&ERA
United States
Environmental Protection
Agency
Engineering  Issue
                       In-Situ Chemical Oxidation
                                   Scott G. Huling1 and Bruce E. Pivetz2
                  Index
  I.  PURPOSE
  II.  INTRODUCTION
     II.A.  Background
     II.B.  Definition
     II.C.  Process Fundamentals
         II.C.1. In-Situ Permanganate Oxidation
         II.C.2. In-Situ Fenton Oxidation
         II.C.3. In-Situ Ozone Oxidation
         II.C.4. In-Situ Persulfate Oxidation

  III. TECHNOLOGY DESCRIPTION AND TECHNOLOGY
     SELECTION FACTORS
     III.A. Bench-Scale Studies
         111.A.I. Objectives
         III.A.2. General Guidelines
         III.A.3. Oxidant Demand
     III.B. Pilot-Scale Studies
         111.B.I. Objectives
         III.B.2. General Guidelines
     III.C. Technology Applicability
         III.C.1. Location of Oxidant Application
         III.C.2. Contaminant Characteristics
         III.C.3. Subsurface Characteristics
     III.D. Site Requirements and Operational Issues
         III.D.1. Site Characterization Data
         III.D.2. Required Site Infrastructure
         III.D.3. Regulatory Constraints on Injection
             of Reagents
     III.E. Field-Scale Implementation and
         Engineering Design Considerations
         III.E.1. Treatment Objectives
         III.E.2. General Conceptual Approach to ISCO
         III.E.3. Oxidant Delivery
         III.E.4. Monitoring
         III.E.5. Summary of Contaminant Transport
             and Fate Mechanisms during ISCO
         III.E.6. Safety Issues
         III.E.7. Treatment Trains
     III.F. Limitations/Interferences/Impacts
         III.F.1. Untreated COCs/Rebound
         III.F.2. Toxic Reaction Byproducts
         III.F.3. Process Residuals
         III.F.4. Geochemical Impacts
         III.F.5. Impact of ISCO on Natural
             Attenuation and Biodegradation
     III.G. Summary

  IV. ACKNOWLEDGMENTS

  V. NOTICE/DISCLAIMER

  VI. QUALITY ASSURANCE STATEMENT

  VII. ACRONYMS, ABBREVIATIONS, AND SYMBOLS

  VIM. REFERENCES


1  U.S. EPA/ORD/NRMRL/GWERD, P.O. Box 1198,
  Ada, OK 74820
2  Dynamac Corporation, 3601 Oakridge Blvd, Ada, OK 74820
                          I. PURPOSE

                          The U.S. Environmental Protection Agency (EPA) Engineering Issue
                          Papers are a series of technology transfer documents that summarize
                          the latest available information on specific technical issues,  including
                          fate and transport, specific contaminants, selected treatment and site
                          remediation technologies, and related other issues. This Engineering
                          Issue Paper is intended to provide remedial project managers (RPMs),
                          on-scene coordinators (OSCs), contractors, and other state, industry,
                          or private remediation  managers with information to assist  in the
                          evaluation and possible selection of appropriate in-situ chemical oxi-
                          dation (ISCO) remedial alternatives.

                          This Engineering Issue Paper provides an  up-to-date overview of
                          ISCO  remediation technology and fundamentals, and is developed
                          based on peer-reviewed  literature, EPA reports, web sources, current
                          research, conference proceedings, and other pertinent information.

                          II. INTRODUCTION

                          II.A. Background

                          In-situ chemical  oxidation involves the introduction of a chemical
                          oxidant into the subsurface for the purpose of transforming ground-
                          water or soil contaminants into less harmful chemical species.  There
                          are several different forms of oxidants that have been used for ISCO;
                          however, the focus of this Engineering Issue Paper will be on the four
                          most  commonly  used oxidants: permanganate (MnO4~),  hydrogen
                          peroxide (H2O2) and iron (Fe) (Fenton-driven, or H2O2-derived oxi-
                          dation), persulfate (S2O82"), and ozone (O3) (Table 1). The type and
                          physical form  of the oxidant indicates the general materials handling
                          and injection  requirements.  The persistence of the oxidant  in the
                          subsurface is important  since this affects the contact time for  advec-
                          tive and diffusive transport and ultimately the delivery of oxidant to
                          targeted zones in the subsurface.  For example, permanganate persists
                          for long periods of time, and diffusion into low-permeability materi-
                          als and greater transport distances through porous media are possible.
                          H2O2  has been reported to persist in soil  and aquifer material for
                          minutes to hours, and the diffusive and advective transport distances
                          will be relatively limited.  Radical intermediates  formed using some
                          oxidants (H2O2,  S2O82", O3) that are largely responsible for various
                          contaminant transformations react very quickly  and  persist for very
                          short periods of time (<1 sec).

-------
Table 1. Oxidant Form, Stability, Stage of Development and Oxidation Potential for Oxidants
Oxidation
Oxidant
Permanganate
Fenton's
Ozone
Persulfate

Reactive Species Fo
Mn04- powder/I
•OH,-02vH02,H02- liquid
03,-OH gas
•S042' powder/I

Oxidant and Reactions
Permanganate
Mn04- + 4H+ + 3e- -

	 >. Mn02 + 2H20
rm Persistence'1)
quid >3 months
minutes- hours
minutes- hours
quid hours -weeks

Electrode Potential (Eh)<2>

1 .7 V (permanganate ion)
Used for In-Situ Chemical
Stage of Development
developing
experimental/emerging
experimental/emerging
experimental/emerging



(1)
Fenton's (H2O2 Derived Reactants)
H202 + 2H+ + 2e- —
2-OH + 2H+ + 2e- -
•H02 + 2H+ + 2e- —
•02-+4H+ + 3e- 	
H02-+H20 + 2e- —
Ozone
03 + 2H+ + 2e- 	
203 + 3H202 	 >
Persulfate
S2082- + 2e- 	 >-
	 »• 2H20
	 >• 2H20
	 * 2H20
— »• 2H20
— >. 30H-

-x 02 + H20
- 402 + 2-OH + 2H20

2 S042-
•S04- + e- 	 >- S042-
1 .8 V (hydrogen peroxide)
2.8V(hydroxyl radical)
1 .7 V (perhydroxyl radical)
-2.4V(superoxide radical)
-0.88 V (hydro peroxide anion)

2.1 V (ozone)
(2)
(3)
(4)
(5)
(6)

(7)
2.8 V(hydroxyl radicalsee rxn 3) (8)

2.1 V(persulfate)
2.6 Vfsulfate radical)

(9)
(10)
1 Persistence of the oxidant varies depending on site-specific conditions. Durations specified here are based on general observations.
2 Reactive species in parentheses; reduction potential is negative.
Permanganate-based ISCO is more fully developed than
other oxidants. Widespread use of in-situ permanganate
oxidation involving a diversity of contaminants and geo-
logical environments under well-documented pilot- and
field-scale  conditions  (in  conjunction  with long-term
monitoring data and cost information)  has contributed
to the development  of the infrastructure needed to sup-
port decisions to design and deploy permanganate ISCO
systems. However, additional research and development
is needed.  Fenton-driven ISCO has been deployed at a
large number of sites and involves a variety of approaches
and  methods involving the use  of hydrogen peroxide
(H2O2) and iron (Fe). In general,  Fenton chemistry and
in-situ Fenton oxidation is complex, involves numerous
reactive intermediates and mechanisms, and technology
development has been slower.  Ozone  (O3) is  a strong
oxidant that has been used in the subsurface but in much
more limited application than permanganate and Fenton-
driven oxidation.  Persulfate (S2O82~) is a  relatively new
form  of oxidant that has mainly been investigated at
bench-scale.  However, considerable research and applied
use of this oxidant at an increasing number of field sites is
resulting in  rapid development.  The electrode (oxida-
tion) potential of the oxidant and reactive species (Table
1) is a measure of the oxidizing strength of the reactive
species, but is not a measure of the reaction rate with dif-
ferent organic compounds.

Site-specific  conditions and parameters,  in conjunction
with oxidant-specific characteristics,  must be  carefully
considered to determine whether ISCO is a viable tech-
nology for deployment relative to other candidate tech-
nologies,  and  to  determine which  oxidant  is  most
appropriate.  These issues and the advantages and disad-
vantages of each oxidant are discussed in detail.

The breadth of ground-water contaminants amenable to
transformation  via various  oxidants  is large.   That is,
many environmental contaminants react at moderately
high rates with these oxidants. Therefore, a wide range
of contaminant classes  are amenable to chemical oxida-
tive treatment (Table 2).  Mixtures of contaminants may
       Engineering Issue
                            Chemical Oxidation

-------
Table 2. Assessment of the Amenability of Various Contaminant and Contaminant Classes to Oxidation Transformations
Contaminant
Petroleum hydrocarbons
BTEX
Benzene
Phenols
PAHs
MTBE
tert-butyl alcohol
Chlorinated ethenes
Carbon tetrachloride
Chloroform
Methylene chloride
Chlorinated ethanes5
Trichloroethane5
Dichloroethane5
Chlorobenzene
PCBs
Energetics (RDX.HMX)
Explosives
Pesticides
1,4-dioxane7
Oxidant
MnO4-
Fenton's
(H202/Fe)
s2cv-(i)
•SCy
(Activated
persulfate)1
Ozone
Ozone/
H202
(Peroxone)(2)
Rating sources
a
G4

P4
G
G
G

E
P


P



P
E



b

E4
G4
E
E


E
P
P


P


P

E
G

c

E4
P4
E
E


E
P
P
P

P
P
P
P

E
G

d

E


E


E



P



P




a
E4

E4
E
E
G

E
P/G


G/E



P
E



b

E

E
G
E
E
E
G
P
G

E
G
E
G

E
P
E
c

E4
E4
E
G


E
P
P
G

P
G
E
P

G
P

d

E

E
E


E



P



P




e

E
E
E
E
E
G
E
P
P
P
P
P
P
E
E


G/E6
E
a
G/E4

G4
P/G
G
P/G

G
P


P



P
G



a
E4

G/E4
G/E
E
E

E
P/G


G/E



P
E



b

E

E
G
E
E
E






E
P

G
G
E
c3

E4
E4
E
G/E


E
P/E
G/E
G/E

P/E
G/E
E
P/E

G/E
G/E

a
E4

E4

E
G

E
P/G


G




E



b

E

E
G
E
E
E
P
P
G

P
G
E
P

E
P

c

G4
G4
E
E










E

E
E

d

E

E
E


E



P



P




a



E











G




b

E

E
G
E
E
E
G
P
G

E
G
E
G

E
P
E
e

E
E
E
E
E
G
E
P
P
P
P
P
P
E
E


G/E6
E
Key: P = poor, G = good, E = excellent. While the different sources used slightly different terminology for rating the amenability, in general, they each used a
three-tiered ranking represented here by the P,G, and E terminology.
Sources a,e: P = poor, G = good.E = excellent
Source b: P = recalcitrant, G = reluctant, E = amenable
Source c: P = recalcitrant, no/low reactivity; G = reluctant, medium reactivity; E = amenable, high reactivity
Source d: P = difficult to treat, E = susceptible
Notes:
1 Persulfate/sulfate radical reactivity studies with 66 organic compounds and isomers under various conditions have been conducted elsewhere (FMC,2005).
(http://www.envsolutions.fmc.com/Klozur8482/ResourceCenter/tabid/356/Default.aspx).
2 The reaction between O3 and H2O2 produces -OH. Therefore, the ratings from source (e) by Fenton's (H2O2/Fe) apply equally to the O3/H2O2 (Peroxone)
technology.
3 Source (c) rated Fe-catalyzed and heat-catalyzed persulfate separately; the lower rating applies to Fe-activated and the higher rating applies to heat-activated
persulfate.
4 Benzene was rated separately from TEX or petroleum hydrocarbons; thus, the BTEX or petroleum hydrocarbons rating excludes benzene.
5 TCA and DCA were rated separately by some sources; the other sources rated chlorinated ethanes as a class of contaminant.
6 A detailed summary of second-order reaction rate constants between pesticides and -OH is reported in Haag and Yao( 1992).
7 Brown etal. (2004) present experimental results indicating that permanganate, Fenton's reagent, persulfate, and ozone are effective in oxidizing 1 ,4-dioxane.
Sources:
a Sperry, K,L,and J.CooksonJr. 2002. In Situ Chemical Oxidation: Design & Implementation. ITRC Presentation to New Jersey Department of Environmental
Protection, October 30, 2002. http://www.state.ni.us/dep/srp/trainina/sessions/insitu2002 10c.pdf
b ITRC. 2005. Technical and Regulatory Guidance for In Situ Chemical Oxidation of Contaminated Soil and Groundwater, Second Edition. Interstate Technology
and Regulatory Cooperation WorkGroup, In Situ Chemical Oxidation Work Team.
c Brown, R.A. 2003. In Situ Chemical Oxidation: Performance, Practice, and Pitfalls. AFCEE Technology Transfer Workshop, February 24-27, 2003, San Antonio, TX.
http://www.afcee.brooks.af.miI/products/techtrans/workshop/postworkshop03/Tuesdav/pm/sourcezoneremediation/4 brown.pdf
d Siegrist, R.L, M.A. Urynowicz, O.R. West, Ml. Crimi, and K.S. Lowe. 2001 . Principles and Practices of In Situ Chemical Oxidation Using Permanganate. 367 pp.
Battelle Press, Columbus, OH.
e Rating based on the second-order reaction rate constants between contaminants and -OH reported in Buxton et al. (1988) and Haag and Yao( 1992): Excellent
(> 1 09 L/mol-s), Good ( 1 0s - 1 09 L/mol-s), Poor (< 1 0s L/mol-s).
In-Situ Chemical
Engineering Issue

-------
require treatment trains involving the sequential applica-
tion of technologies to accomplish the treatment objec-
tive. Chemical oxidation can be deployed under a variety
of applications, i.e., in either the unsaturated or saturated
zones, or possibly above-ground, and under a variety of
hydrogeologic environments.  In this  Issue Paper, the
focus is on ISCO.  There are potential advantages and
disadvantages of ISCO that should be assessed when con-
sidering the deployment of this technology.

Advantages:
     •  Applicable to a wide range of contaminants.
     •  Contaminants are destroyed in-situ.
     •  In-situ treatment may reduce costs incurred
       by other technologies such as pump and treat,
       MNA, etc.
     •  Aqueous, sorbed, and non-aqueous phases of
       contaminants are transformed.
     •  Enhanced mass transfer (enhanced desorption
       and NAPL dissolution).
     •  Heat from H2O2 reactions enhances mass
       transfer, reaction rates, and microbial activity.
     •  Potentially enhances post-oxidation microbial
       activity and natural attenuation.
     •  Cost competitive with other candidate
       technologies.
     •  Relatively fast treatment.

Disadvantages:
     •  Oxidant delivery problems due to reactive
       transport and aquifer heterogeneities.
     •  Natural oxidant demand may be high in some
       soil/aquifers.
     •  Short persistence of some oxidants due to fast
       reaction rates in the subsurface.
     •  Health and safety issues regarding the handling
       of strong oxidants.
     •  Potential contaminant mobilization.
     •  Potential permeability reduction.
     •  Limitations for application at heavily
       contaminated sites.
     •  Contaminant mixtures may require treatment
       trains.
     •  May have less oxidant/hydraulic control relative
       to other remedial technologies.

II.B. Definition

Chemical oxidation is a process  in which the oxidation
state of a substance is increased.  In general, the oxidant
is reduced by accepting electrons released from the trans-
formation (oxidation)  of target and non-target reactive
species.   Oxidation  of non-target   species,  including
reduced inorganics in the subsurface, also involves the
loss of electrons; however, the main target during ISCO
involves organic chemicals. Oxidation of organic com-
pounds may include oxygen addition, hydrogen abstrac-
tion (removal), and/or withdrawal of electrons with or
without the withdrawal of protons. The main objective
of chemical oxidation is to transform undesirable chemi-
cal  species into species  that are harmless or nonobjec-
tionable.  For  example, oxidation of trichloroethylene
(TCE) and perchloroethylene (PCE) may produce reac-
tion byproducts that include dichloroacetaldehyde  and
dichloroacetic acid, compounds with  lower  toxicity.
Similarly, oxidation of phenolic  compounds may pro-
duce an  assortment of  carboxylic acids (Huling et al.,
1998) that are nontoxic. Oxidation of these byproducts
to CO2 and H2O could be accomplished through addi-
tional oxidative treatment and expense, but may not be
practical for economic purposes.  These reaction byprod-
ucts may also serve as  microbial substrate for natural
attenuation processes.

II.C. Process Fundamentals

In oxidative treatment systems, numerous reactions could
potentially occur, including acid/base reactions, adsorp-
tion/desorption, dissolution,  hydrolysis,  ion exchange,
oxidation/reduction, precipitation, etc. In environmen-
tal systems there is  a wide array of reactants and condi-
tions that influence reaction rates and pathways that vary
from site to site. Often, numerous reactions are required
to achieve innocuous end products, and many of the reac-
tion intermediates are never identified. The general reac-
tions presented in this Issue Paper represent a simplified
set of reactions; however, a much broader and more com-
plex set of reactions is expected under field conditions.

II.C.I. In-Situ Permanganate Oxidation

II.C.I.a. Chemical Reactions

Reaction  1 (Table 3) is the 3-electron half reaction for
permanganate  (MnO4~) oxidation under most  environ-
mental conditions  (pH 3.5 to 12).  One of the reaction
byproducts is MnO2, and in the pH range  of 3.5 to 12 it
is a solid precipitate. Under acidic conditions (pH <3.5),
Mn in solution or  in colloidal  form may be present in
different redox-dependent oxidative states  (Mn+2> +4> +7).
Additionally, under strongly alkaline conditions, Mn may
be present as Mn+6. Under conditions where pH is <3.5
and  >12, 5-electron and 1-electron  transfer reactions
occur,  respectively  (Table 3, half reactions  2 and 3).
Reactions 1 to 3 illustrate the general permanganate reac-
tions in the subsurface. Overall, permanganate oxidation
       Engineering Issue
                     In-Itu Chemical Oxidation

-------
potentially involves various electron transfer reactions
(reactions 1 to 3), but is generally considered indepen-
dent of pH in the range of 4 to 8.

Reactions 4 to 7  (Table 3) illustrate the oxidation of per-
chloroethylene (PCE),  trichloroethylene  (TCE), dichlo-
roethylene (DCE), and vinyl chloride (VC), respectively.
Examination of  these  balanced  redox reactions indicate
that the oxidant demand is inversely correlated with chlo-
rine substitution.   For example,  the stoichiometric
requirement for PCE,  TCE, DCE, and VC is 1.33, 2.0,
2.67, and 3.33  mol KMnO4/mol contaminant, respec-
tively.

Although MnO4~ will oxidize a wide range of contami-
nants, there are  notable exceptions for compounds that
are recalcitrant, including 1,1,1-trichloroethane (TCA),
1,1-dichloroethane (DCA), carbon  tetrachloride (CT),
chloroform (CF), methylene chloride (MC),  chloroben-
zene (CB), benzene, some pesticides, PCBs, and others.
The rate  of MTBE oxidation by KMnO4 is two to three
orders of magnitude slower than other oxidation pro-
cesses indicating that  oxidation by KMnO4 limits the
applicability of the process  for rapid cleanup (Damm et
ai, 2002).

ll.C.l.b. Reaction Rate

Contaminant oxidation by MnO4~ occurs by electron
transfer rather than through the  rapid H?O? reaction and
radical attack characteristic of Fenton oxidation.  The
relatively slow reaction rate of MnO4~ in subsurface sys-
tems offers  advantages to permanganate-driven ISCO.
The slow rate of reaction allows for greater transport dis-
tances  of MnO4" during injection delivery in medium
and high permeability materials.  MnO4" persistence  in
the subsurface is proportional  to the concentration  of
MnO4" injected, and  inversely proportional  to the oxi-
dant demand by the aquifer material and contaminant(s).
MnO4" generally persists in  the subsurface for months;
however, persistence varies based on the concentration
and volume of oxidant injected and from site to site.  The
long-term persistence of MnO4~ contributes to diffusive
transport of the oxidant into low-permeability materials,
such as silty clay (Struse etal., 2002a) and fractured shale
(Parker, 2002).  Consequently, this permits deeper pene-
tration of the oxidant into aquifer materials that contain
slow-moving contaminants.

II.C. I.e. Natural Oxidant Demand

A wide range of naturally occurring  reactants other than
the target  contaminant(s) also  react with MnO4~  and
impose a background oxidant demand.  The background
oxidant demand reduces oxidation efficiency and is gen-
erally greater than the demand  imposed by the target
compound(s).    Non-target  reactants  mainly  include
organic matter  and reduced chemical species (e.g., fer-
rous,  manganous, sulfidic species).  In aquifer material
containing low quantities of organic  carbon and reduced
materials, the background oxidant demand  can be  low.
However,  under  highly  reduced  conditions  and/or
organic-rich aquifer  materials, the background oxidant
demand can be high, suggesting that  the mass and cost of
MnO4" to achieve the treatment objectives will be high
(refer also to Section III.A.3. Oxidant Demand).

ll.C.l.d. Permanganate Salts

There are two  forms of permanganate, potassium  per-
manganate  (KMnO4)   and   sodium   permanganate
(NaMnO4). Various grades of KMnO4 are available from
different chemical suppliers.  The average prices of reme-
diation grade KMnO4 and NaMnO4 are $1.80/lb and
$6.50/lb ($2.50/lb aqueous 40% solution),  respectively.
KMnO4 is available as a solid that must be mixed with
water before injection and is soluble  at approximately 60
g/L (6%).  KMnO4 concentrations are generally injected
Table 3. General Permanganate Oxidation and Related Chemical
Mn04- + 2 H20 + 3 e- 	 > Mn02(s) + 4 OH-
Mn04- + 8 H+ + 5e- 	 »• Mn+2 + 4 H20
Mn04- + e- 	 >• Mn04-2
4 KMn04 + 3 C2CI4 + 4 H20 	 »> 6 C02 + 4 Mn02 + 4 K+ + 8 H+ + 1 2 Cl
2 KMn04 + C2HCI3 	 >• 2 C02 + 2 Mn02 + 2 K+ + H+ + 3 C|-
8 KMn04 + 3 C2H2CI2 	 >• 6 C02 + 8 Mn02 + 8 K+ + 2 OH- + 6 Q- + 2 H
1 0 KMn04 + 3 C2H3CI 	 > 6 C02 + 1 0 Mn02 + 1 0 K+ + 7 OH- + 3 C|- +
Reactions
(pH 3,5-12)
(pH <3.5)
(pH>12)


2o
H20

(D
(2)
(3)
(4)
(5)
(6)
(7)
 In-Situ Chemical
                        Engineering Issue

-------
at 0.5  to  2.0%  and occasionally  up to 4% (40 g/L).
Precipitation of KMnO4 in mixing tanks, delivery lines,
or in the subsurface may occur at high KMnO4 concen-
trations and/or in  conjunction with  low temperatures.
The simultaneous presence of VOCs and MnO4~ in water
samples is  uncommon but may occur at low temperature
when NAPL is present in the water sample.  NaMnO4 is
more soluble (400 g/L; 40%) than KMnO4, and is sup-
plied as a  liquid.  The high concentration of NaMnO4
provides flexibility  in oxidant  delivery to the subsurface
and eliminates the potential for KMnO4 granules/dust
exposure during oxidant handling and mixing into solu-
tion.   The density  of permanganate  solutions  is often
greater than water (1.00 g/mL). For example, KMnO4 is
generally injected as a 2 to 4% solution which has a den-
sity of 1.02 to 1.04 g/mL, respectively. Density-driven
transport of MnO4~ facilitates the vertical transport of the
oxidant both in porous and fractured media, and enhances
distribution and  contact between oxidant and  contami-
nants. This transport mechanism has been  documented
in several  field-scale studies (Parker,  2002). NaMnO4
solutions at higher  concentration have even  greater den-
sity and also undergo density-driven transport. The form
of the oxidant (NaMnO4 vs. KMnO4) has little effect on
oxidant consumption or  filterable   solids  production
(Siegrist «^/., 2002).

ll.C.I.e. Impact of MnOJs) Accumulation

Mass  Transfer: The accumulation of MnO,(s) at the
NAPL interface  may interfere with  mass transfer, and
excessive accumulation  in  porous  media may result  in
permeability reduction.   A laboratory  study involving
visualization experiments has shown MnO2(s) to form a
rind around high DNAPL saturation zones (Conrad et
a/,., 2002;  Li and Schwartz, 2004a).   The DNAPL was
sequestered, and  a reduction both  in the delivery of the
oxidant and in contaminant (TCE) oxidation was mea-
sured.  MnO2(s) formed around the PCE DNAPL, and
appeared to cement sand  particles together forming a
rock-like material with low permeability. Correspondingly,
advective transport of the oxidant solution  adjacent  to
the PCE DNAPL was reduced. Under this condition, it
was proposed that diffusive transport of MnO4~ and PCE,
to and  from the DNAPL, respectively, was the only trans-
port mechanism that could facilitate chemical oxidation.
In a model aquifer, MnO2(s)  deposits on or near  PCE
decreased  both the velocity of water  directly above the
pool and the overall mass transfer from the remaining
PCE pool (MacKinnon and Thomson, 2002).  Results
indicated that MnO4~ was capable of removing a substan-
tial mass from the PCE DNAPL pool. However, perfor-
mance of  ISCO as a pool  removal technology will be
limited by the formation and precipitation  of hydrous
MnO2 that occurs during the oxidation process. In other
studies,  soluble chlorinated VOCs and TCE DNAPL
were oxidized by MnO4~, and  negative impacts from
excessive accumulation of MnO2(s)  were  not observed
(Chambers etal., 2000b; Struse etal., 2002a).

Permeability Reduction: A few cases have been reported
where a loss in permeability was attributed to excessive
MnO2(s) accumulation. Clarification of the different
mechanisms  and other possible causes  is useful  to
prevent  or  minimize  this  condition.  Calculations
involving MnO2(s)  precipitation and deposition  in
aquifer  pores  under  a  wide  range  of conditions,
including porosity (0.2 to 0.4), bulk density (1.6  to
2.13 g/cm3), and oxidant demand (1  to  60  g/Kg),
indicate  that  8%) or less of the voids  are  filled  by
MnO2(s)  (Luhrs et  al., 2006).  These calculations
suggest that blockage of ground-water flow by filling
aquifer  pores with MnO2(s) in  porous  media is  an
unlikely  explanation.   Analysis  of the  MnO2(s)
content  in   aquifer  core  samples  yielded  similar
conclusions (Siegrist et al., 2002).

Reductions in permeability are more  likely attributed to
the nonuniform accumulation of MnO2(s)  in  porous
media   due   to  mechanical  straining,   electrostatic
interactions,  chemical bridging,  or specific  adsorption.
Aquifer or well-pack media near the injection point may
"ripen" nonuniformly with MnO2(s) resulting in localized
high levels of MnO2(s).  Further, MnO4~ distribution
under field conditions is not uniform and also contributes
to the nonuniform accumulation of MnO2(s).

MnO2(s) accumulation at NAPL interfaces is well-doc-
umented in lab studies (Li and Schwartz, 2000; Reitsma
and Marshall, 2000;  Reitsma and  Randhawa,  2002);
however, permeability reductions are rarely measured
and reported  in field studies. One explanation for this
apparent discrepancy is that  reductions in the permea-
bility are localized near the DNAPL zones  (Reitsma and
Randhawa, 2002).  Under field conditions, DNAPLs
are distributed heterogeneously and screened intervals
are rarely  entirely completed  in  DNAPL-saturated
porous media and may only  be minimally impacted  by
MnO2(s) accumulation. Permeability reductions and/or
a decline in mass transfer may occur from the formation
of a rind on DNAPL, as described above, but field tests
and site  characterization tools  may be insensitive  to
detect the effect. However, localized DNAPL-dependent
accumulation of  MnO2(s) and localized permeability
reduction attributed to this mechanism may play a sig-
nificant role in the mass transport and mass transfer of
       Engineering Issue
                     fri-Slu Chemical Oxidation

-------
MnO4~ and VOCs, contaminant oxidation, and reme-
diation time frames.  Excessive MnO2(s)  accumulation
near high DNAPL saturation areas, and  the associated
reduction in NAPL mass transfer described above, pro-
vides a practical explanation for contaminant rebound
which is veiy common in source areas (refer to Section
III.F.l. Untreated COCs/Rebound).

Permeability reduction may also be attributed to particu-
lates in the injected fluids and/or gas production or injec-
tion (Luhrs et al, 2006).   ISCO  systems that  involve
permanganate injection, extraction,  permanganate  re-
amendment and re-injection  may inadvertently inject
suspended MnO2(s). Assuming the extracted fluid con-
tains   oxidizable  material   that  reacts  with  MnO4~,
MnO9(s) can form in solution before it  is injected.
Additionally, some MnO4~  mixtures have high silicate
content. Therefore, injecting large volumes of unfiltered
oxidant solutions into a well,  as in the case with injec-
tion/re-injection wells, may result in the accumulation of
solids  (MnO2(s), silicates) in or near  the well pack.  In
one study involving a 5-spot pattern where MnO4~ was
injected  in a central well, and extracted  in four adja-
cent  wells,  increasing  injection  well pressures were
required to maintain constant flow (Siegrist et a/.., 2002).
Recirculation rates declined shortly after  MnO4~ injec-
tion.  Although the precise cause for the fouling was not
determined, re-injection of a highly concentrated TCE
solution (600 mg/L) with MnO4" (3000 mg/L) may have
resulted  in the formation and deposition of MnO2(s)
within, or near the well screen and/or filter pack. The
reaction of MnO4~ in  or near the well, rather than in the
formation, may  have  prevented the dispersal of MnO4~
and  MnO2(s) into the  aquifer.  Another factor which
may have contributed was  the high  clay/silt content of
the aquifer  sediments in conjunction with Na+ (i.e.,
NaMnO4), which could have resulted in  dispersed soil
colloids conditions leading to a reduction in macropores
and a decline in permeability.  Permeability loss was also
reported in another field study where horizontal well-to-
well flushing of percent levels MnO4~ was performed
(Westrf*/, 1998).

KMnO4 is produced as a crystalline solid that is dissolved
in water prior to injection.  The solubility of KMnO4 is
temperature-sensitive.  Typical injection concentrations
(2 to  3 g/L) are well below the solubility (6.5 g/L @ 20
°C).  However,  differences in temperature between the
KMnO4 solution in the mixing tank  and  in the aquifer
could  result in precipitation of  KMnO4  in the  aquifer
where it is cooler. Energy is needed to achieve dissolu-
tion of KMnO4 by mixing prior to injection. If the agi-
tation applied  in the  mixing process is too  low or
insufficient time is allowed to fully dissolve the crystalline
solids, the injection solution will contain a significant
quantity of KMnO4(s) particles.  Accumulation of these
particles in the well, in the sand and gravel pack around
the well, and in the formation  near the well, can cause
loss in permeability.  Given sufficient time, the entrained
KMnO4(s) will dissolve into solution and the permeabil-
ity can be restored.  NaMnO4 is highly  soluble (40%;
400 g/L), produced and delivered as a solution, and only
requires dilution (if desired) before injection. Therefore,
precipitation of NaMnO4 is not possible. At the time of
this report, no documented cases were found where per-
meability reductions occurred at sites where NaMnO4
was injected, suggesting that KMnO4(s) precipitation is
one probable explanation for permeability  reduction.

Carbon dioxide (CO,)  is  a byproduct  from die oxidation
and mineralization of organic chemicals and  natural organic
matter (rxns 4 to 7, Table 3). In column studies, permeabil-
ity reduction and flushing efficiency decreased as a result of
MnO2(s) precipitation and from the  formation of CO2(g)
(Li and Schwartz, 2000; Reitsma and Dai,  2000; Reitsma
and Marshall, 2000; Reitsma and Randhawa, 2002). The
relative mechanistic contributions of MnO2(s) accumulation
and CO2(g) entrapment were differentiated using a pH buff-
ering method to minimize the formation of CO2(g) (Dai and
Reitsma, 2002).  In subsurface systems involving significant
reaction between MnO4~ and high concentrations of organic
chemicals, large quantities of CO2(g) can be produced in the
aquifer.  Due to the similarities  between air sparging and
CO2(g) formation during  ISCO, it is reasonable to assume
that CO2(g) entrapment may result in permeability reduc-
tion in the aquifer.  For example, increased  gas saturations
from air sparging (generally above 20% gas  saturation) can
cause  significant hydraulic conductivity reductions which
would  be detrimental to flow-through operations (Salanitro
etaL, 2000). Therefore, CO2(g) accumulation and entrain-
ment  in porous media could also result  in  blockage of
ground-water flow and permeability reduction. CO2 is solu-
ble in water and given sufficient time,  it will dissolve into
solution, thus restoring the permeability. Air in the injection
lines and equipment can be inadvertendy injected causing a
similar effect.

Increases in the permeability due to dissolution of CO2(g)
and KMnO4(s) indicate that the mechanisms responsible
for permeability reduction are reversible under ambient
conditions. Permeability reduction can be avoided during
ISCO  by filtering re-injected fluids,  selection of KMnO4
with low silicate content, and assuring adequate mixing of
KMnO4 solution before  injection.  Redevelopment of a
well may be needed to restore the permeability where the
responsible mechanism is not reversible under ambient
 In-Situ Chemical
                         Engineering Issue

-------
conditions.  Injection  of chemical reagents (organic and
inorganic acids, EDTA) into MnO2(s)-enriched aquifer
material could dissolve MnO2(s) into solution and reduce
the negative impact of MnO2(s) accumulation  (Li and
Schwartz, 2004a).

II.C.Lf. Metals Mobilization/Immobilization

There are  two main mechanisms for increasing concen-
trations of metals in the ground water during ISCO: (1)
the KMnO4 or NaMnO4 provided by the manufacturer
may contain elevated levels of the heavy metals,  and (2)
mobilization of pre-existing redox- or pH-sensitive heavy
metals (in-situ) by the oxidant.  The  content of heavy
metals in  permanganate is dependent on the type and
source of the oxidant (note: NaMnO4 has lower concen-
trations of metals than KMnO4). Both forms of the oxi-
dant are manufactured in the U.S., Germany, and China.
There is only one manufacturer of permanganate in the
U.S., and they provide analytical data for the heavy metal
impurities in their products. Remediation grade KMnO4
has been  developed  containing minimal quantities of
metal impurities. Chromium (Cr) and arsenic (As) have
historically been the impurities of concern.  Due to the
low  maximum  contaminant level  (MCL)  in drinking
water established by EPA for these metals (0.1 mg/L total
Cr MCL; 0.01 mg/L As MCL) (U.S. EPA, 2002), injec-
tion of technical grade KMnO4 may result in exceeding
the MCL for these elements. Generally, natural attenua-
tion of these metals has been achieved within acceptable
transport distances and time frames.  Due to the possibil-
ity of exposure  pathways and potential receptors, moni-
toring of these  parameters may be needed  under some
conditions.  A site-specific evaluation of the potential
impact  of heavy metals should be conducted to assess
whether ground-water monitoring  for  these metals is
needed.

Enhanced  transport of pre-existing or naturally occurring
redox or pH-sensitive metals may occur as a result of per-
manganate injection. Oxidation of Cr(III) to Cr(VI) by
MnO4"  and subsequent mobilization has been demon-
strated  in  the  laboratory (Li  and  Schwartz,  2000;
Chambers et al, 2000b).  Additionally,  Cr(Vl)  and Ni
mobilization has been  observed  under field conditions
where MnO4"  has been injected (Crimi and Siegrist,
2003).  However, Cr(VI) undergoes natural attenuation
through several mechanisms (McLean and Bledsoe, 1992;
Palmer and Puls, 1994), including adsorption to MnO2(s)
and  various iron minerals.  Several field studies have
reported anomalously high post-oxidation concentrations
of Cr(VI), but natural attenuation of Cr(VI) was observed
(Crimi and Siegrist, 2003), and cleanup concentrations
have been achieved within acceptable transport distances
and time frames (Chambers et al., 2000a). The potential
exists at any site for metals to be introduced as an oxidant
impurity, and/or pre-existing or naturally occurring met-
als to be mobilized by ISCO.  Site conditions can pro-
vide insight into whether metals mobilization could occur
including oxidant dosing, pH, buffer capacity, electrode
potential (Eh), permeability,  cation exchange  capacity,
natural metals, oxidant impurities, and local uses for the
ground  water (Siegrist et al.,  2002).  Bench-scale treat-
ability studies can be used to assess the potential signifi-
cance of metals  mobility, whether metals mobilization
may occur  under field  conditions (Chambers et al.,
2000b), and whether attenuation mechanisms immobilize
metals.  Pilot-scale studies can also be used to evaluate
metals  mobilization and attenuation prior to  full-scale
implementation. Ground-water  monitoring of metals
may be  needed  to assess whether metals mobilization
occurs at ISCO sites and whether attenuation is achieved
within an acceptable transport distance.

MnO2(s) behaves as a sorbent for numerous heavy metals
including, but not limited to Cd, Co, Cr, Cu, Ni, Pb, Zn
(Suarez  and Langmuir,  1976;  Fu et al.,  1991; McLean
and Bledsoe,  1992;  Siegrist et al., 2002)  and has been
demonstrated to  oxidize pentachlorophenol (Petrie et al.,
2002) and aromatic amines (Li et al., 2003). MnO2(s) is
the primary electron acceptor for the oxidation of As(III)
to the less soluble As(V) (McLean and Bledsoe, 1992).
Cr(VI)  adsorption and immobilization in soils is posi-
tively correlated  with  free iron oxides, total manganese,
and soil pH (Korte et al., 1976).  Adsorption of metals
onto Mn oxides increases with increasing pH, and is sig-
nificant even under acidic conditions (Fu  et al., 1991).
Adsorption of metals  onto either  Mn  or Fe oxides will
immobilize metals and restrict their transport in ground
water.  It has been  reported that Cr(VI) formed and
mobilized during oxidation undergoes natural attenua-
tion within acceptable time frames and distances; how-
ever, it is not entirely clear what role MnO2(s) has in the
oxidation of Cr(III) to Cr(VI). Under some conditions,
different oxides  of Mn  may catalyze  the oxidation of
Cr(III) to Cr(VI) (Nico  and Zasoski, 2000).  Therefore,
since permanganate-based ISCO is applied under a wide
range of geochemical conditions, this  underscores the
importance of pilot-scale testing and ground-water mon-
itoring to assess metals mobilization.

II.C.Lg. Advantages

Numerous bench-, pilot-, and full-scale studies have been
conducted resulting in a significant amount of informa-
tion leading to the documentation of fundamental mech-
       Engineering Issue
                     fri-Slu Chemical Oxidation

-------
anisms  and  the  demonstration  of this  technology.
Considerable field experience has been obtained from the
application of this technology at a wide range of sites and
conditions. The chemistry involved with MnO4~ oxida-
tion of organic contaminants is relatively simple,  and
information  and guidelines needed to effectively  and
safely inject MnO4~ into the subsurface have been well-
documented and disseminated.  MnO4~ is highly  soluble,
and high concentrations of the oxidant  can  be injected.
The long-term persistence of MnO4" in the subsurface per-
mits both advective and diffusive transport and can result
in good distribution of the oxidant.  High concentrations
of MnO/j" can result in a density greater than ground water,
causing the density-driven vertical transport of the  oxidant
into the subsurface. This also contributes to  good distri-
bution of the oxidant, especially in low-permeability mate-
rials.  MnO4~ has been successfully delivered  into a wide
range of hydrogeologic environments (i.e., aquifers com-
prised of sands, clays,  sand-clay mixtures, alluvial  materi-
als, fractured shale,  fractured  bedrock,  etc.).  Several
important environmental contaminants (chlorinated  eth-
enes)  are vulnerable to oxidation, and the reaction  rate
between MnO4" and target contaminants in  the aqueous
and NAPL phases is fast.  Short-term reduction in micro-
bial activity results from the injection of MnO4~.  However,
this does not  appear to be permanent and post-oxidation
increases in microbial numbers, activity, and contaminant
attenuation is  often  reported.  Visual confirmation of
MnO4~ presence in ground-water samples and semi-quan-
titative analysis is  possible due to  the characteristic purple
color  of the  oxidant.  Considerable field experience has
resulted in well established health  and safety guidelines.

ll.C.l.h. Disadvantages

Hydraulic short circuiting and/or preferential pathways
may result in the delivery of the  oxidant into non-target
zones.  Some important environmental contaminants are
not vulnerable to oxidation by MnO^. Some grades
of KMnO4 contain heavy metal impurities that when
injected could result in unacceptable ground-water con-
centrations. MnO2(s), the main reaction byproduct,  may
accumulate near the injection well or at the DNAPL inter-
face (i.e., encrustment) resulting in mass transport (per-
meability  reductions)  and  mass transfer  limitations,
respectively. Permeability reductions may also result from
CO2(g) releases. Ion exchange of Na+ in NaMnO4 for
divalent cations in the aquifer matrix may result in  dis-
persed  soil  colloids  and  contribute to permeability
reductions. Permeability reduction is  rarely reported
and can largely be avoided by adhering to  design  and
operational  guidelines.  A  high  background oxidant
demand in aquifer and soil  material attributed to natu-
rally occurring non-target reactants may result in exces-
sive  and costly  oxidant requirements.   High oxidant
concentrations resulting in the density-driven transport
of the oxidant from the targeted zone may result in the
inefficient utilization of oxidant.  EPA has established a
secondary maximum  contaminant  level  for drinking
water for manganese (0.05 nig/L) based on color, stain-
ing,  and taste. Relatively little information is available
regarding the  long-term impact of the manganese resid-
ual on ground-water quality.

II.C.2. In-Situ Fenton Oxidation

//.C.2.a. Fenton and Related Reactions

In general, Fenton chemistry and Fenton oxidation treat-
ment systems  are more complex than the permanganate
oxidation treatment system.  This is mainly attributed to
numerous reaction intermediates, side  and competing
reactions, phases  (gas, liquid,  solid, NAPL), and the
numerous parameters which directly and indirectly affect
Fenton-driven transformation reactions. A detailed and
rigorous field demonstration of in-situ Fenton oxidation
has not been conducted.  Studies are needed to quantify
reaction mechanisms, clarify technical  issues, and opti-
mize the treatment process.  A brief summary of Fenton
chemistry is presented to elucidate process fundamentals
and mechanisms.

The classic Fenton reaction specifically involves the reac-
tion between H2O2 and ferrous iron (Fe(II)) yielding the
hydroxyl radical  (-OH) and ferric (Fe(III)) and hydroxyl
ions (OH-) (Table 4, rxn 1). Fe(III) reacts with H2O2
(Table 4, rxn 2) or the superoxide radical (-O2") (Table 4,
rxn 3) yielding Fe(II). This general sequence of reactions
continues to  occur until the H2O2 is fully consumed.
Since H2O2 injected into the subsurface reacts with many
chemical species  other than Fe(II), this technology  is
often referred to  as catalyzed hydrogen peroxide (CHP).

•OH has an unpaired electron  making it a highly reac-
tive, nonspecific oxidant (Table 4, rxns 4 and 5). Corres-
pondingly,  quasi  steady-state concentrations of -OH in
Fenton  systems are very low (10~14 to 10~16 M) (Huling
etaL 1998; 2000; 2001). Due to the fast reaction rates
of-OH, the transport distance of-OH is limited  to only
a few nanometers.  Therefore, a basic tenet of Fenton oxi-
dation is that the contaminant, Fe(II), and H2O2 must be
in  the same location at the same time.

Nonproductive reactions are  represented by the  general
disproportionation reaction (Table 4, rxn 6) where H2O2
is consumed,  -OH is not produced, and O2 is a reaction
 In-Situ Chemical
                         Engineering Issue

-------
 Table 4.  Fenton and Related Chemical Reactions
   H202 + Fe(ll) 	»• Fe(lll) + -OH + OH-
   H202 + Fe(lll) 	> Fe(ll) + -02- + 2 H+
   •02- + Fe(lll)	>• Fe(ll) + 02(g) + 2 H+
   •OH + target contaminant	>• reaction byproducts
   •OH + H202	>• -H02 + H20
(1)
(2)
(3)
(4)
(5)
(6)
byproduct (Huling etal., 1998).  Examples of these reac-
tants include catalysts such as catalase, a microbial enzyme
found ubiquitously in the subsurface environment, and
some transition  metals such as manganese.  Mn  cycles
between oxidation states similar to  Fe, but -OH  is not
produced (Pardieck et d., 1992).

ll,C.2.b. Contaminant Transformations

A  wide  range of organic compounds  of environmental
significance have moderate to  moderately high second-
order reaction rate  constants with -OH, indicating fast
reaction rates (Table 2).  A comprehensive compilation of
reaction rate constants  has  been published  for a wide
range of reactants with -OH (Dorfman and Adams, 1973;
Buxton et al., 1988; Haag and Yao, 1992). Among the
contaminants represented in these references are haloge-
nated and  non-halogenated volatile organics  (ketones,
furans), halogenated semivolatile organics  (PCBs, pesti-
cides, chlorinated benzenes and chlorinated phenols) and
non-halogenated semi-volatile  organics  (PAHs, non-
chlorinated  phenols). Fenton  oxidation, therefore,  has
potential applicability at a large  number  of hazardous
waste sites.  Compounds with double bonds are especially
vulnerable to -OH oxidation, e.g., TCE, PCE.

In general, oxidized (halogenated) compounds  without
double bonds are poorly reactive with -OH, including
carbon  tetrachloride,  chloroform, methylene  chloride,
1,1,1-  and  1,1,2-trichloroethane.  Although  Fenton
oxidation may lead to complete mineralization of organic
contaminants, this  is  usually  performed under ideal
laboratory conditions where process limitations have been
minimized.   In  the subsurface environment, non-ideal
conditions  (discussed  below)  contribute  to  process
inefficiency and incomplete mineralization. Consequently,
residual concentrations  of  the target  compound may
occur,  and  reaction  intermediates may  accumulate.
Reaction intermediates are commonly less hazardous than
the target  compound.   For example,  carboxylic and
chloroacetic acid compounds are relatively nontoxic and
may accumulate from  the  oxidation of 2-chlorophenol
(Huling et al.,  2000) and TCE, respectively.  Incomplete
oxidation of MTBE may result  in tertiary butanol (TBA),
acetone, and tert-buty\  formate (Chen et al., 1995; Yeh
and  Novak,  1995;  Huling  et  al.,  2005)  which  are
considered less toxic than MTBE; however, they may be
unacceptable in some  situations.  Although TBA and
acetone  also  undergo  transformation,  acetone  may
accumulate relative to TBA because it has a lower reaction
rate constant with -OH (1.1 x 108 M'1 sr1) than TBA (6x 108
M'V1) (Buxton etal., 1988) and is a byproduct from the
oxidation  of  TBA and other MTBE  transformation
intermediates  (Stefan et al., 2000).   The products  of
MTBE oxidation also include a variety of carboxylic acids
(Stefan and Bolton,  1999; Stefan et al, 2000)  and,
ultimately, CO2.

II.C.2.C. Other Transformations

Increasing information suggests that reductive transfor-
mations in Fenton-driven oxidation systems  may play a
role in the degradation of heavily  chlorinated and  nitro-
substituted compounds (Peyton etal., 1995; Watts etal.,
1999).  These reactions may  be  attributed to various
reductants, including superoxide  radical (-Cv), hydro-
peroxide anion (HO2~)  (rxns 1  to 3, Table 5), and  possi-
bly Fe(II).  It has been reported  that the perhydroxyl
radical (-HO7) is not a significant reductant (Watts et al.,
1999); however, the pKa for -HO2 and -Of is 4.8, indi-
cating that some -O2" would be present under most envi-
ronmental  conditions  where  in-situ Fenton oxidation
(ISFO) is implemented.

A  review of Fenton-driven  reductive reactions indicates
that quinones, nitrobenzenes, nitrogen heterocycles, car-
bon  tetrachloride,  and chloroform  are vulnerable  to
superoxide radical transformation  (Watts et al.,  1999).
Many halogenated and nitro-substituted contaminants,
such as PCE and nitrobenzene, react with both -OH and
reductants  at  near-diffusion-controlled rates; therefore,
their degradation in vigorous Fenton-like reactions may
proceed  through  parallel  oxidations  and  reductions
(Watts etal., 1999).  This has several important implica-
Tables
. Formation of Reductant Chemical
Fenton-Driven Chemical Reaction
Species in
System
.OH + H202 	 x.H02 + H20
•H02<-
•H02 +
	 »>H
•cv —
* + -02- (pKa = 4.8)
-^> H02- + 02




(1)
(2)
(3)
       Engineering Issue
                           In-Itu Chemical Oxidation

-------
tions regarding subsurface remediation.  Reductive trans-
formations, when combined with oxidation, yield greater
potential for overall  contaminant transformation.  For
example, TCE is vulnerable to reductive transformation
and -OH oxidation  (k.OH = 4.2xl09).  The reaction
byproduct, chloroacetic acid, is vulnerable to reductive
transformation (ke_ = 6.9xl09).  It is not uncommon for
hazardous waste sites to contain a mixture of contami-
nants  vulnerable to oxidative treatment (benzene, tolu-
ene,xylene) and reductive treatment (1,1,1 -trichloroethane
(1,1,1-TCA), carbon tetrachloride (CT)).  Under this
condition, Fenton oxidation may be an effective remedial
technology for the contaminant mixture.  The  radicals
responsible  for  contaminant  oxidative and  reductive
transformations are highly reactive and are nonspecific,
indicating that radical scavenging may be a potential lim-
iting factor for both reactive pathways. Site-specific tests
are needed to assess the overall role of both oxidative and
reductive contaminant  transformations for mixed waste
conditions.

II.C.2.d, Scavenging

•OH will react  with  naturally  occurring and  anthropo-
genic  non-target  chemical species present in soil and
aquifer material, e.g., H2O2 (rxn 5, Table 4).  The non-
target chemical  species  "scavenge" -OH which may oth-
erwise  oxidize  the  target  contaminants.    Common
ground-water anions (NO3', SO42', Ch HPO42', HCO3',
CO32-) react with -OH (Buxton et al, 1988; Pignatello,
1992; Lipczynska-Kochany et al., 1995) and  may  be a
source of treatment inefficiency. Because H2O2 is gener-
ally present at high concentrations in Fenton systems and
has a moderate rate constant  for reaction with  -OH
(2.7x107 M-V1, Buxton et aL, 1988), H2O2 is itself a
primary source  of inefficiency in Fenton-driven  systems
(Hiding #*/. 1998).

II.C.2.&. OJg) Generation and Exothermic Reaction

In-situ Fenton oxidation involves the injection  of high
concentrations of H9O2, a chemical that is 94.1%  oxy-
gen.  For example, assuming 1 mol O2(g)/2 mol H,O,,
there is approximately 1400 ft3 O2(g) (standard tempera-
ture and pressure) released from 55 gal of 50% H2O2.
O2(g) produced in the  subsurface as a result of H2O2
reactions sparges the saturated zone  and perfuses the
unsaturated  zone.  Air sparging  (Ahlfeld  et al, 1994;
Hein  et al.,  1997; Johnson, 1998) is a technology that
has been rigorously investigated and shares many  similar-
ities with O2(g) sparging that occurs in Fenton systems.
A review of air sparging literature provides insight to mass
transport and mass transfer mechanisms involving O,(g)
sparging resulting from ISFO systems.  The production
of O2(g) in saturated porous media during ISFO may be
problematic.   A significant complication of air sparge
wells  used to intercept a  ground-water plume is the
decline in permeability of the formation due to entrapped
air and  air channels  (Ahlfeld  et al.,  1994).  This could
result in a 95% reduction in conductivity for many aqui-
fers.  The low permeability barrier would impede the nat-
ural gradient of ground-water flow,  and could  result in
the flow  of ground water around  the sparged zone
(Ahlfeld et al., 1994; Rutherford and Johnson,  1996).
Increased gas saturations (generally above 20%  gas satu-
ration)  can  cause  significant  hydraulic  conductivity
reductions which would be detrimental to flow-through
operations (Salanitro  et al., 2000).  Reduced permeability
was caused by colloidal fouling and O9(g) binding due to
H,O2 decomposition in porous  media (Weisner et al.,
1996).  Due to the similarities between air sparging and
O2(g) sparging from injected  H2O2, it is reasonable to
assume  that O2(g) entrapment and O2(g) channels  may
interfere with ground-water  transport, the delivery of
H2O2,  rebound,  and  delayed  or poor  mass  transfer
between aqueous, NAPL, and  sorbed (solid) phases (refer
to Section II.C.3.b. In-Situ Application).

O2(g) sparging can enhance volatilization of environmen-
tal contaminants from the ground water. The pressure
buildup from O,(g) production can pneumatically trans-
port ground water and NAPL away from the treatment
area and cause artesian conditions in nearby monitoring
wells.  Although an  in-depth field investigation of gas
flow blockage, mass transfer,  and  transport of contami-
nated ground water/NAPL away from the treatment zone
has not  been conducted, qualitative  information at  sites
where ISFO has been implemented indicate these mech-
anisms have occurred.

Fenton  and related reactions are exothermic, resulting in
heat release and  elevated  temperatures during ISFO.
Heat  accumulation  near the  injection well is common
due to rapid decomposition of H2O2 and the slow dissi-
pation of heat. Injection wells and nearby monitoring
wells  constructed  of PVC  have  melted during ISFO.
Since the melting point of PVC is 200  °C, this suggests
that very high localized temperatures have resulted.  The
elevated temperature and production of steam (100 °C)
represents a safety hazard when performing ISFO-related
field activities. Heat production  is functionally depen-
dent on the volume and concentration of H2O2 injected,
the rate of H2O2 injection, and  H2O2 reactants in the
subsurface. Stainless or carbon steel injection and moni-
toring wells have been  used to withstand elevated tem-
peratures during ISFO.
 In-Situ Chemical
                        Engineering Issue

-------
ti.C.2,1. Injected Reagents

Various reagents are injected during ISFO to either facil-
itate or enhance contaminant oxidation, including H2O2,
ferrous  Fe, acid, and stabilizers. The volume  of H2O2
solution injected should be sufficient  to fully contact the
targeted zone. The H2O2 concentration should be opti-
mally balanced to minimize -OH scavenging and  to pro-
vide sufficient oxidative treatment. The reaction between
Fe(II) and H9O, is veiy fast. Therefore, each of these solu-
tions should  be  injected  into different wells or injected
separately into the same well (pulsed), but not co-injected
into the same well.  Due to  rapid H2O2 reaction in the
subsurface, high injection rates, shorter injection well
spacing, and  lower pH can improve  H2O2  distribution.
The use of other reagents (Fe, acid, stabilizers) with H2O2
should be based on documented and demonstrated cost
effectiveness and should  be  evaluated on a case-by-case
basis.

II.C.2.f(l)H2O2

H2O2 solutions are clear and can be mixed with water in
any proportion.  Field and laboratory colorimetric analy-
sis of H2O2 can be performed using various methods and
field kits.  The concentration range  using field kits are
low  and  require significant  dilution of  the sample.
Colorimetric  methods involving higher calibration ranges
involve  titanium sulfate (Schumb etal., 1955; Huling et
al., 1998)  and iodometric titration (Schumb etal., 1955).
H2O2 is available throughout the U.S. and is usually pro-
duced and purchased in  bulk at 35% or 50% by weight.
In many early ISFO projects, H2O2 had been injected at
or near these  concentrations despite the high  potential
for -OH scavenging (U.S. DoD, 1999). Injection of H2O2
at lower concentrations (1 to 10%) would reduce H2O2
scavenging,  increase the  volume  of oxidant solution
injected and,  thus, the volume of aquifer contacted, and
result in lower temperatures  (and H2O2 reaction  rate) at
the injection well head.

There is an abundance of reactive species that  will react
with H2O2 including, but not limited to, heavy and tran-
sition metals  (Ca, Cr, Mn, Fe, Co, Ni, Cu, Zn,  As, Se,
Mo, Rh, Pd, Ag, Cd, W,  Os, Ir, Pt, Au, Hg, Pb, Bi, Po),
halogens (Cl, Br, I), microbial enzymes (catalase,  peroxi-
dase), and organic matter (Schumb etal., 1955).   In the
subsurface, a  sufficient abundance  of  reactive  species
exists, mainly iron; and  rapid H2O2 decomposition is
often observed limiting the persistence of H2O2 to short
periods  (1  to  12 hours). Assuming good contact between
H2O2 and the targeted zone is achieved, the rapid rate of
H2O2 reaction may be considered advantageous since
this leads to short-term disruption of commercial activi-
ties at the site.  However, the rapid H2O2 reaction rate
will  impede  H2O2 transport  and  delivery to targeted
zones. The short duration of H?O2 in the subsurface also
may prevent  the diffusive transport of H,O2 into low-
permeability materials containing contaminants.

Numerous physical and chemical differences between
bench- and field-scale conditions affect the reaction rate
of H2O2 in both systems. Therefore, it is recommended
that H2O2 reaction rate kinetics generated from bench-
scale treatability studies not be used to design injection
well spacing  at field-scale.   The transport  distance,  or
radial influence of H2O2 from the injection point, is best
determined by  monitoring ground water for  H2O2  in
monitoring wells during pilot-scale ISFO. This informa-
tion can be used to design the radial distance between
injection wells  for adequate  coverage  during full-scale
ISFO.

II.C.2,f(2) Iron (Fe)

Ferrous  sulfate (FeSO4) and other salts of Fe(II) have been
co-injected with H2O2 to facilitate the Fenton reaction.
The concentration of Fe(II) injected into the  subsurface
has generally  been above  background concentrations but
low (e.g.,  20 to 100 mg/L) relative to [H2O2].  Under
this condition, the relative abundance of Fe(II) may con-
tribute to -OH  production and contaminant  oxidation.
Due to  the slow Fe(III)  reduction reactions  relative  to
the rapid  Fenton reaction (i.e., Fe(II) oxidation), a less
efficient and slower rate of-OH production occurs after
Fe(II) is initially reacted with H2O2. Consequently, one
disadvantage  of Fe(II) amendment is that stoichiometric
quantities are required. Fe(II) is vulnerable to  numerous
reactions (complexation, oxidation, precipitation) which
immobilize the  catalyst and minimize transport distances
and distribution in the aquifer. For example, Fe(II) sorp-
tion and saturation of the Fe(III) surface can occur from
the high  stability of the  Fe(III)-O-Fe(II) interaction
(Roden  and Zachara,  1996). Fe(III) is an unstable form
of Fe which is vulnerable to precipitation and complex-
ation, thus  becoming  immobile. Fe(III)  precipitates
above ~ pH 3.5 to  hydrous ferric oxide (ferrihydrite)
which behaves  as a poor Fenton catalyst relative to the
soluble  form. Fe(II) is also involved in various chemical
and physical  reactions which may  immobilize the cata-
lyst and limit the transport distance  from the injection
well.

The reaction between Fe(II) and  H2O2 is rapid, and
simultaneous injection  (mixing)  of  Fe(II) and  H2O2
before injection or in the  injection well results in the
       Engineering Issue
                     fri-Slu Chemical Oxidation

-------
Fenton  reaction occurring in or very near the injection
well.  This is an inefficient use of both Fe(II) and H2O2
and can be dangerous.

Iron is one of the most common elements found in soils
in the U.S. (average [Fe]SOIL =  26,000 mg/kg, n=1318)
(Shacklette and Boerngen, 1984).  Not all Fe present in
soil and aquifer material is available for reaction with
H2O2. Nevertheless, at many sites, there is an abundance
of naturally occurring heterogeneous forms of Fe which
serves as  the predominant  source of catalyst  for the
Fenton mechanism.  Under low pH and/or reduced con-
ditions, some of the total Fe may be Fe(II), which if avail-
able, could react with H2O2, yielding -OH. At most sites
where Fenton oxidation is carried out, naturally occur-
ring Fe, not the Fe(II) co-injected with H2O2, is predom-
inantly responsible for H2O7 reactions.

Reduced permeability or fouling of injection  wells attrib-
uted to  Fe injection may have occurred at some  sites but
is not well-documented.  Precipitation of Fe(II), forma-
tion of colloidal Fe particles, and entrapment in  the pore
throats  of porous media  could result in a permeability
reduction. For example,  the simultaneous injection of
H2O2 and Fe(II) (which  is  not  recommended)  would
likely result in Fe precipitation and immobilization in or
near the well  screen and/or sand pack.  Fe oxidation and
precipitation  resulting from  H2O, injection  alone could
not explain the  reduced permeability observed in injec-
tion wells at  a bioremediation field site (Weisner et al.,
1996).  Colloidal clay particles mobilized during injec-
tion have resulted in permeability losses (Weisner et al.,
1996, and references therein).

II.C.2.f(3) Acidification

H2O2 stability, contaminant oxidation efficiency, and Fe
solubility and availability are greater under acidic condi-
tions (pH 3 to 4) than in the near neutral pH range (pH
6 to 8) or higher.  These effects  are desirable in ISFO;
therefore, pretreatment  via acid injection  or acidification
of the injected  H9O9 solution is common.   The overall
Fenton-driven oxidation reaction is acid-generating, which
also contributes  to  acidification.  Most aquifer  and soil
materials are  well buffered in the  near-neutral pH range
which resists  acidification. Similar to H2O2  and  Fe(II),
reaction  of the  injected  acid with  naturally occurring
chemicals will limit  the transport distance from the injec-
tion well. Acidification of ground water is poorly docu-
mented but generally persists over a short time frame (<1
to 3 days) and rebound to background pH may occur very
rapidly in well-buffered systems. At the time  of this pub-
lication,  no information  was available which indicated
long-term, post-ISFO  persistence  of acidic conditions.
In poorly buffered systems, acidification and acid trans-
port will be less problematic or may not be needed alto-
gether.

Enhanced  transport of some  pH-sensitive  metals may
occur under acidic conditions. Bench-scale treatability
studies can be used to assess the potential significance of
metals mobility  and whether  metals mobilization may
occur under field conditions.  Ground-water monitoring
at pilot- and/or full-scale ISFO sites is needed to assess
metals  mobilization and whether  attenuation  occurs
within an acceptable transport  distance.

II.C.2,f(4) Stabilizers

Various  reagents  have  been injected to enhance ISFO
performance.  Mainly, these are intended to enhance the
transport distance of H2O2 and  Fe(II) in the aquifer.
The most common H2O2 stabilizer involves  various
forms of phosphate which reduces the availability of inor-
ganic reactants (i.e., Fe, Mn,  etc.) via complexation or
precipitation reactions.  By design, the stabilizer itself is
immobilized through these reactions and the  transport
and areal influence of the phosphate stabilizer may be sig-
nificantly limited depending on the  composition of the
aquifer  material.   H2O2 degradation  rates have  been
shown  to  decline in the presence  of some stabilizers
(Britton,  1985; Kakarla and Watts,  1997;  Watts et al.,
1999) relative to unamended controls in laboratory stud-
ies.  However, significant degradation (97%)  of H2O2
(15 M) occurring over short  transport distances (5 in,
12.5 cm) using high concentrations of phosphate  (>10 g/
L as P) stabilizer (Kakarla and  Watts, 1997) suggests this
form of stabilization may be  impractical. Unsuccessful
H2O2 stabilization in field studies from phosphate  addi-
tion was attributed to microbial enzyme and Fe catalysts
in the porous media (Spain et al., 1989; Hiding et al.,
1990; Hinchee etal., 1991; Aggarwal and Hinchee et al.,
                           CJC?
1991).  Such naturally occurring enzymes, found  ubiqui-
tously and often abundant, are highly efficient in H2O,
disproportionation and are unaffected by phosphate sta-
bilizers.  Field-scale transport   of  stabilizers and  their
impact on H2O2 transport and reaction in ISFO systems
have not been demonstrated.

Stabilizers  also include ligands and chelators that  com-
plex Fe(II) in the  near neutral pH range allowing it to
remain  in solution  and ideally to  enhance the transport
distance in the  aquifer.  Numerous  ligands have  been
tested in conjunction with the Fenton mechanism (Sun
and  Pignatello, 1993),  but two ligands, nitrilotriacetate
and  N-(2-hydroxyethyl)  iminodiacetate, appear to  be
 In-Situ Chemical
                         Engineering Issue

-------
most effective  (Pignatello  and Baehr, 1994;  Pignatello
and  Day, 1996). Subsurface transport and effectiveness
of the ligand in ISFO systems may be limited either by
reaction  with -OH  (Sun and Pignatello,  1993) or  by
sorption of  the ligand  to soil  and  aquifer  material.
Enhanced transport of an Fe-ligand complex was demon-
strated in a soil column relative to Fe in an acidified solu-
tion (Kakarla etal., 2002).  In this case, a 13% reduction
in the  initial Fe(ll)  concentration  (685  mg/L as  Fe)
occurred over 7.9 in  (20 cm) column of aquifer material.
One potential advantage of injecting an  Fe-ligand com-
plex into the aquifer in the near  neutral pH  range  is to
avoid the need to adjust the subsurface pH.   However,
while the transport  distance of the Fe-ligand complex
may increase, the transport of H2O2 can be significantly
limited in the near-neutral pH range due to rapid decom-
position. Field documentation of the simultaneous trans-
port  of H2O9  and  an  Fe-ligand complex, and  cost
information for the Fe-ligand solution,  have not been
reported.

Given the abundance of naturally occurring Fe, the sig-
nificant limitations for Fe(II) transport, and the unproven
performance and documentation of Fe and H2O2 stabili-
zation in ISFO systems,  it is currently unclear whether
the injection of Fe(II), and the injection of stabilizers for
Fe(II) and H2O2 during ISFO are cost-effective.

II.C.2.g. Advantages

Potential advantages of in-situ  Fenton  oxidation  are
included in Table 6.

Potential limitations of Fenton-driven  oxidation,  dis-
cussed below, may also be potential advantages. For exam-
ple, the heat and O2(g) released during Fenton oxidation
may enhance mass transfer via the dissolution of NAPL,
 Table 6. Potential Advantages of In-Situ Fenton Oxidation
    •OH is a powerful nonspecific oxidant that will react rapidly with
    many environmental contaminants.
    Reactions involving H202 are rapid,and it generally persists for
    <12 hours.
    Intermediate chemical species (-02-, H02-) may reductively
    transform contaminants. Fenton oxidation could address
    complex mixtures of organic compounds.
    Enhanced natural attenuation may be attributed to 02(g)
    and heat. Oxidized inorganics may also serve as terminal
    e~ acceptors (refer to Section III.F.5. Impact of ISCO on Natural
    Attenuation and Biodegradation).
   Low cost of H202 ($0.26/lb; $39/1000 equivalents).
desorption  from  the solid  phase, and volatilization.
However, mobilization of NAPLs could increase the sur-
face area of the NAPL for mass transfer, and the heat/
O2(g) could increase mass transfer to the aqueous phase.
Several  cases have  been reported  where post-oxidation
ground-water contaminant concentrations were  elevated
(U.S. DoD, 1999) and could be attributed to DNAPL
mobilization. Enhanced volatilization during ISFO could
result  in  unacceptable exposure  pathways  and  risks.
However, ISFO in conjunction with a vacuum extraction
system could be used to enhance  the control, capture, and
removal of volatile emissions, and thus prevent  the dis-
persal of volatile contaminants in the environment.  Mass
transfer of contaminants is often an important  limiting
factor to subsurface remediation. Therefore, these enhanced
mass transfer mechanisms could enhance remediation effi-
ciency, but have not been  rigorously investigated or doc-
umented.

II.C.2.h. Disadvantages

There  are several potential limitations  to Fenton-based
remediation strategies which  should be  evaluated (Table
7).  Understanding the limitations of ISFO will allow sci-
entists and engineers to better understand  the strengths
and weaknesses of the technology and allow greater oppor-
tunities for improvements in  the technology. Qualitative
information of these limitations and mechanisms has been
reported in  case studies (U.S. DOE, 1997; U.S. DoD,
1999; ITRC, 2005) and is summarized as follows: post-
oxidation  increases in soil gas  contaminant concentra-
tions; steam production; mass flux  of volatiles from wells
near the injection  zone; heat released,  asphalt upheaval,
explosions, fire; overflowing wells; post-oxidation redistri-
bution of contaminants, etc. Several undocumented exam-
ples of excessive heat and gas releases have been reported
elsewhere  (Nyer and Vance, 1999).  Early applications of
Fenton  oxidation led to these problems and occurred as
the technology was developing and guidelines  for design
and operations were limited.  Improvements in the state
of the practice of ISFO over the last few years have con-
tributed to a reduction in the number of reported prob-
lems  and  health   and  safety  incidents  from   field
applications.

II.C.3. In-Situ Ozone Oxidation

ll.C.S.a. Overview

O3  is a gas and a strong oxidant that is sparingly soluble
in water and upon reaction does  not leave a residual (i.e.,
SO42', MnO2(s)) other than O2.  Analysis of dissolved
O3  in aqueous solutions can be performed using an
       Engineering Issue
                             Chemical Oxidation

-------
    Table 7. Potential Limitations of In-Situ Fenton Oxidation*
        Excessive H202 decomposition via nonproductive reactions,*
        Radical scavenging.*
        Low reaction rate between some target contaminant and -OH, -02-, H02- .*
        pH modification (acidification) is problematic in well buffered aquifers.*
        Problematic delivery of H202 Fe(ll), acid, and stabilizers due to reactive transport.
        Production of 02(g) contributes to reductions in permeability. This may reduce the flow of ground water and injected reagents
        through the targeted contaminated zones. It also results in sparging which contributes to volatilization and redistribution of
        contaminants.
        Pneumatic transport of volatiles, NAPL,and contaminated ground water away from the injection point; heaving asphalt, excessive
        pressure.
        Incomplete oxidation and mobilization of metals.
        Excessive release of heat and elevated temperatures associated with high H202 concentrations may damage/melt PVC/plastic
        wells, screens and enhance volatilization, NAPL transport, H202 reaction.
        Unproven use of stabilizer reagents.
        Health and safety issues regarding release of volatiles, steam, strong oxidant solutions.
    'These limitations lead to a decline in process efficiency.
indigo colorimetric method  (Method No. 4500-O3  B   10%, respectively. Compression of O3 gas is required to
(U.S. EPA, 1989), APHA, AWWA, WEF, 1989) or "the   inject the oxidant under pressure. Under this condition,
indigo-based HACH Ozone Accuvac Mid-range Test Kit   hydraulic seals and other  materials used in the remedial
(HACH Co., Loveland, CO). The solubility of O3 is rel-   equipment must be  compatible to  withstand  oxidant
atively low and is functionally dependent on temperature   deterioration.  Teflon, Viton, and 316 stainless steel have
and the partial pressure of O3 in the gas phase. At 1.5%   been used for this purpose (Jensen etal., 1999).
O3 by weight in air, the solubility of O3 (pH 7) at 5 °C,
10 °C, 15 °C, and 20°C, is  11.1, 9.8, 8.4, and 6.4 mg/L,   ll.C.S.b. In-Situ Application
respectively.  Decomposition  is much more rapid in the
aqueous phase than in the gas phase due to  the strong   In-situ O3 oxidation  involves the injection of a mixture
catalyzing  reaction  by the hydroxide  ion (OH-). For   of air and O} gas directly into the unsaturated and/or sat-
example, the typical half-life of gaseous O3 and aqueous   urated zones. Air sparging (Ahlfeld et al,  1994; Hein et
O3 (pH 7) at 20 °C is three days and 20 minutes, respec-   */.,  1997; Johnson, 1998; Brooks et at., 1999) is a tech-
tively. These values are based on thermal decomposition   nology that  has  been rigorously investigated and shares
only, and no wall effects, humidity, organic  loading, or   many similarities with O5 sparging and provides insight
other catalytic effects are considered.   Decomposition   to mass transport and mass transfer mechanisms with in-
increases with increasing temperature and is catalyzed by   situ O3 sparging, which has not been rigorously investi-
several substances including solid alkalis,  metals,  metal   gated in subsurface systems. Injection of air beneath the
oxides, carbon, and moisture in the gas phase.  Depending   water table promotes volatilization, supplies oxygen for
on the reactivity and concentration of reactants, tempera-   aerobic  degradation, and may induce ground-water mix-
ture,  and pH, the persistence of O3 in the environment   ing  (Johnson, 1998).  In addition to these benefits of air
and the extent of contaminant oxidation will vaiy signifi-   sparging, oxidative transformations also occur during O3
cantly.  The instability of O3 requires that it be generated   sparging. Soil vapor extraction is commonly used to cap-
on  site.  This  is accomplished  using  a simple process   ture volatile  emissions in the unsaturated zone  during air
where electrical generators produce O3 from O2(g) pres-   sparging and should also  be an important consideration
ent in the air. Air, dty air, or O2 is drawn into an ozone   and design component in  in-situ O3 sparging.  Air sparg-
generator and the air is charged with high voltage or UV   ing, in general, does not result in a uniform distribution
irradiation where  O2 molecules  split into oxygen atoms   of air bubbles extending radially from the injection well.
that react quickly to form O3. Air and pure O2 can  be   Rather,  air sparging results in the formation of a limited
used to produce O3 concentrations of about 1% and 4 to   number of air channels  in which the  majority of the
 In-Situ Chemical
Engineering Issue

-------
injected air is transported. In an ideal system, as the air
moves upward due to buoyancy and outward due  to
applied pressure, the air channels form a V-shaped net-
work of interconnected air channels (Elder and Benson,
1999, and references therein).  Most sites are character-
ized as nonideal systems where air channels are heteroge-
neously distributed, difficult to characterize and predict
(Ahlfeld etaL, 1994; Hein etaL, 1997), and allow the air
to bypass a significant cross-section of die aquifer into
which it is injected.  This conceptual model is illustrated
for an O3 sparging system in Figure 1. During air sparg-
ing,  air bubbles form in coarse-grained size porous media
and  air  channels  form in  fine-grained  size porous media
(Brooks etaL, 1999; Elder and Benson, 1999).  Since the
majority of remedial sites are composed of media smaller
than coarse sand, air channels should prevail (Brooks et
al., 1999).  Buoyancy forces on the bubble introduce a
vertical transport component which  restricts the lateral
transport of bubbles. Coalescing of small bubbles forms
larger  bubbles  and eventually a continuum of gas  (air
channel) in the saturated media.  O3 sparging in the sub-
surface is analogous to air sparging,  and therefore, it is
reasonable to assume that die transport and distribution
processes are similar.

It  is generally assumed that  mass transfer of volatile
organics from the aqueous phase to the gas phase occurs
by diffusion very near the air channels  at a rate  that is
                                                Pressure
                                                 Gauge
                                            Regulator Flowrneter
        Air      te^J    Ozone
     Compressor         Generator
               Oxygen
              Generator
                           Vapor   Vacuum
                         Treatment   Pump
                           Unit
         Ground-Water Flow
 (After Reddy et al., 1995)
                                              O3/Air
                                             Injection
Figure 1. General conceptual model of in-situ ozonation in the saturated zone with soil vacuum extraction to capture volatile
         emissions and O3(g).On-site O3 generation and injection into the ground water results in oxidation of ground-water
         contaminants and other reduced chemical species. O3/air sparging results in the formation of O3/air channels which
         contact a small cross-section of the aquifer. Close spacing of injection wells is required to accomplish  a high density of
         air channels for adequate distribution of the oxidant.
       Engineering Issue
                      In-Situ Chemical Oxidation

-------
rapid in comparison to removal in water-saturated regions
around the channels (Johnson, 1998).  It is reasonable to
assume that  contaminant oxidation occurs by similar
mass transfer mechanisms, (1) the diffusion and volatil-
ization of contaminants into the air/O3 channels where
gas-phase oxidation reactions occur, and (2) the diffusion
of O3 into the aqueous phase where contaminant oxida-
tion reactions occur.  In one  air sparging  study, mass
transfer was restricted to a zone veiy near the air channel.
The results indicated that for remediation to be success-
ful, air channels during sparging must be as close as pos-
sible where mass transfer zones overlap each other.  The
concentration of  VOCs just outside  the mass transfer
zone remained fairly constant  (Braida and Ong, 2001).
Low O3 content in the injected air and the abundance of
non-target reactants also  contribute to process ineffi-
ciency.  Air/O3 channel density is related to the rate of
remediation;  the  greater  the  density  of air  channels
achieved during in-situ O3 sparging, the greater the mass
transfer and rates  of reactions will occur between O3 and
contaminants in the aqueous and gas phases.

The radius of influence  in the context of air sparging is
ambiguous because air  channels are not uniformly  or
radially symmetric about  a  sparge  injection  point.
Further, heterogeneously distributed air channels leave
large volumes of water in  between  the air  channels
untouched by the  air stripping (volatilization) mass trans-
fer mechanisms (Ahlfeld et ai, 1994).  Again, drawing
similarities between air sparging and O3 sparging, it is
reasonable to assume that treatment is not uniform  (i.e.,
same rate  of remediation)  between wells where sparging
(air channels) is observed. Few O3 sparging cases actually
report  monitoring data  for dissolved O3  in the ground
water or for sparging activity observed in wells. Therefore,
the loosely defined radius of influence for in-situ O3 oxi-
dation is not well-documented.

Due to the  low dissolved concentrations  of O3 in  the
ground water  and poor transport of O3 bubbles, long-
term delivery of O3 into the saturated zone is required for
sufficient O3 mass delivery. The concentration of O3 in
the ground water can be used to assess the radius of influ-
ence of injected  O3. At the time of this publication, no
case study or examples were available that demonstrated
the radius of influence or the transport distance of dis-
solved O3 or O3  microbubbles in ground water.

The transport  of O3  gas in unsaturated porous media is
impacted by various parameters. The water content, soil
organic matter, and metal oxides were found to be  the
factors most influential in the fate and transport of gas-
eous O3 in unsaturated porous media (Choi etal., 2002).
The higher  the water content,  the  faster  the  break-
through.  This was attributed  to less contact with  the
metal oxide and  organic matter reactants associated with
the solid phase material.   Nevertheless, O3  was  readily
delivered and  transported through  unsaturated  porous
media where  phenanthrene and diesel range organics
(C10 to C24/
were oxidized.
ll.C.S.c. Ozone Demand

The O3 demand was measured in the laboratory for four
different soil materials (Table 8). In this study, it was
shown  that increasing  the  water content of  the soil
material resulted in greater O3 demand due to dissolution
into water (probably due to the strong catalyzing reaction
by  OH') and subsequently,  self-decomposition.  The
water content  under field conditions varies considerably,
Table 8. O3 Demand and Energy Costs to Meet the Demand of Uncontaminated Geological Material (Masten and Davies,
1997)
Geological Material
Ottawa sand
Wurtsmith AFB,Oscoda,MI
Metea subsoil, E.Lansing, Ml
Borden sand.Borden AFB,Ontario
03 Demand
(mg 03/g soil)
<0.04
0.022 to 0.21 5 3
1.4
2.0
Energy Cost/Ton
kWh1 $U.S.2
<0.22
<4.3
31
44
<0.013
<0.26
1.85
2.64
1 Based on energy cost for O3 generation of 10 kWh/lb
2 Based on a cost for electricity of $0.06 per kWh
3 0.022 mg O3 /g at 3.2% moisture content, 0.21 5 mg O3 /g at 6.8% moisture content
 In-Situ Chemical
                        Engineering Issue

-------
especially  near the  water table,  which will strongly
influence O3 transport, O3 demand, and costs (refer to
Section 111. A. 3. Oxidant Demand).

ll.C.S.d. Contaminant Transformations

Environmental contaminants can be oxidized either by
direct reaction with O3, or indirectly via O3 decomposition
and formation of the hydroxyl radical (-OH), a stronger
oxidant (Hoigne and Bader, 1976; 1979a; 1979b) (Table
9).  O3 reacts  rapidly with  electron-rich  olefins  and
aromatic compounds.  Increasing  chlorine  substitution
will decrease the rate constant of O3 addition to olefins,
and in the case of TCE and PCE,  the rate constants are
already so low  that at short reaction times common in
treatment processes  there  is  very little   destruction
(Dowideit and von Sonntag, 1998).  PCE reacts so slowly
that it cannot be oxidized  by a direct O3  reaction within a
day,  and  TCE will  react only  during extended  O3
treatment (Hoigne and Bader, 1979a). In contrast, DCE
and vinyl chloride react quickly (Clancy etal., 1996)  due
to the free  C=C double bond. The  reaction  rate of
benzene is low, requiring hours for its oxidation even at
high O3 concentration. However, the rate increases with
increased substitution of functional groups that  elevate
the  electron   density  of  the  ring   (e.g.,  phenols,
chlorophenols). Aliphatic alcohols, aldehydes, and organic
acids generally react so slowly that the reaction rates are of
little interest. However, formic acid, in the ionic form as
formate ion, can  be oxidized  rapidly. NH3  and amines
show an appreciable  reaction rate when  in the non-
protonated form. During ozonation, only those functional
groups of contaminants  which are especially  reactive
towards an electrophylic reactant (i.e., O3), (non-halogen-
substituted olefinic  compounds, phenols and phenolate
ions,   PAHs,   non-protonated  amino  groups,  thio
compounds, etc.)  can be  easily oxidized directly by  O3.
Only  the more  reactive -OH  may attack molecules
containing less reactive functional groups, such as aliphatic
hydrocarbons, carboxylic  acids, benzene, chlorobenzene,
nitrobenzene,   perchloroethylene   or trichloroethylene
(Hoigne and Bader,  1979b).  Ozone has been used for
 Table 9.  General Ozone Oxidation and Related Chemical
         Reactions
 Direct Oxidation
03+ C2HCI3
            H20
2 C02 + 3 H+ + 3 Cl-
 •OH Formation
 03+H20	>
 203 + 3H202
             02 + 2 -OH  (Slow)
                                 (Fast)
(D


(2)
(3)
PAHs (Masten and Davies, 1997; Cambridge and Jensen,
1999; Wheeler^/., 2002), BTEX(Black, 2001), MTBE
(Black, 2001), and chlorinated compounds such as PCE,
TCE, and DCE (Masten and Davies, 1997). Pyrene and
phenanthrene degradation was greater than 90% in one
hour in a loamy sand soil, while degradation of 100 mg/
kg chrysene was 50% in four hours (Masten and Davies,
1997).

The addition of H,O2 to O3 in water generates -OH,
thereby increasing the oxidative capabilities of the treat-
ment system  (Table 9).  Increased  rates of contaminant
oxidation  have  been reported for MTBE (Mitani et al.,
2002) and TCE and PCE (Glaze and Kang, 1988; Clancy
et al., 1996) when O3 is combined with H7O2.  At the
time of this publication, no information  was obtained
where O3 and H2O2 were co-injected into the subsurface
in an ISCO treatment system.

ll.C.S.e. Other Considerations

In-situ ozonation  may involve feasibility testing (bench-
scale testing)  to assess whether the target  contaminants
can be oxidized under site-specific  conditions  using rea-
sonable quantities of O3, without deleterious side effects,
such as metals mobilization or unacceptable reaction
byproducts.   Determination  of in-situ ozone  design
parameters can be determined through pilot-scale testing.
For example, O3  distribution can be measured to assess
whether a sufficient quantity of O3  can be produced and
delivered throughout the targeted zone.  Additionally, a
decline in contaminant concentration and an increase in
reaction byproducts (i.e., CVOCs and  Cl") can be mea-
sured to assure that  the treatment objectives can  be
achieved and volatilization is not a significant loss mecha-
nism. Fugitive  O3 emissions during production or injec-
tion may represent unacceptable risks to human health
and to the environment. Where SVE is needed  to capture
the off-gas  from O3 injection, a nickel catalyst is used to
decompose O3.  Engineering and safety controls are, there-
fore, required to prevent unacceptable exposure pathways.
In addition, high O2 content in confined spaces may rep-
resent  unacceptable health  and safety  conditions and
should also be  monitored and managed.  For example,
monitoring air  quality across the site should assure that
the O^ concentrations meet OSHA requirements.  The
delivery of O3(g) into the subsurface may displace volatile
organics from the injection zone. Consequently, control
of fugitive volatile emissions may be necessary if unaccept-
able exposure pathways  are predicted or determined. A
vapor extraction system can be used to enhance the radius
of influence of the O3 and to capture volatile organics and
unreacted O3 (Jensen etal., 1999).
       Engineering Issue
                                                                                 Chemical Oxidation

-------
II.C.3.f. Advantages

O 5 reacts with many, but not all important environmen-
tal contaminants. Advantages of in-situ ozonation in the
unsaturated zone relative to the saturated zone include:
higher concentrations of O3 can be injected, O3 is more
stable in gas than in water, diffusive transport is greater,
and higher velocities (mass delivery rates) can be achieved.
Co-injection and reaction of H2O2 and O3 can yield -OH,
a strong, nonspecific oxidant.  However, no information
was available regarding the demonstration and documen-
tation of this co-injection process.

II.C.3.g. Disadvantages

O 5 has a short retention time in the subsurface because it
reacts rapidly with  a wide range of naturally occurring
non-target chemical species (reduced  minerals, organic
matter, etc.), including the hydroxide ion (OH"). O3 has
a relatively low solubility in water  and  is highly vulnera-
ble to  hydraulic short  circuiting as a gas in the unsatu-
rated zone. Based on similarities between air sparging and
O3 sparging, it is reasonable to conclude that O3 bubbles
injected into the saturated zone are poorly/nonuniformly
distributed and are transported  very  short  distances.
Transport and distribution of O3(g) in the saturated zone
is most likely restricted to veiy short distances from the
gas channels (i.e., O3(g), O2(g)) that form in the subsur-
face. Consequently,  O3 mass transport and mass transfer
limitations are likely to be significant. On-site generation
and compression of the corrosive  O3 gas is required to
inject under pressure in the saturated zone and results in
the rapid deterioration of remediation piping and plumb-
ing  materials  if  incompatible  materials  are  used.
Specialized oxidant-resistant materials  are likely to be
required.  Enhanced volatilization  of contaminants may
result from sparging the ground water with  O3(g) and
O2(g).  Since volatile organics and O3 both represent a
threat  to human health, collection of volatile emissions
(off-gases)  using  a  vacuum extraction system may be
required to minimize potential exposure pathways.  O3
does not react at an appreciable rate with some important
environmental contaminants.

II.C.4. In-Situ Persulfate Oxidation

II.C.4.a. Physical and Chemical Characteristics and
        Chemical Reactions

Persulfate is the newest form of oxidant currently  being
used for ISCO.  Persulfate salts dissociate in aqueous solu-
tions to form the persulfate anion  (S2O82"). S2O82" is a
strong oxidant and can degrade many environmental con-
taminants, or it can be catalyzed with various reactants to
form the sulfate radical (-SO^"), a more powerful oxidant.
Catalysis of S2O82~ to -SO^ can be achieved at elevated
temperatures (35 to 40 °C), with ferrous iron (Fe(II)), by
photo (UV) activation (Table 10, rxns 1 to 3), with base
(i.e., elevated pH),  or with H2O2. In addition to Fe, other
general activators include the ions of copper, silver, man-
ganese, cerium, and cobalt (Liang et al., 2004a, and refer-
ences therein). Persulfate-driven oxidation by -SO4~ has a
greater oxidation potential  (2.6 V) than S2O82"  (2.1 V)
(Table 1) and can degrade a wider range of environmental
contaminants at faster rates. Formation of-SO4" may ini-
tiate the formation of-OH  (rxn 4, Table 10) and a series
of radical propagation and termination  chain reactions
where organic compounds can be transformed (Huang et
al., 2002, and references therein).
Table
s2 2-S04-
I- Fe+2 	 *• Fe+3 + -S04- + S042-
hv
	 >• 2 -S04-
H20 	 ^-OH + HSCV
Fe+2 	 >> Fe+3 + -S04- + S042-

(D
(2)
(3)
(4)
(5)
The solubility  of potassium persulfate is  too low  for
environmental applications, and the reaction of ammonium
persulfate will result in an ammonia residual, an undesirable
reaction product. Therefore, sodium persulfate (Na2S2O8)
is the most common and feasible form used  in ISCO.
Sodium persulfate costs approximately $1.20/lb  (Brown
and Robinson, 2004).  The solubility of Na9S2O8 is high
(73 g/100 g H,O @ 25 °C) and the density of a 20 g/L
solution (1.0104 g/mL) (FMC,  2006) at 25 °C is greater
than water.  Therefore, the density-driven transport of a
high concentration solution of Na2S2O8 would occur in
the subsurface.  Persulfate is more stable in die subsurface
as compared to H2O2 and O3 (Huang et al., 2002), and
can persist in the subsurface for weeks, suggesting that the
natural  oxidant demand  for persulfate  is  low.   The
persulfate anion (S2O82~) is not significantly involved in
sorption  reactions. These characteristics make persulfate
an attractive oxidant  because it persists in the subsurface,
can be injected at high  concentrations, can be transported
in  porous media,  and will undergo density-driven and
diffusive transport into low-permeability materials.
 In-Situ Chemical
                         Engineering Issue

-------
Co-injection of persulfate and Fe+2 could be performed to
accomplish the catalysis of S2O82~ to -SO^.  The transport
of Fe+2 in the subsurface can be problematic, as discussed
previously (Section II.C.2.f, Injected Reagents). Oxidation
of Fe+2 to Fe+3 by either S2O82" or -SO4" could limit the
effectiveness of either the injected catalyst or the oxidant.
In one persulfate field study, naturally occurring ferrous
iron was  used to catalyze  the  persulfate anion while
maintaining slightly reduced conditions and soluble Fe+2
(Sperry et al.,  2002).  The scavenging reaction between
Fe+2 and -SO4" (rxn  5, Table 10)  and other  non-target
reducible  reactants  represents a potential  sink  for the
sulfate radical.  A  balance  must be  achieved  between
adding sufficient Fe to accomplish -SC^" production and
excessive Fe which may result in -804" scavenging (Liang
etal., 2004a).  Scavenging of -SC^" and -OH and a decline
in the persulfate  oxidation rate of MTBE were attributed
to  naturally  occurring  carbonate and bicarbonate in
ground  water (Huang  et  al,  2002). The persulfate
oxidation rate of MTBE decreased with increasing pH
and increasing  ionic strength (Huang  et  al.,  2002).
However, the  decline in the oxidation rate from pH 2.5
to near neutral was only  30%, indicating that persulfate
oxidation is pH-dependent, but only moderately sensitive
to this parameter.

II.C.4.b. Contaminant Transformations

Several environmental contaminants have  been oxidized
in laboratory  experiments using persulfate and  various
catalysts.   Fe+2-assisted  sodium  persulfate oxidation of
TCE (60 mg/L) removed 47% of the TCE at a persulfate:
iron:TCE  molar ratio  of 20:5:1 (Sperry et al, 2002;
Liang et al, 2004a).  TCE oxidation also  occurs  via
chelated  Fe+2-assisted treatments  (Liang et al, 2004b).
TCE was significantly oxidized at 40 to 60 °C, and TCA
at 60 °C, within several hours (Liang et al., 2001; Liang
et al., 2003).  In a field test involving a glauconitic (iron
rich,  3   to  15  mg/L  Fe+2) sandy  aquifer,  persulfate
catalyzed by naturally occurring ferrous iron oxidized 30
to  50%)  of a  TCE and cis-l,2-DCE concentration of
about 7 to 9 mg/L and vinyl chloride (Sperry etal., 2002).
Heat-assisted sodium persulfate (at about 8 g/L) oxidized
MTBE in the  ppm range (half-life <1  hour; 40 °C)  in a
buffered  laboratory  solution. A  much slower  MTBE
degradation rate was observed in ground-water samples,
perhaps due to  radical scavenging by bicarbonate ions
(Huang etal., 2002).  Persulfate (0.0357 mg/L; persulfate:
organic matter ratio of 12 g/g; 70 °C) oxidation of PAHs
(<200 mg/kg;  16 EPA PAHs) at bench-scale tests with a
three-hour reaction time varied widely in 14 different soils
and sediments. Loss of the 16 PAHs varied between 0 to
80%; 0 to 85% loss of 2- and 3-ring PAHs; 0 to 75% loss
of 4-ring PAHs; and 0 to 70% of 5- and 6-ring PAHs
(Cuypers etal., 2000).  The reaction  between Na2S2O8
and  66 organic compounds  (and isomers)  in aqueous
solution at various temperatures (room temperature, 20
°C, 35 °C, 40 °C, or 45 °C), persulfate concentrations (1,
5,11 g/L), presence of Fe+2 or catalyst, and time of contact
(3, 14, 20, 21, 90 days) has been evaluated (FMC, 2005).

Overall, heat-assisted persulfate oxidation is  rapid,  and
raising the temperature of aquifer material and ground
water  is technically feasible;  however,  the  economic
feasibility has not been established. Methods used to raise
the temperature  in subsurface  systems  include radio
frequency heating, steam  injection, six phase, electrical
resistance, etc.,  but have not been demonstrated at field-
scale in conjunction with persulfate oxidation.

Liang etal.  (2001) hypothesized that sodium persulfate
(being a fairly strong oxidant  at ambient temperatures)
could have  an  important role in the oxidation of soil
organic carbon.  If so, one potential use of persulfate in
in-situ remediation would be for the oxidation of soil
organic matter,  prior to  use  of a different  oxidant  for
oxidation  of  contaminants.   The destruction  of  soil
organic matter is  important  since it will decrease  the
natural  oxidant demand  (NOD) of  the soil,  allowing
subsequent oxidant additions to be used more efficiently
for target contaminants. However, Brown and Robinson
(2004)  questioned the effectiveness  of persulfate  for
decreasing  soil  oxidant  demand, stating  that it  was
relatively unreactive toward naturally occurring organic
matter.

II.C.4.C. Advantages

Persulfate is more stable in the subsurface than H2O2 and
O3, and the radical intermediate, -SO4~, is more stable
than -OH.  This suggests fewer mass transfer and mass
transport limitations. Persulfate will react with benzene,
while permanganate does not, thus allowing this form of
oxidant to be used in the remediation of fuel spills and
BTEX-contaminated ground water.  Persulfate does not
appear to react as readily with soil organic matter as per-
manganate (Brown and Robinson, 2004). This may not
be an advantage over permanganate  in aquifer material
where  the  oxidant demand  is  predominantly  due to
reduced mineral species.

//.C.4.of. Disadvantages

In-situ  chemical oxidation involving persulfate is  an
emerging technology and, in  general, the peer-reviewed
literature is  limited, and there are few reports of bench-
       Engineering Issue
                     fri-Slu Chemical Oxidation

-------
and field-scale studies.  The lack of information pertain-
ing to the fundamental chemistry and applications  in
subsurface systems suggests there is also a limited infra-
structure of knowledge and experience upon which  to
design successful remediation systems.  This limitation/
disadvantage will diminish with time based on ongoing
fundamental and applied research.

Persulfate is less stable than permanganate and will not
persist as long in subsurface systems. Catalysts are required
in the persulfate reaction to produce the more powerful
sulfate radical.  There will likely be difficulties in achiev-
ing the optimal mix of reagents (i.e., Na2S2O8, catalysts)
in the subsurface due to the lack of naturally occurring
catalyst, and due to the difference in transport behavior of
these  reagents  upon  injection.  -SC^2"  scavenging  is a
source of process inefficiency that is currently not  well
understood  nor documented.  Na2S2O8 costs  approxi-
mately $2.70/kg, which is more than KMnO4 and H2O2.
This cost of oxidant may be offset by the lack of oxidant
demand by non-target aquifer materials.
III. TECHNOLOGY DESCRIPTION AND
   TECHNOLOGY SELECTION FACTORS

III.A. Bench-Scale Studies

Bench-scale treatability studies  can  be  useful to gain
insight on the feasibility of contaminant  oxidation prior
to field-scale applications. In complex, heterogeneous sys-
tems it is difficult to predict specific reactions, oxidation
efficiency, oxidation  byproducts, or whether any of the
potential limitations  apply.  The methods and materials
of bench-scale treatability studies may vaiy based on the
oxidant used and the objectives. It is important to recog-
nize  the physical differences  between bench- and field-
scale systems.  The use  of bench-scale treatability results
from simplified systems to design field-scale ISCO sys-
tems must be heavily scrutinized.

III.A. 1. Objectives

One objective of a  bench-scale treatability study is to
establish proof of concept that the target  compound can
be transformed by oxidative treatment(s) given the poten-
tial limitations. Another objective in  MnO4" bench-scale
studies is to measure the oxidant demand.  This data and
information is used to assess the feasibility of ISCO and
to assist in the design of oxidant injection  at pilot- or
field-scale  (refer  to  Section III.E.2.a.   Permanganate
Oxidation).  For all oxidants, significant contaminant
reduction should be demonstrated using reasonable quan-
tities of oxidant and reagents under reasonable simulated
conditions. For example, with Fenton oxidation, the pH
of the test reactors should not be conducted at the opti-
mal pH (pH 3.5 to 4) unless the acidic condition can be
accomplished at field-scale.  Otherwise, the results may
overestimate treatment effectiveness.  Assessment of the
reaction byproducts may also be an important objective
since the oxidative treatment of some target compounds
or complex chemical  mixtures may be  poorly docu-
mented or unknown.  Side and competing reactions may
yield undesirable byproducts.  Historically, mobilization
of redox- and pH-sensitive metals has been a concern and
may be an  important objective in  the bench-scale treat-
ability study.  Although attenuation of metals generally
occurs over short  transport distances, this may be an
important issue where high metals concentrations exist
(i.e., naturally occurring or co-disposed  with organics)
and/or where potential receptors are nearby and could be
impacted over short transport distances.

Under some conditions,  bench-scale treatability studies
may not be feasible. At one  site underlain by a coarse
alluvial  aquifer,  it  was difficult and  costly to obtain
unconsolidated aquifer media  using direct push or hol-
low stem auger. Air rotary drilling was  used to install
wells at the site but  this method of acquiring aquifer
material would significantly impact (pulverize) the physi-
cal integrity of the  samples (i.e., the surface area,  reactive
species, redox, and  contaminant concentrations would be
altered significantly).  Consequently, the MnO4~  oxidant
demand would be highly variable and would likely yield
erroneous results that would confuse the feasibility assess-
ment. A small field-scale pilot  treatability study was con-
ducted which was designed to limit cost and to assess the
feasibility of ISCO.

III.A.2. General Guidelines

Components of the  bench-scale reactor should  include
the aquifer material since  it will contain the majority of
the contaminant(s) and other parameters that will largely
influence oxidant demand and the success or failure of the
treatment process. Disturbed aquifer material is generally
used in this procedure.  Use of ground water from the site
is  ideal  but  is generally not critical.   Capture  and
quantification  of contaminant losses from the reactor is
necessary to  maintain a  mass  balance  and to assess
treatment performance. These losses  include volatiles,
displacement of aquifer material, aqueous solutions, or
DNAPLs.  Reactions  involving  H2O2   may  release
significant  quantities  of  heat  and  O2(g) and  enhance
volatilization. Volatile losses can be captured and quantified
using inert  gas  bags or  an  activated carbon  trap. A
 In-Situ Chemical
                         Engineering Issue

-------
nonvolatile contaminant analog can also be amended to
the reaction vessel and its loss can be used to predict the
oxidative transformation of the target compound (H tiling
etal., 2000). Failure to capture volatile losses could result
in an overestimate of oxidative treatment.

Recommended  monitoring parameters that are a direct
indicator of oxidative treatment include the target com-
pounds, reaction byproducts, metals, and the oxidant  (i.e.,
H2O2, MnO4",  S9O82", O3).  Indirect indicators such as
CO2,  dissolved oxygen  (DO),  total  organic  carbon
(TOC), chemical oxidant demand (COD),  and tempera-
ture have been  used but are generally unreliable and not
recommended. Control reactors can be used  to help quan-
tify non-oxidation losses and are recommended. Measuring
pre- and post-oxidation concentrations of the target com-
pound in the aqueous, solid,  and gaseous phases allows
mass balance calculations which serve as the basis for per-
formance evaluation.

III.A.3. Oxidant Demand

"Natural oxidant demand" generally refers to the demand
attributed to naturally occurring materials  (i.e., reduced
inorganic species + organic matter).  The  total oxidant
demand includes both the natural  oxidant  demand and
the demand due to anthropogenic contaminants. In this
document, the term oxidant demand is synonymous with
total oxidant demand.

The oxidant demand  for  H2O2 is not measured since
H2O2 decomposition involves catalytic  reactions (Table
4, rxns 1 to 2) and the oxidant demand would be infinite.
The oxidant demand in O3 systems may also be difficult
to quantify due to reactions between O3 and H2O  and
OH". Under some conditions, the consumption of O3 by
these reactants  may be relatively small.  Four different
geologic materials excited a limited O3 demand (Table 8),
i.e., the rate of O3 degradation in soil columns was slow
after the  immediate O3  demand was met  (Masten  and
Davies,  1997).  Bench-scale measurements of the  O3
demand may successfully separate the high O3 demand
associated with  reducible  mineral species from a longer-
term demand of lower magnitude attributed to OH" and
H2O. However, the long-term O^ demand  under in-situ
conditions has not been well-documented and may be sig-
nificant.

The permanganate oxidant demand is often measured at
bench-scale in batch reactors prior to field-scale applica-
tions and many variations  of the  test procedures have
been reported.  Testing procedures also include column
tests (Drescher  et al., 1998; Mumford et al., 2004)  and
field push-pull tests (Mumford etal., 2004).  A standard-
ized method to  measure the oxidant demand  is under
review by the American Society for Testing and Materials
(ASTM) (Vella etal., 2005) and can be reviewed at an
EPA website http://www.epa.gov/ada/topics/oxidation
issue.html. The  oxidant demand results can be used to
assess the preliminary feasibility of in-situ permanganate
oxidation. The test involves two tiers of testing.  Tier 1 is
an inexpensive, rapid (minimum 48-hour test) prelimi-
nary screening test used to estimate the  oxidant demand
of aquifer materials  and  contaminants.  Tier  1 testing
involves a minimum  of 250 grams of soil and 250 mL of
site ground water per  soil sample,  a minimum of 50
grams  of soil per reactor vessel,  3  KMnO4  loading rates
(5, 15, and 30 g KMnO/(/kg soil), a soihwater ratio of 1:2
(wt.:wt. wet basis), and a minimum of 48 hours reaction
time. Guidelines for soil  sample handling  (preparation,
mixing, compositing, drying, storage), testing procedures
(reactor vessel, MnO4" measurements, mixing, etc.), and
oxidant demand  calculations are provided.

Oxidant demand results from laboratory studies are gen-
erally used in conjunction with other site-specific  treat-
ability results to assess the preliminary feasibility of ISCO
including projecting the cost for oxidant, assess technical
and economic feasibility, compare  ISCO with other can-
didate technologies, assess implementability, etc.  Oxidant
demand data may also  be used to  assist in the design of
pilot- or full-scale injection systems.  However, there  is
not a uniform approach or method used for either of these
purposes (refer to Section III.E.2. General Conceptual
Approach to ISCO).

Tier 2 of the proposed  ASTM method  is more complex
and, in theoiy,  could  provide  comprehensive  site and
testing data, including extensive sampling methodology,
longer test duration,  increased reaction  data points, and
reaction kinetics. Tier 2 is currently under  development
and  has not yet been submitted to ASTM  for review.
Other  guidelines for performing oxidant demand tests
are also available (Haselow et al., 2003), and continued
development is needed.

Long-term oxidant demand testing reveals  that the oxi-
dant demand extends much longer than  48 hours, result-
ing in higher oxidant demand values than measured over
a shorter time periods (Figure 2).  The long-term persis-
tence of MnO4"  (»2 days) in aquifer material has been
well-documented suggesting  that long-term  testing may
be more representative of testing conditions.  This may be
important in cases where contaminants persist  for long
periods. For example, contaminants present as NAPLs or
contaminants in  low-permeability materials may have
       Engineering Issue
                            Chemical Oxidation

-------
                                                                                                   160
                                            20      25
                                                Time (day)
                          30
                  35
40
45
50
           1B34'bgs
           3D 32.51 bgs
2A33.1'bgs
4B 9' bgs
     3C 10.5'bgs             —B—  3D 10'bgs
-*• -  3C (high dosage) 10.5'bgs  --O- -  4B (high dosage) 9'I
Figure 2. The total oxidant demand (TOD) was measured by amending aquifer samples (30 to 50 g dry weight) with KMnO4
        solutions (0.5 L;2.1 g/L (low dosage) or 10.5 g/L (high dosage)) (Huling and Pivetz, 2003). The aquifer samples were
        collected in an area of approximately 50x50 ft and between 9 to 34ft bgs. The data represent the average of two split
        samples at each depth and location. TOD values increase with time (28 to 44 days) and a significant difference in TOD
        occurs between short-term (two days) and long-term testing periods. Although replicate samples were collected
        from adjacent sections of the core (a few inches apart), variability in TOD (> 100%) was measured in replicates at both
        locations 2A and 4B. Aquifer core samples 3C and 3D (shallow) were collected 10 ft apart, from the same vertical interval.
        However, the average oxidant demand was different by nearly a factor of 5  between locations. Further, greater oxidant
        demand was measured at the high oxidant dosage (4B—151 g/kg vs.43g/kg;3C—14g/kg vs. 5g/kg). TOD values
        resulting from testing at high dosages for samples 3C and 4B are indicated  by dashed lines and are read using the right-
        hand y axis.
limited contact with the oxidant.  Under  these mass
transfer and/or mass transport-limited conditions, long-
term persistence of the oxidant may be required to achieve
the treatment objective.  Conversely, the oxidant demand
measured in long-term tests may be an overestimate when
mass  transfer/transport  is not limited and oxidation is
rapid.  Long-term bench-scale tests may be used to quan-
tify the upper value of the oxidant demand. Alternatively,
results from a  short-term test (such as ASTM, Tier 1)
must be interpreted in a manner that considers the effects
of long-term contact between oxidant and aquifer mate-
rial.  Continued development  of standardized  oxidant
demand testing is needed.
                   The permanganate oxidant demand measured in bench-
                   scale tests can be impacted by various parameters. Spatial
                   variability (depth,  location) in the composition of the
                   aquifer material may affect the oxidant demand (Figure
                   2). Collection of aquifer samples in sufficient number at
                   various depths and locations may be needed to represent
                   the variability in oxidant demand.  This information can
                   be used to establish  a  correlation  between the oxidant
                   demand and  different  lithologic or geochemical zones.
                   Since the composition of organic matter varies in sands,
                   silts, and clays, variability in the oxidant demand will also
                   vaiy. Soil with 1.6% TOC prior to oxidation had 0.6%
                   TOC remaining after oxidation by KMnC^ in a column
 In-Situ Chemical Oxidation

                                            Engineering Issue

-------
study,  indicating that organic matter will exert an  oxi-
dant demand (Drescher et aL, 1998).  Naturally occur-
ring organic matter exhibits various chemical compositions
and therefore, fractions of organic matter may range from
easily oxidizable to recalcitrant.  Reduced chemical  spe-
cies (i.e., sulfides, ferrous Fe, manganous Mn, etc.) in
subsurface media can be oxidized during I SCO and con-
sume oxidant.  Geochemical environments rich in organic
matter and/or reduced mineral species may exeit a signifi-
cant oxidant demand.  Bench-testing can be used to assess
whether excessive natural oxidant demand will be exerted.
The type (Table 3, rxns 4 to 7) and phase (aqueous, sorbed,
NAPL) of organic contaminants will also affect the  oxi-
dant demand.  Exposure of the aquifer material to air and
loss of volatile organics during handling will artificially
lower  the measured oxidant demand.   Such exposure
should be minimized since this could result in an under-
estimate of the actual oxidant demand.

The oxidant demand is functionally dependent on the
concentration  of MnO4~  used   in the  test  (Figure  2).
Under comparable  conditions,  the  oxidant  demand
increases  with  an increase in the oxidant concentration
(Siegrist et aL, 2002).  An  oxidant  demand  of 2.8 g
KMnO^/kg was measured when amended with a 500  mg/
L KMnO4  solution, and increased to 10.8 g KMnO4/kg
when amended with 5000 mg/L KMnC^ (Struse et aL,
2002b).  The concentration  of oxidant most representa-
tive of actual field conditions should be used in the bench
test to obtain the  most  accurate results. The oxidant
demand is also dependent on  (1) the contact time between
the oxidant and  soil, (2) mixing, and (3) the solids:solu-
tion ratio.  The oxidant demand decreased as the solids:
solution ratio increased (Mumford etal.,2004). However,
given sufficient time, the oxidant demand converged  to a
similar value regardless of the mass of aquifer material
used.   The maximum NOD value depends  only on the
mass of oxidizable matter and the stoichiometry between
that oxidizable matter and the oxidant,  and that it does
not depend on the mass of aquifer material (Mumford et
aL, 2004).  In summary,  for short reaction  periods, the
solids:solution  ratio is likely to have a significant effect on
the results;  for longer reaction periods, the solids:solution
ratio is likely to be less important.

The objectives and guidelines presented for permanganate
bench-scale testing are also applicable  to persulfate.  For
example, in soil slurry laboratory experiments, the soil with
the lowest fraction of organic carbon (foc) had the highest
TCE and TCA oxidation efficiency by sodium  persulfate
(Liang et aL, 2001).  In general, the persulfate oxidant
demand is lower than the permanganate oxidant demand,
suggesting that lower oxidant loading may be required.
III.B. Pilot-Scale Studies

Pilot-scale treatability studies provide useful information
to help design and plan full-scale ISCO implementation.
Specifically, due to the spatial variability of samples col-
lected and used in bench-scale tests, pilot-scale studies can
provide data and information from the oxidative treat-
ment over a larger aquifer volume.  The methods and
materials of the study may vaiy based on the oxidant used
and the objectives.

III.B.I. Objectives

The objectives may include the following: determine the
injection rate vs.  injection pressure; assess various injec-
tion strategies; assess the travel times, distribution (verti-
cal/horizontal), and persistence of the oxidant and reagents
(Fe,  acid,  stabilizers, chelators); determine  whether
ground-water contaminants are mobilized or are volatil-
ized; assess the mobilization of metals; assess contaminant
rebound; determine reaction byproducts; conduct a pre-
liminary  performance evaluation of contaminant oxida-
tion; assess  the adequacy of the  monitoring  program;
anticipate well fouling problems; and assess  the potential
difficulties  in scaling  up  a treatment  system.  Multiple
injections of oxidant and/or reagents under different con-
ditions can be used to accomplish different treatment and
testing objectives.

III.B.2. General Guidelines

A detailed assessment of ISCO performance evaluation
and rebound generally requires extended periods of time
due  to the slow mass transfer and mass transport  pro-
cesses in  conjunction with the slow rate of ground-water
movement.  Additionally, in Fenton  systems, a signifi-
cant disturbance  results  from H2O2  injection  and the
subsequent release of heat and O2(g).  It is common to
see significant increases in total dissolved solids in ground-
water  samples collected  soon  after  H2O2 injection.
Therefore, the ground-water quality  is highly disturbed
(transient)  and requires an extended  period of time to
approach chemical equilibrium. Both the  detailed and
general  information  acquired  through the  pilot-scale
study can be used to help design and plan subsequent
injection events.  Monitoring data and information are
useful to design the monitoring system for  the full-scale
system including  appropriate locations and depths of
monitoring wells and appropriate monitoring parameters
and frequency.

The following  general guidelines  for  ISCO pilot-scale
studies are applicable for all oxidants, and also apply to
       Engineering Issue
                     fri-Slu Chemical Oxidation

-------
full-scale implementation.  Conceptually, an  outside-in
injection strategy involves initial oxidant injections on the
periphery of the known contaminant zone. Subsequent
injections in the middle of the source zone may transport
contaminants into adjacent zones already containing oxi-
dant and/or contaminants. Ideally, this reduces the trans-
port  of  contaminants  from  the   source  zone  into
uncontaminated areas.  Ground-water samples represent
an integrated measure of contaminants present in the sub-
surface  and provide valuable  insight regarding perfor-
mance evaluation from the oxidative treatment. However,
due to the slow mass transfer and mass transport processes
which occur in the subsurface, sufficient time should be
allowed (after  ISCO is performed) before ground-water
samples  are  collected   for   performance   evaluation.
Assuming potential receptors are located close to the injec-
tion area, an expedient ground-water monitoring program
(rapid turn-around) may be needed.  Soil core samples
may provide immediate feedback on performance,  but
variability in contaminant concentrations may require
that numerous soil core samples be collected to minimize
uncertainty and allow for an accurate assessment of treat-
ment performance.  Aquifer samples can be highly effec-
tive in performance evaluation where contaminants (i.e.,
DNAPL)  have accumulated at distinct  lithologic units
and contaminant distribution is more easily defined.

Pilot- and full-scale ISCO should be implemented in a
manner that recognizes and minimizes the transport of
contaminated ground water or NAPLs from the source
area into low contamination/clean areas. Pilot-scale stud-
ies are sometimes deployed in or downgradient from a
source zone. In this case, it can sometimes be difficult to
distinguish between rebound and upgradient flushing of
contaminants into the study area.  An outside-in approach
can be used to help minimize this complication  (i.e., a
wedge that extends from the upgradient edge to the cen-
tral area of a source zone). Contaminant transport from
upgradient of the remediated area and possible recontam-
ination underscores the need to design the oxidant deliv-
ery system for full coverage.

In-situ oxidant push-pull tests may  be used  to evaluate
the permanganate oxidant demand (Mumford  et al.,
2004), and presumably persulfate oxidant demand, over
relatively large aquifer volumes. This involves the injec-
tion and recovery of a solution containing an oxidant and
a conservative tracer.  Measurement of oxidant demand
is determined  from the analysis of  the  recovery break-
through curves of the oxidant and tracer.  There are some
design and operational limitations of this technique, but
potentially it could be  used with existing monitoring
wells and may provide accurate oxidant demand data.
III.C. Technology Applicability

III.C.I. Location of Oxidant Application

///. C. 1. a. Saturated Zone

ISCO  involving MnO4~, Fenton's, O3,  and S2O82~ has
predominantly  been  applied  in  saturated, unconsoli-
dated, highly contaminated (source zone)  porous media at
hazardous waste sites (refer also to Section III.F.5 Oxidant
Delivery). The injection of O 5 gas into the saturated zone
has been  used  but  is vulnerable to  nonideal transport
mechanisms including  preferential  pathways (refer to
Section II.C.3.b. In-Situ Application).   Few scientific
investigations of in-situ persulfate oxidation have  been
reported; however, numerous studies are underway and it
is anticipated that more information will  soon  become
available regarding S2O82~. Presumably, persulfate oxida-
tion will be applicable to both the saturated and unsatu-
rated zones, similar to permanganate.

Environmental  contaminants present either as a NAPL,
adsorbed onto aquifer material, or in the aqueous phase
(dissolved) are all vulnerable  to chemical  oxidation trans-
formations.  Due to competition by naturally occurring
non-target reactants in the aquifer material, the most effi-
cient use of chemical oxidation occurs where the concen-
tration  of the  target  contaminants is  highest.  The
oxidation efficiency (f|), defined as mass  of contaminant
transformed/mass oxidant reacted, is highest in source
zones.  Therefore, the most  cost effective contaminant
oxidation occurs in source zones. This is consistent with
the common remedial objective that initially targets the
source zone(s) at a  site. The downgradient contamina-
tion plume  is  often a secondary priority  and may not
involve ISCO due to the large area of contamination rel-
ative to the  source area, lower oxidation efficiency, and
greater cost.

ISCO  is not commonly applied in the  ground-water
plume extending downgradient from the source zone.
This trend is mainly attributed to the small size of source
zones  compared  to the downgradient  plume.   In the
source zone, the oxidant can be applied  at high concen-
tration, focused in specific source area locations, and can
achieve greater oxidation efficiency relative to downgra-
dient zones.  Under this set of conditions,  larger quanti-
ties of contaminant  can  be transformed  using  lower
quantities of oxidant and at lower cost.   However, treat-
ment objectives vary between sites. The treatment objec-
tive in some cases is to prevent the off-site migration of
contamination.   Given  this  objective, periodic applica-
tions of MnO4" have been used to form  a downgradient
 In-Situ Chemical
                        Engineering Issue

-------
(oxidation) barrier to oxidize ground-water contaminants
migrating into  the treatment zone.  Under this condi-
tion, the oxidation efficiency is expected to be lower rela-
tive to the source zone due to lower concentrations of
contaminants.

Despite the lower oxidation efficiency expected in down-
gradient (non-source) zones, it is estimated that the oxi-
dant demand required to meet the treatment objective
may not be as high compared to source zone applications.
Mass  transfer and mass transport limitations  between
MnO4" and NAPLs,  and between MnO^" and high con-
centrations of sorbed contaminants require longer resi-
dence  times  of the  oxidant and  consequently higher
oxidant demand. Fewer limitations occur between MnO^;
and soluble contaminants, resulting  in faster oxidation,
shorter  oxidant residence times,  and a  lower oxidant
demand.  Therefore,  at least in theoiy, lower concentra-
tions of oxidant and fewer applications are required to
meet the treatment objective. Oxidation of soluble con-
taminants in  low-permeability materials  require longer
residence times for diffusive transport and will result in a
higher oxidant demand relative to more permeable aqui-
fer materials.

ISCO is more often used in unconsolidated porous media
than in fractured media.  This is partially attributed to a
greater  number of hazardous  waste  sites  in unconsoli-
dated porous media and more in-depth knowledge of the
flow system. In-situ permanganate oxidation has been car-
ried out in fractured shale (Parker, 2002)  and fractured
bedrock.  Fenton oxidation has significant limitations in
fractured systems due to significant differences between
the reaction and the transport rates of H2O2.  Specifically,
contamination in fractured media may be found in both
the primary porosity  (matrix) and in the secondary poros-
ity (fractures). The volume of matrix porosity is generally
greater (10 to lOOOx) than fracture porosity. Therefore,
the mass of contamination  contained within the matrix
porosity is potentially much greater than  in the fracture
porosity. Transport of contaminants, O^, H7O7, and Fe+2
within the matrix porosity is predominantly by diffusion.
Due to the rapid rates of reaction of O3, H2O2 and Fe+2
relative to diffusive transport, there is insufficient time for
the oxidant and reagents to  penetrate the  contaminated
media.  Consequently, O3  H2O2,  and Fe+2 transport  is
restricted  mainly  to  fracture porosity, the  contaminants
predominantly reside within the matrix porosity, and poor
contact occurs  between the oxidant, reagents,  and the
contaminant.  Although poorly documented, a few unsuc-
cessful attempts of Fenton-driven oxidation  in fractured
systems have occurred.  No cases were found where O3 or
S2O82" was injected into fractured media.
Ill.C.l.b. UnsaturatedZone

ISCO is less frequently used in the unsaturated zone than
in the saturated zone. The delivery of permanganate solu-
tion into the unsaturated zone has occurred via a variety
of methods including,  but not limited to, application to
surface soils, emplacement into trenches/excavations, for-
mer surface impoundments, deep soil mixing, and injec-
tion  into  hanging wells/injectors  (screened in  the
unsaturated  zone).  Similar to the saturated zone, direct
push injection  over short screened  intervals,  at least in
theory, could be used to deliver the permanganate solu-
tion into the unsaturated zone. Presumably, the injection
of persulfate and permanganate solutions would be simi-
lar; however, no reports were  found describing persulfate
application in the unsaturated zone.

Conceptually, due to the slow reaction rate of permanga-
nate and persulfate, vertical transport of the injected oxi-
dant solution would result in the delivery of the oxidant
to areas underlying the  injection zone.  Due to the lack of
buoyancy forces, the rate of vertical transport in the unsat-
urated zone  would be greater than in the saturated zone.
Most subsurface systems exhibit some  degree of anisot-
ropy where the  ratio of horizontal to vertical conductivity
is  10 or greater. Under this condition, vertical transport
could require long residence times of the oxidant.  This
suggests that the fast reaction rate of H2O2 would result
in short vertical transport distances and poor distribution
within the unsaturated  zone.

In-situ O3 oxidation in the unsaturated zone  has several
potential advantages over that in the saturated zone: (1)
the concentrations of O3 that can be achieved in the gas
phase are orders of magnitude higher than is obtained in
aqueous solutions, (2) O^ is more stable in the gas phase
than water, (3) O3 diffusive transport is much greater than
in water, and (4) higher flow velocities can be achieved in
the unsaturated zone than are possible in ground water
(Masten and Davies, 1997). Additionally, the mass deliv-
ery of gaseous  O3  in the unsaturated  zone is generally
much greater and potentially more effective than  in the
saturated zone.  However, delivery and transport  of O3
gas in the unsaturated  zone is much more vulnerable to
preferential pathways attributed to heterogeneities in per-
meability. Consequently, short circuiting of O3 gas may
prevent adequate delivery of O3 to targeted zones. Spatial
monitoring of O3 in the unsaturated zone is required to
accurately assess the areal distribution of O3 and to make
the appropriate  adjustments in the injection design and in
O3 delivery.  Although targeting contaminants in the
unsaturated zone limits  oxidative treatment to zones above
the water table, temporary depression of the water table
       Engineering Issue
                     fri-Slu Chemical Oxidation

-------
exposes additional porous media and contaminants that
may be vulnerable to in-situ O3 oxidation.  Unsaturated
zone  remediation may  primarily  use  ozone  injection
points in the unsaturated zone,  but sparging points just
below the  water table would help deliver  O3 into  the
smear zone associated with the capillary fringe.

Dry KMnO4 and various concentrations of KMnO4 solu-
tions  have been applied to the surface soil in the unsatu-
rated zone.  For example, TCE-impacted soil was treated
with solutions of KMnO4 in one-foot lifts (Balba et al.,
2002).  Two injection techniques, (1) low pressure injec-
tion of KMnO4 into wells and (2) high pressure injection
into nozzles, were used to deliver KMnO4 into the unsat-
urated zone.  More uniform  and better distribution of
the oxidant was observed with the high pressure delivery
method (McKay and Berini, 2002). Nested, hanging
wells  were  used to deliver KMnO4 into the unsaturated
zone, which eventually  drained to  the saturated  zones
(Viellenava et al., 2002). High initial injection pressures
were  used  in these  wells to  hydro fracture  the porous
media,  which created more flow pathways and greater
oxidant delivery.

III.C.2. Contaminant Characteristics

III.C.2.a. Phase

Organic contaminants may be  present in  the aqueous
phase  (dissolved  in  water),  solid phase (adsorbed onto
                                  aquifer material or soil), and as a non-aqueous phase liquid
                                  (NAPL). Chemical oxidation of organic contaminants can
                                  occur in all three phases.  Many variables and site-specific
                                  conditions play a role in terms of which phase of the con-
                                  taminant is oxidized.  When contaminants are oxidized in
                                  the aqueous phase, increased concentration gradients and
                                  enhanced mass  transfer and oxidation  of contaminants
                                  occur from both the solid phase  (desorption) and from the
                                  NAPL phase (dissolution) into solution.

                                  In  addition  to  the  background oxidant  demand, the
                                  quantity of oxidant and number of oxidant applications
                                  needed is dependent on the phase of contaminant present
                                  (NAPLs > solid >  aqueous) and  the associated phase-
                                  dependent mass transfer and mass  transport limitations
                                  (Figure 3). Assuming mobile NAPL is determined to be
                                  present at  a site,  NAPL  removal  is considered  an
                                  important first step in the remediation process.  Removal
                                  of NAPLs by other methods (such as thermal) can usually
                                  be accomplished more cost effectively than by chemical
                                  oxidation and can be conducted in a manner to minimize
                                  NAPL  mobilization into  undesired  areas.  Chemical
                                  oxidation would then be more  appropriate to use on the
                                  immobilized NAPLs (residual saturation) that remain in
                                  the porous media.

                                  Due to the heat and large quantities of O2(g)  released
                                  during Fenton  oxidation, fire, explosion  hazards,  and
                                  safety issues may become important if the NAPL is vola-
                                  tile and flammable, such as gasoline.
         High

   Mass of
   Oxidant
  Required,  o>
    and/or  to
 Number of  ^
   Oxidant  —
Applications

        Low
                                                                            Limitations

                                                                        Dissolution mass transfer
                                                                        Desorption mass transfer
                                                                        Diffusive mass transport
                                                                        Desorption mass transfer
                                                                        Diffusive mass transport
                                                                        Diffusive mass transport
                                      NAPL +
                                      Solid +
                                      Aqueous
                             Solid +        Aqueous
                             Aqueous       (low f,,,.)
                        Contaminant Phases
Figures. Impact of contaminant phases, mass transfer, and mass transport limitations on the mass of oxidant and/or the number
         of oxidant applications needed for ISCO. The presence of all three contaminant phases (NAPL, solid (adsorbed), aqueous
         (soluble)) represents the most challenging set of conditions, potential limitations, and mass of oxidant and/or number of
         oxidant applications.
 In-Situ Chemical Oxidation
                                                           Engineering Issue

-------
III.C.2.b. Concentration

Chemical oxidation can be described by second-order
reaction rate kinetics (Eqn 1).  It is evident that the oxi-
dant can react with either the target contaminant, scav-
engers (i.e.,  non-target  reactants),  and intermediates.
The extent to which the oxidant reacts with either is
dependent on the reaction rate constant of the reactant,
and  the concentration of the reactant.  Therefore,  the
greater the concentration  of contaminants  relative to
other potential reactants, the faster the reaction rate of
the  target contaminant.  This  partially explains why
greater oxidation efficiency occurs in source zones where
high concentrations of the target contaminants are pres-
ent.   The feasibility of treating relatively low dissolved
concentrations of organic contaminants  may not  be as
favorable, and these concentrations may be more  effec-
tively treated by other  candidate technologies, such as
monitored natural attenuation (MNA).

  dO/dt = k,  [O] [C] + k2 [O] [S] + k3 [O] [I]    (Eqn  1)
where:
kj, k2,
            =  second order reaction rate constant
               (L/mol-s)
       [O]  =  concentration of oxidant (-OH,
               MnCV, O3, S2O82vSO4-) (mol/L)
       [C]  =  target contaminant (mol/L)
       [S]  =  scavenger (mol/L)
       [I]  =  intermediates (mol/L)
III.C.3. Subsurface Characteristics

Physical  and chemical characteristics of the subsurface
environment (hydrogeology, geology, geochemistry) vary
from site to site and impact the fate and transport of the
injected  oxidant  and reagents.  Site characterization  is
critical to the feasibility assessment of ISCO and in the
planning and design of pilot- and full-scale ISCO sys-
tems.

III.C.3. a. Geology

Fractures (cracks, fissures, joints, faults) are characterized
by their  length,  orientation, location, density, aperture,
and connectivity (Berkowitz, 2002). Transport of injected
reagents into fractured media is generally much less pre-
dictable than into unconsolidated  porous media, mainly
due to the heterogeneity and uncertainty in fracture char-
acteristics and the difficulty and expense in characterizing
the fracture network. Tracer studies can be helpful to
identify the interconnectedness between monitoring wells,
the rate  of ground-water transport, and the residence
time of the injected water. Borehole hydrophysics can be
used to assess ambient and stressed flow patterns and con-
taminant transport in fractured systems.  These investiga-
tion methods provide general information  on hydraulic
characteristics and hydraulic control  requirements of a
tracer prior to oxidant injection.  Due to density-driven
transport of MnO4~ and S2O82~ solutions, vertical trans-
port may not be fully represented by a tracer study. Due
to transport limitations of H2O2 and O3 in fractured sys-
tems (i.e., within the matrix porosity, refer to Section III.
C.I.a. Saturated Zone), Fenton and O3 oxidation would
have limited use in fractured systems.

Naturally occurring  subsurface heterogeneities such as
zones of high permeability (e.g.,  sand-filled  paleochan-
nel, fractures), as well as subsurface utility corridors and
other anthropogenic subsurface disturbances, can act as
preferential pathways.   Preferential pathways found in
fractured systems and unconsolidated porous media result
in unpredictable flow patterns (rate, direction) for ground
water and injected oxidant solutions.  This can be a sig-
nificant impediment to effective/uniform delivery of oxi-
dant in the  subsurface. Additionally, under high injection
pressures, hydraulically-induced fractures of the media
and/or "breakout" of the  injected oxidant solution may
occur.  Both  of these conditions result in disproportion-
ate volumes  of oxidant solution  being injected over a
small geologic interval. The lowest removal rates of TCE
and DCE in  ground water were found five feet from an
O3 injection point, compared  to significantly  higher
removal rates in wells up to  twenty feet away from the
injection point (Masten and Davies, 1997).  This exam-
ple indicates  that ozone sparged into ground water may
be transported in  specific and limited  preferential gas
channels that short-circuit and  do not contact a signifi-
cant portion of the region  around a sparge point.

These  problems can  be detected and avoided  early
through  good  site characterization  and  ground-water
monitoring.  Also, shorter distances  between injection
wells minimizes the areal coverage and the oxidant trans-
port distance per well, thus limiting the impact of these
nonideal transport mechanisms.

III.C.3.b. Hydrogeology

The oxidants and  reagents injected into the subsurface
will  undergo advective and  diffusive transport.  The
transport distance is  dependent on the method of deliv-
ery, persistence of the chemical, ground-water flow rate,
density of the solution, and diffusive characteristics of the
chemical and porous media  (Figure 4).  For example,
slow reaction rate and long-term persistence of MnO4~ in
       Engineering Issue
                                                                          In-Itu Chemical Oxidation

-------

   s
   •5  .
                                     Long
           Short
            Stow        Med/um        Fast
               Rate of Ground-Water Flow

Figure 4.  The transport distance through porous media of the
         injected oxidant is dependent on the rate of reaction
         of the oxidant, the rate of ground-water flow, and
         the density of the oxidant solution (density-driven
         transport).

aquifer materials has occurred under  a variety of hydro-
geologic  environments including saturated  sands,  clays,
sand-clay mixtures, alluvial materials, fractured shale, and
fractured bedrock.  Consequently, MnO4" can be trans-
ported longer distances than the other  oxidants.   The
greater the hydraulic conductivity and hydraulic gradient
of the aquifer, the farther the transport distance. Due to
the fast reaction rate of O3, H2O2,  and some reagents
(i.e., Fe(II), phosphate, acid), the ground-water flow rate
and direction under most conditions will have a minimal
impact on the post-injection oxidant transport. The ver-
tical transport of the oxidant is affected by density of the
oxidant  solution and on  vertical  gradients  (upward,
downward) in the aquifer. The transport of MnO4~ will
be greatest when high concentrations are  injected into an
aquifer with  high hydraulic gradient and conductivity.
Under this set of conditions, MnO4" may migrate from
the targeted zone and result  in  a lower oxidation effi-
ciency (i.e., due to lower downgradient contaminant con-
centrations)  or be transported into unintended locations
(i.e., ground-water capture by pump and treat system).
Downgradient drift is not always problematic since pin-
point resolution of source area boundaries is not com-
monly  achieved   and  oxidation   of  downgradient
contaminants may be needed.  Similarly, long oxidant
transport distances may be intentionally  designed to tar-
get large, non-source area plumes.

III.C.3.C. Geochemistry

Permanganate oxidation is generally independent of pH
in the range of 4 to 8 (Seigrist etaL, 2001) and thus will
be effective over the pH range normally found in ground
water.  Acidic pH (pH 3 to  4)  is optimal for Fenton-
driven oxidation of organic contaminants.  The buffer
capacity in most aquifers represents significant acid-neu-
tralizing  capacity and maintains the ground-water  pH
near  neutral,  i.e.,   resistant  to  pH  modification.
Acidification of the target zone is often temporary and
restricted to a zone near the injection well.  In poorly
buffered  systems, acid transport and pH modifications
may be easier to accomplish. Under naturally acidic con-
ditions, pH modification  may not be necessaiy.   For
example, high concentrations of Fe+2  and acidic condi-
tions (pH 2 to 3) were measured at a site where large
quantities of organic  carbon were introduced into  the
ground water (i.e., the aquifer underlying former sludge
drying beds) (U.S. DoD, 1999; Maughon etaL, 2000).

Reduced geochemical conditions favor the presence of
reduced divalent transition metal ions such as Fe or Mn.
Under this condition,  the Fe+2 contributes to the Fenton
reaction, and both Fe+2 and Mn+2 contribute to the acti-
vation of the -SO4~.  For example, significant chlorinated
VOC reduction was achieved using persulfate in a glauco-
nitic  (iron-rich) sand containing 3  to 15 mg/L Fe+2
(Sperry etaL, 2002). Reduced geochemical environments
are also a source of treatment inefficiency due to the abun-
dance of reduced chemical species that consume oxidant
and/or scavenge radicals. Natural organic carbon present
in soil and aquifer material can play an important role in
oxidative reactions  (refer to  Section III.F.4.b.  Natural
Organic Matter).

III.D. Site Requirements and Operational Issues

III.D.I. Site Characterization Data

An effective delivery of the oxidant to the targeted zone(s)
is  a critical  element  to achieve success with  I SCO.
Therefore, an important ISCO design  criteria is to iden-
tify the location(s) of the contaminant(s) in the subsur-
face.  Site characterization data is required to identify the
type(s), distribution, and phase/concentration of the con-
taminants (Table 11).  This information is used in  the
planning and the design of the oxidant  injection program
(i.e., dosage, injection  locations, and rates).

Site characterization and ISCO efforts are often focused
in source areas where NAPL may be present. Removal of
mobile NAPL, if present and practical,  is important since
NAPLs could be mobilized  during ISCO. Assuming the
NAPL (LNAPL, DNAPL) can be located, removal would
probably be more cost efficient using other  technologies
than through ISCO.  A significant improvement in  the
development in site characterization techniques and tech-
nologies to locate and delineate suspected DNAPL source
 In-Situ Chemical Oxidation
                        Engineering Issue

-------
Table 11. Site Characterization Data Needed for ISCO
Parameter
Target Contaminant
Type(s)
Distribution
Phase/concentration
Geology and
Hydrogeology
Aquifer Material/Soil/
Ground Water
pH, Buffer Capacity
Eh (electrode potential)
Reduced Inorganics
Purpose of the Data/Information
To select which oxidant is most suitable for the specific contaminant(s).
To determine where to deliver the oxidant (spatial delivery of the oxidant).
Aqueous/sorbed/NAPL — to identify potential hot spot areas where multiple applications will be required; to
estimate contaminant mass which may be used to estimate the total oxidant mass required (see Section III.C.2).
Estimate rate of oxidant injection based on aquifer hydraulic properties. Information on hydraulic conductivity
and gradient and aquifer heterogeneities can be used to identify post-injection flow direction and rates. Assess
whether nearby receptors could be impacted (see Section III.C.3).
Analyzed for organics to quantify and delineate contaminant distribution; used for bench-scale feasibility
testing (contaminant oxidation, oxidant demand testing); analyzed for metals to identify hot spot or
problematic conditions where pH- or redox-sensitive metals may become mobilized during ISCO,
To assess whether pH modification is needed; acidic pH (3 to 5) is optimal for Fenton oxidation; pH modification
will be difficult to achieve in highly buffered soil/aquifer materials; carbonate and bicarbonate buffer species
act as radical scavengers in Fenton and persulfate oxidation.
General indicator of oxidant demand.
Soluble metals (Fe(ll),Mn(ll)),sulfides indicate reducing conditions.
zones has occurred in recent years.  The cost and level of
accuracy achievable by source zone characterization tools
can only be answered on a site-specific basis (U.S. EPA,
2001). In practice, delineating DNAPL source zones and
providing an accurate estimate of the mass and spatial
distribution of the DNAPL can be challenging due to the
heterogeneous distribution of  the DNAPL. Extensive
sampling and analysis of ground water and aquifer mate-
rial is usually required to obtain a reasonable estimate of
contaminant distribution. Oxidant dosage requirements
have been estimated by  some practitioners based on the
mass of contaminant(s), in conjunction with oxidation
stoichiometry and the  natural  oxidant demand.  Other
methods are also used  to estimate  the  total  oxidant
required  that are not based on contaminant mass esti-
mates  (refer to  Section  III.E.2.  General Conceptual
Approach to ISCO).

Geologic  (lithology,  stratigraphy,  heterogeneities)  and
hydrogeologic (hydraulic conductivity, gradient, poros-
ity) characterization will assist in development of a con-
ceptual model used to assess the fate and transport of the
contaminant(s),  injected oxidant, and reaction  byprod-
ucts. This information can be used, in conjunction with
the contaminant data, to select  the location and vertical
intervals for oxidant injection and to assess oxidant trans-
port. Important parameters that can be easily overlooked
are man-made and naturally occurring preferential path-
ways. Subsurface utilities and high hydraulic conductiv-
ity flow paths may result in the disproportionate transport
of oxidant, heat,  and gas, and unanticipated exposure
pathways. Even well-characterized sites are likely to have
heterogeneities that are difficult to quantify, yet play an
important role  in  ISCO.  Pilot-scale studies  are  useful
because they provide an opportunity to acquire localized
fate and transport data for the injected oxidant and to
refine performance monitoring efforts. Aquifer material,
soil, and/or ground-water samples are collected for site
characterization (organic contaminants), bench-scale fea-
sibility testing,  and analysis for general  parameters.  In
some cases, high background concentrations of metals, or
co-disposal of metals and organics, represent conditions
that are conducive for metals mobilization during ISCO.

Due to the uncertainties in site characterization, estimat-
ing contaminant mass, oxidant delivery, etc., the need for
       Engineering Issue
                     fn-Stu Chemical Oxidation

-------
multiple oxidant injections in the targeted zone cannot
be overemphasized.

III.D.2. Required Site Infrastructure

Infrastructure requirements for permanganate, peroxide,
and  persulfate include a water  supply, for dissolution
and/or  dilution  of the  reagent.   Extraction wells and
above-ground treatment are necessary if oxidant recircu-
lation (injection, recovery, re-injection) is used.  A dis-
posal option for treated/untreated  water may also  be
required.  An open, unobstructed  area is necessary  to
accommodate the oxidant mixing apparatus, batch stor-
age tanks, piping, etc.  An electrical power supply or gen-
erator will be required for pumping and pressure injection
of the oxidant solutions. Direct push injection technolo-
gies  are rig-mounted and site access for large or small
vehicles in the injection area is a consideration.

The  small radius of influence of the Fenton's  reagent
(H2O2,  Fe+2) requires closely-spaced  injection points that
involve  an open area clear of underground  utilities. Due
to the heat, O2(g),  and possible volatile emissions that
could be released, access should be restricted to the injec-
tion  area. Veiy high  temperatures often result from
Fenton  oxidation  and  have  melted polyvinylchloride
(PVC) plastic pipes (the crystalline  melt temperature of
PVC is approximately 200 °C). Injection wells should be
constructed  with stainless or  carbon steel.   A soil vapor
extraction system in  conjunction with an impervious
cover and off-gas treatment may be  necessaiy to  capture
volatile emissions from the injection  area. Excessive pres-
sure  build-up may result in dangerous gas/steam venting
conduits at  the ground surface  (randomly distributed)
and  buckling/heaving  of  asphalt parking lots  (mostly
under high H2O2 concentrations).

In-situ ozonation requires an on-site  ozone  generator, gas
handling and distribution equipment, and injection wells.
Extraction wells (i.e., a soil vapor extraction system) may
be necessary to control O^  flow  directions (Masten and
Davies,  1997). Subsurface heating  and  the associated
remedial technology infrastructure would be needed for
thermal activation of persulfate. Radio frequency heating,
electrical resistance heating, or any other thermal technol-
ogy,  is  technically  feasible for heating the subsurface
(Liang etaL, 2001; 2003).

III.D.3. Regulatory Constraints on Injection of Reagents

Regulatory permitting requirements  for oxidant injection
have been  compiled  and  organized by  state  (ITRC,
2005).  The injection of oxidants and reagents are regu-
lated  primarily  through  the  Underground  Injection
Control (UIC) program of the Safe Drinking Water Act
(SDWA), the Resource Conservation and Recovery Act
(RCRA),  the  Comprehensive  Emergency  Response,
Compensation, and  Liability Act (CERCLA),  and the
Emergency Planning and Community Right to Know
Act (EPCRA).  Through these environmental programs,
regulatory approval is required and an oxidant injection
permit may also be  required from some state environ-
mental agencies.  Some states have issued variances and
permit  exceptions  that may  affect  ISCO activities.
Regulatory examples  of six states (NJ, CA, FL, KS, MO,
TX) are  provided  in which chemical oxidation can be
used  for  soil and  ground-water remediation  (ITRC,
2005). Individual states may have more restrictive regu-
lations than the Federal programs listed above. Regulatory
constraints on each ISCO project should be assessed on a
case-by-case basis.

EPA's secondary maximum  contaminant level  (SMCL)
for manganese in water  (0.05  mg/L)  is  a secondaiy
drinking water standard due to aesthetics (taste, color,
staining)  (Table 12) (U.S. EPA, 1992). The post-oxidation
manganese content of aquifer material can be high.  At
one site  where KMnO^driven  ISCO was  used, post-
oxidation ground-water  concentrations  one year  later
exceeded the EPA SMCL for Mn (Crimi and  Siegrist,
2003). Ideally, due to the insolubility of Mn as MnO2(s),
ground-water  concentrations of Mn  will  be  minimal
under most conditions.

Human  consumption  of MnO4~ by recovery in water
supply wells represents a potential exposure pathway and
a serious health threat (but has not been reported).  The
characteristic  purple color of MnO4', or pink color at low
concentration, could  alert water treatment plant operators
and potential consumers of the oxidant-tainted water.
Releases  to nearby  surface  waters  could have serious
environmental impact on biota and should be prevented.

Technical grade KMnO4 may contain impurities, includ-
ing Cr and As.  Due to the low maximum contaminant
level (MCL) in drinking water set by EPA for these met-
als (0.1 mg/L total Cr MCL; 0.01 mg/L As MCL) (U.S.
EPA,  2002),  injection of technical grade KMnO4 may
result in  exceeding the MCL for these elements in the
injection zone.  Although the attenuation of these metals
has typically  been  achieved within acceptable transport
distances and time frames (see  Section II.C.l.f, Metals
Mobilization/Immobilization),  a site-specific assessment
should be conducted  to determine whether these parame-
ters need to be monitored. KMnO4 is produced by some
manufacturers specifically to lower the concentrations of
 In-Situ Chemical
                        Engineering Issue

-------
these impurities (at additional cost). Additionally, low
concentrations  of the oxidant solution can be  used  to
minimize  metals concentration  in  the ground water.
KMnO4 will exhibit minute amounts of radioactivity.
Radioactive potassium-40 (40K) is a small fraction (about
0.012%) of naturally occurring potassium and is primar-
ily  beta-emitting,  but  also  involves gamma radiation
(ANL, 2002).  This may become  an issue at radionuclide
contamination sites, such as some DOE sites, where it is
important to monitor and limit radioactivity.

ISFO may also  involve the injection of acid(s), iron cata-
lyst, and/or stabilizers, any of which  may be  subject  to
regulatory constraints. Regulators may also be  concerned
about the possibility of contaminant volatilization and
subsequent releases to the unsaturated zone,  basements,
and air. Similarly, releases of both volatile organics and
O.j during in-situ  ozonation are a regulatory concern.
The use of a conservative insoluble tracer gas (e.g., helium
(He)) during air sparging/injection and/or O3 sparging/
injection could  provide fate and transport information of
volatile  organics, Ov and the proper placement of soil
vapor monitoring points.  Sulfate is a reaction  byproduct
from  in-situ  persulfate  oxidation. Since persulfate  is
injected at high concentration, and can persist for weeks
to months, the potential for persulfate and/or sulfate  to
migrate to nearby receptors should be assessed.  Although
there are well-documented biotic  and abiotic attenuation
mechanisms, ground-water monitoring is recommended.
EPA has established a secondaiy MCL for sulfate  in water
(250 mg/L) due to a salty taste (Table 12).

III.E. Field-Scale Implementation and Engineering
     Design Considerations

III.E.I. Treatment Objectives

The treatment objectives for I SCO vary from  site to site
and include, but are not limited to the following: reduc-
tion in contaminant toxicity/mass/concentration (risk
based or maximum concentration levels (MCLs)), and/or
a reduction in contaminant mass flux across a site bound-
ary.  ISCO is a source depletion technology that is capa-
ble of removing substantial amounts of DNAPL in source
zones at  sites  with favorable hydrogeologic conditions
(i.e., less  heterogeneous and more permeable subsurface
conditions); however, achievement of drinking  water
MCLs  in  these source zones as well as source zones  in
more challenging  heterogeneous  hydrogeologic  condi-
tions (e.g., bedrock, karst systems, multiple stratigraphic
units) is unlikely (U.S. EPA, 2001). However, ISCO is
capable of achieving partial DNAPL  depletion, which
may provide other performance benefits including elimi-
nating  the mobility of the DNAPL, and reduction in the
mass discharge rate  of DNAPL constituents from the
source  zone, which may reduce environmental risks and
life cycle costs (U.S. EPA, 2001).  One of the alternative
metrics for  judging the performance of  source-mass
depletion  technologies is contaminant mass discharge,
defined as the summation at a point in time of point val-
ues of contaminant  mass flux across a vertical control
plane encompassing the plume and perpendicular to the
mean ground-water flow direction at a location downgra-
dient of the DNAPL source zone. Theoretical  analysis
and  field data indicate that partial DNAPL mass deple-
tion in the source zone  reduces  contaminant mass dis-
charge  (U.S. EPA, 2001).

ISCO is often deployed in source areas to minimize long-
term sources of ground-water contamination.  However,
ISCO has also been deployed at property boundaries for
the purpose of preventing off-site migration of ground-
water  contaminants, and  in weathered plumes where
NAPL  is  present in small volumes, or absent altogether.
In this scenario, the majority  of the contaminants are
present as soluble and sorbed phases at lower concentra-
tions (than in source zones).  Under these operating con-
ditions, ISCO has a much higher probability of achieving
Table 12. EPA Secondary Maximum Contaminant Levels (SMCL) (Abbreviated List)
Contaminant
Iron
Manganese
Sulfate
Secondary MCL*
0.3 mg/L
0.05 mg/L
250 mg/L
* mg/L is milligrams of substance per liter of water
http://www.epa.aov/safewater/consumer/2ndstandards.html

Noticeable Effects Above the SMCL
Rusty color; sediment; metallic taste; reddish or orange staining
Black to brown color; black staining; bitter metallic taste
Salty taste

       Engineering Issue
                     fii-Stu Chemical Oxidation

-------
the MCL objective, than in source zones  that contain
NAPLs.  I SCO has also been used to (1) reduce the mass
flux to pump and treat systems, and (2) to reduce the con-
centration gradient  across a low-permeability  barrier in
hydraulic containment systems.

Due to  significant challenges required of ISCO to meet
stringent clean-up standards, such  as MCLs  in  source
zones, monitored natural attenuation is an integral com-
ponent  in the overall remedial strategy for source zone
and downgradient plume.

III.E.2. General Conceptual Approach to ISCO

It is uncertain how bench-scale oxidant demand  values
relate quantitatively to the actual oxidant demand mea-
sured under field conditions.   It should be recognized
that conditions at bench-scale (e.g., solids:solution ratio,
mixing, contact  between oxidant/reagents and contami-
nants/aquifer material,  etc.) are significantly different
than field-scale, and that samples used in the bench-scale
study may not fully represent field conditions  (i.e., het-
erogeneities).  Correspondingly, there is uncertainty in
data interpretation and how the information is used in
pilot- and field-scale applications.  Review of numerous
ISCO reports reveals that different approaches are used to
design oxidant (MnO^, H2O2) loading for pilot- and full-
scale ISCO systems.  However, economic and infrastruc-
ture resources/limitations also play a strong influence in
the oxidant dosage and injection program.

A single, well-documented,  and well-demonstrated oxi-
dant loading and delivery design approach has not been
established for any oxidant.  Overall, the state of the sci-
ence of ISCO involves the  combined use of best engi-
neering and  scientific judgment (site characterization,
feasibility study testing, remedial design, etc.) in conjunc-
tion with trial and error. Due to the inherent uncertainty
with  contaminant distribution,  subsurface heterogene-
ities, and mass transfer/transport mechanisms that  occurs
at most sites, ISCO requires multiple iterations between
oxidant  application  and  performance   monitoring.
Through this process, clean areas can be  identified
that require no  further treatment, and hot spot zones
can be identified which permits the strategic delivery
of additional oxidant to  accomplish the treatment
objectives.

III.E.2.a. Permanganate Oxidation

There are two general approaches used with MnO4~: the
"high oxidant loading" and the "iterative oxidant load-
ing" approaches. Some approaches reported in  the litera-
ture use  various  combinations  of the two methods
described below.

III.E.2.a (1) High Oxidant Loading

The "high oxidant loading" approach seeks to apply all
necessaiy oxidant mass in one initial application.  This
approach utilizes oxidant demand results and informa-
tion from bench- and pilot-scale oxidation testing in con-
junction with empirical factors to estimate the oxidant
loading for an equivalent mass of contaminated aquifer
material  at field-scale in the targeted zone(s). Empirical
factors used in the calculation have been based on a mar-
gin of safety, field experience, types of aquifer materials,
oxidation stoichiometry, estimated mass of contaminant(s),
reaction rate kinetics, differences between short-term and
long-term oxidant demand tests, etc.   Subsequently the
oxidant load is estimated and delivered into the targeted
zone(s).  It is assumed that the  aquifer samples used in
the bench test are representative of the targeted zone(s).
This approach may potentially result in fewer iterations
between  oxidant injection and ground-water monitoring,
fewer field mobilizations, and potentially lower cost. Due
to heterogeneous contaminant distribution, variability in
background oxidant demand,  and the inability to uni-
formly deliver the oxidant, this  approach may result in
higher oxidant loading (and cost) than needed.   Higher
oxidant loading/concentration may result  in the trans-
port of oxidant  from the targeted zone, higher oxidant
demand, greater metals mobilization,  greater Mn resid-
ual, and may eventually require the same number of iter-
ations  and  mobilizations  as  in the  iterative  oxidant
loading approach.

III.E.2.a (2) Iterative Oxidant Loading

This approach recognizes that multiple iterations between
oxidant injection and post-oxidation monitoring will be
required.   Results  and  information  from site-specific
bench-scale oxidation testing may or may not be used. It
is assumed that aquifer samples in bench-scale studies may
not be representative of the target zone; contaminant dis-
tribution is heterogeneous; moderate oxidant loading may
permanently  reduce contaminant  concentrations over
large areas of the target  zones; contaminant concentra-
tions may be reduced and treatment objectives achieved
without satisfying the total oxidant demand of the aquifer
material;  and   subsequent,  perhaps   heavy,  oxidant
loading(s)  may  be  required in  smaller  hot-spot zones
where mass transfer and mass transport limitations exist.
The advantage of this approach  is that potentially lower
total oxidant loading and costs may occur. The disadvan-
tage is that potentially  a greater number  of iterations
 In-Situ Chemical
                         Engineering Issue

-------
between monitoring  and oxidant injections will occur
which could increase cost.

III.E.2.b. Fenton Oxidation

A well-documented and demonstrated oxidant loading
and delivery design approach has not been established and
published.  Several factors strongly influence the mass of
H2O2 required.   Some ISFO vendors base the mass of
H2O2 to be delivered on an estimate of the  mass of con-
taminants,  reaction stoichiometry, and empirical factors.
Empirical factors  used in  these  calculations have  been
based on a margin of safety, field experience, and types of
aquifer  materials.  An estimate of the mass of contami-
nants in conjunction with reaction stoichiometry provides
an initial theoretical estimate of the H2O2 demand under
ideal conditions.  However, due to the difficulty in accu-
rately estimating the mass of contaminants in the subsur-
face, this approach should  be considered a veiy rough
approximation of the  minimum  amount  of H2O2
required. Additionally, other factors and  nonideal  mass
transfer and mass  transport conditions in  the subsurface
must also be considered.  H2O2 reactions that do not pro-
duce radicals (nonproductive reactions) and reactions that
occur between radicals and non-target  species  (scaveng-
ing) lower the  oxidation efficiency. H2O2 and -OH deple-
tion in these reactions results in greater quantities of H2O,
required for ISFO.

Once hydroxyl radicals are formed, they are transported
only  a  few nanometers due to their high  reactivity.
Therefore, a fundamental tenet of ISFO is that the H2O2,
Fe(II), and contaminant should be in the same location at
the same time for oxidation to occur.  The  fast reaction
rate of  H2O2 represents a significant  limitation  in  its
delivery. Veiy simply, the rate of H2O2 transport must be
greater than the reaction rate. Numerous cases have been
reported where H2O2 delivery was restricted to locations
near the injection wells/points (i.e., within a few feet),
which was attributed to rapid H2O, reaction rates.

An analytical  solution was  used to simulate  H2O2 trans-
port in porous  media for steady-state, radial  flow in
homogeneous porous media with pseudo first-order deg-
radation of H2O2 (Figure 5). In this theoretical analysis,
the same volume  and concentration of H2O2 solution
were injected  at different rates.   A low H2O2  injection
rate limited H2O2 distribution and increasing the injec-
tion rate resulted in increasing the transport  distance and
lowering the injection time.

Acidification of the injected oxidant solution will reduce
the reaction rate of H9O9. Fast delivery of the H2O, solu-
tion at  lower pH will increase the transport distance of
H2O2 into the aquifer. Potential limitations of high injec-
tion rates may involve either low hydraulic conductivity or
excessive injection pressure. Shorter injection well spacing
will  reduce die lateral H2O2 transport distance  required
per well and is an additional design parameter that could
be used to increase H2O2 coverage in the targeted zones.

III.E.2.C. Ozone Oxidation

Few details are available that describe the steps and criteria
used in  estimating the mass of O5 needed for in-situ ozon-
ation. One estimate of the average stoichiometric demand
                                a  Q = 4 L/min

                                •- Q = 12 L/min
        0      0.2     0.4     0.6     0.8     1
              Radial Distance from Injection Well (m)


Figure 5. H2O2 undergoes reactive transport in porous media.
         The relative concentration of H2O2 is illustrated for
         two oxidant injection rate conditions (4 L/min and
         12 L/min; steady-state, radial flow, pseudo first-order
         degradation kinetics; same volume (30,000 L) and
         concentration of H2O2 solution injected). The faster
         the injection rate, the greater the transport (radial)
         distance of H2O2. The equation was solved for radial
         distance (R). Units were converted for dimensionless
         terms ([H2O2]t, [H2O2]0 = time dependent and ini-
         tial H2O2 concentration, respectively; KH202 = H2O2
         degradation rate (0.91 hr1); R = Radial distance (m);
         Z = Vertical interval (3 m); n = Porosity (0.3); Q = Flow
         rate (L/min)).
       Engineering Issue
                      fn-Situ Chemical Oxidation

-------
for phenanthrene was 8.69 g ozone per g phenanthrene
(Kim and Choi, 2002). The ratio of ozone mass to PAH
hydrocarbon mass at a former fuel oil distribution terminal
was 7.6:1 (Wheeler et al., 2002).  An ozone dosage of
about 0.5 g ozone/kg soil resulted in the removal of 81%
of 100 rng/kg pyrene (Masten and Davies, 1997). These
preliminary bench-scale results, in conjunction with esti-
mates of the contaminant mass,  could provide a general
indication of the  mass of O3 required. However, under
field conditions, nonideal mass transfer and mass transport
mechanisms,  reactive  transport,  and reactions between
non-target species and  O3 not represented in the bench-
scale  studies  will  result in  a  greater  O3  demand.
Remediation of a former manufactured gas plant site used
a total 52 Ibs O3/day of 5 to 10% O3 and 90 to 95% O2,
injected at 7.5 scfm and 20 psi,  with occasional increases
to 8 to  12%  O3 for contaminant hot spots (Cambridge
and Jensen, 1999). The O3 concentration in the gas stream,
the gas  flow rate, and  the injection pressures should be
included when specifying the O3 mass delivery rate.

O3 will react with  OH' in water,  which continuously
results in an O3 demand, suggesting that the O3 demand
is essentially infinite. This is similar to Fenton oxidation
where H2O2  reaction is infinite.  Therefore, in-situ O3
oxidation involves a trial and error approach where O3 is
delivered and ground-water monitoring  is used to assess
the extent of O3  transport and distribution and contami-
nant destruction.  Modifications to the O3 injection strat-
egy to  enhance  O3  coverage and/or transport include
closer injection well spacing, and/or faster O3 mass deliv-
ery rate  (increasing O3 content, gas flow rate,  or O3 pres-
sure).

III.E.2.d. Persulfate Oxidation

There are similarities between persulfate and permanga-
nate oxidation (see Section III.E.2.a, above).  For exam-
ple, the  oxidant demand measured in bench-scale studies,
in conjunction with empirical factors, has been used to
estimate the oxidant loading for an equivalent mass of
contaminated  aquifer  material  at  field-scale. A  similar
approach could possibly be used with persulfate. However,
few details are available that describe the steps and criteria
used in estimating the mass of Na2S2O8 to inject, the per-
sistence of Na2S2O8,  and the development of oxidant
injection guidelines. Some information about application
rates and loading approaches can be gleaned. The sequen-
tial use  of Na2S2O8 and MnO4~  in a pilot-scale field test
was conducted involving an aquifer contaminated  with
residual TCE DNAPL  (Droste et aL, 2002).  It was pro-
posed that Na?S?O8 would satisfy the majority of the nat-
ural oxidant demand, thus reducing the MnO4~ demand.
Two treatment zones (each about 8000 m3) were treated
with a total of 8200 kg Na2S2O8 in 4,300,000 L water (1.9
g/L solution) in a 64-day long injection period. Assuming
uniform distribution of Na2S2O8, a soil bulk density of 1.7
g/cm3, and a porosity of 0.36, the Na2S ,O8 loading rate
was approximately 0.3 g Na2S,O8/kg soil. The Na2S2O8
application was soon followed by permanganate injection,
indicating that the iterative oxidant loading approach was
not conducted in this test, at least not for Na2S2O8.

The sequential use of Na2S2O8 and  MnO4" at pilot-scale
was conducted to assess the  efficacy of each oxidant for
the destruction of CVOCs (e.g., TCE, DCE,  and VC)
(Sperry et aL, 2002).  A mixture of 645 kg Na2S2O8 and
ground water/potable water (40 g/L) was injected into a
340 m3 test zone into three  wells (2.5 to 3.0 L/min per
well) for 4 days at 8 hrs/day.  Assuming uniform distribu-
tion of Na,S,O8, a soil bulk density of 1.7 g/cm3, and a
porosity of 0.36, the Na2S2O8 loading rate was approxi-
mately 1.1 g Na2S2O8/kg soil. Again, the Na2S2O8 injec-
tion was immediately followed  by injection of KMnO4
indicating that the iterative oxidant loading approach was
not conducted in this test,  at  least not for Na2S2O8.

The Na2S2O8/kg loading rate was 0.3 g Na2S2O8/kg soil
in the  test where the oxidant was intended to satisfy the
natural oxidant demand, and 1.1 g Na2S2O8/kg soil in
the test where the oxidant was intended to treat the con-
taminants. The Na2S9O8  concentration varied widely in
these two tests (1.9, 40 g/L)  indicating that a "high oxi-
dant loading" approach over smaller areas  is a design
option, and  may  have been intended to  significantly
decrease the contaminant mass in just one application.

III.E.3. Oxidant Delivery

The minimum volume  of H2O2 injected should be suffi-
cient for  full coverage (saturation; pore volume  injection)
of the  targeted zone.   Multiple pore volumes of H2O2
may be required. Simple calculations can be used to pro-
vide a quick check on this important design criteria but it
is often overlooked.  Similarly, full coverage of the target
area is  required in O3 oxidation involving either dissolved
O3 (O3(aq)), O3 sparging, or  O3(g) injection in the unsat-
urated  zone.  Due to the high concentrations of MnO4"
and S2O82' and subsequent density-driven transport (and
diffusive transport), full coverage of the targeted zone may
be achieved without injection of one  pore volume.

Injection of any oxidant solution into a source area can
result in the displacement  of contaminated ground water
from  the source  area  and  transport into potentially
uncontaminated areas. An outside-in  delivery design could
 In-Situ Chemical
                        Engineering Issue

-------
minimize the impact  from the lateral displacement of
contaminated ground-water. Other site-specific strategies
could be developed to prevent the spread of contaminated
ground water during ISCO.   Similar delivery methods
are used for both permanganate and persulfate.  Although
persulfate  oxidation  is  an  emerging  technology,  the
following discussion is also applicable to the delivery of
persulfate into the subsurface.

MnO4" delivery techniques include direct push technol-
ogy, injection wells (with or without recirculation), injec-
tion and recovery (push-pull in wells), hydrofracturing,
and application/infiltration at the ground  surface.  Direct
push technology (GeoProbe, lance permeation, etc.) pro-
vides wide flexibility both in the location and the vertical
interval for oxidant injection.  Ideally,  oxidant injection
(1 to 2 gpm,  20 to 30 psi) occurs in short (approximately
0.5 to 2 ft) screened intervals resulting in the delivery of a
thin layer of  oxidant solution into the aquifer (Figure 6).
This injection strategy minimizes the lateral displacement
of contaminated ground water in the injection zone. The
oxidant is injected at 5- to 10-ft intervals over  the depth
of the targeted zone  and density-driven and  diffusive
transport of the oxidant between the vertical injection lay-
ers results in interlayer distribution of the oxidant solu-
tion.   Nonideal  transport  of  the   oxidant  due  to
heterogeneities can result in  various patterns of oxidant
distribution and  the actual distribution should be  con-
firmed with a good monitoring system. Oxidant leakage
at the ground surface  and breakout  into non-targeted
zones  can  result  from  excessive injection pressures.
Injectors and wells  should  be sealed to prevent short-cir-
cuiting. Injection locations are sometimes backfilled with
a cement/bentonite material immediately after the direct
push tool is extracted to minimize short circuiting of con-
taminants.  Subsequent injection of oxidant in between
previous injection points and at different vertical intervals
allows flexibility in oxidant distribution.  The direct push
technology may not be possible in some geologic environ-
ments where rocks/cobbles/boulders prevent the tool from
advancing into the subsurface.

Injection wells can be constructed with a wide range of
materials (PVC, stainless steel, etc.).  Wells can be  used
Figured. In-situ permanganate oxidation involving the emplacement method of oxidant delivery. Direct push technology can
         be used to inject the permanganate solution over short-screened intervals. Delivery of the oxidant over short-screened
         intervals can reduce the displacement of contaminated ground water relative to injection over longer-screened
         intervals. Stacked, intermittent layers (5 to 15ft) of oxidant will disperse vertically and laterally with time.
       Engineering Issue
                      In-Situ Chemical Oxidation

-------
                                                   IS
to  re-inject  oxidant  solutions  but  re-injection
restricted to the same physical location.  The vertical
interval for oxidant injection  can be varied either  by
using nested wells or by the use of packers to utilize spe-
cific screened intervals.  Oxidant injection at high pres-
sures in nested or packed wells may result in hydraulic
short-circuiting into  the adjacent nested wells above
and/or  below  the  well  where  oxidant is  injected.
Additionally,  excessive  injection pressures  should also
be avoided to minimize oxidant breakout into non-tar-
geted zones via hydraulic fracturing of the porous media.
Ground-water pumping in adjacent wells (until break-
through of the oxidant occurs) may be used  to control/
enhance   the   directional  transport   of  the  oxidant.
Oxidant  recirculation between injection and recovery
wells is occasionally used to enhance delivery/distribution
(Figure 7). This injection method is usually deployed in
low-permeability materials to enhance the delivery of the
oxidant.   Recirculation can involve significant additional
expenses to pump and treat the ground water before it is
re-injected. Expenses with recirculation systems include
but are not limited to engineering, air stripping, filter-
ing of solids (MnO,(s), silicates, etc.), re-amendment
with KMnO4/NaMnO4, disposal of excess water, pip-
ing, tanks, electricity, etc.

Oxidant flooding of former surface impoundments (or
topographical depressions) with underlying contaminants
can be used to deliver MnO4~ solutions over large areas.
Subsequently, infiltration of the oxidant solution into the
contaminated sediments can result in effective delivery
and oxidative treatment. The infiltration surface  should
be well-graded to allow uniform distribution of the oxi-
dant solution over the  ground surface,  and to prevent
accumulation and infiltration of the oxidant solution over
smaller areas. Large diameter (8-ft) augering was used to
deliver KMnO4 to 47 ft below grade into the unsaturated
and saturated zones (Gardner et ai, 1998). Horizontal
wells in either the saturated or unsaturated zone may also
provide an effective delivery technique.  Emplacement of
KMnO4 into fractured  silty clay soils to treat TCE was
investigated (Siegrist et ai,  1999).  Laboratory  experi-
ments and modeling were used to investigate the feasibil-
ity of emplacing solid KMnO4 into vertical wells. Ideally,
Figure 7. In-situ permanganate oxidation involving the recirculation method. Injection and extraction wells are used to
         deliver and recover the oxidant solution. Above-ground treatment is required to remove particulate matter (i.e.,
         MnO2(s), sand, silicates) and possibly COCs/ VOCs, and to re-amend the ground water with permanganate before
         re-injection.
 In-Situ Chemical Oxidation
                                                                                Engineering Issue

-------
the MnO,4~  would slowly diffuse into the ground water
and form a treatment zone to treat contaminant plumes
(Li and Schwartz, 2004B).

H2O2 injection requires injection wells that can withstand
elevated pressures and temperature. Exothermic reactions
and elevated  temperatures that  result  during ISFO
weakens  PVC material and the  elevated pressure causes
material  failure.  Consequently,  wells  should   not  be
constructed  of PVC under these conditions (refer to
Section  II.C.2.e.  O2(g)  Generation  and  Exothermic
Reaction).  Injection  wells constructed  of  stainless or
carbon steel  have been  used successfully under  these
conditions.  Injection wells should be sealed to  prevent
hydraulic short-circuiting along the  well bore  and the
potential release of H2O2 solution at  the ground  surface.
During  H,O,  injection,  O2(g)  expansion,  elevated
pressure, and  breakout of H2O2 can occur.  This can
occur more easily when injecting at shallow intervals.
Breakout could expose site workers to  high concentrations
of H2O2, acid, high  temperature liquids and/or steam,
and contaminants.   Careful   monitoring  should  be
performed during injection to detect this condition and
trigger cessation of injection activities.  Due  to the rapid
reaction  of H2O2, closely spaced injection wells and fast
injection rates are required for the oxidant to be distributed
into the aquifer  (refer  to Section  III.E.2.b.  Fenton
Oxidation).    In-situ Fenton  oxidation often  involves
vendor-specific  reagent   mixes,  injection   methods,
equipment, pressures, and strategies.  A critical review of
the oxidant and  reagent injection  program should be
conducted to assure adequate delivery of these chemicals
under site-specific conditions.

Delivery of O3 into the subsurface mainly involves injec-
tion wells (refer to Section II.C.3.b. In-Situ Application).
Extraction wells may be useful to help control the trans-
port direction of O3 in the subsurface (Masten  and
Davies, 1997). Horizontal wells have been used to intro-
duce O3 into the saturated zone  (Nelson and  Brown,
1994), and may be useful in combination with sparging
technologies.  However, it  is not clear that the potential
advantages of installing horizontal  wells warrants the
additional costs (iMasten and Davies,  1997).  O3 remedi-
ation at a former manufactured  gas plant site used both
vertical injection  wells in  the saturated zone as ozone
sparge  points and  a horizontal well 6 ft under the water
table that had a 135-ft screened section  (Cambridge and
Jensen, 1999). Injection strategies could include an ini-
tial phase of operation consisting of sparging with air
while collecting the  vapors with a soil  venting  system,
such as  at a  PCE-contaminated ground-water site in
Carson City, NV (Masten and Davies, 1997).
III.E.4. Monitoring

///.£.4. a. Ground-Water Monitoring

Ground-water contaminant concentrations represent an
integrated measure of the type, phase, and magnitude of
contaminants in the subsurface. Therefore, the  ground
water can  generally be used as a reliable indicator of
treatment performance, as a diagnostic tool to design and
direct oxidant applications, and  to  determine optimal
oxidant delivery location(s). For example, a reduction in
the concentration of organic contaminants  between pre-
and  post-oxidation, or simply  whether  the resulting
concentrations approach established cleanup  levels  are
often  used as  metrics for  performance evaluation.
Additionally, persistence of VOCs indicates a source area
that  requires additional oxidant application. Fortunately,
oxidants,  especially KMnO4 and  Na2S2O8, can  be
distributed in  source areas in a manner that does not
require  pinpoint accuracy  of  contaminant mass  and
location.  Several  mass  transfer   and  mass  transport
limitations present significant challenges to effective and
efficient ISCO.  These include  slow oxidant transport
through  low-permeability  layers, preferential oxidant
transport  through  high hydraulic  conductivity zones
(inability to deliver the oxidant to the target  area), fast
oxidant reaction rates, background  oxidant demand,
excessive demand in hot spot areas, etc. These conditions,
and  others, will  ultimately  limit ISCO.  The extent to
which  each  of  these  potential  limitations  manifest
themselves at  a  site may  never be  accurately assessed.
However, the  combined effect can be assessed through
pre-  and post-oxidation ground-water monitoring.

Rebound  in post-oxidation ground-water  contaminant
concentrations is time dependent and involves  (1) the
mass transfer from adsorbed and DNAPL phases into the
ground  water, and  (2)  contaminant  mass transport in
ground water to wells where it can be sampled and ana-
lyzed. Collection of ground-water samples immediately
after oxidant injection and/or consumption does not allow
sufficient time for  rebound and would likely represent
transient  (nonequilibrium) conditions.  Ground-water
monitoring should be delayed for these processes to occur
after ISCO has been implemented.  Site-specific contami-
nant transport calculations  can be used to estimate the
time required  for contaminant  transport to monitoring
well  locations  after  oxidant consumption.   Estimates of
the duration of rebound are not well-documented and
may  easily  require months  to fully rebound.  In  Fenton
and O3 oxidation, O2(g) in the porous media may inter-
fere  with mass transfer  and mass transport and  require
longer times for rebound.
       Engineering Issue
                     fri-Situ Chemical Oxidation

-------
Conversely, ground-water samples collected immediately
after ISCO deployment may  be useful for operational
evaluation. For example, ground-water samples containing
significant  contaminant concentrations indicate hot spot
areas that will likely require additional oxidant. Therefore,
decision-making and additional oxidant injections could
occur before rebound is fully developed and documented.
However,  nondetect   or low  concentrations  of the
contaminants  in ground-water samples  are generally
inconclusive until  sufficient time has  been  allowed for
rebound.   Ground-water  sampling  immediately  after
ISCO   may   provide  an  early warning  for metals
mobilization.  Attenuation of mobilized metals is time
dependent  and subsequent sampling may be  necessary to
assess transport and potential exposure pathways.

Ground-water samples collected for analysis of contaminants
that contain the oxidant should  be avoided.  The oxidant
could interfere with analysis of the ground-water sample or
could continue reacting with die contaminant(s) and interfere
widi data interpretation. In the case of MnO4~, reductants
are added to react widi MnO4" and eliminate die oxidant
(e.g.  sodium thiosulfate, sodium bisulfite).  Ground-water
sampling and analysis for contaminants after the oxidant has
fully reacted is recommended.

Ground-water monitoring wells, piezometers, or tempo-
rary  well points can be  used to measure  direct and/or
indirect    parameters   of   treatment   performance.
Performance monitoring will require ground-water sam-
ples collected  from monitoring wells that are appropri-
ately constructed and  strategically placed.  Sentry wells
may be used to assess contaminant transport after ISCO.
This is especially applicable to Fenton systems where var-
ious  enhanced transport mechanisms could result from
H2O2 injection (refer to Section II.C.2.H. Disadvantages).
Pre-  and post-oxidation monitoring of injection  wells
may provide  useful  information on treatment perfor-
mance. However, optimal ISCO treatment performance
will occur near the injection well. Consequently, ground-
water contaminant and  reaction byproduct concentra-
tions will be low, and high, respectively, in ground-water
samples collected from injection wells.  Ground-water
monitoring data representative of an injection well should
be qualified for this reason.

lll,E.4.b. Aquifer/Soil Sampling

The  majority of contaminant mass in source areas where
ISCO is deployed is present as NAPL or sorbed in the solid
phase.  Extraction and analysis of aquifer material and soil
samples collected immediately after oxidant consumption
may potentially provide rapid feedback on ISCO treatment
performance and spatial distribution of contaminants. A
delay in sampling for contaminant rebound is not necessary
for these samples.  Numerous pre- and post-oxidation soil
core samples, distributed horizontally and vertically,  are
often required  to  quantify the  spatial distribution  of
DNAPL  and to assess whether significant reduction  in
treatment was accomplished.  Accumulation of DNAPL
on distinct lithologic units  may  result in  a predictable
pattern of DNAPL  distribution.  Under this condition,
pre- and post-oxidation contaminant  characterization is
more straightforward and economically feasible.

III.E.4.C. Other

Due to the heat and O2(g) released during H2O2 injec-
tion,  volatilization  of organic contaminants  is highly
probable.  Therefore, performance monitoring at in-situ
Fenton oxidation sites may involve monitoring volatile
emissions.  The extent of volatilization during in-situ
Fenton oxidation has not been adequately investigated.
Volatile releases, especially from source areas containing
DNAPL or fuels, represent a high potential for volatiliza-
tion, exposure pathways, and health  and safety hazards.
Volatilization may also result from the sparging effect
that occurs during in-situ ozonation especially if air sparg-
ing/injection  is a component  in  the treatment process.
O3 itself is a hazardous oxidant and human contact and
inhalation could have serious health effects. Until  these
potential exposure  pathways can be documented  and
evaluated, volatile and O3 emissions should be measured,
controlled, and captured.  Correspondingly, soil vacuum
extraction, soil vapor, soil gas pressure, and off-gas moni-
toring may be  needed.  This should  be evaluated on a
case-by-case basis.

Elevated temperatures and pressures during H2O2 injec-
tion can result  in the release of steam and volatile emis-
sions during ground-water monitoring (i.e., opening  of
monitoring well caps).  To avoid this potential health
and safety hazard, sufficient time should elapse to allow
the release of O2(g), dissipation of pressure, and reduc-
tion in temperature.

III.E.4,d. Process and Performance Monitoring

Process and performance monitoring parameters are sum-
marized in Table 13.

III.E.5. Summary of Contaminant Transport and Fate
      Mechanisms during ISCO

During  ISCO, there  are  several transport  and  fate
mechanisms  that  may  occur   simultaneously  with
 In-Situ Chemical
                         Engineering Issue

-------
 Table 13. Process and Performance Monitoring Parameters
            Parameter
                          Purpose of the Data/Information
 Process Monitoring

 Target Contaminant


 Oxidants (H202, Mn04-, S2082-,03)


 Metals


 Fe, Phosphate, Chelators

 pH

 Alkalinity/Buffer Capacity

 Eh (electrode potential)

 Ground-Water Level
Existence or persistence of the target contaminant can be used to determine whereto design the
spatial delivery of the oxidant.

Estimate radius of influence of oxidant injection; Evaluate oxidant distribution; Calculate reaction
kinetics to help design monitoring program.

Assess whether redox- and/or pH-sensitive metals are mobilized; Assess attenuation of mobilized
metals.

Assess radial influence/distribution of injected reagents.

Assess whether pH is optimal; Assess impact of acid injection/pH modification.

Anticipate acid requirements for pH modification.

General indicator of oxidant distribution.

Assess hydraulic connection between injection monitoring wells and potential transport pathways.
 Performance Monitoring

 Target Contaminant
 Reaction Byproducts
Assess treatment performance via reduction in concentrations and/or mass in ground water,
aquifer material,and NAPL; Assess whether cleanup objective is being achieved and if the site can
be closed; Monitoring soil gas or off-gas reveals the fate and transport and potential exposure
pathways of volatile organics; Assess changes in the plume dimension.

Confirm the presence and transformation of the target compound(s); Estimate the mass of
contaminant transformed (for example, oxidation of chlorinated organic compounds releases Q-);
Evaluate the presence of sorbed and non-aqueous phases; Assess secondary MCLs.
oxidation. The  role  of  these  mechanisms should be
evaluated on a site-specific basis.  In some cases,  these
mechanisms can be desirable, i.e., enhanced mass transfer.
In other  cases, it is  not desirable (i.e.,  volatile losses,
exposure pathway, health risk, etc.), and appropriate steps
should  be  taken  to  minimize the  impact.  CO,  (g)
production during permanganate  oxidation in heavily
contaminated zones may mobilize NAPL and  transport
NAPL  vapors (Reitsma  and  Marshall,  2000). Fenton
oxidation involves the most complex range of transport
and fate mechanisms of the oxidants (Figure  8).  Thermal
effects resulting from the Fenton  reaction  enhance the
fate and transport mechanisms listed below.

Transport mechanisms
     •   Advective transport of contaminants in ground
        water results from ground-water displacement
        during injection, ground-water transport
        under natural  and/or an induced gradient, and
                                pneumatically-driven ground-water movement
                                from O2(g) production/expansion.
                             •  Diffusive transport of contaminants in ground
                                water from high to low concentration is
                                influenced by I SCO activities/mechanisms.
                             •  NAPL transport may result from changes in the
                                hydraulic gradient attributed to I SCO activities.
                             •  Advective and diffusive transport of volatile or-
                                ganics in the gas phase. Enhanced transport of
                                volatile emissions due to O2(g) evolution during
                                Fenton oxidation.

                        Fate mechanisms
                             •  Oxidation of dissolved, sorbed, and NAPL phase
                                organic contaminants (i.e., the treatment objective).
                             •  Transport of volatile compounds into the
                                unsaturated zone/atmosphere represents the
                                phase transfer of contaminants rather than a
                                destructive loss mechanism.
       Engineering Issue
                                               In-SIfu Chemical Oxidation

-------
Figure 8. Conceptual model of in-situ Fenton oxidation and potential fate and transport mechanisms. (A) Cross-section of
        hazardous waste site containing DNAPL in the saturated and unsaturated zones. Injection well is constructed in the
        source area and two monitoring wells located in the upgradient direction (downgradient monitoring wells not shown);
        (B) H2O2 is injected and reacts producing heat and O2(g). Contaminants are transformed via oxidation and other
        possible mechanisms (reductive transformation, hydrolysis). The pneumatic pressure from the O2(g) and from H2O2
        injection results in mounding of the ground water and displacement of the ground water away from the injection point.
        DNAPL movement, and enhanced volatilization of contaminants by O2(g) sparging + heat may also occur; (C) O2(g)
        sparging of the ground water in monitoring wells, artesian conditions, and continued ground-water displacement and
        enhanced volatilization may occur; (D) H2O2 injection ceases and is fully reacted. Loss of the target contaminant(s) in
        the source zone is achieved by oxidation transformation but may not be differentiated from other fate and transport
        mechanisms. Contaminant mass transfer and transport results in rebound.
       Biodegradation may occur under pre- and post-
       oxidation conditions.
       Sorption (adsorption, desorption) is impacted
       by changes in the contaminant concentrations in
       either the ground water and/or aquifer materials.
       Dissolution of organics from NAPL into the
       ground water.
       Hydrolysis (and possibly other abiotic reactions)
       is a destructive mechanism for organic
       contaminants.
III.E.6. Safety Issues

H2O2, MnO4", S2O82", and O3 are all strong oxidizing
agents and should be handled using appropriate methods
and personal protective equipment to prevent the risk of
chemical burns, fire, and explosions. Oxidant compatibil-
ity with all  materials  used in the remediation  process
should be reviewed and evaluated to minimize equipment
deterioration, leaks, and failure. Health and safety plans
(HASPs) should be reviewed by persons involved in the
 In-Situ Chemical Oxidation
                         Engineering Issue

-------
oxidant handling and other on-site ISCO procedures. A
Project Safety and Occupational Health Officer (or equiv-
alent) should review field operation plans and personal
protection equipment or other procedures that will pro-
tect worker safety and health while handling these oxidiz-
ers. Oxidant manufacturers provide up-to-date materials
on compatibility, handling, storage, and health and safety
information.   Such information is  easily  accessible on
manufacturers' websites. All on-site personnel should be
trained according to the requirements specified in 29 CFR
1910.1200  (h)  (OSHA's   Hazard   Communication
Standard) for the specific oxidizer.  Materials safety data
sheets (MSDS)  can provide useful information regarding
health and safety  issues  and should be on-site when the
work is performed.  Coordination with local emergency
response service providers will assure they are prepared in
the event of a spill/release emergency.

The work site should  be set up so the oxidant, whether
liquid or solid, will be  contained and  (1) will not mix
with organics or other incompatible material, or (2) flow
uncontrolled  into  the environment (i.e.,  lake, stream,
etc.) if an accident were to occur.  Protective safety equip-
ment including a portable eyewasher and shower  could
be used assuming an accidental exposure to oxidant(s) or
other reagents occurred.  These should  be set up on site
where oxidizers and other chemical materials will be han-
dled or potentially  contacted  by on-site  personnel.
Standards for protective safety equipment are available
through  the American  National Standards  Institute
(ANSI) (i.e., ANSI Z358.1-1998, Eyewash and Showers).
O3 generation  equipment   and associated  plumbing
should be operated and maintained in the open air or a
well-ventilated building/temporary structure  so that O3
from a  leak or improperly operated equipment cannot
build up to levels that will be hazardous to workers. The
storage of liquid and solid oxidizers should comply with
standards established  by the National Fire Protection
Association (NFPA 430: Code for the Storage of Liquid
and Solid Oxidizers) (NFPA, 2006).

A fatal dose of permanganate ingestion for an adult has
been estimated to be  10 g (Carus Chemical Company,
2004),  however, deleterious  effects will occur  from
ingestion of much lower doses. The characteristic pink
color of MnO4" can be observed  at low concentrations
(several mg/L), providing an easy detection method and
prevention of potential exposure pathways.  Precautions
should be taken to prevent inhalation  of KMnO4 dust
during  handling  and mixing. Commercially  available
permanganate mixing equipment can be used to  mini-
mize potential exposures to the oxidant. Skin and eye
contact, and ingestion of MnO4~ must also be avoided.
Unused oxidant  solution should  be neutralized by a
reductant such as sodium thiosulfate or sodium bisul-
fite.  Neutralization chemicals should be available when
the oxidant is delivered to the site.  Due to the poten-
tially violent reactivity (especially at higher sodium per-
manganate solution concentrations), the neutralization
should be  conducted  carefully, using adequate safety
precautions.  Serious  burns  to an  individual  resulted
from an   accident  in  2000  during  ISCO  at  the
Portsmouth Gaseous Diffusion Plant, Piketon, OH in
which  sodium thiosulfate was improperly added  to a
concentrated NaMnO4 solution.

Reactions involving  H2O2 are exothermic and release
large volumes of O2(g), especially at high H2O2 concen-
tration.  This reaction  (and O3 sparging)  can result in
enhanced volatilization and dispersal of organic vapors
which may represent an exposure risk or an explosion or
fire  hazard.  Specifically,  accumulation  of flammable
vapors  in basements, buildings, crawl spaces, etc., could
result in unacceptable indoor air quality and exposures to
humans, or explosion or fire hazards. In 1997, an explo-
sion and fatality was reported in Wisconsin at a residence
near a chemical oxidation project at a petroleum-contam-
inated site where  gasoline vapor migration had occurred
in sewer lines.

O3 has adverse respiratoiy effects, and exposure to harm-
fid levels must be avoided.  O3 injected during ISCO
should be fully reacted in the subsurface.  O3  released
into the air could be inhaled  by site workers or others in
the area.  O^ monitoring in air at the ISCO site should
be an integral component to  the routine monitoring to
prevent unacceptable  exposures. Additionally, O3  pro-
duction, storage equipment, and delivery lines should be
monitored routinely for leaks through appropriate detec-
tion and pressure  testing procedures.

Na2S2O8 can decompose in storage under conditions of
moisture and/or excessive heat, causing release of oxides
of sulfur and O2 that support combustion.  Decomposition
could result in high-temperature conditions.  Airborne
persulfate dust may be irritating to eyes, nose, lungs,
throat, and skin upon contact and may cause difficulty in
breathing (FMC,  2006).  A review of potentially incom-
patible ISCO  materials (with Na2S2O8) should be per-
formed prior to use of this oxidant.

III.E.7. Treatment Trains

Treatment   trains involving   other  technologies  used
before, during, or after ISCO can be used to enhance
treatment performance.
       Engineering Issue
                     fri-Slu Chemical Oxidation

-------
NAPL Removal: Removal of NAPL prior to ISCO is an
important first step in source area treatment and is gener-
ally more efficient and effective than ISCO.   This  will
decrease  the mass of oxidant required for contaminant
destruction. Additionally,  this will increase the NAPL
surface area for chemical oxidation, reduce the potential
mass transfer limitations that results from MnO2(s) depo-
sition at the NAPL interface, and reduce NAPL mobility
during ISCO.  Different technologies have been  devel-
oped  for the purpose of NAPL removal (Newell  et ai,
1995; API, 1996; U.S. EPA, 1996; U.S. EPA, 2001b).

Excavation: Soil excavation  of heavily contaminated soils
may be a viable option at some sites to reduce contami-
nant mass and leaching of contaminants to ground water.
The topographical depression in the subsurface may  also
serve as an infiltration trench for the delivery of oxidant
to contaminants residing at lower elevations.

Soil Vapor Extraction (SVE): SVE can  be used prior to
ISCO  to control, capture,  and remove volatile organics
from the unsaturated zone.  Prior to permanganate injec-
tion at the Union Chemical Company Superfund  Site
(South Hope, ME), SVE was used to remediate the unsat-
urated soils (Connelly, 2003). Remedial objectives  for
SVE are to remediate the unsaturated zone and to prevent
recontamination of  ground water  during/after  ISCO.
Due to the enhanced volatilization effects that occur dur-
ing in-situ Fenton oxidation and O3 sparging, SVE could
be  used  to prevent  volatile emissions, reduce exposure
pathways, and minimize uncontrolled loss and accumula-
tion of flammable vapors.   SVE can also provide  better
distribution of O3 in  the unsaturated zone.

Oxidation:Reduction:  Contaminant mixtures may be com-
prised of chemicals that are not entirely vulnerable to  oxi-
dation. Under this condition, another reaction mechanism
such as reductive transformation may be needed to fully
transform the  contaminant mixture.  Sequencing oxida-
tion and  reduction reactions, or, reduction and  oxidation
reactions, may  be needed.  Chemical reduction  using
dithionite followed by chemical oxidation (H2O2 and Fe)
was laboratory-tested  for a mixture of halogenated alkanes,
halogenated alkenes, and  aromatics  (Tratnyek  et  al.,
1998). Dithionite is a reductant that reduces Fe(III) to
Fe(II). Subsequently, carbon tetrachloride (CT) is reduc-
tively transformed via Fe(II), but is poorly oxidizable by
chemical  oxidants.   Ideally, dithionite  is used first to
reductively transform CT, and is followed by Fenton  oxi-
dation to transform the other contaminants. However, it
proved difficult to  achieve successful  results  with  the
sequential application of reductant and oxidant due to the
reaction between the  oxidant and residual reductant. The
sequential application of reductants  and oxidants, and
vice versa, involves a broad range of competing redox reac-
tions. Ideally, this approach has a wide range of applica-
tion, but is not currently a well-developed technology.

Bioremediation:  Sequencing  oxidation  and  reduction
reactions may be achieved  through biological reductive
treatment. Electron donor reagents (i.e., hydrogen releas-
ing compounds, lactate, vegetable oil, etc.) injected into
the subsurface can biologically produce reducing condi-
tions and reductive transformations. One site where car-
bon sources were injected to create an anaerobic reductive
dechlorination environment after  permanganate ISCO
was the Union Chemical Company Superfund Site near
South Hope, ME (Connelly, 2003; ITRC, 2005).

Enhanced or naturally occurring reductive transformation
of oxidation-resistant compounds may occur downgradi-
ent from the ISCO area provided that sufficient separation
can be achieved to allow the predominant redox zones to
develop.  Post-oxidation enhanced bioremediation may
include aerobic biodegradation or aerobic cometabolic bio-
degradation  with  addition  of co-substrates.  Site-specific
feasibility testing is necessary to address the  uncertainties
of sequential oxidation and bioremediation.

It  is anticipated that in  nearly all  cases, natural attenua-
tion will be an integral component to ISCO because it is
not economically feasible for ISCO alone to achieve the
low cleanup standards  specified at many sites for  the
source area,  and/or for the entire plume (refer to Section
III.F.5.  Impact of ISCO  on Natural Attenuation and
Biodegradation).

Oxidant  Combinations: H2O2 is a reactant of S2O82" that
produces the sulfate radical and may be injected as an acti-
vator during in-situ persulfate oxidation. Injected H2O2, a
low cost oxidant, will react widi naturally occurring Fe(II)
to  form -OH and oxidize contaminants,  and will oxidize
reduced aquifer materials, thus lowering the  TOD. This
pre-treatment step would lower the reaction rate and oxi-
dant demand for MnO.^ and may be more cost effective.

The sequential injection of  persulfate and permanganate
was intended to satisfy the NOD with S2O82~ and oxidize
the VOCs with MnO4- (Droste etal., 2002). The intent
of this approach was to minimize the amount of MnO4~
required  to  meet the treatment objectives, and subse-
quently to minimize the accumulation of MnO2(s) and
the potential  for permeability loss.   However,  it was
unclear  to what extent  TCE, DCE  and VC were oxi-
dized.  The use of combining oxidants  requires further
investigation, since the impact of persulfate on NOD is
 In-Situ Chemical
                        Engineering Issue

-------
unclear.  Others have reported that persulfate is relatively
unreactive  toward naturally  occurring  organic  matter
(Brown and Robinson, 2004).

Ground-Water Pumping: Ground-water recirculation (and
oxidant replenishment)  or  simply ground-water  extrac-
tion can be used during ISCO to enhance the transport
rate and direction of the injected oxidant. Ground-water
pump-and-treat may be used  to control  and contain the
migration of contaminants  and the  injected oxidant, to
improve delivery of the oxidant, or to prevent migration
to potential receptors.

Thermal Treatment: Thermal treatment, such as radio
frequency heating, is  likely to be necessary to heat the
subsurface  sufficiently so that thermally-activated persul-
fate oxidation can be effective (Liang etal., 2001).  The
use of solar energy may also  serve to economically heat
the ground water in warm climate regions.

III.F. Limitations/Interferences/Impacts

III.F.l. Untreated COCs/Rebound

Organic  chemicals that  are poorly reactive  with specific
oxidants will  persist  in the  ground  water (Table 2).
However, these compounds may be vulnerable to other
remediation  technologies   (refer to  Section  III.E.7.
Treatment Trains) or may attenuate naturally with accept-
able risk. Persistence of the organic chemicals may also be
due to insufficient delivery of oxidant. Causes of insuffi-
cient delivery include (1) reactive transport and consump-
tion of the oxidant prior to fully reaching the target zone,
(2) underestimating the total oxidant demand, and (3)
delivery of oxidant to non-targeted zones.  Good site char-
acterization and performance monitoring are needed to
determine why contaminants  persist and to take steps to
assure adequate oxidant delivery.

Contaminant  rebound involves the  condition in which
contaminant concentrations in the presence of oxidants
are low or nondetectable, but steadily increase (rebound)
in the ground water  after  oxidant concentrations  have
diminished.  Rebound is attributed  to extended periods
of slow mass transfer and mass transport mechanisms of
the residual contaminants.  These mechanisms include
the slow dissolution  of contaminants  from  NAPL or
through MnO2(s) precipitate  on NAPL,  slow desorption
from aquifer materials, slow advective transport in ground
water,  and slow diffusive transport of  contaminants
usually from low-permeability materials. If monitoring
wells are not located near the source zone/ISCO area, the
prolonged  time required   for ground-water  transport,
sampling, and detection of contaminants can be delayed
and  contribute to rebound. In Fenton systems,  O2(g)
entrapped  in  porous  media  may  also  interfere  with
ground-water flow, mass transport, or mass transfer.

Few sites where ISCO has  been  implemented, if any,
have achieved  the treatment objectives in a single  appli-
cation. Because of the high probability of rebound, mul-
tiple applications should be  budgeted and planned.  In
general, this involves an iterative approach of monitoring
diagnostics and  reapplications of the oxidant (refer to
Section III.E.2. General Conceptual Approach). Rebound
underscores the  importance  of establishing an efficient
monitoring  well  network, long-term monitoring,  and
multiple oxidant applications.

III.F.2. Toxic Reaction Byproducts
                                                  in
Metals:  An  increase in heavy metals concentration
ground  water  may  result from heavy metals impurities
contained in  the  permanganate, and mobilization of
pre-existing  redox- or pH-sensitive heavy metals (in-situ)
by the oxidant.  Field investigations generally reveal  that
these metals attenuate through various mechanisms  and
within acceptable  transport distances (refer to Section
Il.C.l.f. Metals Mobilization/Immobilization).

Organics: Reaction  byproducts are  generally less  toxic,
more biodegradable, and more mobile than the parent
compound.  For example, reaction byproducts from the
oxidation of MTBE by Fenton's or persulfate include  tert-
butyl formate, TEA, and acetone (refer to Section II.C.2.b.
Contaminant Transformations).  Ketones (e.g., acetone)
and  alcohols are reaction byproducts from the oxidation
of petroleum hydrocarbons  by MnO4' (Fatiadi, 1987).
These byproducts may not be acceptable and should be
monitored and evaluated on a site-specific basis. Further
transformation of reaction byproducts is possible, assum-
ing sufficient oxidant is available.  Enhanced  biodegrada-
tion and/or natural attenuation  may be feasible  and
acceptable under some conditions.

Other: Nitrate (NO3~) is one of the  nitrogen byproducts
from the oxidation of high  explosives HMX and RDX
(Zoh and Stenstrom, 2002). Oxidation of high concen-
trations  of these contaminants  could potentially result in
the accumulation of NO3' in excess of the U.S. EPA max-
imum concentration level (10 mg/L,  as nitrogen).

III.F.3. Process Residuals

Process residuals from KMnO4 can include a sludge  that
accumulates in  mixing, storage,  or distribution tanks.
       Engineering Issue
                     In-Itu Chemical Oxidation

-------
The sludge may contain water-insoluble sand/silica solids
that are additives to "free-flowing" grade of dry oxidant
granules to prevent clumping during handling. The sludge
may also include MnO2(s) and particulate permanganate
(KMnO4(s)). Causes for KMnO4(s) include  (1) inade-
quate mixing of die oxidant solution, (2) exceeding the
solubility  of MnO4~ during preparation,  and  (3) reduc-
tions in temperature during storage (i.e.,  this  lowers the
KMnO4  solubility, resulting in precipitation  reactions).
Injection or precipitation of KMnO4(s) can cause permea-
bility reductions in or near die injection well.  However,
given  sufficient time,  KMnO4(s)  will dissolve into the
aquifer. Redevelopment of a well may be needed to restore
the permeability where the responsible mechanism is not
reversible under ambient conditions.

III.F.4. Geochemical Impacts

The immediate geochemical impact of injecting oxidants
is to increase the oxidation state of the aquifer.  Oxidation
reactions change the solubility of many inorganic species,
resulting in the precipitation of soluble mineral species.
The four most common reduced inorganic species are Fe,
Mn, As,  and sulfides  (Brown and Robinson,  2004).
Long-term post-oxidation rebound in reducing condi-
tions,  possibly  through  microbially-driven  reactions,
results  in the dissolution of precipitated (solid phase)
minerals.  The extent to which this pattern of redox con-
ditions and geochemical reactions occurs is dependent on
site-specific conditions.  The elements O, S, Fe, and Mn
are injected at high concentrations during ISCO and are
predominant participants in ground-water redox pro-
cesses.  Secondary effects from the injection of these oxi-
dants  include  the  geochemical  impact by  enhanced
microbial activity  (refer to  Section III.F.5.  Impact of
ISCO  on Natural Attenuation  and  Biodegradation).
Numerous site-specific  inorganic  and organic reactants
exist in the subsurface that strongly influence a wide range
of geochemical outcomes. It is  beyond the scope of this
Engineering Issue Paper to address the complex and broad
nature  of geochemical  impacts of ISCO.  Probable chem-
ical reactions and byproducts involving these elements are
covered in detail elsewhere (Stumm and Morgan, 1996).
The long-term geochemical impact of these oxidants and
reagents has not been well-documented.

III.F.4.a, Oxidants

After oxidation is complete (i.e., MnO4" reacted), MnO2(s)
and Mn+2  are  the   predominant  manganese species.
MnO2(s) can be found under a wide pH  range (pH 2 to
14) and will likely be the predominant form of manganese.
MnO9(s) can dissolve under reduced conditions (pH <8),
resulting in an increase in Mn+2 in the ground water. The
persistence of MnO2(s) or the impact of Mn+2 in aquifers
have not been rigorously established; however, dissolution
under reducing conditions in sediment/surface water sys-
tems has  been reported (Hem,  1985; Drescher  et al,,
1998). At pH >8, various manganese solid phase minerals
(oxides, hydroxides, carbonate species) form.  Mobilized
manganese species (Mn+2, colloidal MnO2(s)) could be an
aesthetic concern in drinking water if recovered in  water-
supply wells  (refer  to  Section Il.C.l.g. EPA  Drinking
Water Standard). MnO2(s) may impact NAPL mass trans-
fer and permeability, and serve as a sorbent for heavy met-
als  (refer  to  Sections  II.C.I.e.  Impact  of  MnO2(s)
Accumulation  and  Il.C.l.f.   Metals   Mobilization/
Immobilization).

In general, in-situ Fenton  oxidation involves injection
of large quantities of H9O2, injection of various chemi-
cal reagents,  and release of heat and O2(g)  in the sub-
surface. After H2O2  is fully reacted, the predominant
residuals in  the injection  zone are dissolved oxygen,
O2(g), and heat.  Similarly, in O3 oxidation, after O3 is
fully reacted, the predominant residuals are O2(g)  and
dissolved oxygen.  In  either case, O2(g) will slowly dis-
solve into the ground water and maintain elevated dis-
solved oxygen concentrations and  redox  conditions in
the injection  area.  With Fenton oxidation, the heat will
slowly dissipate and ultimately  the site will return to
ambient temperature.  Although it is unknown how long
these conditions will persist, the potential geochemical
impact is site-specific  and involves many potential reac-
tions and  mechanisms (refer to Section II.C.2.e.  O2(g)
Generation and Exothermic Reaction; Section II.C.2.E
Injected Reagents; and Section III.F.5. Impact of ISCO
on Natural Attenuation and Biodegradation).

Persulfate  oxidation results in  high  concentrations of
SO42" in the  aquifer.  Under reducing conditions,  SO42"
can be reduced to sulfide (HS~).  Both SO42" and HS" are
highly soluble and  mobile in ground  water.  Elevated
concentrations of these species in  ground water  could
exceed the secondary  drinking water standard (refer to
Section III.D.3.  Regulatory Constraints on Injection of
Reagents).  In calcium-rich environments, the mineral
gypsum (CaSO4»2H2O) may form  which is a relatively
insoluble form of sulfate.

III.F.4.b. Natural Organic Matter

Permanganate is most reactive with  natural organic mat-
ter (NOM) of the four oxidants used in ISCO, and per-
sulfate is relatively unreactive towards NOM (Brown and
Robinson, 2004). The role of natural organic material in
 In-Situ Chemical
                        Engineering Issue

-------
Fenton-driven oxidation reactions has not been systemati-  IH-F.4.C. pH
cally investigated  and remains uncertain.  Watts  et al.
(1990) found that the ratio of pentachlorophenol (PGP)  pH is a master variable in geochemical equilibrium and
and  H2O2 consumption rates in soil  suspensions was  can  be significantly impacted  during  I SCO either  by
inversely related to NOM, suggesting that NOM success-  injection of an acid reagent or by the acidity/alkalinity
fully competed with PCP (and H7O2) for -OH.  In a dif-  produced by  chemical  reactions.  A decline in pH  is
ferent study involving  FeSO4 and H2O2 addition, the  generally undesirable due to the potential for enhanced
oxidation of benzo(a)pyrene was inhibited by the presence  transport  of  some  pH-sensitive  metals  under  acidic
of non-target organics  including glucose, cellulose, and  conditions.  The magnitude, direction, and permanence
lignin (Kelley etaL, 1990).  Conversely, humic acids pro-  of the pH change are dependent on the buffer capacity of
moted O, decomposition and -OH formation, and any  the aquifer material, the amount and type of contaminant
activity of humic  acids as -OH scavengers was  offset by  oxidized, and the mass of oxidant and/or acid injected.
the enhanced formation of free radicals (Masten, 1990).  All of these parameters are site-specific, suggesting that
Fenton-driven degradation  of 2,4,6-trinitrotoluene was  pH changes that occur during ISCO can be varied.
greater in a system amended with fulvic acid  than in
another containing humic acid.   Enhanced kinetics was  Permanganate oxidation can affect  the pH differently
attributed to accelerated Fe(III) reduction (Li etaL, 1997).  depending on the target analyte. Oxidation of PCE and
Fenton oxidation  and -OH production was enhanced in  TCE will lower the pH by release of H+; and oxidation of
the presence of peat by one or more peat-dependent mech-  DCE and VC raises the pH by release of OH- (Table 3,
anisms (Hiding^/., 2001). Fe concentration and avail-  reactions 4 to 7).  The direction of pH  change could
ability in the  peat, reduction of Fe(III) to Fe(II) by the  therefore increase or decrease depending on the type and
organic  matter,  and  reduction  of organic-complexed  quantity of contaminants present (Siegrist et al., 2001,
Fe(IlI) to Fe(II) were  probable causes.  Contaminated  and references therein).
aquifer material containing this type of organic material
may exhibit similar mechanisms that enhance Fenton oxi-          PCE:        8 moles Hf /3 moles PCE
dation.  Humic material (i.e., NOM) can facilitate elec-          TCE:         1  mol H+/1 mol TCE
tron  transfer  for Fe(III) reduction in microbial systems          DCE:       2 mol OH-/3 mol DCE
(Lovley etaL,  1996; Scott etaL, 1998), and similar  redox          VC:         7 mol OH-/3 mol VC
coupling can be provided by constituents of humic  mate-
rials (quinones, hydroquinones) in Fenton systems (Chen  Reduction in  pH is consistently observed as a result of
and Pignatello,  1997).   Such reactions may  provide an  ISFO.  pH modification during ISFO can  result  from
additional mechanism  for  Fe(III) reduction to Fe(II),  injecting  acid to enhance Fenton oxidation  (refer  to
resulting in more efficient-OH production.               Section  II.C.2.f(3)  Acidification),  from  the   Fenton
                                                     reactions  (Table 4, reactions 1  to  3),  and from the
It has been proposed that NOM is oxidized during ISCO,  oxidation of organic compounds.  Reduction in pH may
resulting in the release of sorbed contaminants.  An  also occur from  the  oxidation of organic compounds  by
apparent release of RDX from the sorbed phase was mea-  S2O82~ (Liang etaL, 2001; Huang etaL, 2002) or O3.
sured during  KMnO4  oxidation (Struse et al., 2002b).
Increased  bioavailability of chlorinated compounds by  III.F.4.d. Cation Exchange Capacity
NOM oxidation was proposed for the observed increased
rate of biological reductive dechlorination (Droste  etaL,  The cation exchange capacity (CEC) is a measure of the
2002). A portion  of sorbed PAHs were released from the  aquifer material's ability to adsorb exchangeable cations.
oxidation  of  NOM by S2O82~  (Cuypers et  al., 2000).  Common cations include  H+,  K+,  Na+,  NH4+, Ca+2,
Additionally,  a  lag time was reported before TCE and  Mg+2, and Al+3.  Changes in  ground-water chemistry
TCA oxidation  by S,O82" and was attributed to the oxi-  caused by the injection of oxidants and from  various
dation of soil organic carbon followed by contaminant  chemical reactions can impact the CEC.  For example,
transformations (Liang etaL, 2001; 2003). Overall, these  injection  of  NaMnO4,  KMnO/,,  or Na2S2O8  at  high
studies suggest that NOM oxidation corresponds with a  concentration may displace some of the common cations
release (desorption) of compounds, especially those less  or possibly heavy metals. In porous  media containing a
amenable to oxidation, such as TCA.   Consequently, a  high  percentage of clay,  displacement of cations  by Na+
temporary increase in contaminant concentrations  could  (i.e.,  NaMnO,,}, Na2S2O8)  could  contribute   to  the
result, but would  decline from the application of addi-  dispersion  of soil particles, elimination of macropores,
tional oxidant.                                        and deterioration of soil structure  making the media
       Engineering Issue
fri-Slu Chemical Oxidation

-------
impervious to water penetration. Although reduction in
permeability  is rarely measured/reported in ISCO, the
potential consequences of injecting high concentrations
of sodium-based oxidants and excessive Na+ in clay-rich
environments should be evaluated.

III.F.5. Impact of ISCO on  Natural Attenuation and
      Biodegradation

III.F.S.a. Impact on Natural Attenuation

Natural attenuation is expected to play an important role
in the overall  remedy at  most  sites where  ISCO  is
deployed. Permanent inhibition of microbial activity by
the injected oxidants is undesirable since biotic processes
are  an  integral  component in  natural  attenuation.
However, H2O9,  MnO4~, O3, and  S?O82~ are antiseptics
and  will  inhibit or kill microorganisms at much lower
concentrations  than   are  typically  used  in  ISCO.
Additionally, oxidant-induced changes in redox potential
and  pH  can  also inhibit some microbial species.  The
post-oxidation decline  in the diversity of microbial spe-
cies suggests that  some microorganisms are also sensitive
to pH (Kastner et al., 2000)  or elevated redox potential
and become inactive or die. Historically, H2O2 injection
has been problematic for enhanced bioremediation  due
to rapid H2O2 decomposition, microbial toxicity, limited
solubility of  O7,  loss of  O2(g) to  the unsaturated zone
(Spain et al, 1989; Hiding et al,  1990; Pardiek et al.,
1992), reduction  in permeability (Weisner et al., 1996),
and excessive heat.

Short-term laboratory oxidation studies involving com-
plete-mix soil slurry batch reactors, and flow-through col-
umn  studies  allow  complete hydraulic control  and
excellent contact  between oxidant,  aquifer material,  and
microbes.  These testing  conditions result in  a  high
impact of the oxidant on microbial activity (Hrapovic et
al., 2005).  Laboratory studies conducted in this manner
provide insight into the potential effect of the oxidant on
microbial activity. However, they do not fully represent
the nonideal mechanisms  and long-term  time frames
associated with ISCO under field conditions that strongly
influence microbial survival and activity under harsh oxi-
dative conditions. Differences between  laboratory  and
field  conditions  help explain  discrepancies   between
microbial inhibition  results from laboratory studies  and
the seemingly low impact of ISCO on microbial activity
at field-scale.
at hem-scale.

Preferential pathways in heterogeneous  porous media
result in hydraulic short-circuiting of the injected oxi-
dants.  Hydraulic short-circuiting and microniches pre-
 In-Situ Chemical
vent full  contact  between the oxidant and microbes,
providing shelter and permitting the survival of microor-
ganisms   during   rigorous  applications  of oxidants.
Laboratory oxidation studies involving complete-mix soil
slurry batch reactors allow excellent contact between oxi-
dant, aquifer  material,  and microbes, which generally
results in a high impact of the oxidant on microbial activ-
ity.  Laboratory investigations  are useful for several rea-
sons and  provide  insight to the potential effect of the
oxidant on microbial activity, but they do not fully repre-
sent the nonideal mechanisms under field conditions that
strongly influence  microbial survival under harsh oxida-
tive conditions.

In sequential  H2O2-driven  oxidation and  biodegrada-
tion,  more extensive  PAH  and PCP degradation was
measured than from  biodegradation  alone.  An initial
decline in the  microbial population occurred after H2O2
(1 to 2%) was applied, but was followed by a significant
increase a week later that suipassed the original microbial
numbers  (Allen and Reardon, 2000).  In a field study
where large volumes and high  concentrations of H2O2
were injected,  the abundance and activity of microorgan-
isms declined,  but  rebounded in six months (Chapelle et
al., 2005). Microbial activity in TCE- and  cis-l,2-DCE-
contaminated  soil and ground water was measured before
and after treatment with KMnO4 (11,000  gal., 0.7%)
(Klens etal., 2001). Two weeks after injection, the redox
potential  of the ground water was >800 mV and viable
populations of anaerobic  heterotrophs,  rnethanogens,
and nitrate- and sulfate-reducing microbes were present,
but at lower levels  than under  pre-oxidation conditions.
Three months  after injection, the nitrate-reducing micro-
bial populations had increased. Six months after injec-
tion, the  redox potential of the ground water was about
100 mV, MnO4" was absent, and the aerobic heterotroph
population in  ground water had  increased  by several
orders of magnitude greater than the pre-oxidation  pop-
ulation.  In other studies involving KMnO4, the rate of
biological reductive  dechlorination  of TCE increased
after treatment (Rowland etal., 2001), and  no changes in
the site's  microbial community structure were measured
(Azadpour-Keeley  et al., 2004).  In  three case studies
involving the  injection of high concentrations and/or
large quantities of  permanganate, ISCO did  not sterilize
the aquifer, nor was microbial activity permanently inhib-
ited (Luhrs et al., 2006). At one site, reductive dehaloge-
nation of CVOCs  following biostimulation with sodium
lactate, and at  the other two sites, a significant increase in
the post-oxidation  microbial biomass, and the post-oxi-
dation presence of a viable and diverse microbial consor-
tia capable of degrading a wide range of organic chemicals,
were measured.
                                                                               Engineering Issue

-------
Sequential injection of S2O82~ and MnO4~ was originally
proposed to  satisfy  the  natural oxidant  demand with
S2O82", oxidize the VOCs with MnO4~, and minimize the
amount of MnO4"  required and  the accumulation of
MnO2(s) (Droste et  ai, 2002).  It was unclear to what
extent TCE, DCE, and VC were oxidized; however, mon-
itoring data including reductive dechlorination daughter
products, Cl" mass  balance, and measurements of H2,
SO42', dissolved  iron,  and  phospholipid  fatty acids,
strongly suggested that enhanced reductive dechlorination
had occurred.  It was proposed this could have resulted
from  (1) increasing  the  amount of SO42" and  sulfate-
reducing bacterial activity, (2) increased microbial activity
due to simpler substrates from VOC  oxidation, and (3)
increased VOC bioavailability from the oxidation of sor-
bate (natural organic  matter) (Droste etal., 2002).

ISCO is often deployed in a source area which is gener-
ally small in size relative to the ground-water plume that
extends downgradient.  In the source area where the oxi-
dant is injected, direct contact between the oxidant and
microorganisms  is possible and could inhibit microbial
activity.  Additionally, the elevation  in redox potential
can inhibit reductive dehalogenation activity until termi-
nal electron accepting conditions shift back to iron- and
sulfate-reducing  conditions.  Consequently,  in systems
where anaerobic and  reductive transformations play a sig-
nificant role, the post-oxidation impact of ISCO on nat-
ural attenuation would be greater and  sustained than
aerobic  conditions.   Reaction of the oxidant near the
injection location (source area(s)) lessens the downgradi-
ent impact of the oxidant on microbial activity where
polishing effects of natural attenuation occur.

A  localized decline in microbial activity will result from
direct contact between the oxidant and microbe, or from
the highly oxidizing conditions. Microbes that are sensi-
tive to oxidative treatments will decline in population and
activity, while others  that are insensitive to the change in
redox potential may be unchanged or may respond favor-
ably.  Due to spatial separation between the source area
where the oxidant is injected and downgradient areas, the
impact of the oxidant may be low or nonexistent in down-
gradient microbially active areas. The length of time for
"microbial rebound"  is unclear, but given sufficient time
after ISCO, microbial populations, microbial activity, and
the rate of biodegradation increase, possibly to levels above
pre-oxidation conditions. Proposed theories for increased
bioactivity include improved bioavailability of trace con-
stituents, lower concentration  of challenging chemicals,
more  available simple substrate resulting  from contami-
nant or natural organic matter oxidation, less competition
(with  other microbiota) for available nutrients and  sub-
strate, die off of microbial predators (Allen and Reardon,
2000), elevated temperatures, and more available terminal
electron acceptors (TEAs).  No cases  were found where
aquifer material was sterilized or where microbial activity
had been permanently inhibited.

III.F.S.b. Mechanisms Which Potentially Enhance
       Biotically-Driven Natural Attenuation

Microorganisms obtain energy and carbon for new cell
material through biochemical redox reactions in which
electrons are transferred  from organic contaminants to
terminal electron acceptors.   Under aerobic conditions,
oxygen is the most energetically favorable TEA.  However,
due to the  low solubility of oxygen, dissolved oxygen
(DO) is rapidly depleted in ground water.  Subsequently,
anaerobic conditions may result where the biochemical
oxidation  of  organic compounds  continues  to occur
(Lovley and Phillips, 1986; Suflita etal., 1988; Hutchins
et al., 1998).  The sequential order of TEA utilization
under anaerobic conditions is nitrate (NO3~), manganese
(Mn(IV)), ferric iron  (Fe(III)), sulfate (SO42'), and car-
bon dioxide (CO2).  The TEAs O2,  NO3', SO42', and
CO2 are generally found in the aqueous phase.  Fe and
Mn species and reducible organic matter are the primary
sources  of solid-phase TEA  in aquifers  (Heron et al.,
1994). The total TEA (aqueous + solid phase) available
in the subsurface is predominantly attributed to the solid-
phase fraction, and the aqueous phase constitutes a minor
fraction (Huling etal., 2002). Currently, the reducibility
of aquifer organic matter  is poorly understood  and
unquantifiable, and manganese species typically only con-
tribute 2 to 5% of the total transferable electron equiva-
lents.  Ferric iron is found in abundance and is present in
the solid phase since it is slightly soluble in the near neu-
tral pH range.  In sulfate-rich environments, SO42" may
also be derived from aquifer sediments.

There are several mechanisms in which ISCO could be
beneficial  to  natural attenuation. The exothermic reac-
tion of H2O2 will raise the temperature and increase bio-
activity.  The injection of phosphate stabilizers during
ISFO by some vendors introduces phosphorus, an essen-
tial element in the production of  microbial  energy and
new cell material. Acidification resulting from the injec-
tion of acid or as a reaction byproduct may lower the pH
and increase bioavailability of some microbial  nutrients.
The injection of each oxidant results  in the  addition of
various TEAs including DO from  H2O2  and O3; SO42"
from S2O82"; and Mn^ from MnO4". The injection of Fe
during ISFO also  contributes to the  total TEA, but  is
expected to be minor in most cases relative to the quantity
of naturally occurring Fe. Oxidation of reduced aquifer
       Engineering Issue
                     fri-Slu Chemical Oxidation

-------
sediments via H2O2 proceeds rapidly relative to microbi-
ally mediated decomposition (Barcelona and Holm, 1991;
Korom et ai, 1996).   Each of the oxidants involved in
ISCO can oxidize  the reduced TEAs (i.e., Fe+2, Mn+2,
etc.) in the aquifer sediments. Therefore, while oxidant
injection is intended for immediate contaminant oxida-
tion and could result in short-term, localized microbial
inhibition, it also introduces TEAs into the aquifer and
oxidizes aquifer sediments. At one site, a shift from sul-
fate-reducing to  Fe(III)-reducing conditions  resulting
from H2O2  injection may have decreased the efficiency
of  reductive  dechlorination in  the  injection  zone
(Chapelle etal., 2005). However, it could also shift the
predominant terminal electron accepting process from an
inefficient one (methanogenesis)  to more efficient  pro-
cesses such as aerobic biodegradation and/or Fe, Mn, and
SO42" reduction (Huling et al., 2002) and provide a sus-
tained long-term source of TEA.

III.G. Summary

Over the last 10 years, significant development of ISCO
has dramatically advanced the state of the science, state of
the practice, and  the effectiveness  of this technology.
ISCO has been used at thousands of hazardous waste sites
in the U.S. and is the  fastest growing subsurface remedial
technology used today. Wide application of this technol-
ogy under variable  site conditions has provided valuable
field experience.   Such field experience in conjunction
with continued research and development will improve
both the state of the practice and state of the science.
These improvements  will  include the identification  of
site-specific parameters needed to assess the feasibility of
ISCO, the development of predictive tools that allow an
improved assessment of the potential benefits and adverse
effects prior to ISCO deployment.  Ultimately, this will
contribute to more effective and efficient use of the tech-
nology and lower remedial costs.
IV.  ACKNOWLEDGMENTS

This Engineering Issue Paper was prepared for the U.S.
Environmental Protection  Agency, Office of Research
and Development, National Risk Management Research
Laboratory.  The authors  were  Dr. Scott G. Huling
(Ground Water and Ecosystems  Restoration  Division,
National Risk Management Research Laboratory, Office
of Research and Development, U.S. EPA) and Dr. Bruce
E. Pivetz (Dynamac Corporation). The authors acknowl-
edge the peer review comments provided by Dr. Robert
Norris,  Dr. Ben Shiau, and the U.S. EPA Engineering
and Ground-Water Forum members. The authors also
thank Ms. Carol House (Dynamac  Corporation)  for
desktop publishing  and graphics  preparation support,
and Ms. Kathy Tynsky (Computer Sciences Support) for
graphics preparation support.

For additional information, contact:
   Dr. David Burden, Director
   Ground Water Technical Support Center
   Ground Water and Ecosystems Restoration Division
   National Risk Management Research Laboratory
   Robert S. Kerr Environmental Research Center
   P.O.Box  1198
   Ada, OK 74820
   (580) 436-8606

   Mr. David Reisman, Director
   Engineering Technical Support Center
   Land Remediation and Pollution Control Division
   National Risk Management Research Laboratory
   26 W Martin Luther King Drive
   Cincinnati, OH 45268
   (513) 569-7588
V.  NOTICE/DISCLAIMER

The  U.S. Environmental Protection Agency through its
Office of Research and Development partially funded and
collaborated in the research described here under Con-
tract No. 68-C-02-092 to  Dynamac Corporation. This
document has been subjected to the Agency's peer and
administrative review and has been approved for publica-
tion  as an EPA document.  Mention of trade names or
commercial products does not constitute endorsement or
recommendation for use.
VI.  QUALITY ASSURANCE STATEMENT

All  research projects making conclusions or recommen-
dations based on environmentally-related measurements
and funded by the Environmental Protection Agency are
required to participate in the Agency Quality Assurance
(QA)  program.  This project did not involve physical
measurements and as such did not require a QA plan.
VII. ACRONYMS, ABBREVIATIONS, AND SYMBOLS
(g)
(s)
f|
AEB
          solid
          oxidation efficiency
          Air Force Base
 In-Situ Chemical
                        Engineering Issue

-------
aq        aqueous
BTEX    benzene, toluene, ethylbenzene, xylene
CB       chlorobenzene
CEC      cation exchange capacity
CERCLA Comprehensive Emergency Response,
          Compensation, and Liability Act
CF       chloroform
CHP      catalyzed hydrogen peroxide
cm       centimeter
COC     contaminant of concern
COD     chemical oxygen demand
CT       carbon tetrachloride
CVOC    chlorinated volatile organic compound
DCA     1,1 -dichloroethane
DCE      dichloroethylene
DNAPL   dense nonaqueous phase liquid
DO       dissolved oxygen
EDTA    ethylenediaminetetraacetic acid
Eh        oxidation-reduction potential
EPA      U.S. Environmental Protection Agency
EPCRA   Emergency Planning and Community Right
          to Know Act
Eqn       equation
foc        fraction of organic carbon
ft         foot
g         gram
gpm      gallons per minute
HMX     octogen, or cyclotetramethylenetetra-
          nitramine (high explosive)
hr        hour
I SCO     in-situ chemical oxidation
ISFO     in-situ Fenton oxidation
kg        kilogram
kWh      kilowatt-hour
Ib         pound
L         liter
LNAPL   light nonaqueous phase liquid
m         meter
MC       methylene chloride
MCL     maximum contaminant level
mg       milligram
MGP     manufactured gas plant
min       minute
mL       milliliter
MNA     monitored natural attenuation
mol       gram molecule
MTBE    methyl tertiary butyl ether
mV       millivolt
n         porosity
NAPL    nonaqueous phase liquid
NOD     natural oxidant demand
NOM     natural organic matter
NTA     nitrilo-triacetic acid
°C        degrees Celsius
ORNL    Oak Ridge National Laboratory
OSC      on-scene coordinator
OSHA    Occupational Safety and Health
          Administration
OU      operable unit
PAH      polynuclear aromatic hydrocarbon
PCB      polychlorinated biphenyl
PCE      perchloroethylene
PCP      pentachlorophenol
pH       negative log of hydrogen ion concentration
pKa      negative logarithm of the acid dissociation
          constant, K,a
ppm      parts per million
psi        pounds per square inch, unit of pressure
PVC      polyvinyl chloride
RCRA    Resource Conservation and Recoveiy Act
RDX     cyclotrimethylenetrinitramine, (cyclonite),
          (hexogen), (T4), high explosive
RPM     remedial project manager
rxn       reaction
scfm      standard cubic feet per minute
SDWA    Safe Drinking Water Act
sec        second
SMCL    secondary maximum contaminant level
SVE      soil vapor extraction
TEA      tertiary butanol
TCA      1,1,1-trichloroethane
TCE      trichloroethylene
TEA      terminal electron acceptor
TOC     total organic carbon
TOD     total oxidant demand
U.S. EPA United States Environmental Protection Agency
UIC      Underground Injection Control
UV      ultraviolet
V        volt
VC       vinyl chloride
VOC     volatile organic compound
WEF      Water Environment Federation
wt        weight
VIII. REFERENCES

ANL. 2002. Human Health Fact Sheet, Potassium-40.
Argonne National  Laboratory,  http://www.ead.anl.gov/
pub/doc/potassium.pdf (Accessed 6/7/04).

APHA, AWWA, and WEF. 1989. Standard Methods for
the Examination of Water and Wastewater. Indigo Colon-
metric Method (Method No. 4500-O^ B). Clesceri, L.S.,
A.E Greenberg, and  R.R.  Trussell (Eds.), 17* edition,
pp. 3-102 to 3-106.
       Engineering Issue
                     In-S/fu Chemical Oxidation

-------
API. 1996. A Guide to the Assessment and Remediation of
Underground Petroleum Releases, 3rc1 Edition, API Publi-
cation  1628, Washington, DC.

Aggarwal, P.K. and R.E. Hinchee. 1991. Monitoring in-
situ biodegradation of hydrocarbons by using stable car-
bon  isotopes.  Environmental  Science  &  Technology.
25(6):1178-1180.

Ahlfeld, D.P., A. Dahmani, and W. Ji.  1994. A concep-
tual model of field  behavior of air sparging and its impli-
cations for application.  Ground  Water Monitoring  &
Remediation. Fall, 14(4):132-139.

Allen,  S.A.  and K.F. Reardon. 2000.  Remediation  of
contaminated soils by combined chemical and biological
treatment.  In: Physical and Thermal  Technologies,  The
Second International Conference on Remediation of Chlori-
nated  and Recalcitrant  Compounds.  Wickramanayake,
G.B. and A.R, Gavaskar (Eds.), Battelle Press, Columbus,
OH. pp. 301-306.
line. Texas  Research  Institute. API  Publication 4389,
Washington, DC.

Brooks, M.C.,  W.R.  Wise,  and M.D. Annable. 1999.
Fundamental changes in in  situ air  sparging flow pat-
terns. Ground Water Monitoring & Remediation. Spring,
19(2):105-H3.

Brown, R.A. 2003. In Situ Chemical Oxidation: Perfor-
mance, Practice, and Pitfalls.  AFCEE Technology Trans-
fer Workshop, February 24-27, 2003, San Antonio, TX.
http://www.afcee.brooks.af.mil/products/techtrans/
workshop/postworkshop03/Tuesday/pm/sourcezonere-
mediation/4 brown.pdf

Brown, R.A. and D. Robinson. 2004. Response to natu-
rally occurring organic  material:  Permanganate versus
persulfate. In: Proceedings of the Fourth International Con-
ference on Remediation  of Chlorinated and Recalcitrant
Compounds. Gavaskar, A.R,  and  A.S.C.  Chen. (Eds.),
Battelle Press, Columbus, OH. Paper  2A-06.
Azadpour-Keeley, A., L.A. Wood, T.R. Wood, and S.C.   Buxton, G.V., C. Greenstock, W.P. Hellman, and A.B.
Mravik. 2004.  Microbial  responses to in-situ chemical   Ross. 1988. Critical review of rate constants for reactions
oxidation, six-phase heating, and steam injection remedi-   of hydrated electrons, hydrogen atoms, and hydroxyl rad-
ation technologies in groundwater. Remediation. Autumn   icals (-OH/-O")  in aqueous solution. Journal of Physical
(Fall), 14(4):5-17.                                      and Chemical Reference Data. 17(2):513-886.
Balba, M.T., F. Blickle, D. Coons, G. Hotchkiss, C. Lin,
and A.F. Weston. 2002. Soil remediation by potassium
permanganate. In: Proceedings of the Third International
Conference on Remediation of Chlorinated and Recalcitrant
Compounds. Gavaskar, A.R,  and  A.S.C. Chen (Eds.),
Battelle Press, Columbus, OH. Paper 2C-23.

Barcelona, M.J. and T.R. Holm. 1991. Oxidation-reduc-
tion capacities of aquifer solids. Environmental Science &
Technology. 25(9): 1565-1572.

Berkowitz, B. 2002. Characterizing flow and transport in
geologic media: A review. Advances in Water Research.
25:861-884.

Black, H. 2001. Ozone as a cleanup tool. Environmental
Science & Technology. 35(13):283A-284A.

Braida, W.J. and S.K. Ong. 2001.  Air sparging effective-
ness:  Laboratory  characterization of  air-channel  mass
transfer zone for VOC volatilization. Journal of Hazard-
ous Materials. B87:241-258.

Britton, L.N. 1985. Feasibility Study on the Use of Hydro-
gen Peroxide to Enhance Microbial Degradation of Gaso-
Cambridge, M., and R, Jensen. 1999. In situ ozonation
to degrade recalcitrant compounds. Tech Trends. 33:3-4.
U.S. Environmental Protection Agency, Office of Solid
Waste and Emergency Response, Washington, DC. EPA
542-N-99-003.

Carus Chemical Company. 2004. iMaterial Safety Data
Sheet for CAIROX® Potassium Permanganate, http://
www.caruschem.com/pdf/new files/CAIROX MSDS.
pdf

Chambers, J., A.  Leavitt, C. Walti,  C.G. Schreier, J.
Melby,  and L. Goldstein. 2000a.  In  situ destruction of
chlorinated solvents with KMnO4 oxidizes chromium.
In: Chemical Oxidation and Reactive Barriers, The Second
International Conference  on Remediation of Chlorinated
and Recalcitrant Compounds. Wickramanayake,  G.B.,
A.R. Gavaskar, and A.S.C. Chen (Eds.), Battelle Press,
Columbus, OH. pp. 49-55.

Chambers, J., A. Leavitt,  C. Walti, C.G. Schreier, and J.
Melby.  2000b.  Treatability   study:  Fate of chromium
during  oxidation of chlorinated solvents.  In: Chemical
Oxidation and Reactive Barriers, The Second International
Conference on Remediation of Chlorinated and Recalcitrant
 In-Situ Chemical
                        Engineering Issue

-------
Compounds. Wickramanayake, G.B., A.R. Gavaskar, and   ence on Remediation of Chlorinated and Recalcitrant Com-
A.S.C. Chen (Eds.), Battelle Press, Colunibus, OH. pp. 57-   pounds. Gavaskar, A.R. and A.S.C.  Chen (Eds.), Battelle
65.                                                  Press, Columbus, OH. Paper 1A-13.
Chapelle,  F.H., P.M. Bradley, and C.C.  Casey. 2005.
Behavior of a chlorinated ethene plume following source-
area treatment with Fenton's reagent. Ground Water
Monitoring & Remediation. Spring, 25(2): 131 -141.

Chen, C.T., A.N. Tafuri, M.  Rahman,  M.B. Foerst, E.
Pfetzing, and M. Taylor. 1995. Oxidation of methyl-t-
butyl ether (MTBE) using Fenton's reagent. Presented at
Air and Waste Management Association,  San Antonio,
Texas, June 18-23, 1995. 95-WA91.03, p. 16.

Chen, R. and J.J. Pignatello. 1997. Role of quinone inter-
mediates as electron shuttles in Fenton and photo-assisted
Fenton oxidations of aromatic compounds. Environmen-
tal Science & Technology. 31(8):2399-2406.

Choi H., H. Lira, J. Kim, T. Hwang, and J. Kang. 2002.
Transport  characteristics of gas phase ozone  in unsatu-
rated porous media for in-situ chemical oxidation. Journal
of Contaminant Hydrology. 57(1 -2):81 -98.

Clancy, P.B., J. Armstrong, M. Couture, R, Lussky, and
K. Wheeler. 1996. Treatment of chlorinated ethenes in
groundwater with ozone and  hydrogen  peroxide. Envi-
ronmental Progress. 15(3):187-193.

Connelly, T. 2003. SVE and ISCO used after pump and
treat for multimedia removal of VOCs. Technology News
and Trends. Issue 9. U.S. Environmental  Protection
Agency. EPA 542-N-03-006.

Conrad, S.H., R.J. Glass,  and WJ. Peplinski. 2002.
Bench-scale visualization of DNAPL  remediation pro-
cesses in analog heterogeneous aquifers: Surfactant floods
and  in  situ oxidation using permanganate, journal of
Contaminant Hydrology. 58:13-49.

Crimi, M.L. and R.L. Siegrist. 2003. Geochemical effects
on metals following permanganate oxidation of DNAPLs.
Ground Water. 4l(4):458-469.

Cuypers, C., T. Grotenhuis, J. Joziasse, and W. Rulkens.
2000. Rapid persulfate oxidation predicts PAH bioavail-
ability in soils and sediments. Environmental Science &
Technology. 34 (10): 20 57-2063.
Damm, J.H., C. Hardacre, R.M. Kalin, and K.P. Walsh.
2002. Kinetics of the oxidation of MTBE by potassium
permanganate. Water Research. 36:3638-3646.

Dorfman, L.M. and G.E. Adams.  1973. Reactivity of the
Hydroxyl Radical. National Bureau of Standards, Report
No. NSRDS-NBS-46.

Dowideit, P. and C. von Sonntag.  1998. Reaction  of
ozone with ethene and its methyl- and  chlorine-substi-
tuted derivatives in aqueous solution. Environmental Sci-
ence 6- Technology. 32(8): 1112-1119.

Drescher,  E., A.R. Gavaskar, B.M. Sass, L.J. Gumming,
M.J. Drescher, and T.K.J. Williamson. 1998. Batch and
column testing to evaluate chemical oxidation of DNAPL
source  zones. In:  Physical,  Chemical,  and  Thermal
Technologies,   The  First  International   Conference  on
Remediation of Chlorinated and Recalcitrant Compounds.
Wickramanayake, G.D. and R.E. Hinchee (Eds.), Battelle
Press, Columbus, OH. pp. 425-432.

Droste, E.X., M.C.  Marley, J.M. Parikh, A.M.  Lee,
P.M.  Dinardo, B.A.  Woody, G.E.  Hoag,  and  P.V.
Chheda. 2002. Observed enhanced reductive dechlori-
nation  after in situ chemical oxidation pilot test. In:
Proceedings  of the Third International  Conference  on
Remediation of Chlorinated and Recalcitrant Compounds.
Gavaskar,  A.R. and A.S.C. Chen (Eds.), Battelle Press,
Columbus, OH. Paper 2C-01.

Elder, C.R. and C.H. Benson. 1999. Air channel forma-
tion, size,  spacing, and  tortuosity during air sparging.
Ground Water Monitoring  & Remediation.  Summer,
Fatiadi, A. J. 1987. The classical permanganate ion: Still
a novel oxidant in organic chemistry. Synthesis: Journal of
Synthetic Organic Chemistry. 2:85-127.

FMC.  2005.  Bulletin  1 -General Efficacy Chart.  FMC
Environmental Resource Center, Environmental Solutions.
http://envsolutions.fmc.com/Klozur8482/tabid/356/
Default.aspx
Dai, Q. and S. Reitsma. 2002. Two-dimensional experi-   FMC. 2006. Persulfates Technical Information, http://
mental  studies of permanganate  flushing of pooled   www.fmcchemicals.com/Content/CPG/Images/AOD
DNAPL. In: Proceedings of the Third International Confer-   Brochure Persulfate.pdf
       Engineering Issue
                     fri-S/fu Chemical Oxidation

-------
Fu, G, H.E. Allen, and C.E. Cowan. 1991. Adsorption
of cadmium and copper by manganese oxide. Soil Science.
152(2):72-81.

Gardner, F.G., N.  Koite,  J.  Strong-Gunderson,  R.L.
Siegrist, O.K. West, S.R. Cline, and J. Baker. 1998. Imple-
mentation of Deep Soil Mixing at the Kansas City Plant.
ORNL/TM-13532.

Glaze, W.H. and J.W. Kang. 1988. Advanced oxidation
process for treating groundwater contaminated with TCE
and PCE: laboratory studies. Journal of American Water
Works Association. 88(5):57-63.

Haag, W.R. and C.C.D. Yao.  1992. Rate constants for
reaction of hydroxyl radicals with several drinking water
contaminants.  Environmental Science  &  Technology.
26(5):1005-1013.

Haselow, J.S.,  R.L.  Siegrist, M. Crimi, and T.  Jarosch.
2003. Estimating the  total  oxidant demand for in situ
chemical  oxidation   design.  Remediation.  Autumn,
13(4):5-16.

Hein, G.L., J.S.  Gierke, N.J. Hutzler, and R.W. Falta.
1997. Three-dimensional experimental testing of a  two-
phase flow-modeling approach for air sparging. Ground
Water Monitoring & Remediation. Summer, 17(3):222-
230.

Hem,  J.D. 1985. Study  and  Interpretation  of the
Chemical Characteristics of Natural Water, 3rd Edition.
U.S. Geological  Survey Water-Supply  Paper 2254.
263 pp.

Heron,  G., T.H.  Christensen, and  J.C.  Tjell. 1994.
Oxidation capacity of aquifer sediments.  Environmental
Science 6" Technology. 28(1): 153-158.

Hinchee, R.E., D.C. Downey, and P.K. Aggarwal. 1991.
Use of hydrogen peroxide as an oxygen source for in situ
biodegradation. Part I. Field studies, journal of Hazard-
ous Materials. 27(3):287-299.

Hoigne, J. and H. Bader. 1976. The role of hydroxyl rad-
ical reactions in ozonation processes in aqueous solutions.
Water Research. 10:377-386.

Hoigne, J.  and H. Bader. 1979a. Ozonation of water:
Oxidation-competition values of different types of waters
used in  Switzerland.  Ozone: Science and Engineering.
1:357-372.
Hoigne, J.  and H. Bader. 1979b. Ozonation  of water:
Selectivity and rate of oxidation of solutes. Ozone: Science
and Engineering. 1:73-85.

Hrapovic, L.A., B.E. Sleep, D.J. Major, and E. Hood.
2005. Laboratory study of treatment of trichloroethene
by chemical oxidation followed by  bioremediation.
Environmental Science  &  Technology.  39(8):2888-
2897.

Huang, K.C., R.A. Couttenye, and G.E. Hoag. 2002.
Kinetics of heat-assisted persulfate oxidation of methyl
feTt-butyl ether (MTBE). Chemosphere. 49(4):413-420.

Hiding, S.G., B.E.  Bledsoe,  and M.V. White.  1990.
Enhanced Bioremediation Utilizing Hydrogen Peroxide
as a Supplemental Source of Oxygen: A Laboratory and
Field Study. EPA/600/2-90/006.

Hiding, S.G., R.G.  Arnold,  R.A. Sierka,  M.A. Miller.
1998. Measurement of hydroxyl radical activity in a soil
slurry using the spin trap CC-(4-pyridyl-l-oxide)-N-tert-
butyl-nitrone. Environmental  Science  &  Technology.
32(21):3436-344l.

Huling, S.G., R.G. Arnold, P.K. Jones, and R.A. Sierka.
2000. Predicting the rate of Fenton-driven 2-chlorophe-
nol transformation using a contaminant analog. Journal
of Environmental Engineering. 126(4):348-353.

Huling, S.G., R.G.  Arnold,  R.A. Sierka,  M.A. Miller.
2001. Influence  of peat on  Fenton oxidation.  Water
Research. 35(7): 1687-1694.

Huling, S.G., B. Pivetz, and R. Stransky. 2002. Terminal
electron acceptor mass balance:  NAPLs  and  natural
attenuation.  Journal  of Environmental  Engineering.
128(3):246-252.

Huling, S.G. and B.  Pivetz. 2003. Total oxidant demand
testing and  results, former Southern Solvents, Inc. Super-
fund Site OU#1, Tampa, Hillsborough County, Florida,
18pp.

Huling, S.G., P.K. Jones, W.P. Ela, and R.G. Arnold.
2005. Fenton-driven chemical regeneration of MTBE-
spent  granular   activated  carbon.   Water  Research.
39:2145-2153.

Hutchins, S.R., D.E. Miller,  and A. Thomas.  1998.
Combined laboratory/field study on the use of nitrate for
in-situ bioremediation  of a fuel-contaminated aquifer.
Environmental Science dr Technology. 32(12): 1832-1840.
 In-Situ Chemical
                        Engineering Issue

-------
ITRC. 2005.  Technical and Regulatory Guidance for In
Situ  Chemical Oxidation  of Contaminated Soil  and
Groundwater 2nd  Edition,  ISCO-2. Washington, DC:
Interstate Technology & Regulatory Council, http://
www.itrcweb.org/gd  ISCO.asp

Jensen, J., J.  Dablow,  B.  Marvin, and D. Grasmick.
1999.  Application  of full scale  in-situ  ozonation to
degrade recalcitrant PAH compounds at a former MGP
site. In: Proceedings of the 12th International Symposium on
Environmental and Site Remediation Technologies, and Gas
Industry Environmental Issues and Strategies. Institute of
Gas Technology.

Johnson, P.C. 1998. Assessment of the contributions of
volatilization and  biodegradation to in-situ  air sparging
performance.  Environmental  Science  &   Technology.
32(2):276-281.

Kakarla, P.K.C. and R.J. Watts.  1997. Depth of Fenton-
like oxidation in remediation of surface soil. Journal of
Environmental Engineering.  123(1):11-17.

Kakarla,  P.K., T. Andrews, R.S.  Greenberg, and D.S.
Zervas. 2002. Modified Fenton's  processes  for effective
in-situ chemical oxidation-laboratory  and field evalua-
tion. Remediation. Autumn, 12(4):23-36.

Kastner, J.R., Domingo, J.S., Denham, M,  M. Molina,
and R. Brigmon. 2000.  Effect of chemical oxidation on
subsurface microbiology and trichloroethylene biodegra-
dation. Bioremediation Journal. 4(3):219-236.

Kelley, R.L., W.K.  Gauger, and V.J.  Srivastava. 1990.
Application  of Fenton's  Reagent as a Pretreatment Step
in Biological Degradation  of Polyaromatic Hydrocar-
bons. Institute of Gas Technology. 16 pp.

Kim, J. and  H. Choi. 2002. Modeling in situ ozonation
for the remediation  of  nonvolatile  PAH-contaminated
unsaturated  soils. Journal  of Contaminant Hydrology.
55:261-285.

Klens, J., D. Pohlmann,  S. Scarborough, and D. Graves.
2001. The effects of permanganate oxidation on subsur-
face microbial populations. In:  Natural Attemiation of
Environmental Contaminants,  The Sixth International In
Situ and On-Site Bioremediation  Symposium.  Leeson, A.,
M.E.  Kelly,  H.S. Rifai, and V.S. Magar (Eds.). Battelle
Press, Columbus, OH. pp. 253-259.

Korte, N.E., J. Skopp, W.H. Fuller, E.E. Niebla,  and
B.A. Alshii. 1976. Trace element movement  in soils:
Influence of soil chemical and physical  properties. Soil
Science. 122:350-359.

Korom,  S.F.,  MJ.  McFarland, and  R.C. Sims. 1996.
Reduced sediments: A factor in the design of subsurface
oxidant delivery. Ground Water Monitoring & Remedia-
tion. Winter, 16(1):100-105.

Li, H., L. Lee, D. Schulze, and C. Guest. 2003. Role of
soil manganese in the oxidation of aromatic amines. Envi-
ronmental Science & Technology. 37(12):2686-2693.

Li, X.D. and F.W. Schwartz. 2000. Efficiency problems
related to permanganate oxidation schemes. In: Chemical
Oxidation and Reactive Barriers, The Second International
Conference on Remediation of Chlorinated and Recalcitrant
Compounds. Wickramanayake, G.B., A.R, Gavaskar, and
A.S.C. Chen (Eds.), Battelle Press, Columbus, OH. pp.
41-48.

Li, X.D. and F. Schwartz. 2004a. DNAPL  remediation
with in situ chemical oxidation using potassium perman-
ganate. Part I. Mineralogy of Mn oxide  and its dissolu-
tion in organic acids. Journal of Contaminant Hydrolology.
68:39-53.

Li, X.D. and F.W. Schwartz. 2004b. In situ chemical
oxidation of contaminated ground  water: Permanent
reactive barrier systems for the  long-term treatment of
contaminants.  The 227th  American  Chemical  Society
Meeting, Anaheim, CA.

Li, Z.M., P.J. Shea, and S.D. Comfort. 1997. Fenton
oxidation of 2,4,6-trinitrotoluene in  contaminated soil
slurries.  Environmental Engineering  Science.  14(1):55-
66.

Liang, C., C.J.  Bruell, M.C. Marley, and K. Sperry. 2001.
Kinetics of thermally activated persulfate oxidation of tri-
chloroethylene (TCE) and 1,1,1-trichloroethane (TCA).
Conference  on Oxidation  and Reduction Technologies
for In-Situ Treatment of Soil and Groundwater. Niagara
Falls, Ontario,  Canada, June 25-29, 2001.

Liang, C., C.J. Bruell, M.C. Marley, and K.L. Sperry.
2003. Thermally activated persulfate oxidation of trichlo-
roethylene  (TCE) and 1,1,1-trichloroethane (TCA) in
aqueous systems and soil slurries. Soil and Sediment Con-
tamination. 12(2):207-228.

Liang, C., C.J. Bruell, M.C. Marley, and K.L. Sperry.
2004a. Persulfate oxidation for in situ  remediation of
TCE. I. Activated by ferrous ion with  and without a per-
       Engineering Issue
                            Chemical Oxidation

-------
sulfate-thiosulfate  redox  couple.  Chemosphere.  55(9):
1213-1223.

Liang, C., C.J. Bruell, M.C.  Marley, and  K.L. Speny.
2004b.  Persulfate oxidation for in situ remediation of
TCE. II. Activated by chelated ferrous ion. Chemosphere.
55(9):1225-1233.

Lipczynska-Kochany, E., S. Gregor, and S. Harms. 1995.
Influence of some groundwater and surface waters con-
stituents on the degradation  of 4-chlorophenol  by the
Fenton reaction. Chemosphere. 30(1):9-20.

Lovley,  D.R. and  E.J.P.  Phillips.  1986. Availability of
ferric iron for microbial reduction in bottom sediments
of the freshwater tidal Potomac river. Applied and Envi-
ronmental Microbiology. 52(4):751-757.

Lovley, D.R., J.D. Coates, E.L. Blunt-Harris, E.J.P. Phil-
lips, and J.C. Woodward. 1996.  Humic  substances as
electron  acceptors  for  microbial  respiration. Nature
382:445-448.

Luhrs, R, R. Lewis, and S.G. Huling. (In press 2006). In-
situ permanganate oxidation: Impact on aquifer permea-
bility, natural attenuation, and contaminant variability.
In: Proceedings  of the Fifth International Conference on
Remediation of Chlorinated and Recalcitrant Compounds.
Battelle Press, Columbus, OH.

MacKinnon, L.K. and N.R. Thomson. 2002. Laborato-
ry-scale in situ chemical  oxidation of a perchloroethy-
lene pool using permanganate. Journal of Contaminant
Hydrology. 56(l-2):49-74.

Masten, S.J.  1990. Ozonation of VOCs in the presence of
humic acid  and soils. Ozone: Science  and Engineering.
13:287-312.

Masten, S.J. and S.H.R. Davies. 1997. Efficacy of in-situ
for the remediation of PAH contaminated soils, journal
of Contaminant Hydrology. 28:327-335.

McKay, D.J. and C.M. Berini. 2002. A comparison of
permanganate delivery methods in an unsaturated set-
ting. In: Proceedings of the Third International Conference
on  Remediation  of Chlorinated  and Recalcitrant  Com-
pounds.  Gavaskar, A.R. and A.S.C. Chen (Eds.), Battelle
Press, Columbus, OH. Paper 2H-47.

McLean, J.E. and B.E. Bledsoe. 1992. Behavior of Metals
in Soils. U.S. EPA Ground Water Issue Paper. EPA/540/
S-92/018. 25pp.
Mitani, M.M., A.A. Keller, C.A. Bunton, R.G. Rinker,
and  O.C.  Sandall.   2002.  Kinetics and  products  of
reaction  of MTBE with  ozone and ozone/hydrogen
peroxide in  water, journal  of Hazardous  Materials.
689:197-212.

Maughon,  M.J.,  C.C. Casey, J.D.  Bryant,  and J.T.
Wilson. 2000. Chemical oxidation source reduction and
natural  attenuation  for  remediation  of chlorinated
hydrocarbon in groundwater. In: Physical and Thermal
Technologies,  The Second  International  Conference  on
Remediation of Chlorinated and Recalcitrant Compounds.
Wickramanayake,  G.B.  and  A.R.  Gavaskar  (Eds.).
Battelle Press, Columbus, OH. pp. 307-314.

Mumford,  K.G.,  C.S. Lamarche, and N.R. Thomson.
2004.  Natural  oxidant demand of aquifer  materials
using the push-pull technique, journal of Environmental
Engineering. 130(10): 1139-1146.

Nelson, C.H. and R.A. Brown. 1994. Adapting ozonation
for soil  and ground  water  cleanup.  Environmental Engi-
neering,  a Supplement  to  Chemical Engineering. EE20-
EE24.

Newell, C., S. Acree, R. Ross, and S. Huling. July, 1995.
Light Non-Aqueous Phase Liquids. U.S. EPA Ground
Water Issue Paper. EPA/540/S-95/500.

NFPA. 2006. National  Fire Protection Association web-
site. http://www.nfpa.org

Nico, P. and R.  Zasoski. 2000. Importance of Mn(III)
availability  on the rate of Cr(III) oxidation on G-MnO2.
Environmental Science & Technology. 34(l6):3363-3367.

Nyer, E.K.  and  D. Vance.  1999.  Hydrogen peroxide
treatment: The good, the  bad, the  ugly.  Ground Water
Monitoring & Remediation.  Summer, 19 (3): 54-57.

Palmer, C.D. and R.W. Puls. 1994.  Natural Attenuation
of Hexavalent Chromium  in Ground Water and Soils.
U.S. EPA  Ground  Water  Issue  Paper.  EPA/540/S-
94/505.

Pardieck, D.L.,  E.J. Bouwer,  and A.T. Stone.  1992.
Hydrogen peroxide  use to  increase  oxidant capacity for
in situ bioremediation of contaminated soils and aqui-
fers:  A  review.  Journal  of Contaminant Hydrology.
9(3):221-242.

Parker, B.L. 2002. Full-scale permanganate remediation
of a solvent DNAPL source zone in a sand  aquifer. Pre-
 In-Situ Chemical
                        Engineering Issue

-------
sented at the EPA Seminar, In Situ Treatment of Ground-
water Contaminated with Non-Aqueous Phase Liquids,
Chicago,  Illinois,  http://cluin.org/studio/napl  121002/
agenda.cfm

Petrie, R., P. Grossl, and R.  Sims. 2002. Oxidation  of
pentachlorophenol  in   manganese  oxide  suspensions
under controlled Eh and pH environments. Environmen-
tal Science 6- Technology. 36(17):3744-3748.

Peyton, G.R., O.J. Bell, E. Girin, and M.H. Lefaivre.
1995. Reductive destruction of water contaminants dur-
ing treatment with hydroxyl radical processes. Environ-
mental Science 6" Technology. 29(6)1710-1712.

Pignatello, J.J. 1992. Dark and photoassisted Fe3+-cata-
lyzed degradation of chlorophenoxyl herbicides by hydro-
gen  peroxide.  Environmental  Science &   Technology.
26(5):944-951.

Pignatello, J.J. and K. Baehr.  1994.  Ferric complexes  as
catalysts for Fenton degradation of 2,4-D and  metola-
chlor in soi\. Journal of Environmental Quality. 23(2):365-
370.

Pignatello, J.J.  and M.  Day.  1996. Mineralization  of
methyl parathion insecticide in soil by hydrogen peroxide
activated widi iron(III)-NTA or -HEIDA complexes. Haz-
ardous Waste and Hazardous Materials. 13(2):237-244.

Reddy, K.R., S.  Kosgi, and J. Zhou. 1995.  A review  of
in-situ air sparging for the remediation  of VOC contam-
inated saturated soils and groundwater. Hazardous Waste
and Hazardous Materials. 12(2):97-118.

Reitsma, S. and Q. Dai. 2000. Reaction-enhanced mass
transfer from NAPL pools. In: Chemical Oxidation and
Reactive Barriers, The Second International Conference on
Remediation of Chlorinated and Recalcitrant  Compounds.
Wickramanayake, G.B., A.R Gavaskar, and A.S.C. Chen
(Eds.). Battelle Press, Columbus, OH. pp. 33-40.

Reitsma, S. and M. Marshall.  2000.  Experimental study
of oxidation of pooled NAPL. In: Chemical Oxidation and
Reactive Barriers, The Second International Conference on
Remediation  of Chlorinated and Recalcitrant  Compounds.
Wickramanayake, G.B., A.R Gavaskar, and A.S.C. Chen
(Eds.). Battelle Press, Columbus, OH. pp. 25-32.

Reitsma, S. and J. Randhawa. 2002.  Experimental inves-
tigation of manganese dioxide plugging of porous media.
In: Proceedings of the Third International Conference on
Remediation of Chlorinated and Recalcitrant  Compounds.
Gavaskar, A.R. and A.S.C. Chen (Eds.). Battelle Press,
Columbus, OH. Paper 2C-39.

Roden, E.E. and J.M. Zachara. 1996. Microbial reduc-
tion of crystalline iron(III) oxides: Influence of oxide sur-
face area and  potential  for cell growth.  Environmental
Science 6- Technology. 30(5):1618-1628.

Rowland, M.A., G.R. Brubaker, K. Kohler, M. Westray,
and D. Morris. 2001. Effects of potassium permanganate
oxidation on subsurface  microbial activity. In: Anaerobic
Degradation of Chlorinated Solvents, The Sixth International
In Situ and On-Site Bioremediation Symposium. Magar, V.
S., D.E. Fennell, J.J. Morse, B.C. Alleman, and A. Leeson
(Eds). Battelle Press, Columbus, OH. 6(7): 1-12.

Rutherford, K.W.  and P.C. Johnson. 1996. Effects of
process control  changes on  aquifer oxygenation rates
during in situ air sparging  in homogeneous aquifers.
Ground'WaterMonitoring & Remediation. Fall, 16(4):132-
141.

SalanitroJ.P., P.C.Johnson, G.E. Spinnler, P.M. Maner,
H.L. Wiseniewski, and C. Bruce. 2000. Field-scale dem-
onstration of enhanced MTBE bioremediation through
aquifer bioaugmentation  and  oxygenation. Environmen-
tal Science 6" Technology. 34(19):4l52-4l62.

Schumb, W.C.,  C.N. Satterfield, and RL. Wentwoith.
1955.  Hydrogen Peroxide. American Chemical  Society
Monograph Series, Reinhold Publishing Corporation
New York, NY. 759 pp.

Scott,  D.T.,  D.M.  McNight, E.L.  Blunt-Harris,  S.E.
Kolesar, and D.R. Lovley. 1998. Quinone moieties act as
electron acceptors in the  reduction of humic substances
by  humics-reducing  microorganisms.  Environmental
Science 6- Technology. 32(19):2984-2989.

Shacklette,  H.T. and J.G. Boerngen. 1984. Elemental
Concentration in Soils and Other Surficial Materials of
the Conterminous United States. U.S. Geological Survey
Professional Paper 1270, 105 pp.

Siegrist, RL., K.S. Lowe, L.C. Murdoch, T.L. Case, and
D.A.  Pickering. 1999.  In situ  oxidation  by  fracture
emplaced reactive solids. Journal of Environmental Engi-
neering. 125(5):429-440.

Siegrist, RL., M.A. Urynowicz, O.R. West, M.L. Crimi,
and K.S. Lowe.  2001. Principles and Practices of In  Situ
Chemical Oxidation Using Permanganate.  Battelle Press,
Columbus, OH.
       Engineering Issue
                            Chemical Oxidation

-------
Siegrist, R.L.,  M.A. Uiynowicz, M.L. Crimi,  and K.S.
Lowe.  2002. Genesis and effects of particles produced
during in situ chemical oxidation using permanganate.
Journal of  Environmental Engineering.  128(11): 1068-
1079.

Spain,  J.C.,  J.D.  Milligan,  D.C.  Downey,  and J.K.
Slaughter. 1989.  Excessive bacterial  decomposition  of
H>O2  during enhanced biodegradation. Ground Water.
27(2): 163-167.

Sperry, K.L., M.C. Marley, C.J. Bruell, C. Liang, and J.
Hochreiter.  2002. Iron catalyzed persulfate oxidation of
chlorinated solvents. In: Proceedings of the Third Interna-
tional Conference on Remediation of Chlorinated and Recal-
citrant Compounds. Gavaskar, A.R., and A.S.C. Chen.
(Eds.). Battelle Press, Columbus, OH. Paper 2C-22.

Stefan, M.I. and J.R. Bolton.  1999.  Reinvestigation  of
the acetone  degradation  mechanism  in dilute aqueous
solution by the UV/H9O7 process. Environmental Science
& Technology. 33(6):870-873.

Stefan  M.L, J. Mack, and J.R Bolton.  2000. Degrada-
tion pathways during the treatment of methyl  tert-buty\
ether by the  UV7 H2O2 process. Environmental Science &
Technology. 34(4):650-658.

Struse, A.M., RL. Siegrist, H.E.  Dawson, and  M.A.
Uiynowicz.  2002a. Diffusive transport of permanganate
during in situ oxidation. Journal of Environmental Engi-
neering. 128(4):327-334.

Struse, A.M., B.K.  Marvin,  S.T.  Harris, and  W.S.
Clayton. 2002b. Push-pull  tests: field evaluation of in
situ chemical oxidation of high explosives at the Pantex
plant.  In: Proceedings of the  Third International Confer-
ence on Remediation of Chlorinated and Recalcitrant Com-
pounds. Gavaskar, A.R.  and A.S.C. Chen (Eds.), Battelle
Press, Columbus, OH. Paper 2G-06.

Stumm, W.  and J.J. Morgan. 1996. Aquatic Chemistry -
Chemical Equilibria and Rates in Natural Waters, 3rd edi-
tion. A Wiley-Interscience Publication, John Wiley and
Sons, Inc. 1022 pp.

Suarez, D.L. and Langmuir, D. 1976. Heavy metal rela-
tionships in  a Pennsylvania soil. Geochimica et Cosmochi-
micaActa. 40(6):589-598.

Suflita, J.M.,  SAGibson, and RE. Beemann.  1988.
Anaerobic biotransformation of pollutant chemicals  in
aquifers. Journal oflndustial Microbiology. 3:179-194.
Sun, Y. and J.J. Pignatello. 1993. Organic intermediates
in the degradation of 2,4-dichlorophenoxyacetic acid by
Fe3+/H2O2 and Fe3+/H2O2/UV. Journal of Agricultural
and Food Chemistry. 41:1139-1142.

Tratnyek, P.G., T.L.  Johnson, S.D.  Warner,  H.S.
Clarke, and J.A. Baker.  1998.  In situ treatment of
organics  by sequential  reduction and  oxidation. In:
Physical,  Chemical,  and Thermal Technologies, The First
International Conference on Remediation of Chlorinated
and  Recalcitrant Compounds. Wickramanayake,  G.B.
and  R.E.  Hinchee (Eds.),  Battelle Press, Columbus,
OH. pp.  371-376.

U.S. Department of Defense.  1999. Technology Status
Review:  In  Situ Oxidation.  Environmental  Security
Technology Certification Program. 42 pp.

U.S. Department of Energy. 1997. Final Report for
Demonstration of  In   Situ  Chemical  Oxidation  of
DNAPL  Using the Geo-Cleanse Technology. WSRC-
TR-97-00283, 23 pp.

U.S. EPA. 1989. Standard Methods for the Examination
of Water and  Wastewater. Method 4500-O^  B. Indigo
Colorimetric Method. L.S.  Clesceri,  A.E  Greenberg,
and  R.R. Trussell, (Eds.),  17rh edition, pp. 4-162 to
4-165.

U.S. EPA. 1992. Secondaiy  Drinking  Water Regula-
tions: Guidance for Nuisance  Chemicals. Office of
Water. EPA 810/K-92-001. http://www.epa.gov/safewa-
ter/consumer/2ndstandards.html

U.S. EPA. 1996. How to Effectively Recover Free Prod-
uct at Leaking Underground Storage Tank Sites. Office
of Solid Waste and Emergency Response, National Risk
Management  Research  Laboratory,  Cincinnati, OH
45268. EPA 510-R-96-001.

U.S. EPA. 2001. The DNAPL  Remediation Challenge: Is
There a Case for Source Depletion? Office of Research
and Development, National Risk Management Research
Laboratory, Cincinnati, OH 45268. EPA/600/R-03/143,
lllpp.

U.S. EPA. 2002. National Primary Drinking Water Stan-
dards. Office of Water. EPA 816-F-02-013. http://www.
epa.gov/safewater/mcl.html#mcls

Vella, P.  et al. 2005. Standard Test Method for Deter-
mining the Permanganate Soil Oxidant Demand (Screen-
ing Phase, PSOD-1) Draft Procedure, Carus Chemical,
  In-Situ Chemical
                        fngineering Issue

-------
14  pp. http://www.epa.gov/ada/topics/oxidation issue.
html

Viellenave, J.H., J.P. Lauer,  and J.V.  Fontana.  2002.
Using risk based cleanup goals for in situ chemical oxida-
tion of PCE in vadose zone soils  under a voluntary
cleanup program.  9th  Intl.  Petroleum Environmental
Conference, Albuquerque, NM.  http://ipec.utulsa.edu/
Conf2002/tech sessions.html

Watts, R.J., M.D. Udell, P.A. Rauch, and S.W. Leung.
1990. Treatment of contaminated soils using catalyzed
hydrogen  peroxide.  Hazardous  Waste  and  Hazardous
Materials. 7(4):335-345.

Watts, R.J., B.C. Bottenberg, T.F. Hess,  M.D. Jensen,
and A.L. Teel. 1999. Role of reductants in the enhanced
desorption  and transformation of chloroaliphatic com-
pounds by modified Fenton's reactions. Environmental
Science 6" Technology. 33(19):3432-3437.

West, O.R., S.R. Cline, W.L. Holden, F.G. Gardner,
B.M.  Schlosser, J.E. Thate,  D.A. Pickering, and T.C.
Houk. 1998. A Full-Scale Demonstration of In-Situ Oxi-
dation through Recirculation  at the  X-701B  Site: Field
Operations and TCE Degradation. ORNL/TM-13556.
Oak Ridge National Laboratory, Oak Ridge, TN.

Wheeler, K.P., S.A. Miller, and J.C. Dey. 2002. In  situ
ozone remediation of adsorbed PAHs in  soil. In: Proceed-
ings of the  Third International Conference on Remediation
of Chlorinated and Recalcitrant Compounds.  Gavaskar,
A.R., and A.S.C. Chen. (Eds.), Battelle Press, Columbus,
OH. Paper2C-15.

Wiesner, M.R., M.C. Grant,  and  S.R. Hutchins.  1996.
Reduced permeability in groundwater remediation  sys-
tems:  Role  of mobilized colloids and injected chemicals.
Environmental Science & Technology. 30(11):3184-3191.

Yeh, C.K.  and Novak J.T.  1995. The effect of hydrogen
peroxide on the degradation  of methyl and  ethyl tert-
butyl   ether  in soils.   Water  Environment  Research.
67(5):828-834.

Zoh, K.D. and M.K. Stenstrom.  2002. Fenton oxidation
of hexahydro-l,3,5-trinitro-l,3,5-triazine   (RDX)   and
octahydro-l,3,5,7-tetranitro-l,3,5,7-tetrazocine  (HMX).
Water Research. 36:1331-1341.
       Engineering Issue
Chemical Oxidation

-------
v°/EPA
      United States
      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 HERE I  I
detach, or copy this cover, and return to the address in the
upper left-hand corner.
PRESORTED STANDARD
 POSTAGES FEES PAID
 EPA PERMIT No. G-35
      Office of Research and Development
      National Risk Management
        Research Laboratory
      Cincinnati, OH 45268

      Official  Business
      Penalty for Private Use
      $300

      EPA/600/R-06/072
      August 2006
      www.epa.gov
                                ^^   ^
                                w
         Recycled/Recyclable
         Printed with vegetable-based ink on
         paper that contains a minimum of
         50% post-consumer fiber content
         processed chlorine free

-------