HomeMy WebLinkAboutNCD003200383_20021119_Koppers Co. Inc._FRBCERCLA RA_Private Well Sampling-OCRMichael F. Easley, Governor
William G. Ross Jr., Secretary
Dexter R. Matthews, Director
Ms. Beverly Hudson
Remedial Project Manager
Superfund Branch
Waste Management Division
November 19, 2002
United States Environmental Protection Agency
Region IV
61 Forsyth Street, 11 th Floor
Atlanta, GA 30303
Re: Evaluation of the Source of Dioxins and Furans Detected in Private Water-Supply Wells
& Evaluation of the Soil Remedial Goals for Ground-Water Protection
Koppers Company NPL Site
Morrisville, Wake County
Dear Ms. Hudson:
The Superfund Section of the North Carolina Department of Environment and Natural Resources
(NC DENR) has received the Evaluation of the Source of Dioxins and Furans Detected in Private Water-Supply Wells & Evaluation of the Soil Remedial Goals for Ground-Water Protection for the Koppers Company National Priority List (NPL) Site. The Superfund Section has reviewed
this document and offers the attached comments.
We appreciate the opportunity to comment on this document. If you have any questions, please
feel free to call me at (919) 733-2801, extension 349.
Attachment
Sincerely,
Davi.ti /3. !Y/_a;t;u_s~ / dk
David B. Mattison, CHMM
Environmental Engineer
Superfund Section
1646 Mail Service Center, Raleigh, North Carolina 27699-1646
Phone: 919-733-4996 \ FAX: 919-715-3605 \ Internet: www.enr.state.nc.us
AN EQUAL OPPORTUNITY\ AFFIRMATIVE ACTION EMPLOYER -50% RECYCLED/ JO% POST CONSUMER PAPER
Ms. Beverly Hudson
November 19, 2002
Page I
• •
EVALUATION OF THE SOURCE OF DIOXINS AND FURANS
DETECTED IN PRIVATE WATER-SUPPLY WELLS & EVALUATION OF
THE SOIL REMEDIAL GOALS FOR GROUND-WATER PROTECTION
KOPPERS COMPANY NPL SITE
Part A Sources of Dioxins and Fu rans Detected in Private Water-Supply
Well Samples
Table of Contents
I. Please correct the Table of Contents to indicate that the title of Section 5.2.4 is
"Comparative Analysis of PCDD/PCDF Distributions: Burned Wastes, Ashes and
Related Contaminated Media versus Domestic Wastewater/Sewage Sludges and
Chlorophenol Wood Preservatives".
2. Please correct the Table of Contents to indicate that the title of Section 7 is "Summary
and Conclusions-Part A of the Report".
Part B Evaluation of the Soil Remedial Goals for Ground-Water
Protection
3. Please correct the Table of Contents to indicate that the title of Table I is "Approximate
Congener-Specific Soil Cleanup Levels Based on the Record of Decision Total
PCDD/PCDF Soil Remedial Goal for Ground-Water Protection, Adjustment to
Account for the Congener TEFs, and Congener Proportions in Site Soils".
4. Please correct the Table of Contents to indicate that the title of Table 7a is "Comparison
of Soil Remedial Goals from Table 6 to the Measured Fire Pond PCDD/PCDF Sediment
Concentrations, RI Data".
•
Mr. Dave Madison
REGION 4
ATLANTA FEDERAL CENTER
61 FORSYTH STREET
ATLANTA, GEORGIA 30303-8960
October 9, 2002
•
North Carolina Department of Environment
and Natural Resources
401 Oberlin Road
Raleigh, North Carolina 27605
Dear Mr. Madison:
Enclosed is a report entitled "Evaluation of the Source of
Dioxins and Furans Detected in Private Water-Supply Wells and
Evaluation of the Soil Remedial Goals for Groundwater Protection"
for the Koppers.Company Superfund Site. EPA will meet with the
Shiloh Community to discuss the findings of the report in the
near future. If you have any questions regarding the report,
please call me at (404) 562-8816.
Enclosure:
cc: Hope Taylor
Sincerely,
{JJ-~J-~
Beverly T. Hudson, RPM
Waste Management Division
Internet Address (URL) • http://www.epa.gov
Recycled/Recyclable• Printed with Vegetable on Based Inks on Recycled Paper (Minimum 30% Poslconsumer)
•
; ••• ___ .J ••
"{; ·t, J} )
· 1Fl-W-;.,-;_ c,\,r,::1v,"~ ,~:\ r..;,rv,,;,. ,~t,'1,"'''J.V; ,,:)<.(,;!-'~':\.-,•,; 1-·L~1;,1~·.; ~1:; ;:-1 i •• ,,,;._ ,~_\,.;~.::;,:,.;,. r;;tg.~1 •~·.1t·,7;'_j
•
'.).
''.}
.'-i/t~-i: •• -/J ·:,:.:
'<?)
Executive Summary
Part A.
Section
l.
2.
3.
4.
5.
4.1
4.2
5.1
5.2
5.3
• •
Table of Contents
Source of Dioxins and Fu rans Detected in Private
Water-Supply Well Samples
Subject
Introduction and Purpose of the Report
Site History
Site Investigations
Hydrogeologic Setting
Regional and Local Geology
Hydrogeology·
4.2. J Site Conceptual Hydrogeologic Model
4.2.2 Rate and Direction of Ground-Water Flow
4.2.3 Hydraulic Connection between Shallow (A Zone)
and Deep (B Zone) Monitoring Intervals
Contaminants of Concern
Sources of Polychlorinatcd Dioxins and Furans
Distributions and Concentrations of PCDDs and PCDFs in
2
4
4
5
5
6
22
22
23
Wastes and Contaminated Media 23
5.2.1 PCDDs and PCDFs in Burned Wastes, Ashes, and
Related Contaminated Media 24
5.2.2 PCDDs and PCDFs in Domestic Wastewaters and
Sewage Sludges 26
5.2.3 PCDDs and PCDFs as Contaminants in
Chlorophenol Wood Preservatives 28
5.2.4 Comparative Analysis of PCDD/PCDF Distributions:
Burned Wastes, Ashes and Related Contaminated Media{ vu~->
Domestic Wastewaterfdf!G-Sewage S!udgeyand
Chlorophenol Wood Preservatives 3 I
Contaminant Propenies 32
5.3. J Solubility of the PCDD/PCDF Congeners 34
5.3.2 Organic Carbon Pa11itioning Coefficient of the PCDD/
PCDF Congeners 34
5.3.3 Soil-Water Panitioning and Retardation of Ground-
Water Contaminants 36
6.
7.
8.
5.4
6.1
6.2
• •
5.3.3.1 Partitioning Analysis 36
5.3.3.2 Retardation Factor 41
5.3.4 Macromolecules and Potential PCDD/PDCF
Ground-Water Contaminant Transpon
Dioxins and Furans in Ground Water
Distribution of Dioxins and Furans in the Soil and Ground Water
44
48
at and Around the Koppers Site 49
On-Site Contamination 49
6.1.l Soil Contamination 49
6.1.2 Ground-Water PCDD/PCDF Contamination 49
6.1.2.l On-Site Monitoring Well PCDD/PCDF
Ground-Water Contamination
Off-Site Contamination
6.2.1 Off-Site Monitoring Well PCDD/PCDF
Ground-Water Contamination
6:.2.2 PCDD/PCDF Contamination in Private
Water-Supply Well Samples
51
53
Summary and Conclusions-Pa11 A ,i 4 'E.<-ro~+
References
53
66
74
76
Appendix l. PCDD and PCDF Structure and Physical/Chemical Propcnies
Part B.
Section
l.
2.
3.
4.
5.
6.
• •
Evaluation of the Soil Remedial Goals for Ground-Water Protection
Subject Page
Introduction and Purpose of Part B of the Report l
Relevant Sampling Data 3
Procedure for Determining Soil Remedial
Goals for Ground-Water Protection 10
Calculations of Soil Remedial Goals for
Ground-Water Protection and Comparison to
Fire Pond Contaminated Soil PCDD/PCDF
Concentrations
Summary and Conclusions
References
17
25
28
Appendix l. PCDD and PCDF Concentration Data, Fire Pond Sediment Samples from the
Remedial Investigation
· Appendix 2. PCDD and PCDF Concentration Data,-Fire--P-end-Sedimeut (Soil) Samples from
the EPA 2000 Investigation
Appendix 3. Technical Review Comments and Responses to the Technical Review Comments
on the Report
Figure
I
2
3
4
5
6
7
8
9
10
I I
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
• •
Figures-Part A
Title··
Key Site Areas of Potential Environmental Contamination
Remedial Investigation July 3, 1990 Shallow Ground-Water Levels
Remedial Investigation October 2, 1990 Shallow Ground-Water Levels
Remedial Investigation March 21, 1991 Shallow Ground-Water Levels
Remedial Investigation October 2, 1990 Deeper Ground-Water Levels
Remedial Investigation March 21, 1991 Deeper Ground-Water Levels
Regional C-Zone Water Levels, October 1990 '
Regional C-Zone Water Levels, March 1991
Aquifer Test Location and Observed Drawdown
Remedial Investigation October 2, 1990 Deeper Ground-Water Levels
with Flow Lines Reflecting an Anisotropic Aquifer
Per Cent of Total Mass for Combined PCDDs and PCDFs, Burned Waste,
Domestic Wastewater and Sewage Sludges, and Pcntachlorophenol
Contaminant Sources
Average Aqueous Solubility (ug/L) of the PCDD/PCDF Congeners
Average Organic Carbon Pa11itioning Coefficient of the PCDD/PCDF
Congeners
Estimated Ground-WaterTranspo11 Velocity of the PCDD/PCDF
Page
3
8
9
10
11
12
14
15
16
20
33
35
37
Congeners 45
Macromolecule Effect on the Mobility of Hydrophobic Compounds 47
Remedial Investigation Soil ,Samples with Dioxin and Furan Analyses,
Relative Concentrations of Dioxin and Furan Congeners with TEFs 50
Remedial Investigation Ground-Water Samples with Dioxin and Furan
Analyses, On-Site Monitoring Wells 52
Recent Ground-Water Samples with Dioxin and Furan Analyses, On-Site
Monitoring Wells 54
Remedial Investigation Ground-Water Samples with Dioxin and Furan
Analyses, Off-Site Monitoring Wells _ 55
2000 Ground-Water Samples with Dioxin and Furan Analyses, O,ff-Site
Monitoring Wells 56
Median Site (Source Area) Concentrations Compared to Off-Site Ground-
Water Concentrations of PCDDs/PCDFs; RI Data for Cl6C and Cl9C 59
Median Site (Source Area) Concentrations Compared to Off-Site Ground-
Water Concentrations of PCDDs/PCDFs; RI Data for C20C and C21C 60
Private Well Samples with Dioxin and Furan Detects, 1998 Sampling 67
Private Well Samples with Dioxin and Furan Detects, 1999 Sampling 68
Private Well Samples with Dioxin and Furan Detects, 2000 Sampling 69
Comparison of l,2,3,4,6,7,8-H7CDD and OCDD Concentrations in
Private Well Samples to l,2,3,4,6,7,8-H7CDD and OCDD
Concentrations in Off-Site Monitoring Well Samples 72
Table
I
2
3
4
5
6
• •
Tables-Part A
Title
Hydraulic Gradient Analyses
Organic Carbon Contents of Various Sedimentary Rocks
Kd Estimates for the Aquifer at the Koppers Site
Estimated Ground-Water Contaminant Transport Velocities
of PCDDs and PCDFs
Summary of Private Well Sampling, 1998, 1999, and 2000
Congener-Specific PCDD/PCDF Detections in Potable Well Samples
Page
7
39
40
43
70
71
\=;·f·;
Figure
1
2
3
4
Table
1
2
3
4
5
6
7a
7b
8
9
•
Figures and Tables-Part B
Title
Locations of RI Fire Pond Sediment Samples
Remedial Investigation Sediment Samples with Dioxin and Furan
Analyses-Relative Concentrations of Dioxin and Furan Congeners
with TEFs
· Approximate Locations of EPA December 2000 Investigation Fire
Pond Sediment Samples
EPA December 2000 Fire Pond Soil Samples with Dioxin and Furan
Analyses-Relative Concentrations of Dioxin and Furan Congeners
with TEFs
Title
Approximate Congener-Specific Soil Cleanup Levels Based on the
Page
4
5
7
8
Record of Decision Total PCDD/PCDF Soil Remedial Goal and . Al .,,1_J -J.. A,.00.~-I
C P . . S. S 'I (..../'., (,.u 'i>.,f,.,/.,-,-. l ongener rop01t1ons In Ite 01 s . ~-~ c. ')'·~L. ,CF,
Comparison of Sediment Sample Concentrations to the · · 1
Approximate Congener-Specific ROD Soil Cleanup Goal for
Ground-Water Protection · 9
Kd Estimates for the Buried Fire Pond Sediments at the Koppers Site 13
Toxicity Equivalency Factors (TEFs) for PCDDs and PCDFs 14
TE_F.s for PCODs/PCDFs in.Ground Water at the Practical.
Quantitation Limit Concentration 15
Initial Soil Remedial Goals for Ground-Water Protection 17
Comparison of Soil Remedial Goals from Table 6 to the Measured
. c; Ve"" I re,t>I' . Fire Pond Sechment Concentrat10ns, RI Data 19
Comparison of Soil Remedial Goals·from Table.6 to the Measured
Fire Pond PCDD/PCDF Sediment Concentrations, EPA 2000 Data 20
Statistical Breakdown on Congener Concentrations with Respect to
the Soil Remedial Goals Calculated in Table 6 22
Soil Remedial Goals for Ground-Water Protection with the North
Carolina Ground-Water 2,3,7,8-TCDD TEQ as the Basis for the
Remedial Goals 24
PCDD
PCDF
RI
TCDD
TCDF
PeCDD
PeCDF
H6CDD
H6CDF
H7CDD
H7CDF
OCDD
OCDF
foe
Koc
Kd
·"
\7.f;'i
KO\\'
TEF
TEQ
ng
pg
PCP
DOC
•
List of Acronyms
polychlorinated dibenzo-p-dioxin
polychlorinated dibenzofuran
Remedial Investigation
tetrachlorodibenzo-p-dioxin
tetrachlorodibenzofuran
pentachlorodibenzo-p-dioxin
pcntachlorodi benzofuran
hexachlorodibenzo-p-dioxin
hexachlorodibenzofuran
heptachlorodibenzo-p-dioxin
heptachlorodibenzofuran
octachlorodibenzo-p-dioxin
octachlorodibenzofuran
fraction of organic carbon
•
organic carbon pa11itioning coefficient, a measure of the tendency of an organic
contaminant to adsorb (bind) to pa11iculate organic carbon
soil-water pa11itioning coefficient a measure of the tendency of an organic
contaminant to adsorb (bind) to the soil or solid phase; the product of the Koc and
the fraction of organic carbon in the soil
octanol-water pa11itioning coefficient, a measure of the tendency of an organic
compound to pal1ition between water and octanol
toxicity equivalency factor, a measure of the toxicity of a specific dioxin or
furan compound relative to the toxicity of 2,3,7,8-TCDD
toxic equivalency, the product of the TEF and the dioxin or furan concentration
nanogram or one billionth (0.00000000 I) gram
pico gram or one trillionth (0.00000000000 I) gram
pentachlorophenol
dissolved organic carbon
• •
Executive Summary
This report presents the results of an evaluation of the probable source(s) of polychlorinated
dibenzo-p-dioxins/polychlorinated dibenzofurans (PCDDs/PCDFs) detected in samples from
private water-supply wells near the Koppers Superfund Site, Morrisville, North Carolina. This
report also presents a reevaluation of the soil remedial goals for ground-water protection for the
Koppers Site, focusing on soil contamination in the former Fire Pond.area at the Koppers Site.
The evaluation of the source(s) of PCDDs/PCDFs in samples from private water-supply wells
concludes that the Koppers Site is an unlikely source for that PCDD/PCDF contamination. This
analysis is primarily based on the environmental mobility of the PCDDs/PCDFs detected in those
private water-supply well samples. The relative concentrations of PCDDs/PCDFs detected in
ground-water samples found at various distances from likely PCDD/PCDF source areas at the
Koppers Site was also considered as possible evidence of whether or not the Koppers Site was
the source of PCDDs/PCDFs in private well samples. However, this analysis could only be used
to evaluate if the principal source of that contamination was probably something other than
contamination migrating from the Koppers Site. The source(s) of PCDDs/PCDFs in the private
well samples is indeterminate, but the ubiquitous nature of these compounds in the environment
indicates there are numerous potential sources of PCDD/PCDF contamination that could account
for the concentrations observed in the private well samples. )
The evaluation of soil remedial goals for ground-water protection determined that the soil
remedial goal presented in the Koppers Site Record of Decision is not a valid measure of the
potential threat to ground water from the soil contamination in the former Fire Pond area.
Regardless of that conclusion, the evaluation indicates the contaminated buried sediments in that
part of the site are not likely to be a threat to ground-water quality. Revised congener-specific
soil remedial goals for ground-water protection are presented that should apply to those buried
sediments in the Fire Pond area.
• •
Part A. Source of Dioxins and Fu rans Detected in
Private Water-Supply Well Samples
1. Introduction and Purpose of the Report
This report presents an analysis of the source or sources of ground-water contamination by
polychlorinated dibenzo-p-dioxins and polychlorinated dibenzofurans (PCDDs/PCDFs) detected
in samples from several private water-supply wells in the vicinity of the Koppers Company, Inc.
NPL Site, MoITisville, North Carolina (hereafter termed the Koppers Site or the Site). A second
section of the report presents soil remedial goals for ground-water protection at the Koppers Site.
The detections of dioxin or furan compounds at low levels (sub nanogram/L concentrations) in
samples from several water-supply wells has prompted a concern that this low-level
contamination is a result of contaminant migration from the Koppers Site. A second concern
about the Koppers site is that sediment contamination by dioxins and furans was left in a pai1 of
the Koppers Site that was a former pond receiving contaminated runoff. This contamination,
now buried under a layer of clean fill material, may constitute a long-term source for low-level
dioxin or furan ground-water contamination .
. A soil remedial goal for ground-water protection was calculated prior to issuance of the 1992
Record of Decision for the Koppers Site. However, that soil remedial goal for ground-water
protection used the contaminant-specific properties of the 2,3,7,8-tetrachlorodibenzo-p-dioxin
congener to calculate a soil remedial goal appl_icable to the total dioxin and furan contamination,
expressed as the toxic equivalency (TEQ). This approach was clearly overly conservative,
because it used the dioxin/furan compound that is considered both the most toxic and one of the
more environmentally mobile of the "toxic" dioxins/furans (see the discussion in Section 5 below
regarding the definition of toxic dioxins/furans) to calculate the soil remedial goals. Data
obtained during a subsequent 1996 EPA investigation of residual dioxin/furan concentrations in
sediments that were covered during the site remedial actions indicated that the 2,3,7,8-
tetrachlorodibenzo-p-dioxin congener represents less than 0.0 I% of the average total
dioxin/furan mass present in these sediment samples.
2. Site History
The Koppers NPL site was operated by Koppers Company, Inc. (Koppers) from 1962 until 1986
(Keystone Environmental Resources, Inc., 1992). The site produced glue-laminated wood
products during this period and site operations included a wood-treatment plant which operated
in the southeastern section of the property from 1968 until 1975 (Keystone Environmental '
Resources, Inc., 1992). Wood treatment involved the use of pentachlorophenol (PCP) dissolved
in a liquified butane carTier with isopropyl ether or a glycol-based compound used as a co-solvent
(Keystone Environmental Resources, Inc., 1992). Waste liquid from this process was
• •
-2-
sent to two lagoons at the site an.ct in 1977, wastewater from the lagoons was reclaimed by land
treatment conducted in two areas at the northern end of the site (Keystone Environmental
Resources, Inc., 1992). During operation of the lagoons, treated water from the lagoons was
discharged to an on-site pond (U.S. EPA, 2000). Figure I shows the locations of key site
features of concern with respect to sources or areas of environmental contamination at the site.
Several remedial actions to address site contamination have occurred. Between 1980 and 1986,
Koppers removed soils from the former lagoon, wood treatment plant, and other areas at the site
(Keystone Environmental Resources, Inc., 1992). From September 1995 until December 1996,
further site remedial actions were conducted, including removal and off-site incineration of 775
tons of contaminated soil, construction and commencement of a groundwater pump and treat
system, and pond dewatering and backfilling with clean soil (U.S. EPA, 2000a). The pond fill is
primarily a silty clay to clayey silt material with some granular material with fines included
(Cummings Riter Consultants, Inc., 1997). The backfilled area is graded to promote drainage
and is currently vegetated with native plants and shrubs (U.S. Army Corps of Engineers, 2000).
3. Site Investigations
Numerous investigations of the Koppers Site and suri·ounding areas have occurred since 1980.
Koppers conducted several site investigations in 1980, and EPA Region 4 conducted one site
investigation at that time. Based on information presented in the Remedial Investigation Report
(Keystone Environmental Resources, Inc., 1992), there were no evaluations of potential Site
contamination by either polychlorinated dibenzo-p-dioxins or by polychl01inated dibenzofurans
(PCDDs/PCDFs) during these early site investigations.
Following the 1980 investigations, continued site investigation was performed by Koppers, or an
environmental consultant to Koppers, in 1981, 1984, and• 1986. Beginning in 1986, more
concentrated sampling of off-site water-supply wells by either Beazer East, Inc., a consultant to
Beazer, or the state of North Carolina occu1Ted. As a result of those water-supply well
investigations, replacement water supplies were provided for a number of households around the
Site.
The Koppers Site Remedial Investigation (RI) began in 1990 and continued into early 1992. This
investigation was conducted under EPA oversight, followed an EPA-approved work plan, was
comprehensive in terms of environmental media that were sampled, and included analyses of
PCDDs/PCDFs in some of the samples. Samples of soil, grnund water, surface water, sediment,
and fish tissue were collected and analyzed. PCDD/PCDF analyses for soils, ground water,
surface water, sediment, and fish tissue were done by EPA method 8290 (with the exception of
three soil samples analyzed by EPA method 1613, which has a comparable to somewhat higher
quantitation limit than that obtained by method 8290).
Following completion of the Remedial Investigation, EPA Region 4 has conducted several
• -3-• Figure 1. Key Site Areas of Potential Environmental Contamination
() ,,.~ (? ;>
Key to Figure
G ~ property owned by Beazer East. Inc.
---3000 ---Cl ~ property owned by Unit Structures, Inc.
scale, feet -=c.--==--
0 500
N
A
2500
-_:: :-,CJ ---□--::-
\ --------,,.':!!" ,
2000 '----·---.
1500
1000
500
a o-1--~~~~~~~~~~~~~~~~~~
0 500 1000 1500
base map from Figure 1-2, RI Report, Keystone Environmental
Resources, Inc., 1992.
2000· 2500
• •
-4-
investigations of the Koppers Site. In 1996, EPA sampled sediments in the Fire Water Pond that
had been covered by clean fill as a part of the site remedial actions. Concentrations of toxic
dioxins and furans were established in three of these buried sediment samples. EPA has also
conducted several rounds of ground-water sampling focusing on private water-supply wells.
Sampling was performed in December 1998, March 1999, and December 2000. In the last of
these ground-water sampling events, dioxins and furans were also monitored in four key off-site
monitoring wells. These off-site monitoring wells are considered iniportant because they provide
well documented ground-water quality data from areas between the Koppers Site and potentially
impacted water-supply wells or are otherwise located at distances from the Koppers Site that are
closer to the Site than the water-supply wells. Also in December 2000, additional buried
sediment (now subsurface soil) samples were obtained from the Fire Pond area.
Periodic monitoring of ground water has also been done as a part of the site's ground-water
remedial action program that was implemented in I 997. The remedial action program is
designed to contain contaminated ground water and eventually remediate ground water to attain
performance standards. Ground-water extraction is from a well completed in the upper bedrock
and located immediately east of the former lagoon area (well PW I; reference Figure 9). The_
ground-water monitoring has involved quaiterly analysis of pentachlorophenol and 2,4-
dichlorophenol from several monitoring wells and semiannual monitoring of PCDDs/PCDFs
from a subset of monitoting wells.
4. Hydrogeologic Setting
The hydrogeologic setting is impo11ant to the understanding of the environmental mobility of
PCDDs/PCDFs, and thus is impo11ant to both the evaluation of sources of these compounds in
private water~supply wells and to the evaluation of soil remedial goals for ground-water
protection. The hydrogeologic setting is described in the Remedial Investigation Report, with
supporting data available from ground-water monitoring done as a pa11 of the site remedial action
(e.g. Fluor Daniel GT!, 1997).
4.1 Regional and Local Geology
The Koppers Site is located in the eastern part of the Piedmont physiographie province. The
Piedmont is primarily underlain by a variety of igneous and metamorphic rocks with varying
degrees of resistence to weathering.
The Koppers Site is primarily underlain by interbedded elastic sedimentary rocks that were
deposited in a basin structure. Weathering of these rocks has produced predominantiy silty and
clayey soils. These soils have a laboratory-detetmined ve11ical hydraulic conductivity of roughly
l0-6 cm/s near the surface, generally decreasing to between approximately 10-7 to 10-8 cm/sec at
depths of several feet (Keystone Environmental Resources, Inc., 1992). Beneath the near-surface
soils, there is a zone of weathered bedrock. Data from the RI Repm1 (Keystone Environmental
• •
-5-
Resources, Inc., 1992) indicate this weathered rock zone ranges in thickness from a few feet to
approximately 18 feet and averages approximately JO or 12 feel in thickness. In addition to the
sedimentary rocks and their weathering products, there are several igneous dikes at or near the
Site. While these igneous rocks are a limited part of the entire volume of earth materials of
concern, either the igneous rocks themselves or contact-metamoqihosed sedimentary rock
sun-ounding the igneous rocks may have some significant influence on ground-water flow
pattens.
Geologic structure at the site includes horizontal to moderately dipping bedrock with a strike
direction of approximately N45E, and rock fracturing (jointing of various orientations and
angles). Fracture trace analysis by EPA (as referenced in Keystone Environmental Resources,
Inc., 1992) indicates a number of fracture traces that are oriented parallel or subparallel to
regional strike (roughly N40E to N50E) with another predominant sttike direction of joints at
between N35W to N50W. Subordinate fracture traces were observed in other orientations.
According to the RI Repo11 (Keystone Environmental Resources, Inc., 1992), "Fractures were
observed in wells which ranged in strike and dip directions ... " with dip angles from horizontal to
75° or more.
4.2 Hydrogeology
The hydrogeology of the site encompasses the rate and direction of ground-water flow, the
hydraulic communication between the surface, shallow subsurface (soil and weathered rock
zone) and deeper subsurface (bedrock), the hydraulic properties of the geologic materials at the
site, and ground-water recharge and discharge zones. These factors have significance with
respect to the lranspot1 and fate of dioxin and furan contaminants.
Preliminary to a specific discussion of the Site hydrogeology, a Site conceptual hydrogeologic
model should be developed. The conceptual model considers generalized factors such as the
overall lithologies (rock types) present, the locations of surface-water bodies around the Site, the
Site topographic setting and other factors to present an overview of the likely hydrogeologic
conditions present.
4.2.1 Sile Conceptual Hydrogeologic Model
As noted in Section 4.1, the Site is ptimarily underlain by a series of elastic rocks, primarily
shaley to silty in texture, that have weathered in place to form an overlying weathered rock zone
and a near-surface soil zone that reflects the grain size distribution of the parent rock material.
The silty to clayey shallow soils are expected to generally very slowly transmit water to the
underlying earth materials and saturated zone, although near-surface soils may have a locally
higher hydraulic conductivity due to the presence of bun-owing organisms, roots, and other
features that enhance the porosity of the soil. Particularly in the weathered rock zone, there may
be relict structures such as rock joints that, if not clay filled, may result in a relatively higher
hydraulic conductivity than the overlying more deeply weathered soils or the underlying
• •
-6-
unweathered bedrock. Within the bedrock, rock fractures are expected to be laterally
discontinuous and irregularly spaced, with possible higher fracture densities in the more b1ittle
lithologies. As is typical of bedrock in the Piedmont region, rock fractures arc anticipated to
decrease in both density and aperture with increasing depth. The probable significant structural
control on ground-water flow in the bedrock may be associated with strongly anisotropic
hydraulic conductivity, meaning that the intrinsic ability of the bedrock to transmit water differs
in different compass directions.
The Site topographic setting provides an indication of likely recharge and discharge areas for the
aquifer(s) beneath the Site. Based on a review of the U.S. Geological Survey 7.5-Minute
topographic map that covers the area, the Site occupies an upland topographic position at an
elevation of approximately 350 to 380 feet. The topographic map indicates that surface water
from the Site generally drains to the south-southeast, although part of the Site may drain to the
west-northwest. Perennial streams in the area, indicative of the regional base level, suggest a
regional base level elevation of approximately 300 feet. A number of constructed ponds occupy
positions in the headwaters of various ephemeral or interrnittcnt streams in the area. These ponds.
may represent local discharge points for interflow (ground-water flow in near-surface layers that
arc briefly saturated following some precipitation events). Ground-water flow is probably to
some degree related to the land-surface elevation but is likely to be locally controlled by the
presence and orientation of rock fractures or relict geologic structures in weathered or pai1ially
weathered bedrock.
4.2.2 Rate and Direction of Ground-Water Flow
The rate of ground-water flow is a function of the intrinsic hydraulic properties of the earth
materials, which relate to the type of porosity present (primary, or intergranular versus secondary,
or fracture porosity), the density and width of rock fractures, and the grain size of the geologic
materials; and the hydraulic gradient in the aquifer. The site Remedial Investigation and
subsequent Site investigative activities provide data that allow for a reasonable approximation of
the rate and direction of ground-water flow. Monitoring wells at the Site are completed as either
shallow well_s (A-zone wells) open to the weathered bedrock or upper pm1s of the bedrock
(approximate monitoring depths below ground surface of 30 to 50 feet; monitoring elevations in
the range of 322.6 to 362.2 feet); interrnediatc depth (B-zone) wells completed at depths of
generally 60 to 70 feet below ground surface; and deep C-zone wells, completed at depths of
greater than 100 feet (reference Keystone Environmental Resources, Inc., 1992). Most of the A-
Zone wells are screened in bedrock and have I 0-foot screened intervals. B-zone wells arc
generally paired with A-zone wells, are double cased, and most of the wells have 10-foot
screened interva_ls. On-site C-zone wells are double cased, screened, and paired with shallower
wells, but most of the C-zone wells are unpaired off-site wells having an open-hole completion
with several tens of feet of open interval. Such well completions suggest a low well yield for r'he
deeper bedrock (hence the long open hole intervals that would be required to obtain sufficient
yield for monitoring purposes) and also indicate that the deeper wells may be of limited utility for
• •
-7-
defining the direction of grou~·d-water flow, since water levels from most of these wells may
represent a composite of water levels at different elevations. Testing of the ability of sections of
several open holes completed at the Site to accept injected water indicated that generally, the
hydraulic conductivity of the earth materials at the Site decreased substantially below depths of
between approximately 60 and 100 feet (reference Keystone Environmental Resources, Inc.,
1992).
Water-level data from both shallow and deep wells indicate a potentiometric (water'level) high
area around the Koppers Site, with ground-water flow in multiple directions from the Site. This
condition probably reflects the topographic high at the 'site, as shallow ground-water flow
typically reflects the topography of an area. Local_ized site conditions such as buildings, other
impervious surfaces and soil types may affect the specific position of the potentiometric high at
the Site.
Figures 2 through 6 show water levels measured in shallow and deep monitoring wells during the
Remedial Investigation and present an approximation of the potentiometric contours for shallow
A zone wells and deeper B/C zone wells. These figures show a general consistency of ground-
water flow directions for the different seasons in which measurements occurred. The figures also
show a somewhat northwest to southeast elongation of potentiometric contours for both the
shallow and the intermediate/deep (B/C) monitoring zones. These potentiometric contour
orientations may reflect an anisotropic aquifer condition with a greater hydraulic conductivity
along a northwest-southeast axis.
Based on Figures 2 through 6, analyses of the aquifer hydraulic.gradients (slope of the water
table, or of a potentiomet1ic surface) were made. Table 1 presents the results of these analyses.
Table 1. Hydraulic Gradient Analyses
Shallow Monitoring Zone (A Zone)
Upgradienl Downgradient Distance· Estimated
Orientation Date Water Level. ft Water LcvCt ft Between Wells Avdraulic Gradient
NW-SE 7/3/90 373.83, CSA 363.82, C3A 601.94 ft 0.0166
NW-SE 7/3/90 372.76, C4A 366.42, C2A 545.02 ft 0.0116
NW-SE 7/3/90 371.82, C6A 359.03, C31A 439.36 ft 0.0291
SW-NE 7/3/90 372.76, C4A 365.7, CI0A 453.92 ft 0.0156
NW-SE 10/2/90 369.27, CSA 358.18, C3A 601.94 ft 0.0184
NW-SE 10/2/90 369.91, C4A 362.66, C2A 545.02 ft 0.0133 NW-SE 10/2/90 364.28, C6A 354.04, C3 IA 439.36 ft 0.0233.
SW-NE 10/2/90 369.~I. C4A 362.25, ClOA 453.92 ft 0.0169
NW-SE 3/21/91 373.49, CSA 361.92, C3A 601.94ft 0.0192
NW-SE 3/21/9 I 372.87, C4A 365.49, C2A 545.02 ft 0.0135
NW-SE 3/21/91 369.51, C6A 358.35, C31A 439.36 ft 0.0254
SW-NE 3/21/91 372.87, C4A 364.38, ClOA 453.92 ft 0.0187
Table 1 is continued on page 13
• -8-•
Figure 2. Remedial Investigation July 3, 1990 Shallow Ground-Water Levels
3500
3000
2500
2000
1500
1000
500
.\
\ ,1
36 -~
C3
0
363.88
C9A
Q •
'?>~66.95
. c1AO
66.42 --· ·ec2A s .
500
data soufce: Figure 3-11 RI Report
1000 1500
[
----
55.88
Key to Figure
t SC ALE, FT N o 200 =
• C11A
monitoring well
location
C11A
2000 2500
• -9-•
Figure 3. Remedial Investigation October 2, 1990 Shallow Ground-Water Levels
3500-
3000
2500
2000
1500
1000
500
I;
360.01
C9A
;;/~( I
<Jl/ l:
/ Medl/J·\I : Pond
I
, Key to Figure
\
t SC ALE, FT N o 200 =
• C11A
monitoring well
1ocation
L:=~Li 0-"-----~-------,----~-~-.L_--="----------------__J
0 500 1000 1500 2000 2500
data source: Figure 3-12 Al Report
• -I 0-•
Figure 4. Remedial Investigation March 21, 1991 Shallow Ground-Water Levels
3500
3000
2500
_ ------2000
1500
1000
500
0
364.09
C9A
364.53
C1A.
365.49
C2A.
500
---
data source: Figure 3-13 RI Report
-----°",
A~' 67c
\. ~~" \__:: __ _ ,., ., ---, ---
"'---
Key to Figure
t SC A LE, FT N o 200 =
• C11A
monitoring well
location
351\.06
• C11A
C
1000 1500 2000 2500
• -11-•
Figure 5. Remedial Investigation October 2, 1990 Deeper Ground-Water Levels
4000
3500
3000
2500
1500
1000-
500
• /19c
347.0B • C20C
357.13
355.62
;j!j
C'?
.•. C98
C9C
Ii
\
I:_·
,I
dala sources: Figure 3-14 a~d Table 3-2a, RI Report
1500
Key to Fig_ure
t SC ALE, FT . N o 200 =
e C11B
monitoring well
location
343.73
--------•--
C16C
2000 2500
• -12-•
Figure 6. Remedial Investigation March 21, 1991 Deeper Ground-Water Levels
4000
3500
3000
2500
1500
1000
500
~~O~·---------------------------~
7
357.3 • C20C
0
361.7
.C9B
C9C 360.89
500 1000
data sources: Figure 3•15 and Table 3-2a, Al Aeporl
1500
Key to Figure
t SC Al E, FT N o 200 =
e C11 B
monitoring well
location
2000 2500
•
-13-
Table 1, Ilydraulic Gradient Analyses, continued
Deeper Monitoring Zone (B Zone)
Upgradient Downgradicnt Approximate Dist~nce Estimated
Orientation Date . Water Level. ft Water Level, fl Between Wells Hydraulic Gradient
NW-SE l0/2/90 368.15, Cl2B 362.64, C2B 940.985 ft 0.0059
NW-SE l0/2/90 363.89, C28B 350.52, Cl5B 753.22 ft 0.0178
SW-NE 1012190 368.15, Cl2B 349.69, C26B 765.66 ft 0.0241
SW-NE 10/2/90 362.56, C29B 359.6, Cl 1B 265.37 ft 0.0112
NW-SE 3/21/91 367.46, CI 2B 363.27. C2B 940.985 ft 0.0045
NW-SE 3/21/9 I 370.65, C28B 351.97, Cl5B 753.22 ft 0.0248
SW-NE 3/21/91 367.46, Cl2B 352.05, C26B 765.66 ft 0.0201
Regional ground-water flow patterns in the C zone are depicted on Figure 7.and Figure 8. These
figures indicate a generally eastward lateral flow direction in the deeper bedrock. Some C-zone
water-level data are not included in these figures because they are anomalously low water levels
that are inconsistent with the general range of 312 to 350 feet water-level elevation for the C
zone. Such anomalous water levels may indicate the wells are at locations where the ability of
the rock to yield water is minimal, and thus equilibration of water levels in those wells to the
regional hydraulic head is very slow.
With respect to the direction of ground-water flow away from potential on-site contaminant
source areas, the following observations.apply:
► Under ambient conditions, contaminants originating in the former wood treating area, fire
pond or lagoon areas (reference Figure 1) would either be expected to move downward or
laterally to the south or east.
► Under ambient conditions, contaminants originating in the former land farm area
(reference Figure 1) would be expected to either migrate downward or northward
(potentially no11hwest to east-northeast). ,
► Contamination that reaches the deepest pai1 of the geologic materials that are monitored
around the Site may be transported eastward.
The hydraulic properties of the earth materials at the Site have been evaluated during the RI by
slug testing of monitoring wells, and by conducting an aquifer test on a well completed to a depth
of 49 feet with 17 feet of open hole. Figure 9 shows the aquifer test location and the drawdowns
observed during the aquifer test after approximately one day of test well pumping. Observations
relating to this aquifer test are summarized as follows (analysis as either presented in the RI
Report or in this report):
► In the vicinity of the test well, there can be a relatively high degree of vertical hydraulic
connection and/or leakage between the weathered bedrock and underlying fractured
bedrock, as evidenced by the immediate response and significant drawdown in an A-zone
well (C-27 A). However, the hydraulic connection between the uppermost part of the
saturated zone and the shallow bedrock in the pumped interval is not consistent (compare
A-zone well C-27 A drawdown to A-zone well drawdown at CllA, Cl3A and M-4, on
10000
9000
8000
7000
6000
5000
4000
3000
2000
1000
• -14-•
Figure 7. Regional C-Zone Water Levels, October 1990
C1 BC •• 354.08
\ 39.84 ~ .o
. \\·
I Key to Figure
SCALE, FT
tN o soo 1000
e C11B
monitoring well
location
• 315. 2
24
31~
0-t-------,-----,--------,----.-----;---~--'----r-'--__J
0 1000 2000 3000 4000 5000 6000 7000
data sou1ces: Figure 3-14 and Table 3-2a, Al Report
10000
· 9000
8000
7000
6000
5000
4000
3000
2000
1000
• -15 -•
Figure 8. Regional C-Zone Water Levels, March 1991
C1 BC \.
356.41
& C")
\
(.,)
()'\
0
358.54
I
3
I
0 ~
•
Key to Figure
tN SCALE, FT
500 1000 0
e C11B
monitoring well
location
• 312. 8
C24 • 310.73
0-+----~-----.-----'--r---~----;-----'----------'
0 1000 2000 3000 4000 5000 6000 7000
data sources: Figure 3-15 and Table 3-2e. RI Report
• • -1 6-
Figure 9. Aquifer Test Location and Observed Drawdown
SCALE (FEET)
0 150 300 450
l.E6BI)
-t -IIONITIJUNG llEU. LOCATION * PW LOCATIOS OF THE PU~'PED WELL
,, ,,,,,
-8£AZ£R EAST. INC. PROPERTY 80LNlARY
-lllfl SlRUCTUAES 111:. PROPERTY BOUNOAAY
-fRACTlllE TRAa:
-STRIKE OF SEDIMENTARY UNITS AT N 45 E (AT 500 FT. HORIZCMAL JNTUIVALS)
-EXTENT OF IIVIIIJOIIN OF 25 n .
-----£lCTEHT OF lf!AICJOIIN OF 20 FT.
----EXTENT OF ORAIOlWN OF LO FT.
-----EXTEkT OF IIVINDOIIN OF 0 .25 FT
!CONTOUR IS DASHED WrlEM INftRRED\
129.251 OBS£RYEO ORAWOOWN AT 30 HOURS
ORAWOOWN CROSS SECTION.
PRESENTED AS RI REPORT FJGURE 3-19
\)
a
• •
-17-
Figure 9). This point is also supported by Table 3-3a of the RI Report. Table 3-3a shows
wells with a qualitative description of having high hydraulic connection, moderate
hydraulic connection, and low hydraulic connection between wells (based on aquifer test
observations).
► Figure 9 shows a water-level response to pumping that clearly demonstrates a structural
control on ground-water flow and anisotropic conditions within the upper bedrock around
the Site.
► According to the RI Report the time-drawdown curve for the pumping well and the
"responsive wells" (presumably wells showing multiple feet of drawdown) indicated a
leaky aquifer condition
► Aquifer test analysis for fractured bedrock is more complicated than for porous media,
although a general indication of aquifer hydraulic properties can be obtained.
Referencing Table 3-5 of the RI Report, the hydraulic conductivity of the aquifer as estimated
from the aquifer test ranges from a low value of 0.1526 ft/d to a hi gh of 4.12 ft/d with a median
of 0.33 ft/d (applicable to the aquifer fractures). There is a much lower estimated hydraulic
conducti vity for the aquifer matrix (median value of 0.0012 ft/d), using one method of aquifer
test analysis. For compaiison, the Theis recovery method of aquifer test analysis yielded an
aquifer transmissivity of between 11.15 and 224.6 ft2/d, with a median transmissivity of
approximately 11 .5 ft2/d. Considering that the effective thickness of the aquifer is approximately
60 to 100 feet (based on the analyses of the ability of the aquifer to accept water, as discussed
above), the overall hydraulic conductivity of the aquifer is estimated from the aqui fer
transmissivit y using the relationship transmissivity (T) = hydraulic conductivity (K) • aquifer
thickness (b). Solving for the aquifer hydraulic conductivity yields a value of approximately
0.11 5 ft/d to 0.19 ft/d.
The hydraulic conductivity of the earth materials at the Site was also estimated by slug tests on
several monitoring well s. According to Table 3-5 of the RI Report, these slug test data indicated
a range in the hydraulic conductivity of 0.004 ft/d to 0.64 ft/d with a median value of 0.245 ft/d
(Bouwer and Rice method of analysis). Most of these data were from A zone wells (6 of 9 test
results).
To conclude the basic analysis of hydra~lic conductivity, the various methods that have been
used to estimate the hydraulic conductivity of the aquifer materials at the Koppers Site have all
indicated that the average (median) hydraulic conducti vity of the aqui fer materials in roughly the
upper 100 feet of the saturated zone is on the order of 0.2 to 0.3 feet per day. The hi ghest
estimated hydraulic conductivity for the Site is for A-zone well C-27 A. The various methods of
aquifer test analysis indicate a hydraulic conductivity at this well of between 1.09 ft/d (Cooper-
Jacob analysis of T; assumed effective aquifer thickness of 100 feet) and 4.12 ft/d. Various data
and data analyses presented in the RI Report indicate the hydraulic conducti vity of the deeper
parts of the bedrock monitored by the C-zone wells is generally very low.
As is apparent from Figure 9, the response of wells to pumping at well PW-1 varies with respect
to not only their proximity to the pumped well but also varies with respect to direction and
• •
-18-
therefore indicates an anisotropic aquifer (hydraulic properties vary with respect to the direction).
As the best example of the anisotropy, the water-level responses for deeper wells Cl4B and
C27B are compared. C14B is about half the distance from the pumped well than is C27B, yet
there was almost three times as much drawdown noted at C27B compared to Cl4B. The
hydraulic conductivity at C27B (considered as more or less representative of the hydraulic
conductivity in the B monitoring zone along the major axis of flow, termed Kx) and the hydraulic
conductivity at C14B (considered as more or less representative of the hydraulic conductivity
along the minor axis of flow, termed K,) are, respectively reported as 0.8482 ft/d and 0.3427 ft/d
(reference RI Report Table 3-5; pumping test results using the Moench analysis).
For purposes of defining the principal direction(s) of ground-water flow in this anisotropic
medium, the procedures presented in Section 5.1 of Freeze and Cherry (1979) are followed.
Specifically, proper scaling of the anisotropic flow net was performed as follows:
l.
2.
3.
4.
5.
Draw two lines on the drawdown contour map (Figure 9) corresponding to the piincipal
direction of drawdown (one line along the long axis of the 20 and 25-foot drawdown
contours; this is the x direction, and one line drawn at a 90° angle to that principal
direction; this is the z direction.
Copy the October 2, 1990 deeper monitoring zone potentiometric contour plot onto a new
plot file in the Surfer® software program.
Cut the two lines added to the drawdown contour map and paste on the new plot file.
Rotate all elements on the new plot file so the two lines added to the potentiometric
contour map in step 3 plot parallel to the x and y coordinates of the figure. This step is
done to make stej 5 easier, since all elements are now aligned with the x-y coordinate
system of Surfer .
Use equation 5.11 in Freeze and Cherry (1979) to adjust the z direction so that Z =
zF,
F
where: K_, is the hydraulic conductivity in the x direction (the major axis of flow) and K: is the
hydraulic conductivity in the z direction.
6. Add a third line to the new plot file that parallels the x direction on the Figure 9
drawdown contour map.
7. Insert flow lines perpendicular to equipotentials (equivalent to step (2) in Freeze and
Cherry). For convenience, four flow lines are added; these flow lines are initiated as
parallel to the x and z directions at the point of highest hydraulic head.
8. Rotate all elements back so that the x line added to the new plot file in step 3 is parallel to
the new line (added in step 6) that represents the x direction on the drawdown contour
map.
9. Per Freeze and Cherry, invert the scaling ratio.
10. Copy the four flow lines onto the original contour map. This process now results in flow
lines that represent a truer direction of ground-water flow for the principal directions of
• •
-19-
anisotropy than flow lines that would be drawn if an isotropic condition had been
assumed. Figure 10 shows the corrected positioning of the flow lines for the principal
directions of anisotropy, along with flow lines that would have been drawn if an isotropic
condition had been assumed.
Considering the flow lines that reflect the anisotropic condition, the rate (or potential range in
rate) of ground-water flow along the x and z directions (i.e. probable maximum and minimum
hydraulic conductivity tensors in the horizontal plane) in the B monitoring zone can be estimated
by considering the estimated hydraulic conductivities associated with observation wells oriented
in the x and z directions away from the pumped well (specifically, wells C27B and C14B) and
the hydraulic gradients along flow lines shown in Figure 10.
The ground-water velocity is estimated as the product of the hydraulic conductivity and the
hydraulic gradient, divided by the aquifer effective porosity. For a fractured rock aquifer, the
effecti ve porosity is potentially a combination of both fracture, or secondary porosity and matrix,
or primary porosity. Considering the lithologies noted in RI geologic logs, along with the aquifer
test results presented in Table 3-5 of the RI Report, the primary porosity of the geologic materials
at the Site is probably very low. Thus, the effective porosity of the aquifer can probably be
represented by the fracture porosity. While fracture porosity varies with depth and with specific
geologic mate1ials, an overall reasonable estimate of the fracture porosity in the predominantly
shaley to siltstone to sandstone earth materials present can be estimated on the basis of the
aquifer storage coefficient and equation (21) in Lohman (1972), which relates the storage
coefficient to the aquifer porosity. Table 3-5 in the RI Report presents several estimates of the
specific storage for the aquifer at the Site. The specific storage, S5, is related to the storage
coefficient, S as follows:
S = S/b, where bis the aquifer thickness. Typical values of S for confined aquifers are in the
range of 10-3 to 10-5. Thus, given an effective aquifer thickness at the Site of between
approximately 60 tolO0 feet, a minimum value for S5 is on the order of 10-5•60 = 6.0 E-4. A
specific storage of this magnitude was not observed for any of the wells listed in Table 3-5 of the RI Report; the_B-zone well with the highest specific storage value is C-14B, with an S5 estimate
of 4.08E-7, three orders of magnitude lower. Given the lowest realistic storage coefficient
estimate based on Lohman (1972), equation (21) in Lohman is used to estimate an aquifer
porosity as follows:
a S = Hyb (/J + -) e
where Sis the storage coefficient (minimum typical value for confined aquifers is 0.00005 or
5E-5).
8 is the aquifer porosity (unknown)
y is the specific weight per unit area for water (0.434 lb in-2)
bis the aquifer thickness (assumed equal to 60 feet, the approximate effective thickness of the
aquifer, based on RI Report data; or in units of inches, 720 inches) /3 = 1/E'", where E'" is the bulk modulus of elasticity of water (3E5 lb in-2 , per Lohman, 1972)
• • -20-
Flgure 10. Remedial Investigation October 2, 1990 Deeper Ground-Water Levels
with Flow Lines Reflecting an Anisotropic Aquifer
40
3
2
1
1000
347.08
$
C20C
\
0 500
data IIOUnlOS: Frgunt 3-14 and Table 3-2■, RI Report
SCA\.E,FI'
~00
.Jl i
lledl Pon
l__
1000 1500 2000 2500
- - - • flow line fof l<x direction, anisotropic condition
---• flow line for Kz direction, anisotropic condition
• flow llne for K,r direction, Isotropy assumed
• flow line for Kz direction, Isotropy assumed
• •
-21-
a= 1/Es where Es is the bulk modulus of elasticity of the aquifer, as confined in situ, in units of
lb in-2
The bulk modulus of elasticity of the aquifer is estimated from information presented in Hunt,
E 1984. According to this reference, Es =----
3(1-2 v)
where Es is the bulk modulus of elasticity of the aquifer
Eis Youngs Modulus
and vis the poisson ratio
For the type of geologic materi~ls present at the Site, Eis likely to be approximately 4E5 Kg/cm2
and vis likely to be approximately 0.25 to 0.33 (based on Table 3.33 in Hunt, 1984). Given
these estimated values, Es is estimated to be in the range of 266,667 Kg/cm2 to 392,157 Kg/cm2.
Converting to units of lb in-2, Es is estimated to range from 3.78E6 lb in-2 to 5.56 E6 lb in-2.
From equation (21) in Lo~man (1972), the aquifer porosity is then estimated to be 0.001 (within
an order of magnitude). This porosity is extremely low but such low values have been noted in
the literature (reference Freeze and Cherry, 1979, p. 408). Effective fracture porosities are
typically several orders of magnitude less than effective porosi ties for most granular aquifers
(Esposi to and Thomson, 1999). A low value for fracture porosity is also consistent with the low
yield of the aquifer materials based on the aquifer test and slug tests, and provides a conservative
-::-;. . ., estimate of the ground-water velocity in the bedrock.
Considering the aquifer test data, hydraulic gradient estimates, and a fracture porosity of
approximately 0.001, as presented above, estimates of the ground-water velocity in the bedrock
at the Site are presented as follows:
For ground-water flow along the major hydraulic conductivity tensor, using the equation v =
KI/0, where:
K is the hydraulic c~nductivity
I is the hydraulic gradient,
and 0 is the aquifer porosity,
(from the RI pumping test results using the Moench
analysis; reference RI Repo11 Table 3-5)
the estimated ground-water velocity is as follows:
v = 0.8482 ft/d • 0.00777 (hydraulic gradient from the vicinity of well C9B to well Cl2B ; see
Figure 10)/ 0.001 = 6.59 ft/d.
For ground-water flow along the minor hydraulic conductivity tensor, the estimated ground-water
velocity is as follows:
v = 0.3427 ft/d • 0.012 (hydraulic gradient from the vicinity of the approximate 354 elevation
contour to well Cl2B; see Figure 10)/ 0.001 = 4.11 ft/d.
• •
-22-
These estimates are based on the assumed flow of water through porous media (Darcy equation
for fluid flow in porous media). In Section 5 .3.3.2, a further di scussion of ground-water flow in
fractured rock conditions is presented.
4.2.3 Hydraulic Connection Between Shallow (A Zone) and Deep (B Zone) Monitoring Intervals
Movement of ground water and ground-water contaminants has both a horizontal and a vertical
component. At the Koppers Site, the vertical ground-water flow component is primarily
downward, based on a comparison of paired A and B-zone monitoring well water levels. Such a
downward vertical flow component is anticipated around an area where there is a potentiometric
(water level) high (a ground-water recharge area). Some exceptions to such a downward flow
component are noted around the Site. These exceptions are generally small upward gradients
that are not consistent between different measurement events (reference RI Report, Table 3-3).
Such ex·ceptions may reflect complex distributions of fractures or other flow paths that result in
perturbations to the distribution of the hydraulic head. However, as a generalization, across most
of the Site, there is a potential for vertical contaminant mi gration into the deeper pa11 of the
bedrock where there is effecti ve hydraulic conductivity (i.e. generally the upper 60 to 100 feet of
the saturated zone).
The RI aquifer test provided some qualitati ve information on the hydraulic connection between
the A and B zone. Some A-zone wells showed a notable response to the pumping well, whereas
other A zone wells showed little or no response to pumping. Some of the non-responsiveness
may be due to the fact that A-zone wells are not ali gned along the major hydraulic conductivity
tensor. However, at least one well (C-13A; reference Figure 9) is aligned more or less along the
principal hydraulic conductivity tensor and is as close to the pumped well as several other A-
zone well s showing more than a foot of drawdown, yet there was only a 0.03-foot water level
decline in C-13A after 24 hours of pumping. This observation is consistent with the in-egular
and di scontinuous nature of bedrock fractures, as suggested by the Site conceptual model. As a
generali zation, there are areas at the Site where there is some significant vertical hydraulic
connection and downward flow from the A to B zones, and other areas where there is minimal
hydraulic connection and thus little or no potenti al for significant vertical ground-water fl ow and
contaminant transport. Such cor,ditions can explaifl apparent4£. unusual patterns of ground-water
contaminant distributions.
5. Contaminants of Concern
For purposes of this analysis the contaminants of concern are two classes of compounds known
as polychlorinated dibenzo-p-dioxins (PCDDs) and polychlorinated dibenzofurans (PCDFs).
There are 75 possible PCDDs (75 congeners) and 135 PCDFs (U.S. EPA, 1998). Of these large
number of compounds, only seven of the PCDDs and ten of the PCDFs are considered by EPA, to
have toxicity of particular concern. These are the compounds with chlorine substitutions in the
2,3,7 and 8 positions (U.S. EPA, 1998). Appendix l to this report presents a summary of
physical an d chemical properties and structures of these seventeen PCDDs and PCDFs. These
PCDDs and PCDFs have been periodically monitored in soil and ground water at and around the
• •
-23-
Koppers Site and are the only PCDDs and PCDFs that are further di scussed in this report.
For risk assessment purposes, a system has been developed that defines the relative toxicity of
the seventeen PCDD and PCDF congeners. This system is based on the toxicity of the most
environmentally significant toxic compound, 2,3,7,8-tetrachlorodibenzo-p-dioxin (2,3,7,8-
TCDD), which is assigned a toxicity equivalency factor (TEF) of l. The remaining
PCDDs/PCDFs which are considered to also have toxicity are assigned TEFs that range from
0.001 to 0 .5. A congener with a TEF of 0.01 would have a toxicity of one hundredth th at of
2,3,7,8-TCDD. For a mixture of dioxins and furans, the toxicity equivalency (TEQ) is calculated
by multiplying the concentration of each individual congener by its TEF and summing the
products. Thus, for two mixtures of dioxins and furans, a much higher concentration of a
congener with a low TEF in one sample may represent an equivalent, or lesser ri sk, than a much
lower concentration of congeners with higher TEFs. A further discussion of the TEF and TEQ is
included in Part B, Section 3 of this report.
5 .1 Sources of PCDDs and PCDFs
There are wide variety of sources of PCDDs and PCDFs in the environment. Combustion of a
variety of mate1ials can produce dioxins and furans. A combustion source for PCDDs and
PCDFs is probably the most ubiquitous in the environment. There are a number of additional
sources of di oxins and furans in the environment, including:
► Composting of household or garden wastes (Krauss et al, 1994; Oberg et al, 1994)
► Use or production of organochlorine pesticides (ATSDR, 1988)
► Manufacture of a variety of chlorinated organic compounds (Ree et al, 1988)
► Dyes, pigments and printing inks (Santi et al, 1994; Williams et al, 1992)
► Municipal and household wastewaters and sewage sludge (Horstmann et al , 1992;
Horstmann et al, 1993, Horstmann and McLachlan, 1994)
► Pulp and paper manufacturing plants (ATSDR, 1988)
► Dry cleaning residues (Umlauf et al, 1993)
It is now known that PCDDs and PCDFs are also naturally occurring compounds and have been
found in soils where there is little or no contribution to the contaminant mass from anthropogenic
sources. A discussion of how such naturally occuning PCDD/PCDF contamination may
contribute to ground-water contamination around the Site is included in Section 6.2 of this
report.
5.2 Distributions and Concentrations of PCDDs and PCDFs in Wastes and Contaminated Media
Investigati on of the specific composition of PCDDs and PCDFs from different sources has
revealed that specific PCDDs and PCDFs predominate in tested samples. Of particular interest is
the prevalence of PCDDs and PCDFs in samples consisting of or contaminated by the products
of combustion, since the burning of various materials is commonplace and a number of studies
have considered the di stribution of PCDDs and PCDFs in burned materials, ashes and the like.
,
·...,.;(:.;.
• •
-24-
The presence of PCDDs and PCDFs in domestic wastewaters or sewage sludges is also of
interest, since these potential PCDD/PCDF sources may be present in close proximity to
household water-supply wells. Also of interest is the di stribution of PCDDs and PCDFs in
chlorinated wood preservative wastes, since these contaminants are potential sources of PCDDs
and PCDFs at the Koppers Site. Characteristic relative amounts of PCDDs/PCDFs in different
source materials may be compared to environmental PCDD/PCDF data, to help evaluate the
source of the environmental contamination.
5.2.1 PCDDs and PCDFs in Burned Wastes, Ashes, and Related Contaminated Media
In a study of burned residue from a municipal waste landfill fire, Ruokojarvi et al (1995) reported
that OCDD represented most of the total concentration of the toxic PCDDs/PCDFs in the
residue. 1,2,3,4,6,7,8-H7CDD and l,2,3,4,6,7,8-H7CDF were the other two predominant
congeners noted in the burnt waste samples. Similar relative distributions of dioxin and furan
congeners were also generall y observed in unburnt landfill waste samples.
In a separate study of PCDDs and PCDFs around a landfi ll where wastes had been burned in the
open, (Alawi et al, 1996), contaminated soil samples indicated the following median
co ncentrations (ng/Kg) of the toxic PCDDs and PCDFs:
2,3, 7,8-TCDD <10
1,2,3,7,8-PeCDD 37
l ,2,3,4,7,8-H6CDD 40
1,2,3,6,7,8 H6CDD 68
1,2,3,7,8,9 H6CDD 45
1,2,3,4,6,7,8 H7CDD 440
OCDD 509
2,3,7,8-TCDF <10
1,2,3,7,8-PeCDF 100
2,3,4,7,8-PeCDF 87
1,2,3,4,7,8-H6CDF 58
1,2,3,6,7,8 H6CDF 55
1,2,3,7,8,9 H6CDF <10
2,3,4,6,7,8-H6CDF 67
1,2,3,4,6,7,8 H7CDF 242
1,2,3,4,7,8,9 H7CDF <10
OCDF 52
As for the study by Ruokojarvi et al (1995), this independent study of landfi ll-de1ived
PCDD/PCDF contamination indicated that the principal congeners present were OCDD,
1,2,3,4,6,7,8-H7CDD and 1,2,3,4,6,7,8-H7CDF.
Based on data repo11ed in U.S. EPA (1990), analyses of municipal waste combustion ash samples
from five sources indicated median PCDD/PCDF median concentrations (arranged in order of
decreasing concentration in ng/Kg) as follows:
•
OCDD 544
1,2,3,4,6,7 ,8 H7CDF 539
1,2,3,4,6,7,8 H7CDD 319
1,2,3,6,7,8 H6CDF 279
2,3,7,8 TCDF 263
OCDF 243
1,2,3,4,7,8 H6CDF 218
1,2,3,7,8,9 H6CDF 127
1,2,3,7,8,9 H6CDD 79
1,2,3,7,8 PeCDF 64
2,3,4,7,8 PeCDF 56
2,3,4,6,7,8 H6CDF 54
1,2,3,4,7,8,9 H7CDF 48
1,2,3,4,7,8 H6CDD 40
1,2,3,7,8 PeCDD 35
1,2,3,6,7,8 H6CDD 34
2,3,7,8 TCDD 16
•
-25-
These medi an concentrations show a general accordance with data from the other muni cipal
waste contamination studies of PCDD/PCDF concentrations, as the principal congeners present
were OCDD, l ,2,3,4,6,7,8-H7CDD and l ,2,3,4,6,7,8-H7CDF.
From a study of wood combustion residues by Wunderli et al (2000), samples of bottom ash and
'"::~ .. :~ fly ash from burned native wood and waste wood samples, indicated the following approximate
concentrations (ng/Kg) of the six toxic PCDDs/PCDFs found in the highest concentration in each
ash sample type:
Native Wood
Bottom Ash
OCDD-30
OCDF-10
I ,2,3,4,6,7,8 H7CDD-8
1,2,3,4,6,7,8 H7CDF -7
2,3,4,7,8 PeCDF-5
l,2,3,7,8PeCDF-4
Waste Wood
Fly Ash Bottom Ash Fly Ash
OCDD-20 OCDD-85 OCDD-7500
1,2,3,4,6,7,8, H7CDD-9 1,2,3,4,6,7,8 H7CDD-20 1,2,3,4,6,7,8 H7CDD-5000
OCDF-6 1,2,3,7,8 PeCDF-8 1,2,3,7,8 PeCDF-2500
1,2,3,4,6,7,8 H7CDF~l.8 l,2,3,4,6,7,8H7CDF-7 2,3,4,7,8 PeCDF-2000
2,3,4,7,8 PeCDF~ 1.4 OCDF-5 1,2,3,4,6,7,8 H7CDF -1700
1,2,3,7,8 PeCDF-1.1 2,3,4,7,8 PeCDF-5 2,3,7,8 TCDF-110
As with the data from the municipal landfill burned waste residues and soil samples, OCDD was
the most prevalent toxic PCDD/PCDF in all ash samples. l,2,3,4,6,7,8-H7CDD and
1,2,3,4,6,7,8-H7CDF were also in the group of the six most predominant toxic PCDD and PCDF
congeners. The 2,3,7,8-TCDD congener was one of the least concentrated PCDDs in the ash
samples.
As reported by Nestrick and Lamparski (1983), analysis of chimney soot from residenti al
fireplaces found that the concentration of OCDD was hi gher than the concentration of any other
2,3,7,8-PCDD congener (1,2,3,7,8-PeCDD was not reported). 1,2,3,4,6,7,8-H7CDD had the next
highest concentration. Three soot samples were tested.
• • -26-
5.2.2 PCDDs and PCDFs in Domestic Wastewaters and Sewage Sludges
Some investigations have been performed to determine the concentrations of toxic PCDD and
PCDF congeners in domestic wastewater and sewage sludges. A study by Horstmann and
Mclachlan (1995) of wastewater discharged from an apartment building (approximately 750
residents) found the following distributions of PCDDs and PCDFs in the wastewater:
PCDD/PCDF Congener
2,3,7,8-TCDD
1,2,3,7,8-PeCDD
1,2,3,4,7,8-H6CDD
1,2,3,6,7,8 H6CDD
1,2,3,7,8,9 H6CDD
1,2,3,4,6,7,8 H7CDD
OCDD
2,3,7,8-TCDF
1,2,3,7,8-PeCDF
2,3,4,7 ,8-PeCDF
1,2,3,4,7,8-H6CDF
1,2,3.6,7,8-H6CDF
1,2,3, 7,8,9-H6CDF
2,3,4,6, 7 ,8-H6CDF
1,2,3,4,6,7,8-H7CDF
1,2,3,4,7,8,9-H7CDF
OCDF
Median Concentration. pg/L (eight samples)
0
0.23
0
1.9
0.84
42
525
1.4
0.5
0.7
0.55
0.47
0
0.34
8.1
0.33
9.8
For these wastewater samples, OCDD averaged almost 90% of the total PCDD/PCDF congener
mass. OCDD and l ,2,3,4,6,7,8-H7CDD together were almost 96% of the total congener mass.
Horstmann and Mcl achl an (1995) also sampled washing machine effluent from different sources
and found ;he following median concentrations of PCDDs/PCDFs:
PCDD/PCDF Congener
2,3,7,8-TCDD
1,2,3, 7 ,8-PeCDD
1,2,3,4, 7,8-H6CDD
I ,2,3,6,7,8-H6CDD
l ,2,3,7,8,9-H6CDD
l ,2,3,4,6,7,8-H7CDD
OCDD
2,3,7,8-TCDF
1,2,3, 7,8-PeCDF
2,3,4,7,8-PeCDF
1,2,3,4, 7 ,8-H6CDF
l ,2,3,6,7,8-H6CDF
1,2,3, 7 ,8,9-H6CDF
2,3,4,6,7,8-H6CDF
1,2,3,4,6, 7 ,8-H7CDF
1,2,3,4, 7,8,9-H7CDF
OCDF
Median concentration, pg/L
(four samples. different cloth types)
0.75
1.8
1.7
22.5
6.85
555
4250
6.7
3.2
4.8
4.6
7
0
9
250
4.2
340
• •
-27-
The washing machine wastewater showed somewhat higher relative and absolute concentrations
of congeners other than OCDD, compared to the wastewater samples. However, as for the
wastewater samples in general, OCDD was most of the total PCDD/PCDF mass for the washing
machine effluents.
Horstmann and McLachlan (1995) also sampled shower water, theorizing that PCDDs/PCDFs
from contaminated textiles (a major source of the PCDDs/PCDFs in household wastewater)
could be transferred to the skin surface and then washed off during showering. The median mass
of PCDDs/PCDFs in shower water (units presented in pg/shower) were as follows:
PCDD/PCDF Congener
2,3,7,8-TCDD
1,2,3,7,8-PeCDD
1,2,3,4,7,8-H6CDD
l,2,3,6,7,8-H6CDD
l ,2,3,7,8,9-H6CDD
1,2,3,4,6,7 ,8-H7CDD
OCDD
2,3,7,8-TCDF
1,2,3,7 ,8-PeCDF
2,3,4,7 ,8-PeCDF
1,2,3,4,7 ,8-H6CDF
l ,2,3,6,7,8-H6CDF
l,2,3,7,8,9-H6CDF
2,3,4,6,7,8-H6CDF
l ,2,3,4,6,7,8-H7CDF
l ,2,3,4,7,8,9-H7CDF
OCDF
Median Quantity, (pg/shower; five samples)
0
0
0
40
0
670
7600
9.6
0
19
11
0
0
0
230
0
230
The shower samples had relative amounts of PCDD/PCDFs that were similar to those observed
in the washing machine effluents.
Analyses of digested sewage sludges by Sewart et al (1995) provide another indication of the
distributions of PCDDs/PCDFs associated with domestic wastewaters. Their studies found that
for urban sludges, PCDD/PCDF congener profiles indicated a major input to the observed
contamination from pentachlorophenol. Less contaminated sewage sludges were observed in
samples from rural areas of England. Two samples of these sludges are identified as
"domestic/rural" versus "urban" or "industrial/urban." Analyses of the two samples of
"domestic/ rural" sewage sludges indicated the following distribution of PCDDs/PCDFs:
'r .. _,.,: .,._
•
Sample 1, ng/Kg
2,3,7,8-TCDD 0
1,2,3,7,8-PeCDD 0
l,2,3,4,7,8-H6CDD 0
1,2,3,6,7,8-H6CDD 6
l ,2,3,7,8,9-H6CDD 0
1,2,3,4,6,7,8-H7CDD 210
OCDD 3000
2,3,7,8-TCDF 18
1,2,3,7,8-PeCDF 0
2,3,4 ,7 ,8-PeCDF 0
1,2,3,4,7,8-H6CDF 0
1,2,3,6,7,8-H6CDF 0
l ,2,3,7,8,9-H6CDF 0
2,3,4,6,7,8-H6CDF 0
1,2,3,4 ,6,7,8 H7CDF 70
1,2,3,4,7 ,8,9 H7CDF 0
OCDF 20
-28-
•
Sample 2, ng/Kg
0
0
0
16
0
760
6400
12
0
0
0
0
0
0
90
0
80
These samples both contained inconsequential amounts of most of the PCDD/PCDF congeners.
OCDD and 1,2,3,4,6,7,8-H7CDD accounted for more than 96% of the total toxic PCDD/PCDF
mass in both samples.
5.2.3 PCDDs and PCDFs as Contaminants in Chlorophenol Wood Preservatives
PCDDs and PCDFs are significant impurities in chlorophenols used as wood preservatives.
Treatment of wood with pentachlorophenol (PCP) occuJTed at the Koppers Site from 1968 until
1975 (Keystone Environmental Resources, Inc., 1992). Thus, the Site is a potential source of
PCDDs and PCDFs found in samples from nearby water-supply wells.
Of the PCDD and PCDF impurities in pentachlorophenol, it is the more highly chlorinated
PCDDs and PCDFs that have been found as the predominant contaminants (U.S. EPA, 2000b).
Several studies have considered the specific distribution of PCDDs and PCDFs in
pentachlorophenol, C6HCl5O or sodium pentachlorophenate, C 6Cl5NaO, a chemically similar
compound. Hagenmaier and Brunner (1987) found that in two samples of sodium
pentachlorophenate wood preservative, the more highly chlorinated PCDDs and PCDFs were the
predominant PCDDs and PCDFs present. PCDDs and PCDFs were present as impurities in
relatively high concentrations in the wood preservative formulations that were analyzed. Specific
congener concentrations were as follows:
2.3. 7 .8-TCDD
1,2,3,7,8-PeCDD
1,2.3.4,7,8-H6CDD
I ,2,3,6,7,8-H6CDD
I ,2.3,7,8,9-H6CDD
I ,2,3,4,6,7,8-H7CDD
OCDD
2,3,7,8-TCDF
1,2,3,7,8-PeCDF
2.3,4,7,8-PeCDF
I ,2,3,4,7,8-H6CDF
l,2,3,6,7,8-H6CDF
1,2,3,7,8,9-H6CDF
2,3.4,6,7,8-H6CDF
1,2,3.4,6,7,8-H?CDF
1,2,3,4,7,8.9-H7CDF
OCDF
•
Sample 1, ug/Kg
0.23
18.2
28.3
2034
282
9100
41600
1.8
8.2
6.6
48
69
0
87
699
675
37200
-29-•
Sample 2, ug/Kg
0.5 1
3.2
13.3
53
19
3800
32400
0.79
1.9
I.I
4.6
1.3
1.3
4.6
197
36
4250
As with samples of burnt wastes ashes and so fo1th, 2,3,7,8-T CDD was only a small proporti on
of the total PCDD and PCDF mass present in these wood preservati ve samples.
Two samples of pentachlorophenol were also analyzed and concentrations of PCDDs/PCDFs
repo1ted in Hagenmaier and Brunner (1987). These PCP samples indicated the followin g
amounts of PCDDs/PCDFs were present as impurities:
Sample I, ug/Kg Sample 2, ug/Kg
2,3,7,8-TCDD O 0
1,2,3,7,8-PeCDD I 2
1,2.3,4,7,8-H6CDD O 0
1,2,3,6,7,8-H6CDD 831 1480
I ,2,3.7,8,9-H6CDD 28 53
1,2,3,4,6,7,8-H?CDD 78000 99900
OCDD 733000 790000
2.3,7.8-TCDF O 0
1,2,3.7,8-PeCDF 0.5 0.2
2,3,4,7,8-PeCDF 1.5 0.9
l,2,3.4,7,8-H6CDF 125 163
1,2,3,6,7,8-H6CDF O 0
I ,2,3,7,8,9-H6CDF 32 146
2,3.4.6.7.8-H6CDF O 0
1,2,3,4,6,7,8-H7CDF 11 280 19940
l,2,3,4,7,8,9-H7CDF 637 980
OCDF 118000 137000
As with the two samples of sodium pentachlorophenate reported in the Hagenmaier and Brunner
reference, the principal PCDDs and PCDFs present in these PCP samples were the most hi ghly
chlorinated congeners, with OCDD being the bulk of the total PCDD/PCDF mass in both
samples.
Santi et aJ (1994) found a similar proportion of PCDDs/PCDFs in a sample of sodium
pentachlorophenate. In this sample, the more highly chlorinated PCDDs and OCDF were present
in very high concentrations, relative to the sodium pentachlorophenate samples described in the
Hagenmaier and Brunner reference:
•
PCDD/PCDF Congener
2,3,7,8-TCDD
1,2,3,7 ,8-PeCDD
1,2,3,4,7 ,8-H6CDD
1,2,3,6,7 ,8-H6CDD
1,2,3,7,8,9-H6CDD
1,2,3,4,6,7 ,8-H7CDD
OCDD
2,3,7,8-TCDF
1,2,3,7 ,8-PeCDF
2,3,4,7 ,8-PeCDF
1,2,3,4,7 ,8-H6CDF
1,2,3,6,7,8-H6CDF
1,2,3,7,8,9-H6CDF
2,3,4,6,7,8-H6CDF
1,2,3,4,6,7 ,8-H7CDF
1.2,3,4,7 ,8,9-H7CDF
OCDF
-30-
Concentration. ug/Kg
0.076
18.7
96
4410
328
175400
879000
0
0
0
27.6
21.9
9.8
103
9650
2080
114600
•
Jn another study of the PCDD/PCDF contamination in a sample of sodium pentach lorophenate,
Palmer et al (1988) found a similar distribution of PCDDs and PCDFs as in the other studies of
PCDDs and PCDFs in PCP or sodium pentachlorophenate fo rmulations:
PCDD/PCDF Congener
2,3,7,8-TCDD
1,2,3,7,8-PeCDD
I ,2,3,4,7,8-H6CDD
1,2.3,6,7,8 H6CDD
1,2,3,7,8,9 H6CDD
1,2,3,4,6,7,8 H7CDD
OCDD
2,3,7,8-TCDF
1,2,3, 7 ,8-PeCDF
2.3,4,7,8-PeCDF
l ,2,3,4,7 ,8-H6CDF
1,2,3,6,7,8 H6CDF
1,2,3,7,8,9 H6CDF
2,3,4,6,7,8-H6CDF
1,2,3,4,6,7,8 H7CDF
1,2,3,4,7,8,9 H7CDF
OCDF
Concentration. ug/Kg
0
28.3
0
4050
0
33800
81000
149
3 19
324
0
22'5-
480
0
6190
154
36000
Overall, for samples of _PCP and sodium pentachlorophenate, the most highly chlorinated
PCDD/PCDF congeners constitute almost all of the contaminant mass. Considered as per cent of
the total PCDD/PCDF mass present in these samples, the median per cent of the total mass for
each congener is as follows:
•
PCDD/PCDF Congener
2,3,7,8-TCDD
1,2,3,7,8-PeCDD
1,2,3,4,7,8-H6CDD
1,2,3,6,7,8 H6CDD
1,2,3,7,8,9 H6CDD
1,2,3,4,6,7,8 H7CDD
OCDD
2,3,7,8-TCDF
1,2,3,7,8-PeCDF
2,3,4,7,8-PeCDF
1,2,3,4,7,8-H6CDF
1,2,3,6,7,8 H6CDF
1,2,3,7,8,9 H6CDF
2,3,4,6,7,8-H6CDF
1,2,3,4,6,7 .8 H7CDF
1,2,3,4,7,8,9 H7CDF
OCDF
•
-31-
PCP/Sodium Pentachlorophenate
Median Per Cent of Total PCDD/PCDF Mass (6 samples)
0.0000032
0.00472
0.00405
0.256
0.01636
9.712
74.7
0.00097
0.00234
0.0014
0.0123
0.0025
0.0033
0.00434
1.006
0.094
12.79
The tabulation above demonstrates that OCDD is approximately 75% of the total PCDD/PCDF
!;._,_,,? mass in the PCP and sodium pentachlorophenate samples. OCDD, OCDF and 1,2,3,4,6,7,8
H7CDD combined represent approximately 97% of the total PCDD/PCDF mass in PCP and
sodium pentachlorophenate samples.
5.2.4 Comparative Analysis of PCDD/PCDF Distributions: Burned Wastes, Ashes and Related
Contaminated Media versus Domestic Wastewater/Sewage Sludges and Chlorophenol Wood
Preservatives
Investigations of PCDDs and PCDFs in burned wastes, ashes and related contaminated media,
domestic wastewaters/sewage sludges, and wood preservative formulations indicate that
generally, all three classes of materials or wastes contain the more highly chlorinated PCDDs or
PCDFs as the predominant contaminants. However, the relative proportions of the different
toxic PCDD and PCDF congeners may vary between these three classes of wastes or materials.
Therefore, environmental contamination may be more likely to be attributable to one or another
of these sources, based on contaminant distributions in the environment. This possibility was
investigated by preparing pie chart graphs for each material or waste type showing the relative
proportion of each PCDD and PCDF both collectively, and separated into dioxin and furan
subgroups.
Figure 11 shows the per cent of total PCDD/PCDF mass for each of the seventeen toxic PCDD
and PCDF congeners, for the three categories of PCDD/PCDF contaminant sources discussed
above. There are apparent characteristic congener distributions for each of the three potential
contaminant source categories. The patterns are summarized as follows:
----------------------------------------• •
-32-
Burned Wastes, Ashes, and Related Contaminated Media
► OCDD is the PCDD/PCDF generally found in the highest concentration, comprising as
much as approximatel y 55% of the total PCDD/PCDF congener mass.
► OCDF is less than 9% of the PCDD/PCDF congener mass in these samples and is
typically present at a lower concentration than l,2,3,4,6,7,8-H7CDF.
► 1,2,3,4,6,7,8-H7CDF represents at least 5% of the total PCDD/PCDF congener mass
► 1,2,3,4,6,7,8-H7CDD is typically more th an 12% of the PCDD/PCDF congener mass and
is almost always present at a higher concentratj on than OCDF.
► Typically, the toxic H6CDD congeners and 1,2,3,7,8-PeCDD are present at a combined
concentration of more than 5% of the total PCDD/PCDF mass.
► Typically, the toxic PeCDF and H6CDF congeners are present at a combined
concentration of more than 15% of the total PCDD/PCDF mass. If present, the toxic
PeCDF and H6CDF congeners have a higher total mass than mass of the toxic H6CDD
congeners and 1,2,3,7,8-PeCDD.
Domesti c Wastewaters and Sewage Sludges
► OCDD is more than 75% of the total toxic PCDD/PCDFs.
►
►
►
►
OCDF is less than 7% and typically less than 3% of the PCDD/PCDF mass rresent.
1,2,3,4 ,6,7 ,8-H7CDF represents less than 5% of the total PCDD/PCDF congener mass
1,2,3,4,6,7 ,8-H7CDD is 6% to 10% of the PCDD/PCDF congener mass and is always
present at a hi gher concentration than OCDF.
The combined mass of 1,2,3,7,8-PeCDD, H6CDD, PeCDF and H6CDF congeners
represent approximately 1 % or less of the total PCDD/PCDF mass.
Contaminants in Chlorophenol Wood Preservatives
► OCDD is the PCDD/PCDF generall y found in the highest concentration, comprising
approximately 45% to 80% of the total toxic PCDD/PCDF mass.
► OCDF is approximately 10% or more of the total PCDD/PCDF mass.
► OCDF is typically present in concentrations that equal or exceed the concentrations of
l ,2,3,4,6,7,8-H7CDD. .
► The combined mass of 1,2,3,7,8-PeCDD, H6CDD, PeCDF and H6CDF congeners
represent from less than l % to approximately 3% of the total PCDD/PCDF mass.
5.3 Contaminant Properties
Several properties of the toxic PCDD/PCDF congeners are important for the evaluation of the
source of those contaminants that are detected in samples from private water-supply wells around
the Koppers Site. These properties are related to the subsurface environmental mobility of the
PCDDs/PCDFs and the environmental persistence of the PCDDs/PCDFs in the subsurface.
• -33-• Figure 11. Per Cent of Total Mass for Combined PCDDs and PCDFs,
Burned Waste, Domestic Wastewater and Sewage Sludges,
and Pentachrorophenor Contaminant Sources
Burned wastes, Ashes, and Related Contaminated Media
Alawl et al, Alawl et al, Alawl et al, Alawi et al, Alawl et al, Alawl et al,
contaminated soil near contaminated soH near contaminated soil near contaminat.d soil near contamlnatad soil near contaminated soil near
burned solld waste burned solid waste burned solld waste burned solld waste burned solld waste burned solid waste
U.S. EPA, 1990
municipal solid
waste ash
U.S. EPA, 1990
municipal solid
waste ash
U.S. EPA, 1990
municfpa.l solid
wasttash
U.S. EPA, 1990
municipal solid
wuteash
U.S. EPA, 1990
municipal solid
waste ash
Wunderll et al,
waste wood
grate ash
Domestic Wastewaters and Sewage Sludges
Wundefll et al,
was1ewood
fty~h
Hofstmann and
Mclachlan,
apartment building
wastewater
Horstmann and
Mclachlan,
washing machine
effluent
Horstmann and
Mclachlan,
shower water
Pentachlorophenol and Sodium PentachJorophenate
Hagenmeler and Brunner, HaganllHlier and Brunner,
sodium pentachlorophenate sodium pentachlorophenate
Hagenmeier and Brunner,
pentachloJOphenol
Santi Mal,
sodium penlacfllorophenate
Hagenmeler and Brunner,
pentachlorophenol
Palmer et al,
sodium pentachloJOphenate
see refefences section for citations
Sewartetal,
domestic/rural
sewage sludge
Sewart et al,
domestic/rural
sewage sludge
Kev to Figure
■2,3,7,8-TCDD
■1,2,3,7,8-PeCDD
□1,2,3,4,7,8-H6CDD
□1 ,2,3,6,7,8 H6CDD
□1,2,3,7,8,9 H6CDD
■1,2,3,4,6,7,8 H7CDD
■OCDD
□2,3,7,8-TCDF
■1,2,3,7,8-PeCDF
■2,3,4,7,8-PeCDF
□ 1,2,3,4, 7 ,8-H6CDF
□1,2,3,6,7,8 H6CDF
□1,2,3,7,8,9 H6CDF
Gl2,3,4,6, 7,8-H6CDF
□1,2,3,4,6,7,8 H7CDF
■1,2,3,4,7,8,9 H7CDF
■OCDF
• •
-34-
Traditionally, PCDDs and PCDFs as a group were considered to be immobile, or virtually
immobile in ground water, because of their extremely hi gh affinity for organic carbon (the
presumption being that PCDDs and PCDFs would not be found in measurable concentrations in
ground water). This presumption is based upon both the solubility of the PCDDs/ PCDFs and
the soil-water partitioning properties of the PCDDs and PCDFs, where theory predicts virtually
complete partitioning of contaminants onto aquifer material, rather than the PCDD/PCDFs being
transported via ground water. Solubility and partitioning coefficients for the PCDD/PCDF
congeners are discussed in Sections 5.3.1 and 5.3.2 of this report.
5.3.1 Solubility of the PCDD/PCDF Congeners
The aqueous solubility of the PCDD/PCDF congeners defines the mass of the compound that
may be dissolved in water. Appendix 1 presents PCDD/PCDF aqueous solubility data derived
from a literature review. Several of the PCDD and PCDF congeners have a limited amount of
solubili ty data. As is apparent from Figure 12, the solubility of the PCDDs and PCDFs generally
decreases with increasing molecul ar weight (additional chlorine substitutions). Some of the
apparent exceptions to thi s generalization are likely a fu nction of the limited avai lability of data
and disparities between reported congener solubilities for the two principal data sets referenced
in this report. Regardless of this situation, both data sets show a general decrease in solubility
with increasing chlorine substitution, for both dioxin and furan congeners. The most highly
chl orinated PCDDs and OCDF have very low aqueous solubilities, averagin g approximately 1
part per trillion.
5.3.2 Organic Carbon Partitioning Coefficient of the PCDD/PCDF Congeners
The organic carbon partitioning coefficient is a measure of the affinity of an organic molecule for
organic carbon. Essentially, it represents the ratio of the solid phase (sorbed to organic carbon)
concentration to the aqueous (dissolved) phase concentration of an organic compound under
equilibrium conditions. Appendix 1 presents PCDD/PCDF organic carbon partitioning
coefficient data derived from a literature review. Similar to the case for aqueous solubility data,
several of the PCDD/PCDF congeners have a limited amount of organiccarbon partitioning
coeffi cient data. Organic carbon partitioning coefficient data could not be found for some
congeners. For these compounds, the organic carbon partitioning coefficient for a similar
congener with available data was used as a smTogate value in the environmental mobility
determinations (reference Table 3 below).
As can be seen from Figure 13, the more highly chlorinated PCDD and PCDF congeners
generally have a hi gher organic carbon partitioning coefficient than the less chlorinated
congeners. The notable exception to this generalization appears to be the 1,2,3,4,7,8-H6CDF
congener, which has the highest partitioning coefficient shown on the figure. There is only one
data point for this congener, while other congeners with partitioning coefficient data have more
available data. Considering information on sediment-water partitioning of the PCDDs/PCDFs
0
Figure 12. Average Aqueous Solubility (ug/L) of the PCDD/PCDF Congeners
0.2 0.4 0.6
2,3,7,8 TCDD
1,2,3,4,6,7,8, H7CDD
OCDD
--":"---r-1 ----1
OCOF
2,3,7,8 TCDF
1,2,3,7,8 PeCDF
0.8 1.2 1.4
1,2,3,6,7,8, H6CDD
dioxin congeners shown in red
furan congeners shown in blue
CDF
1,2,3,6,7,8 H6CDF
See Appendix 1 for Information on the Data Sources Used to Prepare this Figure
1.6
2,3,4,6,7,8
H6CDF
1.8
I w \J't
•
• •
-36-
presented in Govers and Krop (1998), the very high apparent organic carbon partitioning
coefficient value for the 1,2,3,4,7,8-H6CDF congener is probably an outlier or unreliable value,
and the actual organic carbon partitioning coefficient for this PCDF is probably less than the
organic carbon partitioning coefficients for the more chlorinated PCDFs. Note that Figure 13
suggests that the most highly chlorinated PCDDs have higher organic carbon partitioning
coefficients than the most chlorinated PCDFs, while the less chlorinated PCDDs have low
partitioning coefficients compared to their PCDF analogues.
5.3.3 Soil-Water Partitioning and Retardation of Ground-Waler Contaminants
5.3.3.l Partitioning Analysis
Under equilibrium partitioning theory, the ground-water mobility of organic contaminants is
reduced, relative to the movement of water, due to the partitioning of the organic compounds
onto aquifer materials. Typically, in near-surface environments and for low mobility compounds
such as the PCDDs/PCDFs, partitioning of organic contaminants onto aquifer materials is the
partitioning to the organic carbon present in the aquifer. Thus, the measure of the partitioning of
organic compounds in the field is the product of the contaminant-specific organic carbon
partitioning coefficient and the fraction of organic carbon that is present in the aquifer materials.
This product is termed the soil -water partitioning coefficient, or Kd. In some hydrogeologic
settings, organic molecules partition mostl y to inorganic material, because of a limited amount of
organic carbon present in the aquifer materials. This partitioning to inorganic aquifer materials is
more significant for the more mobile organic compounds. McCarty et al (1981) developed an
equation to calculate the critical level of organic carbon in the subsurface(/ oc), below which the
organic molecules will primarily be adsorbed onto mineral surfaces, rather than onto organi c
carbon in the aquifer. The cri tical organic carbon content depends upon the surface area of the
aquifer materials (Su) and the octanol-water partitioning coefficient (K0w) of the specific organic
compound. Typical surface areas for some geologic materials are reported in Olsen and Davis
(1990). The surface area of calcite (calcium carbonate, the principal component of limestone)
is l 2.5 m2/g; the surface area of clays varies with the type of clay (expanding clay minerals such
as montmorillonite have reportec1 surface areas of 50 to 150 m2/g, while kaolinite, a non-:SWelling
clay mineral, has a surface area of between 10 and 50 m2/g. Iron hydroxide, which could be a
component of rock fracture or granular coatings has a reported surface area of 300 m2/g. Organic
matter has a surface area of 1,900 m2/g. The specific surface area for samples of sandstone was
determined by Kleineidam et al (1999) to be in the range of 1 to approximately 4 m2/g.
The surface area of the aquifer materials at the Koppers Site, particularly in the anticipated zone
of principal horizontal contaminant transport (part of the A zone and the B zone; see Section 4.2
of this report), is expected to vary from very low for relatively clean, fractured sandstone bedrock
with open fractures to high, for weathered shale or more clayey bedrock units, particularl y with'
fractures containing some clay filling or clay or iron oxide coatings along the fracture walls. The
importance of clays in fracture fills is illustrated in an article by Wefer-Roehl et al (2001), where
an analysis of fracture filling material in a chalk formation in Israel indicated that with a 15%
0
Figure 13. Average Organic Carbon Partitioning Coefficient of the PCDD/PCDF
Congeners
5000000
2,3, 7,8 TCDD
1,2,3,7,8 PeCDD
I
1,2,3,4,7,8 H6CDD
I
1,2,3,6,7,8-H6CDD no data
1,2,3,7,8,9-H6CDD no data
10000000 15000000
OCDD
20000000 25000000
dioxin congeners shown in red
furen congeners shown in blue
30000000
1,2,3,4,7,8 H6CDF
1,2,3,6,7,8-H6CDF no data
1,2,3,7,8,9-HSCDF no data
2,3,4,6,7,8-H6CDF no data
I
1,2,3,4,6,7,8 H7CDF
OCDF
See Appendix 1 for Information on the Data Sources Used to Prepare this Figure
I vl .....:i I
• •
-38-
clay content (comprised of expanding smectite and palygorskite clays), the specific surface area
was 82.6 m2/g, while fresh chalk with less than 5% clay (80% calcite) had a specific surface area
of 11.7 m2/g.
The critical organic carbon content at which the PCDD/PCDF compounds are expected to
partition mostly to the inorganic fraction of the aquifer materials is anticipated to be very low,
because of the very high octanol-water partition coefficient of these compounds. Using a value
of 100 m2/g for the surface area and considering the PCDD/PCDF with the lowest octanol-water
partitioning coefficient (2,3,7,8 tetrachloro-dibenzo-furan; reference Mackay et al 1992; Govers
and Krop, 1998), the method of McCaity et al (1981) indicates anf oc of 0.00000175, or an
organic carbon concentration of 1.75 mg/Kg. This organic carbon content is likely to be
exceeded in even the deeper, predominantly inorganic aquifer materials at the Koppers Site. The
organic carbon content of deeply buried geologic materials has mostly been measured in studies
of petroleum source rocks, where organic carbon content would be expected to be relatively
significant. Several examples of the organic carbon content of various sedimentary rocks are
shown in Table 2. These examples are primarily from petroleum-related investigations, but
represent rocks that are relatively low in organic content.
Even in rocks of a non-sedimentary ori gin, some organic carbon has been noted. From Table 2
in Kleineidam et al (1999), the organic carbon content of igneous rock fragments (in sand or
gravel) was determined to be in the range of 0.00007 to 0.0001. Papelis (2001) reported an
organic carbon content of 30 mg/Kg (fraction of organic carbon of 0.00003) for a granite
specimen. The lowest fraction of organic carbon reported in Kleineindam et al (1999) for
sedimentary rock was 0.00004 (40 mg/Kg) for an unweathered, Triassic-aged sandstone.
Although a site-specific estimate could not be made of the organic carbon present in the aquifer
materials around the Koppers Site, a literature review indicates that at least some organic carbon
is likely present and would result in contaminant sorption. The RI Report (Keystone
Environmental Resources Inc., 1992) detected a minimum fraction of organic carbon of 0.000269
for a soil sample obtained from a depth of 6 to 8.5 feet below ground surface (reference
A-ppendix J, Table 3-2). The lowest literature-reported fraction of organic carbon that was found
for geologic materials is 0.00003, or roughly an order of magnitude less than the lowest Koppers
Site organic carbon content for near-surface soil samples. The organic carbon content in the rock
matrix in the vicinity of the Koppers Site is anticipated to be small, based on the environment of
deposition and the presence of a considerable thickness of reddish sedimentary rock near the
surface, indicative of an oxidizing depositional environment. Such a condition is not conducive
to the preservation of large amounts of organic material. Somewhat higher organic carbon
contents may be associated with the rock fracture walls, where some organic carbon transport
from a near-surface environment and subsequent deposition along the fracture walls may have
occurred.
~---------------------------------------~---• •
-39-
Table 2. Organic Carbon Contents of Various Sedimentary Rocks
reference rock type environment of sample depth fraction of
deposition organic carbon
Wefer-Roehl et Eocene chalk not specified rock sample from 0.0129 for rock;
al, 2001 · approximately 0.0038 for
280 feet deep; fracture filling
depth of fracture and 0.0026 for
fillings not fracture wall
reported coating
Akande et al, Yolde Formation, deltaic to >150 feet 0.001 to 0.129,
1998 a Cretaceous age nearshore marine median foe of
shale, siltstone, 0.0056 (9
sandstone and samples).
calcareous
mudstone
sequence
Jendrzejewski et Cretaceous age marine, depth of 0 to >600 feet approximately
al, 2001 gray marls deposition 0.002 to 0.006;
probably < 600 m mean foe of
0.0037 (n umerous
sa mples).
Erlich et al, 1999 Late Cretaceous marine deposition not specified Average orga nic
inner to outer carbon content of
shelf/slope deep 0.0094
water limestones
to shale, siltstone
and sandstones
Thornton et al, Triass ic age not specified not specified; Non-calcareous
2000 sandstone sample obtained sandstone had a -from a rock fraction of -quarry organic carbon of
0.00026.
Ishi ga et al, 1999 Miocene non-marine, not specified range from
sandstone, shallow-water 0.0005 in non-
mudstone and lacustrine marine sample to
conglomerate transitional to 0.0042 in marine-
shallow marine transitional
sample
If an organic carbon content of 0.0000027 (0.0lx the minimum observed site-specific soil
fraction of organic carbon; approximately 0. l x the lowest literature-reported fraction of organic
carbon) is used to estimate the aquifer-specific Kd for the PCDDs/PCDFs, the Kd estimates
r ,.. ... , -:~-~
• •
-40-
shown in Table 3 are obtained. For contaminant transport evaluation purposes, the Table 3 soi l-
water partitioning coefficients are conservative.
Table 3. Kd Estimates for the Aquifer at the Koppers Site
Organic Carbon Partitioning
Congener Coefficient (Koc)• Estimated Kd
2,3,7,8-TCDD 1,659,587 4.48
1,2,3,7,8-PeCDD 367,865 0.99
l ,2,3,4,7,8-H6CDD 1,047,129 2.83
l ,2,3,6,7,8-H6CDD no data 2.83'
1,2,3, 7 ,8,9-H6CDD no data 2.83'
l ,2,3,4,6,7,8-H7CDD 4,897,778 13.32
OCDD 12,022,644 32.46
2,3,7,8-TCDF 2,238,721 6.04
1,2,3,7,8-PeCDF 3,126,079 8.44
2,3,4,7,8-PeCDF 3,126,079 8.44
l ,2,3,4,7,8-H6CDF 25,118,864+ 8.44 ~
l ,2,3,6,7,8-H6CDF no data 8.44 ~
l ,2,3,7,8,9-H6CDF no data 8.44 ~
2,3 ,4,6, 7 ,8 -H6CDF no data 8.44 ~
1,2,3,4,6,7 ,8-H7CDF 2,344,229 6.33
l,2,3,4,7,8,9-H7CDF 707,946 1.9 1
OCDF 5,623,413 15.18
• data from Appendix I
The Kd for this congener is assumed to be equal to the estimated Kd for the l,2,3.4,7,8-H6CDD congener.
+ This value is considered an outlier or unreliable value (see Section S.3.2) and an estimate for the Kd equi valent to that
estimated for the PeCDF congeners is assigned to the H6CDF congeners, on the basis of the Kd value generally increasing with
increasing chlorination.
~ The Kd for this congener is assumed to be equal to the estimated Kd for the 2,3,4,7,8-PeCDF congeners.
• •
-41 -
5.3.3.2 Retardati on Factor
The retardation of a contaminant is the result of contaminant sorption to the aquifer materials. It
is a convenient method of describing the velocity of the contaminant in ground water relati ve to
the ground-water velocity. For saturated porous medi a, the retardati on factor, R1, is calculated as
follows:
where :
R1 is the retardation factor
Kd is the soil-water partition coefficient
A is the soil bulk density
nc is the aquifer effecti ve porosity
(1)
Equation (1) applies to saturated porous media. For fractured media, the retardation of a
contaminant is more complex. This complexity ari ses because of not only the presence of
secondary (fracture) porosity, w hich o ften is the primary, if not sole feature contributing to the
aquifer hydraulic conductivity, but also due to the frequentl y dual-porosity nature of fractured
rock aquifers. In the dual porosity aquifer, there is the potential for contaminant movement into
or though both rock fractures and pores within th e aquifer matrix. For contaminant transport in a
dual porosity aquifer, contaminants may be (a) sorbed onto fracture surface, (b) sorb onto the
inner pore surfaces or (c) may diffuse into "stagnant" water in pores of the rock.
For the Koppers Site, it is conservatively assumed that the dual porosity model does not apply
and that contaminant transport is entirely within the fracture porosity. With this assumpti on, the
velocity of a solute that is reversibly sorbed onto a fracture surface (u) is estimated usi ng
equation (3) from Moreno et al (1997):
LI o S
"s = 8 + 2 K.
where:
us is the solute velocity
u0 is the Darcy velocity;
S is the average spacing between fractures
o is the fracture aperture
and K,, is the sorption coefficient.
the water velocity u , may be expressed as a functi on of the average spacing between fractures, S
and the fracture aperture, o, such that u = (ucfi)/f, (Moreno et al, 1997).
• •
-42-
Equation (3) from Moreno et al requires an understanding of the sorption (partition ing)
coefficient, the fracture aperture, the spacing between rock fractures, and the Darcy velocity. The
water velocity in the fractures is calculated from the Darcy velocity, spacing between fractures,
and fracture aperture.
The Darcy velocity can be estimated from the aquifer hydraulic conductivity and the hydraulic
gradient. The sorption coefficient is estimated for the PCDDs/PCDFs in Table 3. The fracture
spacing and aperture are unknown, but can be estimated, within reasonable bounds, based on the
Koppers Site aquifer hydraulic properties, descriptions of geoph ysical logs from the Koppers Site
Remedial Investigation Report, and literature data for rock fracture apertures in relatively shallow
elastic sedimentary rock formations.
The Darcy velocity is the product of the aquifer hydraulic conductivity and the hydraulic
gradient. As noted in Section 4.2.2, the ground-water fl ow along the major hydraulic
conductivity tensor is estimated to occur under conditions of a hydraulic conductivity of
0.8482 ft/d and a hydraulic gradient of 0.00777. These values equate to a Darcy flow velocity of
0.00659 ft/d (0.00201 mid).
Based on the Remedial Investigation descriptions of the geophysical logs and hydraulic
,.~:.c conductivity data, the spacing between rock fractures at the Koppers Site is probably best Cr,.,.•:-, ,-.... characterized as being in increments of feet, rather than inches. Spacing of fractures is by no
means uniform, and some spatially isolated bedrock zones may have fracture spacings of
centimeters or inches . However, as a generalization, fracture spacings on the order of one foot or
greater likely apply to the upper bedrock at the Koppers Site.
The geologic literature provides some data on both the fracture spacings and apertures noted for
sedimentary rocks similar to those found at the Koppers Site. Wealthall et al (2001) report that
in a Permo-Triassic sandstone in no1thwest England the mean rock fracture ape1ture was 5.4
millimeters (mm) and the mean aperture spacing was 2.07 meters (m) for rock to a depth of 40
meters. This sandstone is a major aquifer. In that study, larger fractures were sedi~t filled,
which would impede ground-water flow and contaminant transport. Parker et al (1994)
summarize field-measured fracture apertures for different sedimentary rocks. They report
apertures of between 0.08 and 0.2 mm for a sandstone aquifer and apertures ranging from 0. J 2 to
0.2 mm for a shale aquitard.
As an approximation, a fracture aperture of l mm (0.001 m) with a fracture spacing of 5 feet
(1.524 m) was used to estimate the water velocity through the fractured rock at the Koppers Site.
With these values, the water velocity is estimated to be 3.06 mid using the equation cited in
Moreno et al, 1997.
The contaminant transport velocity can then be estimated from equation (3) in Moreno et al
(1997). An alternate method for evaluating the contaminant transport velocity in fractured medi a
• •
-43-
is presented as equation 9.18 in Freeze and Cherry (1979). In this equation:
~ =l +2Kn
V e b
where: V is the ground-water velocity
V c is the contaminant transport velocity
Kn is the sorption coefficie nt and
b is the fracture ape11ure.
For a ground-water fracture flow velocity of 3.06 mid, the estimated contaminant transport
velocities are shown in Table 4, using both approaches for calculating the contaminant transport
velocity. These estimates should be considered as unce11ain approximations of the contaminant
transp011 velocities, because of the inherent uncertainties in calculating contaminant transpon
velocities in fractured rock. For thi s reason, conservative assumptions were used in estimating
values for the va1iables th at were used to estimate the contaminant transport velocities.
Table 4. Estimated Ground-Water Contaminant Transport Velocities
of PCDDs and PCDFs
Estimated Ground-Water
Estimated Ground-Water Contaminant Transpo11
Contaminant Transpo11 Velocity using Equation
Velocity using Equation (3) (9.18) from Freeze and
PCDDIPDCF Congener from Moreno et al, 1997 Cherry (1979)
2,3,7,8-TCDD 3.57E-4 mid 3.41E-4 mid
1,2,3,7,8,-PeCDD l .62E-3 mid l.54E-3 mid
1,2,3,4,7,8-H6CDD 5.65E-4 mid 5.41E-4 mid
1,2,3,6,7 ,8-H6CDD 5.65E-4 mid 5.41E-4 mid
l ,2,3,7,8,9-H6CDD 5.65E-4 mid 5.41E-4 mid
1,2,3,4,6,7,8-H7CDD l .20E-4 mid l.15E-4 mid
OCDD 4.93E-5 mid 4.7 lE-5 mid
2,3,7,8-TCDF 2.65E-4 mid 2.53E-4 mid
1,2,3,7,8-PeCDF l .90E-4 mid l.81E-4 mid
Table 4 is continued on page 44
'
• •
-44-
Table 4, continued
Estimated Ground-Water
Estimated Ground-Water Contaminant Transport
Contaminant Transport Velocity using Equation
Velocity using Equation (3) (9.18) from Freeze and
PCDD/PDCF Congener from Moreno et al, 1997 Cherry (1979)
2,3,4,7,8-PeCDF l .90E-4 mid 1.81 E-4 mid
l ,2,3,4,7,8-H6CDF 1.90E-4 mid 1.81 E-4 mid
l ,2,3,6,7,8-H6CDF l.90E-4 mid 1.81 E-4 mid
l ,2,3,7,8,9-H6CDF l.90E-4 mid 1.81 E-4 mid
2,3,4,6,7,8-H6CDF l .90E-4 mid 1.81 E-4 mid
l ,2,3,4,6,7,8-H7CDF 2.52E-4 mid 2.42 E-4 mid
1,2,3,4,7,8,9-H7CDF 8.38E-4 mid 8.01 E-4 mid
OCDF l.05E-4 mid 1.01 E-4 mid
While the analysis in Table 4 is a rough approximati on, it probably indicates a reasonable, order
of magnitude approximation of the ground-water transport velocities for the PCDDs and PCDFs.
F igure 14 graphically shows the relative magnitude of the ground-water mobility of the seventeen
dioxin and furan congeners, using the higher of the two estimates of contaminant velocity shown
in Table 4. Based on the analysis of ground-water mobility, 1,2,3,7,8-PeCDD is the most mobile
of the PCDDIPCDF congeners. Even at the estimated velocity shown for this congener the
annual rate of ground-water transport of 1,2,3,7,8-PeCDD is approximately 0.59 mly ( 1.94
ft/year). The retardation factors for the PCDDs and PCDFs are calculated as the ratio of the
ground-water velocity to the ground-water contaminant transport velocity. Considering that the
ground-water velocity in the bedrock is potentially over 3 meters per day (assuming ground-water
flow is generally through narrow, widely-spaced rock fractures), the retardation factor for the
PCDDslPCDFs ranges from a low of approximately 1890 for the 1,2,3,7,8-PeCDD congener to
approximately 62,070 for the OCDD congener.
5.3.4 Macromolecules and Potential PCDDIPCDF Ground-Water Contaminant Transpo1t
Typically, aqueous solubilities and organic carbon partitioning coefficients are useful indicators
of the subsurface environmental mobility of organic compounds. By these measures, the PCDJ?
and PCDF congeners would be considered as having extremely low mobility. At one ti me,
chlorinated dioxins (and by inference, chlorinated fu rans) were, as a group, considered to be
"immobile" (Olsen and Davis, 1990). As indicated in Section 5.3.3, there is the potential for
0
Figure 14. Estimated Ground-Water Transport Velocity of the PCDD/PCDF
Congeners
See Appendix 1 and Section 5.3.3.2 text for Information
on the Data Sources Used to Prepare this Figure
I
2,3,4,6,7,8, H6CDF
1,2,3,7,8,9 H6CDF
I
1,2,3,6,7,8 H6CDF
I
1,2,3,4,7,8 H6CDF
2,3,4,7,8 PeCDF
1,2,3,7,8 PeCDF
I
1,2,3,4,6,7,8 H7CDF
I
2,3,7 8 TCDF I
dioxin congeners shown In red
furan congeners shown in blue
2,3,7,8 TCDO
0.0002 0.0004
1,2,3,7,8,9 H6COD
1,2,3,6,7,8 H6CDD
I I
1,2,3,4,7,8 H6CDD
0.0006 0.0008 0.001 0.0012
Contaminant Transport Velocity, mid
0.0014
1,2,3,7,8
PeCDD
0.0016
~ Vo I
0.0018
• •
-46-
slow movement of PCDDs and PCDFs in the ground water beneath the Koppers Site, due to a
presumably very low organic carbon content, combined with a relatively high potential ground-
water flow velocity. However, a paper by Enfield and Bengtsson (1988) noted that organic
macromolecules (naturally occurring organic particles) may be transported through larger pore
spaces in an aquifer. A hydrophobic organic compound such as a PCDD or PCDF may bind to
such organic macromolecules, potentially greatly increasing the mobility of the hydrophobic
compound. The research by Enfield and Bengtsson suggested that for the 2,3,7,8-TCDD
congener, organic macromolecules present in ground water at a concentration of 100 mg/L would
cause the 2,3,7,8-TCDD mobility to be approximately 3x the contaminant mobility in the purely
. dissolved phase. Figure 15 shows this relationship. For higher organic macromolecule
concentrations, the mobility of 2,3,7,8-TCDD could increase substantially, relative to its
dissolved-phase mobility in ground water.
Although no·measurement of organic carbon in ground water was made during the Koppers Site
Remedial Investigation, measurements of organi c carbon in a number of hydrogeologic settings
(e.g. Starr and Gillham, 1993; Thurman, 1985) suggests that at the depth of the monitoring
intervals at the Koppers Site, the mobile organic carbon is unlikely to exceed a concentration of
10 to 20 mg/L. This concentration of organic carbon is equivalent to approximately 17 to 34
mg/L of organic matter, which could conceivably be present as organic macromoleCl!les. Thus,
the mobility of 2,3,7,8-TCDD is potentiall y approximately twice the 2,3,7 ,8-TCDD mobility for
conditions where no mobile macromolecules are present, based on Figure 15 . For any less
environmentall y mobile PCDDs/PCDFs, the effect of the organic carbon could be greater. For
example, considering Figure 15, a congener with a log Kow of 8 may, for an organic carbon
content of 10 to 20 mg/L (organic matter of 17 to 34 mg/L), have a relative mobility
approximately 3 to 4x greater than the dissolved-phase mobility of that congener.
Considering the potential sources of PCDD/PCDF contaminati on detected in water-supply wells
near the Koppers Site, there may be a substantially enhanced environmental mobility of any
PCDDs/PCDFs in septic tank effluent or sewage that may be present in leaking pipes. Wilhelm
et al (1994) reported that domestic wastewater contains approximately 0.2 g/L to 0.6 g/L (200 to
600 mg/L) of organic matter by_ weight. Depending on the subsuiface treatment of this organic
matter by microorganisms, varying amounts of this organic matter load may be converted to
byproducts such as carbon dioxide and thus varying amounts of the organic matter may reach the
ground water. In aerobic environments, septic tank effluent may be virtually stripped of organic
carbon. Wilhelm et al (1996) report analyses where septic tank effluents containing average
dissolved organic carbon of several tens of mg/L were reduced to background or near background
levels (less than 5 mg/L) at the water table. Ptacek (1998) reports that for a septic tank effluent
with a dissolved organic carbon (DOC) of 3 1.8 mg/L, 60 to 80% removal of the DOC occurred
between the septic tank drain field and proximal parts of the identified contaminant plume.
Similar reduction in DOC carbon was observed by Reemtsma et al (2000) for ground water
contaminated by infiltration from discharge of combined sewer overflow water (mixed domestic
sewage and storm water runoff).
~A,~) • -4 7-• Figure 15. Macromolecule Effect on the Mobility of Hydrophobic Compounds ......---.... • 1000 -:, u • • 0 :, c.> e • 0 -.. 0 u E • 0 E .. u ... • 0 E . .. 100 u :, C 0 • .I:. • .. >, . --.. . -CL 41 C • 0 -c.> :I -• E • u • > E .I:. .. c.> • . ... • -:; 0 a: >, 10 ... o: u >-0 .. --. : > i .. > C • -.. C • • CL .. Q. :< Q. ~ 1 2 • C • N C • C •• • u CC C • •• 0 .. N :I .. .I:. c-::I .. •o C • 0 D 0 a .. 0 0 I-.I:. CD u -" ,... ... IC -0 • I') GI I-Q < 0 % N I I I I I macromol•cul• conc•ntratlon (mg/L) 10 100 I I I 1000 / -·-·-10000 / 1/ 1(160 / / I I ; I / I / / i I , / I / • I I , t I ,, / ,' ;-- ---i20 . I I 1 i ,' / / I , / ,,, I .' I -------------I~ ----------,·.'.. ----- -----/ I / I /I , I ' / ,,,, I ,' I I / -I I / I/_, -/----r----,13 I .I / /'I ' ,/ ,,,, / / ,, / I ,,./ ,,,," .,, " /' ... -.,,,.,,, ,--.,,,,. ....... ~---: _______ _,,,-.-~ -I /, 1,2 I 4 e 8 log kow Fig. 2. Mobility of a hydrophobic compound relative to the moa;ility ·of the same comp~n_d without the presence of a macromolecule as a function of octanol-water partition coefficient and amount of organic carbon in the mobile phase. figure modified from Enfield and Bengtsson, 1988
• •
-48-
In anaerobic environments, oxidation of organic matter is much less significant, and thus more
significant organic carbon concentrations might be expected to reach the saturated zone from
percolation of septic tank effluent from a drain field, or out of a leaki ng sewer pipe. Presuming
that in a poorly functi9ning septic tank drain field or a leaky sewer pipe in a very anoxic
environment, an organic carbon concentration of 50 mg/L reached the ground water, the
concentration of organic macromolecules in the recharge to the ground water would potentially
be approximately 86 mg/L. This concentration would undoubtedly be reduced through
dispersion and other subsurface processes, but would result in mobility of hydrophobic
contaminants such as PCDDs/PCDFs that is somewhat hi gher than would be the case for
conditions where there is no point source anthropogenic organic carbon input.
5.4 Dioxins and Furans in Ground Water
There has been limited documentation in the literature of ground-water contamination by dioxin
and furan compounds. At the Koppers Company Superfund Site in Oroville, California, dioxins
and/or furans were found at a TEQ concentration of approxi mately 122 ng/L (0.122 ug/L) in an
area of highl y contaminated ground water with a large volume· of free-phase creosote
contamination (U.S. EPA, 1999). In another incidence of ground-water contamination by
PCDDs/PCDFs, a TEQ of 0.0158 ug/L was associated with free-phase contaminati on at the
Midland Products Site, a wood-preservative operation in Arkansas (U.S. EPA, 2001a). In EPA
Region 4, several ground-water samples obtained at the Escambia Wood Treating NPL Site
contained 2,3,7,8 dioxins and/or furans in concentrations of less than 0.001 ng/L (CDM Federal
Programs Corporation, 2000). In an interesting case·of significant source area dioxin and furan
contamination without accompanying ground-water contamin ation, a study of sediment and fish
tissue samples from a lake in Finland with significant levels of chlorophenol contamination
indicated that the source of that contaminati on was ground-water discharge to the lake. This
conclusion was made because of the absence of above-background concentrations of PCDDs
and PCDFs in the sediment and fish tissue samples. The wood preservative (fungicide) th at was
the source of the chlorophenol contamination contained a total of approximately 39,600 ug/Kg of
2,3,7,8 PCDD and PCDF congeners. The findings from this Finnish study were believed to
demonstrate that the lessenvironmentally mobile contaminants (PCDDs and PCDFs) had been
trapped by adsorption to aquifer materials and had therefore not reached the lake (Vartiainin et al,
1995).
Traditionally, PCDDs and PCDFs as a class were considered to be immobile, or vi11ually
immobile in ground water, because of their extremely high affinity for organic carbon. However,
as noted by Enfield and Bengtsson (1988), organic macromolecules that may be transported
through larger pore spaces in an aquifer can significantly alter the relative mobility of extremely
hydrophobic compounds, even when the amount of macromolecule is in concentrations typically
found in some ground water. The limited amount of documentation of PCDDs and PCDFs in
ground water may be a result of their being frequently overlooked as potential ground-water
contaminants during site investi gations.
• •
-49-
6. Distribution of Dioxins and Fu rans in the Soil and Ground Water at and Around the
Koppers Site
6.1 On-Site Contamination
6.1.1 Soil Contamination
Soil samples obtained during the Remedi al Investi gation were analyzed for 2,3,7,8-PCDDs and
PCDFs. Figure 16 shows the distribution of dioxin and furan congeners in soil samples
collected during the Remedial Investi gation. As is apparent from Figure 16, OCDD constitutes
the majority of the PCDD/PCDF mass in 19 of 20 soil samples shown and is the most prevalent
congener in all of the samples. The remainder of the PCDD/PCDF mass in many samples is
primaril y 1,2,3,4,6,7,8-H7CDD, which is more prevalent th an OCDF in 16 of the 20 soil
samples. OCDF is generally the third most prevalent congener, foll owed by l,2,3,4,6,7,8-
H7CDF. One sample (X48, 2 to 4 feet) contained a number of H6CDD and H6CDF congeners in
nearly equal concentrati ons, and in th at sample, 1,2,3,7 ,8 ,9 H6CDD was the third most prevalent
congener, at a concentration nearly equal to the 1,2,3,4,6,7 ,8-H7CDD mass. This sample is
atypical for the Site soils and more closely resembles contamination associated with some type of
burned waste rather than contamination associated with pentachlorophenol waste.
Compared to literature-reported distributions of PCDD and PCDF congeners in
pentachlorophenol and sodium pentachlorophenate, the soil contamination at the Koppers Site
shows generally higher concentrations of l ,2,3,4,6,7,8-H7CDD relative to OCDF. Also, unlike
the typical literature values for pentachlorophenol or sodium pentachlorophenate where fi ve of
six samples contained OCDF as a constituent representing more than 10% of the total
PCDD/PCDF mass, OCDF in the Koppers soil samples is typically between I % and 10% of the
total PCDD/PCDF mass (14 of 20 samples). With the one exception noted in the previous
paragraph, Koppers contaminated soil samples still represent material with a chlorophenol
PCDD/PCDF distribution signature, compared to a characteristic PCDD/PCDF signature for
burned wastes or wastewater.
6.1.2 Ground-W ater PCDD/PCDF Contamination
Ground-water contamination by PCDDs/PCDFs has been investigated at the Koppers Site on
several occasions. As far as is known, the first time these compounds were an alyzed in ground-
water samples was the 1992 Remedi al Investi gati on. Samples from monitoring wells both on
and off Site were analyzed for the 2,3,7,8-PCDD/PCDF congeners. Total concentrations of
PCDDs/PCDFs were also evaluated. For the on-Site moni toring wells only a subset of well s
were checked for the presence and concentrati ons of the PCDDs/PCDFs. Two rounds of
sampling occurred during the Remedi al Investigation. Three on-Site wells were sampled twice,
whereas other wells were sampled in either round 1 or round 2. All off-Site wells were sampled
once.
• •
Figure 16. Remedial Investigation Soil Samples with Dioxin and Furan Analyses
Relative Concentrations of Dioxin and Furan Congeners with TEFs
e
□12368H6CDD
■123789H6CDD
~ft • ■1234678H7CDD
.,,,,, 100{, aocoo
□2378TCOF
■ 12378PeCDF
• ■23478PeCOF
□ 123478H6CDF
□ 123678H6CDF
■234678H6CDF
■ 123789H6CDF
/ ■ 1234878H7CDF \
■ 1234789H7CDF
■OCDF
' 1000
1500
\ Uft
Magnitude of Total DioxinlFuran Contamination, nglKg
• •
-51-
Following the Site Remedial Investigation, the four off-S ite monitoring wells sampled during the
Remedial In vestigation were resampled in 2000. Periodic monitoring of PCDDs/PCDFs has
occurred at a few on-Site monitoring wells during the remedial acti on period. In addition to
monitoring well sampling, private water-supply wells have been sampled on several occasions
following the Remedial Investi gation and tested for the presence and concentration of the 2,3,7,8-
PCDD/PCDF congeners. PCDDs/PCDFs detected in those private wells are the focus of Part A
of this report.
Seven on-site wells were checked fo r the presence of PCDDs/PCDFs during the Remedial
Investigation. For at least some of this sampling, Table 4-32 in the RI Report indicates that
analytical results were rejected during data validati on, because of detection of those contaminants
in a blank sample. Assuming th at any Site-related contaminati on in these samples woul d have
been no greater than the reported concentrati ons that were rejected (a reasonable assumption,
based on resampling results, total concentrations reported, and other congener concentrations),
the concentrati ons of 2,3,7 ,8 PCDD/PCDF congeners in the ground water were, with exceptions
for one monitoring well , in the range of 0.001 ng/L to 0.66 ng/L. Some concentrations were also
repo,ted as an "estimated maximum potential concentrati on." These "estimated maximum
potential concentration" va lues were considered as valid representations of the sample
concentrations.
{·::l> Figure 17 shows the concentrations of 2,3,7,8-PCDDs/PCDFs detected in on-Site ground-water
samples obtained during the Remedial In vestigation. This figure omits data that were rejected
during the data validation process (some of the OCDD and 1,2,3,4,6,7 ,8-H7CDD results). The
figure indicates that in most of the samples, the contaminants found in the maximum
concentrations were OCDD, l ,2,3,4,6,7,8-H7CDD or OCDF. The exception to this condition is
at the C30A location, where a variety of PCDDs and PCDFs were found in relati vely similar
concentrations. OCDD data were rejected for both C30A samples, so it may have been present in
higher concentrations than the concentrations of the congeners that are shown on Figure 17.
Figure 17 implies that with the exception of the samples from C30A (particularly the round 2
sample), other monitoring locati ons yielded ground water that essenti ally contai ned only one or
more of the five congeners OCDD, 1,2,3,4,o,7,8-H7CDD, 2,3,4,6,7,8-H6CDF,
l ,2,3,4,6,7,8H7CDF, and OCDF. The only other monitoring location where congeners other than
these five congeners were detected was at C28A, where relatively low levels of several other
PCDDs/PCDFs were detected during both round 1 and round 2 sampling. This well also
contained the hi ghest concentration of total 2,3,7,8-PCDD/PCDFs, and, along with samples from
C-30A, the total toxicity equivalency values (TEQ) in ground water were highest for the C-28A
samples. Interestingly, RI data indicate that ground-water contamination by phenoli c compounds
in the C-28A and C-30A samples was relati vely inconsequential. It is unknown if samples that
were more highly contaminated by chlorophenols would have also contained relati vely high
concentrations of PCDDs/PCDFs. In the other five wells that were monitored for
PCDDs/PCDFs, the maximum concentration of any one congener was 0.06 ng/L.
2000
150
1000
• •
Figure 17. Remedial Investigation Ground-Water Samples with
Dioxin and Furan Analyses, On-Site Monitoring Wells
0
concentration, ng/L
0 O.G2 0.04
SCALE
250 500
0
round 1
concentration, ng/L
0 0.04 0.08
round 2
concentration, ng/L
0 0.1 0.2 0.3
iiE
Key to Figunp
1000 2000 2500
data source: Table 4-32, RI Report
E9 Round 1
concentration, ng/L
0 0.005 0.01
S Round 2
• Round 1 and 2 (note that rouid 1 data for well C4A ar-e not
shown becat.ae al ~rsweni non detect)
.. Approximate direction of ~ ground-waler c:ontmninant transport rn
concentration,ng/L
0 0.02 0.04
EB
concentration, ng/L
0 0.02 0.04 0.06
concentration, ng/L
0 0.01 0.02
round 1
concentration, ng/L
0 0.4 0.8
round 2
concentration,ng/L
0 6 12 18
• •
-53-
Figure 17 shows that PCDD/PCDF contamination was detected on Site in both the A and B zone.
The highest contaminant concentrations were found in two of the A zone monitoring wells, but
there are too few data points to make any definitive statement about the relative degree of
contamination in the two monitoring zones.
Based on the distribution of hydraulic head in the aquifer (reference Figure 2 through Figure 6),
the direction of the horizontal component of ground-water contaminant transport from the areas
of identified PCDD/PCDF ground-water contamination is also shown on Figure 17. There is
· some vertical component to contaminant transport, and that vertical component is generally
downward in the immediate vicinity of the Koppers Site, based on comparison of hydraulic head
at a number of paired A and B-zone wells. Figure 17 shows generalized directions of ground-
water contaminant transport, rather than presenting a more detailed analysis of the direction of
ground-water flow as was done in Section 4.2.2 above.
Following the Remedial Investigation, monitoring of PCDDs/PCDFs at a few of the on-Site
wells has occurred. Figure 18 shows the results of the most recent analyses from the on-Site
wells that are cuJTently monitored for PCDDs/PCDFs. Of the wells monitored during the
Remedial Investigation, only well C29B also has PCDD/PCDF data from the most recent
investigation. The 2001 sample from well C29B contained a reported 0.2 ng/L concentration of
OCDD, versus an OCDD concentration of 0 .06 ng/L reported from the Remedial Investigation.
Other monitoring wells that were sampled in 2001 indicated the presence of primarily the OCDD
congener in the ground water. The highest congener concentrations were found in the A zone
well that was monitored and the results from this well indicate a TEQ of 0.0244 ng/L using the
WHO TEF values (see Patt B, Section 3 and Pa,t B Table 4 for a di scussion and tabulation of
the TEF values for PCDDs/PCDFs). The 0.0244 ng/L TEQ exceeds the orth Carolina ground-
water standard for the 2,3,7,8-TCDD TEQ and provides some evidence of a remaining concern
about ground-water contamination at the Site.
6.2 Off-Site Contamination
6.2.1 Off-Site Monitoring Well PCDD/PCDF Ground-Water Contamination
Off-Site monito1ing wells have been sampled for PCDDs and PCDFs during both the Remedial
Investigation and during a follow-up EPA investigation in December 2000. Figure 19 shows the
detections of 2,3,7,8-PCDD/PCDF congeners for the RI sampling. Figure 20 shows detections of
2,3,7,8-PCDDs/PCDFs for the December 2000 samples. Both figures show that generally,
OCDD was the 2,3,7,8-PCDD/PCDF congener detected in the hi ghest concentrations in the off-
Site ground-water sarnp.les. OCDF was the congener with the highest concentration in one off-
Site RI ground-water sample. With the exception of .OCDD, OCDF, and a detection of
1,2,3,4,6,7 ,8H7CDD in the December 2000 sample, all of the off-Site PCDD/PCDF
congeners detected in ground water were found in concentrations of less than or equal to 0.006
ng/L.
• •
Figure 18. Recent Ground-Water Samples with
Dioxin and Furan Analyses, On-Site Monitoring Wells
concentratlon,ng/L
0 1 2 3 4
0 500 1000
Key to figure
1500 2000 2500
data source: Table 2, First Quarter 2001 Groundwater Monitoring Report, ThermoRetec Corporation
• A Zone Well
0 BZoneWell
$ Pumped Well (Ground-Water Remedial Well)
no detections
concentration, "?'-
0 0.2 0.4
concentration, ng/L
0 0.03 0.06 0.09
• • -55-
Figure 19. Remedial Investigation Ground-Water Samples with
Dioxin and Furan Analyses, Off-Site Monitoring Wells
concentration, ng/L
0 0.01 0.02
3500
3000
250
2000
s
C19C
s
C20C
SCALE
0 250 500
0
concentration, ng/L
0 0.02 0.04
500 1000
data source: Table 4-32, RI Report
concentration,ng/L
0 0.05 0.1 0.15
s
C21C
Key to Figure
2378TCDO
12378PeCOO
,...._ _ _, 123478H6CDD
1-----1 123678H6CDD
123789H6CDO
1234678H7CDD
OCDD
2378TCDF
12378PeCDF
23478PeCDF
r,..;,.;.,..a--...i 123478H6CDF
........ "--'--' 123678H6CDF
~~~234678H6CDF
i,.;...-..~,.,.. 123789H6CDF
1234678H7CDF
1234789H7CDF
OCDF
PCDDJPCDF Monitonng
e Round 1
s Round 2
• Round 1 and 2
concentration, ngJL
o 0.02 o.04 o.oe
s--+---. ;-----1--+---1
C16C
<.
1500 2000 2500
-7
• • -56-
Figure 20. 2000 Ground-Water Samples with Dioxin and
Furan Analyses, Off-Site Monitoring Wells
* C20C
SCA LE
0 250 500
0
concentration, ng/L
0 0.25 0.5 0.71 1
500 1000
* C21C
* C16C
1500 2000 2500
data source: Koppers Company Site, Morrisville, North Carolina Draft Field Investigation Report, May 2001,
U.S. EPA Region 4 Science and Ecosystem Support Division Project No. 01-0078
Key to figure
2378TCDO
12378PeCDD
1----1 123478H6CDD
1----1 123678H6CDD
123789H6CDD
1234678H7CDD
OCDD
2378TCDF
12378PeCDF
23478P..CDF
i-:.;.;~-1 123478H6CDF
'-"'" _ _.123678H6CDF
,._._~..., 234678H6CDF
123789H6CDF
1234678H7COF
1234789H7CDF
OCDF * non detects
• •
-57-
Off-Site monitoring well sample PCDD/PCDF concentrations from the Remedial Investigation
were compared to on-Site PCDD/PCDF monitoring well sample concentrations from the RI. It
was assumed that because there are only seven on-Site monitoring points, the overall degree of
ground-water contamination at the Site during the RI would be best represented by a composite
of all of the on-Site ground-water quality data, or by the maximum contaminant concentration
observed in a specific sample, rather than subdividing the Site into areas upgradient of a
particular off-Site monitoring well and comparing the upgradient to the downgradient ground-
water quality. This approach was considered to be conservative with regard to the contamination
potentially mi grating northward, since the available surface soil PCDD/PCDF data suggest a
more heavily contaminated area of surface soil above ground-water fl ow with a southward
hori zontal flow component (compare Figure 16, Remedial Investigation Soil Sample data, to
figures showing ground-water flow).
An additi onal element of conservatism was added to the evaluation of potential ground-water
PCDD/PCDF transport from on-Site to off-Site locations. In evaluation of contaminant transport
ti mes from on-Site to off-Site locations, distances were assumed to be from the closest part of the
Koppers Site to the off-Site well location. The principal sources of on-Site ground-water
contamination are in the fo1mer lagoon and wood treating areas north of the Fire Pond
((reference Figure 16) and thus the estimated time for on-Site contaminati on to travel from a
potential source to an off-Site well is probably overestimated.
Data from on-Site and off-Site wells were compared as follows: The median on-Site ground-
water concentration of each detected congener was compared to its concentration in each off-Site
monitoring well. While a compari son to the mean concentration in on-Site monitoring wells
would be more conservative, it is unlikely that the mean concentration for some of the congeners
is representati ve of the Site (source area) ground-water concentration, because of the extreme
difference between the mean and median values for those congeners, as a result of relatively hi gh
congener concentrations in one (or less commonly two) samples. Where such relatively high
values were not observed in any on-Site ground-water sample, a text discussion of the mean on-
Site ground-water concentration to the off-Site monitoring well concentration is included below.
It should be noted that this comparison of the median on-Site congener concentrations to
observed off-Site concentrations presumes that the calculated median on-Site concentrati ons are
actually representative of a true median concentration. This assumption is probably valid if the
calculated median s are considered to be rough estimates of true median concentrations.
Specific approaches that were used in calculation of the median Site ground-water concentrati ons
were as follows:
a. The congeners that were not detected in a sample were assigned a concentration of O.Sx
the reported detection limits.
b. Data from C28A and C30A (2 samples for each of these wells) were averaged for each
• •
-58-
well. Those well-speci fic averages were then used as a single well-representative value in
the calculations of the median Site concentrations. Al though not shown on Figure 17,
round l sample data were also obtained for monitori ng well C4-TP (C4A) and these
results were non-detect for all the 2,3,7,8 PCDD/PCDF congeners. These values were
also averaged with the round 2 C4A data to produce an average C4A concentration,
similar to the case for C28A and C30A.
c. "R" qualified data were omitted from the calculation of the on-Site median.
d. Data from the RI that were identified as being an "estimated maximum potential
concentration" were assumed to be valid concentrati on data.
e. Data from C30A were used in the calcul ations, despite the observation that because of the
congener di stribution, the C30A samples may represent PCDD/PCDF ground-water
contamination derived, or partly deri ved, from some burned material, which may or may
not be tied to the Site environmental contamination.
Figures 21 and 22 show the compari son of th e estimated source area (Site average) ground-water
concentration to the off-Site ground-water concentrations of detected congeners. A well by well
discussion follows:
C l 6C comparison
The 1,2,3,4,6,7,8-H7CDD and OCDD congeners were actually detected in the C l6C sample,
whil e the 2,3,7,8-TCDD and 2,3,4,6,7,8-H6CDF congeners were each reported as an "esti mated
potential maximum concentration." Figure 21 shows that the C l 6C concentration of 2,3,7,8-
TCDD exceeded the on-Site median concentration of that congener. The on-Site mean
concentrati on of 2,3,7,8-TCDD was 0.00268 ng/L, a value that is slightly less than the estimated
potential maximum concentration of 0.003 ng/L in the C l6C sample. For this congener, it
appears unlikely that the Site is the primary source of the contamination that may have been
present in the C l6C sample. The on-Site median concentration of 2,3,4,6,7,8-H 6CDF was also
slightly less than the concentration of 2,3,4,6,7 ,8-H6CDF observed in the C 16C sample. T he
mean on-Site concentration of 2,3,4,6,7,8-H6CDF was 0.01 578 ng/L, which is approximately
2.5x the concentration in the Cl6C sample. While thi s congener may have an on-Site source, it
appears unlikely the Site is the primary source of that congener, based on the comparison of the
median on-Site concentration to the Cl6C concentrati on. The Cl6C concentration of
l,2,3,4,6,7,8-H7CDD was much lower than the l ,2,3,4,6,7,8-H7CDD concentration th at is likely
to be characteristic of the Site ground water. This congener is more likely to have a significant
on-Site source, based on the comparison of the on-Site medi an concentration to the Cl 6C sample
concentration. It is noteworthy, however, th at the OCDD concentrati on in the Cl6C sample is
almost equal to the median on-Site OCDD concentration (0.05 ng/L versus 0.06 ng/L). Based on
the relative contaminant transp01t velocity and environmental mobility of OCDD versus
l ,2,3,4 ,6,7,8-H7CDD (reference Table 3 and Table 4 above), it would be anticipated that the ,
concentration of 1,2,3,4,6,7,8-H7CDD in the C16C sample would be more similar to the on-Site
concentration of that compound than would the concentration of OCDD in the C l 6C sample
compared to the median on-Site OCDD concentration. The fact that the C 16C 1,2,3,4,6,7,8-
• • -59-
Figure 21. Median Site (Source Area) Concentrations Compared to o,f-Site
Ground-Water Concentrations of PCDDs/PCDFs; RI Data for C16C and C19C
0
2378TCOO
12378PeCDD
123478H6CDD
123678HeCDD
123789H6CDO
1234678H7COO
OCDD
2378TCDF
12378PeCDF
23478PeCDF
123478H6CDF
123678H6CDF
234678H6CDF
.i.......__,1~
123789H6CDF
1234678H7CDF
1234789H7CDF
OCDF
2378TCOO
12378PeCOD
12.3478HGCDO
123678H6CDD
123789H6COD
1234678H7CDD
OCDD
2378TCDF
12378PeCDF
23478PeCDF
1234'78H8CDF
123678H6CDF
234678H6CDF
123789H6CDF
1234678H7CDF
1234789H7CDF
OCDF
0
0.01 0.02
0.01 0.02
*C16C concentration
conc»1 lb ation, nglL
0.03 0.04 0.05
*C19C concentration
concentration, nglL
0.03 0.04 0.05
0.06 0.07
0.06 0.07
• -60-•
Figure 22. Median Site (Source Area) Concentrations Compared to Off-Site
Ground-Water Concentrations of PCDDs/PCDFs; RI Data for C20C and C21C
2378TCDD
12378P.CDO
123478H8CDD
1231S78H6COO
123789H6CDO
1234678H7CDO
OCDO
2378TCDF
12378PeCDF
23478PeCDF
123478HeCOF
123878HICOF
234871HeCDF
123789HSCDF
1234S78H7COF
1234789H7CDF
OCDF
2378TCDD
12378PeCDD
123471HeCDO
1231r78H6CDO
123789H6CDO
1234678H7COD
OCDO
2378TCDF
12378P.COF
23478PeCDF
123478HeCOF
123CJ78H6COF
234678tteCDF
123789H8COF
1234678H7CDF
1234780H7CDF
OCDF
0 0.01 0.02
0 0.01 0.02
* C20C concentration
c:oncenlndion, nglL
0.03 0.04 0.05
* C21 C concentration
conce.ib ation, nglL
0.03 0.04 0.05
0.06 0.07
0.0& 0.07
• •
-61-
H7CDD concentration is an order of magnitude less than the on-Site median concentration does
not appear to be consistent with the relative mobility of that congener, compared to the mobility
of OCDD. Note that the estimated ground-water transport velocity of l,2,3,4,6,7,8-H7CDD is
0.00012 mid. C-16C is almost 360 meters distant from the Koppers Site boundary. If the
estimated contaminant transport velocity of 0.00012 mid for l ,2,3,4,6,7,8-H7CDD is an
underestimate of the actual transport velocity by a factor of 100, it would still require
approximately 82 years for 1,2,3,4,6,7,8-H7CDD to migrate from the Site boundary to the Cl6C
location. Thus, it appears improbable that the contamination at Cl6C is deri ved from a Koppers
Site source.
Cl 9C comparison
The Cl9C sample contained 0.04 ng/L of OCDD and an estimated maximum potential
concentration of 0.003 ng/L 2,3,4,6,7,8-H6CDF. These concentrations compare to median on-
Site concentrations of 0.06 ng/L OCDD and 0.005 ng/L 2,3,4,6,7,8-H6CDF. Conceivably, since
the median on-Site concentrations of both OCDD and 2,3,4,6,7,8-H6CDF exceed the Cl 9C
concentrations of these congeners, an on-Site source for the contamination detected in the Cl 9C
sample is possible. The absence of l ,2,3,4,6,7,8-H7CDD in the Cl9C sample (non-detect at
0.003 ng/L detection limit) is somewhat inconsistent with an on-Site source for the other
congeners in the C19C sample. This observation is made because (1) 1,2,3,4,6,7,8-H7CDD has
almost the same median on-Site ground-water concentration as OCDD, (2) 1,2,3,4,6,7,8-H7CDD
is more mobile than OCDD, and (3) 1,2,3,4,6,7,8-H7CDD is present in on-Site ground water at a
median concentration 8x greater than the on-Site median concentration of 2,3,4,6,7 ,8-H6CDF,
and is only slightly less environmentally mobile than th at congener.
The distance from C 19C to the closest part of the Koppers Site is more than 550 meters.
Considering the estimated ground-water contaminant transport velocities reported in Table 4, it
would require more than 30,000 years for OCDD to reach the location of Cl9C from the closest
part of the Koppers Site. Even if the ground-water contaminant transport velocity for OCDD was
two orders of magnitude greater than that estimated in Section 5.3.3.2 (0.00493 mid), it would
still require over 300 years for OCDD contamination, moving at the average estimated
contaminant transport velocity multiplied by a factor of 100, to reach the location of Cl9C. In a
similar analysis, the time for 2,3,4,6,7,8-H6CDF to move from the Site boundary to the Cl9C
location would be 79 years at a contaminant transport velocity lOOx the value presented in Table
4. Thus, although detected congener concentrations at Cl9C are less than median on-Site
concentrations of those congeners, it is unlikely that the contamination detected at this off-Site
monitoring location is derived from the Koppers Site.
C20C comparison
The RI sample from monitoring well C20C contained 2,3,7,8-TCDF at a concentration of 0.001
ng/L, l ,2,3,4,6,7,8-H7CDF at a concentration of 0.003 ng/L, OCDF at a concentration of 0.02
• •
-62-
ng/L, and 2,3,4,6,7,8-H6CDF at an estimated maximum potential concentration of 0.004 ng/L.
These concentrations are slightly less than the median on-Site concentration of those four
congeners, with the exception of OCDF in the C20C sample. The OCDF in the C20 sample is
equal to the on-Site median OCDF concentration. Considering both the ground-water mobility of
the OCDF congener (Table 4) and the fact that concentrations in the C20C sample are equal to
the median on-Site concentration of OCDF, it is unlikely that the OCDF contamination in the
C20C sample is derived from the Koppers Site. The same conclusion basically applies to the
other three congeners, although it is more likely that they could be derived from the Koppers
Site. However, the most likely congener to be derived from the Koppers Site based on relative
concentration and environmental mobility considerations is 2,3,7,8-TCDF, at a concentration of
2/3 the on-Site median concentration. The estimated ground-water transport velocity for this
congener is 0.000265 mid (0.097 m/y), and for a di stance of approximately 250 meters from the
closest part of the Koppers Site, it would require approximately 2,577 years for 2,3,7,8 TCDF
from the Site to reach the C20 well location. Even at a contaminant transport velocity lOOx
greater than the estimate from Table 4, the contaminant travel time would be over 25 years, or
longer than the repo1ted date when pentachlorophenol wood treatment began at the Site
(approximately 22 years before the RI sample collection).
C21C comparison
t _., The monitoring well C21C sample possibly contained a variety of PCDD/PCDF congeners.
With the exception of OCDD, all of the concentrations in the RI are repo1ted as "estimated
maximum potential concentrations." OCDD was detected in the C21C sample at a concentration
of 0.14 ng/L. This concentration is 2.3x the median OCDD concentration of the on-Site ground
water, makjng th e Site an unlikely source for most of the OCDD contamination, regardless of
environmental mobility considerations. Of the remaining PCDDs/PCDFs, the congener with the
estimated maximum potential concentration that was the lowest in the C21C sample relative to
the on-Site median concentration was l ,2,3,4,6,7,8-H7CDD at an estimated maximum potential
concentration of 0.005 ng/L. The estimated travel time for this congener to move from the Site
boundary to the location of C21C (approximately 500 meters) is more than 11 ,000 years. Even
allowing for some uncertainty in the contaminant transport velocity, and the fact that the leading
edge of contamination could precede the highest concentration that could ultimately be detected
at C21C due to longitudinal dispersion of the contaminant, it is unlikely that the 1,2,3,4,6,7,8-
H7CDD detected in the C21C sample is derived from a Koppers Site source.
It is useful to compare EPA's 2000 off-Site ground-water quality data to the RI ground-water
quality data as a further indication of the potential source of off-Site ground-water contamination.
Figure 20 shows that the only off-Site monitoring well sample with detected PCDD/PCDF
contaminants in the 2000 sampling was the sample from well Cl9C. This sample contained an
estimated 1 ng/L concentration of OCDD and 0.095 ng/L of l ,2,3,4,6,7,8-H7CDD. These
contaminants would be most likely to be detected as a result of Site-related ground-water
contaminants migrating to an off-Site location, based on their on-Site median ground-water
concentrations. However, they are some of the least likely congeners to migrate to such
• •
-63-
distances, based on their relative ground-water mobility (Table 4). They were also detected in
higher concentrations in the monitoring well sample than the median on-Site concentrations of
those two congeners found during the RI. In the Cl9C RI sample from 1990, OCDD was
detected at a concentration of 0.04 ng/L and l ,2,3,4,6,7,8-H7CDD was not detected. The
appearance of l,2,3,4,6,7,8-H7CDD in the 2000 sample from C19C following an earlier
detection of OCDD in the Cl9C RI sample would not be consistent with those ground-water
contaminants being transported from the Koppers Site. This conclusion is made because
1,2,3,4,6,7,8-H7CDD is a more environmentally mobile congener than OCDD (reference Table 3
and Table 4 above) and was present on-Site duting the RI at a median concentration of 0.04 ng/L,
versus a median 0.06 ng/L OCDD concentration. Therefore, l ,2,3,4,6,7,8-H7CDD should appear
first as a contaminant migrating with the ground water due to advective transport. Again, the
ground-water mobility of the congeners detected in either the RI C19C sample or EPA's 2000
sample is such that none of the PCDDs/PCDFs detected in Cl9C samples are likely to have been
derived from a Koppers Site source.
Considering th at the source of PCDDs/PCDFs in the off-Site ground-water samples from the
monitoring wells is probably not the Koppers Site, it is reasonable to ask why the PCDD/PCDF
contamination was detected in those off-Site samples. One possible source of PCDDs/PCDFs is
contaminant "can-y down" during monitoring well construction. This process may create
apparent ground-water contamination in the subsurface, due to the transport of surface soil
contamination downward during well constnJction, with subsequent entrapment of those surface
\·:z' soil particles in the collected ground-water samples. This apparent low-level ground-water
contaminati on would not trul y be present as dissolved or mobile contamination ; i.e. the detected
concentrations in the ground water are an artifact of well construction. This explanation is most
plausible for the three monitoring locations where some PCDD/PCDF contamination was
detected during the RI but not in the follow up sampling in 2000. Considering just those th ree
monitoring well s, the confirmed ground-water contamination (not including the repo11ed
"estimated maximum potential concentration" values) is as follows:
C l6C-l ,2,3,4,6,7,8H7CDD 0.004 ng/L; OCDD 0.05 ng/L
C20C-2,3,7,8-TCDF 0.001 ng/L; 1,2,3,4,6,7,8-H7CDF 0.003 ng/L; OCDF 0.02 ng/L
C21C-o e DD 0.14 ng[b.
According to EPA draft guidance (EPA, 2000c), typical background concentrations of dioxins (as
a TEQ) in rural soils are in the range of 1 to 6 pg/g and in urban soils, typical background
concentrations of dioxi ns range from 7 to 20 pg/g (as a TEQ). Given that the area around the
Koppers Site is not urban, but is not greatly removed from a number of possible PCDD/PCDF
sources (including, but not limited to, the Koppers Site), it is reasonable that a background soil
concentration (TEQ) is as much as approximately 10 pg/g (10 ng/Kg). This concentration is
presented as a TEQ value, rather than a congener-specific value. It would be equivalent to l 0
ng/Kg of 2,3,7,8 TCDD, the most toxic of the PCDDs/PCDFs, but would also be equal to a
proportionally greater concentration of a less toxic congener, such as OCDD (estimated toxicity
relative to 2,3,7,8-TCDD of 0.001, such that a TEQ of 10 ng/Kg would be equal to an absolute
concentration of 10,000 ng/Kg of OCDD).
• •
-64-
The relative amounts of the various congeners likely to be present in background soils was not
thoroughly evaluated for this report. A paper by Hoekstra et al (1999) reported that for four soil
samples from a beech forest, the natural concentrations of the PCDDs/PCDFs were as much as
95 ng/Kg 1,2,3,4,6,7,8-H7CDD, 360 ug/Kg OCDD, 11 ng/Kg 2,3,7,8-TCDF, 88 ng/Kg
1,2,3,46,7,8-H7CDF and 120 ng/Kg OCDF. Lower concentrations of PCDDs/PCDFs were
reported for four soil samples from a Douglas Fir forest. Taken together, the median
concentration of these congeners in the eight soil samples discussed in the Hoekstra et al paper
were: 1,2,3,4,6,7,8-H7CDD, 27.5 ng/Kg; OCDD, 105 ng/Kg; 2,3,7,8-TCDF, 3.4 ng/Kg;
1,2,3,4,6,7,8-H7CDF, 25 ng/Kg; and OCDF, 38.5 ng/Kg. For these soil samples the TEQ ranged
from 0.31 to 18 with a median value of 5.5 ng/Kg.
Taking the median concentrations of the PCDD/PCDF congeners detected in off-Site ground-
water samples around the Koppers Site and considering that the background concentrations of
PCDDs/PCDFs in the soils near the monitoring wells could be somewhat higher than the TEQ
for the median PCDF/PCDD concentrations reported in the paper by Hoekstra et al (potentially a
TEQ of 10 ng/Kg versus 5.5 ng/Kg), the estimated potential concentrations of the congeners in
off-Site soils around the Koppers Site was de,ived. This derivation involved multipl ying the
median concentrations from the Hoekstra et al reference by a higher potential TEQ that may
apply to soils around the Koppers Site (median congener concentration from Hoekstra et al •
10 ng/Kg TEQ/5.5 ng/Kg TEQ). Usin g this procedure, potential soil background concentrations
of the four congeners detected in the off-Site ground-water samples from C 16C, C20C, and
:·f .= C2 l C (RI data) are:
V
OCDD 191 ng/Kg
1,2,3,4,6,7,8 H7CDD 50 ng/Kg
1,2,3,4,6,7,8-H7CDF 45 ng/Kg
OCDF 70 ng/Kg
2,3,7,8 TCDF 6 ng/Kg
Contaminant carry down could result in a ground-water sample with some small amount of
surface soils material included in the sample. While it is unknown how much surface soil could
be incorporated in a ground-water sample from an incompletely developed monitoring well, it
would be improbable that more than a gram of surface material could be present in such a
ground-water sample. Assuming that a gram of surface soil material containing the potential
background concentration of PCDDs/PCDFs was incorporated in each ground-water sample
obtained from the off-Site wells during the RI, the total contaminant concentration in the ground-
water sample (assume a one-liter sample bottle for collection of dioxin samples, per U.S. EPA,
1997) would be, for OCDD (the congener with the highest concentration), 191 ng/Kg • 0.001 Kg
• 1 L = 0.191 ng/L. This concentration is slightly higher than the maximum OCDD
concentration detected in an off-Site monitoring well sample. For 2,3,7,8 TCDF, the congener
with the lowest concentration in both the soil and the off-Site ground-water samples, the
potential concentration in a one-liter sample containing one gram of surface soil material could
be approximately 0.006 ng/L. This concentration exceeds the maximum concentration of
2,3,7 ,8-TCDF detected in an off-Site monitoring well sample. This analysis suggests that surface
soil carry down could be a source of PCDD/PCDF contamination detected in the ground-water
' --,~-;)
• •
-65-
samples, if there was enough surface soil material entrained in a ground-water sample. The
absence of detected contamination in samples from C 16C, C20C and C2 l C when they were
resampled by EPA in 2000 is consistent with either a contaminant carry down explanati on or
more generally, a contamination adsorbed to particulate matter scenario, where sampling
technique has a potentially profound influence on PCDD/PCDF sample concentrations.
A second possibility for the observed PCDD/PCDF contamination in the off-Site ground-water
samples is that the dioxin is naturally occurring and is incorporated in some of the soil or aquifer
materi als. A paper by Ferrario et al (2000) presents the results of an analysis of clay samples
obtained from an open mining pit at a depth of 50 to 60 feet below land surface. The average
concentrations of 2,3,7,8-PCDDs in the raw clay samples included 20,640 ng/Kg of OCDD and
2,383 ng/Kg of 1,2,3,4,6,7 ,8-H7CDD. PCDFs were present a much lower concentrations in these
clay samples, and notable concentrations of PCDF congeners in the ground-water samples are
unlikely to be derived from such a potential source. Thus, while PCDD ground-water
contamination could be a result of naturally occurring PCDDs present in some aquifer materials
at depth, it is less probable that notable PCDF contamination·would be deri ved from such a
source. The paper by Ferrario et al repo1ts an average OCDF concentration of 11 ng/Kg in the
clay samples that were tested. Assuming that 1 gram of aquifer material contaminated with
OCDF at this concentration was present in a one liter ground-water sample, the anti cipated
reported concentration of OCDF in that sample would be 0.011 ng/L. This level is close to the
0.02 ng/L concentration of OCDF reported in the ground-water sample from C20C, and could be
hi gher, if there was a concentration of OCDF in deeper aquifer materials higher than the average
OCDF concentration repo1ted in the paper by Ferrario et al. Potential concentrations of all of the
PCDDs in the aquifer materials would be sufficient to account for the concentrati ons of those
congeners detected du1ing the RI in off-Site well samples, provided there was sufficient aquifer
material containing PCDDs entrained in the ground-water samples. Note, however, that the
analysis in the paragraph below di scussing the 2000 EPA sample from Cl9C makes the point
that such levels of PCDDs or PCDFs in the deeper aquifer materials are questi onable.
Finally, there could be some outside anthropogenic source for the PCDDs found in the off-Site
monitoring well samples. Such an outside anthropogeni c source cannot be identified from the
available data, but as noted above, there are many sources of PCDDs and PCDFs, some of which
can be associated with residential settings.
A notable concentration of 1 ng/L of OCDD was detected in the Cl 9C ground-water sample
obtained during the EPA sampling conducted in 2000. For this sampling acti vity, off-Site
monitoring wells were sampled and purged in a manner that minimized sample turbidity. T his
situation reduces the probability that the 1 ng/L concentration of OCDD in the ground-water
sample results from suspended solids present in the sample. Consider, however, that for OCDD
in the subsurface, the soil-water partitioning coefficient could be as low as 32.46 (reference Table
3 above) Even for low levels of OCDD in the aquifer materials, a 1 ng/L OCDD di ssolved-phase
OCDD concentration is possible. The same point applies to other congeners and other wells.
However, the absence of consistent detections of PCDDs/PCDFs in other off-Site wells from the
RI samples to the EPA 2000 samples suggests that for those wells, either contaminant carry down
• •
-66-
from the RI monitoring well construction or more generally, collection of samples with
particulate matter containing adsorbed PCDDs/PCDFs are the most probable scenarios
accounting for most of the PCDD/PCDF contamination found in off-Site moni toring well s
during the Remedi al Investigation.
6.2.2 PCDD/PCDF Contamination in Pri vate Water-Supply Well Samples
The focus of Part A of this report is on the potential source or sources of PCDDs and PCDFs
detected in private water-supply well s around the Koppers Site. Numerous pri vate wells have
been sampled during one or more periods foll owing the Remedial In vestigation. Figures 23, 24,
and 25 show the sampled pri vate wells and contaminant detecti ons from sampling in 1998, 1999,
and 2000. Tables 5 and 6 summari ze th e private well sampling th at occurred in 1998, 1999, and
2000.
T he figures and tables show several important facts with regard to the private well
contamination:
(1)
(2)
(3)
No PCDFs were detected in any of the private well samples
Detections of PCDDs are li mited to 1,2,3,4,6,7 ,8-H7CDD and OCDD
Detections of PCDDs are generally sporadi c for samples from a specific pri vate well. For
the ten private wells that were sampled more than once, onl y one well yielded more than
one sample with a detectable concentrati on of a PCDD.
The concentrations of l ,2,3,4,6,7,8-H7CDD and OCDD in the pri vate well samples were
compared to the concentrations of those congeners in off-Site monitoring wells. The off-Site
monitori ng wells ranged in distance from the Site from approximately 250 meters to 550 meters.
The monitoring wells are closer to the Koppers Site source areas than any of the private well s
with PCDD detects.
Figure 26 presents a comparati ve analysis of off-Site monitoring well sample concentrations to
private well sample concentrati ons for the 1,2,3,4,6,7,8-H7CDD and OCDD congeners. The
C-19C 2000 sample contained the highest concentrati on of both congeners. This monTtoring well _
is potentially upgradient of several private well s to the north of the Koppers Site, including wells
KP020, KP021, KP023, KP025, KP026, and KP027. The closest private well to C-19C that
contained measurable concentrations of one of the PCDDs is either KP025 or KP026. The 1998
samples from both of these wells contained no detectable PCDDs. The 2000 samples from these
wells contained OCDD. The OCDD concentration in the 2000 KP25 sample was 0.26 ng/L,
while the 2000 KP026 sample contained 0.007 ng/L of OCDD. The 2000 KP25 sample
contained a higher concentration of OCDD than the 1998 C-19C sample (or the OCDD
concentration in any other monitoring well sample from the RI), but OCDD in the KP25 was at
lower concentration t~an OCDD in the 2000 sample from C-19C.
• •
Figure 23. Private Well Samples with Dioxin
and Furan Detects, 1998 Sampling
concentration. ng/l concentration, nglL
0.000 0.oo1 0.I02 0.ool 0.0000 o.aot2 OJ10:M l.0036
Ell
1
1
1
t~~~~~-~
• . • ✓-. . rc.:~~~--
l "",;..~ ~-, ~ '
' ' .. , .,
r'' • , . ~ '1 •
.,: . '
...
concentration, nglL
0 0.04 0.08 0.12
~,';.1:-::--;, ......
base map: U.S. Geological SUrvey Cary. NC 7 .5 minute topowaphic map,
photorevised 1987
data sources: RI Report Figure 1-2
cor,ce.1bation, nglL
0.0000 0.0005 0.0010 0.0015
Key to Figure
2378TCDD
12378P8CDD
!-----1 123478H6CDD
.---'---I 123678H6CDO
123789H6CDD
123a78H7CDD
OCDD
237STCDF
12378PeCDF
23478PeCDF
i,;.,,;..,;.;,_.....,i 123478HICDF i,,..;.;.......,__, 123S78HICDF
µ.;....;.--'--4,,-1 234678HfCDF
123719HICDF
1234678H7CDF
1234789H7CDF
OCDF
* non detects * dloxln/furan detects
wi1h no TEQ defined
concentration, ng/L
0 0.07 0.14 0.21 0.28
concentration, ng/L
0 0.0012 0.0024
concentratlon,ng/L
0 0.0009 0.0018 0.0027
concentration, nglL
0 0.0008 0.0016 0.0024
1
1
1
• • -68-
Figure 24. Private Well Samples with Dioxin and
Furan Detects, 1999 Sampling
0 2000
concentratlon,ng/L
0 0.001 0.002 0.003
4000 6000 8000 10000 12000
base map: U.S. Geological Survey cary, NC 7.5 minute ~le map,
photorevised 1987
data M>Uroes: RI Report Flg\X& 1-2
Key to Figure
2378TCDD
12378PeCDD
i-----123478H6COO i-----123678H6CDD
123789H6COO
1234878H7CDD
OCOD
2318TCDF
12378PeCOF
23478PeCDF
.---123478H6COF
........, ........ .......--1 123678H6CDF
~ .......... --i 234678H6CDF
123789H6CDF
1234678H7CDF
1234789H7CDF
OCOF
* non detects
* dioxinlfuran detects
with no TEQ defined
• • -69-
Figure 25. Private Well Samples with Dioxin
and Furan Detects, 2000 Sampling
concentration, ng/L
0 0.008 0.016 0.024
concentration, nglL concentration, nglL
o 0.1 o.2 u o o.oo:s o.ooa o.ooe
1
1
0 2000 4000 6000 8000
~~ . ) '(
·---~·. . j ct i
t ·,,.01 >·.
~, .. "..,-~;; . ..,., : . ,~~ ·. . )
fflMN ~~d
., L
10000 12000 14000
base map: U.S. Geological Survey Cary, NC 7.5 minute topogn1phlc map,
photorevlsed 1987
data $OUfC8$: RI Report figure 1·2
Key to Figure
2378TCDD
12378PeCDD
1------1 123478H6CDD
..----1 123678H6CDD
123789H6COD
1234678H7COO
OCDD
2378TCDF
12378PeCDF
23478PeCDF
123478H6CDF
_,._,._.., 123678H6CDF
234678H6CDF
~........,...._. 123789H6CDF
1234678H7CDF
1234789H7CDF
OCDF
* non detects ' * dioxin/Juran detects
with no TEQ defined
concentration,ng/L
0 0.008 0.018 O.OZA
EH]
concentration, nglL
0 0.008 0.016 0.024
concentration,ng/L
0 0.008 0.016 0.024
• • -70-
Table 5. Summary of Private Well Sampling, 1998, 1999, and 2000
(reproduced from Tab le 4 in U.S. EPA, 2001b
Sample Number
ll" IHI()! l>\V
KP002PW
KP003PW
KPOO-IPW
KP005PW
KPOOSPW
Kl'009PW
KPOI IPW
KP012PW
KP015PW
K.P016PW
KP017PW
KP018PW
KP0191'W
KP020PW
KP021PW
KP022PW
KP023PW
KP024PW
KP025PW
KP026PW
KP0271'W
KP029PW
KP030PW
KP032PW
KP033PW
Table 4
Potable Well .Sampling, TEQ Value Comparison
Koppers Company
Morrisville, North Carolina
Sampling Event
12/01/1998 03/3011999
TIO() V ,\ I r JC t-.ir. 'I ',!I; nn
TEQ VALUE NG/L NS NS
TEQ VALUE NGIL 0 . 00002J 0.0
TEQ VAUJE NG/L .'IS NS
TEQ \' A LUE NG/L 0.00002] 0.0
TEQ V,\LlJE NG/L 1'S NS
TEO \'.\LUE NGIL 0.(Xl025 NS
TEQ \',\LUE NG/L NS NS
TEO \',\LUE NG/I. NS :--is
TEO VALUF. 1'G/L O.U(XXJ2J 0.0
TEQ VALUE NG/I. NS NS
TEQ VALUE NG/L NS NS
TEQ VALUE NG/L NS NS
TEO VALUE NG/L NS NS
'l'EQ VALUE NG/L 0.000181 0 . 0000 3 J
TEO VALUE NG/L 0 . ()()()() IJ 0.0
TEO \' ALUE NG/L 0.000031 0.0
TEO VALUE NG/L 0.0000::?J u.o -TEQ VALUE NG/I. NS 0.(l.._
TEQ VALUE NG/L NS NS
TEQ VALUE NG/L NS NS
TE() V /\LUE NG/L NS NS
TEQ VALUE NG/L NS NS
TEQ VALUE NG/L NS NS
TEO VALLIE NGIL NS 1'S
Tr.() VA LUE NG/L NS NS
Data Qualifiers J-Estimated value. NS -Not sampled
12/20/2000
... .,
0.0
0.0
0.0
OJI
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
NS
0.0
0.0
o.u
0.0
0.0003
0.00001 J
0. 00002 J
0.00002 J
0.00002 J
0.00002 J
0.0
• •
-7 1-
KP025 and KP026 are both approximately 1050 meters west-northwest from C-19C, and, at the
estimated ground-water transport velocity of 4.93E-5 mid, it would require over 58,000 years for
OCDD to migrate from C-19C to the location of those wells. There is some uncertainty in the
ground-water transport velocity for PCDDs and PCDFs in the vicinity of the Koppers Site.
However, it appears very improbable that OCDD at C-19C, even if it was related to a Koppers
Site source (that possibility is itself considered improbable), could have migrated another 1050
meters fu1ther from the Site to contaminate the ground water withdrawn from the closest water-
supply wells to the north .
Table 6. Congener-Specific PCDD/PCDF Detections in Potable Well Samples
Sample Approximate 1998 1998 1999 1999 2000 2000
Number Distance from 1,2,3,4,6, 7,8-OCDD, 1,2,3,4,6,7,8-OCDD, 1,2,3,4.6, 7 ,8-OCDD,
Site boundary, H7CDD. ng/L H7CDD, ng/L H7CDD, ng/L meters ng/L ng/L ng/L
KP003 1160 0.0025 0 0 0 0 0
KP005 744 0.0023 0 0 0 0 0
KP009 628 0 0.25 NS NS 0 0
KP015 907 0.0019 0 0 0 0 0
KP020 1825 0.0063 0.12 0.003 0 NS NS
KP021 1850 0.0013 0 0 0 0 0
KP022 1845 0.0033 0 0 0 0 0
KP023 1950 0.0015 0 0 0 0 0
KP025 1414 0 0 NS NS 0 0.26
KP026 1333 0 0 NS NS 0 0.007
KP027 1858 NS NS NS NS 0 0.024
KP029 1240 NS NS NS NS 0 0.02
KP030 1160 NS NS NS NS NS 0.019
KP032 1530 NS NS NS NS NS 0.02 ·
Other private wells north of the Koppers Site contained either l ,2,3,4,6,7,8-H7CDD or OCDD in
concentrations that were lower than the highest concentration in a C-19C monitoring well
• •
-72-
Figure 26. Comparison of 1,2,3,4,6, 7,8-H7CDD and OCDD
Concentrations in Private Well Samples to 1,2,3,4,6,7,8-H7CDD
and OCDD Concentrations in Off-Site Monitoring Well Samples
0.001
OCDD
0.01 0.1
1,2,3,4,s,1,s-t I i Ii iii ~ ~ II
H7CDD
0.001 0.01 0.1
Concentration, ng/L
e Monitoring Well 1,2,3,4,6, 7,8-H7CDD
e Private WeH 1,2,3,4,6,7,8-H7CDD
• Monitoring Well OCDD
A Private Well OCDD
1
1
• •
-73-
sample. However, these private wells are at greater di stances from the Koppers Site than are
K.P025 and K.P026, and it therefore appears extremely unlikely there is a Koppers Site source for
the PCDD detections in those samples.
Private wells to the east or northeast of the Koppers Site that yielded samples with measurable
amounts of OCDD include K.P029, K.P030, and K.P032. The OCDD concentrations in these
wells were lower than the OCDD concentration in the RI sample from monitoring well C-16C,
which is located approximately 360 meters east of the Site boundary. C-16C was resampled in
2000 and no OCDD was detected. The closest of these three private wells to C-16C is
approximately 800 meters further downgradient from the closest pa11 of the Site. Considering the
estimated ground-water velocity of OCDD (Table 4), it is also improbable that the private wells
to the east of the Koppers Site have yielded samples with contamination that is related to on-Site
ground-water contamination. Furthermore, the inconsistent presence of PCDD contamination at
C-16C suggests there is no Site-related plume of mobile-phase ground-water contamination that
extends as far as that location from the Site boundary.
As indicated by Figure 25 , no significant PCDD contamination (i .e. no contaminati on by a
2,3,7 ,8 congener) was detected in any of the thirteen pri vate wells to the south or southwest of
the Site when the wells were sampled in 2000. The 2000 sample from the private well at the
greatest distance southward from the Site did contain some H7CDF congener(s) other than those
,. with chlo1ine atoms in the 2,3,7,8 positions. " :~/-... .)-.t'I.. ..
Figure 24 shows that four wells south of the Koppers Site were sampled in J 999. Two wells
contained some TCDD congeners other than 2,3,7,8-TCDD while two of the four wells contained
no PCDDs.
Several private well samples from wells south or southwest of the Site did contain either
1,2,3,4,6,7,8-H7CDD or OCDD when they were sampled in 1998. Concentrations of
1,2,3,4,6,7,8-H7CDDD were lower in the private well samples south of the Site than the Site
ground water estimated median concentration of that congener, which could be consistent with a
Koppers Site source for that contamination. In contrast, the concentration of OCDD in the 1998
sample from KP009 exceeded the Koppers Site ground-water estimated median_ OCDD
concentration by a factor of more than 3. K.P009 is the closest to the Site of the private wells
located south of the Site (approximately 628 meters distant, per Table Sb). Considering (J) the
higher OCDD concentration in the 1998 KP009 sample relative to the Site median OCDD
ground-water concentration, (2) the estimated ground-water contaminant travel time from the
Site boundary to the location of K.P009 (more than 30,000 years, based on Table 4), and (3) the
apparent absence of OCDD contamination in the 2000 KP009 sample, it is improbable that the
KP009 OCDD detection is related to ground-water contaminant migration from the Koppers Site.
For the other private wells south or southwest of the Site, the travel time required for Site-related
1,2,3,4,6,7,8-H7CDD contamination to migrate to those wells via the ground water (an estimated
minimum time of almost 17,000 years, based on Table 4) also makes a Koppers Site source for
the 1998 detections of 1,2,3,4,6,7,8-H7CDD in those wells improbable.
• •
-74-
Another factor that suggests the Koppers Site is not the source for contaminants detected in th e
private water-supply wells is the complete absence of any PCDF congeners in any of the pri vate
well samples with PCDD contamination. As presented in Section 6.2 (Figure 21 and Figure 22),
the median OCDF ground-water concentration from on-Site monitoring wells is 0.02 ng/L.
While this concentration is less than either the median l ,2,3,4,6,7,8-H7CDD or OCDD ground-
water·concentrations at the Site, and OCDF is somewhat less environmentally mobile than
l,2,3,4,6,7,8-H7CDD (but more mobile than OCDD), it would be anticipated that some low-level
detection of OCDF would characterize at least some of the private well sample contamination, if
that contamination was related to a Koppers Site source. Note that although it is considered
improbable that PCDDs/PCDFs detected in the off-Site monitoring well samples are related to a
Koppers Site contaminant source, OCDF was detected in two of those off-Site monitoring well
samples. In at least one of those samples, OCDF was found in a hi gher concentration, relative to
the on-Site median concentration of OCDF, than the concentration of l ,2,3,4,6,78-H7CDD,
relative to the on-Site median concentration of that congener. Thus, if Site-related ground-water
contamination had moved to or beyond the off-Site monitoring wells, it is reasonable to conclude
that some of the private well samples with PCDD detections would have contained some
measurable concentration of OCDF.
7. Summary and Conclusions-Part A of the Report
f.-:,.~z, This repo1t presents an analysis of the source or sources of ground-water contamination by
~::-~,-polychlorinated dibenzo-p-dioxins and polychlorinated dibenzofurans (PCDDs/PCDFs) detected
in samples from several private water-supply wells in the vicinity of the Koppers Company, Inc.
NPL Site, Morrisville, No,th Carolina. The analysis of sources of these ground-water
contaminants considered the following suite of factors:
►
►
►
►
►
►
►
►
The Koppers Site history
The hydrogeologic setting as it relates to the rate and direction of ground-water
contaminant transport
Potential sources of PCDDs/PCDFs in the environment, and the different proportions and
concentrations of the 2,3,7,8-PCDD/PCDF congeners associated with contamination from
those sources
PCDD/PCDF properties as they relate to the environmental transport of those
contaminants
Koppers Site soil PCDD/PCDF contamination-congener distribution
Comparison of on-Site to off-Site ground water PCDD/PCDF distribution in monitoring
well samples
Comparison of <?ff-Site monitoring well samples to private well samples
Evaluation of alternative potential sources of private well PCDD/PCDF contamination
The following points summarize this evaluation of the PCDD/PCDF contamination of private
well samples around the Koppers Site:
• •
-75-
(1) Ground-water flow away from the Koppers Site is in multiple directions but with
principal northwest and southeast directions of flow due to anisotropic conditions. Along
the major directions of flow, the ground-water velocity in the upper part of the bedrock is
potentially as much as 6.59 ft/d. This high ground-water velocity is possible due to
ground-water flow p1imarily being through fractures which may be widely spaced, as
suggested by a low fracture porosity.
(2) The effective depth of significant aquifer hydraulic conductivity is probably in the range
of 60 to 100 feet. C-zone wells completed at depths greater than 100 feet typically
indicate minimal ground-water flow.
(3) Both Site monitoring data and the Site conceptual model indicate there are areas at the
Site where there is significant ve11ical hydraulic connection from near-surface weathered
bedrock and saturated soils into the shallow bedrock, and other areas where there is
limited potential for vertical contaminant transport.
(4) A literature review indicates there are numerous potential anthropogenic sources of
environmental contaminati on by PCDDs and PCDFs. Specific potential sources of these
contaminants that may have applicability to the private well contamination around the
Koppers Site include pentachlorophenol wood preservati ve waste, various types of
burned materials, and domestic wastewater. Each of these specific sources of
PCDDs/PCDFs have a characteristic signature with respect to the relative concentrations
of PCDDs/PCDFs in source or waste materials. These characteristic signatures were
defined as a possible indicator of PCDD/PCDF sources for the private well
contamination.
(5) PCDDs and PCDFs are considered as organic ground-water contaminants with relatively
low environmental mobilities. A review of the prope11ies of PCDDs/PCDFs related to
their environmental mobilities indicates that the likely ground-water transport velocities
of these compounds in the bedrock earth materials around the Koppers Site is probably in
the range of approximately 0.0000471 meters per day to 0.00154 meters per day (0.0564
feet per year to 1.844 feet per year). T hese velocities could be enhanced somewhat by
transport of PCDDs bound to large organic molecules that may be transpo1t ed through
more significant fracture zones. This process of enhanced contaminant transpo11 would
-enly be significant in situations where there is a significant mobile organic carbon load in
the ground water, such as ground water contaminated by untreated sewage.
(6) The rate of potential PCDD/PCDF ground-water transport from the Koppers Site was the
primary factor used to evaluate if the Koppers Site was the likely source of PCDDs/
PCDFs detected in the private well samples. Evaluation of the relative concentrations of
PCDDs and PCDFs in on~Site monito1ing wells, off-Site monitoring wells, and private
well samples was used to evaluate if the Koppers Site was potentially the primary source
of PCDD/PCDF contamination found in the water-supply well samples.
(7) T he available data indicate that the Koppers Site is an unlikely source of PCDD/PCDF
contamination detected in samples from the private water-supply wells. Alternate
potential sources of the PCDDs/PCDFs in private well samples include contaminant carry
down during well construction, naturally occurring PCDD/PCDF contamination in
geologic materials, or some localized anthropogenic source.
• •
-76-
8. References
Akande, S.A., 0. J. Ojo, B.D. Erdtmann and M. Hetenyi, 1998, "Paleoenvironments, Source
Rock Potential and Thermal Maturity of the Upper Benue Rift Basins, Nigeria: Implications for
Hydrocarbon Exploration," Organic Geochemistry, Volume 29, Issues 1-3.
Alawi, M.A., H. Wichmann, W. Lorenz, and M. Bahadir 1996, "Dioxins and Furans in the
Jordanian Environment," Chemosphere, Volume 32, No. 5.
ATSDR (Agency for Toxic Substances and Disease Registry), 1988, Toxicological Profile for
2,3, 7,8-Tetrachlorodibenzo-p-dioxin, A TSDR/TP-88/23.
CDM Federal Programs Corporation, 2000, Phase I Data Evaluation Report, Remedi al
Investigation/Feasibility Study for the Escambia Treating Company Site Operable Unit 2,
Pensacola, Florida, report prepared for U.S. EPA Region 4.
Cummings Riter Consultants, Inc., 1997, Remedial Action Report, Surface Water, Soil and
Groundwater Remediation, Koppers Company, Inc. Superfund Site, MoJTi sville, Not1h Caro lina.
Enfield, C.G., and G. Bengtsson, 1988, "Macromolecular Transpo1t of Hydrophobic
Contaminants in Aqueous E nvironments," Ground Water, Volume 26, No. I.
Erlich, R. N., 0. Macsotay I., A. J. Nederbragt and M. Antonieta Lorente, 1999, "Palaeoecology,
Palaeogeography and Depositional Environments of Upper Cretaceous Rocks of Western
Venezuela," Palaeogeography, Palaeoclimatology, Palae'?ecology, Volume 153, lssuesl-4.
Esposito, S.J., and N. R. Thomson, 1999, "Two-Phase Flow and Transpo11 in a Single
Fracture-Porous Medium System," Journal of Contaminant Hydrology, Volume 37, Issues 3-4.
FeJTario, J., C. Byrne, and D. Cleverly, 2000, "Summary of Evidence for the Possible Natural
Formation of Dioxins in Mined Clay Products," Organohalogen Compounds, Volume 46.
Fluor Daniel GTI, 1997, Six-Month Groundwater Remedial Action Monitoring Report, Former
Koppers Company, Inc. Superfund Site, Morrisville, North Carolina.
Freeze, R.A. and J.A. Chen·y, 1979, Groundwater, Prentice-Hall, Inc., Englewood Cliffs, NJ.
Govers, H.A.J., and Krop, H.B., 1998, "Partition Constants of Chlorinated Dibenzofurans and
Dibenzo-p-Dioxins," Chemosphere, Volume 37, Nos 9-12.
Hagenmaier H ., and H. Brunner, 1987, "Isomer-Specific Analysis of Pentachlorophenol and
Sodium Pentachlorophenate for 2,3,7,8-Substituted PCDD and PCDF at sub-ppb Levels",
Chemosphere, Volume 16.
.-,., ........ :::;)
• •
-77-
Hoekstra, E.J., H. de Weerd, E.W.B. De Leer and U.A.T. Brinkman, 1999, "Natural Formation
of Chlorinated Phenols, Dibenzo-p-Dioxins, and Dibenzofurans in Soil of a Douglas Fir Forest,"
Environmental Science and Technology, Volume 33, No. 15.
Horstmann, M., M. McLachlan, and M. Reissinger, 1992, "Investi gation of the Origin of
PCDD/CDF in Municipal Sewage," Organohalogen Compounds, Volume 9.
Horstmann, M , M .. McLachlan, and M. Reissinger, 1993, "Further Investigations of the Sources
of PCDD/CDF in Municipal Sewage,"Organohalogen Compounds, Volume 11.
Horstmann M., and M. McLachlan, 1994, "Polychlorinated Dibenzo-p-Dioxins and
Dibenzofurans in Textiles and Their Transfer to Human Skin, Sewage Sludge and other
Matrices," Organohalogen Compounds, Volume 20, 1994.
Horstmann M., and M. McLachlan, 1995, "Concentrations of Polychlorinated Dibenzo-p-
Dioxins (PCDD) and Dibenzofurans (PCDF) in Urban Runoff and Household Wastewaters,
Chemosphere, Volume 31, No. 3.
Hunt, R.E., 1984, Geotec/mical E11gi11eeri11.g Investigation Manual, McGraw Hill, New York.
Ishiga, H., K. Dozen and Y. Sampei, 1999, "Geochemical Constraints on Marine Invasion and
Provenance Change Related to the Opening of the Japan Sea: an Example from the Lower
Miocene Shales in the Hoda Section, Shimane Peninsula, SW Japan," Journal of Asian Earth
Sciences, Volume 17, Issue 4.
Jendrzejewski , L., R. Littke and J. Rullkotter, 2001 , "Organic Geochemistry and Depositional
History of Upper Albian Sediments from the Kirchrode I Borehole, Notthem Germany,"
Paleogeography, Paleoclimatology, Paleoecology Volume 174.
Keystone Environmental Resources, Inc., 1992, Final Revised Remedial Investigati on Report,
Former Koppers Company, Inc. Superfund Site, Morrisville, North Carolina.
Kleineidam, S., H. Rugner, B. Ligouis, and P. Grathwohl, 1999, "Organic Matter Facies and
Equilibrium Sorption of Phenanthrene," Environmental Science and Technology Volume 33, No.
10.
Krauss, T., P. Krauss and H. Hagenmaier, 1994, "Formation of PCDD/PCDF During
Composting?" Chemosphere Volume 28, No. 1.
Lohman, S.W., 1972, Ground-Water Hydraulics, U.S. Geological Survey Professional Paper
708, US. Government Printing Office, Washi ngton.
• •
-78-
Mackay, D ., W.Y. Shiu, and K.C. Ma, 19 92, Illustrated Handbook of Physical-Chemical
Properties and Environmental Fate f or Organic Chemicals, Volume II, Polynuclear Aromatic
Hydrocarbons, Polychlorinated Dioxins, and Dibenw furans, Lewis Publishers, Chelsea
Michigan.
McCarty, P.L., M . Reinhard, and B.E. Rittman, 1981, "Trace Organics in G roundwater,"
Environmental Science and Technology, Vol. 15, No. 1.
Moreno, L., B. Gylling, and I. Neretnieks, 1997, "Solute Transport in Fractured M edia -the
Important Mechanisms for Performance Assessment," Journal of Contaminant Hydrology,
Volume 25, Issues 3-4.
Nestrick, T.J. and L.L. Lamparski, 1983, "Assessment of Chlorinated Dibenzo-p-Dioxin
Formation and Potential Emission to the Environment from Wood Combustion," Chemosphere,
Vol 12, N o. 4/5.
Oberg, L.G., N . W agman, M. Koch and C. Rappe, 1994, "Polychlorinated Dibenzo-p-D ioxins.
Dibenzofurans and non-O1tho PCBs in Household Organic-Waste Compost and Matu re Garden
Waste Compost," Organohalogen Compounds, Volume 20.
Olsen, R.L., and A. Davis, 1990, "Predicting the Fate and T ranspo1t of Organic Compounds in
Groundwater," H azardous Materi als Control, May/June 1990.
Palmer, F.H., R.A. Sapudar, J.A. Heath, N J . Richard and G.W. Bowes, 1988 "Chlorinated
Dibenzo-p-Dioxin and Dibenzofuran Contamination in Californi a from C hlorophenol Wood
Preservative Use," California State W ater Resources Control Board, Report No. 88-5WQ.
C. Papelis, 2001, "Cation and Anion Sorption on Granite from the P roject Shoal Test Area, near
Fallon, Nevada, USA," Advances in Environmental Research, Volume 5, Issue 2.
Parker, B .L., R.W. Gillham, and J .A. Cherry, 1994, "Diffusive Disappearance oflmmiscible-
Phase Organic Liquids in Fractured Geologic Media," Ground W ater, Volume 32, No. 5.
Ptacek, C.J., 1998, "Geochemistry of a Septic-System Plume in a Coastal Barrier Bar, Point
Pelee, Ontario, Canada," Journal of Contaminant Hydrology, Volume 33, Issues 3-4.
Ree, K.C., E.H.G. E vers, and M. Van Der Be rg, 1988, "Mechanism of Formation of
Polychlorinated Dibenzo(p)dioxins and Polychlorinated Dibenzofurans (PCDFs) from Potential
Industrial Sources," Toxicological and Environmental Chemistry Volume 17, 1988.
Reemtsm a, T., R. Gnirb, and M. Jeke l, 2000, "Infiltration of Combined Sewer O verflow and
Tertiary Municipal Wastewater: an Integrated Laboratory and Fie ld Study o n Nutrients and
D issolved Organics," W ater Research, Volume 34, No.4.
• •
-79-
Ruelle, P., and U.W. Kesselring, 1997, "Aqueous Solubility Prediction of Environmentally
Important Chemicals from the Mobile Order Thermodynamics," Chemosphere, Volume 34,
No.2.
Ruokojarvi, P., M. Ettala, P. Rahkonen, J. Tarhanen, and J. Ruuskanen, 1995 "Polychlorinated
Dibenzo-P-Dioxins and -Furans (PCDDs and PCDFs) in Municipal Waste Landfill Fires,"
Chemosphere Volume 30 No. 9.
Santi, H ., L. Gruber, and E. Stohrer, 1994, "Some new Sources of Polychlorinated
Dibenzodioxins (PCDDs) and Dibenzofurans (PCDFs) in Waste Papers and Recycled Pulps,"
Chemosphere Volume 29.
Starr, R.C., and R.W . Gillham, 1993, "Denitrification and Organic Carbon Availability in Two
Aquifers," Ground Water Volume 31, Number 6.
ThermoRetec Corporation, 2001, First Qua1ter Groundwater Monitoring Repo1t for the Former
Koppers Industri es, Inc. Site, Morrisville, North Carolina; report submitted to U.S. EPA Region
4.
T hornton, S.F., M.I. Bright, D.N. Lerner and J.H. Tel lam, 2000, "Attenuation of Landfill
Leachate by UK T1iassic Sandstone Aquifer Mateiials; 2. Sorption and Degradation of Organic
Pollutants in Laboratory Columns," Journal of Contaminant Hydrology, Volume 43, Issues 3-4.
Thurman, E.M., 1985, Organic Geochemistry of Natural Waters, M. Nijhoff and W. Junk
Publishers, Boston.
Umlauf, G., M. Horstmann, P. Klein and J. Kurz, 1993 "Mass Balance of PCDD/F in a dry
Cleaning Machine," Organohalogen Compounds, Volume 11.
U.S. Army Corps of Engineers, 2000, First Five-Year Review Report for Koppers Co. Inc.
(Morrisville Piant) Morrisville, W ake County, North Carolina, draft repo1t Rrepared for U.S.
EPA Region 4.
U.S. EPA, 1990, Characterization of Municipal Waste Combustion Ash, Ash Extracts, and
Leachates, EPA-530-SW-90-029A, Office of Solid Waste and Emergency Response,
Washington.
U.S. EPA, 1997, EPA Region 4 Science and Ecosystem Support Division Environmental
Investigations Standard Operating Procedures and Quality Assurance Manual, Internet avai lable
at address http://www.epa.2:ov/reeion4/sesd/sesdpub guidance.html.
U.S. EPA, 1998, The Inventory of Sources of Dioxin in the United States, External Review Draft
Report, Internet available at address http://www.cpa.gov/ncca/pdfs/dioxin/dioxin.pdf.
• ••
-80-
U .S. EPA, 1999, Amendment #2 to the Record of Decision for the Soil and Ground Water
Operable Unit Koppers Company, Inc. Superfund Site, Oroville, California U.S. Environmental
Protection Agency Region 9 S an Francisco, Cali fornia.
U.S. EPA, 2000a, North Carolina NPL Site Summaries, Koppers Company, Inc., (Monisville
Plant), Internet available at http://www.epa.gov/region4/waste/npl/np1nc/koppernc.ht m.
U.S . EPA 2000b, U.S. EPA Toxicological Review of Pentachlorophenol (CAS No. 87-86-5) in
Support of Summary Information on the Integrated Risk Information System (IRIS), Preliminary
Draft Report, Internet available at address http://www.epa.gov/oppad001/pentach1 oropheno1.pd f.
U.S. EPA 2000c, Exposure and Human Health Reassessment of 2,3,7,8-Tetrachlorodibenzo-p-
Dioxin (TCDD) and Related Compounds, EPN600/P-00/001Ag, External Review Draft.
U.S. EPA, 2001a, Site Status Summary for the Midland Products Superfund Site, Internet
avail able at address http://www.epc.urnv/Region06/6sf/pdffiles/midland.pdf.
U.S. EPA, 2001b, Koppers Company Supcrfund Site, Monisville, North Carolina, Draft Field
Investi gati on Project, SESD Project No. 01 -0078, repo11 prepared for Beverly Hudson, Remedial
Project Manager, U.S. EPA Region 4.
Vmtiainin, T ., P. Lampi, K. Tolonene and J. Tuomisto, 1995, "Polychlorodibenzo-p-Dioxin and
Polyc hlorodibenzofuran Concentrations in Lake Sediments and Fish after a Ground Water
Pollution with Chlorophenols," Chemosphere Volume 30 , No. 8.
Wealthall, G.P., A. Steele, J .P. Bloomfield, R.H. Moss and D .N. Lerner, 2001 , "Sediment Filled
Fractures in the Pe1mo-Triassic Sandstones of the Cheshire Basin: Observati ons and
Implications for Pollutant Transpo11," Journal of Contaminant Hydrology, Volume 50, Issues
1-2.
Wefer-Roehl, A., E.R. Graber, M.D. Boiisover, E. Adar, R. Nativ, and Z. Ronen, 2001,
"Sorption of Organic Contaminants in a Fractured Chalk Fo1mation," Chemosphere, Volume 44,
No. 5.
Wilhelm, S.R., S.L Schiff, and J.C. Cherry, 1994, "Biogeochemical Evolution of Domestic
Waste Water in Septic Systems: 1. Conceptual Model," Ground Water, Volume 32, No. 6.
Wilhelm, S.R., S.L. Schiff, and W.D. Robertson, 1996, "Biogeochemical E volution of Domestic
Waste Water in Septic Systems: 2. Application of Conceptual Model in Sandy Aquifers," Ground
Water, Volume 34, No. ·5_
Wunderli , S., M. Zennegg, I.S. Dolezal, E. Gujer, U. Moser, M . Wolfensberger, P. Hasler, D.
Noger, C. Studer, and G. Karlaganis 2000, "Determination of Polychlorinated Dibenzo-p-
Dioxins and Dibenzo-Furans in Solid Residues from Wood Combustion by HRGC/HRMS,"
Chemosphere Volume 40.
• •
Appendix 1. PCDD and PCDF Structure and PhysicaVChemical Properties
•
2,3,7,8-tetrachlorodibenzo-p-dioxin (2.3,7,8-TCDD)
Cl
Cl
0
0
aq ueous solubility (ug/L):
(reference Mckay et al, 1992)
maximum reported solubility 19.4 ug/L
Cl
Cl
median reported solubility (16 values) 0.2 ug/L
organic carbon partitioning coefficient
(reference Mckay et al, 1992)
minimum repo1ted value 1148
median reported value (29 values) 1,659,587
•
lower quartile (75% of observations equal or exceed value) 467,735
1.2.3.7,8-pentachlorodibenzo-p-dioxin O ,2.3,7,8-PeCDD)
Cl
Cl
0
0
Cl
Cl
Cl
aqueous solubility (ug/L; data for the 1,2,3,4,7-PeCDD congener):
(reference Mckay et al, 1992)
maximum reported solubility 8.16 ug/L
median reported solubility (9 values) 0.19 ug/L
organic carbon partitioning coefficient (data for the 1,2,3,4,7-PeCDD congener):
(reference Mckay et al, 1992)
minimum reported value 70,795
median reported value (4 values) 367,835
• •
1,2.3.4,7,8-hexachloro-dibenzo-p-dioxin (l,2,3,4,7,8-H6CDD)
Cl
Cl
0
0
Cl
Cl
Cl
CI
aqueous solubility (ug/L):
(reference Mckay et al, 1992)
maximum repo11ed solubility 0.044 ug/L
median reported solubility (8 values) 0.005 1 ug/L
organic carbon partitioning coefficient
(reference Mckay et al, 1992)
minimum repo1ted value 104,712
median reported.value (29 values) 1,047,129
1,2,3.6,7 .8-hexachloro-dibenzo-p-dioxin ( I .2,3.6,7,8-H6CDD)
Cl
Cl
Cl
Cl
0
0
aqueous solubility (ug/L): 0.88
(reference Govers and Krop, 1998)
Cl
Cl
organic carbon partitionin g coefficient: no reliable data fou nd
• •
1,2,3,7,8,9-hexachloro-dibenzo-p-dioxin ( l .2.3,7,8,9-H6CDD)
Cl3C6HO2C6HC13
Cl
Cl
Cl
0
0
Cl
Cl
Cl
aqueous solubility (ug/L): no reliable data found
organic carbon partitioning coefficient: no reliable data found
l,2,3.4,6.7,8-heptachloro-dibenzo-p-dioxin ( I .2.3.4,6,7.8-H7CDD)
Cl
Cl
Cl
0
0
aqueous solubility (ug/L):
(reference Mckay et al, 1992)
Cl
Cl
maximum reported solubility 0.848 ug/L
Cl
Cl
median reported solubility (10 values) 0.0024 ug/L
organic carbon partitioning coefficient
(reference Mckay et al, 1992)
minimum reported value 295,121
median reported value (3 values) 4,897,788
•
Octachloro-dibenzo-p-dioxin (OCDD)
Cl
Cl
Cl
Cl
0
0
aqueous solubility (ug/L):
Cl
Cl
(reference Mckay et al, 1992)
maximum reported solubil ity 0.1 8 ug/L
Cl
Cl
medi an reported solubility (17 values) 0.0004 ug/L
organic carbon pa1titioning coefficient
(reference Mckay et al, 1992)
minimum repo1ted value 831,764
median repo11ed value (3 values) 12,022,644
2,3,7,8-tetrachlorodibenzo-furan (2,3,7,8-TCDF)
Cl
Cl
0
aqueous solubility (ug/L):
(reference Mckay et al, 1992)
Cl
Cl
maximum reported solubility 3.51 ug/L
median reported solubility (3 values) 0.4 ug/L
organic carbon partitioning coefficient
(reference Mckay et al, 1992)
minimum reported value 158,489
median(geometric mean) reported value (2 values) 2,238,721
•
•
1,2,3,7,8-pentachlorodibenzo-furan ( 1,2,3,7,8-PeCDF)
Cl
CI
0
Cl
Cl
Cl
•
aqueous solubility ug/L (data for the 2,3,4,7,8-PeCDF congener):
(reference Mckay et al, 1992)
maximum reported solubility 0.52 ug/L
mean reported solubility (2 values) 0.376 ug/L
organic carbon partitioning coefficient (data for the 2,3,4,7 ,8-PeCDF congener):
(reference Mckay et al, 1992) ·
minimum reported value 389,045
median(geometric mean) reported value (2 va lues) 3,126,079
2.3,4,7 .8-pentachlorodibenzo-furan (2.3.4.7 .8-PeCDF)
Cl
CI
0
aqueous solubility ug/L:
CI
(reference Mckay et al, 1992)
maximum reported solubility 0.52 ug/L
CI
Cl
mean reported solubility (2 values) 0 .376 ug/L
organic carbon partitioning coefficient:
(reference Mckay et al, 1992)
minimum reported value 389,045
median(geometric mean) reported value (2 values) 3,126,079
•
1,2,3,4,7,8-hexachlorodibenzo-furan ( l ,2,3,4.7,8-H6CDF)
Cl
Cl
0
aqueous solubility ug/L:
Cl
Cl
Cl
Cl
(reference Mckay et al, 1992; Govers and Krop, 1998)
median reported solubi lity (3 values) 0.801 ug/L
maximum solubility 2.654 ug/L
organic carbon pa1titioning coefficient:
(reference Mckay et al, 1992)
repo1ted value (1 value) 25,118,864
1,2,3,6,7,8-hexachlorodibenzo-furan (l .2.3.6,7.8-H6CDF)
Cl
Cl
Cl
0
aqueous solubility ug/L:
Cl
Cl
Cl
•
(reference Ruelle and Kesselring, 1997; Govers and Krop, 1998):
maximum reported solubility 2.26 ug/L
median reported solubi lity (4 values) 0.994 ug/L
organic carbon partitioning coeffi cient: no reliable data found
•
1,2,3,7,8,9-hexachlorodibenzo-furan O ,2,3,7,8.9-H6COF)
Cl
Cl
Cl
aqueous solubility ug/L:
Cl
0
(reference Govers and Krop, 1998):
reported solubility 0.859 ug/L
Cl
Cl
organic carbon partitioning coefficient: no reliable data found
2.3.4,6.7.8-hexachlorodibenzo-furan (2,3,4,6,7.8-H6CDF)
~,,..:. Cl3C6H2OC6HCl3
'>
Cl
Cl
Cl
0
aqueous solubilitYog!L:
Cl
(reference Govers and Krop, 1998):
reported solubility 1.563 ug/L
Cl
Cl
organic carbon partitioning coefficient: no reliable data found
•
. . ~;:;:'
• •
1,2,3,4.6,7 ,8-heptachlorodibenzo-furan (l,2,3,4,6,7,8-H7CDF)
Cl
Cl
Cl
aqueous solubility ug/L:
0
Cl
Cl
Cl
Cl
(reference Mckay et al, 1992; Govers and Krop, 1998)
maximum reported solubility 0.711 ug/L
median repo,ted solubility (4 values) 0.087 ug/L
organic carbon partitioning coefficient:
(reference Mckay et al, 1992)
minimum repo1ted value 1,000,000
median reported value (3 values) 2,344,229
1,2.3,4,7,8,9-heptachlorodibenzo-furan ( l ,2,3,4.7,8,9-H7CDF)
Cl3C6H2OC6HCI~
Cl
Cl
Cl
0
aqueous &olubi lity ug/L:
Cl
Cl
(reference Govers and Krop, 1998):
reported solubility 0.258 ug/L
organic carbon partitioning coefficient:
(reference Mckay et al, 1992)
minimum reported value 100,000
Cl
Cl
geometric mean reported value (2 values) 707,946
•
Octachloro-dibenzo-furan (OCDF)
C6Cl40C6Cl4
Cl Cl
Cl Cl
Cl Cl
0
Cl Cl
aqueous solubility ug/L:
(reference Mckay et al, 1992; Govers and Krop, 1998)
maximum repo1ted solubility 0.233 ug/L
median reported solubility (5 values) 0.00154 ug/L
organic carbon partitioning coefficient:
(reference Mckay et al, 1992)
minimum reported value 1,000,000
median reported value (3 values) 5,623,413
•
• •
Part B. Evaluation of the Soil Remedial Goals
for Ground-Water Protection
1. Introduction and Purpose of Part B of the Report
Soil remedial goals for ground-water protection at the Koppers Site were calculated in the 1992
Remedial Investigation Report (Keystone Environmental Resources, Inc., 1992). Appendix J to
that report presents those soil remedial goals. The approach used in that evaluation combined
PCDDs/PCDFs into a composite soil contaminant, with the soil cleanup goal developed based on
the properties of 2,3,7,8-TCDD. The RI Report notes that the more highly chlorinated congeners
are less environmentally mobile than is 2,3,7,8-TCDD. Thus, according to the RI Report, using
2,3,7,8-TCDD as the representative composite PCDD/PCDF contaminant imparts a high degree
of conservatism to the calculated soil remedial goal for ground-water protection. This asse,tion
is partiaJly correct, as there are more highly chlorinated congeners that are apparently both more
and less environmentally mobile than 2,3,7,8-TCDD (reference Figure 14, Part A of this repo11).
Other considerations that went into the RI Report's calculation of soil remedial goals for ground-
water protection were the selection of the fo1mer lagoon area as the location of the potential soil
contaminant source of ground-water contamination, and the No11h Carolina promulgated ground-
water quality standard of 0.00022 ng/L for "dioxin".
The RI Report's calculation of soil remedial goals for ground-water protection considered a
three-zone conceptual contaminant transpo11 model. In this analysis, contamination leached from
the unsaturated zone was considered to be mixed through both the weathered bedrock and the
fractured bedrock and then was potentially found at a hypothetical receptor point in the fractured
bedrock.
The Summers model was used to calculate soil remedial goals for ground-water protection. A
description of this model is found in Summers et al (1980), as well as in U.S. EPA (1989a) and
in U.S. EPA (1996). The Summers model considers one-dimensional advective transport with
equilibrium contaminant sorption in the unsaturated zone. It is inherently conservative, because
it assumes no contaminant degradation, an infinite source term, and no contaminant dispersion in
the unsaturated zone.
The Summers model does consider dilution of contaminated recharge by ground water flowing
beneath the area of unsaturated zone contamination, provided that ground water is
uncontaminated (or less contaminated) and that significant dilution of the contaminated recharge
occurs. This element of the model was considered in the 1992 Remedial Investigation Report.
Subsequent to the Remedial Investigation Report, site remedial action occurred that addressed
• • -2-
the soil contamination in the lagoon area, through excavation and off-Site treatment. A soil
cleanup goal of 7 ug/Kg·"Dioxins/furans" was established in the Record of Decision (U.S. EPA,
1992). Although the Record of Decision is not specific on this point, this 7 ug/Kg
"Dioxins/furans" cleanup goal should be interpreted as meaning 7 ug/Kg expressed as the TEQ
(see Part B, Section 3 for further discussion of the TEQ). Given the relative proportions of the
various PCDD/PCDF congeners in the Site soils (reference Figure 16, Part A), this 7 ug/Kg soil
remedial goal is generally equivalent to the congener-specific remedial goals shown in Table l(as
determined from the average proportions of the seventeen 2,3,7,8 PCDD/PCDF congeners in the
soil samples illustrated on Figure 16, with adjustments made to account for the congener-specific
TEFs ). Note that because the total of all dioxins/furans must exceed 7 ug/Kg as the TEQ in
order to exceed the soil cleanup goal for ground-water protection, the congener-specific soil
cleanup goals in Table 1 do not necessarily represent a concentration of concern with respect to
ground water.
Table 1. Approximate Congener-Specific Soil Cleanup Levels Based on the Record of
Decision Total PCDD/PCDF Soil Remedial Goal for Ground-Water Protection, Adjustment
to Account for the Congener TEFs, and Congener Proportions in Site Soils
Congener
2,3,7,8-TCDD
1,2,3,7,8-PeCDD
1,2,3,4,7,8-H6CDD
1,2,3,6,7,8 H6CDD
1,2,3,7,8,9 H6CDD
1,2,3,4,6,7,8 H7CDD
OCDD
. 2,3,7,8-TCDF
1,2,3,7,8-PeCDF
2,3,4,7 ,8-PeCDF
l ,2,3,4,7,8-H6CDF
1,2,3,6,7,8 H6CDF
1,2,3,7,8,9 H6CDF
2,3,4,6,7,8-H6CDF
1,2,3,4,6,7,8 H7CDF
1,2,3,4,7,8,9 H7CDF
OCDE.
Approximate Soil Remedial Goal, ng/Kg
1
20
250
450
310
118,300
5,312,000
<10
20
<2
50
30
<10
20
8800
1000
285,000
As a part of the remedial action, clean fill replaced the contaminated soils from the lagoon area.
Clean fill was also placed in the Fire Pond and Medlin Pond areas (locations shown on Figure 2
from Part A).
Although the soil contamination that was evaluated in the RI Report and addressed in the Site
remedial action has been effectively remediated, the emplacement of fill materials over
contaminated sediments in the Fire Pond has resulted in the effective burial of earth materials
that may exceed the Record of Decision (ROD)-based soil remedial goals for ground-water
protection. Thus, a concern has been expressed.that the contaminated earth materials in the Fire
Pond area may contribute to further ground-water contamination and may therefore need to be
remediated to eliminate this potential soil to ground-water contaminant transport pathway. Part
B of the Report considers the potential threat to ground-water quality from contaminated earth
• •
-3-
materials in the Fire Pond part of the Koppers Site.
2. Relevant Sampling Data
Sediment samples from the Fire Pond were obtained during the Koppers Site Remedial
Investigation. Sixteen sediment samples were obtained from within the Fire Pond. PCDD/PCDF
analyses were run on samples from two depths obtained at five of those sixteen sample locations.
The sample locations with PCDD/PCDF analyses from the RI are shown on Figure 1. Congener
concentration data are presented in Appendix 1 to Part B. With two exceptions out of 76 paired
analyses, shallow sediment sample congener concentrations exceeded deeper congener
concentrations (this statistic excludes nine non detect shallower and deeper sediment congener
concentration pairs). This situation is consistent with the concept that vertical movement of
PCDDs/PCDFs is limited by the adsorption of the PCDDs/PCDFs to organic carbon in the
sediment or soil. Organic carbon concentrations in the pond sediment are likely to have been
much higher than the organic carbon concentrations measured in the surrounding soils (soil
organic carbon in the range of approximately 2400 mg/Kg to 3558 mg/Kg; reference Keystone
Environmental Resources, Inc., 1992, RI Report,. Appendix J, Section 3.3.3). Sediments in
surface water bodies such as the Fire Pond are typically organic carbon sinks. In such water
bodies, accumulating sediments typically contain more organic carbon than the surrounding
upland soils (U.S. EPA, 2000a). A default organic carbon content for bottom sediments of 0.03
(30,000 mg/Kg) has been cited as a value to be considered in estimating impacts of dioxin-like
compounds to bottom sediments (U.S. EPA, 2000a). Table 4-42a in the RI Report presents data
for five sediment samples obtained from the Fire Pond. For these samples, the mean organic
carbon content was 39,480 mg/Kg and the median organic carbon content was 23,900 mg/Kg.
Figure 2 shows the relative RI sediment sample concentrations of the seventeen PCDD/PCDF
congeners with chlorines in the 2,3,7,8 positions. Consistent with data from RI surface soil
sampling, the sediment samples show that the PCDD/PCDF distribution in Fire Pond sediments
is primarily OCDD (86.4% to 99.8% of the total congener mass in these samples), with
subordinate amounts of 1,2,3,4,6,7,8-H7CDD and OCDF (respectively, from 0.2% to 9.4% and
from <0.1 % to 3.7% of the total congener mass). l,2,3,4,6,7,8-H7CDF was present in
concentrations of up to approximately 0.7% of the total congener mass. Other PCDDs/PCDFs
were present in minor concentrations relative to these four congeners.
OCDD is somewhat enriched in the sediment samples, compared to the OCDD fraction of the
total congener mass in the soil samples obtained during the RI. The mean OCDD proportion of
the total congener mass was 92.7% for the ten sediment samples while the OCDD proportion of
the total congener mass was 80.7% for the twenty soil samples. This difference may be related ,
to the higher molecular weight and higher affinity of OCDD for organic carbon compared to the
other congeners that were also present in significant concentrations in the contaminant source
material. Thus, of the PCDDs/PCDFs that entered the pond environment, relatively more OCDD
• -4-•
Figure 1. Locations of RI Fire Pond Sediment Samples
dala source: Figure 3-11 RI Reporl
Key to figure + RI sediment sample with
S1 PCDO/PCOF analysis
0 SB *RI sediment sample without
PCDO/PCDF analysis
+ S13A
* S5
Scale, ft
0 30 60
• •
Figure 2. Remedial Investigation Sediment Samples with Dioxin and Furan Analyses-
Relative Concentrations of Dioxin and Furan Congeners with TEFs
S13A
•su
*S14
+ S15
• 811
Scale, ft
*s12
Key to Figure
♦ RI Ndtment •ample with S1 PCODIPCOF analysis
•RI Ndlment sample without
S8 PCDDIPCDF analysis
2-4 Sample depth, feet
Key to Pie Charts
□2378TCDD D 12378PeCDD
□ 123478H6CDD □ 123678H6CDO
■ 123789H6CDD ■ 1234678H7CDD
aocoo □2378TC0F
■ 12378PeCDF ■23478PeCDF
□ 123478H6C0F ■ 123678H6CDF
■234678H6CDF ■ 123789H6CDF
■ 1234678H7CDF ■ 1234789H7CDF
Magnitude cf Total Dioxinlfuran Contamination, nglKg ■OCDF
• •
-6-
bound to organic matter and settled to the pond bottom. Another possibility is that OCDD is less
mobile than most of the other congeners, and thus the apparent enrichment of OCDD in the pond
sediment relative to the Site soil is in fact a depletion of the more mobile congeners due to
leaching of those congeners from the sediment. Notably, the amount of OCDD as a percentage
of the total congener mass was generally higher in deeper sediment samples compared to the
paired shallower sediment samples. This observation suggests that the greater mobility of the
other congeners results in their continued movement through the soil/sediment column, whjle
OCDD is not so readily transported through the subsurface.
Following the soil remedial action at Koppers, four samples were obtained by an EPA contractor
from the former Fire Pond area, as noted in Part A, Section 3. The specific locations and depths
of these samples are unavailable and therefore these data are not considered in this evaluation of
the soil remedial goals for ground-water protection.
A follow up sampling was conducted in December 2000 to obtain additional samples from the
now-covered sediments in the Fire Pond area for PCDD/PCDF analyses. Ten additional sample
locations were selected and samples were obtained from below the clean fill material
(identifiable through visual inspection; reference EPA 2001). The upper six inches of buried
sediment was sampled at all the locations. At six locations the sediment from a depth of 24 to 30
inches below the fill-sediment interface was collected. This deeper sampling allows for an
assessment of the environmental mobility of the PCDDs/PCDFs in the buried sediment samples.
Figure 3 shows the locations of the samples obtained by EPA in 2000. The sample locations
were selected to cover areas of the Fire Pond where RI sediment sampling did not include
PCDD/PCDF analyses. Primarily, locations within the former Fire Pond that were relatively
close to former plant production and high soil contamination areas noted during the RI were
sampled by EPA, along with areas in the central to south central parts of the former Fire Pond
where there had been no PCDD/PCDF analyses conducted on RI sediment samples.
Appendix 2 to Part B presents the congener concentration data for the EPA 2000 samples. At the
six sample locations with paired shallow and deeper sediment samples, 74 of 79 paired analyses
had shallow sediment sample congener concentrations that exceeded deeper congener
concentrations. This statistic excludes 23 non detect analyses from both shallower and deeper
sediment sample congener pairs. The higher shallower sediment concentrations compared to
paired deeper sample concentrations are consistent with the conditions observed during the RI.
Figure 4 shows the relative amounts of PCDDs/PCDFs in the 16 samples obtained by EPA.
Consistent with the sediment sample observations during the RI, OCDD comprises most of the
PCDD/PCDF mass in the EPA 2000 sediment samples. For the 16 sediment samples collected
by EPA, OCDD constituted 51.5% to 99 .2% of the total congener mass (the mean value was 85%
of the total congener mass). 1,2,3,4,6,7,8-H7CDD was typically between 2% and 7% of the total
PCDD/PCDF mass but was mtire than 109°0 of the total congener mass in 3 samples. OCDF was
present at up to 11.1 % of the total congener mass but was not detected in 6 of the 16 sediment
• -7 -•
Figure 3. Approximate Locations of EPA December 2000 Investigation
Fire Pond Sediment Samples
dela source: Figure 3-1 1 RI Report
Key to Figure Scare, ft
0 30 60
I
• • -8-
Figure 4. EPA December 2000 Fire Pond Soil Samples with Dioxin and Furan Analyses-
Relative Concentrations of Dioxin and Furan Congeners with TEFs
Z--SLA . ~ \
Fire 3-Sl.
Pond
*
KeytoRquc,
Sedllaafrt ..... + Jocdorl wldl depth 1-SLA belowta••tt-t
,/ lntarface(JNt)
./ 2-2.5 Sample dj th, ft
Scale, n l ..... i
Key to Pit Charts
□2378TCDD Bl 12378PeCDO
□ 123478H6C0D □ 123678H6CDD
■ 123789H6CDD ■ 1234678H7CDD
□OCDD O2378TCDF
■ 12378PeCDF ■23478PeCDF
□123478H6C0F ■ 123678H6C0F
■234678H6CDF ■ 123789H6CDF
■ 1234678H7CDF ■ 1234789H7C0F
■OCDF
• •
-9-
samples, and was generally less than 3% of the total congener mass. 1,2,3,4,6,7,8-H7CDF was
generally less than 2% of the total congener mass. The remaining congeners were detected in
relatively minor concentrations in most of the remaining sediment samples, although four .
shallower sediment samples contained relatively high concentrations of several other congeners,
principally H6CDDs, and high relative high ratios of 1,2,3,4,6,7 ,8~H7CDD, OCDF and
1,2,3,4,6,7,8-H7CDF to OCDD. These sediment samples may represent PCDD/PCDF
contamination derived from two principal sources, such as penlachlorophenol contamination and
burned material.
Of some interest is the sediment contamination compared to the approximate congener-specific
soil cleanup levels presented in Table 1 of Part B. The approximate congener-specific ROD
cleanup goal shown in Table I was exceeded for at least one congener in half of the sediment
samples obtained during the RI and in EPA's follow-up sampling (note that this condition does
. not necessarily mean that the sum of the PCDD/PCDF contamination for a sample would exceed
the ROD-specified soil remedial goal of 7 ug/Kg for all of the PCDDs/PCDFs). However,
-
Table 2. Comparison of Sediment Sample Concentrations to the Approximate
Congener-Specific ROD Soil Cleanup Goal for Ground-Water Protection
Conoener
2,3,7.8,-TCDD
1.2,3,7,8,-PcCDD
1.2,3,4, 7 ,8-H6CDD
l,2,3,6,7,8-H6CDD
l ,2,3.7,8,9-H6CDD
1,2,3,4,6,7,8-H7CDD
OCDD
2,3,7,8-TCDF
1.2,3,7,8-PeCDF
2,3,4,7,8-PeCDF
l ,2,3,4,7,8-H6CDF
1,2,3,6,7,8-H6CDF
l,2,3,7,8,9-H6CDF
2,3,4,6, 7 ,8-H6CDF
1,2,3,4,6,7,8-H7CDF
1,2,3,4,7,8,9-H7CDF
OCDF
----
Estimated ROD
cleanup goal
(based on the
Table l analysis)
I
20
250
450
310
118.300
5,312,000
<10
20
<2
so
30
<10
20
8,800
1,000
285,000
..
Mean ng/Kg
congener
concentration of all
sediment samples•
36.17
111.12
251.68
479.65
574.84
15,883.45
63,467.31
5.09
8.86
8
97.25
131.74
65.73
14.88
2,333.4
..
323.3 I
5,110.63
Median ng/Kg Congener-Specific
congener Snmples Above
concentration of all ROD Cleanup Goal/
sediment samnlcs"' Total Samoles
I.I I 13/26
1.65 5/26
6.05 4/26
18.6 4/26
14.5 4/26
793 1/26
21,090 0/26
0.515 3/26
0 3126
,
0 6126
1.35 4/26
3.5 4/26
2,45 5126
0 4/26
147 1/26
3.85 2/26
375 on6 . . .. ., -· .... -. ---·-*assumes that non-detect nlucs are at a concentration of 0.
• •
-10-
when considered in aggregate, the concentrations of all seventeen toxic congeners do not add to
more than 7 ug/Kg TEQ in any of the twenty-six sediment samples.
A statistical summary of the cleanup goal exceedances is presented in Table 2. Included in the
table are the mean (arithmetic average concentration of each congener) and median
concentration, from all of the RI and EPA 2000 sediment samples. These measures of the
average concentration would be an overall indication, of the threat of sediment samples to the
ground-water quality, if the ROD-specified performance standard for ground-water protection is
used as the measure of threat to ground-water quality for the covered PCDD/PCDF contaminated
sediments in the Fire Pond area.
The Table 2 analysis suggests that Fire Pond sediment samples do not represent a significant
threat to ground-water quality. This possibility is explored in Section 4 of Part B of this report.
3. Procedure for Determining Soil Remedial Goals for Ground-Water Protection
Use of the Summers model to develop soil remedial goals for ground-water protection is retained
in this report. The decision to retain this model for the analysis is based upon the fact that of the
models available, Summers is the most conservative models (or is as conservative as any model
available) and was previously used to derive a soil remedial goal for ground-water protection at
the Koppers Site. The Summers model includes the following conservative assumptions:
a. The Summers model assumes no contaminant degradation.
b. The Summers model assumes an infinite source term.
The Summers model requires that the following four conditions be met (U.S. EPA 1996):
1. There is no contaminant loss due to volatilization or degradation. This requirement is
I . o met, because for 2,3,7,8-TCDD, the median reported vapor pressure (Pa at 25 C) of 2.02
E-7 (15 measurements) and median Henry's Law Constant (Pa m3/mol) of 1.63 (l.66E-05
atm-m3/mol) (9 measurements) (Mackay et al, 1992) demonstrate the low volatility of
this compound. Comparable values for these properties are reported for several other
PCDDs or P,CDFs (Mackay et al, 1992).
2. Adsorption is linear with concentration. Where this assumption typically is not met is in
cases where there are either very high concentrations of a contaminant in the soil, or for
contaminants that very weakly adsorb to a soil. •For low mobility PCDDs/PCDFs at low
concentrations (part per billion magnitude), this assumption is reasonable.
3. The system is at equilibrium with respect to adsorption. As noted in U.S. EPA (1996),
this assumption is conservative, and if it is not met, concentration in recharge will be less
than that predicted by the model.
• •
-1 1-
4. Adsorption is reversible. As noted in U.S. EPA (1996), this assumption is also
conservativb, and if it. is. not met, the concentration in recharge will be less than that
predicted by the model.
The Summers model assumes that soil contaminated by organic compounds or other non-
background contgminants must be uniformly cleaned up to a value predicted by the equation:
SCL = (Q, + Q,,JIQ, • K d • gwps
where
SCL = soi I cleanup level
Q, = ground-water recharge volume
Qgw = ground-water flow volume
Kd = soil-water partitioning coefficient
gwps = the ground-water protection standard
(I)
The volumetric terms in equation (I) represent the dilution factor.
The vertical mixing zone in the aquifer, cl, (the aquifer thickness where ground water will dilute
/.Y·., the contaminated ground-water recharge) is calculated using equation (38) from U.S. EPA ,,,;,,> ( 1996):
cl= (2 avL)05 + da{ 1-exp[(-Ll)/(V, n,da)l}
where cl = mixing zone thjckness
I= recharge rate (meters/year)
av= vertical dispersivity (meters)
(2)
L = length of waste disposal area parallel to the ground-water flow direction (meters)
da = the aquifer thickness (meters)
V, = the ground~water seepage velocity (meters/year)
ne = the effective porosity
The aquifer vetiical dispersivity is a function of the ground-water flow path length (length of
waste disposal area parallel to the ground-water flow direction. Generally, the longitudinal
dispcrsivity is approximately 10% of the flow path length; the vertical dispersivity is
conservatively assumed to be I% of the longitudinal dispersivity.
The ground water flowing beneath the site from upgradicnt is calculated by Darcy's law, Q =
KIA, where:
Q = ground water flow
K = the aquifer hydraulic conductivity
I= the hydraulic gradient,
and A = the cross sectional area of ground water flow.
• •
-12-
The cross sectional area of ground-water flow is the dimension of the area of soil contamination
aligned at a right angle to the prevailing ground-water flow direction, multiplied by the vertical
mixing zone in the aquifer.
In the soil cleanup level analysis presented in the 1992 RI Repott, there was a determination of
the dilution of contaminated recharge by ground water flowing beneath the presumed area of soil
contamination. This procedure is not followed in this analysis. This decision was made because
there is likely to be significantly contaminated ground water entering the area of Fire Pond
contamination from locations upgradient of the Fire Pond, based on the both the RI and the
recent ground-water quality data (reference Figure 17, Figure 18, and Patt A, Section 6.1.2.) and
the analysis of PCDD/PCDF fate and transpott in the subsurface. Thus, equation (I) above is
reduced to the following equation:
SCL= (Q,+ Q,w)/Q,• K, • gwps
Qgw = 0
SCL = (Q, + 0)/Q, • K, • gwps
SCL = (Q, )/Q, • K" • gwps
SCL = I, K, • gwps
t/ where
K" = soil-water partitioning coefficient
gwps = the ground-water protection standard
(I)
The Kd is a congener-specific measure of the propensity of contaminants to partition to soils (or
aquifer materials) versus ground water (dissolved-phase contamination). Table 3 in Part A of
this report presents estimated Keis for the seventeen toxic PCDD/PCDF congeners, considering
the aquifer materials at the Site. The Keis for contaminants in the buried sediment samples in the
former Fire Pond area should be _higher than the Keis _listed in Table 3, because the sediment
samples are considered to have a higher organic carbon content than the aquifer materials (sec
paragraph I, Section 2 of Patt B for a discussion of organic carbon in sediment).
The Site-specific sediment organic carbon content has been measured at the Koppers Site. RI
Report Table 4-42a presents analyses of five sediment samples from the Fire Pond, where the
organic carbon ranged from 13,000 mg/Kg (0.013) to 120,000 (0.12), with a mean concentration
of approximately 39,480 mg/Kg and a median concentration of 23,900 mg/Kg. The median
organic carbon content of 23,900 mg/Kg was selected for this analysis of soil remedial goals for
ground-water protection, considering the former Fire Pond contaminated sediment samples. The
lowest measured organic carbon in the Fire Pond sediments (13,000 mg/Kg) was also evaluated,
for a comparative, more conservative analysis of the soil remedial goals for ground-water
protection. Table 3 presents the sediment Kds for the seventeen PCDDs/PCDFs, using the
contaminant-specific Koc values presented in Part A, Table 3.
• •
-13-
Table 3. Kd Estimates for the Buried Fire Pond Sediments at the Kou1iers 1tc
Organic Carbon Estimated Sediment Kd Estimated Sediment Kd
Partitioning (sediment organic carbon (sediment organic carbon
Congener Coefficient (Koc) 23,900 mg/Kg) 13,000 mg/Kg)
2,3,7,8-TCDD 1,659,587 39,664 21,575
1,2,3,7,8-PeCDD 367,865 8,792 4,782
l,2,3,4,7,8-H6CDD 1,047,129 25,026 13,613
l ,2,3,6,7,8-H6CDD no data 25,026" 13,613'
1,2,3,7,8,9-H6CDD no data 25,026' 13.613'
1,2,3,4,6,7 ,8-H7CDD 4,897.778 117,05,7 63,671
OCDD 12.022.644 287 .34'1 156.294
2,3,7 ,8-TCDF .2.238.721 53.505 29.103
1.2,3,7 .8-PeCDF 3,126.079 74,713 40,639
2.3,4,7,8-PeCDF 3. I 2(,.079 74.713 40.639
1,2,3,4,7,8-H6CDF 25,118,864. 74,713 40,639
1,2,3.6,7,8-H6CDF no data 74.713. 40,639.
l,2,3,7,8,9-H6CDF no data 74,713. 40,639.
2.3 .4,6,7 .8-1·16CDF no data 74.7 I 3• 40,639.
1.2,3,4,6,7 ,8-l-17CDF 2.344,229 56,027 30,475
l .2,3,4,7,8,9-I-17CDF 1 707.946 16,920 9,203
OCDF 5,623.413 134,400 73.104
• The Kd for this congener is assumed to be equal to the estimated Kd for the 1,2,3,4,7 ,8-H6CDD congener.
• This value is considered an outlier or unreliable value (sec.Part A, Section 5.3.2) and a low estimate for the Kd,
equivalent to that estimated fur the PeCDF congeners, is assigned to the H6CDF congeners. on the basis of the Kd
value generally incrcising with increasing chlorination.
• The Koc for this· congener is assumed to be equal to the Koc for the PcCDr congeners
The ground-water protection standard for the PCDDs/PCDFs congeners is either based on the
state of North C_arolina ground-water protection standard for "dioxin" or is based on the practical
quantitation limit (PQL), if the PQL exceeds the promulgated dioxin standard.
The state of North Carolina is proposing a modification to their promulgated ground-water
quality standard for the 2,3,7,8-TCDD TEQ (North Carolina Depat1ment of Environment and
Natural Resources, 2001). In the proposed new standard, the standard for "dioxin" will be
replaced with text indicating the rule applies to the total of the TEFs for the seventeen
•• •
-14-
PCDDs/PCDFs with chlorines in the 2,3,7,8 positi?ns, and the standard will become 0.00023
ng/L, replacing the old standard of 0.00022 ng/L (as the TEQ). This value is virtually identical to
the old standard anl:I the principal intent of the rule change is most likely to clarify that the
ground-water standard represents the 2,3,7,8-TCDD TEQ.
In order to determine the appropriate contaminant-specific ground-water protection standard for
the PCDDs/PCDJrs, the congener-specific TEFs at the practical quantitation limit must be
calculated, summ1cd, and compared to the promulgated ground-water protection standard for
"dioxin."
The TEFs of the seventeen toxic PCDDs and PCDFs arc presented in Table 4.
Table 4. Toxicity Equivalcncy Factors (TEFs) for l'CDDs and PCDFs
TEF TEF (World Health Organization:
con°cncr '·"' (US EPA 1989b) ' ., , Van den Bcr0 cl al 1998)
2.3.7,8,-TCDD I I
1.2,3,7.8,-PeCDD 0.5 I
' I .2,3.4.7.8-H6CDD 0.1 0.1
1,2,3.6,7,8-H6CDD 0.1 0.1
1.2,3.7,8,9-HGCDD 0.1 0.1
1,2.3.4.6,7 ,8-H7CDD 0.01 0.01
OCDD 0.001 0.0001
2,3.7,8-TCDF 0.1 0.1
1,2,3,7,8-PcCDF 0.05 0.05
2,3.4,7,8-PcCDF 0.5 q5
1,2,3,4.7,8-HGCDF 0.1 ' 0.1
1,2,3,6,7,8-HGCDF 0.1 0.1
1,2,3,7,8,9-H6CDF 0.1 0.1
2,3,4,6,7,8-HGCDF 0.1 0.1
1,2,3,4,6,7,8-H?CDF 0.01 0.01
1,2,3,4,7,8,9-H?CDF 0.01 0.01
OCDF 0.001 0.000 l
• •
-15-
Table 4 shows that for three congeners, the World Health Organization (WHO) TEFs differ from
the EPA TEFs. Currently, EPA uses the U.S. EPA (1989b) TEFs in human health risk
assessments; however, it is anticipated that in the near future, the WHO TEFs will be substituted
in EPA human health risk assessment (Akin, 2001). Thus, in order to provide a comprehensive
analysis of the ground-water protection standard for the calculation of soil remedial goals for
ground-water protection, both sets ofTEFs are used in the calculation of congener-specific TEFs
associated with individual congener PQLs.
Table 5 presents the TEQs for each individual PCDD/PCDF in ground water, using the TEFs
from Table 4 multiplied by the practical quantitation limit (PQL) for each congener as repo11ed in
U.S. EPA (2000b).
Congener
2.3.7.8.-TCDD
1.2.3.7.8.-PcCDD
l,2,3.4,7,8-H6CDD
1,2,3,6,7 ,8-H6CDD
\,2J,7,8,9-H6CDD
1,2,3,4.6,7 ,8-H7CDD
OCDD
2.3,7.8-TCDF
1,2.3.7 ,8-PcCDF
2,3,4,7,8-PcCDF
I ,2,3,4,7,8-H6CDF
I ,2.3,6.7,8-H6CDF
1,2,3,7 ,8,9-HGCDF
2,3,4,6,7,8-H6CDF
l ,2,3,4,6,7,8-117CDF
1,2,3,4,7 ,8,9.J-17CDF
OCDF
Table 5. TEFs for PCDDs/PCDFs in Ground Water at the
Practical Quantitation Limit Concentration
PQL from TEQ al the PQL
TEF-TEF U.S. EPA, Concentration
(El'A, 1989) (WIIO, 1998) 20001, (ng/L) (EPA, 1989)
I I 0.01 0.01
0.5 I 0.01 0.005
0.1 0.1 0.025 0.0025 ·
0.1 0.1 0.025 0.0025
0.1 0 I 0.025 0.0025
0.01 0.01 0.()25 0.00025
0.001 0.0001 0.05 ' 0.00005
0.1 0.1 0.01 0.001
0.05 0.05 0.01 0.0005
0.5 0.5 0.01 0.005
0.1 0.1 0.025 0.0025
0.1 0.1 0.025 0.0025
0.1 0.1 0.025 0.0025
0.1 0.1 0.025 0.0025
0.01 0.01 0.025 · 0.00025
0.01 0.01 0.025 0.00025
0.001 0.0001 0.05 0.00005
I
TEQ (sum of all
congcncr-spccilic
TEQs) 0.02985 nolL
TEQ al the l'QL
Concentration
(\\'IIO, 1998)
0.01
0.01
0.0025
0.0025
(l.()025
{).{)0025
0.000005
0.001
0.0005
0.005
0.0025
0.0025
0.0025
0.0025
0.00025
0.00025
0.000005
0.0]476 m!IL
• •
-16-,
Regardless of whether the TEFs reported in U.S. EPA (1989) or the 1998 WHO TEFs are used,
the sum of the congener-specific TEQs at the PCDD/PCDF PQL concentrations exceeds the
North Carolina promulgated ground-water standard for dioxin (or the pending standard for the
2,3,7.8-TCDD TEQ). Therefore, the PQLs for the PCDDs/PCDFs should be used to calculate
the soil remedial goals for ground-water protection, rather than calculating the soil remedial goals
on the basis of the promulgated state ground-water quality standard for dioxin or the 2,3,7 ,8-
TCDD TEQ.
Three points are made regarding the ground-water protection standard used to calcu'iate soil
remedial goals for ground-water protection:
I.
2.
If some of the congeners of concern were not present in soils and ground water, the
analysis might differ from that presented in this report. As a simplified example, if
OCDD was t~e only congener of concern, the TEF multiplied by the PQL would be well
below the promulgated state standa_rd for dioxin and the state standard would therefore be·
applied.
For a larger number of congeners that would not sum to a TEQ exceeding the state I
promulgated ground-water standard for dioxin or the 2,3,7,8-TCDD TEQ, the comparison
becomes complicated. Because each congener in ground water represents a proportion of
the total TEQ, there are an extremely large number of multiple PCDD/PCDF congener
mixes that would allow for a total TEQ equal to the standard. For example, consider a
case where there are only two congeners to be considered, one with a TEF of 0.01 and the
second with a TEF of 0.001. A ground-water concentration of 0.01 ng/L for the first
congener would equal a 0.0001 ng/L contribution to the TEQ for that congener. The
remaining "allowable concentration" of 0.00013 TEF (state standard 2,3,7,8-TCDD TEQ
of 0.00023 ng/L minus the TEQ contribution from the first congener) would equal a
concentration of 0.13 ng/L for the congener with a TEF of 0.001 (0.01 TEF •0.1 ng/L)+
(0.001 TEF •0.13 ng/L) = 0.00023 ng/L= allowable TEQ. In but one of numerous
alternative scenarios, a ground-water concentration of 0.015 ng/L for the congener with a
TEF of 0.0 I would allow for a 0.08 ng/L concentration of the congener with the TEF of
0.001, using the same type of analysis. As should be apparent from this simplistic
example, as more congeners figure into the calculations, the number of congener
combinations that could be present in ground water and that would equal a TEQ not
exceeding the state ground-water standard would_ become enormous. Thus, calculation of
the soil remedial goals for ground-water protection using a congener-specific approach
would be very complicated. A scheme for assigning a "proportionate representation" for
each congener of interest, probably based on their relative mobility or prevalence of the
congeners in Koppers Site soils, would have to be developed in order to perform such a
calculation.
Another point is that the PQLs for most of the PCDD/PCDF congeners are O.Sx or less
3.
• •
-17-'
the "contract required quantitation limit" forPCDDs/PCDFs used in Superfund Site
investigative work. Thus, by using the PQLs as the ground-water protection standard, the
soil cleanup levels for ground-water protection at the Koppers Site are generally based on
ground-water protection standards that are lower than the ground-water concentrations
that would be reported, with a sufficient degree of certainty, in any analysis of Site
ground water that might be contaminated by PCDDs/PCDFs. This condition adds to the
conservatism in calculation of the soil remedial goals.
Some PCDD/PCDF congeners may be present in concentrations in the sediments that are
well below the soil remedial goals for ground-water protection calculated using the PQL
as the ground-water protection standard. If so, these congeners can be "taken off the
table" in terms of calculating the risk from soil contaminant migration to ground water,
because the modeled leachate concentration of those congeners reaching the ground water
would be immeasurable, as it is below the PQL. In this case, the remaining congeners
for which the calculation of soil remedial goals for ground-water protection was done
using the PQL as the ground-water protection standard may actually have an allowable
concentration higher than that soil remedial goal. Therefore, following the calculation of
soil remedial goals for ground-water protection using the PQL approach, the congeners
for which soil concentrations indicate a concern should be subject to further evaluation
using the "proportionate representation" concept cited in point l above.
4. Calculations of Soil Remedial Goals for Ground-Water Protection and Comparison to
Fire Pond Contaminated Soil PCDD/PCDF Concentrations
The initial soil remedial goals for ground-water protection at the Fire Pond part of the Koppers
Site are calculated as the product of the contaminant soil (sediment)-water partitioning
coefficient and the congener-specific practical quantitaiion limit. These values are presented in
Table 3 and Table 5 respectively. Table 6 presents the soil remedial goals for ground-water
protection.
Table 6. Initial Soil Remedial Goals for Ground-Water Protection
Kd (using Kd (using ground-water Soil remedial
11_1edian organic minimum organic protection goal (median Soil remedial
carbon in Fire carbon in Fire standard organic goal (minimum
congener Pond se!liments) Pond sediments) (PQL, ng/L) carbon) organic carbon)
2,3,7,8,-
TCDD 39,664 21,575 0.01 397 ng/Kg 216ng/Kg
1,2,3,7,8,---.
PeCDD 8,792 4,782 0.01 88 ng/Kg 48 ng/Kg
Table 6 is continued on the following page
• •
-18-
Table 6, continued
Kd (using Kd (using ground-water Soil remedial
median organic minimum organic protection goal (median Soil remedial
carbon in Fire carbon in Fire standard organic goal (minimum
congener Pond sediments) Pond sediments) (PQL, ng/L) carbon) organic carbon)
1,2,3,4,7 ,8-
H6CDD 25,026 13,613 0.025 626 ng/Kg 340 ng/Kg
1,2,3 ,6,7 ,8-
H6CDD 25,026 13,613 0.025 626 ng/Kg 340 ng/Kg
1,2,3,7,8,9-
H6CDD 25,026 13,613 0.025 626 ng/Kg 340 ng/Kg
1,2,3,4,6,7,
8-H7CDD 117,057 63,671 0.025 2926 ng/Kg 1592 ng/Kg
OCDD 287,341 156,294 0.05 · 14,367 ng/Kg 7815 ng/Kg
2,3,7,8-
TCDF 53,505 29,103 0.01 535 ng/Kg 291 ng/Kg
1,2,3,7,8-
PeCDF 74,713 40,639 0.01 747 ng/Kg 406 ng/Kg
2,3,4,7 ,8-
PeCDF 74,713 40,639 0.01 747 ng/Kg 406 ng/Kg
1,2,3,4,7,8-
H6CDF 74,713 40.639 0.025 1868 ng/Kg 1016 ng/Kg
1,2,3,6,7 ,8-
H6CDF 74,713 40,639 0.025 1868 ng/Kg 1016 ng/Kg
-
1,2,3,7,8,9-
H6CDF 74,713 40,639 0.025 1868 ng/Kg 1016 ng/Kg
2,3,4,6,7,8-
H6CDF 74,713 40,639 0.025 1868 ng/Kg 1016 ng/Kg
1,2,3,4,6,7,
8-H7CDF 56,027 30,475 0.025 1401 ng/Kg 762 ng/Kg
1,2,3,4,7,8,
9-H7CDF 16,920 9,203 0.025 423 ng/Kg 230 ng/Kg
OCDF 134,400 73,104 0.05 6720 ng/Kg 1828 ng/Kg
,.s.' ;::,\/':"
• •
-19-
Note that the soil remedial goals for ground-water protection calculated in Table 6 are
· considerably lower than the soil remedial goals for ground-water protection that would apply if
the 7 ug/Kg TEQ concentration is the target soil remedial goal. This discrepancy arises because
for the Fire Pond sediments, no dilution factor is applied to the analysis and also, the protected
ground water is assumed to be the entire saturated zo_ne. Using the median organic carbon
content, Table 6 derives a soil remedial goal of 1.91 ug/Kg as a TEQ.
The congener-specific soil remedial goals for ground-water pro'tection at the Fire Pond are then
compared to the contaminant concentrations detected· in sediment samples collected during the
RI and in the EPA
follow-up investigation. This comparison is presented in Table 7a and Table 7b.
Tables 7a and 7b indicate the concentrations of some congeners in the buried sediment samples
I are a potential concern with respect to the soil remedial goals for ground-water protection. Table
8 summarizes the statistical breakdown on congener concentrations with respect to the soil
remedial goals calculate_d in Table 6.
Table 7a. Comparison of Soil Remedial Goals from Table 6 to the Measured Fire Pond
PCDD/PCDF Sediment Concentrations, RI Data
Sample (with sample depth in feet) ,
-.:oil remedial goa ~oil rcmedtal goa:
rom•ener (median) (minimum) Sl 0-0.5 Sl 2,4 S4 0-2 S4 2-4 S10 0.2 S10 2-4
378TCDD 397 ~K" 216 ~K0 12.9 2.5 0.92 0.5 9.4 0
2378PeCDD 88 no/K11 48 no/Ko 23. l 1.8 0.69 0 8.1 0
23478H6CDD 626 Il" Ko 340 n° Kcr 67.6 7.9 3.5 0 24.4 0
23678H6CDD 626 tic Ko 340 n• Ko 205 20.2 7.6 0.71 80.3 2.4
23789H6CDD 626 n• Ko 340 Ill! Ko 290 17.2 l 1.8 0.93 115 3.4
234678H7CDD 2926 ng/Kg 1592 ng/Kg 16,210 1610 586 150 5020 175
DCDD 14.367 ng/Kg 7815 ng/Kg 215,930 33,730 24,140 81,100 58,020 11,960
378TCDF 535 n• Ko 291 "' Kg 6.2 0.88 0.52 0 5.7 1.4
2378PeCDF 747 ng· Ke 406 mi: Ko 4.8 0 0 0 4.5 1.3
3478PeCDF 141 nc K• 406 nc Ko 2.8 0 0 0 3.8 1.4
23478H6CDF 1868 "' Ko 1016 n, K• 29.4 4.9 1.6 0 20.7 3.2
23678H6CDF 1868 nc Kcr 1016m Ko 23 4 1.5 0 11.8 2.1
34678H6CDF 1868 "' Ko 1016n Ko 17.8 4.3 2.5 0.47 18 1.7
23789H6CDF 1868 n• K• 1016 n K• 0 0 0 0 0 0
234678117CDF 1401 ng/Kg 762 ng/Kg 1300 144 59.9 4 408 16.4
234789H7CDF 423 no/Ko 230 no/Ko 161 13.2 3.7 0 19 0
riCDF 6720 Il"1K" 1828 no-1K0 "' 773 301 27.5 "'"" 61.7
Entries in bold type indicate the soil concentration exceeds the minimum soil remedial goal calculalcd in Table 6. Entries in bold and
italic type indicate the soil concentration exceeds the soil remedial goal calculated using the median sediment organic carbon
concentration.
Table 7 is continued on the following page
• •
-20-
Table 7, continued
·Sample (with sample depth in feet) ,
:>01 ! remedial goa Soll remedial goa
,.;OOQ'.CllCr {median) (minimum) · Sl3A 0-2 SBA 2-4 SIS 0-0.5 SIS 2-4
378TCDD 397 ng/Kg 216 ng/Kg 6.2 3.8 2.5 I
2378PeCDD 88 no/Ko 48 ng/Kg 16.8 8.1 3.4 I
123478H6CDD 626 Ill Kg 340 ng/Kg 55 26.9 12.5 I
23678H6CDD 626 ne Kg 340 ng/Kg 188 89.3 28.3
23789H6CDD 626 n< Ko 340 ng/KP-145 81.9 31.8
234678H7CDD 2926 ng/Kg 1592 ng/Kg 12,800 6790 1820 41.
DCDD 14,367 ng/Kg 7815 ng/Kg 117,670 80,550 22,180 537
378TCDF 535 n< Ko 291 m K• 1.2 1.5 1.1 0.51
2378PeCDF 747 m Kg 406 ni Kg 2.6 0 0
3478PeCDF 747 ni Kg 406 n.11 Kg 1.6 0 0
23478H6CDF 1868 ll!! Kg 1016 n Kg 28.8 13.7 3.5
23678H6CDF 1868 n, Kg 1016 ni;i Kg 16.1 11.3 3.4
34678H6CDF 1868 n< Ko 1016m Ko 12.1 9.9 7
23789H6CDF 1868 n, Kg 1016 11!! Kg 0 0 0
234678II7CDF 1401 ng/Kg 762 ng/Kg 890 574 155 2.
234789H7CDF 423 n.E!/Kg 230 ng/Kg 46.3 48.3 10.2
OCDF 6720 n~Kg 1828 n~Kg 43 " 919 14.1
Table 7b. Comparison of Soil Remedial Goals from Table 6 to the Measured Fire Pond
PCDD/PCDF Sediment Concentrations, EPA 2000 Data
Sample
:>Oil remed1a Soil remedial goa
Congen~r goal (median) (minimum) KPOOISLA KP002SLA KP003SLA KP003SLB ..
378TCDD 397 ng/Kg 216 ne/K, 1.3 1.7 0 0
12378PeCDD 88 ng/Kg 48 ng/Kg 4 3.6 0.42 0
23478H6CDD 626 ne/Kg 340 ng/Kg 8.6 10 1.4 0.76
23678H6CDD 626 n, Ke 340 ne/Ke 29 28 2.7 . 2.6
23789116CDD 626 ng Kg 340 ng/Kg 26 30 4.8 . 2.9
234678H7CDD 2926 ng/Ke 1592 ng/Kg 1300 1100 190 150
OCDD 14,367 ng/Kg 7815 ng/Kg 16,000 14,000 7100 5800
378TCDF 535 ng/Kg 291 n1/Kg 0 0 0 0
2378PeCDF 747 ng/Ke 406 n, Ke 0 0 0 0
3478PeCDF 747 ne/Ke 406 ni Kg 0.51 0.54 0.16 0
23478H6CDF 1868 ng/Kg l016 ng/Kg 0 : 0 0 0
23678H6CDF 1868 ne/K• 1016 n /Ke 6.4 6.7 0.75 0.47
34678H6CDF 1868 ng Kg 1016 n Kg 2.3 2.6 0.38 0
23789H6CDF 1868 n, Ke l016 n Ke 15 0.7 4.8 0
234678H7CDF 1401 ng Kg 762 ng/Kg 180 170 19 17
234789H7CDF 423 ne/Kg 230 ne/Ke 15 18 3.9 0
C-.<"n~ ,, /Ka ,x" '"" "" 40n 7i n .. Entries m bold type md1cate the sod concentration exceeds tbe IIUmmum soil remedial goal calculated m Table 6. Entries m bold and
. italic type indicate the soil concentration exceeds the soil remedial goal calcul_ated using the median sediment organic carbon
concentration.
Table 7b is..conti_~mcd on the following page
• •
-21-
Table 7b, continued
Samole
Soil remeo1a :::;011 remedial goa
!cone:ener oal (median) (minimum) KP004SLA KP004SLB KPOOSSLA KPOOSSLB
378TCDD 397 ne/Kg 216 ng/Kg 69 0 3.5
2378PeCDD 88 ng/Kg 48 ng/Kg 410 1.5 5.6
23478H6CDD 626 ng/Kg 340 ng/Kg 1700 4,2 14 0.71
,23678H6CDD 626 ng/Kg 340 ng/Kg 1900 17 45 0.88
123789H6CDD 626 ng/Kg 340 ng/Kg 4100 11 49 L
234678H7CDD 2926 ng/Kg 1592 ng/Kg 79,000 1000 2500 97
:JCDD 14,367 ng/Kg 7815 ng/Kg 140,000 20 000 'i6,000 12,001
378TCDF 535 ne/Ke 29 I ne/Ke 0 0 0.54 ' 2378PeCDF 747 ne/Kg 406 ne/Kg 0 0 0
3478PeCDF 747 ne/Ke 406 ne/Ke 55 0 0
23478H6CDF 1868 ne/Ke 1016 ne/Kg 600 0 0 '
23678H6CDF 1868 ng/Kg 1016 ng/Kg 690 3,6 5.9 '
34678H6CDF 1868 ne/Ke 1016 ne/Ke 470 l.6 4,1 '
23789H6CDF 1868 ne/Kg IOI 6 ng/Kg 83, 0,58 0.85 '
234678H7CDF 1401 ng/Kg 762 ng/Kg 7200 150 310 '
234789H7CDF 423 ng/Kg 230 ng/Kg 3700 0 0 ' 7CDF 6720 ne/Ke 1828 ne/Ke l'i 330 890 '
Sam 1]e
_ongener :Soil remectta so,1 remeoiat goa KPOOOSLA t<sP006SLB KPUU7SLA
oal (median) (minimum)
378TCDD 397 ne/Ke 216ne/Kg 0,31 0 4•
2378PeCDD 88 ng/Kg 48 ng/Kg I 0 22
23478H6CDD ' 626 ng/Kg 340 ng/Kg 2.7 0,56 57
23678H6CDD 626 ng/Kg 340 ng/Kg 8.3 1.6 1101
123789H6CDD 626 ng/Kg 340 ng/Kg 9 LB 160l
234678H7CDD 2926 nglKg 1592 ng/Kg 520 110 27.00l
uCDD 14,367 ng/Kg 7815 ng/Kg 11,000 5800 66 001
2378TCDF 535 ng/Ke 291 ne/Kg 0 0 15
2378PeCDF I 747 ne/Kg 406 ng/Kg 0.16 0 35
,3478PeCDF 747 ne/Ke 406 ne/Kg 0.19 0 31
23478H6CDF 1868 ne/Ke 1016 ne/Ke I.I 0 411
123678H6CDF 1868 ne/Kg 1016 ng/Kg 1.9 0,33 281
s34678H6CDF 1868 ne/Ke 1016 ne/Kg 0,87 0 151
l23789H6CDF 1868 ng/Kg 1016 ne/Kg 0 0 2
1234678H7CDF 1401 ng/Kg 762 ng/Kg 61 6.6 3401
1234789H7CDF 423 ng/Kg 230 ng/Kg. 2.9 0.85 961
OCDF 6720 ng/Kg 1828 ng/Kg 160 0 7001
Entries in hold type indicate the soil concentration exceeds the minimum soil remedial goal calculated in Table 6. Entries in bold and
italic type indicate the soil concentration exceeds the soil remedial goal calculated using the median sediment organic carbon
concentration. ·
• •
-22-.
Table 8. Statistical Breakdown on Congener Concentrations
with Respect to the Soil Remedial Goals Calculated in Table 6
I
~ongencr er Cent ot Samples Exceedmg er Lent ot Samples Exceedm[
Median Soil Remedial goal Minimum Soil Remedial goal
,378TCDD 0 0
2378PeCDD 9.52 9.52
23478H6CDD 4.76 9.52
23678H6CDD 9.52 9.52
123789H6CDD 9.52 9.52
1234678H7CDD 28.57 38.09
uCDD 61.9 80.95
378TCDF 0 0
2378PeCDF 0 0
3478PeCDF 0 0
23478H6CDF 0 0
23678H6CDF 0 0
34678H6CDF 0 0
23789H6CDF 0 0
234678H7CDF 9.52 19.05
1234789H7CDF 9.52 9.52
LJCDF 9.52 28.57
,;. Table 7 and Table 8 suggest that many of the buried sediment samples may represent a threat to
ground-water quality, given the congener-specific concentrations for ground-water protection.
While Table 7 and Table 8 suggest a concern about the buried Fire Pond sediments, the true
indication of the degree of concern would be the TEQ for each specific sediment sample,
compared to the 1.91 ug/Kg TEQ calculated as the sum of individual congener concentations
protective of ground water. Conside,ing this 1.91 ug/Kg TEQ criterion, for the case where the
median organic carbon concentration is assumed to control contaminant leaching from the buried
sediment samples, none of the RI sediment samples and two of sixteen samples collected by EPA
in 2000 might pose a concern with respect to contaminant leaching to ground water.
If a risk-based soil remedial goal for ground-water protection is considered (i.e. a soil remedial
goal based on the North Carolina 2,3,7,8-TCDD TEQ ground-water standard), the relative
contribution of each congener present in significant concentrations in the buried sediment to the
overall TEQ can be considered. This is done by taking the total allowable TEQ in ground water
(0.00023 ng/L) and assigning a weighting factor to each sediment congener on the basis of the
frequency of detection of that congener at a concentration above the soil remedial goals for
ground-water protection that are listed in Table 6. For the calculations of those soil remedial
goals using the median organic carbon concentration, the weighting approach is done as follows:
1. Define the number of congener analyses that exceed the median soil remedial goal for
ground-water protection (32 analyses, from Table 7a and Tabl_e 7b).
2. Apportion the allowable "TEQ units" based on the number of congener-specific
3.
B.
• -23-•
observations that exceed the median soil remedial goal for ground-water protection:
1,2,3,7,8-PeCDD = 0.0625
1,2,3,4,7,8-H6CDD = 0.03125
l,2,3,6,7,8-H6CDD = 0.0625
1,2,3,7,8,9-H6CDD = 0.0625
l,2,3,4,6,7,8-H7CDD = 0.1875
OCDD'= 0.4063
1,2,3,4,6,7,8-H7CDF = 0.0625
l ,2,3,4,7,8,9-H7CDF = 0.0625
OCDF = 0.0625
Note that in relative terms, the allowable "TEQ units" calculated in step 2 somewhat
con-espond to the average on-Site ground water quality (reference Part A of the report,
Figure 21). That is, the two congeners that are most prevalent in Fire Pond area soils in
concentrations above soil remedial goals for ground-water protection (OCDD and
l,2,3,4,6,7,8-H7CDD) are found in the highest concentrations in on-Site ground water.
· Considering that if these two congeners were '.100% of the congener mass of concern in
the Fire Pond area soils, the allowable "TEQ units" would be divided between these two
congeners, the fact that they are most of the congener contamination of significance
present in the soils means that they are proportionately weighted to account for their
overall prevalence in the environment. Note also that this weighting procedure generally
allows a relatively higher soil remedial goal for the less environmentally mobile and less
toxic congeners such as OCDD and a relatively lower soil remedial goal for the more
toxic and more environmentally mobile congeners.
Calculate the congener-specific contribution to the TEQ in terms of the unadjusted
ground-water concentration (herein termed the "TEQ factor"). For OCDD, the allowable
contribution to the TEQ is 0.4063 (from step 2 above). Thus, the "TEQ factor" for this
congener is calculated as the total allowable TEQ (0.00023 ng/L) • 0.4063 = 0.000093449
ng/L. Considering that the TEF for OCDD is 0.001 (using the U.S. EPA 1989 value), the
ground-water protection standard for OCDD, weighted for the TEF of that congener, is
0.00009344 ng/L-;. 0.001 = 0.09345 ng/L. The calculations for all nine congeners are as
follows:
Calculations Using the U.S. EPA 1989 TEF
"TEO factor"
1,2,3,7,8-PeCDD 0.000014375
l ,2,3,4,7,8-H6CDD 0.000007394
l,2,3,6,7,8-H6CDD 0.000014375
1,2,3,7,8,9-H6CDD 0.000014375
1,2,3,4,6,7,8-H7CDD 0.000043125
OCDD 0.000093449
1,2,3,4,6,7,8-H7CDF 0.000014375
Weighted Ground-Water
Protection Standard
0.00002875 ng/L
0.00007394 ng/L
0.00014375 ng/L
G.000i4375 ng/L
0.0043125 ng/L
0.09345 ng/L
0.0014375 ng/L
____ __,,,2.3_,4.LS.JhH7CDE 0 000014375
TEF
0.5
0.1
0.1
0.1
0.01
0.001
0.01
00]
0.001
Q 0014375 uga
OCDF 0.000014375 0.014375 ng/L
---··--· .. ---
'
• •
-24-
C. Calculations Using the WHO TEP Values
"TEO factor"
1,2,3,7,8-PeCDD 0.000014375
1,2,3,4,7,8-H6CDD 0.000007394
1,2,3,6,7,8-H6CDD 0.000014375
' l,2,3,7,8,9-H6CDD 0.000014375
l,2,3,4,6,7,8-H7CDD 0.000043125
OCDD 0.000093449
l,2,3,4,6,7,8-H7CDP 0.000014375
l,2,3,4,7,8,9-H7CDP 0.000014375
OCDP 0.000014375
Weighted Ground-Water
TEP Protection Standard
0.000014375 ng/L
0.1 0.00007394 ng/L
0.1 0.00014375 ng/L
0.1 0.00014375 ng/L
0.01 0.0043125 ng/L
0.0001 0.9345 ng/L
0.01 0.0014375 ng/L
0.01 0.0014375 ng/L
0.0001 0.14375 ng/L
The weighted ground-water protection standard values calculated in step 3 above are then
multiplied by the congener-specific sediment-water partitioning coefficients (median values)
listed in Table 3 to derive the modified soil remedial goals for ground-water protection if the
North Carolina ground-water 2,3,7,8-TCDD TEQ is used as the basis for the calculation of the
soil remedial goals. Table 9 presents the results of this analysis.
Table 9. Soil Remedial Goals for Ground-Water Protection with the North Carolina
Ground-Water 2,3,7,8-TCDD TEQ as the Basis for the Remedial Goals
-. . _____ ,,,_,_ ···-· . ··--· .. -... ·--·-···-····· -· Soil Rem-fdiii! Gcilll for SOi1-RCniecliiii'GO.i1 fof -·---.... ---
Ground-Water Ground-Water .Est.ima.t..e _, Eroteclion..(K.., -Erotection-(K•1 ....Mean-ng.i-K-g--J\lledian-nglK0 Sediment Kd multiplied by the multiplied by the congener congener
(sedirhent weighted ground-water weighted ground-water concentration concentration of
Congener organic carbon protection standard, protection st::mdard, of all sediment all sediment
23.900 mg/Kg) calculated using the calculated using the samples samples .
EPA 1989 TEF) •WHOTEF)
1.2.3.7 .8-PeCDD 8,792 0.253 ng/Kg 0.1264 ng/Kg 111.12 1.65
1.2.3.4.7.8-H6CDD 25.b26 1.85 ng/Kg 1.85 ng/Kg 251.68 6.05
1.2.J.6.7.8-H6CDD 25,026 3.6 ng/Kg 3.6 ng/Kg 479.65 18.6
1.2.3.7.8.9-H6CDD 25.026 3.6 ng/Kg 3.6 ng/Kg 574.84 14.5
1.2.3.4.6.7.8-117.057 505 ng/Kg 505 ng/Kg 15.883.45 793
H7CDD
OCDD 287.341 26.852 ng/Kg 268.520 ng/Kg 63.467.31 21.G90
1.2.3.4.6.7.8-56.027 80.5 ng/Kg
H7CDF
80.5 ng/Kg 2333.4 147
1.2.3.4.7.8.9-16.920 24.3 ng/Kg
H7CDF
24.3 ng/Kg 323.31 3.85
--OCDF 134 400 1932 ne/K0 19.320 n£/K'1'. 5110.63 375
assumes non-detect samples have a concentration of 0.
-·• --·-. ---.
• •
-25-
In this analysis, the TEQ soil concentration using the EPA 1989 TEFs would be only about 0.04
ug/Kg. This modeling approach is considered un_realistic, because the analysis predicts that for
soils contaminated at the congener concentrations shown in Table 9, the only measurable
resultant congener 1concentrations in the ground water would be OCDD and OCDF (if the latter is
considered using the WHO TEF). The analysis predicts that the remaining congeners would be
present at sub-PQL concentrations.
Relative to the calculated soil remedial goals for ground-water protection that are presented in
Table 6, the soil remedial goals for ground-water protection that are presented in Table 9 show
lower congener-specific soil remedial goals would apply to all of the congeners with the
exception of OCDD and OCDF (if the latter is considered using the WHO TEF). Table 6 is
considered the appropriate presentation of soil remedial goals for ground-water protection that
are applicable to the buried Fire Pond sediments.
When the Table 6 soil remedial goals are summed to a TEQ and that TEQ is compared to
sediment concentration data, the analysis indicates there is a small amount of the buried sediment
in the former Fire Pond area that may represent a continuing impact to ground-water quality.
Sample KP004SLA (EPA 2000 sample) has a TEQ of 2.31 ug/Kg, which is slightly higher than
the 1.91 ug/Kg TEQ considered protective of ground water for the buried Fire Pond area
,,.·:,, sediments. EPA sample KP009SLA has a calculated TEQ of 6.92 ug/Kg, which is considerably
'"0 --' higher than the 1.91 ug/Kg TEQ concentration considered protective of ground water. However,
the deeper sediment samples collected at both the KP004 and KP009 locations have considerably
·-·-·· -·······10werTEQconcentrationnhan· their shallow sample c·olmtetparts:--At the.KP004Tocaticiii~the· ····
deeper KP004SLB sample has a TEQ concentration of only 0.04 ug!Kcr, while at the KP009
location, KP009SLB has a TEQ concentration of less than 0.02 ug/Kg. These much lower TEQ
concentrations for the deeper samples probably indicate that the high concentrations observed in
the shallower samples are a result of the shallower buried sediments containing a higher than
average amount of organic carbon, to which the PCDDs and PCDFs are muc_h more strongly
bound than is indicated by the Table 6 analysis. Considering this interpretation, there is probably
not a concern about leaching of PCDDs/PCDFs from the buried Fire Pond sediments.
5. Summary and Conclusions
Soil remedial goals for ground-water protection at the Koppers Site were calculated in the 1992
Remedial Investigation Report. The approach used in that evaluation combined PCDDs/PCDFs
into a composite soil contaminant, with the soil cleanup goal developed based on the properties
of 2,3,7,8-TCDD. The RI Report's calculation of soil remedial goals for ground-water protection
considered the former (now remediated) lagoon area as the location of the potential soil
contaminant source of ground-water contamination. The North Carolina promulgated ground-
water quality standard of 0.00022 ng/L for "dioxin,, was set as the ground-water target
concentration. The Summers model was used to calculate the soil remedial goals. This model is
inherently conservative, because it assumes no contaminant degradation, an infinite source term,
· ··-and no contaminant dispersion in the unsaturated zone.
• •
-26-
As a part of the Site remedial action, clean fill replaced the contaminated soils from the lagoon
area. Clean fill was also placed in the Fire Pond area. The emplacement of clean fill in the Fire
Pond area has buried earth materials that may exceed soil remedial goals for ground-water
protection. Part B of the Repo11 considers the potential threat to ground-water quality from these
contaminated earth materials in the Fire Pond part of the Koppers Site.
In the Pat1 B analysis, to evaluate threat to ground-water quality from the buried sediment
contamination in the former Fire Pond area, the Summers model was retained. No dilution of
soil water contamination by uncontaminated ground-water was considered. Congener-specific
soil remedial goals for ground-water protection were determined, as well as a revised TEQ
applicable to the Fire Pond buried sediments.
Fire Pond area bu1ied sediment sample data from the RI and data from sampling by EPA in 2000
were evaluated. The environmental mobility of the PCDDs/PCDFs in the buried sediment was
consid.ered by calculating soil-water pa11itioning coefficients using organic carbon data
applicable to the buried sediments. The organic carbon content of the bu1ied sediments is
generally considerably higher than that of soils at the Koppers Site, resulting in lower
environmental mobility of the PCDDs and PCDFs in the bu1ied sediments compared to the
Koppers Site soils .
..._.,,;, Consistent with RI surface soil data, most of the PCDD/PCDF buried sediment contamination is
by OCDD, with subordinate contamination by 1,2,3,4,6,7,8-H7CDD and OCDF. Most of the
-··---------·bufiect·sedimentsamples"lrnct congener i!istfibutions·consistenCwith apeiifiichforopnenor ------· · ··--· --· ·
contaminant source, although four sediment samples appeared to be somewhat representative of
other sources of contamination, possibly indicating a mixture of both pentachlorophenol-derived
PCDD/PCDF contamination and burned material.
Estimated congener-specific soil cleanup values were initially calculated based on the Record of
Decision total dioxin soil remedial goal of 7 ug/Kg (for "Dioxins/furans") arid the proportion of
individual congeners in site soils. This preliminary analysis was performed to evaluate the
potential magnitude of the sediment contamination prnblem, using the Record of Decision soil
remedial goal for ground-water protection as a starting point. The mean congener concentrations
for all of the sediment samples exceeded the estimated congener-specific goals for ground-water
protection, and several median congener concentrations exceeded the estimated congener-
specific goals for ground-water protection. However, when concentrations were converted using
TEFs and the resultant values summed to obtain a sample-specific TEQ, none of the sediment
samples exceed the R'OD-specified soil remedial goal for ground-water protection.
'
Reevaluation was then done of the soil remedial goals for ground-water protection.
Modifications to the procedures used to calculate the soil remedial goal in the ROD were that
(1) no dilution of cdntaminated recharge was considered, (2) the organic carbon in the buried
sediment samples was used to estimate the contaminant-specific soil-water partitioning
coefficient, and (3) congener-specific soi_! _remedial goals w_ere calculated.
• •
-27-
The North Carolina ground-water protection standard for dioxin was again used as the ground-
water protection standard in the calculation of soil remedial goals for ground-water protection.
However, the state regulations specify that the "practical quantitation limit" (PQL) is to be used
if the PQL exceeds the promulgated dioxin standard. The TEQ (toxic equivalency) at the
PCDD/PCDF PQL concentrations exceeds the North Carolina promulgated ground-water
standard for dioxin. Therefore, the PQLs for the PCDDs/PCDFs should be used to calculate the
soil remedial goals. This conclusion presumes that all (or at least most) congeners are present in
soils and ground water, such that the sum of congener-specific TEFs in the soil could result in a
ground-water concentration that exceeds the state promulgated 2,3,7,8-TCDD TEQ.
Because several con~eners are not present in the buried sediments in concentrations of concern
using the PQLs as the ground-water target concentrations, a more complicated analytical
procedure was considered to determine the soil remedial goals for ground-water protection. This
more complicated procedure is based on the concept of proportionate representation. In this
approach, only the congeners are considered that were detected in Fire Pond buried sediment I .
(soil) at concentrations above the congener-specific soil remedial goal for ground-water
protection determined using the PQLs as the ground-water target concentration. Of those
congeners, the ones that were more frequently detected in those soil samples are given a greater
weight in the calculation of soil remedial goals for ground-water protection. This procedure also
weights the soil cleanup goals such that relatively less mobile and less toxic congeners generally
have a higher soil remedial goal for ground-water protection while more toxic and
environmentally mobile congeners have a lower soil remedial goal for ground-water protection.
· ·· ----·Howevet;thisproceaure·generateasoirc1eanup-goalslhat werecoiisiaered uiii-ealistic-;-becai.i"se it·--
presumed that most of the congeners present at concentrations equal to the soil cleanup goals
calculated using this method would result in ground-water concentrations that would be present,
but in concentrations lower than the PQL.
The Fire Pond buried sediment soil remedial goals for ground-water protection calculated in this
report were converted to a TEQ value and that value was compared to buried sediment
concentration data. That analysis indicates there may be some small areas of buried Fire Pond
sediment that represent a continuing impact to ground-water quality. However, for the two
sediment samples where the TEQ exceeded the Fire Pond buried sediment TEQ considered
. protective of ground water, paired deeper samples showed much lower TEQ concentrations.
This condition is considered to indicate the shallower samples contain above-average
concentrations of organic carbon which is effectively trapping PCDDs/PCDFs in the shallow part
of the buried sediment, thus accounting for the high PCDD/PCDF concentrations. With this
interpretation, the conclusion of this report is that there 1s probably not a concern about leaching
of PCDDs/PCDFs from the buried Fire Pond sediments.
(
• -28-•
6. References
Akin, E.W., 2001, personal communication, Chief, Office of Technical Services, EPA Region 4
Waste Managemer Division.
Keystone Environmental Resources, Inc., l 992, Final Revised Remedial Investigation Report,
Former Koppers Company, Inc. Superfund Site, Monisville, North Carolina.
Mackay, D., W.Y. Shiu, and K.C. Ma; 1992, Illustrated Handbook of Physical-Chemical
Properties and Environmental Fate for Organic Chemicals, Volume II, Polynuclear Aromatic
Hydrocarbons, Polyclzlorinated Dioxins, and Dibenzofurans, Lewis Publishers, Chelsea,
Michigan.
North Carolina Department of Environment and Natural Resources, 2001, Advanced Notice of
Groundwater Quality Standards Rulemaking -ISA NCAC 2L .0202.
Summers, K., S. Gherini, and C. Chen, 1980, Methodology to Evaluate the Potential for
Groundwater Contamination from Geotlzennal Fluid Release, EPA-600/7-80-117.
U.S. EPA, 1989a, Detennining Soil Response Action Levels Based on Potential Contaminant
Migration to Ground Water: A Compendium of Examples, EPA/540/2-89/057.
U.S. EPA, 1989b, Interim Procedures for Estimating Risks Associated with Exposures to -----·---·-Mutiiresof Chlorinatea Dibeffzo=p:Dioxirzs cil'ldDioenzafaraiis ( CDDsaiid CDFs)aiTd 1989 -· ·--·
Update, EPA 625/3-89/016.
U.S. EPA, 1996, Soil Screening Guidance Technical Background Document, Office of
Emergency and Remedial Response Publication 9355.4-17 A.
U.S. EPA, 2000a, Exposure and Human Health Reasses,sment o/2,3,7,8-Tetrachlorodibenzo-p-
dioxin (TCDD) and Related Compounds, Part/: Estimating Exposure to Dioxin-Like
Compounds, Draft Final, Volume 4: Site-Specific Assessment Procedures, (EPA/600/P-0O/
OO!Ad).
U.S. EPA, 2000b, SW-846 On-Line, Test Methods for Evaluating Solid Wastes, Physical/
Chemical Methods, Internet address http://www.epa.gov/epaoswer/hazwaste/test/uncler. htm.
U.S. EPA, 2001, Koppers Company Superfund Site, Monisvilie, North Carolina, Draft Field
Investigation Project, SESD Project No. 01-0078, report prepared for Beverly Hudson, Remedial
Project Manager, U.S. EPA Region 4.
Van den Berg, M., L. Birnbaum, ·A.T.C. Bosveld, B. Brunstrom, P. Cook, M Feeley, J.P. Giesy,
A. Hanberg, R. Hasegawa, S.W. Kennedy, T. Kubiak, J.C Larsen, F.X.R. van Leeuwen, A.K.D.
Liem, C. Nolt, R.E. Petersen, L. Poellinger, S. Safe, D. Schrenk, D. Tillitt, M Tysklind, M. Younes, F. Warn, T. Zacharewski, 1998, "Toxic Equivale~cy Factors (TEFs) for PCBs, PCDDS:-____ _
PCDFs for Humans and Wildlife," Environmental Health Perspectives, Volume 106.
•
Appendix 1. PCDD and PCDF Concentration Data, Fire Pond Sediment
Samples from the Remedial Investigation
• •
Sediment Concentration Data in ng/Kg
Congener Sample (with sample depth, in feet)
SI 0-0.5 SI 2-4 S4 0-2 S4 2-4 srn 0-2 S10 2-4
,3,7,8, TCDD 12.9 2.5 0.92 0.5 9.4 0
1,2,3,7,8 PeCDD 23.1 1.8 0.69 0 8.1 0
1,2,3,4,7,8 H6CDD 67.6 7.9 3.5 0 24.4 0
1,2,3,6,7,8 H6CDD 205 20.2 7.6 0.71 80.3 2.4
1,2,3,7,8,9 H6CDD 290 17.2 11.8 0.93 I I 5 3.4
1,2,3,4,6,7,8 ll7CDD 16210 1610 586 150 5020 175 ncoo 215930 33730 24140 81100 58020 11960
,3,7,8 TCDF 6.2 0.88 0.52 0 5.7 1.4
1,2,3,7,8 PeCDF 4.8 0 0 0 4.5 1.3
,3,4,7,8 PeCDF 2.8 0 0 0 3.8 1.4
1,2,3,4,7,8 H6CDF 29.4 4.9 1.6 0 20.7 3.2
1,2,3,6,7,8 H6CDF 23 4 1.5 0 11.8 2.1
,3,4,6,7,8 H6CDF 17.8 4.3 2.5 0.47 18 1.7
1,2,3,7,8,9 H6CDF 0 0 0 0 0 0
1,2,3,4,6,7,8 II7CDF 1300 144 59.9 4 408 16.4
1.2.3,4,7,8,9117CDF 161 13.2 3.7 0 19 0
"ICJW 5970 773 30 I 27.5 1980 61.7
Sediment Concentration Data in ng/Kg
Congener Sample (with sample depth in feet)
S13A 0-2 S13A 2-4 SIS 0-0.5 SIS 2-4
,3,7,8 TCDD 6.2 3.8 2.5
1,2,3,7,8 PeCDD 16.8 8.1 3.4
1,2,3,4,7,8 H6CDD 55 . 26.9 12.5
1,2,3,6,7,8 H6CDn 188 89.3 28.3
1,2,3,7,8,9 ll6CDD 145 81.9 31.8
1,2,3,4,6,7,8 H7CDD 12800 6790 1820 41.
JCDn 117670 80550 22180 537
2,3,7,8 TCIW 1.2 1.5 I. I 0.51
1,2,3,7,8 PcCD~-2.6 0 0
2,3,4,7,8 PeCDF 1.6 0 0
1,2,3,4,7,8 H6CDF 28.8 13.7 3.5 (
l,2,3,6,7,8116CDF 16.1 11.3 3.4 (
,3,4,6,7,8 H6CDF 12. I . 9.9 7 (
~.2,3,7,8,9 H6CDF 0 0 0 (
,2,3,4,6,7,8 H7CDF 890 574 155 2.(
1,2,3,4,7,8,9 ll7CDF 46.3 48.3 10.2 I
"CDF 4380 1910 919 14.1
'
• •
Appendix 2. PCDD and PCDF Concentration Data, Samples from the EPA
2000 Investigation
Congener
•
Sediment Concentration Data in ng/Kg
Sample (with sample depth, in feet)
KP00lSLA KP002SLA KP003SLA KP003SLII KP004SLA
2,3,7,8 TCDD
1,2,3,7,8 PcCDD
1,2,3,4,7,8 H6CDD
1,2,3,6,7,8 116CDD
1,2,3,7,8,9 H6CDD
1,2,3,4,6,7,8 H7CDD
'"lCDD
,3,7,8 TCDF
1,2,3,7,8 PcCDF
,3,4,7,8 PcCDF
1,2,3,4,7,8 H6CDF
1,2,3,6,7,8 H6CDF
2,3,4,6,7,8 H6CDF
1,2,3,7,8,9 H6CDF
1,2,3,4,6,7,8 117CDF
l,2,3,4,7,8,9117CDF
"CDF
Congener
1.3 1.7 0 0
4 3.6 0.42 0
8.6 10 1.4 0.76
29 28 2.7 2.6
26 30 4.8 2.9
1300 1100 190 150
16000 14000 7100 5800
0 0 0 0
0 0 0 0
0.51 0.54 0.16 0
0 0 0 0
6.4 6.7 0.75 0.47
2.3 2.6 0.38 0
15 0.7 4.8 0
ISO 170 19 17
15 IS 3.9 0
530 420 0 0
Sediment Concentration Data in ng/Kg
Sample (with sample depth in feet) ,
69
410
1700
1900
4100
79000
140000
0
0
55
600
690
470
83
7200
3700
15000
KP00SSLA KP00SSLII Kl'006SLA K 1'1106S LIi Kl'007SLA
,2,3,7,8 TCDD 3.5 0 0.31 0 4(
1,2,3,7,8 PcCDD I 5.6 () I 0 22(
1,2,3,4,7,8 H6CDD' 14 0.71 2.7 0.56 571
1,2,3,6,7,8 H6CDD 45 0.88 8.3 1.6 110(
1,2,3,7,8,9 H6CDD 49 1.6 9 1.8 1601
1,2,3,4,6,7,8 H7CDD 2500 97 520 110 27001
'"lCDD 56000 12000 11000 5800 66001
,3,7,8 TCDF 0.54 0 0 0 15
,2,3,7,8 PcCDF 0 0 0.16 0 35
~,3,4,7,8 PcCDF 0 0 0.19 0 31
1,2,3,4,7,8 H6CDF 0 0 I. I 0 411
1,2,3,6,7,8 H6CDF 5.9 0 1.9 0.33 281
,3,4,6,7,8 116CDF 4.1 0 0.87 0 15(
,2,3,7,8,9 H6CDF 0.85 0 0 0 2.
1,2,3,4,6,7,8 H7CDF 310 0 61 6.6 3401
1,2,3,4,7,8,9 H7CDF 0 0 2.9 0.85 961
"CDF 890 () 160 0 7001
KP004SLll
0
1.5
4.2
17
II
1000
20000
0
0
0
0
3.6
1.6
0.58
150
0
:no
Congener
',3,7,8 TCDD
1,2,3,7,8 PcCDD
1,2,3,4,7,8 H6CDD
1,2,3,6,7,8 H6CDD
1,2,3,7,8,9 H6CDD
l,2,3,4,6,7,8 H7CDD
OCDD
2,3,7,8 TCDF
1,2,3,7,8 PcCDF
,3,4,7,8 l'cCDF
1,2,3,4,7,8 ll6CDF
1,2,3,6,7,8 ll6CDF I
2,3,4,6,7,8 H6CDF
l ,2,3, 7,8,9 H6CD F
l,2,3,4,6,7,8 H7CDF
l,2,3,4,7,8,9 H7CDF
--,CDF
• •
Sediment Concentration Data in ng/Kg
Sample (with sample depth, in feet)
KPUU8SLA KPUU8SLB KP009SLA Kl'U09SLB Kl'0lUSLA
0 0 650 0 13
0 0 1900 l 28
0 0 3300 3 73
3.2 3 7500 7.7 120
0 3.1 7100 9.7 130
110 130 210000 560 4400
4600 4300 550000 7900 7900
0 0 71 0.66 2
() 0 120 0 6,
0 0 69 0 37
0 0 .980 l.6 43
1.5 2.2 1800 2.4 55
0 2.4 560 I 44
0 0 120 0 14(
15 20 38000 66 750
0 0 2800 3.8 60(
0 0 75000 210 1700(
.
...
\:(~:-
• •
Appendix 3. Technical Review Comments and Responses to the Technical
Review Comments on the Report
•
UNITED STATES ENVIRONMENTAL PROTECTION AG~NCY
NATIONAL RISK MANAGEMENT RESEARCH LABORATORY
SUBSURFACE PROTECTION AND REMEDIATION DIVISION
P.O. Box 11.98 Ada, OK 74820
May 29, 2002
OFFICE OF
RESEARCH AND DEVELOPMENT
MEMORANDUM
SUBJECT: Comments on the Evaluation of the Source of Dioxins and Furans Detected in
Private Water-Supply Wells and Evaluation of the Soil Remedial Goals for
Ground-Water Protection, Koppers Company Superfand Site; Morrisville, North
Carolina (02-R04-004)
FROM: David S. Burden, Ph.D., Director
Ground-Water Technical Support Center
TO: William O'Steen, Environmental Scientist
Office of Technical Services, Waste Management Division
U.S. EPA Region IV
This memorandum is in response to your request for technical assistance with the
Koppers Company Superfund Site, Morrisville, NC, in the form of a technical review of the
report entitled Evaluation of the Source of Dioxins and Furans Detected in Private Water-Supply
Wells and Evaluation of the Soil Remedial Goals for Ground-Water Protection. The review was
co.nducted under my oversight by Dr. Hai Shen, Mr. Tom Sunderland, and Mr. Steve Yarbrough
of the Dynamac Corporation, an off-site contractor supporting EPA' s Ground-Water Technical
Support Center in Ada, OK. I have reviewed their comments and concur with them. If you have
any questions concerning these comments, please contact me at your convenience at 580-436-
8606 or by email at burden.david@epa.gov. Specifically, the Region requested a thorough
review of the above-referenced .report and requested we address two subjects of concern. These
concerns included:
1)
2)
Is the Koppers Site a potential source rif dioxin and furan contaminants detected in local
potable water supply wells? Part A of the report evaluates this concern.
The presence of dioxins/furans in .capped Fire Pond sediments in concentrations above . .
the specified soil remedial goal for ground-water protection established in the ROD may
pose an ongoing threat to ground waier. Part B of the report addresses this concern.
• •
General Technical Review Conclusions:
A conclusion of the report should state that the generated data do not rule out the Koppers
site as a potential source of the polychlorinated dibenzo-p-dioxin (PCDD) and polychlorinated
dibenzofuran (PCDF) contamination detected in the private water-supply wells. Additional
research and data is needed. to arrive at and defend such a conclusion. Suspected anthropogenic
sources of PCDDs/PCDFs in closer proximity to the potable water supply wells should be
evaluated to ascertain their viability as contaminant sources.
It is noteworthy that the PCDDs and PCDFs were identified as ground-water
contaminants having relatively low mobilities due to aquifer properties and congener partitioning
factors. However, in the event that PCDDs are bound to large (macro) organic molecules
associated with septic system releases, off site transport of these contaminants could
preferentially escalate along fractured discrete pathways (e.g., solution channels, fractures, joints,
etc.), resulting in the transport of congeners bound to large organic molecules through such
fractured systems at unpredictable velocities. Thus, there is a significant degree of uncertainty
inherent with the hydraulic conductivity and hydraulic gradient assumptions. Existing scientific
methodology used to identify specific flow paths which contaminated ground water migrates
though fractured bedrock has significant limitations. Any conclusions concerning time
requirements for off-site migration of various congeners to reach off-site wells should
acknowledge these limitations.
Another contention of the report used to discount the Koppers Site as a potential
contamination source is the detection of higher concentrations of some congeners in off-site
wells than in on-site wells. As indicated in the report, there are several potential sources of
PCDD/PCDF contamination, including the Koppers Site. The detection of higher concentrations
of some contaminants, such as 2,3,7,8-TCDD and 2,3,4,6,7,8-H6CDF, in the off-site wells than
in the on-site wells may only exclude site waste sources as being solely responsible for the
contamination. The data do not, however, definitively rule out the contamination detected in the
private water-supply wells as potentially being partially attributable to site waste sources.
Most importantly,Jhe report's conclusions were largely based on the comparative
analysis of the on-site and off-site ground-water contaminant median concentrations, which were
culled from no more than two rounds of ground-water samples (p. 56-62). Ground-water
contaminant concentrations fluctuate considerably with seasonal changes. The report raises this
issue in stating that the ground-water contaminant concentrations fluctuate enough to invalidate
selected statistical analysis (p. 56). In addition, the report indicates that the use of only seven on-
site monitoring points could render a comparison of up-gradient to down-gradient ground-water
concentrations unrepresentative (p. 56). This underscores the difficulty in identifying off-site
plume transport trends given the current remedial monitoring program. Greater time increments
and additional monitoring points are necessary to validate ihe specific approach of the report to
calculate on-site and off-site median ground-water concentrations. Therefore, it is recommended
that further remedial investigation be conducted to establish and/or confirm data credibility. It is
-2-
• •
also recommended that the potential for magnitude and extent of congeners bound to large
organic molecules at the site be evaluated.
Two general topics which warrant further explanation in Part A include: 1) specific
ground water recharge mechanisms driving the site's potentiometric high, and, 2) the intrinsic
permeability of the clean fill utilized to cap the Fire Pond and corresponding infiltration
(contaminant transport via the soil to ground water migration pathway) ramifications.
In addressing the Part B concern, a more rigorous method of estimating remedial goals
for protection of ground water than that utilized in the ROD was followed, and thus it is
concluded that a potential for release of dioxins and furans to ·ground water is a legitimate threat.
The assumptions and logic in arriving at this conclusion are conservative and more protective
than the ROD goals. Future work should focus on the risk to receptors from potential releases,
and the determination whether removal or contai11ment may be necessary to safeguard human
health and the _environment. Specific comments concerning these, and other more minor
technical and editorial considerations, are offered below.
Technical Review Comments -Part A:
Text:
-/pg 1
' pg 1
pg2
pg2
pg2
pg4
pg6
Comment:
Second paragraph: Describe intrinsic permeability of the "clean fill material."
Last line: Specify which solvents were used, and presumably, are bein·g
remediated along with PCP.
Second paragraph, last line: Descri_be intrinsic permeability of "clean soil," i.e .. is
it clay? Was the cap engineered? Was it compacted? Is it a uniform thickness?
Is it maintained?
Fifth paragraph:. It would be helpful to elaborate on the nature of the analyses
performed under EPA oversight. Were they multimedia analyses? Were rigorous
Contract Laboratory Program (CLP) sample collection and analysis protocols
followed? Were the dioxin/furan congener analyses performed using SW-846,
Method 8290 (as opposed to Method 8280 in that lower reporting levels are
achievable)? If so, a brief summary would serve to render associated data
reliability·and usabilicy issues more defensible.
• Sixth paragraph, last line: Describe why the four off site monitoring wells are
considered "key" or call out"section of report that does.
First paragraph: With respect to the remedial action program, which ground-
water contaminants of concern is the program designed to address/capture (PCP?
Solvents? BNAs? Metals?). More importantly, state whether the remedial
program, by design or not, functions to mitigate off-site PCDD/PCDF migration
via the ground-water pathway.
Second paragraph: Include discussion regarding monitoring well specifications.
Are any«A," "B," and "C" zone wells nested? Clarify that some A zone wells
also have significant open hole completion (i.e., C27 A has 17 feet). The
statement is made (pg 38, second paragraph) that the lower portion of the A zone
. -3-
pg6
pg27
pg57
pg63
•
and the B zone represent the "anticipated interval of principal horizontal
contaminant transport." The reader is then directed to Section 4.2. This summary ·
ground-water flow conclusion should be stated as one of the bullets starting at the
bottom of page 7.
Third paragraph: The discussion about subsurface mechanisms (aquifer presence,
depth, fracture orientation, anisotropy, ·porosity, etc.) and their control over
ground-water flow is excellent. Additional discussion of surface recharge
mechanisms would be helpful. The opening statement raises several questions not
addressed in the ensuing narrative. Specific recharge mechanisms (natural,
seasonal, climatic) resulting in the year-round "potentiometric high" should be
explained. If gravity is the dominant driving force in ground-water movement
away from the Koppers site in multiple directions, and, in effect, the water table is
a subdued replica of the topographic surface, then this conclusion should be
stated. What other surficial physical characteristics of the site (soil, plant cover,
slope, water content of surface materials, rainfall intensity, anthropogenic, etc.)
may significantly affect recharge? How do on-site and nearby surface water
bodies (Western Drainage Ditch, Medlin Pond, Eastern Drainage Ditch) influence
ground-water recharge? Do the ditches have perennial flow? What influence, if
any, do paved surfaces and storm water runoff diversion systems associated with
the facility have on localized ground-water recharge? Has air entrapment during
recharge of the shallow unconfined aquifer been noted? As part of the remedial
program, have any wells been used for injection purposes to dilute or control the
flow of contaminated ground water? ·~0:/'!'
Showers and associated units of measure_ (pg/shower) are not inherently
comparable. Comparable units (pg/L) should be reported if available .
. Cl6C comparison: The distance between CI6C and the "site boundary" (-360 m)
is used to estimate the time required for offsite contaminant transport to the well.
The "site boundary" is closer to off-site wells than specific on-site waste sources
and somewhat arbitrary. Measuring from it, rather than spatially defined waste
source(s) for which contaminant attribution has been established, introduces yet
another inherent component of conservatism to the overall conclusion. This
comment also applies to the CI9C and C20C comparisons (pg 60) and the C2IC
comparison (pg 61). This consideration should be adaressed in thc;_preface to the
well by well discussion.
Third paragraph: Regarding the contaminant carry down postulate; are filtered vs.
unfiltered metals results from the RI sample event available? If so, the results ·
should be evaluated to help confirm or refute the presence of significant turbidity
(sediment) entrained in the sample matrix.
-4-
• • •
Figures -Part A:
9
9
Dashed segments of the 0.25-foot drawdown contour appear to be inferred and
should be labeled accordingly in the legend.
An A-A' axis descriptor should be included in the legend.
Tables -Part A:
1 Suggest that orientation reflect general direction of flow. Z data is accurate to
within 0.01 foot. This precision suggests that a wellhead survey has been
accomplished. X and Y data (disiance between wells) is approximated (rounded
off) to nearest 10-foot increments. If greater precision is available (survey, GPS,
or tape), it should be utilized to estimate hydraulic gradients.
Technical Review Comments -Part B:
pg 11
Table 9
,,..Table 9
Equation (2): The preface to equation (2) defines "m" as the vertical mixing zone
of the aquifer. Equation (2) solves for "d" -the mixing zone thickness. It appears
that both letters are intended to represent the same aquifer parameter. If this is
indeed the case, use one or the other in both locations. Otherwise, explain their
diffe_rence i~ the preface.
The "Values for Soil Remedial Goal for Grnund-Water Protection"·(EPA
and WHO), for the 1,2,3,4,6,7,8 H7CDD congener appear to have
incorrect results calculated. A concentration of 117 lng/kg is reported; the
correct calculation appears to be 504.81 ng/kg for each value.
It would be useful to include values of mean and median congener
concentration in sediment samples from the Fire Pond.
Editorial Review Comments -Part A:
Text:
Lof A
Lof A
p_g 1
pg 1 _.,,
pg4
p_g 6
pg6
pg 17
Comment:
Add 'foe' to the acronym list (stated on Table 2).
Add 'K.,w' to the acronym list (stated on pg 46 and Fig 15).
Add page number.
Third paragraph, last sentence: Delete extra space between "that the."
Third paragraph, fifth sentence: Delete extra reference parenthesis.
First paragraph, last word: Change to "bedrock."
Third paragraph, third sentence: Replace the word "periods" with "seasons."
Fourth paragraph, first sentence: Replace "the direction" with "their proximity to
-5-
,
pg23
pg24
pg25
pg40
pg40
pg64
pg64
pg65
pg72
pg74
pg 75
pg77
pg77
pg79
App. I
PW-1..,'
Section 5.2, first sentence: Replace "certain of the" with "specific."
Second list of congeners: For document consistency purposes, format second list
similar to first list, i.e., insert commas between congener numbers and delete units
(provide units to reader in narrative preface).
See preceding comment. ·
Third paragraph: Delete extra space between "Kd" and "estimates."
Third paragraph: Add statement to the effect that, for contaminant transport
evaluation purposes, the Table 3 soil-water partitioning coefficient determination
assumptions are conservative.
Third paragraph, second sentence: Change the words "are be" to "can be."
Last sentence: Change the words "of-Site" to "offcSite."
Second to last line: Delete second period at end of sentence.
Third paragraph, second sentence: Change "congener" to "congeners."
Conclusion 7: Change text from "private well samples contaminant..." to."both
private and monitoring well samples include contaminant..."
References: The order of the first two references (Alawi et al. and Akande et al.)
should be switched (alphabetized).
Fourth reference (Nestrick et a.I): Delete space between the comma following the
word "Chemosphere." ·
Seventh reference: Insert space prior to the word "California."
First reference: Place period at end of reference.
Revise the spelling of the reference "Mckay et al, 1992" to "McKay. et al., 1992."
Editorial Review Comments -Part B:
Text:
pg2
pg 10
pg27
Comment:
First paragraph: Change '"fable 1( as ... " to '"fable 1 (as ... "
Third paragraph, third sentence: Change "Summer" to "Summers."
Third reference: Insert comma following "Chelsea."
App. 1 and 2 For document consistency purposes, insert commas between congener numbers.
cc: Rich Steimle (5102G)
John M. Cunningham (5204G)
Kay Wischkaemper, Region 4
Felicia Barnett, Region4
-6-
• •
Notes: (I)
B.
In this response to comments this report is refe,,-ed to as "the Report"
Comments may cite Report sections that have been changed as a result
of additional Report modifications.
I. General Technical Review Conclusions
The principal comment contained under the heading General Technical Review Conclusions is
that there is a greater degree of uncertainty regarding the source(s) of PCDDs/PCDFs in the
private water-supply well samples than is indicated in the draft report. Several points are offered
as reasons for why there is a greater degree of unce,1ainty regarding the source(s) of the
PCDDs/PCDFs in the private well samples:
a.
b.
C.
d.
The potential°for PCDDs/PCDFs to be bound to large organic molecules may result in
accelerated transport of the contaminants via the ground-water pathway.
There is a significant degree of uncertainty inherent with hydraulic gradient and hydraulic
conductivity assumptions, when applied to a fractured bedrock hydrogeologic setting.
Thus, ground-water and contaminant transport velocities in such settings arc often not
predictable with any degree of certainty.
Higher concentrations of PCDDs/PCDFs in off-Site wells when compared to on-Site wells
may be indicative of other contaminant sources but does not rule out the Site as
contributing to the observed off-Site ground-water contamination.
There are inadequate data to fully characterize on-Site ground-water concentrations and
thus the comparison of on-Site to off-Site ground-water quality is a tenuous line of
evidence for the evaluation.
Responses to these points are presented below. Report modifications have been made to address
at least some of the comments, as indicated in the responses.
b. The potential for enhanced PCDD/PCDF ground-water transport via organic
macromolecules is noted i.n Section 5.3.4 of the repo,1. As that discussion indicates, the
degret;. of enhanced contaminant transpo11 resulting from the presence of organic
macromolecules is-related to the concentration of organics present in the ground water.
The concentration of dissolved organic carbon in the ground water at the Koppers Site has
not been measured; however, as noted in the reference, the anticipated dissolved organic
carbon, based on literature data, is in the range of 10 mg/L to 20 mg/L. As noted in the
report, such a dissolved organic carbon concentration in ground water might result in a
PCDD or PCDF with a log octanol-water pm1ition coefficient of 8 having an
environmental mobility 3 to 4x greater than would be the case if the dissolved organic
carbon was not present. PCDDs/PCDFs in samples from water-supply wells around the,
Koppers Site have been either l,2,3,4,6,7,8-H7CDD or OCDD. According to Mackay et
al (1992), the median log octanol-water partition coefficient for l,2,3,4,6,7,8-H7CDD is
10.52 (23 estimates) and the median octanol-\vater partition coefficient for OCDD is 10.07
. .
(39 estimates). Considering these values, along with Figure 15 in the report, the ground-
water mobility of these PCDD congeners may be on the order of 25x higher than would be
• •
the case if organic macromolecules were not present. Using Table 4 as a basis for the
mobility of the PCDD congeners in the absence of any organic macromolecules (see the
discussion in the response to point b below regarding the estimation of the ground-water
velocity), the potential transport velocity of l,2,3,4,6,7,8-1-17CDD and OCDD in the
presence of between 10 and 20 mg/L diss·olved organic carbon as organic macromolecules
would be approximately 3E-3 mid and l.23E-3 mid respectively. Over the time period
when the Koppers Site initially began operations to the present (I 962 to 2002; a
conservative estimate of the time available for ground-water transp011 of the PCDDs to
off-Site locations), 1,2,3,4,6,7,8-H7CDD, the more mobile of the two PCDDs detected in
· water-supply well samples, would have migrated an estimated distance of 43.8 meters
(144 feet) from a source area. Allowing for even more uncertainty in the ground-water
velocity than was considered in the draft repon for contaminant-transport calculations,
there is a low probability that l,2,3,4,6,7,8-H7CDD and OCDD would be transported from
the source areas at the Site for a distance of more than 2000 feet to the nearest water-
supply well with any detectable PCDD contamination.
c. The comment about uncertainties regarding ground-water movement in fractured bedrock
is valid. The evaluation of the ground-water velocity in bedrock is complicated and
estimates of the velocity may be very crude approximations of the actual ground-water
velocity.
In the Koppers RI Repo11, the maximum estimated ground-water velocity in the bedrock
was 164 feet per year. This value compares to an estimated velocity of approximately
2400 feet per year presented in the draft report (Section 4.2.2). The discrepancies are a
result of:
(I) using a more conservative (lower) effective porosity than that used in the RI (lower
by a factor of 5). The RI effective porosity was based on four estimates of the
aquifer storage coefficient calculated using the Cooper-Bredehoeft and
Papadopulos method of aquifer test analysis, which, according to Lohman, 1972.
probably provides only a rough estimate of the storage coefficient (effective
porosity).
(2) using a hydraulic conductivity of 0.8482 ft/d (the maxi-mum reported hydraulic
conductivity from on-Site aquifer testing) which compares to a maximum average
hydraulic conductivity of 0.393 I ft/d cited in the RI Report, Table 3-6.
(3) using a hydraulic gradient of 0.00777 versus a hydraulic gradient of 0.00591 along
the principal hydraulic conductivity tensor (i.e. in the northwest flow direction).
Thus, the estimated bedrock ground-water velocity used to calculate the contaminant
transport velocity is more conservative, by more than an order of magnitude, than that
cited in the RI Report.
Further conservatism in the estimate of the ground-water contaminant transport velocity is
incorporated into the analysis in the report. Referring to report Section 5.3.3, an org~nic
carbon content of approximately 10% of the lowest literature-reported organic carbon
content in geologic materials and approximately I% of the lowest observed organic carbon.
• •
content for near-surface (6 to 8.5-foot deep) earth materials was used to estimate the soil
(solid)-water partitioning coefficient for the PCDDs/PCDFs. Use of this low estimate of
the organic carbon content available for contaminant sorption in the bedrock resulted in a
high estimate of the ground-water contaminanttransport velocity for l,2,3,4,6,7,8-H7CDD
of l.2E-4 mid (repmt Table 4). Even if this velocity is underestimated by two orders of
magnitude, the transport of l,2,3,4,6,7,8-H7CDD over a 40-year period would be
approximately 575 feet from the source area. The water-supply well closest to the
Koppers Site with a detection of l,2,3,4,6,7,8-1-17CDD is well KP005 (reference Figure
23). KP005 is approximately 2700 feet from the Site boundary. Thus, assuming
·contaminant transport at a rate more than IOOx that used in the draft repo11 would not
i-esult in l,2,3,4,6,7,8-H7CDD migration to the private well in a 40-year period. Note that
the calculated 2700-foot transport distance is the straight-line distance between the Site
boundary and the water supply well. Ground-water flow in fractured bedrock will be
along a more tmtuous pathway of intersecting fractures. Thus, calculation of the tr:ivel
time required for PCDDs/PCDFs to migrate from the Site boundary to any water-supply
wells has additional conservatism built into the calculations.
Finally, in the draft report, the calculation of solute velocity using equation (3) from
Moreno ct al (1997) or equation 9.18 from Freeze and Cherry (1979) (original reference
was Section 5.3.3.2 of the draft repo1t) assumed that the soil-water partitioning coefficient,
Kd, was applicable to the analysis of contaminant partitioning onto the walls of rock
fractures. As noted in Freeze and Cherry ( 1979), the equations relating the solute velocity
to the ground-water velocity require an estimate of the distribution coefficient on a per unit
surface area basis (K,), rather than on a per unit mass basis. The soil-water partitioning
coefficient is typically calculated as the product of the fraction of organic carbon and the
contaminant-specific organic carbon pa1titioning coefficient (Koc). As such, the published
values of the Koc for PCDDs/PCDFs represent the distribution coefficient on a per-surface
area basis for organic carbon. Thus, by considering the fraction of organic carbon
available along the rock fracture walls, the available surface area is accounted for by the
fact that the pmtitioning is assumed to be between water and the organic carbon along the
rock fractures, rather than between water and the entire surface area available for
contaminant sorpti6n.
In the report, calculations ofground-water velocity assumed a fracture aperture of I mm.
A larger fracture aperture would result in a higher rate of ground-water flow and thus
higher potential contaminant transpo1t. According to Appendix J to the RI Report, the
aperture of a fracture at PW-I (the aquifer test well at the Site) was at least 0.25 feet. A
0.1-foot aperture width was assumed in RI Report Appendix J calculations. Such widths
of fracture aperture appear to be highly unrepresentative of the aquifer materials, as the
aquifer test data from RI Report Table 3-5 indicate an aquifer transmissivity on the order
of approximately IE-02 ft/d (14.4 ft2/d), which indicates an.aquifer with moderately lo\\:
yield. This moderately low yielding aquifer is fu1ther supported by the specific capacity at
the pumped well, which was approximately 0.2 gpm/ft at t = 29.7 hours. The open
fracture at PW-1 is therefore considered to be unrepresentative of the bedrock as a whole.
\
A statement about the uncertainties inherent in the calculations of ground-water
• •
contaminant transpo11 in fractured bedrock has been added to the repott as text
immediately preceding Table 4, Estimated Ground-Water Contaminant Transport
Velocities of PCDDs imd PCDFs.
c. The comment regarding a comparison of on-Site concentrations to off-Site·concentrations
as a basis for concluding that the Site is not a source for PCDDs/PCDFs is technically
valid. That is, an off-Site concentration higher than the median on-Site concentration of a
PCDD or PCDF may only be indicative of a greater contaminant contribution from a
source other than the Koppers Site to the PCDDs/PCDFs detected in the water-supply well
sample. The wording of the rcpott has been changed where appropriate to indicate that
this line of evidence provides a clue as to the principal source of PCDD/PCDF
contamination detected in the water-supply well samples.
d. There arc a limited amount of on-Site PCDD/PCDF ground-water concentration data.
However, when considered collectively, the available data allow for an approximation of
the overall degree of PCDD/PCDF ground-,vater contamination represented by the Site.
For purposes of this report, the median concentration data from the on-Site wells were
compared to individual off-Site well concentrations of PCDDs/PCDFs, and an off-Site
concentration that was higher than or approximated the median on-Site concentration was
considered as evidence that the Koppers Site was an unlikely source (or an unlikely
primary source, per the revised report) for the off-Site contamination. The presumption
that the median on-Site concentration may be a very crude approximation of the true
median due to the limited amount of data available to calculate a Site median docs not
invalidate this point. This statement is made because the comparison of on-Site median to
off-Site concentrations assumes no dilution of contamination migrating from the Site.
Given the Site hydrogcologic setting and contaminant fate and transpott considerations, it
is vi1tually ce1tain there would.be a significant degree of dilution and concentration
reduction by the time any Site-related contamination reached one of the off-Site well
locations. Thus, an off-Site concentration of PCDDs/PCDFs from wells hundreds or
thousands of feet from the Site would have to be much less than the estimated on-Site
median concentration in order to conclude that the Koppers Site was potentially the
principal source of that contamination.
The first paragraph under the heading General Technical Review Conclusions states "the
generated data do not rule out the Koppers site as ·a potential source of the polychlorinated
dibenzo-p-dioxin (PCDD) and polychlorinated dibenzofuran (PCDF) contamination detected in
the private water-supply wells." This point is not made in the draft repott nor in the revised
repott. The repo11 makes conclusions about the likelihood of the Koppers Site being the source
for that contamination. There is a possibility that the repo,t's conclusions drawn from the
available data are incon'ect and that the Site is the source of the contamination detected in the
water-supply well samples. This possibility is considered unlikely but can probably never be
completely dismissed, regardless of the amount of monitoring data available and number of ddta
analyses performed.
A point that is not made in the repo1t is that of the six private wells that have been sampled during
'ill three monitoring events (I 998, 1999 and 2000), none of the wells has yielded more than one
• •
sample with a detectable concentration of one of the seventeen congeners that contribute to the
calculated TEQ. This situation is inconsistent with the potential presence of a plume of Site-
related contamination representing the migration of mobile PCDDs/PCDFs in the ground water.
This observation, along with the absence of any PCDFs in the water-supply well samples are
additional evidence (albeit not conclusive evidence) that the Site is probably not the source of the
contamination detected in the watcr-suppl.Y wells.
2. Technical Review Comments
2.A Technical Review Comments; Table and Figures, Part A of the Report
I. The intrinsic permeability of the clean fill material can only be described in general terms,
as no permeability testing of the backfill ha's occuJTed. This backfill material is described
as primarily a silty/clay clayey silt; however, roughly 30% of the backfill is described as a
"granular matctial with fines." Given these descriptions of the backfill textures, it is
probable that most of the backfill has a permeability (hydraulic conductivity) in the range
of 10-5 to 10-6 emfs. The relevant text on page 2 of the draft report (reference the
response to comment 3 below) has been modified to indicate the texture of the fill
materials.
5. Solvents that were used in the wood treatment process arc rep0t1cdly liquified butane,
isopropyl ether, and a "glycol-based co-solvent reportedly also was used for a sho11 period
of time" (Keystone Environmental Resources, lnc., RI Report, 1992). The text has been
changed to indicate these solvents were used.
6. Available records indicate the bulk of the backfill is a silty clay to clayey silt material,
which probably has a hydraulic conductivity in the range of 10-5 to 10-6 cm/s. The fill is
graded to promote drainage, the cap was engineered but apparently was not compacted
and is currently covered with native vegetation. Changes have been made to the report to
indicate the nature of the fill and current conditions of the backfilled area.
7. The text of the repot1 has been changed to provide a more detailed description of the
analyses that were performed.
8. The-text of the repot1 has been changed to explain why the off-site monitoring wells are
considered as particularly impottant.
9. Text modifications have been made to prese11t more detail on the ground-water remedial
action and remedial action monitming program.
IO_ Text has been added to more fully describe the monito1ing well completions. With regard
to the second pat1 of this comment, the text on draft report page 17 was modified to
indicate the hydraulic conductivity of the deeper bedrock (monitored by the C zone wells)
is generally very low. This condition means that the A and B zone represent the '
anticipated interval of principal horizontal contaminant transport, as stated on repot1 page
36.
11. Some additional discussion of recharge and its influence cin the water table have been
• •
added to Section 4.2.2 of the repo11. The comment notes many factors that can, and
probably do, influence recharge at the Site. The specific influence of these conditions
such as soil type, vegetative cover, paved surfaces and so fo11h can only be discussed in
general terms. The site is located on a topographically high area at the head of several
drainage basins, and it is probable that the potentiometric high reflects-the topographic
position of the site. This point has been added to Section 4.2.2.
Other questions included with the eighth comment ask about air entrapment during
recharge of the shallow ground water and wells that may have been used for injection
purposes to dilute or control contaminated ground-water flow. Air entrapment during
·recharge events has not been investigated at the Site and does not influence the overall
observation of a potentiometric high being present. No injection of water into the
subsurface has occurred. The current ground-water extraction (sec response to comment 6
above) and backfilling of both the Fire Pond and modification of the Medlin Pond area
(now a wetland; reference U.S. Arrny Corps of Engineers, 2000) have had some effect on
the water table and distribution of hydraulic head at the site, but there is still a
potcntiometric high or series of potentiometric high areas in part of the Site, based on recent water-level data collected during routine monitoring of the remedial action.
12. The cited reference presents concentrations in units of pg/shower and states "As the
concentrations depend on the water quantity used for each shower, the rcsults ... are given in pg/shower and not pg/L." No changes were made to the repo11 text.
13. Changes have been made to the text in the suggested location to address this comment.
14. There are no metals ground-water quality data available from off-Site wells sampled
during the Remedial Investigation. This omission of metals data from off-Site wells was
apparently based on an evaluation of on-Site .monitoring data that eliminated inorganic
constituents as ground-water contaminants of concern. No changes to the repo11 text have
been made.
15. Figure 9 has been modified to address the comments.
16. Survey data were found in Appendix B to the RI Repo11 and modified distances between
wells were calculated and are used in Table I. The first pa11 of this comment states
"Suggest the orientation reflect general direction of flow." Ground-water flow at the Site
is in multiple directions, with the apparently most significant flow in the nol1hwest or
southeast directions, consistent with the p1incipal hydraulic conductivity tensor. However,
for the sake of completeness, the hydraulic gradient in the direction of the minimum
hydrau'lic conductivity value was also estimated.
2.B Technical Review Comments; Table and Figures, Pan B of the Repo11
I. The term "m" is generally used to represent aquifer thickness; while "d" in Equation (2)
represents the aquifer thickness across which ve11ical mixing is estimated to occur ("d" is
s"m"). The term "m" has been replaced by "·cl."
• •
2. Table 9 has been corrected to address the comment about the soil cleanup value for the
1,2,3,4,6,7,8-H7CDD congener.
3. Table 9 _has been revised to add two columns listing the mean and median congener
concentrations.
Editorial review comments: Text changes have been made, as appropriate.