Method, the accuracy of the contaminant mass estimate can
be improved.
                                     6-10

-------
                   Location of separate
                   phase contaminant
                   in soil
                                               Dissolved and adsorbed
                                               contaminant
                                               in soil core sample
                                   Soil cores for
                                 segment analysis   =
/  /  / / ^^m  /  /

       Separate Phase
                                                       f

                                                        JVadose Zone Soil
                                                       f i~t   f   •^•'••"i f i  r
        Contaminant
                                   Dissolved
                                 Contaminant
                                     Saturated Soil
Figure 6-1.  Analysis of soil core segments accurately defines location of
separate phase and dissolved phase contaminants.
                                6-11

-------
The Thiessen Method involves the identification of specific sam-
pling locations within a sampling network, and the determination of
associated areas  based on  the construction of polygons sur-
rounding the sampling points (monitoring wells in this example).

Polygons are constructed in the following manner:

Step 1.  Define the outer boundary of the contaminant plume.

Step 2.  Connect all monitoring wells to all adjacent wells to create
a group of triangles.

Step 3.  Place a perpendicular line at the bisection point of each line
used to connect monitoring wells.  Extend the perpendicular lines to
intersect one another.

Step 4.  Connect the intersection lines to form polygons.

The area within each polygon is calculated and used to calculate
the volume of soil or groundwater within the area assigned to each
monitoring well. For quality assurance, the total area of the plume
and the summed areas of the polygons should agree  within
5 percent.

The contaminant concentration in each well is used  to determine
the total mass  of contaminant present  within  the  polygons
associated with the monitoring wells. The center of mass (centroid)
can also be determined  using this approach. Trends observed in
the results collected with time can reveal the occurrance of natural
attenuation.

Summary of  concentration  mass and center of mass trends as
indicators of natural attenuation (Dupont, 1995).
Contaminant Mass
Increasing
Constant
Constant
Decreasing
Decreasing
Center of Mass
Moving downgradient
Moving downgradient
Stable
Moving downgradient
Moving upgradient
Interpretation
Continuous source, unstable
plume, contaminant migration
Finite source, plume
migrating, minimal natural
attenuation
Continuous source, stable
plume, contaminant
attenuation
Finite source, plume
migrating, contaminant
attenuation
Finite source, plume
attenuation, rapid contaminant
attenuation
Steps 1 and 2

 -^\.
  Monitoring Points
                  Plume Periphery
StepS
                Perpendicular
                Bisectors
                                                                  Step 4
                                                                                  Area Associated With
                                                                                  Each Monitoring Point
Figure 6-2.  Contaminant mass estimation using the Thiessen Method for area
assignments.
                                             6-12

-------
Contaminant Physical and Chemical Properties
Physical and chemical properties of contaminants affect their distribution and move-
ment in groundwater and soil gas. The octanol-water partition coefficient, Henry's Law
Constant, and aqueous solubility of contaminants can be used to estimate the interac-
tion of the contaminant with air, water, and soil. The accuracy of these estimates is
usually questionable; however, general ranges can be useful for predicting the behavior
of the contaminant (Example Calculations 6-3 and 6-4). Retardation coefficients that
reflect the relative mobility a contaminant in an aquifer are useful for comparing the ex-
pected plume migration rate with observed plume movement (Example Calculation 6-5).

Biological Monitoring Parameters
Biological monitoring parameters are indicators of microbiological activity in the subsur-
face. In all cases, microbial activity is the most likely if not the only explanation  for the
observed changes.  Therefore, changes or activity observed in a contaminated  area
that do not occur or do not occur to the same magnitude in uncontaminated soil can
often be attributed to biodegradation.

Microbial Respiration
Microbial respiration is the biochemical process that leads to the oxidation of reduced
organic carbon.  Examples of reduced organic carbon include every organic compound
that contains carbon and hydrogen in its chemical composition. Demonstrating aerobic
respiration is relatively simple in the vadose zone. However, it is not very practical in
the saturated zone. Therefore, laboratory evaluations may be used as a surrogate to in
situ demonstrations.

The principle benefit of demonstrating  microbial respiration in laboratory studies is that
they indicate that microbial respiration  can occur in the contaminated area. Although
the observed respiration rate from laboratory microcosms is frequently higher than that
observed in situ, the laboratory microcosm can set an upper limit for in situ respiration.
Table 6-2 outlines a simple microcosm study that evaluates aerobic and anaerobic
respiration and abiotic contaminant loss.

Conversion/Consumption of Respiratory Substrates
In situ observation of the results of microbial respiration is often obvious when the con-
centration of respiratory substrates (electron acceptors) are compared within and out-
side of a contaminated area. The biodegradation potential of an aquifer [also known as
the "expressed biodegradative capacity" (Wiedemeier, 1994)] can be estimated using
the concentration of respiratory substrates and products (Example Calculation 6-6).

In addition to estimating  the biodegradation potential of the aquifer, respiratory sub-
strate and product concentrations are critical data used to support the occurrence of
intrinsic bioremediation.  The observations described below are indicative of intrinsic
                                      6-13

-------
Site-Specific Data:

Type of contaminant

Mass of NAPL present

Mass fraction of target contaminants in the NAPL

The maximum amount of contaminant that will dissolve from the NAPL into ground-
water is calculated using the fuel-water partition coefficient and the concentration of
individual chemicals in the NAPL. The fuel-water partition coefficient is given by:

                         l°9Kfw = 6.099  -  1.15 logSt

                           K^ = io6-099 - 1.15 logSt



Where:

Kfi,, = fuel-water coefficient, the ratio of the concentration of a chemical in the NAPL
to the concentration dissolved in water.

The concentration of a chemical in NAPL is calculated by multiplying the mass frac-
tion of the chemical by the density of the NAPL.
                                    = Ft Pf
Where:
Ct = volumetric concentration of chemical in the NAPL
Ft = mass fraction of the chemical on the NAPL
Pf = density of the NAPL                                         :

The amount of chemical that can dissolve into the groundwater is:
                                                               i,
     Example Calculation 6-3. Contaminant dissolution from nonaqueous-
                           phase liquids (NAPL).	
                                   6-14

-------
                    Example Calculation 6-3 (continued)
                                 C  =
                                  w
Where:


Cw == concentration of the target chemical in the groundwater.



From Bruce et al., 1991.
                                   6-15

-------
The diffusive movement of contaminant in NAPL into the groundwater constrains the
time required for natural attenuation to restore an aquifer. So long as NAPL is pres-
ent and contributing dissolved contaminants to the groundwater, the aquifer will
continue to be contaminated.  The rate of dissolution is limited by the diffusion of
contaminant into the water.  Ideally, all NAPL would be physically extracted from the
soil; however, in practice only 40 to 60 percent is typically removed. This leaves a
large amount of NAPL in the ground which will continue to contaminate groundwater.
The diffusion of target contaminants in the residual NAPL becomes an important pa-
rameter for estimating the time required to remediate the aquifer because the water-
will be constantly recontaminated until the NAPL is depleted.

The diffusivity of the target contaminant into water can be calculated using several
methods (Handbook of Chemical Property Estimation Methods, Chapter 17). The
Hayduk and Laudie method is presented because of its simplicity and small absolute
average error (5.8 percent).

                             D   =  13.26  x 10'5
                              BW      1.14  ,,0.589
                                    "w    VB
Where:

DBW = diffusivity of compound "B" into water "W"
nw = viscosity of water
VB = molar volume calculated using the method of LeBas (see Table 17.5, Handbook
of Chemical Property Estimation Methods, for details).

Using the diffusivity value generated above, the mass flux of contaminant B into
groundwater can be calculated based on Pick's Second Law of Diffusion.

                                      (CB1 ~ CB2)
Where:

NB = mass flux of compound B into water                            ;
DBW = diffusivity of B into water
(CB1 - CB2) = concentration gradient of compound B

     Example Calculation 6-4.  Mass flux of contaminant in NAPL to water.

                                   6-16

-------
                     Example Calculation 6-4 (continued)
(z2 - z.,) = distance
ft =  tortuosity, a soil property typically ranging from 2 to 100 as applied above (Hand-
book of Chemical Estimation Methods and Jury et al., 1991)

The results of the mass flux calculation are in units of mass/(area * time); therefore,
the time required for the available contaminant to diffuse into the groundwater can be
estimated.
                                  T =
                                      ANB
Where:

T = time to complete diffusion of compound B into water
MB = total mass of compound B available to diffuse into water
A = area covered by NAPL or separate-phase compound B
                                   6-17

-------
The retardation coefficient for a chemical species describes the movement of a dis-
solved chemical relative to the advection movement of groundwater.  Groundwater
contaminants frequently migrate at a slower rate than the groundwater itself. Re-
tardation results in exposure of dissolved contaminants to more electron acceptors
than would otherwise occur if contaminant and groundwater moved in unison. The
retardation coefficient also can affect the calculation of anaerobic biodegradation
rates when data from monitoring wells lying along a flow vector are used. If contam-
inant takes twice as long as the groundwater to travel from one monitoring well to the
next/the biodegradation rate will be inaccurately estimated using groundwater veloci-
ty to calculate contaminant travel time between the wells. The retardation coefficient
for a dissolved chemical may be estimated as shown:
                             ne
Where:

R = retardation coefficient
Pb = bulk density of the aquifer (mass per volume)
Kd = distribution coefficient                                         ;
ne = effective porosity

The distribution coefficient describes the distribution of dissolved and sorbed con-
taminant.  In simple terms, the distribution coefficient indicates the amount of
contaminant that is dissolved and the amount that is stuck to the soil in the aquifer.
The distribution coefficient is the product of the mass fraction of total organic carbon
in the aquifer and the soil adsorption coefficient.
                                  Kd = Kocfoc
Where:
   = soil adsorption coefficient (usually determined using published values, Jeng et
al., 1992)
foc = mass fraction of organic carbon in the soil expressed as mass of organic carbon
per mass of soil                                                   ;

               Example Calculation 6-5. Retardation coefficient.
                                    6-18

-------
Table 6-2. Simple Microcosm Study to Define Aerobic and Anaerobic
Biodegradation.	
              Replicates
  Treatment   (minimum)                        Description
 Aerobic
 Anaerobic
 Abiotic loss   3
Microcosm vessel containing contaminated groundwater.
Use a manometer or respirometer to measure oxygen
consumption. Analyze for contaminant concentration and
carbon dioxide produced overtime.

Establish microcosm vessels in an anaerobic chamber,
fill the vessel to eliminate headspace, sparge water with
inert, oxygen-free gas, or add an oxygen scavenger to
generate and maintain anaerobic conditions.  Depending
on the setup, multiple vessels may be required for inter--
mediate contaminant and electron acceptor analysis.
Additional vessels are required if denitrification, sulfate
reduction, or methanogenesis are evaluated.

The loss of contaminant due to laboratory manipulations
is quantitated using an abiotic control treatment. Micro-
cosms are established as described for either of the other
treatments.  The control is incubated at 4°C and sampled
like the other instruments. In some cases, before and
after treatment sampling is adequate to define abiotic
losses.  Inhibitors such as azide or mercury may be used
to inhibit biological activity.
                                     6-19

-------
Site-Specific Data:

Maximum and average contaminant concentration within the plume.

Electron acceptors and respiratory product concentrations in the plume and at back-
ground locations. Electron acceptors include dissolved oxygen, nitrate, and sulfate.
Respiratory products include ferrous iron (Fe II) and methane.

The following calculations are based on benzene. Minor disparities will exist if com-
pounds other than benzene are the major contaminants; however, these differences
are generally insignificant relative to the larger error associated with analytical data,

Constants:

Mass ratio of benzene to electron acceptor for complete mineralization:

Benzene/Oxygen = 1/3.1 = 0.32

Benzene/Nitrate = 1/4.8 = 0.21

Benzene/Sulfate = 1/4.6 = 0.22

Mass ratio of benzene to respiratory products for complete mineralization:

Benzene/Iron (Fe II) = 1/15.7 = 0.064

Benzene/Methane = 1/0.8 = 1.25

The assimilative capacity of the aquifer (the amount of contaminant that can be bio-
degraded) is determined by the general equation:

                         °o + Nc+ +/c + Sc + MC = Tc

Where:

Oc = assimilative capacity of dissolved oxygen
Nc = assimilative capacity of nitrate


       Example Calculation 6-6.  Biodegradation potential of an aquifer.	
                                    6-20

-------
                      Example Calculation 6-6 (continued)
!c = assimilative capacity of iron based on the amount of ferric iron (Fe II) converted
to Fe II
Sc = assimilative capacity of sulfate
Mc == assimilative capacity of methane based on the amount of methane produced

The assimilative capacity of each electron acceptor or respiratory product is calcu-
lated by:           .
                                           Oc
Where:

B/O2 = the benzene-to-nitrogen mass ratio
[C] = background concentration of nitrate
                                     [C] -
The assimilative capacity of each electron acceptor or respiratory product is calcu-
lated in like manner.  The total assimilative capacity is calculated by summing the
assimilative capacities of each individual electron acceptor or product. If the total
assimilative capacity is greater than the maximum contaminant concentration, intrin-
sic biodegradation has the potential to remediate the aquifer.  On the other hand, if
the total assimilative  capacity is less than the maximum contaminant concentration,
intrinsic biodegradation is less likely to be successful.
                                     6-21

-------
 biodegradation and are important aspects for evaluating the potential of successful
 intrinsic bioremediation at a site. Because of the energy derived from each substrate,
 they are preferentially used in the order shown in Figure 6-3.  In addition to preferential
 utilization, respiration tends to be exclusive. For example, in the presence of oxygen,
 other respiratory substrates are not used. Similarly, when oxygen is depleted and ni-
 trate is available, denitrification will prevail with no iron and sulfate reduction or meth-
 anogenesis.  This pattern of respiratory exclusion continues as high energy respiratory
 substrates are expended.

 Oxygen Consumption
 Groundwater undergoing intrinsic bioremediation will usually contain much less dis-
 solved oxygen than a nearby "clean" groundwater. An area of oxygen depleted ground-
 water usually coincides with the dissolved contaminant plume and it often extends be-
 yond the contaminant plume in the downgradient direction. The difference in dissolved
 oxygen between "clean" and contaminated groundwater is a useful indicator of intrinsic
 biodegradation. Three to 3.5 pounds of oxygen are required to completely biodegrade
 one pound of hydrocarbon (Wiedemeier et al., 1994).                   :

 Nitrate Reduction
 The nitrate content of groundwater varies with local land use practices. Agricultural
 areas, groundwater downgradient of land fills, and industrial areas often have elevated
 nitrate concentrations in the groundwater. However, within areas with measurable ni-
 trate, the groundwater within a contaminant plume will often have a much lower nitrate
 concentration than surrounding "uncontaminated" groundwater.  This concentration
 change indicates nitrate reducing activity. The absence or much lower rate of nitrate
 reduction in clean groundwater indicates that the presence of contaminant is required to
 support nitrate reduction.  About four pounds of nitrate are required to biodegrade one
 pound of hydrocarbon (Ehrlich, 1981, Kuhn etal., 1988, Major etal., 1988, and
Wiedemeier et al., 1994).

 Iron Reduction
 Oxidized iron is insoluble  and its presence is indicated by the red coloration of oxidized
soil. Soil within contaminant plumes is often reduced. The biological conversion of oxi-
dized  iron, Fe(lll), to reduced iron, Fe(ll), can support hydrocarbon biodegradation. Iron
 reduction is not an efficient biochemical process  requiring about 42 pounds of Fe(lll)
per pound of hydrocarbon degraded.  Nevertheless, iron reduction is an important
process because of the abundance of iron in the soil. Unfortunately, the speciation and
analysis of Fe(lll) and Fe(ll) is very challenging and usually not accurate.  Field mea-
surement of Fe(ll) in groundwater samples using colorimetric test kits usually yields
satisfactory results (Ehrlich, 1981, Lovely, 1991, and Wiedemeier et al., 1994).
                                      6-22

-------
   Redox
  Potential
    (mv)
  Respiration
   Pathway
Electron
Acceptor
    400
    200
       0
   -200
   -400
               Aerobic
               Respiration
Nitrate
Reduction

Iron and Manganese
Reduction
iSuIfate Reduction

Methanogenesis
   N03


   Mn(IV)
   Fe(III)
   SO4
Figure 6-3. Electron acceptors for common microbial respiration pathways and
approximate oxidation/reduction potential (Redox or Eh) where each pathway
occurs. (Atlas and Bartha, 1981)
                        6-23
                        •-c-

-------
Sulfate Reduction
Sulfate reduction can contribute to hydrocarbon biodegradation in the absence of
oxygen, nitrate, and iron [Fe(lll)].  Sulfate reducing conditions are characterized by a
low oxidation reduction potential and the presence of sulfide.  Approximately 4.8
pounds of sulfate are required to biodegrade one pound of hydrocarbon (Ehrlich, 1981,
Edwards et al.,  1992, and Wiedemeier et al., 1994 ).

Methanogenesis
Under strictly anaerobic conditions and low redox, biodegradation of hydrocarbon leads
to the production of methane.  About 0.8 pounds of methane are  produced during the
biodegradation  of one pound of hydrocarbon (Wiedemeier et al.,  1994, Wilson et al.,
1993, Grbic-Galic and Vogel, 1987).

Manganese Reduction
Oxidized manganese (Mn+4) can be reduced by microbial respiration (Ehrlich, 1981,
Lovely, 1991, and Wiedemeier et al., 1994 ). As seen in Figure 6-2, manganese reduc-
tion occurs under conditions similar to those that support iron  reduction. There is some
evidence that manganese reduction can support hydrocarbon biodegradation (refer-
ence); however, the difficulty in speciating manganese ions usually precludes any
serious attempts to evaluate the effect of manganese on intrinsic biodegradation. A
practical approach to dealing with manganese reduction is to evaluate the other (easy
to analyze) respiratory substrates or products.  If contaminant degradation appears to
be occurring but can not be accounted for using more common respiratory substrates,
manganese should be considered as a possible substrate supporting biodegradation.

Biodegradation Rates
Biodegradation  rates may be determined using field data or laboratory data.  Biodegra-
dation rates are usually calculated with satisfactory results using a first order rate equa-
tion (Example Calculation 6-7). Laboratory data have the unfortunate disadvantage of
usually overestimating biodegradation rates.  Collection of biodegradation rate data in
the field with the accuracy to calculate a degradation rate in the absence of other phys-
ical loss mechanisms is challenging. The calculation of biodegradation rates corrected
for physical loss using a conserved compound that does not biodegrade anaerobically
is shown in Example Calculation 6-2.  Laboratory derived biodegradation rates may
also be used to define the percentage of the overall loss that can be attributed to bio-
logical activity.

Physical/Chemical Monitoring Parameters
Physical/chemical monitoring parameters reveal abiotic processes that can result in
contaminant dissipation. Unless these parameters are accounted for in the overall con-
taminant mass balance, biodegradation may be credited for more contaminant loss
than actually occurs. Because most of the physical loss mechanisms do not result in
contaminant degradation, these processes may not be acceptable approaches for site

                                     6-24

-------
Site-Specific Data:

Groundwater velocity -- determines the minimum time required for dissolved con-
taminants to travel between two points.

Contaminant concentration data ~ contaminant concentration must be determined in
monitoring wells positioned along a vector in the direction of groundwater flow. Con-
taminant concentration data from two or more points is required. The points may be
temporal or spatial. Selecting useful data can be difficult. Sequential analytical data
for the same well or data from two wells separated by a known groundwater travel
time may provide useful data, depending on site conditions described below.

Speciation of dissolved contaminants - contaminants monitored should include
target compounds as well as one or two compounds that do not readily degrade
anaerobically.  For refined petroleum products, 2,3-dimethyl pentane, trimethylben-
zenes, and tetramethylbenzenes have been used to evaluate the rate and magni-
tude of nonbiological contaminant loss. The choice of an appropriate conserved
chemical is determined by conducting gas chromatography and mass spectroscopy
on groundwater samples near or in the source and at a few locations downgradient.
Chemicals that persist in the groundwater at elevated levels-presumably due to
poor anaerobic biodegradability-and that have chemical properties similar to target
compounds can serve as conserved indicators of contaminant dispersion, dilution,
volatilization, and sorption.

Sequential contaminant data covering several months to several years

Distance between monitoring wells - used to calculate hydraulic retention time be-
tween wells

Seasonal groundwater elevation in each monitoring well - seasonal fluctuations can
cause significant changes in groundwater flow direction and velocity.

A first-order decay equation is usually an accurate model for contaminant bio-
degradation:

                                 Ct = CoQ-kt
 Example Calculation 6-7. Biodegradation rate and target compound half-life.
                                    6-25

-------
                       Example Calculation 6-7 (continued)

Where:                                                             ;
Ct = concentration of contaminant at time T
C0 = concentration of contaminant at time "zero"
t = lapsed time from C0 to Ct
-k = biodegradation rate constant

From the biodegradation rate constant "k", the half-life of a contaminant can be
calculated:

                                    t
                                    1/2
                                          k

Where:                                                             '•
t1/2 = contaminant half-life, the time required for the concentration of the contaminant to
decline by one-half

Calculation of "t" when contaminant concentration data are collected from two wells
located along a downgradient flow vector:
                                                                   i

                                    . t = -
Where:
x = distance between the two monitoring wells (sample collection points)  ;
v = groundwater velocity or seepage velocity.  Groundwater velocity can be a tricky
variable to calculate.  The hydraulic gradient may change dramatically with the season.
Some aquifers actually change flow direction.  Such changes in the hydraulic gradient
seriously affect groundwater velocity calculations and must be considered when calcu-
lating the groundwater retention time between two points.                \

NOTE: When NAPL is present and contributing dissolved contaminant to the ground-
water, the use of groundwater data collected from the same well, but at  different times,
is not useful for calculating biodegradation rates, because the groundwater is continually
replenished by newly dissolved contaminant. If NAPL is the  present, this type of analyt-
ical data can be used to calculate biodegradation rates, because all of the groundwater
should be experiencing approximately the same conditions without the input of fresh
contaminant.
                                       6-26

-------
remediation.  Chemical processes usually do not result in contaminant degradation;
however, abiotic chemical reactions in soil and groundwater may convert a contaminant
into another compound. Chemical transformations may be beneficial if they result in a
compound that is less hazardous or more biodegradable.

In addition to accounting for abiotic contaminant loss mechanisms, chemical monitoring
is the ultimate parameter that determines when a site is remediated.  Therefore, moni-
toring to determine contaminant concentration  reduction, contaminant mass loss, and
possibly the appearance of degradation products is critical to define the success of
natural attenuation.                                                ,

Dilution by Diffusion and Dispersion
Dilution is an inevitable physical event. The significance of dilution is site dependent
and difficult to predict with certainty; therefore,  a measure of the dilution rate is useful
for defining physical changes in contaminant concentration.  Dilution affects the con-
centration of a contaminant, not the mass of contaminant. Diffusion and dispersion are
the most common mechanisms of contaminant dilution; however, in shallow aquifers,
infiltrated rain water can also be a significant contributor.

Diffusion is a minor component of dissolved contaminant dilution because the tortuosity
of soil greatly reduces the diffusion of rate.  Although diffusion is usually insignificant in
the dilution of dissolved contaminant (Freeze and Cherry, 1979), it may be one of the
principle factors in the dissolution of contaminant from NAPL into the groundwater. The
rate of dissolution depends on the major rate limiting components of natural attenua-
tion, namely, how long is required to expend the source of dissolved contaminant in
separate phase contaminant. Example Calculation 6-4 shows one  approach for de-
fining diffusion-limited contaminant dissolution.

Dispersion is a more significant physical process resulting in contaminant dilution.  It is
the movement of contaminant in a direction that is not the flow vector of groundwater.
Dispersion commonly results in the side-to-side spread of a contaminant plume along
the downgradient flow path.  Like diffusion, dispersion does not result in a change of
contaminant mass; although  it does result in a concentration change. Dispersion is
often estimated by examining plume dimensions in the downgradient and cross-
gradient directions.  The ratio (cross gradient movement versus downgradient move-
ment) represents the dispersion coefficient. Dispersion can be directly measured using
an in situ tracer; however, this approach is time consuming and expensive.  Dispersion
can also be approximated using published values for defined aquifer conditions.  The
approximated value can be further refined by adjusting it iteratively in a groundwater
transport model until the model matches the in  situ condition.

The net result of both diffusion and dispersion is a reduction in contaminant concen-
tration. The combined effect of dilution and dispersion can be measured using the

                                     6-27

-------
single; procedure. The contribution of each to contaminant concentration reduction
does not need to be segregated; therefore, a single measurement is used to describe
the effect of diffusion and dispersion on the movement of dissolved contaminant.

Volatilization
Volatilization results in a decrease in contaminant mass from one matrix and an in-
crease in another matrix. For example, loss of volatile compounds from groundwater is
the movement of volatile compounds from the water into unsaturated soil and possibly
the atmosphere. Volatilization is controlled by the Henry's Law Constant, octanol-water
partition coefficient, solubility, and density of the contaminant.  Volatilization is most
prominent in the capillary fringe and the groundwater fluctuation space when the con-
taminant is less dense than water or moves along the saturated/unsaturated soil
boundary.

The contribution of volatilization is difficult to predict during natural attenuation, since
multiple conditions interact during volatilization.  Volatilization is probably a minor con-
taminant loss mechanism, unless a soil gas or vadose zone soil treatment technology is
used in conjunction with natural attenuation.  An example is the operation of a biovent-
ing or soil vapor extraction system in the vadose zone above a contaminated aquifer.

Sorption and Retardation
"Sorption" is a general term used to describe adsorptive and absorptive processes that
result in the partitioning of contaminant from the aqueous or dissolved phase to the
solid phase or soil (Freeze and Cherry, 1979).  In simple terms, contaminants stuck to
soil are no  longer dissolved in the groundwater.  An adsorption coefficient can be esti-
mated using the octanol-water partition coefficient and contaminant solubility.  The ad-
sorption coefficient can also be determined experimentally by observing the amount of
contaminant in the soil and water. This type of experimental information usually fits a
Freundlich  adsorption isotherm which can be used to  calculate the adsorption  coeffi-
cient.

The sorptipn and desorption of dissolved contaminant results in contaminant migration
that is slower than the groundwater velocity.  The retardation coefficient describes the
movement of dissolved contaminant relative to groundwater (Freeze and Cherry, 1979).
Example Calculation 6-6 further discusses the retardation  coefficient and shows how to
estimate it. The retardation coefficient is important because it more accurately defines
the retention time of a contaminant between two points along a flow vector.  This
information is central to the calculation of anaerobic biodegradation ratesj

Contaminant Mass Loss
Determining the mass of contaminant lost during natural attenuation is a challenging
but important activity. One of the greatest obstacles associated with measuring
contaminant mass loss is accurately determining the original mass of contaminant.

                                      6-28

-------
This estimate is made by determining the volume of groundwater in the contaminated
area, the contaminant concentration, the three dimensional contamination distribution,
soil porosity, soil contaminant concentration, and the depth of contaminated soil. Mass
loss is determined by comparing the initial mass to the mass remaining at the time of
subsequent contaminant mass estimates. The Thiessen Method discussed above is a
potentially useful and relatively simple approach for tracking mass loss (Dupont, 1995).

Contaminant Concentration Reduction
Reduction in contaminant concentration is the simplest measurement to make when
evaluating natural attenuation, because all the widely accepted analytical methods are
designed to report contaminant as a concentration.  Accepted methods are available for
most common groundwater contaminants. Analytical methods specific for target con-
taminants should be used. General analyses such as Total Petroleum Hydrocarbon
(TPH or TRPH) (U.S. EPA Method 418.1) should be avoided, because the results do
not provide data for individual contaminants and the methods are prone to interferences
and ambiguities.

The use of typical monitoring wells tend to give artifactual results that do not represent
the actual conditions in the aquifer.  The long screened intervals in most monitoring
wells and the non-uniform flow of groundwater into a monitoring well during bailing often
result in samples that poorly reflect actual aquifer water quality.  Groundwater monitor-
ing points positioned in contaminated zones or narrowly screened monitoring wells are
better alternatives to typical monitoring  wells, because they are installed to provide
information within the vertical limits of the contaminated area.

Appearance of Degradation Products
Organic contaminants degrade through a series  of steps, with each step resulting in  a
intermediate product that is not the original compound nor the final product (Atlas,
1983, Rochkind-Dubnisky,  1987). Sometimes these intermediates are so transient that
they can not be detected or they do  not accumulate to a detectable level. After the first
few degradation steps, the products are usually common to several biodegradation
pathways and cannot be correlated to contaminant degradation.

Quantitation of intermediate biodegradation products is analytically challenging, be-
cause the compounds are typically more hydrophilic than the parent contaminant and
are, therefore, more difficult to extract and analyze.  Fatty acids are potentially quanti-
fiable intermediates of petroleum biodegradation.

Defining an Efficient and  Cost-Effective Monitoring Plan for Natural Attenuation
Several approaches can be taken to minimize the cost of long term monitoring of natu-
ral attenuation based  corrective action plans. Once the plume has been delineated and
its behavior has been evaluated, the following issues are important for long term moni-
toring:

                                     6-29

-------
      •     "Sentry wells" may be installed in an intermediate point between the cur-
            rent downgradient edge of the plume and the nearest receptor, property
            boundary, or "point of compliance" well.  Sentry wells are "clean" ground-
            water monitoring wells that are used to indicate movement of the con-
            taminant plume.  If sentry wells become contaminated,  more aggressive
            remedial technologies are invoked to deal with the contamination; other-
            wise, natural attenuation continues to be the environmentalmanagement
            strategy.

      •     The contaminant concentration must be periodically evaluated; however,
            since plume migration is limited, the sentry wells are protecting down-
            gradient receptors, and plume dimension and concentration is well docu-
            mented, the frequency of contaminant analysis can be minimized.  Simi-
            larly, the number of samples can be reduced to the minimum number that
            will confirm satisfactory performance of the natural attenuation remedial
            plan.  There is no need to remap the entire contaminant plume at each
            sampling event.  Sampling events should be used to confirm treatment
            with a negotiated confidence level.

      •     In addition to contaminant concentrations, groundwater parameters
            should be monitored. The choice of which parameters to monitor should
            be determined from the initial site assessment.  For example, if no  nitrate
            was detected in the groundwater during the site assessment, nitrate would
            be a useless monitoring parameter.  Similar arguments can be made for
            each groundwater chemistry parameter.  If the parameter was not initially
            useful in defining the natural attenuation process, it need not continue to
            be analyzed. Field data on water table elevation is an important parame-
            ter that should always be included in the monitoring  program.

Modeling Natural Attenuation
Predicting the long-term results of natural attenuation relative to aquifer restoration,
plume migration, further groundwater contamination, and projected remediation time is
often a critical component in a defining a natural attenuation remedial action plan that is
acceptable to regulatory agencies and the public. Two general approaches are usually
taken to model the future performance of natural attenuation. These are analytical and
numerical modeling.

Analytical Modeling
Analytical modeling relies on relatively simple calculations of groundwater flow, con-
taminant dispersion, adsorption, retardation, partitioning into groundwater, biodegrada-
tion, and abiotic transformations of the contaminant. Site specific values for each of the
above parameters are used to define the likely migration and attenuation of the con-
taminant plume. The accuracy of the prediction is dependent on the  accuracy of the

                                     6-30

-------
defining parameters. Attenuation rates are often modeled using chemical-specific data
one chemical at a time.

In at least one case, multivariate statistical analysis was used to model intrinsic biodeg-
radation of BTEX in an aquifer (Tan, 1994).  The benefit of this approach is that it can
evaluate all pertinent site data as a whole and derive an overall degradation rate for a
multicomponent contaminant plume.  Further refinements of this and other statistical
approaches may lead to a new generation of analytical tools for evaluating and predict-
ing natural attenuation.

If enough site data are available to directly calculate abiotic contaminant losses and a
biodegradation rate, most of the other parameters become unimportant since they con-
tribute to abiotic loss. The octanol-water partition coefficient and the diffusion mass flux
are of particular value because the continued addition of contaminant to the ground-
water from NAPL is an important process governing the rate of remediation. As long as
more contaminant is being added to the aquifer from a source area, the remediation
cannot be completed although plume dimensions should reach an equilibrium based on
the biodegradation rate and the dissolution rate or mass flux into the groundwater.

The mass flux is controlled by the diffusion coefficient and the partition coefficient of the
contaminant.  Example Calculation 6-4 shows how the mass flux can be estimated. As
suggested above, aquifer restoration cannot be completed until all the separate phase
(NAPL) contaminant has been physically removed or dissolved into the water.

Numerical Modeling
Numerical modeling employs computer programs to simulate the behavior of an aquifer.
The most widely used numerical model for aquifer bioremediation is BIOPLUME II  (Rice
University). This computer model requires the input of many of the parameters dis-
cussed above. In addition to simulating current conditions, the model also predicts
what will happen to the contaminant plume in the future. This aspect of the numerical
model is extremely valuable for weighing the benefits and risks of natural attenuation,
defining the location of sentry wells, and approximating the time required to complete
the aquifer remediation.

Computer models are constantly evolving to include more features, improve their accu-
racy, and expand their flexibility.  A major revision of BIOPLUME II is underway that will
permit much more precise definition of anaerobic processes within the dissolved con-
taminant plume.  Other aquifer bioremediation computer models include BIO1D
(GeoTrans, Inc., Washington, D.C.) and BIOTRANS (Environmental Systems & Tech-
nologies, Inc., Blacksburg, VA) (Rafai, 1993).  Each model has useful features and may
be pertinent for a given site.  Because of the subtleties of computer modeling, potential
modelers are referred to software user's manuals for detailed instructions on how to use
each model. Novice users are strongly encouraged to seek the guidance of an

                                     6-31

-------
experienced groundwater modeler.  Incorrectly applied or inaccurately calibrated com-
puter models will provide faulty output that can lead to false expectations for natural
attenuation.

Data Presentation
Integrating monitoring data into clear and concise figures and tables is very useful for
showing the interrelationships between various data and the attenuation process.
Particularly important items include:

      •     The ratio of compounds not subject to anaerobic biodegradation to
            degradable ones

      •     Time versus concentration  comparisons

      •     Calculated biodegradation  rates and mass biodegraded  based on con-
            served marker concentrations

      •     Plume movement, size, and contaminant concentration

      •     Groundwater flow rate and  direction, especially seasonal fluctuations

      •     Groundwater chemistry
                  Dissolved oxygen
                  Nitrate
                  Sulfate and sulfide                              \
                  Methane

Numerous other parameters, calculations, and observations may be included in natural
attenuation reports.

One of the most convincing ways to display natural attenuation data employs super-
imposed contour maps showing the following relationships:

      •     Change in contaminant concentration with time

      •     Relationship between contaminant  concentration and electron acceptor or
            respiratory product  concentration such as dissolved oxygen; nitrate, sul-
            fate, or methane concentration superimposed on the contaminant contour
            map

      •     Relationship between the conserved marker compound and a target  com-
            pound
                                     6-32

-------
      •      Superimposed contour maps of various target compounds.

A simple way to express monitoring well data is to use pie charts positioned on a site
map at each well location. The  pie chart shows the relative chemical composition of
the groundwater sample.  Shifts in relative concentration are easily visualized as the
slices of the pie change in size.  The magnitude or level of contamination in each well is
conveyed by the diameter or area of the pie. During natural attenuation, the size of the
pie is expected to decrease with distance from the source and the chemical composi-
tion should shift to favor compounds that have a slower anaerobic biodegradation rate.
Figure 6-4 is a simplified example of this type of chart.

Regardless of how results are conveyed, the data presentation should accurately por-
tray the natural attenuation process. Important decisions supporting, rejecting, or dis-
continuing natural attenuation will be made from site data.  Inappropriate or erroneous
interpretations of site data will have serious implications for the natural attenuation
process as time passes and performance objectives are not met. Therefore, conserva-
tive, well-documented assessments of candidate sites are essential before decisions
are made to proceed with natural attenuation.  Armed with a thorough understanding of
site conditions, a critical evaluation and presentation of site data, and an appreciation
for public and regulatory sensitivities, natural attenuation can be embraced as a viable
remedial alternative at many sites.

Rigorous evaluation of natural attenuation during aquifer remediation will provide assur-
ance that the process is working. The final result is successful aquifer restoration and
confidence that natural attenuation and intrinsic bioremediation are reliable remedial
alternatives that economically protect human health and the environment

Applying the Principles  of Natural Attenuation for Environmental Management,
Risk Reduction, and Remediation

Natural attenuation is a powerful process that can, with time, result in site remediation.
Natural attenuation faces  three obstacles that will hinder its use as a viable environ-
mental risk management strategy. Recognition of these obstacles and conscious effort
to avoid stumbling over them will  assure responsible and reliable use of natural attenua-
tion.

      •     Overuse.  Natural attenuation will effectively reduce the risk associated
            with contaminated groundwater and even result in aquifer restoration in
            some fraction of contaminated aquifers. Some of the suitable sites will
            also be easily and rapidly remediated using other technologies. When
            cost and timing do not favor natural attenuation, especially at small or
            geologically simple sites, other, more aggressive, technologies may be
                                     6-33

-------
        -  CONTAMINANT
        s>  SOURCE AREA
Benzene
Toluene
Xylenes
Conserved Marker
Figure 6-4. Simplified groundwater elevation and contaminant content
and composition figure showing relative contaminant composition and
content in three monitoring wells.
                              6-34

-------
             preferable to natural attenuation(GW citation). In such cases, natural at-
             tenuation is probably not the best environmental management choice.

       •      Misapplication.  Emerging environmental technologies are often indiscrim-
             inately applied to every site as though the technology can handle any
             problem. This unfortunate practice has compromised the perception of
             many useful technologies including air sparging, soil vapor extraction, soil
             flushing, thermal desorption, and nearly every biological treatment tech-
             nology.  Because of the potential cost savings implied with natural atten-
             uation, this approach will be particularly susceptible to misapplication.

       •      Skepticism. Skeptical, uninformed, or inflexible regulatory agencies can
             thwart the use of natural attenuation at ideal sites.  Cooperation and
             flexibility among all involved parties and realization that even proven
             technologies were once innovative and unproven will help ease the way
             for applications of natural attenuation at suitable sites.

Evaluating, Selecting, and Monitoring Natural Attenuation for Site Remediation
The following outline suggests a  logical progression of data collection, evaluation, and
interpretation for quantifying and applying natural attenuation. The outline is intended
to highlight some of the steps that are usually required to develop a working knowledge
of the natural attenuation processes occurring on site. Site specific conditions, previous
site activity, or regulatory and public involvement may result in significant deviation from
the proposed outline.

Stepwise Process for Evaluating, Selecting, and Monitoring Natural Attenuation for
Groundwater Remediation-

       1.     Collect and evaluate existing site data

      2.     Identify exposure points, water use practices, and receptors of the aquifer
             (RBCA)
                                                                   i
      3.     Determine groundwater flow direction, velocity, and distance to nearest
             receptor

      4.     Define the risk associated with the current groundwater conditions
             (RBCA)

      5.    Assess potential for natural attenuation using existing data and prelimi-
             nary risk evaluation
                                      6-35

-------
6.    Construct a conceptual model for natural attenuation on site
      6.1   If preliminary site data provide evidence that natural attenuation is
            occurring proceed
      6.2   If risk of human exposure or further environmental damage is unac-
            ceptable or if adequate site data indicate the natural attenuation is
            not or cannot occur, evaluate other remedial strategies

7.    Conduct site characterization to specifically support natural attenuation
      7.1   Contaminant mass
      7.2   Contaminant concentration
      7.3   Presence of source areas
      7.4   General groundwater monitoring parameters, e.g., electron accep-
            tors, respiration products, pH, alkalinity, etc.
      7.5   Define abiotic mechanisms that result in change in concentration,
            e.g., dilution, dispersion, dissolution from a source area, retarda-
            tion, etc.

8. Refine the conceptual model, incorporating new site data

9.    Determine if supplemental treatment technologies (e.g. NAPL recov-
      ery/source removal) are required to insure successful and expedient natu-
      ral attenuation

10.   Project performance of natural attenuation using analytical or numerical
      methods
      10.1  Analytical modeling includes the application of the calculations
            presented in this document and other emerging analytical ap-
            proaches such as multivariate statistical analysis
      10.2  Three numerical models are widely available for modeling natural
            attenuation
            10.2.1        BIO-1D, a commercially available one-dimensional
                         computer simulation of biodegradation and contami-
                         nant migration
            10.2.2        BIOTRANS, a commercially available two-dimen-
                         sional computer model that incorporates available
                         electron acceptors
            10.2.3        BIOPLUME II, public domain computer code using
                         the USGS Method of Characteristic (MOC) code to
                         model oxygen and contaminant distribution.  Bio-
                         degradation occurs in stoichiometric proportions when
                         oxygen and contaminant coincide. BIOPLUME II is
                         the most widely used model for oxygen enhanced
                               6-36

-------
                               aquifer bioremediation and natural attenuation. (BIO-
                               PLUME III is anticipated by May, 1995)

      11.   Compare natural attenuation model predictions with long-term risk
            11.1   If risk is acceptable, proceed
            11.2   If risk is unacceptable, evaluate a more protective remedial strate-
                   gy

      12.   Develop a long-term monitoring plan
            12.1   Revise attenuation model as data become available
            12.2   Sampling and analysis to verify continuing site remediation
            12.3   Locate "sentry" wells to delimit the maximum allowable extent of
                   contaminant migration before a contingency plan is executed
            12.4   Define a contingency plan in case natural attenuation does not
                   meet expectations or otherwise fails to protect human health and
                   the environment
      13.   Execute monitoring plan
            13.1   Sample and analyze sentry wells
            13.2   Sample and analyze groundwater from selected monitoring wells
            13.3   Evaluate results and compare with expectations
            13.4   Close site when clean-up goals are reached
            13.5   Default to contingency if sentry wells become contaminated, or if
                   natural attenuation otherwise fails to protect human health and the
                   environment.

Conclusion
This chapter has identified the basic tools currently used to define the natural attenua-
tion process.  Natural attenuation is a combination of physical, chemical, and biological
processes that occur in a complex geological setting. Evaluation techniques for param-
eters affecting  natural attenuation are still evolving. Each parameter contributes com-
plexity and largely unknown error to evaluations of past  performance and predictions of
future performance. The negative impact of these errors can be minimized by a tech-
nically sound monitoring plan, attention to deviations from model predictions, and the
use of conservative assumptions.  Time and experience will improve our understanding
of this phenomenon. This will lead to more accurate performance estimates and
greater acceptance of natural attenuation as a viable, reliable, and satisfactory
environmental  risk management strategy.                             ;
                                     6-37

-------
References for Section 6

Atlas, R. M., 1984, Petroleum Microbiology.  Macmillan Publishing Co., New York.

Atlas, R. M., and R. Bartha, 1981, Microbial Ecology: Fundamentals and Applications,
Addison-Wesley Publishing Co., Reading.

Charbeneau, R. J., P. B. Bedient, and R. C. Loehr, 1992, Ground Water Remediation,
Techomic Publishing Co., Inc. Lancaster.

Cozzarelli, I. M., R. P. Eganhouse, and M. J. Baedeker, 1990, Transformation of Mono-
aromatic Hydrocarbons to Organic Acids in Anoxic Ground Water Environment. En-
viron. Geol. Water Science, 16.

Cozzarelli, I. M., M. J. Baedecker, R. P. Eganhouse, and D. F. Goerlitz,  1994, The Geo-
chemiical Evolution of Low-Molecular-Weight Organic Acids Derived From the Degrada-
tion of Petroleum Contaminants in Groundwater. Geochimica et Cosmochimica Acta, v.
58, No. 2,  p. 863-877.

Dupont, R. R.,  1995, Personal Communication.

Edwards, E. A., L. E. Wells, M. Reinhard, and D. Grbic-Galic, 1992, Anaerobic
Degradation of Toluene and Xylene by Aquifer Microorganisms Under Sulfate-Reducing
Conditions. Appl. Environ. Microbiol., v. 58, p. 794-800.

Ehrlich, H. L., 1981, Geomicrobiology, Marcel Dekker, Inc. New York.

Freeze, R. A., and J. A. Cherry, 1979, Groundwater. Prentice-Hall, Inc., Englewood
Cliffs.

Grbic-Galic,  D., and T. M. Vogel, 1987, Transformation of Toluene and Benzene by
Mixed Methanogenic Cultures.  Applied and Environmental Microbiology, v. 53, p. 254-
260.

Jury, W. A., W. R. Gardner, and W. H. Gardner, 1991, Soil Physics, Fifth Edition, John
Wiley and  Sons, Inc., New York.

Kuhn, E. P., J.  Zeyer, P. Eicher, and R. P. Schwarzenbach, 1988, Anaerobic Degra-
dation of Alkylated Benzenes in Denitrifying Laboratory Aquifer Columns. Applied and
Environmental  Microbiology, v. 54, p. 490-496.                        I

Lovely, D.  R., 1991, Dissimilatory Fe(III) and Mn(IV) Reduction.  Microbiological Re-
views, June  1991, p. 259-287.

                                     6-38

-------
Lyman, W. J., W. F. Reehl, D. H. Rosenblat, 1990, Handbook of Chemical Property Es-
timation Methods. American Chemical Society, Washington.

Major, D. W., C. I. Mayfield, and J. F. Barker, 1988, Biotransformation of Benzene by
Denitrification in Aquifer Sand. Groundwater, v. 26, p. 8-14.            ;

Rafaii, H. S., 1993, Mathematical Modeling of Intrinsic Bioremediation at Field Sites.
Proceeding of the Symposium on Intrinsic Bioremediation of Ground Water.
EPA/540/R-94/515.

Rochkind-Dubnisky, M. L, G. S. Sayler, and J. W. Blackburn, 1987, Microbiological De-
composition of Chlorinated Aromatic Compounds. Marcel Dekker, Inc., New York.

Tan, T. C., 1994, Modeling of a Natural Bioremediation Processing Multivariate Statis-
tical Approach, Master's Thesis, University of Tennessee.              :

Wiedemeier, T. H., D. C. Downey, J. T. Wilson, D. H. Kampbell, R. N. Miller, and J. EE.
Hansen,  1994, Technical Protocol for Implementing the Intrinsic Remediation with
Long-Term Monitoring Option for Natural Attenuation of Dissolved-Phase Fuel Con-
tamination in Ground Water, Air Force  Center For Environmental Excellence, Brooks Air
Force Base, San Antonia, TX.

Wilson, J. T., M. Cook, D.  Kampbell, 1994, Intrinsic Remediation of JP-4 Jet Fuel,
Proceedings of the Third Environmental Restoration Technology Transfer Symposium,
San Antonio, TX.

Wilson, J. T., D.H. Kampbell, J. Armstrong,  1993, Natural Bioreclamation of Alkyl-
benzenes (BTEX) from a Gasoline Spill in Methanogenic Groundwater.  In: Proceed-
ings, Second International Symposium on In Situ and On Site Bioreclamation, San
Diego, CA.

Wilson, J. T., 1993, Testing Bioremediation in the Field, In: In Situ Bioremediation,
When Does It Work?, National Academy Press. Washington.
                                     6-39

-------