&EPA
United Slates
Environmental
Agency 	
          Monitored Natural
          Attenuation of
          Tertiary Butyl Alcohol (TBA)
          in Ground Water at
          Gasoline Spill Sites
          MW-11
                 CH,
               CH3-C-O-CH3

                 CH,
               O
                   OH
              CH3-C-OH

                 CH,
                 H

              CH.-C-OH
                 CH,
                               CH,
                      CH3-C-OH

                         CH,
                               CH2 - OH
                      CH3-C-OH

                         CH,
                      CH3-C
                          O
                         CH,
            MW-7
MW-3
                            MW-14
                              MW-1
              Clays/Silt
                             UST

                              "'
                                      -V
                                 Silty Fine
                                 Sand

-------
                                                      EPA/600/R-07/100
                                                         October 2007
Monitored  Natural  Attenuation of Tertiary Butyl
  Alcohol (TBA) in  Ground Water at Gasoline
                         Spill Sites
                      John T. Wilson and Cherri Adair
                    U.S. Environmental Protection Agency
                    Office of Research and Development
                    National Risk Management Laboratory
                         Ada, Oklahoma 74820
                            Project Officer
                            John T. Wilson
               Ground Water and Ecosystems Restoration Division
                National Risk Management Research Laboratory
                         Ada, Oklahoma 74820
                National Risk Management Research Laboratory
                    Office of Research and Development
                    U.S. Environmental Protection Agency
                         Cincinnati, Ohio 45268

-------
                                Notice
    This document is intended for internal Agency use only. The U.S. Environmental
Protection Agency through its Office of Research and Development funded portions
of the research described here. Mention of trade names and commercial products
does not constitute endorsement or recommendation for use. All research proj-
ects making conclusions and recommendations based on environmentally related
measurements and funded by the Environmental Protection Agency are required to
participate in the Agency Quality Assurance Program. This project was conducted
under a Quality Assurance Project Plan for Task 10013, Fate  of Fuel Oxygenates
in Aquifer Materials.  Work performed by U.S. EPA employees or by the U.S. EPA
on-site analytical contractor followed procedures specified in  these plans without
exception. Information on the plans and documentation of the quality assurance
activities and results are available from Cherri Adair.

-------
                                         Foreword
The U.S. Environmental Protection Agency is charged by Congress with protecting the Nation's land, air,
and water resources. Under a mandate of national environmental laws, the Agency strives to formulate
and implement actions leading to a compatible balance between human activities and the ability of natural
systems to support 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 science 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 investigation of technologi-
cal and management approaches for preventing and  reducing risks from pollution that threatens human
health  and the environment.  The focus of the  Laboratory's research program is on methods and their
cost-effectiveness for prevention and control of pollution to air, land,  water, and subsurface resources;
protection of water quality in public water systems; remediation  of contaminated sites, sediments and
ground water; prevention and control of indoor air pollution; and restoration of  ecosystems.  NRMRL
collaborates with both public and private sector partners to foster technologies that reduce the cost of
compliance and to anticipate emerging problems. NRMRLs research provides solutions to environmental
problems by: developing and promoting technologies that protect  and  improve the environment; advanc-
ing scientific and engineering information to support regulatory and policy decisions; and providing the
technical support and information transfer to ensure implementation of environmental regulations and
strategies at the national, state, and community levels.

This publication has been produced as part of  the Laboratory's strategic long-term  research plan. It is
published and made available by EPA's Office of Research and Development to assist the user community
and to  link researchers with their clients.

Tertiary butyl alcohol (TBA) is one of the most widely distributed  organic contaminants in ground water
at gasoline spill sites.  The U.S. EPA does not have a Maximum Contaminant Level (MCL) for TBA in
drinking water.  Nevertheless, many states have set standards for TBA in  drinking water and clean up
goals for TBA at gasoline spill sites. Because other contaminants, such as benzene and methyl tertiary
butyl ether, are often biologically degraded in anaerobic ground water, the state agencies that imple-
ment the Under Ground Storage Tank program  rely heavily on monitored natural attenuation to clean up
these contaminants at gasoline spill sites. This report reviews the prospects for using monitored natural
attenuation to manage the risk from TBA in ground water at gasoline  spill sites.

The  report reviews the distribution of TBA in ground water at gasoline spill sites, the process that pro-
duces TBA from anaerobic  biodegradation of MTBE, and the prospects  for natural biodegradation of
TBA in ground water. The report presents data from a microcosm study conducted by U.S. EPA on TBA
degradation in sediment from six gasoline spill sites distributed  around the  United States.  Finally the
report reviews the limited knowledge on use of stable carbon and stable hydrogen isotopes to evaluate
natural biodegradation of TBA at field scale.
                                               G. Schmelling, Director
                                        Ground Water and Ecosystems^Piestoration Division
                                        National Risk Management Research Laboratory

-------

-------
                                     Contents
Foreword	iii
Figures  	vii
Acknowledgments 	ix
Abstract 	xi
Introduction	1
Distribution of TEA, MTBE, and Benzene at Gasoline Spill Sites	1
Interpreting TEA Biodegradation at Field Scale	4
    Attenuation with Distance from the Source	4
    Attenuation with Time in Individual Monitoring Wells	5
Biodegradation of TBA 	5
Rates of Biodegradation of TBA in Ground Water 	9
Extent of Biodegradation of TBA 	11
EPA microcosm study of anaerobic TBA biodegradation 	15
    Construction of Microcosms 	15
    Laboratory Analytical Procedures	16
    Data Quality 	16
    Biodegradation of TBA in Microcosms	17
    Use of Stable Isotope Ratios 	20
Summary  	26
Recommendations	27
    Research Needs	28
References 	29
Appendix  	32

-------
VI

-------
                                              Figures
Figure 1.   Frequency distribution of the maximum concentrations of TEA, MTBE, and Benzene
           at gasoline spill sites in Los Angeles County, California	2

Figure 2.   Distribution of MTBE and TBA at gasoline spill sites in Orange County, California, in 2002 and in
           the Eastern United States in 1999	3

Figure 3.   Relationship between the fractionation of carbon isotopes in MTBE produced by biodegradation
           of MTBE to TBA and the ratio of TBA to MTBE in ground water in monitoring wells at 13 gasoline
           spill sites in Orange County,  California	4

Figure 4.   Removal of MTBE and production or removal of TBA in anaerobic sediment  from a gasoline spill
           site in Fountain Valley, California (98UT010) and a gasoline spill site in Laguna Miguel, California
           (91UT086)	9

Figure 5.   Relative importance of electron  acceptors at 25 fuel spill sites in North America	12

Figure 6.   Distribution of sulfate at 77 gasoline spill sites in the Eastern United States	12

Figure 7.   General distribution of terminal electron accepting processes (TEAPs) in ground water down
           gradient from a spill of gasoline	13

Figure 8.   Distribution of the concentrations of TBA and sulfate in selected monitoring wells at gasoline spill
           sites in Orange County, California	14

Figure 9.   Relationship between depth  below the water table and the concentrations of TBA and sulfate in
           ground water at a gasoline spill  site in Port Hueneme, California	14

Figure 10.  Removal of TBA in triplicate  microcosms constructed with material for a gasoline spill site in
           Petaluma, California	18

Figure 11.  Behavior of TBA in triplicate  microcosms constructed with material for gasoline spill sites in Deer
           Park, New York and Parsippany, New Jersey. 	18

Figure 12.  Behavior of TBA in microcosms constructed with material for gasoline spill sites in Boca Raton,
           Florida	19

Figure 13.  Behavior of TBA in microcosms constructed with material from a gasoline spill site at Port
           Hueneme, California and a site  at Vandenberg AFB, California	19

Figure 14.  Expected relationship between the ratio of stable isotopes of carbon in MTBE or TBA and the
           extent of biodegradation of MTBE or TBA	21

Figure 15.  Changes in concentrations of MTBE and TBA in monitoring wells over time at three sites in
           Orange County, California, data plotted on logarithmic scale	22

Figure 16.  Changes in concentrations of MTBE and TBA in monitoring wells over time at three sites in
           Orange County, California, data plotted on arithmetic scale	22

Figure 17.  Changes in concentrations of MTBE and TBA in monitoring wells over time at a site in Orange
           County, California, where a decline in concentration of MTBE was followed at a  later time by a
           decline in concentration of TBA	23
                                                  VII

-------
Figure 18.  Relationship between the isotopic ratio of carbon in TEA in ground water at a manufacturing
           facility in Pasadena, Texas, and the concentration of TEA	24

Figure 19.  Comparison of the range of variation in the isotope ratio of carbon (8 13C) in TEA at monitoring
           wells at sites in Orange County, California, (Kuder et al., 2005) and at a site in South America
           (Zwank et al., 2005) to the range of variation in the isotope ratio of carbon (8 13C) in MTBE in the
           same monitoring wells	24

Figure 20.  Comparison of the range of variation in the isotope ratio of hydrogen (8 D) in TEA at monitoring
           wells at a site in South America (Zwank et al., 2005) to the range of variation in the isotope ratio
           of carbon (8 13C) in MTBE in the same monitoring wells	25
                                              Tables
Table 1.    Rates of Anaerobic Biodegradation of TBA in Aquifer Sediments or Field Scale Plumes at
           Gasoline Spill Sites	8

Table 2.    Distribution of TBA, MTBE, Methane, and Sulfate in Monitoring Wells at the three Gasoline
           Service Stations in Orange County, California that Provided Sediment for the Microcosm
           Study of DeVaull et al. (2003)	10

Table 3.    Comparison of first order rates of anaerobic biodegradation of TBA, MTBE, and Benzene	11
                                           Appendix:



Table 4.    Typical Quality Performance Data for Analysis of TBA in Water	32
Table 5.    Typical Quality Performance Data for Analysis of Sulfate in Water	35
                                                 VIM

-------
                                 Acknowledgments
This document was peer reviewed by Dr. Matthew Caldwell of the National Risk Management Laboratory,
U.S. EPA, by Dr. Kevin T. Finneran of the University of Illinois, and by Dr. Yue Rong, Senior Environmental
Scientist, California Environmental Protection Agency, California Regional Water Quality Control Board
Los Angeles Region.
                                            IX

-------

-------
                                          Abstract
The state agencies that implement the Underground Storage Tank program rely heavily on Monitored
Natural Attenuation (MNA) to clean  up contaminants such as  benzene and methyl tertiary butyl ether
(MTBE) at gasoline spill sites.  This  is possible because the contaminants are biologically degraded in
anaerobic ground water at the site.   Tertiary Butyl Alcohol (TBA) is  generally considered to be more
readily degradable than MTBE, and there is a danger that the state agencies will consider contamination
from TBA a good prospect for MNA. A close examination of the available information indicates that a
default presumption that TBA is readily degraded in anaerobic  ground water is not justified.  Anaerobic
biodegradation of TBA will require a supply of an electron acceptor such as sulfate or biologically available
Iron(lll) or Manganese(IV). The available survey data indicate that ground water in the source area of
the majority of known plumes is devoid of sulfate. Although a procedure to estimate biologically available
Iron(lll) is commercially available, it is not routinely applied to gasoline  spill sites. There is no established
procedure to  estimate biologically available Manganese(IV). To date, the performance of available ap-
proaches to document  anaerobic biodegradation of TBA at specific field sites has been disappointing.
These include field monitoring to show a statistically significant attenuation in concentration with distance
along the flow path, microcosm studies conducted with sediment from the  site, and analysis of stable
isotope ratios in TBA in the plume.
                                               XI

-------
XII

-------
                                             Introduction
Tertiary butyl alcohol (TEA) is one of the most widely distrib-
uted organic contaminants in ground water at gasoline spill
sites. The U.S. EPA does not have a Maximum Contaminant
Level (MCL) for TEA in drinking water. Nevertheless, many
states have set standards for TEA in drinking water and
clean up goals for TEA at gasoline spill sites.

  The state agencies that  implement  the Underground
Storage Tank program rely heavily on  monitored natural
attenuation to clean up organic contaminants  in ground
water at  gasoline spill sites.  There are a variety of pro-
cesses that attenuate the concentrations of contaminants
in ground water, including sorption, hydrodynamic disper-
sion, and biodegradation.  At many gasoline spill  sites,
sorption and dispersion alone are not adequate to prevent
the contaminants from reaching a receptor.  To protect the
receptor, the contaminant must be degraded  in ground
water. Monitored natural attenuation has been selected as
a remedy, or part of the remedy, at certain  fuel spill sites
because it has shown that the contaminants of concern,
such as benzene, are biologically degraded in anaerobic
ground water (Parsons Engineering Science, 1999; Wiede-
meieretal., 1999a; Wiedemeieret al., 1999b; Wiedemeier
etal., 1999c).

Because TEA is very soluble in water, it dissolves readily
out of spilled gasoline into ground water, and can  reach
high concentrations in ground water. Sorption of TEA to the
solids in the aquifer matrix is  negligible (Interstate Technol-
ogy & Regulatory Council, 2005). To use monitored natural
attenuation as a remedy for TEA contamination in ground
water at  many gasoline spill sites, it will be necessary to
demonstrate that TEA is biologically degraded in the ground
water at the site.

This report reviews the distribution of TEA in ground water
at gasoline spill sites,  the process that produces TEA from
anaerobic biodegradation of  MTBE, and the prospects for
natural biodegradation of TEA in ground water. Additionally,
the report evaluates available information  on attenuation
of TEA with distance along  a flow path in  ground water.
This report reviews available information on the rates and
extent of biodegradation of TEA  in ground water. Data is
presented from a microcosm study conducted by U.S. EPA
on TEA degradation in sediment from six gasoline spill sites
distributed around the United States. This report reviews the
limited  knowledge on the use of stable  carbon and stable
hydrogen isotopes to evaluate natural  biodegradation of
TEA at field scale. Finally, the report makes  recommenda-
tions for using monitored natural attenuation to manage the
risk posed by TEA in ground water at gasoline spill sites.
Distribution of TEA, MTBE, and Benzene at
      Gasoline Spill Sites
Tertiary butyl alcohol (TEA) is widely distributed in ground
water that is  contaminated by spills  of gasoline.  Shih
et al., (2004)  compared the distribution of the maximum
concentrations of TEA, methyl tertiary butyl ether (MTBE),
di-isopropyl  ether  (DIPE),  tertiary amyl methyl  ether
(TAME), ethyl tertiary butyl ether (ETBE) and benzene in
ground water at 868 leaking fuel tank sites in Los Angeles,
California. The geometric mean concentration of TBA was
1,730 ug/L. The geometric mean concentrations of MTBE,
benzene, DIPE, TAME, and ETBE were 900, 700, 31, 24,
and 7 ug/L. The concentration of TBA was slightly higher
than the concentrations of MTBE and benzene, and sub-
stantially higher than the concentrations of DIPE, TAME,
and ETBE. The frequency distribution of TBA, MTBE, and
benzene as reported by Shih et al., (2004) is presented in
Panel A of Figure 1.
The frequency distribution of concentrations of TBA and
MTBE in Orange County, California are presented in Panel
B of Figure 1 (Wilson et al., 2005a; personal communication
Seth Daugherty, Supervising Hazardous Waste Special-
ist-Retired, Orange County Health Care Agency). Orange
County is on the Pacific Coast in Southern California, just
south of Los Angeles County.  As might be expected, the
distribution of concentrations of TBA and MTBE in Orange
County are almost  identical to the distribution in Los An-
geles County.
Other regions of the United States have a similar distribution
of concentrations of TBA.  Panel C of Figure 1 presents
data from a survey of 74 gasoline stations in the Eastern
United States (Kolhatkar et al., 2000). A total of 41 sites
from Pennsylvania, 8 sites from New Jersey, 5 sites from
New York, 5 sites from Florida, 7 sites from Indiana, 3 sites
from Maryland, 2 sites from Washington DC, and 3 sites
from Ohio were included in the survey. At least six wells
were sampled from  each site. Panel C of Figure 1 presents
the frequency distribution of the maximum concentration of
TBA,  MTBE, and benzene in any well at the 74 gasoline
spill sites. There was no significant difference between
the frequency distributions in the spill sites in the Eastern
United States and the distribution reported by Shih  et al.,
(2004) for Los Angeles, California.  The concentrations of
TBA were equivalent to the concentrations of MTBE. The
geometric mean concentration of TBA was 1,512 ug/L while
the geometric mean concentrations of MTBE and benzene
were 1,724 ug/L and 447 ug/L respectively.

-------
   100

«   90
I   80
°   70
§   60
I   50
§   40
|   30
I   20
°   10

                               Los Angeles
                               County, California
                                    10
                      100
1000    10000   100000  1000000
Cumulative Percent of Sites
IUU
90
80
70
60
50
40
30
20
10
n .:
^i^'"
Bjgf^ .."
X--""
Orange County, ^*J~ * TBA
California J? / * MTBE
J^ -::r
.JS^*" _?"
Jl^* ^ .!::""
b
A^* s b
***^ *»»= "
~*f
                                    10     100     1000    10000   100000  1000000


.1
co
|
Tz
0)
D.
5
1
i
d

IUU •
90 -
80 -

70 -
60 -
50 -
40 -
30 -
20 -
10 -

• •
J **;*
c A ='
/,.,-
jk~" *
Eastern .j,*,~ » jgy
United States rf'/ * MT
, ff^ * Ber
-I .:•,*
/' ,f
/ !?
1 ,::l
• ' ,- - "" /
• " i •» f
»•»*"*"
*



\
BE
izene





                                    10      100     1000    10000   100000  1000000
                                      Maximum Concentration (M^g/L)
Figure 1.      Frequency distribution of the maximum concentrations of TBA, MTBE, and Benzene at gasoline
              spill sites. Panel A presents data from Los Angeles County, California. Panel B presents data from
              Orange County, California in 2002.  Panel C presents data from the Eastern United States in 1999.
              Concentrations of TBA are represented by red diamonds, MTBE by blue triangles,  and benzene by
              black squares.

-------
Although the frequency distribution of concentrations of TEA
and MTBE are very similar in Southern California and the
Eastern United States, the distributions of concentrations
at individual gasoline spill sites are not similar. Panel A of
Figure 2 compares the maximum concentration of TEA in
any well at a site to the maximum concentration of MTBE
in any well at a site in Orange County, California, in 2002
(Redrawn from Figure 1 in Wilson et al., 2005a).  There
was no correlation between concentrations of MTBE and
TBA.

Wilson et al. (2005a) and Wilson et al. (2005b) used equi-
librium partitioning theory to predict the concentrations of
TBA and MTBE in ground water in contact with gasoline in
                                      an aquifer at reasonable values for residual gasoline. They
                                      assumed that gasoline spilled in Orange County was 11 %
                                      MTBE by weight, and that the technical grade of MTBE used
                                      in the gasoline was 2% TBA by  weight.  The solid curved
                                      lines in Figure 2 are the range of concentrations of TBA and
                                      MTBE that would be expected if the residual concentration
                                      of gasoline in the spill varied from 1000 to 40,000 mg/kg
                                      total petroleum hydrocarbons.  The dashed lines  bound
                                      the range of concentrations that  would be expected if con-
                                      centrations of TBA or  MTBE in contact with gasoline were
                                      attenuated by dilution  and dispersion in ground water, and
                                      the TBA and MTBE did not sorb and were not biologically
                                      degraded in the ground water.
                   10000000


                    1000000 -


                     100000 -


                   -, 10000 -
                   <  1000 -
                   CD
                       100-


                        10
                 Orange County, California
                    A      .
                           «   »«
                                  •V     -r ^     -r       - ^
                          *     »   *       «*.*•*'
                          •:*• •{  ••v/eC*^*x

                 •     .  V V.\%* V^/V^^^*"
                              «*
                «««»•«  ^^J,  ^  ^.^ ^
                           10
                     100     1000    10000    100000   1000000 10000000
IUUUUUUU -

1000000 -

100000 -



~ 10000 -
•a
3.
^ 1000 -
p
100-


m -

Eastern United States

B .. /'I

x S
• » xx X
* * * • xx xx

* *t * ^x' V^
** »*xx » ^X

» ******* v*».x
*•• ^^ x^"^
« X » XX
X «» x'
x" * »X
Figure 2.
             10       100     1000    10000    100000   1000000  10000000

                                   MTBE (Mg/L)

Distribution of MTBE and TBA at gasoline spill sites. Panel A presents data from Orange County,
California, in 2002. Panel B presents data from the Eastern United States in 1999. The figure
compares the maximum concentration of TBA in any well at a particular gasoline spill site in any
sampling period in the year against the maximum concentration of MTBE in any well at any sampling
period in the year.  The dashed straight lines and solid curved lines bound the range of concentrations
of TBA and MTBE expected in water in contact with gasoline containing 11% MTBE and 0.22% TBA.

-------
Many of the data in Panel A of Figure 2 fall above the range
of concentrations of TEA that would be expected for parti-
tioning of TEA from gasoline. TEA is produced as the first
biodegradation product during anaerobic biodegradation of
MTBE (Kolhatkar et al., 2002); the most plausible source of
the high concentrations of TEA is from natural anaerobic
biodegradation of MTBE.  Recent work makes it possible
to recognize and track biodegradation of MTBE in ground
water by a change in the ratio of the stable isotopes of
carbon in MTBE (Hunkeler et al., 2001; Gray et al., 2002;
Kolhatkar et al., 2002; Kuder et  al., 2005; Wilson et al.,
2005b; Zwank  et al.,  2005).  Wilson et al., (2005a) used
this approach to show that MTBE was being degraded at
many of the sites in Orange County, California, and that
the sites where MTBE was extensively degraded had high
ratios of TBA to MTBE (Figure  3).

The distribution of TBA and MTBE in the survey of MTBE
gasoline spill sites in the Eastern United States is pre-
sented in Panel B of Figure 2 (redrawn from Figure 3.10 in
Wilson et al., 2003).  The data are sparse, but the general
distribution of concentrations is similar to the distribution
in Southern California.  As was the case in Southern
California, the concentrations of TBA at many stations was
from one hundred to one thousand-fold higher than would
be expected from partitioning of TBA from gasoline. The
high concentrations of TBA suggest that the TBA is largely
produced by biodegradation of  MTBE.
                     At this writing (June 2006), the standards set by individual
                     states for TBA in ground water vary from 12 ug/L to 3,200
                     ug/L. Based on the frequency distribution of TBA in Orange
                     County, California, approximately 5% of sites have at least
                     one well with concentrations of TBA greater than 110,000
                     ug/L. In the data set reported by Shih et al. (2004) for Los
                     Angeles,  California, 5% of sites had TBA concentrations
                     greater than 97,000 ug/L.  A concentration of 110,000 ug/L
                     would have to be diluted 10,000 fold to meet a standard of
                     12 ug/L, but only 34 fold to meet a standard of 3,200 ug/L.  If
                     natural attenuation is to be used as a remedy in those states
                     that have clean up goals for TBA in the range of 12 to 140
                     ug/L, then natural biodegradation must make a significant
                     contribution to natural attenuation.  The remainder of this
                     report evaluates the prospects for natural  biodegradation
                     of TBA in ground  water at gasoline spill sites.

                     Interpreting TBA Biodegradation at Field
                          Scale

                     Attenuation with Distance from the Source
                     The most direct approach to evaluate biodegradation of
                     TBA at field scale  is to compare the concentration of TBA in
                     highly contaminated wells in the source area to concentra-
                     tions in wells down gradient. Unfortunately, this approach
                     rarely works well for TBA.
               100000
                10000
                 1000
             LU
             CD
             CD
                  100
                   10
                  0.1
    A +
     0
                     -40
-20
20
                                               813C of MTBE
40
60
                            • 99UT015
                            o 99UT032
                            A 96UT028
                            *87UT211
                            *91UT086
                            = OOUT038
                            + 86UT175
                            • 89UT007
                            - 86UT062
                            o88UT138
                            D88UT198
                            A85UT114
                            X98UT010
Figure 3.      Relationship between the fractionation of carbon isotopes in MTBE produced by biodegradation of
               MTBE to TBA and the ratio of TBA to MTBE in ground water in monitoring wells at 13 gasoline spill
               sites in Orange County, California. (Redrawn from Figure 3 in Wilson et al., 2005a)

-------
Buscheck and  Alcantar (1995) developed an equation
that could predict the first order rate of biodegradation of
a compound along an inferred flow path in ground water
from the changes in concentration in monitoring wells along
the flow path and the seepage velocity of the contaminant
along the flow path.  The equation corrects for attenuation
due to dispersion and sorption to aquifer solids.  Kolhatkar
et al., (2000) used the approach of Buscheck and Alcantar
(1995) to  evaluate TBA biodegradation  in ground water
at 74 gasoline  spill sites.  To  correct for dilution effects
in monitoring wells, Kolhatkar et al., (2000) divided the
concentration of TBA by the concentration  of methane in
ground water.

In this approach the first-order biodegradation rate constant
(X) in a steady-state plume is estimated  from (1) a linear
regression of the natural logarithm of TBA  concentration on
distance along the flow path, (2) the coefficient of longitudi-
nal dispersivity and (3) the flow velocity of the contaminant
(equal to the seepage velocity of ground  water divided by
the retardation factor).
                                     -1
                4oc
 Where:  A = first-order biological degradation  constant
         (per year),

     vc = retarded contaminant velocity in the x-direction
         (meters per year),

     cc^ = longitudinal dispersivity (meters)

     k/vx=  rate of attenuation in contaminant concentration
         with distance along the flow path (per meter).

The rate of attenuation in contaminant concentration with
distance along the flow path can be estimated as the nega-
tive of the slope of a regression  of the natural logarithm
of contaminant concentration on  distance along the flow
path (meters). Their criterion for biodegradation of TBA at
a site was a rate of attenuation in contaminant concentra-
tion with distance that was statistically greater than zero.
Specifically, the slope of the regression had to be greater
than zero at 95% confidence.

Of 74 gasoline spill sites that they examined  in the  East-
ern United  States, they found only three sites where the
regression for TBA was significant at 95% confidence. It is
possible that TBA biodegradation was rare in this population
of field sites. However, these sites were not research sites.
The wells were installed at the direction of the regulatory
authorities for monitoring  to support a risk assessment
and to select appropriate remedies. It is also possible that
conventional practice for locating monitoring wells failed to
recognize TBA biodegradation. In either case, monitoring
using conventional  practice provided little  interpretable
evidence that biodegradation of TBA was  important at
gasoline spill sites.

As a practical matter, it is  difficult to distinguish  changes
in concentration  due to biodegradation from changes due
to purely physical processes. The concentration data may
be affected by uncertainties associated with hydrological
factors.  If the ground water flow velocity is slow, the con-
centrations in down gradient wells may not have reached
the maximum concentration. The plume may change flow
direction, and carry the contaminated ground water away
from the monitoring well.  The down gradient well may be
askew of the predominant flow path in the aquifer, and con-
centrations are lower because the well failed to sample the
centerline of the plume. If there is variation in the lithology
at the site, and the plume is confined to one flow zone and
other zones produce clean ground water, concentrations
in a down gradient well may be lower because the plume
was diluted to a  greater extent in the down gradient well.
The concentrations of TBA can also be  attenuated by
hydrodynamic dispersion in the aquifer.  Finally, the con-
centrations of TBA are related to the distribution of MTBE
and BTEX contamination, particularly if the MTBE is being
biodegraded to produce TBA.

Attenuation over Time in Individual
      Monitoring Wells

Because it is often difficult to compare changes in con-
centration of TBA from one well to another, it is important
to maintain a good  continuous monitoring record of TBA
concentrations over time in strategically located wells.  If
the general trend in concentrations over time is down, that
is the best evidence available from conventional monitoring
that natural attenuation processes are reducing concentra-
tions. However,  when the only data are TBA monitoring
data,  it  is impossible to attribute natural  attenuation to
natural biodegradation. Very  likely most of the reduction
in concentration  is  due to physical weathering of TBA, or
its parent compound MTBE from the source area of con-
tamination at the spill site.  If a transport and fate model is
used as part of the risk evaluation, there is no justification to
include biodegradation in the model. Including biodegrada-
tion in the model requires direct evidence that bacteria in
the contaminated aquifer are capable of degrading TBA, or
in fact, have degraded TBA. In the past, evidence that the
aquifer harbored organisms that could degrade an organic
contaminant was attained by conducting microcosm studies.
In recent years, more direct evidence of biodegradation of
benzene, other BTEX compounds, and MTBE  has been
attained from the analysis of stable isotope ratios (Schmidt
et al., 2004b; Meckenstock et  al., 2004).

The following sections  review the literature on  biodegra-
dation of TBA, and then specifically  consider anaerobic
biodegradation of TBA in microcosms constructed from con-
taminated aquifer material. The final section evaluates the
use of stable isotopes to recognize TBA biodegradation.

Biodegradation of TBA

At many gasoline spill sites, TBA occurs at high concen-
trations  in groundwater, and dilution and dispersion alone
cannot be expected to bring concentrations of TBA to clean
up goals before the ground water reaches a receptor. At
these sites, Monitored  Natural Attenuation will only be a

-------
viable option if the TEA is biologically degraded to harm-
less materials.

Organic compounds in ground water can be biologically
degraded through a variety of mechanisms. These mecha-
nisms are biochemical reactions. In a chemical reaction, if
an element or compound loses an  electron it is said to be
oxidized; if the element or compound receives an electron
it is said to be reduced. This old nomenclature goes back
to the chemistry of metal ores.  Molecular oxygen tends
to obtain additional electrons.  The chemical that lost an
electron to oxygen is said to be oxidized.  When a metal
ore is processed to recover the pure metal, the weight of
the pure metal that is produced is  less than the weight of
the original ore. The ore was reduced to produce the pure
metal.  In the process the metal atoms in the ore received
electrons.

Biological  metabolism is a  linked  series of biochemical
oxidation/reduction reactions where one compound loses
electrons and is oxidized and the other compound receives
electrons and is reduced.  If the organic compound is en-
tirely oxidized to carbon dioxide, the process is termed  a
respiration.  If the organic compound is oxidized, something
else must be reduced. That "something else" is described
as the terminal electron acceptor. Oxygen, nitrate, sulfate,
and Iron(lll) minerals can serve as the terminal electron
acceptors during respiration. When molecular oxygen is
reduced to water, the process is termed aerobic respiration.
When sulfate  is reduced to sulfide, the process  is termed
sulfate reduction.  When nitrate is reduced to ammonia or
molecular nitrogen, the process is  called nitrate reduction
or denitrification. When insoluble Iron(lll) minerals are re-
duced to form soluble  Iron(ll) salts, the process  is termed
iron reduction and when insoluble  Manganese(IV) miner-
als are reduced to form soluble Manganese (II) salts, the
process is termed manganese reduction.

In some cases,  organic compounds  can also  serve as
both  an electron donor and an electron acceptor.  These
reactions are  called fermentations. In a fermentation re-
action, an organic compound is the terminal electron ac-
ceptor. Sometimes it is the  same compound that acts as
the electron donor. The most straightforward example of
fermentation is the anaerobic biodegradation of acetate to
produce methane and carbon dioxide. One of the carbon
atoms in acetate is oxidized  to carbon dioxide; the other is
reduced to form methane.

Bacteria that carry out aerobic  respiration are widely  dis-
tributed in soil and sediment and have a great metabolic
diversity. However, oxygen has limited solubility in water and
is usually unavailable in contaminated aquifer sediments.
Ambient concentrations of nitrate  are usually low.  As  a
result, the most important electron acceptors in contami-
nated ground water are sulfate and  Iron(lll) minerals in the
aquifer sediment (Wiedemeier et al., 1999a).

Biodegradation of TBA has  been reviewed by Schmidt et
al., (2004a). Degradation of TBA has been reported with
a variety of terminal electron acceptors including oxygen,
nitrate, sulfate, Iron(lll) and Manganese(IV). Biodegradation
of TBA is also theoretically possible through fermentation
where part of the TBA molecule is reduced to methane or
hydrogen and part is oxidized to carbon dioxide. Schmidt
et al. (2004a) calculated the free energy yield for biodeg-
radation of TBA with oxygen, nitrate, sulfate, or Iron(lll) as
an electron acceptor and for the fermentation of TBA to
methane.  The free energy yield under environmental con-
ditions, ArG°(w), was -2659 kilojoules per mole for oxygen
as electron acceptor, -2360 kilojoules per mole for nitrate
as electron acceptor, -814  kilojoules per mole for Iron(lll)
as electron acceptor, -171  kilojoules per mole for sulfate
as electron acceptor, and -72.06 kilojoules per mole during
methanogenesis.

If the value for  ArG°(w) is  negative, then the reaction is
theoretically possible.  Based on the thermodynamics of
the reactions, biodegradation of TBA should be theoretically
possible under aerobic, nitrate-reducing, sulfate-reducing,
lron(lll)-reducing, and methanogenicconditions. The more
negative the value for ArG°(w), the greater the amount of
energy available from biodegradation. Much more energy
is available from aerobic biodegradation and nitrate reduc-
tion compared to iron  reduction, sulfate reduction, and
methanogenesis.

When oxygen or nitrate is available, biodegradation of TBA
in sediment is rapid  and extensive.  Bradley et  al., (2002)
compared biodegradation of radio-labeled TBA in surface
water sediments. When sediment from Charleston, South
Carolina, was amended with nitrate and incubated at 23°C
for 198 days, 28 ± 5% of the label was recovered as carbon
dioxide. When the sediment was amended with oxygen, 99
± 3% of the label was recovered as carbon dioxide. When
sediment  from  Laurens, South Carolina, was amended
with  nitrate or oxygen, the recovery was 70 ± 10% and
99 ± 2% respectively.

Schirmer et al., (2003) constructed laboratory microcosms
using sediment from the Borden field site in Ontario, Cana-
da. The sediment was collected from a region in the aquifer
where there was evidence of field scale biodegradation of
MTBE. After 22 days of incubation at 10°C, the sediment
degraded  650 ug/L  TBA to a concentration of less  than
5 ug/L. Hunkeler et al., (2001) showed transitory accumu-
lation of TBA in  laboratory cultures derived from sediment
from the Borden field site.  During aerobic biodegradation
of 10 mg/L MTBE, up to 4 mg/L of TBA accumulated. After
the  MTBE was  completely degraded, the TBA began to
degrade. The TBA was completely degraded within 15 days
after the MTBE was no longer available.

Kane et al., (2001)  evaluated MTBE biodegradation un-
der aerobic conditions in sediment from fuel spill sites in
California.  They added 4.5 mg/L of MTBE to sediment
from a  spill site in Palo Alto, California.  When  the MTBE
was degraded, they respiked the sediment with 4.5 mg/L
MTBE. As the MTBE was degraded, TBA accumulated to
a maximum concentration near 2.0 mg/L. Within 25 days,
the TBA was degraded.

Aerobic conditions may exist at the down gradient margin of
a plume, or in sediment where a plume discharges to aero-

-------
bic surface water; however, ground water is often devoid of
oxygen in the source area of plumes where concentrations
of TEA are high.  If Monitored Natural Attenuation is to be
a viable remedy for  TEA, then TEA must be biologically
degraded in the absence of oxygen.

White et al., (1986) evaluated biodegradation of TEA under
anaerobic conditions in sediment from a refinery near the
Schuykill River in Philadelphia, Pennsylvania. They incu-
bated 12 grams of sediment and 4 ml of sterile ground water
without headspace in test tube microcosms at 12°C. After a
period of acclimation, the TEA was completely removed.

Zenker et al., (1999) constructed microcosms using sedi-
ment from a site in North Carolina that had been exposed to
MTBE. The microcosms were constructed under anaerobic
conditions.  After 200 days of incubation, TEA was com-
pletely consumed  in  each of triplicate microcosms.

Bradley et al., (2001) added radio-labeled [14C] TEA to mi-
crocosms prepared using stream and lake-bed sediments
from sites in Charleston, South Carolina; Jacksonville,
Florida; and Picatinny Arsenal, New Jersey.  When the
microcosms  were prepared to simulate anaerobic  condi-
tions,  8%± 1%, 11%  ± 1%, and 8% ± 1% of the label was
recovered as carbon dioxide after 166 days of incubation.

Finneran and Lovley (2001) reported rapid and  extensive
degradation of TEA under mixed iron(lll)-reducing, sulfate-
reducing, and methanogenic conditions in sediment from
the Potomac River. They added radio-labeled  [14C] TEA to
the sediment. After  62 days of incubation, approximately
7% of the label was recovered as methane and 25% of the
label was recovered as carbon dioxide. If the methane had
been produced by fermentation of TEA,  the molar ratio of
methane produced to carbon dioxide produced should have
been three to one. If TEA is degraded by sulfate-reducing
bacteria or lron(lll)-reducing bacteria, all the label should
go to carbon dioxide. The actual ratio was near 0.3 to one.
Finneran and Lovley (2003) concluded that  TEA  in the
sediment from the Potomac River was being degraded by
a variety of terminal  electron accepting  processes, which
included sulfate reduction and Iron(lll) reduction.

In a subsequent  experiment,  Finneran  and Lovly (2003)
added additional sulfate to the sediment  from the Potomac
River. If the addition of sulfate allowed the sulfate-reducing
bacteria to compete with methanogenic bacteria for the
TEA, the fraction of  radio-label going to methane should
be smaller.  When sulfate was not added, approximately
24% of the label was recovered as  carbon dioxide and
10% of the label was recovered as methane.  The ratio of
label in methane to  label in carbon dioxide was near 0.4
to one. When 960 mg/L of sulfate was added to sediment
from the Potomac River, approximately 26%  of the label
was recovered as carbon dioxide and approximately 3% of
the label was recovered as methane, and the ratio of label
in methane to label in carbon dioxide was near 0.1 to one.
The addition of sulfate inhibited the production of methane.
Molybdate is a specific inhibitor of sulfate reduction. When
Finneran and Lovley (2003) added sulfate and molybdate to
sediment from the Potomac River, approximately 23% of the
label was recovered as carbon dioxide and approximately
3% of the label was recovered as methane.  Because the
molybdate did not produce a substantial reduction in the
amount of label in carbon dioxide, Finneran and Lovley
(2003) concluded These results indicate that two processes
are probably responsible for the [14C]-TBA mineralization
- sulfate reduction and Fe(lll) reduction (because nitrate
was not detected in these sediments).

Bradley et al., (2002) compared  anaerobic biodegrada-
tion of radio-labeled [14C] TEA in surface water sediments
under various electron accepting conditions. In sediment
from Charleston, South Carolina, the sediment as collected
was methanogenic. After 200 days of incubation, none of
the 14C from TEA was recovered as carbon dioxide.  They
should have been able to detect 2%  of the original label
as carbon dioxide. When the sediment was amended with
Iron(lll) or Manganese(IV), no 14C from TEA was detected as
carbon dioxide. However, when the sediment was amended
with sulfate, 4 ± 1 % of the 14C from TBA was recovered as
carbon dioxide.

In sediment from  a site  in Laurens, South Carolina, the
unamended sediment was also methanogenic.  After 200
days of incubation, none of the labeled TBA was recovered
as carbon dioxide.  When the sediment was  amended
with Iron(lll), none of the labeled  TBA was recovered as
carbon dioxide.  When the sediment was amended with
Manganese(IV), 75 ± 20% of the  label was recovered as
carbon dioxide.  When the sediment was amended with
sulfate,  5 ±  1%  of the label was recovered as  carbon
dioxide.

Finneran and Lovley (2003) added radio-labeled [14C] ac-
etate to material from a contaminated aquifer near Bemidji,
Minnesota.   Based on the distribution of  label between
methane and carbon dioxide, they concluded that methano-
genesis was the only important electron accepting process
in the sediment. They added radio-label [14C] TEA to this
sediment.  After 23 days  of incubation, approximately 5%
of the 14C from TBA was recovered as methane and ap-
proximately 9% was recovered as carbon dioxide.

The interpretation of their experiments on TBA biodegrada-
tion under sulfate reducing and methanogenic conditions is
not straightforward. Some portion of the 14C from TBA is
recovered as methane, indicating that at a minimum some
transformation product of TBA can be degraded under
methanogenic conditions. However, the yield of methane
is much lower than would be  expected from the complete
fermentation of TBA under methanogenic conditions.
C4H100
H20
                            3CH  + C0
If TBA is fermented to methane according to the stoichiom-
etry above, 75% of label in [14C] TBA should be recovered
as methane. In sediment where  TBA is degraded in the
absence of oxygen or nitrate, approximately 36% to 10%
of the label is recovered as methane. This would suggest
that some portion of the TBA is being oxidized by sulfate,
Iron(lll), or Manganese(IV).

-------
There is not unequivocal evidence that TEA can be me-
tabolized under strictly fermentative conditions.  In their
review, Schmidt et al., (2004a) concluded Thus, in contrast
to MTBE degradation, there is general consensus that
TBA is recalcitrant under methanogenic conditions.  If this
is the case, either sulfate or biologically available Iron(lll)
or Manganese(IV) is required for biodegradation of TBA in
the absence of oxygen or nitrate.

As part of a study to explain the distribution and behavior
of TBA in ground water at gasoline spill sites in Orange
County, California, DeVaull et al., (2003)  collected  sedi-
ment from three  locations at each of three gasoline spill
sites. The sediment was used to construct microcosms.
One experimental treatment  in their report simulated the
natural anaerobic conditions  in the aquifer at their study
sites. A 160 ml  glass vial was filled with  75 ml of water
and  50 g of wet  sediment. The headspace was purged
with nitrogen for 30 minutes and then the vial was sealed
with a septum. The microcosms were spiked with MTBE
and TBA to bring the concentrations in the pore water to
approximately 10 mg/L. The  ratio of water to solids in the
experimental system was approximately 2.1 to 1 (gm/gm).
The ratio of water to solids in a sandy aquifer is near 0.25
to 1 (gm/gm).  Because the experimental system had a high
water to solids ratio, and because both the TBA and soluble
electron acceptors such as sulfate were supplied in water,
the microcosm system was more sensitive than a natural
aquifer to the contribution of soluble electron acceptors and
less sensitive to the contribution of Iron(lll).

The concentrations of TBA, MTBE, methane, and sulfate in
monitoring wells at the sites that provided sediment for the
microcosm study of DeVaull  et al., (2003) are presented
in Table 1.   The water samples were collected  in 2002
and 2003. All the wells at all  three sites had much higher
concentrations of TBA than MTBE. There were high con-
centrations of methane in at least one well from each site.
As will be discussed later, the oxidation of TBA by sulfate
reduction requires 3.9 mg/L sulfate for each 1.0 mg/L of
TBA.  All the wells at the site at Fountain Valley, California,
and Laguna Miguel, California, had adequate concentrations
of sulfate to support complete metabolism of TBA by sulfate
reduction. The site at San Juan Capistrano, California, did
not have adequate sulfate to meet the demand associated
with the TBA.

The terminal electron accepting processes (TEAR) were not
defined in the treatment that simulated natural anaerobic
conditions in the aquifer. Sulfate and nitrate were available
at the beginning of the experiment, but DeVaull et al., (2003)
did not present data on depletion of nitrate or sulfate, or
accumulation of Iron(ll).

In microcosms constructed  with three different sediment
samples from a site in Fountain Valley, California, MTBE
was degraded in all three microcosms after 91  days of in-
cubation. In two of the three microcosms, TBA was either
completely degraded or substantially degraded within 161
days of incubation (Panel A of  Figure 4 presents data from
one of the microcosms).

In microcosms constructed  with three different sediment
samples from a site in San Juan Capistrano, California, TBA
was completely degraded in two of the three microcosms
within 114 days, but MTBE did not degrade within 154 days.
In the third microcosm, degradation followed the opposite
pattern; MTBE degraded within 70 days,  but TBA did not
degrade within 154 days.

In microcosms constructed  with three different sediment
samples from  a  site  in  Laguna  Miguel,  California, TBA
degraded within 90 days in one microcosm and 145 days
Table 1.       Distribution of TBA, MTBE, Methane, and Sulfate in Monitoring Wells at the three Gasoline Service
               Stations in Orange County, California that Provided Sediment for the Microcosm Study of DeVaull et
               al. (2003).
Station

Well

TBA
ml/L
MTBE
ml/L
Methane
ml/L
Sulfate
ml/L

Fountain Valley
Fountain Valley
Fountain Valley
16969 mw-6
16969 mw-15
16969 mw-13
28.0
2.0
14.8
0.079
0.015
0.012
0.0982
0.0293
0.535
183
128
79

Laguna Miguel
Laguna Miguel
Laguna Miguel
30011 mw-7
30011 mw-10
30011 mw-11
45.9
33.2
40.3
0.042
0.028
0.229
1.02
0.953
0.0176
1290
2640
2680

San Juan Capistrano
San Juan Capistrano
San Juan Capistrano
27101 w-2
27101 mw-8
27101 mw-11
10.8
12.0
69.9
0.230
0.002
0.084
4
1.56
0.835
11
17
6

-------
in a second microcosm, but MTBE did not degrade within
196 days in either of the two microcosms.  In the third mi-
crocosm, degradation followed the opposite pattern; TEA
did  not degrade within 196 days, but MTBE completely
degraded within 90 days (Panel B of Figure 4).

The results of the microcosm studies reported by DeVaull
et al., (2003) suggest that conditions that are favorable for
anaerobic biodegradation of TBA are widespread in Orange
County, California. At  least one microcosm from each of
the three sites degraded TBA under anaerobic conditions.
However, the microcosm studies also suggest that the ca-
pacity to degrade TBA may be heterogeneously distributed
at particular sites; TBA failed to degrade in at least one of
the microcosms constructed with one of the three different
sediment samples from each of the three sites.
                                        Microcosm 11
                                        of98UT010
                                               200
                  50        100       150
                    Time of Incubation (days)
                                200
Figure 4.
Removal of MTBE and production or
removal of TBA in anaerobic sediment
from a gasoline spill site in Fountain
Valley, California  (98UT010) and a
gasoline spill site in Laguna Niguel,
California (91UT086). Plotted from data
in DeVaull et al. (2003). Data are plotted
from one of three microcosm studies
constructed with sediment from each of
the sites.
Rates of Biodegradation of TBA in Ground
      Water

The rates of attenuation of TBA in sediment under aerobic
conditions are rapid. Schirmer et al.,  (2003) reported a
first order rate of removal of TBA in sediment from the
Borden field site of 0.12 per day. The initial concentration
of TBA was 0.7 mg/L. Wilson et  al., (2002) conducted a
microcosm study with sediment from a gasoline spill site
at Vandenberg AFB, California. The sediment was spiked
with MTBE, and then spiked again six more times after the
MTBE was degraded. The spiked concentrations of MTBE
were as high as 16 mg/L. The first order rates of removal of
MTBE were 0.05, 0.15, 0.03, 0.07, and 0.08 per day. TBA
was produced by metabolism of MTBE. In each cycle of
spike  and  removal of MTBE, the TBA never accumulated
to any significant extent. The rate of removal of MTBE was
the rate of production of TBA. TBA would  be  expected to
accumulate unless the rate of removal of TBA was faster
than the rate of removal of MTBE.

Kane et al., (2001) spiked and respiked MTBE into sediment
from a gasoline spill site in Palo Alto, California. Each spike
of MTBE was near 5 mg/L.  As the MTBE degraded, TBA
accumulated to concentrations near 2 mg/L. Once MTBE
was entirely degraded, the concentrations of TBA started to
decline. The rate of removal of TBA corresponded to first
order  rates of degradation of 0.10 and 0.13 per day.

The rates  of aerobic biodegradation are so fast that they
are effectively instantaneous in the context of ground water
movement at field scale. Shih et al., (2004) reported the dis-
tribution of the maximum concentration of TBA in monitoring
wells at gasoline spill sites in Los Angeles,  California. The
median of the TBA concentrations was 1,880 ug/L.  A first
order  rate of aerobic biodegradation of 0.1 per day would
bring  the median  concentration of TBA to the  California
Action Limit of 12 ug/L in fifty days.

The reported first order rates  of biodegradation of TBA
under  anaerobic conditions (as reported in Table 2) are
slower than under aerobic conditions. In their study of TBA
degradation in sediment from a refinery near Philadelphia,
Pennsylvania, White et al., (1986) reported their data as
a zero order rate of  utilization.  At initial concentrations of
1.7, 13.5, and 150 mg/L, the utilization rates were 0.057,
0.52 and 2.00 mg/L/day respectively.  A  first order  rate
constant can be thought of as a zero order rate of change
in concentration (AC/At) normalized to the concentration at
the particular instance in time over which the rate is operat-
ing. By definition of a first order rate constant (k),  first order
rate = AC/At = k*C; therefore k = (AC/At)/C). We estimated
the first order rate constants for attenuation of TBA at the
beginning  of the experiment by dividing the reported zero
order  rates of utilization by the initial concentration of TBA
in the pore space of the microcosms. At an initial concen-
tration of 1.7  mg/L and a zero order rate  of  utilization of
0.057 mg/L/day, the first order rate constant for degradation
was 0.057 mg/L/day divided by 1.7 mg/L or 0.0335 per day,
equivalent to a rate of 12 per year. At an initial concentration
of 13.5 mg/L, the first order rate constant was  14 per year.

-------
Table 2.       Rates of Anaerobic Biodegradation of TEA in Aquifer Sediments or Field Scale Plumes at Gasoline
               Spill Sites. If More than one Rate is Reported from Microcosm Experiments Constructed from
               Material from One Location at a Site, all the Rates are Reported in the Table.
Source of Material
Fountain Valley,
California
Laguna Miguel,
California
Potomac River
Philadelphia,
Pennsylvania
New York
Florida
Pennsylvania
San Juan Capistrano,
California
Pasadena, Texas
flow path from location 1 50
Pasadena, Texas
flow path from location 57 to
Pasadena, Texas
flow path from location 1 65
Field
Rate




X
X
X

X
X
X
Lab Rate
X
X
X
X



X



First Order Rate
of Biodegradation
(per year)
26,
17
18,
5.1
15
14,
12,
4.9
8.8
7.3
7.2
5.1
3.5
1.1
0.97
0.26
Reference
DeVaull et al. (2003)
DeVaull et al. (2003)
Finneran and Lovley
(2001)
White etal. (1986)
Kolhatkar et al.
(2000)
Kolhatkar et al.
(2000)
Kolhatkar et al.
(2000)
DeVaull et al. (2003)
Day and Gulliver (2003)
Day and Gulliver (2003)
Day and Gulliver (2003)
At an initial concentration of 150 mg/L, the first order rate
constant was 4.9 per year. As would be expected for a first
order process, the estimated first order rates of attenuation
were essentially independent of concentration.

Finneran and Lovley (2001) conducted spike and respike
experiments with TEA in sediment from the Potomac River.
The spiked concentrations were  approximately 50 mg/L.
The time course  of TEA degradation reported in Figure 3
of Finneran and Lovley (2001) corresponds to a first order
rate constant of 15 per year.

In the microcosm experiment conducted by DeVaull et al.,
(2003) the rates  of TEA biodegradation in the anaerobic
microcosms where TEA degraded were in the range of 26
to 3.5 per year (Table 2).  The rates in sediment samples
from Fountain Valley, California, were 26 and 17 per year.
The rates in sediment samples from San Juan Capistrano,
California, were 3.5 and 5.1 per year. The rates in  sedi-
ments from Laguna Miguel, California were 18 and 5.1 per
year.
There are also a few reports on the rate of TEA biodegra-
dation under anaerobic conditions at field scale (Table 2).
Kolhatkar et al., (2000) evaluated the distribution of TEA at
74 gasoline spill sites in the Eastern United States. They
used the approach  of Buscheck and Alcantar (1995) to
extract first  order rate constants for biodegradation of TEA
along the flow path in the plume. They were able to extract
rate constants that  were statistically significant at 95%
confidence  from three of the 74 plumes. The rate of TEA
biodegradation in  a  plume in New York was 8.8 per year,
the rate in a plume in Pennsylvania was 7.26 per year, and
the rate in a plume in Florida was 7.3 per year.

Day and  Gulliver (2003)  evaluated the natural attenua-
tion of a large plume of TEA at a chemical manufacturing
plant at Pasadena,  Texas. They compared the reduc-
tion in concentration of TEA with distance along the flow
path to the  reduction in concentration of the co-occurring
contaminants 1,1-dichloroethene and 1,1-dichloroethane.
The attenuation of TEA was significantly faster than that
                                                     10

-------
of 1,1-dichloroethene and 1,1-dichloroethane.  They used
the approach of Buscheck and Alcantar (1995) to extract
first order rate constants for natural biodegradation of TEA
along three flow paths in the aquifer. The rate constants
were 0.97, 0.26, and 1.1 per year.

The rates of TEA biodegradation under anaerobic condi-
tions (when biodegradation occurs) vary over  two orders
of magnitude  (Table 2).  The rates are faster than the rates
reported for anaerobic biodegradation of MTBE or for ben-
zene, which is the usual "risk-driver" at gasoline spill sites
(Table 3). If the microbial community at a particular spill
site acclimates to anaerobic biodegradation of TBA, and if
there is an adequate supply of the metabolic requirements
for anaerobic degradation, then natural anaerobic biodeg-
radation of TBA can provide a substantial contribution to
natural attenuation of TBA.

Extent of Biodegradation of TBA

The extent of TBA biodegradation by aerobic respiration,
nitrate reduction, Iron(lll) reduction orsulfate reduction will
depend  on the supply of these electron acceptors in  the
ground water or aquifer sediment. Wiedemeieretal., (1995)
compared the concentrations of oxygen, nitrate, Iron(ll)
and sulfate in ground water in the interior of 25 fuel spills
to the concentrations in the surrounding ground water that
had not been  affected by the fuel spill. They calculated the
consumption of each of the soluble electron acceptors and
the production of soluble Iron(ll) in the  ground  water from
insoluble Iron(lll) compounds  in the  aquifer solids.  Then
they compared the concentrations of electrons that were
transferred by each of the electron accepting processes on
a milliequivalent per liter basis. Their results are presented
in Figure 5.  Sulfate-reduction was the dominant electron
accepting process.

Figure 6 presents data on the concentration of sulfate in
monitoring wells at  77 gasoline spill sites in the Eastern
United States (unpublished data from  study published in
Kolhatkar et al., 2000). The figure compares the maximum
sulfate concentration in any well at each site and the mini-
mum sulfate concentration in  any well at each site.  The
maximum concentration is an estimate of the ambient
                                        concentration in ground water in the aquifer. The minimum
                                        concentration is an estimate of the sulfate concentration
                                        in the LNAPL source area of the gasoline spill.  At 75%
                                        of sites, the ambient concentration of sulfate was at least
                                        56 mg/L, at 50% of sites the concentration was at least
                                        107 mg/L, and at 25% of sites the concentration was at
                                        least 304 mg/L.

                                        Oxidation of TBA by sulfate reduction  requires 3.9 mg/L
                                        sulfate for each 1.0 mg/L of TBA:
C4H100
                                                        3S04-2
3S-2 + 4CO
5H2O
                                        Ambient concentrations of 56, 107 and 304 mg/L sulfate
                                        can support degradation of 14, 27, and 78 mg/L of TBA.

                                        In the most typical scenario at a gasoline spill site, a plume
                                        of MTBE in ground water is produced and sustained by
                                        continual dissolution of MTBE from the residual gasoline to
                                        ground water. The major portion of the TBA is produced by
                                        biodegradation of the MTBE once it is dissolved in ground
                                        water.  Other soluble components of gasoline,  such as
                                        the  BTEX compounds also dissolve in ground water.  The
                                        demand for oxygen, nitrate, and sulfate for biodegradation
                                        of the other components of the gasoline depletes sulfate
                                        and other soluble electron acceptors from the ground water
                                        that are in  contact with the residual gasoline.

                                        Based  on the supply of electron acceptors, and the  free
                                        energy available from the reaction, anaerobic regions
                                        in aquifers tend  to develop distinct  areas dominated by
                                        different electron acceptors (Finneran and Lovley, 2003;
                                        Wiedemeier et al.,  1999a).  When gasoline is spilled  into
                                        an aquifer, oxygen is depleted first, then nitrate is depleted,
                                        then sulfate is depleted, and finally the terminal electron
                                        accepting process  transitions to methanogenesis. As a
                                        consequence, the ground water immediately adjacent to
                                        a gasoline spill is often methanogenic. This region is  sur-
                                        rounded by ground water that is sulfate reducing, which in
                                        turn is surrounded by ground water that is lron(lll)-reducing,
                                        which in turn is surrounded by ground water that  is nitrate
                                        reducing. Although one electron accepting process tends to
                                        dominate, they often proceed concomitantly. If the supply of
                                        biologically available Iron(lll) is adequate, Iron(lll) reduction
Table 3.
Comparison of First Order Rates of Anaerobic Biodegradation of TBA, MTBE, and Benzene.

number of rates
mean (per year)
median (per year)
Reference
TBA
14
9.2
7.3
this report
MTBE
10
1.0
0.41
Wilson
(2003)
Benzene
20
3.7
Not Provided
Suarez and Rifai
(1999)
                                                     11

-------
                                                    Aerobic
                                                    Respiration 16%
                               Sulfate Reduction
                               48%
                                                         Nitrate Reduction
                                                         23%
Figure 5.     Relative importance of electron acceptors at 25 fuel spill sites in North America. Adapted from
              Wiedemeier et al., 1995.
                  10000
               w
                   1000 -
                     100 -
                      10 -
                                     -»- Maximum Concentration
                                     n Minumum Concentation
                                           nn
                          nrm 	mnrnn
                                              n i • Mm n - n m FISH
                                                                    1 1 • 1 1 rii
                                                                                -ffl
                         0
20
40          60
Percent of Sites
80
100
Figure 6.     Distribution of sulfate at 77 gasoline spill sites in the Eastern United States. The maximum
              concentration represents the likely concentration in ambient ground water that was not impacted
              by the spill. The minimum concentration represents the concentration in the LNAPL source area of
              ground water contamination. Unpublished data from Kolhatkar et al. (2000).
                                                   12

-------
can occur in ground water that is also sulfate reducing or
is methanogenic. Sulfate reduction and methanogenesis
can occur together, particularly at lower concentrations of
sulfate. These relationships are depicted diagrammatically
in Figure 7.
The  concentration of sulfate at which sulfate becomes
limiting for sulfate reduction varies widely from one organ-
ism to the next  and with  different environmental condi-
tions. Ingvorsen and J0rgensen (1984) found that the half
saturation constant for sulfate reduction in four strains of
bacteria varied from 0.5 mg/L to 7  mg/L.  Fukui and Takii
(1994) showed that the half saturation constant for sulfate
reduction of a Desulfovibrio desulfuricans was 0.8 mg/L
when the cells were associated  with particles of FeS, and
24 mg/L when the cells were free living. Concentrations of
sulfate less than 5 mg/L are considered sulfate poor (Fukui
and Takii, 1994), and flow of electrons can be expected to
transition to methanogenesis at sulfate concentrations less
than 40 mg/L (Personal Communication Kevin T. Finneran,
University of Illinois, August, 2006).
If sulfate has been depleted in contaminated ground water,
it is reasonable to presume that nitrate and oxygen are also
depleted (Finneran and Lovley,  2003; Lovley, et al., 1994;
Wiedemeier et al., 1999a).  In  62  of  the 77 sites in the
survey of gasoline spills in the Eastern United States, the
concentration of sulfate was less than 1.0 mg/L in the most
contaminated wells in the LNAPL source area (Figure 6).
Based on the available knowledge of anaerobic biodegrada-
tion of TEA, biodegradation of TEA supported by aerobic
respiration, nitrate reduction, or sulfate reduction would not
be expected in the source area  of these plumes.
If the biodegradation of TEA in ground water is limited by
the supply of sulfate, there should be an inverse relation-
ship between the concentration of TEA in  ground water
and the concentration of sulfate. Figure 8 compares the
concentration of sulfate to the concentration of TEA in 58
wells at 13 gasoline spill sites in Orange County, California.
These are the same wells sampled by Wilson et al., (2005c)
to determine whether stable isotope ratios in MTBE could
be used to predict MTBE biodegradation in  ground water.
In general, an inverse correlation did apply to the data set
from these wells (Figure 8), but there is more scatter than
would be expected.

Part of the scatter of the data may be an artifact caused by
mixing of ground water from different plumes in monitoring
wells.  Most  conventional monitoring wells  are screened
over ten or fifteen vertical feet (3.05  to 4.57  meters) in the
aquifer. Many plumes at gasoline spill sites are vertically
heterogeneous over ten to fifteen feet. This  relationship is
illustrated in Figure 9 using data from a gasoline spill site
at Port Hueneme, California. The plume was in a layer of
sands and gravels. Water samples were taken with tempo-
rary push wells with a vertical screened interval of 1.5 feet
(0.46 meters). Samples started just below the water table
and extended to a clay confining layer.  The ground water
in the shallow intervals had low concentrations of sulfate
and high concentrations of TEA while water in the deeper
interval had  high concentrations of sulfate  and  low con-
centrations of TEA.  However, water from a  well screened
across this aquifer would have intermediate concentrations
of both TEA and sulfate. The concentrations of TEA and
sulfate in the water produced by the  monitoring well would
                             Methane formation
                                                            Nitrate Reduction
                                     Iron(lll) Reduction
                 Aerobic Respiration
                                         Sulfate Reduction
Figure 7.     General distribution of terminal electron accepting processes (TEAPs) in ground water down gradient
               from a spill of gasoline.
                                                      13

-------
                 450
                               1000       2000       3000       4000
                                           Concentration Sulfate (mg/L)
5000
6000
Figure 8.      Distribution of the concentrations of TBA and sulfate in selected monitoring wells at gasoline spill sites
               in Orange County, California.
                                3

                               3.5

                                4

                               4.5
                            I
                            "S.
                            Q 5.5
                               6.5


                                7
                                               200           400           600

                                            Concentration of Sulfate (mg/L) or TBA (ng/L)
            800
Figure 9.      Relationship between depth below the water table and the concentrations of TBA and sulfate in
               ground water at a gasoline spill site in Port Hueneme, California.
                                                     14

-------
suggest that sulfate was available for anaerobic biodegra-
dation of TEA.

The TEA at this site was produced through biodegradation
of MTBE that dissolved out of residual gasoline near the
water table (data not shown).  Biodegradation of alkylben-
zenes (the BTEX compounds) associated with the residual
gasoline is probably responsible for the lower concentra-
tions of sulfate near the water table.  The mixed water
sample produced by a conventional monitoring well in this
aquifer would not have produced a water sample that was
representative of the water surrounding a particular sulfate
reducing bacterium.

In general, if water  produced from a monitoring well is
devoid of sulfate, then sulfate would  not be available to
bacteria in the ground water sampled by the well. The in-
verse is not necessarily true. The presence of both sulfate
and TBA in water produced by a monitoring well does not
mean that sulfate is available for organisms to degrade the
TBA. The sulfate and TBA may have come from different
depth intervals.

If sulfate is depleted in the source area of a gasoline spill,
anaerobic biodegradation of TBA will require admixture of
TBA in the contaminated ground water in the plume with
sulfate in the receiving ground water in the aquifer down gra-
dient. Mixing by dispersion is controlled by flow of ground
water and the geometry of  the plume.  The mixing occurs
across the interface  between the plume and the ambient
ground water. If a plume has a large volume relative to the
surface area presented to the ambient ground water, then
mixing by dispersion will be a slow process, and adequate
admixture of sulfate to meet the stoichiometric demand for
biodegradation of TBA will likely occur at some distance
away from the source area.

At most gasoline spill sites, the monitoring wells are on
the property of the gasoline station  or on the property of
the immediate neighbors.  The existing monitoring wells
may  not be located  in the portion  of the plume where
sulfate is available to support biological degradation  of
TBA. At many gasoline spill sites, it  may be impossible to
use monitoring wells that are  in or near the source area
to document natural  biodegradation  of TBA carried out by
sulfate reducing bacteria.

If TBA is not biologically degraded  under methanogenic
conditions, the only plausible agents for biodegradation of
TBA in ground water that has been depleted of sulfate and
the other soluble electron acceptors are Iron(lll) reducing
bacteria or Manganese(IV) reducing bacteria.   If sulfate
and the other soluble electron acceptors are depleted, the
extent of biodegradation of  TBA will be limited by the sup-
ply of biologically available Iron(lll)  or Manganese(IV) in
the aquifer sediments. At present there is a commercially
available assay for biologically available  Iron(lll) in sediment,
(Bioavailable Ferric Iron (BAFe III) Assay,  New Horizons
Diagnostics  Corp., 9110 Red Branch Road, Columbia,
Maryland USA 21045, 1-800-888-5015 ext. 0 or 235, fax:
410-992-0328, NHDiag@aol.com ).  The performance of
the assay to predict biologically available Iron(lll) has been
evaluated by the Environmental Security and Technology
Certification Program  (ESTCP, 2005).  However, to  the
authors' knowledge, the assay has never been  used to
evaluate natural attenuation of TBA in ground water. There
is no commercially available assay for biologically available
Manganese(IV).  Until  techniques to directly evaluate the
supply of biologically available Iron(lll) and Manganese(IV)
in aquifer sediment become standard practice, the potential
contribution of lron(lll)-reducing and Manganese(IV)-reduc-
ing bacteria to the natural attenuation of TBA cannot be
characterized. It is inappropriate to attribute apparent disap-
pearance of TBA from ground water to  lron(lll)-reducing or
Manganese(IV)-reducing bacteria simply because Iron(ll)
or Manganese(ll) accumulates in ground water.

EPA microcosm study of anaerobic TBA bio-
      degradation
To provide  an  independent evaluation of the natural  an-
aerobic biodegradation of TBA in ground water at gasoline
spill sites, EPA/ORD conducted microcosm studies using
material from several gasoline spill sites around the United
States. Sediment was  collected from BP gasoline stations
at  Petaluma, California; Deer Park, New York; Parsippany,
New Jersey; and Boca Raton, Florida. The microcosms
were part of a larger study that examined the anaerobic
biodegradation of MTBE and  ethanol.  Sediment was also
collected from  a motor gasoline spill site at a U.S. Navy
Base at Port Hueneme,  California, and a motor gasoline
spill site at Vandenberg Air Force Base, California. All the
samples were  from shallow aquifers in sandy unconsoli-
dated sediments.  Sediment from Deer Park, Parsippany,
and Boca Raton were selected for  the microcosm study
because field  data collected in 1999 indicated that  the
rates of degradation of MTBE and TBA  in ground water,
using the method of Buscheck and Alcantar (1995), were
statistically  significant at 80% confidence. Sediment from
Parsippany was selected because the field data indicated
that MTBE was being degraded, and TBA was not accumu-
lating. The sediments from Port Hueneme and Vandenberg
AFB were selected as  negative control sites. At the time
the microcosms were constructed, there was no evidence
of  biodegradation of MTBE or TBA  in  the plumes at  Port
Hueneme and Vandenberg AFB.

Construction of Microcosms
The sediment was collected and stored in 1 -L glass  jars.
To  protect the anaerobic microorganisms that might be
present in the samples  from oxygen in the atmosphere, the
head space above the sediment was replaced with ground
water. The  sediment samples were  shipped by air freight
and were stored at 4 °C until they were used to construct
microcosms.
All manipulations to prepare the microcosms were carried
out aseptically in an anaerobic glove box. An oxygen meter
indicated that the concentration of  oxygen in the atmo-
sphere of the glove box was  less than 1 ppm by  volume.
Microcosms were prepared in sterile glass serum bottles
with a volume of 25 ml. When available, ground water from
monitoring wells at the  sites was added to the sediment to
                                                    15

-------
make a slurry, and the sediment samples were stirred to
blend well. If ground water was not available, the slurry was
made with autoclaved reverse osmosis water.  The added
water was 5% or less of the final volume of the slurry. The
slurry was transferred to the serum bottles  with a sterile
scoop. Each microcosm received approximately 45 gm of
slurry and 1.0 ml of an aqueous dosing solution containing
a sterile aqueous  solution of TEA. The concentration of
TEA in the dose solution for microcosms constructed with
sediment from Deer Park, New York; Petaluma, California;
Vandenberg AFB, California; Parsippany,  New  Jersey;
and Boca Raton, Florida, varied from 13 to 15 mg/L. The
microcosms constructed  with sediment from  Port Huen-
eme, California, had 80 mg/L TEA. The microcosms were
sealed with a sterile Teflon-faced gray butyl rubber septum
and a crimp cap.  The microcosms were stored at room
temperature (20 to 22 °C) in the same glove box, under an
atmosphere  containing 2%  to 5% v/v hydrogen.

They were incubated from eighteen months to two years.
Every two to three months, triplicate microcosms were
selected for  analysis. Prior to sampling, the contents of
each microcosm were mixed with  a vortex mixer while the
microcosms were  still sealed, and then the sediment was
allowed to settle. The septum was removed and approxi-
mately 1 ml of the standing water was taken and diluted in
14 ml of distilled water containing 1% trisodium phosphate
as a preservative.  The diluted samples contained approxi-
mately 15 ml of diluted water and 6.0 ml of head space.
The diluted samples were sealed with a septum and crimp
cap, and  then shaken to  bring the water and head space
to equilibrium.

Laboratory Analytical Procedures

The concentrations of TBA were determined by head space
gas chromatography/mass  spectrometry (GC/MS) using
a modification of  EPA Method  5021 A, "Volatile Organic
Compounds in Various Sample Matrices using  Equilibrium
Headspace Analysis," June 2003.  Samples were collected
for analysis with an automated static headspace sampler.
Analytes  were determined  by gas chromatography/mass
spectrometry using an Ion Trap Detector. The  lowest cali-
bration standard was 10  ug/L; the method  detection limit
was 2.4 ug/L. Concentrations of sulfate were determined
with a Waters Quanta 4000 Capillary Ion Analyzer, using
a modification of EPA Method 6500, "Dissolved Inorganic
Anions in Aqueous Matrices by Capillary Ion Electrophore-
sis," January 1998. The method detection limit for sulfate
was 0.172 mg/L.

Depending on the amount of standing water that was
sampled from each microcosm, the pore water in the micro-
cosms was diluted  in a range between 15:1 and 30:1 before
analysis.  In the samples with concentrations of TBA below
the lower calibration limit, the pore water in the microcosms
was diluted 15:1 before analysis.  As a result, the effective
lower limit for calibration of TBA in the original pore water of
the microcosms was 150 ug/L, and the effective detection
limit was 36 ug/L. The effective method detection limits for
sulfate in diluted samples were 1.2 mg/L.
Data Quality

Laboratory analyses for data presented in Panel C  of
Figure 1, in Panel B of Figure 2,  Figure  6, Figures 8
through  13,  and Table 1 were conducted at the R.S. Kerr
Environmental Research Center in accordance with a
Quality Assurance Project Plan prepared for in-house task
10013 (Fate of Fuel  Oxygenates in Aquifer Material).

Major quality  assurance and quality control (QA/QC)
evaluations for the analyses included method blank (MB),
continuing calibration check (CCC),  second source check
(QC) using a sample obtained from the second source as
identified by their designated names, laboratory duplicates
(LD), and matrix spike (MS). A method blank was analyzed
in the beginning and end of a sample set. Continuing cali-
bration check standards (CCC) were analyzed every ten
samples as well as in the beginning and end of a sample
set.  QC checks were analyzed every ten  samples. Lab
duplicates were analyzed every ten samples. Matrix spikes
were analyzed every twenty samples.

The  data quality  objectives for TBA were  as  follows:
The target analyte in the method blank would be below
method  detection limit. The reported concentration of the
continuing calibration check standard (CCC) would agree
with the expected concentration plus or minus 20% of the
known concentration: the matrix spike would agree with the
expected concentration plus or minus 30% of the known
concentration (i.e., Recovery of the expected value would
be in the range of 70-130%).  Laboratory duplicates would
agree with each other with a relative percent difference of
± 25%.

The data quality objectives for sulfate were as follows: The
target analyte in the method blank would  be below method
detection limit. The reported concentration of the continuing
calibration check standard (CCC) and the QC check would
agree with the expected concentration plus or minus 10%
of the known concentration, the matrix spike would agree
with the expected concentration plus or minus 20% of the
known concentration (i.e., Recovery of the expected value
would be in the range of 80-120%). Laboratory duplicates
would agree with each other with a relative percent differ-
ence of  ± 10%.

Performance of the Quality Checks is presented in Tables
4 and 5  in the Appendix.  For analysis of TBA, 80 of the 81
continuing calibration check standards met the goal.  One
of the standards was 126% of the nominal concentration.
For analysis of TBA, 16 of 16 matrix spikes met the goal.
For the analysis of TBA, 16 of 18 laboratory duplicates met
the goal; the relative percent difference of one duplicate
was ± 25.2% and the relative percent difference of another
duplicate was ± 29.5%. In the 38 blanks for analysis of TBA,
reported concentrations were less than the lower calibration
limit, or less than the method detection limit, depending on
which concentration was reported by the analyst.

For analysis of sulfate, 92 of 93 continuing calibration check
standards met the goal.  For the analysis of sulfate, 19 of
19 matrix spikes met the goal. For the analysis of sulfate
                                                    16

-------
19 of 19 laboratory duplicates met the goal.  The reported
concentration of all 19 blanks was less than the method
detection limit.

There were 19 sample sets for analysis of  sulfate; 10 of
the sets were analyzed within 30 days, and the maximum
holding time for any set was 165 days.  There were 20
sample sets for analysis of TEA; 12 of the sets were ana-
lyzed within 30 days, and the maximum holding time for any
set was 145 days. Based on the reproducibility of data in
the container controls (as presented in Figures 10, 11, 12
and 13) the excess holding times do not appear to have
compromised the data quality for TEA.  All the data were
determined to be of acceptable quality, and the data were
used in the report.

Biodegradation of TBA in Microcosms

Of the six sites in the survey, TBA only degraded in micro-
cosms constructed from sediment from Petaluma, California
(Figure 10). There were three experimental treatments in
the study: microcosms constructed with sediment as col-
lected,  microcosms constructed with sediment  that had
been autoclaved to kill organisms that could biodegrade
TBA, and container controls that did not contain sediment.
Live microcosms were spiked to initial concentrations of
approximately 1,400 ug/L. Reductions in concentrations
were apparent after 91 days of incubation. After 182 days
of incubation, the concentration was down to the effective
detection limit in two of  the three replicate microcosms
sampled. After 273, 314 and 456 days of incubation, the
concentration of TBA was below the detection limit in all
the microcosms sampled. Disregarding the lag, the rate of
removal of TBA was 6.5 ±4.1 per year at 90% confidence.
This rate is in good agreement with the rates presented
in Table 2.

In one of the treatments, the sediment was autoclaved at
121 °C overnight in an attempt  to sterilize  the sediment
and then dosed.  There  was not significant removal  of
TBA over 734 days of incubation in the autoclaved control
microcosms.  One of  the experimental treatments was a
control for loss from the microcosm container. This control
consisted of the sterile glass serum bottle filled with sterile
water and dose solution,  and sealed with a sterile septum
and crimp cap.  There was no sediment in  the container
control.  As expected, there was no  loss of TBA from the
container control.

The pore water of the live microcosms contained 11.7 ±
0.2 mg/L sulfate at the start of the experiment. At the time
when the TBA was entirely consumed, the sulfate concen-
tration was 15.9 ± 0.1  mg/L. As was discussed earlier, oxi-
dation of TBA by sulfate reduction requires 3.9 mg/L sulfate
for each 1.0 mg/L of TBA. The sulfate demand to oxidize
1.4 mg/L TBA in the live microcosms would be 5.5 mg/L.
It is not entirely clear why sulfate concentrations appeared
to increase over time. The method for analysis of sulfate
is recalibrated when calibration check standards are off by
10%.  It is possible that sulfate desorbed from the anion
exchange complex in the sediment. In any case, there were
adequate concentrations of sulfate to meet the theoretical
demand for sulfate reduction of the TBA. However, the
sulfate concentrations were low. They were near the half
saturation constant for sulfate reduction.  It is equally plau-
sible that TBA in the sediment from Petaluma was degraded
by Iron(lll) or Manganese(IV) reducing bacteria.

Panel A of Figure 11 presents data for the sediment from
Deer Park, New York. In contrast to the behavior of TBA in
sediment from Petaluma, California, there was no sustained
or significant removal of TBA in the live microcosms, in the
control microcosms,  or in the container controls. The esti-
mated rate of MTBE biodegradation in the plume at Deer
Park was 5.3 ± 3 per year at 95% confidence (Kolhatkar
et al., 2001).  However, the companion microcosm study
on anaerobic biodegradation of  MTBE also failed to show
any degradation of MTBE (Wilson et al., 2005b).
The maximum concentration of TBA in the live microcosms
was 350 ug/L.  If 10% of the TBA is converted to microbial
biomass, and if the average dry weight of a microbial cell
is 10~12 g, then complete consumption of 350 ug/L would
produce 3.8 x 107 cells per liter. The  microcosms were
incubated for 740 days. If initially, there was only a single
TBA degrading bacterium in each  microcosm,  and the
growth rate of the organism was at least 3% per day, the
organisms would have grown to consume the TBA in the
incubation period.
The concentration of sulfate in the pore water of the Deer
Park microcosms was less than 1.0 mg/L at the beginning
of the incubation period and at  the end  of the incubation
period. However, the theoretical  demand for sulfate for TBA
biodegradation would have only  been 1.4 mg/L. It is highly
likely, but not definitely established, that concentrations
of sulfate limited TBA  biodegradation  in the  microcosms
constructed with sediment from the Deer Park site.
In microcosms constructed from  sediment from Parsippany,
New Jersey, the  concentration  of TBA in the live micro-
cosms increased  approximately tenfold  during 744 days
of incubation (Panel B of Figure 11).  This was probably
due to TBA formed by anaerobic biodegradation of MTBE
in residual gasoline in the sediment used to construct the
microcosms (Wilson et al., 2005c). If TBA was degraded
in the sediment, the rate  did  not exceed the rate of TBA
production from biodegradation of MTBE. As expected,
there was no loss of TBA from the killed control microcosms
or the container controls.

Similarly, there was  no evidence of  removal  of TBA over
730 days of incubation in microcosms  constructed with
sediment from the Boca Raton site (Figure 12). The initial
concentration of TBA was 1.6 mg/L.  The theoretical de-
mand for sulfate  was 6.2 mg/L. The initial concentration
of sulfate was 4.8 ±  1.9 mg/L.  The final concentration of
sulfate was 3.4,  1.3, and <0.3  mg/L  respectively in the
triplicate microcosms.  It is likely that sulfate was limiting
for TBA biodegradation in the microcosms constructed with
sediment from the Boca Raton site.

Finally, there was no evidence  of TBA biodegradation in
microcosms constructed with sediment from the Port Huen-
eme site (Panel A of Figure 13) or the Vandenberg AFB site
                                                    17

-------
                               10000
                                1000
                                100 •
                                 10
                                                  Petaluma
                                                                         • Living Microcosms
                                                                        — Living Microcosms
                                                                           Average
                                                                         n Control Microcosms
                                                                         A Container Control
Figure 10.
                                    0     100     200     300     400     500     600     700     800
                                                      Time of Incubation (Days)

                 Removal of TBA in microcosms constructed with material from a gasoline spill site in Petaluma,
                 California.

                               10000
                                1000

                                 100
                                  10
                                                        Living Microcosms
                                                        Control Microcosms
                                                        Container Controls
                                       * i  *
                                       "
                                                                      Deer Park
                                     0      100    200    300    400    500    600    700    800
                              100000
                               10000
                             •3.
                             ^
                                1000
                                 100
                                  10
                                         B
                                       s
                                            Living Microcosms
                                            Control Microcosms
                                            Container Controls
                                                                        Parsippany
Figure 1 1 .
                                     0     100    200    300    400    500    600    700    800
                                                      Time of Incubation (Days)
                 Behavior of TBA in microcosms constructed with material from gasoline spill sites in Deer Park, New
                 York and Parsippany, New Jersey.
                                                            18

-------
                           10000
                            1000
                         I
                             100 •
                              10
                                         » Living Microcosms
                                         n Control Microcosms
                                         ' Container Controls
                                              I
                                                                             Boca Raton
                                      100    200    300    400     500     600
                                                    Time of Incubation (Days)
                                                                                 700     800
Figure 12.     Behavior of TBA in microcosms constructed with material from a gasoline spill site in Boca Raton,
                Florida.

                        100000
                         10000
                          1000
                           100
                            10
                         10000
                                                                 Living Microcosms
                                                                 Control Microcosms
                                                              Port Hueneme
                                        100       200       300       400       500       600
                                  B
§ 8 5
(1/61) V81
S n H B H n
* ! i ' s
• Living Microcosms j
o Control Microcosms Vandenberg |
A Container Control I

                                     100      200      300      400      500
                                                  Time of Incubation (Days)
                                                                                600
                                                                                        700
Figure 13.     Behavior of TBA in microcosms constructed with material from a gasoline spill site at Port Hueneme,
                California and a site at Vandenberg AFB, California.  These are sites where natural anaerobic
                biodegradation of TBA was not expected.
                                                        19

-------
(Panel B of Figure 13). Both of these aquifers are naturally
anaerobic and sulfate reduction is the dominant electron
accepting process. Unfortunately, data on concentrations of
sulfate were not collected in the microcosm experiments.

In summary, the microcosm experiments did not present
compelling evidence that anaerobic TBA biodegradation
was widespread at gasoline spill sites. At three of the sites,
the microcosm studies failed to confirm initial expectations
that TBA was being degraded at the site based on field
monitoring data.

Use of Stable Isotope Ratios

In recent years an alternate approach has been developed
to recognize natural biodegradation of  MTBE in  ground
water (Hunkeler et al., 2001; Gray et al., 2002; Kolhatkar
et al., 2002; Kuder et al., 2005; Wilson et al., 2005b; Zwank
et al., 2005). Organic compounds contain two stable iso-
topes of carbon. Carbon with a weight  of twelve daltons
(12C) is approximately one hundred times more abundant
than carbon with a weight of thirteen daltons (13C). During
biodegradation, MTBE molecules with 12C  in the methyl
group are degraded more rapidly than MTBE molecules with
13C in the methyl group. Over time 13C accumulates in the
residual MTBE molecules that have not been degraded, and
the extent of biodegradation of  MTBE can be inferred by an
increase in the ratio of 13C to 12C in the residual MTBE.

The use of stable isotopes has two important advantages. It
uses the compound of interest as its own tracer. Molecules
composed with 13C and molecules with 12C share a common
source and should have the same behavior with respect to
sorption, hydrodynamic dispersion, or dilution in a monitor-
ing well. A second and more important advantage is that the
stable isotope ratios recognize and validate biodegradation
that has actually occurred in the aquifer. Microcosm studies
done in the laboratory with sediment samples only establish
a potential for biodegradation in the aquifer.

The behavior of stable isotopes is described by two param-
eters: 813C and s. The 813C of a compound is a measure of
the ratio of 13C atoms to 12C atoms in the molecule. It will
be defined formally later  in the text. The usual pronuncia-
tion of 813C is "delta thirteen sea."  During  the course of
biodegradation, the compound that is still remaining will
have more of the heavy isotope 13C, and the value of 813C
will become more positive. The value of & (epsilon) relates
the change in 813C to the fraction of the original contaminant
that is still remaining. As a consequence, & is often called
the isotopic enrichment factor.

The ratio of isotopes is determined with an isotope ratio
mass spectrometer.  The mass  spectrometer does not
measure  the ratio of the  stable carbon  isotopes to each
other.  Rather, it measures the deviation of the  ratio in
the sample from the ratio of a standard  used to calibrate
the  instrument.  The substance used as the international
standard  for stable carbon isotopes has a ratio of 13C to
12C of 0.0112372.
The conventional  notation for the ratio of 13C to 12C in a
sample (813C) reports the ratio in terms of its deviation from
the ratio in the standard.
      813C =
            (13c/12c)    -(13c/12c)
            V   '    /samp/e  V   '    /:
                                   standard
                    (13c/12c)
                    \   '    'stand
xlOOO
The units for 813C are parts per thousand, often represented
as %0, or per mil, or per mill.

The extent of isotopic f ractionation is typically determined by
a linear regression of the 813C in MTBE on the natural loga-
rithm of the fraction of MTBE remaining after biodegrada-
tion. The slope of the regression line is termed the isotopic
enrichment factor (e.).  Fractionation of carbon in MTBE is
much greater during anaerobic biodegradation compared
to biodegradation during aerobic respiration.  Somsamak
et al., (2006) reported values for & during biodegradation of
MTBE under methanogenic conditions of -13.3%0 to -14.6%0,
and values under sulfate reducing conditions of -13.4%0
to -14.6%0.  In contrast, Hunkeler et al., (2001) reported
enrichment factors  for aerobic biodegradation that  varied
from -1.52%0 ± 0.06%0 to -1.97%0 ± 0.05%0, and Gray et al.
(2002) determined the enrichment factors that varied from
-1.4%0 ± 0.1 %0 to -2.4%0 ± 0.3%o.

Figure 14  compares the extent of f ractionation of  MTBE
that would be expected during biodegradation under aero-
bic and  anaerobic  conditions.  The most  negative value
for e was plotted because this value  will predict the least
biodegradation of MTBE with an increase in the value of
813C in MTBE. The most negative value for e is the most
conservative value.

Any change in the  value of 813C in MTBE caused by  bio-
degradation must be compared to the normal variation of
813C in MTBE used to make  gasoline.  Smallwood  et al.,
(2001) reported that the normal range of 813C for MTBE in
gasoline is from -28.3%0 to -31.6%0; more  recent surveys
indicate  that the normal range extends between  -27.5%0
and -33%0 (O'Sullivan et al., 2003). This natural variation
is represented as a filled arrow in Figure 14.  The natural
variation in 813C in MTBE is of the same order as the frac-
tionation that would be expected after 99% of the original
amount of  MTBE had been degraded  under aerobic  condi-
tions.  As a consequence of this weak fractionation, it  has
been difficult to document natural aerobic biodegradation of
MTBE at field scale. However, the variation in 813C during
anaerobic  biodegradation is much larger than the variation
in 813C in  MTBE in gasoline, and determinations of 813C
in MTBE have been used to  document natural anaerobic
biodegradation  of MTBE at field scale (Kolhatkar  et al.,
2002; Kuder et  al., 2005; Wilson et al., 2005a; Wilson et
al., 2005b; Zwank et al., 2005).

Wilson et al., (2005a) used stable isotope ratios to evaluate
production of TBA from natural biodegradation of MTBE at
thirteen gasoline spill sites in Orange County, California. At
                                                     20

-------
                          70
                          50
                          30
                      O
                      n
                      "to
                          10
                         -30
                         -50
                   -A- TBA anaerobic biodegradation

                   -CD- MTBE anaerobic biodegradation

                   -*- MTBE aerobic biodegradation
                                      y = -14.59Ln(x)- 29.01
S                                    Natural Range in
                                    MTBE in Gasoline
                                                   y = -0.73Ln(x) - 28.41
                                                                                      0.001
Figure 14.
             1                   0.1                  0.01

                         Fraction MTBE or TBA remaining (C/Co)

Expected relationship between the ratio of stable isotopes of carbon in MTBE or TBA and the extent
of biodegradation of MTBE or TBA.
these thirteen sites in Orange County, the concentrations
of MTBE and TBA are not constant over time. Figures 15
and 16 present data from wells at three separate sites that
illustrate common patterns.  At some wells, at some sites,
concentrations of MTBE are greater than concentrations
of TBA at the beginning of the  monitoring record.  Over
time, the concentrations of MTBE go down by orders of
magnitude and the concentrations of TBA increase by
orders of magnitude (Figure 15). Occasionally there was
an almost stoichiometric replacement of  MTBE with TBA
(Figure  16 and  17).  An evaluation of the ratio of stable
isotopes of carbon in the residual MTBE demonstrated that
biodegradation of MTBE to TBA could explain the increase
in concentrations of TBA (Wilson et al., 2005a).

In some wells, at four of the thirteen sites, the high con-
centrations of TBA persisted for a period and then the
concentrations of TBA also declined over time (illustrated
by Figure  17). The pattern was very similar to the pattern
seen in the microcosm studies conducted with sediment
from sites in Orange County, California (Figure 4, present-
ing data from DeVaull et al., 2003).

Studies with enrichment cultures typically follow the same
pattern (Bradley et al., 1997; Suflita et al., 1997; Walter,
1997). The initial density  of active organisms in the in-
oculum  is low and the activity of  the organisms  has no
perceptible effect on  the concentration of their substrate.
As the organisms grow and increase in number, they even-
tually reach a density where their activity has a perceptible
effect on the concentration of substrate.  As they continue
to grow and increase in numbers they eventually exhaust
the substrate.
                                        The pattern shown in Figure 4 for a laboratory microcosm
                                        and in Figure 17 for a monitoring well at field scale can be
                                        explained as a lag period of MTBE biodegradation during
                                        microbial acclimation to MTBE, followed by biodegradation
                                        of MTBE to TBA, then a lag during microbial acclimation to
                                        TBA, followed by biodegradation of TBA. However, there
                                        are other explanations for the pattern at field scale. The
                                        supply of MTBE may have been exhausted in the source
                                        area and as a result there was no continuing supply of
                                        MTBE to produce TBA.  As a result, the concentrations
                                        of TBA decreased because the TBA that was previously
                                        produced was swept away by advective flow of  ground
                                        water in the aquifer.  It is  also possible that the direction
                                        of ground water flow in the aquifer changed and the new
                                        flow path to monitoring well MW-4S at site 88UT138 did
                                        not contain MTBE or TBA.  Because there are alternative
                                        explanations for the pattern illustrated in Figure 17, field
                                        monitoring data cannot provide unequivocal evidence that
                                        TBA is being biologically degraded in ground water.

                                        It would  greatly simplify the evaluation of natural biodegra-
                                        dation of TBA at field scale if the ratios of stable isotopes
                                        could be used to recognize and quantify TBA biodegradation
                                        in ground water.  However, the authors  are not aware of
                                        any report comparing fractionation of TBA under anaero-
                                        bic conditions in either microcosm studies or enrichment
                                        cultures. The only data available on  the fractionation of
                                        carbon isotopes during biodegradation of TBA are from a
                                        study of a field-scale plume and the data suggest that the
                                        fractionation coefficient is small.  Day et al.,  (2002) and
                                        Day and Gulliver (2003) describe a detailed evaluation of
                                        natural TBA biodegradation in ground water at a chemical
                                                    21

-------
  1000


   100
 o
 0  0.1
   0.01
  0.001
                               OOUT038 MW-08
     Oct-00   Apr-01    Nov-01    May-02   Dec-02   Jun-03
  1000


   100


la   10
.§
d
8    1

   0.1


  0.01
                         87UT211 MW-5
     Mar-00    Oct-00    Apr-01    Nov-01    May-02   Dec-02
  1000
   100
    10
     1
          85UT114MW-25
     Jun-99
             Jun-00    Jun-01     Jun-02
                       Date of Sampling
Jun-03
Jun-04
Figure 15.
               Changes in concentrations of MTBE
               and TBA in monitoring wells over time at
               three sites in Orange County, California,
               data plotted on logarithmic scale.
                                                                             Nov-01    May-02    Dec-02
                                                                                                        Jun-03
                                                                    Oct-00    Apr-01    Nov-01    May-02    Dec-02
                    Jun-99    Jun-00    Jun-01     Jun-02    Jun-03    Jun-04
                                       Date of Sampling


                Figure 16.     Changes in concentrations of MTBE
                                and TBA in monitoring wells over time at
                                three sites in Orange County, California,
                                data plotted on arithmetic scale.
manufacturing plant at Pasadena, Texas. They compared
the concentrations of TBA in the plume to the 813C of TBA.
Figure 18 in this report plots the data provided in Figure
30-7 of Day and Gulliver (2003).  Following the approach
of Kolhatkar et al., (2002), a regression of 813C of TBA on
the natural logarithm of the concentrations of TBA in the
plume was used to estimate the fractionation coefficient
(e.) for biodegradation of TBA at field scale.  The value of
e was -0.73%o.
                                                        Figure 14 compares the extent of fractionation of TBA that
                                                        would be expected during biodegradation under anaerobic
                                                        conditions to the extent of fractionation of MTBE that would
                                                        be expected under aerobic and anaerobic conditions.  If
                                                        the value of the  enrichment coefficient during anaerobic
                                                        biodegradation of TBA is  -0.73%0, then minimal change
                                                        can be expected in the value of 513C for TBA, even when
                                                        the mass of TBA is reduced by two orders of magnitude
                                                        through biodegradation of TBA.
                                                      22

-------
                          10
                       -B  6
                       |
                       "jo

                       §  4

                       <§
                                   88UT138MW-4S
                          Jan-01
Jan-02            Jan-03
       Date of Sampling
Jan-04
Figure 17.     Changes in concentrations of MTBE and TBA in monitoring wells over time at a site in Orange
               County, California, where a decline in concentration of MTBE was followed at a later time by a decline
               in concentration of TBA.
The prediction of the enrichment factor e for anaerobic bio-
degradation of TBA in Figure 18, as projected in Figure 14,
is consistent with field observations of 813C for MTBE and
813C for TBA at  field sites in  Orange County, California,
(Kuder et al., 2005) and a large field site in South America
(Zwank et al., 2005). Figure 19 compares the total variation
in 813C in TBA in ground water samples in monitoring wells
to the variation in 813C in MTBE in the same well.  In the
data reported by Kuder et al., (2005) for Orange County,
the total  variation of 813C for TBA was 6.76%0 while the
total variation of 813C for MTBE was 89.5%0 (Panel A of
Figure 19). In the data reported by Zwank et al., (2005)
for South America, the total variation of 813C for TBA was
5.6%0 while the total variation of 813C for MTBE was 66.7%0
(Panel B of Figure 19). There was substantial anaerobic
biodegradation of MTBE in these studies, as documented
by extensive  fractionation of  carbon isotopes in MTBE.
There was little fractionation of TBA. This may have been
a result of little biodegradation of TBA.  It may also have
been the result of the much weaker fractionation of TBA
during anaerobic biodegradation.

To properly interpret the variation of 813C for TBA in ground
water, it will be necessary to compare the measured value
of 813C for TBA to the natural variation of 813C for TBA. To
the authors' knowledge, data on the natural variation in 813C
for TBA has not been  published.  In fact, two  distributions
on natural variation would be  necessary to interpret  813C
for TBA in ground water, (1)  the natural variation of 813C in
TBA in gasoline,  and (2) the natural variation of 813C in the
tertiary butyl functional group of MTBE in gasoline. Zwank
et al., (2005) determined the 813C of a sample of MTBE,
then hydrolyzed the MTBE to TBA and determined the 813C
of the TBA produced.  The 813C of the MTBE was -28.13 ±
0.15%0, while the 813C of the TBA was-25.49 ±0.10%0. The
difference between the 813C of MTBE and the 813C of the
tertiary butyl functional group in MTBE is one half as wide
            as the entire reported range of variation of 813C of MTBE
            in gasoline (-27.5%0 to -33%0;   reported in O'Sullivan et
            al., 2003). It will not be appropriate to use the variation of
            813C of MTBE in gasoline as a surrogate for the variation
            of 813C of TBA.

            It is possible that future  research will show that the true
            coefficient of fractionation of TBA during anaerobic bio-
            degradation is  considerably more  negative than -0.73%0.
            Until that work is published, and until the distribution of
            the natural variation of 813C of TBA is published,  it will be
            difficult or impossible to use measurements of 813C of TBA
            in ground water to  estimate the  extent of biodegradation
            of TBA at field sites.

            In contrast to the limited variation in 813C of the TBA, there
            was much wider variation of isotope ratios of hydrogen in
            TBA (expressed as  8 D, the ratio  of deuterium to hydrogen
            one as measured against a standard reference material for
            hydrogen isotopes). Data reported by Zwank et al., (2005)
            are presented in Figure 20. Although values of 813C in TBA
            varied by 3.9%0 in the population of wells, values of 8 D in
            TBA varied by  67.9%0. Future research  may establish  a
            relationship between 8 D  and the extent of biodegradation
            of TBA in ground water; however, there are concerns that
            isotopic exchange of the hydrogen in the alcohol functional
            group  with hydrogen in water may alter 8 D in TBA, mak-
            ing a straightforward interpretation of 8 D in TBA difficult
            (Zwank et al., 2005).

            In summary, there  is  not  a good technique to determine
            biodegradation of TBA at field scale.  At most field sites
            there will not be enough monitoring wells, or the wells will
            not be in the right place to document anaerobic biodegra-
            dation along a flow  path in the aquifer. The application of
            stable isotope ratios to TBA in ground water (at the present
            level of development) cannot resolve TBA biodegradation
            at field scale.
                                                     23

-------
            -20

            -21

            -22 •

            -23

         A  -24

         K  -25
         c
         O  -26
         GO
            -27

            -28

            -29 •

            -30
                                                              y = -0.734Ln(x) - 24.782
                             0.1
                                                          10
                                                       TBA(mg/L)
                                                                        100
                                                                                      1000
Figure 18.
 Relationship between the isotopic ratio of carbon in TBA in ground water at a manufacturing facility
 in Pasadena, Texas, and the concentration of TBA. The slope of the line is an indirect estimate of the
 isotopic fractionation factor (s) for anaerobic biodegradation of TBA at the site.  Adapted from Day and
 Gulliver (2003).
                                o
                                     60 n
                                     40
                                     20
                                    -20
                                    -40
                             Kuderetal. (2005)


                             A

                                      -40     -20      0

                                     60 n	

                                            Zwank et al. (2005)
                                                               20
                                                                       40
                                                                               60
                                     40


                                     20
                                2
                                fS     o
                                "
                                    -20


                                    -40
                             B
                                      -40
                                              -20
                                                       0       20

                                                        13C MTBE (%„)
                                                                       40
                                                                               60
Figure 19.
Comparison of the range of variation in the isotope ratio of carbon (813C) in TBA at monitoring wells at
sites in Orange County, California, (Kuder et al., 2005) and at a site in South America (Zwank et al.,
2005) to the range of variation in the isotope ratio of carbon (813C) in MTBE in the same monitoring
wells.
                                                        24

-------
                        0.0'
                      -20.0
                  Si
                      -40.0'
                      -60.0
                      -80.0
                     -100.0 •
                                                                                 f
                     -120.0
                     -140.0
                        -140.00    -120.00
                                           -100.00    -80.00     -60.00

                                                     813C TBA (%.)
                                                                        -40.00
                                                                                  -20.00
                                                                                            0.00
Figure 20.    Comparison of the range of variation in the isotope ratio of hydrogen (8D) in TBA at monitoring wells
               at a site in South America (Zwank et a/., 2005) to the range of variation in the isotope ratio of carbon
               (8J3Cj in MTBE in the same monitoring wells.
                                                       25

-------
                                               Summary
Tertiary butyl alcohol (TEA) is widely distributed at gasoline
spill sites, and is present at high concentrations.  As an ex-
ample, Shih et al., (2004)  reported that TEA was detected
in ground water at 61.1% of sites in Los Angeles County,
California. The concentration of TEA was equivalent to the
concentrations of methyl tertiary butyl ether (MTBE) and
benzene. The mean and median concentration of TEA was
30,100 and 1,880 ug/L, compared to 44,800 and 1,200 ug/L
for MTBE and 83,800 and 1,370 ug/L for benzene.

At a major fraction of sites,  the concentration  of TBA  in
ground water is greater than can plausibly be expected
from the  TBA that was a  constituent  in the gasoline that
was originally spilled (Wilson, 2003; Wilson et al., 2005a;
this report Figure 2).  Based on the ratio of TBA to  MTBE
at gasoline spill sites in Orange County, California, Wilson
et al., (2005a) determined that TBA resulting from the bio-
degradation of MTBE could explain the concentrations  of
TBA at 85% of the sites.

A review of available literature indicated that  microorgan-
isms can degrade TBA  using  oxygen, nitrate, Iron(lll),
Manganese(IV), or sulfate as a terminal electron acceptor
(Bradley  et al., 2002).  The current consensus  opinion  is
that an electron acceptor is necessary, and that TBA cannot
be directly fermented to produce methane (Schmidt et al.,
2004a). Although TBA does not degrade under  anaerobic
conditions in  many laboratory experimental systems, the
average rate of anaerobic biodegradation of TBA (when it
does occur) is faster than the average rate of  anaerobic
biodegradation of MTBE.  The average first order  rate  of
anaerobic biodegradation  of TBA in the studies reported in
Table 2 was 9.2 per year compared to an average  rate  of
anaerobic biodegradation of MTBE of 1.0 per year (Wilson,
2003; Table 3).

  Data on the availability of electron acceptors in ambient
ground water and the concentrations of electron  acceptors
in the source area of plumes indicates that sulfate is the
most important electron acceptor at fuel spill sites (Wiede-
meier et al., 1995). Data collected for a study published by
Kolhatkar et al., (2000), but presented for the first time  in
this report,  indicates that sulfate, and  presumably oxygen
and nitrate, are entirely depleted in the source area of ap-
proximately 80% of gasoline spill sites.

As a consequence,  anaerobic biodegradation  of TBA  at
many particular field sites will be limited by the availability of
sulfate and the rate of TBA biodegradation will be limited by
the rate that sulfate is supplied to the plume by diffusion and
dispersion.  This rate of  supply can be orders of magnitude
slower than the rate of TBA biodegradation in laboratory
microcosm studies where sulfate is not limiting.

There  is a strong possibility that  Iron(lll) reducing and
Manganese(IV) reducing bacteria degrade TBA in ground
water that is depleted of soluble electron acceptors.  How-
ever, at the current state of practice, it is impossible to evalu-
ate the contribution of Iron(lll) reducing and Manganese(IV)
reducing bacteria to the natural biodegradation of TBA in
ground water at gasoline spill sites.

The OSWER Directive on MNA (U.S. Environmental Protec-
tion Agency, 1999) identifies three lines of evidence that
can be used to support the selection of MNA as a remedy.
The first line of evidence is Historical groundwater and/or
soil chemistry data that demonstrate a clear and meaningful
trend of decreasing contaminant mass and/or concentration
over time at appropriate monitoring or sampling points. The
second line of evidence is Hydrogeologic andgeochemical
data that can be used to demonstrate indirectly the type(s)
of natural attenuation processes active at the site, and the
rate at which such processes will reduce contaminant con-
centrations to required levels. Until techniques are applied
that can estimate the supply of biologically available Iron(lll)
or Manganese(IV) in aquifer sediment,  it is not possible to
compare the supply of these insoluble  electron acceptors
to the demand for electron acceptors provided by TBA or
other organic materials in contaminated ground water. It
is not possible to provide the second  line of evidence for
natural biodegradation of TBA by iron-reducing or manga-
nese-reducing bacteria.
Data on the fractionation of carbon isotopes in TBA during
anaerobic biodegradation are indirect data from a field study
instead of direct measurements from controlled laboratory
studies and the data are available from only one site (Day
et al.,  2002; Day and Gulliver, 2003). The fractionation
reported at the one site available in the literature is weak;
suggesting that fractionation of carbon isotopes will not be
generally useful to recognize anaerobic biodegradation of
TBA at field sites.  Fractionation  of hydrogen isotopes in
TBA is  much stronger  (Zwank et al., 2005) and may be
useful  in the future.  However, there is a concern about
isotopic exchange of the hydrogen in the -OH function of
TBA with hydrogen in water.  At  this writing,  there  is  not
a consensus on the appropriate interpretation of shifts in
the ratio of hydrogen isotopes of TBA.  However, isotopic
fractionation during biodegradation is  an active area  of
research and advances in the state of  knowledge can be
expected in the future.
                                                     26

-------
                                        Recommendations
If it is necessary to evaluate natural biodegradation of TEA
at a gasoline spill site, do not rely on the data from con-
ventional monitoring wells.  It is necessary to obtain data
on the vertical distribution of the  concentrations of TEA,
sulfate, benzene, and methane in  ground water. This can
be accomplished using the following steps. Collect ground
water with push tools that sample a narrow vertical inter-
val (six inches to two feet; 0.15 m  to 0.61 m).  Extend the
vertical profile from the water table into clean water below
the plume. Use the concentrations of benzene as a tracer
for the contaminated ground water that might contain TEA.
If the  benzene has  been biologically degraded and is not
present in the ground water, use concentrations  of methane
as a tracer for the plume. The strongest and  most direct
evidence for natural biodegradation of TEA is  a series of
sampling locations  or monitoring  wells  down  gradient of
the  source area  that have high concentrations of tracer
compounds but are devoid of TEA.

If unacceptable concentrations of  TEA are still present in
ground water from the down  gradient wells, compare the
concentrations of TEA to the supply of sulfate  as an elec-
tron acceptor for biodegradation of TEA.  As a loose rule of
thumb, the stoichiometric demand for sulfate as  an electron
acceptor for complete oxidation of a fuel component is four
times  the concentration of the fuel component. There is a
reasonable prospect for natural attenuation of TEA through
natural anaerobic biodegradation  if  the concentration of
sulfate exceeds four multiplied by  the sum of the concen-
trations of TEA, MTBE, benzene,  toluene, ethylbenzene,
xylenes, and trimethylbenzenes.

Many risk evaluations at gasoline  spill sites use a simple
transport and fate model such as BIOSCREEN (Newell et
al., 1996) or the calculations in ASTM E-1739, Risk-Based
Corrective Action (RBCA) at Petroleum  Release Sites,
issued by the American Society for Testing and Materials
Standards (2002).  Unless or until it has been  shown that
a sufficient supply of electron acceptor is available to meet
the stoichiometric demand for TBA and the other organic
compounds  in  the  plume, assume that TBA  will not be
biologically degraded in the ground water plume and that
natural attenuation of TBA will be due exclusively to hydro-
dynamic dispersion.

Do not assume that TBA is  being biologically degraded
under anaerobic conditions in ground water because there
is evidence that natural anaerobic biodegradation of MTBE
is occurring. The same organisms are not degrading MTBE
and TBA. The initial step in anaerobic MTBE biodegrada-
tion is carried out by acetogenic bacteria that use MTBE
to consume molecular hydrogen as part of their energy
metabolism. The anaerobic biodegradation of TBA is most
commonly carried out by sulfate reducing bacteria.

A risk evaluation for TBA may indicate that it is necessary
to remediate TBA in ground water.  Long term monitoring
data may indicate that the plume of TBA is continuing to
expand  down gradient.  Farther expansion  of the TBA
plume may put a down gradient receptor at risk, or exceed
some concentration-based  goal at a down gradient point
of compliance.  Depending  on the remedial goals,  on the
nature of the source area for TBA in ground water, and on
the geochemistry of the ground water, Monitored Natural
Attenuation may have a role in the overall strategy for risk
management.  Small and Weaver  (1999) reasoned that
plumes of MTBE and benzene expand because the MTBE
or benzene is transferred from the source area to flowing
ground water faster than natural biodegradation can remove
MTBE or benzene from the  flowing ground water. Plumes
of TBA should follow the same pattern.

The capacity for biodegradation  of TBA  is limited  by the
concentration of sulfate in ground water entering the source
area and the amount of biologically available Iron(lll) or
Manganese(IV) associated with the aquifer sediments. The
only remedial approach that can tip balance between the
release of TBA to ground water and the degradation of TBA
in ground water is to reduce the transfer of TBA from the
source to flowing ground water in the aquifer.

There are two common situations that lead to a continuing
source of TBA contamination. In the first situation, MTBE
in  residual gasoline is slowly released to flowing ground
water over time, and the MTBE is biologically degraded to
TBA once it comes into solution in ground water.  In the
second situation, TBA that is present in residual gasoline,
as an oxygenate, is slowly released to flowing ground water
over time because the residual gasoline  is held in clay or
silt by capillary attraction. As a result the TBA must escape
the clay or silt by diffusion through pore water before it can
enter the major channel of ground water flow in the aquifer.
In either situation, the first  remedial response should  be
removal of the residual gasoline by excavation, or some
effective technique for in situ remediation. At some sites,
air sparging has proved effective to remove sources of
MTBE in residual gasoline (Hattan et al.,  2003). Because
air sparging can effectively supply oxygen to ground water,
air sparging should effectively remove TBA as well.
                                                   27

-------
After the source has been remediated, it may be neces-
sary to pump and treat the ground water that still contains
high concentrations of TEA.  It is not reasonable to expect
natural biodegradation of TEA to remediate TEA in a plume
unless the concentration of sulfate in the plume can meet
the stoichiometric demand for complete metabolism of TEA
and other organic compounds  in the  plume.  If a pump
and treat remedy is put in place, it  should continue until
the concentrations of sulfate in the plume are adequate to
degrade TEA and all the other organic compounds in the
ground water.

Research needs

More  research and field studies are needed on the con-
tribution of Iron(lll) reducing and Manganese(IV) reducing
bacteria to the natural anaerobic biodegradation of TEA at
gasoline spill sites. In particular, techniques are needed
to evaluate the supply of biologically available Iron(lll) and
Manganese(IV).

More  research is also needed on techniques to determine
the presence and activity  of naturally occurring microbes
in ground water that  can degrade TEA.  Recently, new
techniques have been developed to sample and evaluate
the microorganisms in water in  monitoring wells (Big-
gerstaff, 2007; Geyer et al., 2005; Sublette  et al.,  2006).
These techniques are built around the use of  Bio-Sep®
beads (Microbial Insights, Rockford, TN).  These  beads
are constructed from a composite of 25% aramid polymer
(Nomex®, DuPont, Wilmington, DE) and 75% powdered
activated carbon. The beads are from  2 to 4 mm in diam-
eter. They have a high porosity (74%) and high specific
surface area (600 m2 /g). The Bio-Sep® beads provide a
surface for the microorganisms to colonize and grow. After
a period of incubation, in ground water in a monitoring well,
the beads are retrieved and the microorganisms that grew
in the beads are extracted and analyzed.

One particularly compelling approach is to amend Bio-Sep®
beads with an organic compound that is mass labeled with
the stable carbon isotope 13C.  If the compound is biologi-
cally degraded, some portion of the mass label should find
its way into the biomass that develops in the bead.  Geyer
et al., (2005) amended the beads with 13C labeled benzene
or toluene by sorbing vapors of benzene or toluene to the
powdered activated carbon. The beads were installed in a
monitoring well at a contaminated site for 32 days and then
recovered. The phospholipid fatty acids in the biomass were
extracted and the concentration of 13C in the fatty acids was
determined using compound specific isotope ratio mass
spectrometry. Selected fatty acids were highly enriched in
13C and the mass label could only have come from metabo-
lism of the mass labeled benzene or toluene incorporated
into the Bio-Sep® beads before they were deployed to the
well. As of this writing,  EPA funded research is applying
the same approach to evaluate the biodegradation of TBA
in contaminated ground water.

Finally, more effort is needed to monitor the lifecycle of
TBA plumes in ground water at gasoline spill sites and
to document the contribution of natural attenuation pro-
cesses,  including natural biodegradation and dilution and
dispersion.
                                                    28

-------
                                             References
American  Society for Testing  and Materials Standards.
    ASTM E-1739-95; Standard  Guide for  Risk-Based
    Corrective Action (RBCA)  Applied at Petroleum
    Release Sites (2002). Available at http://webstore.ansi.
    org/ansidocstore/default.asp.
Biggerstaff, J.R, M. Le Puil, B.L. Weidow, J. Leblanc-Gridley,
    E.  Jennings,  J. Busch-Harris,  K.L.  Sublette,  D.C.
    White and R.S. Alberte. A novel and in situ technique
    for the quantitative detection of MTBE and benzene
    degrading bacteria in contaminated matrices. Journal
    of Microbiological Methods 68(2): 437-441 (2007).
Bradley,  S.N.,  T.B.  Hammill,  and  R.L. Crawford.
    "Biodegradation of agricultural chemicals." In Manual
    of Environmental Microbiology, by  C. J. Hurst, G. R.
    Knudsen, M.  J. Mclnerney,  L. D. Stetzenbach, and M.
    V. Walter: American Society  for Microbiology (ASM)
    Press, Washington, DC, 815-821, (1997).
Bradley, P.M., F.H. Chapelle, and J.E. Landmeyer. Effect of
    redox conditions on MTBE biodegradation in surface
    water sediments. Environmental Science & Technology
    35 (23): 4643-4647 (2001).
Bradley, P.M., J.E. Landmeyer, and F.H. Chapelle. TBA
    biodegradation in surface-water  sediments under
    aerobic  and  anaerobic conditions.  Environmental
    Science & Technology, 36 (19): 4087-4090 (2002).
Buscheck, T.E., and  C.M.  Alcantar.   "Regression
    techniques and analytical  solutions to  demonstrate
    intrinsic bioremediation." In Intrinsic Bioremediation,
    Proceedings  of the Third  International In  Situ and
    On-Site Bioremediation Symposium, San Diego, CA,
    109-116, (1995).
Day, M., R. Aravena, D. Hunkeler, andT Gulliver. Application
    of  carbon isotopes to document  biodegradation of
    tert-butyl alcohol under field conditions. Contaminated
    Soil Sediment and Water.  Oxygenated  Fuels Issue,
    July/August 2002, 88-92.
Day, M.J.  and T Gulliver.  "Natural  attenuation of  tert
    butyl alcohol  at a Texas chemical plant."  In MTBE
    Remediation  Handbook,  by E.E. Moyer  and  PT.
    Kostecki, Amherst, Massachusetts: Amherst Scientific
    Publishers, 2003, 541-560.
DeVaull, G.E.,  RT Sun,  I.A.L. Rhodes,  and D.F. Walsh.
    Study of tert-butyl  alcohol (TBA) at  selected
    underground  storage tank remediation project sites in
    Orange County, California,  Final Report, prepared by
    Shell Global  Solutions US, Inc., for Seth Daugherty,
    Orange County Health Care Agency, November 13,
    2003, 25 pages.  Available by sending a  request to
    George DeVaull, Shell Global Solutions US, Inc., HSC
    Consultancy, 3333 Highway 6 South, Houston, Texas
    77082 or by sending e-mail to Wilson.johnt@epa.gov.
Environmental Security Technology Certification Program
    (ESTCP). Cost and Performance Report Bioavailable
    Ferric Iron  (BAFelll) Assay, ESTCP Project  Number
    CU-0009 (June 2005). http://www.estcp.org/viewfile.
    cfm?Doc=ER%2D0009%2DC%26P%2DFerriclron%
    2Epdf. Accessed July 16, 2007.
Finneran, K.T. and D.R. Lovley. Anaerobic degradation of
    methyl tert-butyl ether (MTBE) and tert-butyl alcohol
    (TBA). Environmental  Science & Technology, 35 (9):
    1785-1790(2001).
Finneran,  K.T.  and  D.R. Lovley.  "Anaerobic  In Situ
    Bioremediation." In MTBE Remediation Handbook, by
    E.E. Moyer and PT Kostecki, Amherst, Massachusetts:
    Amherst Scientific Publishers, 2003, 265-277.
Fukui, M., and T.S. Takii. Kinetics of sulfate respiration by
    free-living and particle-associated  sulfate-reducing
    bacteria. FEMS Microbiology Ecology 13 (4):  241-247
    (1994).
Geyer,  R., A.D.  Peacock,  A. Miltner, H.H. Richnow, D.C.
    White, K.L. Sublette, and M. Kastner. In Situ Assessment
    of Biodegradation Potential Using Biotraps Amended
    with ISC-labeled Benzene  or Toluene. Environmental
    Science and Technology 39(13):4983-4989 (2005).
Gray, J.R., G. Lacrampe-Couloume, D. Gandhi, K.M. Scow,
    R.D.Wilson, D.M. Mackay, and B.S. Lollar. Carbon and
    hydrogen isotopic fractionation during biodegradation
    of methyl tert-butyl ether.  Environmental  Science &
    Technology  36 (9): 1931-1938 (2002).
Hattan,  G.,  B. Wilson, and J.T Wilson. Performance of
    Conventional  Remedial Technology for  Treatment
    of MTBE and  Benzene  at  UST Sites  in  Kansas.
    Remediation: The Journal of Environmental  Cleanup
    Costs,  Technologies  &  Techniques 14(1):85-94.
    (2003).
Hunkeler, D.,  B.J. Butler, R. Aravena, and J.F. Barker.
    Monitoring  biodegradation  of methyl tert-butyl ether
    (MTBE)  using compound-specific  carbon  isotope
    analysis. Environmental Science & Technology 35 (4):
    676-681 (2001).
Ingvorsen, K. and B.B. J0rgensen. Kinetics of sulfate uptake
    by freshwater and  marine species of Desulfovibrio.
    Archives of Microbiology 139 (1): 61-66 (1984).
                                                   29

-------
Interstate Technology & Regulatory Council. Overview of
    Groundwater  Remediation Technologies  for  MTBE
    and TEA, by The Interstate Technology &  Regulatory
    Council, MTBE  and Other Fuel  Oxygenates  Team,
    2005.  available http://www.itrcweb.org/Documents/
    MTBE-1 .pdf
Kane, S.R., H.R.  Beller, T.C.  Legler,  C.J. Koester, H.C.
    Pinkart, R.U. Halden,  and A.M. Happel. Aerobic
    biodegradation of  methyl  tert-butyl ether by aquifer
    bacteria from leaking underground storage tank sites.
    Applied and Environmental  Microbiology.  67 (12):
    5824-5829(2001).
Kolhatkar,  R., J.T Wilson,  and I.E. Dunlap.  "Evaluating
    Natural Biodegradation of MTBE at multiple UST sites."
    In Petroleum and Organic Chemicals in Ground Water:
    Prevention,  Detection and Remediation  Conference
    & Exposition Anaheim,  California November  15-17,
    2000, 32-49.
Kolhatkar,  R., J.T. Wilson, and G.  Hinshalwood. "Natural
    biodegradation of  MTBE  at a site on Long Island,
    NY." In  Proceedings of the Sixth International In Situ
    and On-Site Bioremediation Symposium,  San Diego,
    California, June  4-7, 2001, Bioremediation of MTBE,
    Alcohols, and Ethers  6(1):43-50. Battelle Press,
    Columbus Ohio (2001).
Kolhatkar,  R., T Kuder, P. Philp, J.  Allen, and  J.T. Wilson.
    Use of compound-specific  stable  carbon isotope
    analyses  to demonstrate  anaerobic  biodegradation
    of MTBE  in groundwater at  a  gasoline release site.
    Environmental Science & Technology 36  (23): 5139-
    5146(2002).
Kuder, T, J.T. Wilson, P. Kaiser, R. Kolhatkar, P. Philp, and
    J. Allen. Enrichment of stable  carbon and hydrogen
    isotopes during  anaerobic biodegradation of MTBE:
    microcosm and field evidence. Environmental Science
    & Technology 39 (1): 213-220 (2005).
Lovley, D.R., F.H.  Chapelle  and J.C. Woodward. Use of
    dissolved H2 concentrations to determine  distribution
    of  microbially catalyzed redox reactions in anoxic
    groundwater. Environmental Science & Technology 28
    (7): 1205-1210(1994).
Meckenstock, R., B. Morasch, C. Griebler, and H.H. Richnow.
    Stable isotope f ractionation  analysis as a tool to monitor
    biodegradation in  contaminated aquifers.  Journal of
    Contaminant Hydrology 75: 215-255 (2004).
Newell, C.J., R.K. McLeod, and J.R. Gonzales. BIOSCREEN
    natural attenuation decision support system, EPA/600/
    R-96/087,  Cincinnati,  OH:   U.S.  Environmental
    Protection Agency, 1996.
O'Sullivan,  G., G. Boshoff,  A. Downey, and  R.M. Kalin.
    Carbon isotope effect  during  the abiotic oxidation
    of  methyl-tert-butyl  ether (MTBE). In Proceedings
    of  the  Seventh International  In  Situ and On-Site
    Bioremediation  Symposium,  Orlando, FL: Battelle
    Press, Columbus, OH, 2003, 1-10.
Parsons Engineering Science. Streamlined Risk-Based
    Closure of Petroleum Contaminated Sites, Performance
    and Cost Results from Multiple Air Force Demonstration
    Sites, Technology Demonstration Summary Report, Air
    Force Center for Environmental Excellence, Technology
    Transfer Division, Brooks  AFB, San Antonio,  Texas.
    1999. available http://www.afcee.brooks.af.mil/products/
    techtrans/riskbased.asp.
Schirmer,  M.,  B.J.  Butler, C.D.  Church,  J.F. Barker,
    and N. Nadarajah. Laboratory evidence  of  MTBE
    biodegradation in Borden aquifer material. Journal of
    Contaminant Hydrology 60 (2003): 229-249 (2003).
Schmidt, T, M. Schirmer, H. Weiss, and  S.B. Haderlein.
    Microbial degradation of methyl tert-butyl ether and tert-
    butyl alcohol in the subsurface. Journal of Contaminant
    Hydrology 70 (2004): 173-203 (2004a).
Schmidt, T.C., L.  Zwank, M.  Eisner, M. Berg, R.U.
    Meckenstock, and S.B. Haderlein. Compound-specific
    stable isotope analysis of organic contaminants in
    natural environments: a critical review of the state of
    the art, prospects, and future challenges. Analytical and
    Bioanalytical Chemistry 378:283-300 (2004b).
Shih, T, Y.  Rong, T. Harmon,  and  M. Suffet. Evaluation
    of the impact of fuel  hydrocarbons and oxygenates
    on  groundwater resources. Environmental Science &
    Technology 38 (1): 42-48 (2004).
Small, M.C. and J. Weaver.  An updated conceptual model
    for subsurface fate and transport of MTBE and benzene.
    In:  Proceedings of the Petroleum  Hydrocarbons and
    Organic Chemicals  in  Ground Water: Prevention,
    Detection, and Remediation Conference. API, NGWA,
    STEP Conference and Exposition. November  17-19,
    Houston, Texas, pages 209-220 (1999).
Smallwood, B.J., R.P Philp, T.W. Burgoyne, and J.D. Allen.
    The use of stable isotopes to differentiate specific
    source markers for MTBE.  Environmental Forensics
    2 (3): 215-221 (2001).
Somsamak, P., H.H. Richnow and M.H. Haggblom. Carbon
    isotope fractionation  during anaerobic degradation
    of methyl tert-butyl ether under sulfate-reducing and
    methanogenic conditions. Applied and Environmental
    Microbiology 72 (2): 1157-1163 (2006).
Suarez, M.P and H.S. Rifai. Biodegradation rates for fuel
    hydrocarbons and chlorinated solvents in groundwater.
    Bioremediation Journal 3 (4): 337-362 (1999).
Sublette,  K., A. Peacock, D. White, G. Davis,  D.  Ogles,
    D. Cook, R. Kolhatkar,  D. Beckmann,  and X. Yang.
    Monitoring  subsurface microbial ecology in a sulfate-
    amended,  gasoline-contaminated aquifer. Ground
    Water Monitoring & Remediation 26(2):70-78 (2006).
Suflita,  J.M., K.L. Londry, and G.A. Ulrich. "Determination
    of anaerobic biodegradation  activity." In Manual of
    Environmental  Microbiology, by   C.J.  Hurst, G.R.
    Knudsen,  M.J. Mclnerney,  L.D.  Stetzenbach,  and
    M.V. Walter, Washington,  DC: American Society for
    Microbiology (ASM) Press, 1997, 790-801.
                                                    30

-------
U. S. Environmental Protection Agency. Use of monitored
    natural  attenuation  at  superfund, RCRA corrective
    action, and underground storage tank sites, EPA 5407
    R-99/009. EPA Office of Solid Waste and Emergency
    Response Directive, 9200.4-17P, 1999.
Walter, M.V."Bioaugmentation." In Manual of Environmental
    Microbiology, by   C.J.  Hurst,  G.R. Knudsen, M.J.
    Mclnerney, L.D. Stetzenbach,  and  M.V.  Walter,
    Washington, DC: American Society for Microbiology
    (ASM) Press, 1997, 753-757.
White,  K.D., J.T. Novak, C.D. Goldsmith, and S. Bevan.
    "Microbial degradation kinetics of alcohols in subsurface
    systems." In:  NWWA/API Conference  on Petroleum
    Hydrocarbons and Organic Chemicals  in  Ground
    Water-- Prevention,  Detection,  and Restoration,
    November 13-15,1985, The Westin Galleria, Houston,
    TX,140-159, National Water Well Association, Dublin,
    Ohio, 1986.
Wiedemeier, T.H.,  R.N. Miller,  J.T. Wilson, and D.H.
    Kampbell. "Significance of anaerobic processes for
    the intrinsic bioremediation of fuel hydrocarbons." In
    Petroleum  Hydrocarbons and Organic Chemicals in
    Ground Water: Prevention, Detection, and Remediation
    Conference, November 29-December 1,1995, Houston,
    TX, 49-61, National Ground Water Association, Dublin,
    Ohio, 1995.
Wiedemeier, T.H., H.S. Rifai, C.J. Newell, and J.T. Wilson.
    "Intrinsic Bioremediation of Petroleum Hydrocarbons." In
    Natural Attenuation of Fuels and Chlorinated Solvents
    in the Subsurface,  New York:  John Wiley &  Sons,
    1999a., 189-240.
Wiedemeier, T, J.T. Wilson, D.H. Kampbell,  R.N.  Miller,
    and J.E. Hansen. Technical Protocol for Implementing
    Intrinsic Remediation with Long-Term Monitoring for
    Natural Attenuation  of Fuel Contamination Dissolved
    in  Ground Water Volume  I, Air  Force  Center for
    Environmental  Excellence,  Technology Transfer
    Division, Brooks AFB, San Antonio,  Texas.  (1999b)
    available   http://www.afcee.brooks.af.mil/products/
    techtrans/download/A324248.pdf.
Wiedemeier, T, J.T. Wilson, D.H. Kampbell,  R.N.  Miller,
    and J.E. Hansen. Technical Protocol for Implementing
    Intrinsic Remediation with Long-Term Monitoring for
    Natural Attenuation  of Fuel Contamination Dissolved
    in  Ground Water Volume II, Air  Force Center for
    Environmental  Excellence,  Technology Transfer
    Division, Brooks  AFB, San  Antonio, Texas. 1999c.
    available http://www.afcee.brooks.af.mil/products/
    techtrans/download/A324247a.pdf. and  http://www.
    afcee.brooks.af.mil/products/techtrans/download/
    A324247b.pdf.
Wilson, J.T. "Fate and transport of MTBE and other gasoline
    components." In MTBE Remediation Handbook, by
    E.E. Moyerand PT. Kostecki, Amherst, Massachusetts:
    Amherst Scientific Publishers, 2003, 19-61.
Wilson,  J.T, R. Kolhatkar, T.  Kuder,  P. Philp, and S.J.
    Daugherty. Stable isotope analysis of MTBE to evaluate
    the  source of TBA in  ground  water. Ground Water
    Monitoring and Remediation 25 (4): 108-116 (2005a).
Wilson, J.T, P.M. Kaiser, and C. Adair. Monitored Natural
    Attenuation of MTBE as a Risk Management Option
    at Leaking Underground Storage Tank Sites, EPA-600/
    R-04/179. Cincinnati, OH: U.S. Environmental Protection
    Agency, 2005b
Wilson,  J.T,  C.  Adair, P.M.  Kaiser, and  R. Kolhatkar.
    Anaerobic biodegradation of MTBE at a gasoline spill
    site. Ground Water Monitoring and Remediation 25 (3):
    103-115 (2005c).
Wilson,  R.D., D.M. Mackay, and K.M. Scow. In situ MTBE
    degradation supported by diffusive  oxygen release.
    Environmental Science & Technology 36 (2): 190-199
    (2002).
Zenker, M.J., R.C. Borden, and M.A. Barlaz. "Investigation
    of the intrinsic biodegradation of alkyl and cyclic ethers."
    In In Situ Bioremediation of Petroleum Hydrocarbon
    and Other Organic Compounds. The Fifth International
    In Situ and On-Site Bioremediation Symposium, San
    Diego, CA, 1999: Battelle Press, Columbus, OH, 1999,
    Volume (1): 165-170.
Zwank,  L.,  M. Berg,  M. Eisner,  T.C. Schmidt,  R.P
    Schwarzenbach, and S.B. Haderlein. New evaluation
    scheme for  two-dimensional  isotope analysis  to
    decipher  biodegradation processes: application  to
    groundwater contamination by MTBE. Environmental
    Science & Technology 39 (4): 1018-1029 (2005).
                                                   31

-------
                                            Appendix
Table 4 part 1. Typical Quality Performance Data for Analysis of TEA in Water. All Values are ug/L Unless Otherwise
              Indicated.
Date Collected
Date Analyzed

CCC Standard Nominal
CCC Standard Measured
Percent of Check Standard

CCC Standard Nominal
CCC Standard Measured
Percent of Check Standard

CCC Standard Nominal
CCC Standard Measured
Percent of Check Standard

QC Standard Nominal
QC Standard Measured
Percent of Check Standard

QC Standard Nominal
QC Standard Measured
Percent of Check Standard

Blank 1
Blank 2

Sample Analysis
Laboratory Duplicate
Relative Percent Difference

Spike Concentration
Sample Concentration
Spike Recovery (Percent)
02/01/01
03/22/01 &
03/26/01

200
194
97.0%

200
177
89.0%

20.0
19.0
95.0%

20.0
17.4
87.0%

200
194
97.0%

<10*


31.2
33.8
8.00%

NP
112
83.0%
04/06/01
06/07-14/01

200
188
94.0%

20.0
18.0
90.0%

200
177
88.5%

20.0
16.6
83.0%

200
190
95.0%

<10*
<10*

34.2
33.4
2.37%

NP
<1.00
89.0%
05/31/01
09/27/01 ,
10/18-23/01

200
217
109%

20.0
22.4
112%





20.0
18.9
94.5%

200
189
94.5%

<10*
<10*








07/06/01
08/28-31/01
& 09/1 9/01

200
198
99.0%

20.0
20.5
103%

200
204
102%

20.0
21.1
106%

200
209
105%

<10*
<10*

20.8
19.8
4.93%

NP
17.1
100%
09/05/01
11/02-06/01

20.0
19.8
99.0%

200
171
85.5%

200
182
91 .0%

20.0
19.0
95.0%

200
210
105%

<10*
<10*

91.6
118
25.2%

200
31.4
99.0%
 CCC: Continuing Calibration Check;  QC: Second Source Check; NP: Not Provided
 * Lower Calibration Limit
                                                 32

-------
Table 4 part 2. Typical Quality Performance Data for Analysis of TEA in Water. All Values are ug /L Unless Otherwise
               Indicated.
Date Collected
Date Analyzed

CCC Standard Nozminal
CCC Standard Measured
Percent of Check Standard

CCC Standard Nominal
CCC Standard Measured
Percent of Check Standard

CCC Standard Nominal
CCC Standard Measured
Percent of Check Standard

QC Standard Nominal
QC Standard Measured
Percent of Check Standard

QC Standard Nominal
QC Standard Measured
Percent of Check Standard

Blank 1
Blank 2

Sample Analysis
Laboratory Duplicate
Relative Percent Difference

Spike Concentration
Sample Concentration
Spike Recovery (Percent)
10/11/01
11/28-30/01

20.0
19.7
98.5%

20.0
19.5
97.5%

200
194
97.0%

200
219
110%





<10*
<10*

<20.0
<20.0
0.00%

NP
<20.0
97.0%
10/25/01
11/30/01-
12/04/01

200
226
113%

20.0
24.3
122%





200
237
119%





<10*
<10*

55.2
53.6
2.94%

NP
42.4
126%
11/16/01
11/26
- 27/01

200
241
121%









200
252
126%





<20*


113
110
3.05%

NP
71.8
121%
12/05/01
01/04-09/02

200
219
110%

20.0
21.5
108%

20.0
20.9
105%

20.0
20.3
102%

200
198
99.0%

<10*
<10*








01/04/02
01/18/02

200
213
107%

20.0
24.0
120%

200
201
101%

20.0
22.2
111%





<10*
<10*

57.9
67.3
15.0%

NP
53.9
104%
 CCC: Continuing Calibration Check; QC: Second Source Check; NP: Not Provided
Table 4 part 3. Typical Quality Performance Data for Analysis of TBA in Water. All Values are ug /L Unless Otherwise
               Indicated.
Date Collected
Date Analyzed

CCC Standard Nominal
CCC Standard Measured
Percent of Check Standard

CCC Standard Nominal
1/15/02
1/26/02

100
88.7
88.7%

200
2/12/02
3/6/02

20.0
19.3
96.5%

100
3/12/02
4/8/02

200
189
94.5%

200
4/16/02
5/7/02

200
213
107%

20.0
4/25/02
4/25-26/02

200
223
112%

20.0
                                                   33

-------
Date Collected
CCC Standard Measured
Percent of Check Standard

CCC Standard Nominal
CCC Standard Measured
Percent of Check Standard

QC Standard Nominal
QC Standard Measured
Percent of Check Standard

QC Standard Nominal
QC Standard Measured
Percent of Check Standard

Blank 1
Blank 2

Sample Analysis
Laboratory Duplicate
Relative Percent Difference

Spike Concentration
Sample Concentration
Spike Recovery (Percent)
1/15/02
160
80.0%

20.0
17.9
89.5%

200
162
81 .0%

100
84.9
84.9%

<10*
<10*

19.8
20.3
2.49%

92.6
4.20
88.0%
2/12/02
93.7
93.7%

100
93.3
93.3%

200
208
104%

20.0
19.1
95.5%

<2.4**
<2.4**

<2.40
<2.40
0.00%

98.8
<2.40
98.8%
3/12/02
200
100%

100
116
116%

20.0
17.5
87.5%





<2.4**
<2.4**

<2.40
<2.40
0.00%

94.9
<2.40
94.9%
4/16/02
21.9
110%

200
205
103%

100
117
117%

1000
962
96.2%

<2.4**
<2.4**

88.1
90.0
2.13%

114
<2.40
114%
4/25/02
22.5
113%





200
239
120%





<2.4**
<2.4**

118
105
1 1 .7%

290
276
109%
 CCC: Continuing Calibration Check;  QC: Second Source Check
Table 4 part 4. Typical Quality Performance Data for Analysis of TBA in Water. All Values are ug /L Unless Otherwise
              Indicated.
Date Collected
Date Analyzed

CCC Standard Nominal
CCC Standard Measured
Percent of Check Standard

CCC Standard Nominal
CCC Standard Measured
Percent of Check Standard

CCC Standard Nominal
CCC Standard Measured
Percent of Check Standard

QC Standard Nominal
QC Standard Measured
Percent of Check Standard

5/29/02
6/6/02 to

200
225
113%









20.0
20.8
104%

7/16/02
8/7/02

200
240
120%

100
116
116%

100
106
106%

200
224
112%

12/04/02
12/12/02













20.0
20.9
105%

12/11/02
12/18/02

100
112
112%

100
101
101%

50.0
54.5
109%

100
92.5
93.0%

5/15/03
5/16/03

200
200
100%

20.0
22.5
1 1 3%





100
98.6
98.6%

                                                   34

-------
Date Collected
QC Standard Nominal
QC Standard Measured
Percent of Check Standard

Blank 1
Blank 2

Sample Analysis
Laboratory Duplicate
Relative Percent Difference

Spike Concentration
Sample Concentration
Spike Recovery (Percent)
5/29/02
200
220
110%

<2.4**
<2.4**

38.8
52.2
29.5%

200
40.4
94.0%
7/16/02
100
102
102%

<2.4**
<2.4**

<2.40
<2.40
0.00%

109
<2.40
109%
12/04/02
200
238
119%

<2.4**
<2.4**

259
283
8.90%

256
251
103%
12/11/02
50.0
49.4
99.0%

<2.4**
<2.4**

39.5
43.5
9.64%




5/15/03
200
200
100%

<2.4**
<2.4**

270
270
0.00%




 CCC: Continuing Calibration Check; QC: Second Source Check
Table 5 part 1.  Typical Quality Performance Data for Analysis of Sulfate in Water. All Values are mg/L Unless
               Otherwise Indicated.
Date Collected
Date Analyzed

CCC Standard Nominal
CCC Standard Measured
Percent of Check Standard

CCC Standard Nominal
CCC Standard Measured
Percent of Check Standard

CCC Standard Nominal
CCC Standard Measured
Percent of Check Standard

QC Standard Nominal
QC Standard Measured
Percent of Check Standard

QC Standard Nominal
QC Standard Measured
Percent of Check Standard

Blank 1
Blank 2

Sample Analysis
Laboratory Duplicate
Relative Percent Difference

Spike Concentration
01/05/01
4/15-26/01

29.8
30.8
103%

1.00
1.08
108%

25.0
25.2
101%

29.8
29.5
99.0%

5.00
5.25
105%

<0.50
<0.50

2.85
2.82
1 .06%

NP
5/31/01
11/8-12/01

25.0
25.5
102%

5.00
4.87
97.4%

25.0
26.4
106%

5.00
5.24
105%

25.0
27.4
110%

<1.00


16.0
15.9
0.63%

NP
5/31/01
10/3-4/01

28.3
28.1
99.3%

5.00
4.93
98.6%

5.00
4.93
98.6%

5.00
4.95
99.0%

28.3
28.1
99.3%

<1.00


2.82
2.84
0.71%

NP
6/28/01
10/4-5/01

28.3
27.1
95.8%

5.00
4.67
93.4%

5.00
4.78
95.6%

5.00
4.92
98.4%

5.00
5.45
109%

<1.00


2.50
2.53
1.19%

NP
07/06/01
9/26-27/01

28.3
28.2
99.6%

1.00
0.95
95.0%

50.0
47.9
95.8%

28.3
29.0
103%

50.0
48.3
96.6%

<1.00
<1.00

2.04
2.01
1 .48%

50.0
                                                   35

-------
Date Collected
Sample Concentration
Spike Recovery (Percent)
01/05/01
3.49
100%
5/31/01
12.8
101%
5/31/01
2.77
95.0%
6/28/01
2.39
96.0%
07/06/01
2.28
98.6%
 CCC: Continuing Calibration Check; QC: Second Source Check;  NP: Not Provided
Table 5 part 2.  Typical Quality Performance Data for Analysis of Sulfate in Water. All Values are mg/L Unless
               Otherwise Indicated.
Date Collected
Date Analyzed

CCC Standard Nominal
CCC Standard Measured
Percent of Check Standard

CCC Standard Nominal
CCC Standard Measured
Percent of Check Standard

CCC Standard Nominal
CCC Standard Measured
Percent of Check Standard

QC Standard Nominal
QC Standard Measured
Percent of Check Standard

QC Standard Nominal
QC Standard Measured
Percent of Check Standard

Blank 1
Blank 2

Sample Analysis
Laboratory Duplicate
Relative Percent Difference

Spike Concentration
Sample Concentration
Spike Recovery (Percent)
7/26/01
10/10/01

28.3
28.2
99.6%

5.00
4.83
96.6%

5.00
5.00
100%

5.00
5.31
106%

5.00
5.27
105%

<1.00


2.70
2.75
1 .83%

NP
2.81
98.0%
9/05/01
12/12-14/01

9.87
9.09
92.1%

25.0
24.9
99.6%

5.00
4.97
99.4%

25.0
24.8
99.2%

5.00
5.11
102%

<1.00


3.25
3.34
2.73%

NP
4.97
102%
10/11/01
10/25/01

28.3
27.6
97.5%

25.0
24.8
99.2%

25.0
24.9
99.6%

5.00
4.85
97.0%

25.0
25.9
104%

<1.00


2.00
1.99
0.50%

NP
1.99
95.0%
10/25/01
11/9-12/01

28.3
28.6
101%

5.00
5.18
104%

25.0
25.8
103%

5.00
5.24
105%

25.0
26.1
104%

<1.00


2.69
2.61
3.02%

NP
2.78
97.0%
12/05/01
1/7/02

9.87
9.32
94.4%

25.0
24.4
97.6%

5.00
4.93
98.6%

25.0
25.1
100%

5.00
5.01
100%

<1.00


4.49
4.49
0.00%

NP
2.27
94.0%
 CCC: Continuing Calibration Check; QC: Second Source Check;  NP: Not Provided
Table 5 part 3.  Typical Quality Performance Data for Analysis of Sulfate in Water. All Values are mg/L Unless
               Otherwise Indicated.
Date Collected
Date Analyzed

CCC Standard Nominal
CCC Standard Measured
01/04/02
02/06/02

9.87
9.18
1/15/02
2/4-5/02

9.87
8.91
02/07/02
03/01/02

9.87
8.95
2/12/02
3/2/02

9.87
8.32
3/12/02
4/11/02

9.87
9.24
                                                    36

-------
Date Collected
Percent of Check Standard

CCC Standard Nominal
CCC Standard Measured
Percent of Check Standard

CCC Standard Nominal
CCC Standard Measured
Percent of Check Standard

QC Standard Nominal
QC Standard Measured
Percent of Check Standard

QC Standard Nominal
QC Standard Measured
Percent of Check Standard

Blank 1
Blank 2

Sample Analysis
Laboratory Duplicate
Relative Percent Difference

Spike Concentration
Sample Concentration
Spike Recovery (Percent)
01/04/02
93.0%

25.0
24.5
98.0%

5.00
4.84
96.8%

25.0
25.2
101%

5.00
5.05
101%

<1.00


2.65
2.59
2.29%

3.58
2.54
92.0%
1/15/02
90.3%

25.0
25.1
100%

5.00
4.83
96.6%

25.0
25.4
102%

5.00
5.16
103%

<1.00


12.4
12.3
0.81%

3.67
2.74
92.0%
02/07/02
90.7%









5.00
5.14
103%

25.0
25.8
103%

<1.00


6.60
6.64
0.60%

13.4
1.16
103%
2/12/02
84.3%

25.0
25.1
100%

5.00
5.00
100%

25.0
25.4
102%

5.00
5.25
105%

<1.00


10.6
10.7
0.94%

16.8
9.16
98.0%
3/12/02
93.6%

25.0
25.3
101%

5.00
5.04
101%

25.0
26.7
107%

5.00
5.20
104%

<1.00


8.16
8.15
1 .23%

8.49
12.1
98.0%
 CCC: Continuing Calibration Check; QC: Second Source Check
Table 5 part 4. Typical Quality Performance Data for Analysis of Sulfate in Water. All Values are mg/L Unless
               Otherwise Indicated.
Date Collected
Date Analyzed

CCC Standard Nominal
CCC Standard Measured
Percent of Check Standard

CCC Standard Nominal
CCC Standard Measured
Percent of Check Standard

CCC Standard Nominal
CCC Standard Measured
Percent of Check Standard

QC Standard Nominal
QC Standard Measured
4/16/02
5/14-15/02

9.87
9.04
91 .6%

25.0
25.0
100%

5.00
4.97
99.4%

25.0
25.7
4/25/02
5/23/02

18.6
17.0
91 .4%

25.0
24.6
98.4%

5.00
4.91
98.2%

25.0
25.8
5/29/02
6/25/02

18.6
17.6
94.6%

25.0
24.7
98.8%

5.00
4.85
97.0%

25.0
24.5
7/16/02
8/15/02

18.6
17.9
96.2%

5.00
5.09
102%

25.0
25.7
103%

5.00
5.04
                                                   37

-------
Date Collected
Percent of Check Standard

QC Standard Nominal
QC Standard Measured
Percent of Check Standard

Blank 1
Blank 2

Sample Analysis
Laboratory Duplicate
Relative Percent Difference

Spike Concentration
Sample Concentration
Spike Recovery (Percent)
4/16/02
103%

5.00
5.05
101%

<1.00


10.7
10.7
0.00%

18.1
10.5
103%
4/25/02
103%

5.00
4.91
98.2%

<1.00


0.53
0.51
3.85%

11.5
0.53
90.0%
5/29/02
98.0%

5.00
4.91
98.2%

<0.33


9.78
9.78
0.00%

11.8
<0.331
94.0%
7/16/02
101%

25.0
25.5
102%

<1.00


13.4
13.5
0.74%

18.7
12.1
101%
38

-------
39

-------

-------
SEPA
      United
      Environmental Protection
      Agency
Please make all necessary changes on the below label,
detach or copy, and return to the address in the upper
left-hand corner.

If you do not wish to receive these reports CHECK HERED;
detach, or copy this cover, and return to the address in the
upper left-hand corner.
PRESORTED STANDARD
 POSTAGE & FEES PAID
         EPA
   PERMIT No. G-35
      National Risk Management
        Research Laboratory
      Cincinnati, OH 45268

      Official Business
      Penalty for Private Use
      $300


      EPA/600/R-07/100
      October 2007
                                                                          Recycled/Recyclable
                                                                          Printed with vegetable-based ink on
                                                                          paper that contains a minimum of
                                                                          50% post-consumer fiber content
                                                                          processed chlorine free

-------