NCD003446721_20080606_Celeanse Corporation - Shelby Fiber_FRBCERCLA SPD_Conceptual Site Model 2007 - 2008-OCR

UNITED STATES ENVIRONMENTAL PROTECTION AGENCY REGION 4
61 Forsyth Street
Atlanta, Georgia 30303-3104
4WD-TSS
SUBJECT:
MEMORANDUM May-, 2008
Celanese Conceptual Site Model Report dated September 2007
David N. Jenkins, Environmental Scientist FROM:
THROUGH:
Technical Support Section, Superfund Support Branch
Glenn Adams, Chief, Superfund Support Branch
TO: Beth Walden, Remedial Project Manager
Beth,
I have read portions of the Conceptual Site Model Report for the Celanese Site dated September, 2007
as you requested. Here are my comments. Please call me at 404-562-8462 if you have any questions.
The document reviewed is titled:
EarthTech, 2007, Conceptual Site Model Report, CNA HoldingsfflCONA, Celanese Fiber
Operations, Shelby, NC, EarthTech, Inc., 1455 Old Alabama Road, Suite 170, Roswell,
Georgia 30076, Document# 026SB-258 September, 2007.
COMMENT REGARDING THE EXECUTIVE SUMMARY:
This report is supposed to be a concept model describing the site and the physical factors and features
which caused or control the migration of contamination. The Executive Summary contains statements
about the type and degree of remediation required which are advocacy positions that should not be part
of a Conceptual Site Model report. These are conclusions about remediation which are appropriate for an
FS, not conceptual descriptions of natural and man-made processes which work at the site. The last
paragraph in the Executive Summary (p.ES-2) " ... concludes that long-term monitoring is the most
appropriate alternative for the site." This is report is a site conceptual model, not a feasibility study. This
statement should not be in this report. Further, long-term monitoring is not a remedy. Long-term
monitoring does not reduce contaminant concentrations, prevent contaminant migration or return the
water in the aquifer to unrestricted beneficial use. EPA can not accept the conclusion that long-term
monitoring i~ the most appropriate alternative for the site.
Long-term monitoring is not Monitored Natural Attenuation. Monitored Natural Attenuation is a remedy
with EPA guidelines and nearly two decades of scientific basis. EPA has requested that groundwater
contamination at the site be characterized according to the appropriate EPA guidelines. The EPA
guidelines for MNA have not been followed at this site. This report does describe the groundwater flow
system at the site. It does present a description of the site geology, hydrogeology and important aspects
regarding the extent of groundwater contamination. As such, the report constitutes and excellent
Concept Model report. EPA must decide what to do with the findings of this report.
COMMENT REGARDING VARIOUS PLUMES AND VARIOUS CONTAMINANTS IN GROUNDWATER:
The report describes many important issue regarding groundwater contamination at this site. The
overriding issues seem to be the following.
1.) There are three primary contaminants in two groundwater plumes at this site:
Diethylene oxide also commonly known as 1,4-dioxane;
Ethylene glycol; and
Trichloroethylene (TCE).
2.) There are three known primary source areas for contaminants in groundwater at this site:
Disposal areas in the vicinity of the Inner Tier (IT) recovery wells (1,4-dioxane & ethylene glycol);
The vicinity of monitoring well F-55 (ethylene glycol);
Page 1 Printed: June 6, 2008 (12:01PM) -DRAFT search for..: and resolve issues before final print
C:\Documents and Settings\dmattison\My Documents\Documents\2008\NPL Site Documents\Celanese
Corporation\0804 Conceptual Site Model 0709 DNJ Comments.doc
• •
The vicinity of the SE corner of the main plant building (TCE in monitoring wells TD-3 and TD-4).
3.) There are no remedial measures in effect at the site for any groundwater plume from these source
areas at this time.
The contaminants from the disposal area around the Inner Tier wells are similar to those at F55,
but these are believed to be two separate source areas. No clean-up measures have
been implemented at F55. The extent of contamination at F55 may have been defined
laterally, but not vertically. Clean-up measures at F-55 should be developed and
implemented soon.
Some source removal activities were implemented years ago in soils near the Inner Tier wells. A
pump & treat system operated from the Inner Tier wells between 1989 and 2004 when
the pumping system was shut off with the agreement of EPA for an evaluation of MNA.
The MNA investigation provided some information, but did not identify mechanisms or
agents promoting degradation, did not provide information from monitoring wells in the
plume down gradient from the source areas which documented the progress of
degradation after the pumps were off, and generally did not define the MNA processes at
this site following methods described in EPA guidelines for MNA investigations. The
monitoring well network which was designed for the pump and treat system was not
adequate to monitor the plume once the pumping wells were turned off.
Natural attenuation may effectively limit migration of ethylene glycol, but Celanese states in this
document that diethylene oxide will migrate off Celanese property. EPA has
recommended that pumping at the Inner Tier wells be resumed because migrating
plumes are not candidates for MNA as a remedy. EPA policy expects contamination to
be contained. Water which is currently drinkable should not be allowed to become
contaminated simply because no one is drinking the water. Celanese has not agreed to
the recommendation to restart the pumps, and the plume is not under control at this time.
No remediation of the TCE plume has occurred or has been contemplated because Celanese has
maintained that the TCE observed on site has been trivial and could not be responsible
for the TCE concentrations observed in the off-site wells (HH-48 & HH-78) on the Elliot
property. Celanese maintains this position regarding TCE despite information in this
report which shows TCE concentrations on-site are higher than previously observed and
are increasing. The TCE plume may be migrating from the source area somewhere in
the vicinity of well TD-3. Figure 5-1 shows that the question of whether there is a
connection between wells TD-3, TD-4 and the HH wells on the Elliot property is no longer
the most important question regarding TCE in groundwater at the Celanese site. TCE
concentrations in TD-4 have increased from slightly over the NC ARAR in 1996 to
2,630µg/L in 2006 (Table 5-4).
The TCE concentration in TD-4 is nearly 1,000 times greater than the NC ARAR. TD-4 is the
deeper monitoring well in the TD-3/TD-4 well pair. Between 1996 and 2006, TD-4 has
become the most contaminated well, suggesting that downward migration of TCE is
occurring. EPA must conclude there is a TCE plume on Celanese property. This plume
may or may not be connected to the HH wells. Steps to define the extent of
contamination should be implemented immediately. Measures to remediate this plume
should be implemented in a reasonable time and following a process determined by the
EPA RPM for this site.
COMMENT REGARDING CONTINUED MIGRATION OF DIETHYLENE OXIDE IN GROUNDWATER:
Page 2 Printed: June 6, 2008 (12:01 PM) -DRAFT search for -and resolve issues before final print
C:\Documents and Settingsldmattison\My Documents\Documents\2008\NPL Site Documents\Celanese
Corporation\0804 Conceptual Site Model 0709 DNJ Comments.doc
• •
The Executive Summary (p.ES-1) states "With the exception ofdiethylene oxide, the site chemicals are
attenuated naturally within the site perimeter." First, the distribution of the TCE plume on-site has not
been determined. TCE has been detected in off-site wells (HH-47 and HH-78) down gradient from the
facility. Second, an appropriate MNA program should already be in place to support this statement, but it
is not. Third, exceptions are not acceptable and should not be ignored. Either natural attenuation is
working on the contaminants at this site or it isn't. If diethylene oxide is an exception and is migrating
' from the site, then natural attenuation is not working. If diethylene oxide is migrating from the site,
measures to contain the plume should be implemented. (Please note: diethylene oxide is also commonly
known as 1,4-dioxane.) Whether natural attenuation is working or not, a groundwater monitoring program
which follows EPA guidelines should be implemented.
The Executive Summary (p.ES-1) states" ... diethylene oxide could potentially migrate off site ... to an
area with existing water use restrictions ... [then] discharge to surface water where it can potentially
photodegrade." This statement is a suggestion from Celanese to accept potential diethylene oxide
migration to off-site areas where land use controls may limit exposure and the contamination may
degrade after it discharges to surface water. This proposal does not follow EPA policy described in the
NCP. Allowing contaminant migration is not acceptable, contamination of water which currently is
drinkable is not acceptable, and knowingly permitting contaminated groundwater to discharge to surface
water is not acceptable.
EPA typically does not permit plumes of contaminated groundwater to migrate regardless of existing
water use restrictions or land use controls. Water use restrictions do not cleanup contaminated
groundwater or prevent contaminant migration. EPA should not accept a statement such as" ... discharge
to surface water where it can potentially photodegrade" with any comfort. Neither the potential for
successful photodegradation, nor impacts to potential eco-receptors in surface water before degradation
occurs have been quantified. Also please note the highest 1,4-dioxane value reported on site, about 3
mgll, is nearly 500 times the EPAR9 PRG for 1,4-dioxane which is 6.11 µg/L. Migration of the diethylene
oxide (1,4-dioxane) plume should be controlled.
The Executive Summary (p.ES-1) states "Conservative modeling of the fate of ethylene glycol indicates
that the existing plumes will not expand." An appropriate MNA program should already be in place to
support this statement, but it is not. EPA guidelines for MNA specifically state that demonstrations of
MNA must be based on site-specific observations not models or calculations. The groundwater model
presented in the report appears to be a reasonable simulation of groundwater flow paths at the site. It is
useful for planning purposes, but regulatory decisions must be based on samples and measurements
from the site, not on model results.
COMMENT REGARDING TCE IN GROUNDWATER:
The Executive Summary (p.ES-1) discusses trichloroethene detections at TD-3 and TD-4, and states that
these detections are unrelated to off-site detections at the two HH wells. EPA has never agreed with this
conclusion and has recommended additional investigation (Everett Glover, EarthTech, minutes of
meeting on 01FEB07 pages 3 & 4).
The Executive Summary (p.ES-1) states that wells located up gradient from the HH wells have not
reported comparable trichloroethene detections. The argument is based on the idea that if the TCE
source is at Celanese, some degradation is bound to occur along the pathways from the site to the HH
wells, therefore concentrations up gradient should be greater than concentrations down gradient. But
EPA has noted previously (D. Jenkins e-mail to Beth Walden 30MAY07) because" ... one of the wells
(TD4) ... has TCE at 526 times the MCL. This on-site concentration is high enough to be responsible for
the off-site concentrations observed at the HH wells off site. " (Also see similar comments in D. Jenkins
memo dated 04JUN07). There is no obvious source for TCE at the HH wells. There is not yet an
identified pathway between the TCE wells on site and off site, but information in this report suggests that
Page 3 Printed: June 6, 2008 {12:01 PM) -DRAFT search for -and resolve issues before final print
C:\Documents and Settingsldmattison\My Documents\Documents\2008\NPL Site Documents\Celanese
Corporation\0804 Conceptual Site Model 0709 DNJ Comments.doc
• •
a pathway from the vicinity of monitoring well TD-3 through the vicinity of monitoring well T-35 to the
vicinity of the HH wells is possible.
Regardless of whether the TCE contamination on-site is connected to the HH wells, TCE contamination
on-site is increasing in some monitoring wells. The TCE plume may be migrating deeper into the aquifer.
In this report, the TCE situation is summarized best in Figure 5-1.
i! ln04 Concrptu.il S1tr Mc,,di_,10709 F111urt' 5 1 P'llf • Adobe RNder G]@(E]
'' ,~-I ~> o't',t , r,'i'
r-11-
' -/..!\. fi I
l ~-•\ '
' \ .. 1 .. ~·11nt11, :W,19fol 12;·1,,_ g'1'8'1M Y-!4'2002 2,11®0!! '1J1v,;,,»? ---1-N:;:a.-.-T-lSj I ----'
~· ~------------1 'i'THMI 61•"1W\ l71I~ l;!lt,1006 1211,1m !>?<t.'002 2117.'2\Yl!i 11•, .. ~1
.... p .... -
)--fh<-8-•->ttn-,-,;;21.I
i '' 1------,.,;,","'ic'-_,,?---~
j , .. t-------'l-f-'/-7'----.A--cr=----=----i
0
°''"(~--------------' •·111111 G:0111091 ll11!1KM 1Zl111Wli l/23'1111'11 ~:2•'2002 :!Jt7,~ ,,,,.,~7 _.,...,0o1.
j-..::21. -•---• -♦-10-3 -•-TIH -•--IH• I
NOTE:
(1) HOLLOW SYMBOL INDICATES NON-DETECT
FIGURE 5-1
@EarthTech TCE CONCENTRATION PLOTS
A l)ICO lnltnlf!..,il ltd Conp1rry
SEPTEMllEA 2007
CNA Holdings, lnc/ricooa
Shelby, North Carolina
NOTE the original figure in the report is in color and may be easier to read.
NOTE that the Y-axis in all 3 graphs shown on Figure 5-1 is logarithmic not arithmetic, and the
scale of the Y-axis on the 3 graphs is not the same. There is nothing wrong with the
presentation style, but care is required in interpreting the graphs.
NOTE ALSO that the solid (red in the original graph) horizontal line on all three graphs is the NC
ARAR for TCE concentrations in groundwater.
Figure 5-1 of the Site Conceptual Model report shows the results of TCE samples in three different areas
at or near the site.
•I
TCE concentrations in the HH wells are shown in the graph on the lower left of Figure 5-1. These
2 off-site wells are located approximately 3,500 feet southeast from the main plant
building (see Figure 3-4). HH-48 is 48 feet deep and HH-78 is 78 feet deep. TCE
concentrations have been higher in the deeper well since 1994. Celanese has
Page 4 Printed: June 6, 2008 (12:01 PM) -DRAFT search for -and resolve issues before final print
C:\Documents and Settingsldmattison\My Documents\Documents\2008\NPL Site Documents\Celanese
Corporation\0804 Conceptual Site Model 0709 DNJ Comments.doc
• •
maintained that TCE contamination in these wells is not related to activities at the
Celanese site.
TCE concentrations in well T-35 area well are shown in the graph on the upper left of Figure 5-1 .
T-35 is an on-site well located approximately 1,800 feet southeast from the main plant
building (Figure 3-4). Well T-35 is 35 feet deep. Numerous TCE detections have been
reported, but this well is shallower than the HH wells or TD-4. The groundwater model·
results in this report and the TCE concentrations in other wells suggests the much of the
plume may be deeper than well T-35.
TCE concentrations in wells TD-3 and TD-4 are shown in the graph on the right side of Figure 5-
1. These wells are located near the SE corner of the main plant building (Figure 3-4).
TD-3 is 45 feet deep and TD-4 is 79 feet deep. The TCE concentration in TD-4 has
increased from the ARAR for TCE in 1996 to 2,630µg/L in 2006. TCE concentrations in
on-site wells PEW-4, TD-3 and TD-4 are similar to observed in the off-site HH wells.
Some simple but important observations can be made from the presentation of TCE concentration trend
data shown on Figure 5-1.
TCE concentrations in the HH wells have not changed significantly since 1988. No TCE
degradation is apparent in the graph showing TCE concentrations at the HH wells.
lnstead-TCE concentrations may be increasing in these wells. It is clear that natural
attenuation has made little progress in the last 10 years. The deeper well HH-78 is the
most contaminated well. The well screens in these two wells are approximately 30 feet
apart. The well screen in each well is 5 feet long.
TCE concentrations in TD-3 and TD-4 typically have been much greater that found in the HH
wells (Figure 5-1 ). Note also that the TCE concentration in TD-4 has increased more
than an order of magnitude on the logarithmic scale since 1999. The deeper well has
become more contaminated than the shallow well during this time suggesting that
downward movement of contamination may be occurring. ·
In the TD-3/TD-4 area, the concentration in wells G-88 and PEW-4 historically have been similar
to the concentrations observed in the HH wells in some samples, but much lower in
others (Figure 5-1 ). The argument that the TCE contamination in the down-gradient HH
wells could not be from Celanese because the HH concentrations were notably greater
than found farther up gradient on the Celanese property can not be supported based on
these data because the TCE concentrations in G-88 and PEW-4 are similar to
concentrations in the HH wells.
The TCE concentrations in PEW-4 can not be compared directly with TCE concentrations in wells
G-88, HH48 or HH-77. The well screens in each of these wells are 5 feet long. The
screen in PEW-4 is over 31 feet long (Table 2-1 ). We don't know how thick the TCE
plume is at PEW-4. We don't know where the top and bottom of the plume is. We don't
know the portion of clean water entering the well screen versus the portion of
contaminated water from the plume. It may be that the TCE concentrations in the
contaminated portion of the screen in PEW-4 are much greater than the concentrations in
the HH wells, but when the well is purged for sampling, the contamination is diluted by
clean water from other portions of the screen.
The inconsistent or irregular TCE concentrations in some of these wells, particularly PEW-4 may
mean only that the well is localed near the fringe of the plume and is not always fully in
the pathway from the source area.
Page 5 Printed: June 6, 2008 (12:01 PM) -DRAFT search for -and resolve issues before final print
C:\Documents and Settings\dmattison\My Documents\Documents\2008\NPL Site Documents\Celanese
Corporation\0804 Conceptual Site Model 0709 DNJ Comments.doc
• •
Regardless of the concentration fluctuations in PEW-4, the data on Figure 5-1 and Table 2-1
shows that by 1994 TCE contamination was at least 62 feet below the land surface at this
well at concentrations of about 1 0Oµg/L.
Whether the TCE contamination at the HH wells is related to the Celanese site is unclear. It is clear that
a TCE plume is present beneath the Celanese site. Concentrations are increasing in some wells and the
plume may be expanding downward beneath the plant site. The extent of contamination caused by this
TCE plume should be defined and appropriate remedial measures should be implemented.
The Executive Summary (p.ES-1) discusses DOWTHERM A in well F-55 and concludes that
contamination is localized and not migrating. However, the existing monitoring wells are relatively
shallow and located laterally around F-55. The DPT holes drilled.near by may not have detected
contamination migrating vertically. The water level contour maps, the topography of the site and the
groundwater model data shown in Appendix D of this report all indicate that groundwater flow beneath the
plant is vertically downward. The shallow monitoring wells are located laterally from the source rather
than below it, and may not detect the contamination.
The Executive Summary (p.ES-1) states that the Inner Tier pumping system probably captured less than
30 percent of the contaminated groundwater in the area. It is unfortunate that this observation was not
made sometime during the 15 years in which the pumping system was in operation. The report uses the
poor performance of the pumping system to conclude " ... that some other mechanism must play a
significant roll in limiting the expansion of the ethylene glycol plume." First, the other mechanisms should
be identified and quantified in a MNA investigation which follows EPA guidelines. But an equally valid
conclusion regarding the extent of contamination is that if the monitoring network around the IT wells was
so bad that 70 percent of the contaminated water was not captured, and no one noticed until at least 4
years after the pumps were off, then the plume may be expanding without being detected because the
monitoring network is as ineffective as the pump & treat system.
Figure 3-4 shows that there are relatively few monitoring wells between the Inner Tier (IT) pumped wells
and the Outer Tier (OT) wells which are more than 500 feet down gradient. The distribution of
contamination in the area under the ponds has not been documented. Previous comments from EPA and
EPA guidelines for optimization of pump and treat systems have noted that monitoring well networks for
pump and treat systems must be different from monitoring well networks for MNA. Once the pumps are
off, groundwater doesn't flow toward the pumping wells. Instead, groundwater flow and contaminant
migration returns to the natural pathways. EPA recommended additional monitoring wells in the IT area
at least as early as July 2004, but the monitoring well network has not been improved.
Figures 5-15 through 5-18 make this point indirectly. The titles say these figures show the extent of
diethylene oxide concentration. The figures include maps and cross-sections, but the concentration data
is not contoured on any of the figures so relationships between source areas, groundwater flow
directions, natural discharge areas and contaminant degradation versus distance are not apparent on
these figures. If the data are insufficient to contour, the data probably are insufficient to conclude that the
monitoring well network is adequate. The titles _of these figures say that they show the extent of
contamination, but they do not. Some analytical results are shown, but the extent of contamination is not
shown. The consultant's interpretation of the line showing where groundwater exceeds the NC2L ARAR
for diethylene oxide should be added to each of these figures.
Map view Figure 5-15 shows diethylene oxide reaches the stream north and east of the facility at
concentrations greater that the NC 2L standard. Primary pathways to the stream are not apparent on the
figure.
Page 6 Printed: June 6, 2008 (12:01 PM) -DRAFT search for -and resolve issues before final print
C:\Documents and Settingsldmattison\My Documents\Documents\2008\NPL Site Documents\Celanese
Corporation\0804 Conceptual Site Model 0709 DNJ Comments.doc
• •
Cross-section Figure 5-16 is 4,600 feet long. The subsurface topography between wells 0-25 and Z-78
suggests that the diethylene oxide concentrations in these wells may have little to do with each other.
Primary pathways between the highest concentrations in well V-65 to well OT-5 and Z-78 are not
apparent on the figure. Groundwater flow directions determined from water level elevation equi-potential
lines are not shown on the figure and can not be inferred from Figure 5-15.
Similar sub-surface topographic features are apparent in Figures 5-17 and 5-18. Figure 5-18 shows that
the diethylene oxide concentration in every monitoring well screen deeper than the top of bedrock
exceeds the NC 2L standard for diethylene oxide. The bedrock is contaminated 200-300 feet below the
land surface. The distribution of diethylene oxide at depth in the bedrock aquifer suggests that pathways
for TCE migration also may be present in the bedrock. A groundwater monitoring program for the
contaminants of concern at this site should be implemented soon.
COMMENT REGARDING SOURCE CONTROL AND CONTAMINANT MIGRATION #1:
The section titled "Contaminant Identification and Extent" (page 3) states:
"The migration pathways that exist currently and are anticipated in the future are:
Leaching of contaminants through the soil to groundwater, and
Potential migration of groundwater to off-site properties."
The persistence of both of these pathways shows that conditions at this site remain in serious conflict with
EPA policy and goals. Continued leaching of contaminants through the soil to groundwater indicates that
source control measures are ineffective. Potential migration of contamination to off-site properties
indicates that the plume is expanding, consequently neither long-term monitoring nor Monitored Natural
Attenuation (MNA) are acceptable remedies for this plume.
The site has been under investigation since 1981 (page 2). Source contiol should have been
accomplished by now and source material should no longer be leaching contamination to groundwater.
Instead, this Site Conceptual Model concludes that a significant potential for off-site migration still exists
even after 3,259 tons of material have been removed from the site (page 2), 39 cubic yards of steam
sediment has been excavated (page 2) and groundwater was pumped from August 1989 until February
2004 (page 3).
EPA policy is bound by the NCP which defines source control as" ... generally considered to include the
construction or installation ... of those actions necessary to prevent the continued "release" of hazardous
substances or pollutants or contaminants into the environment .. ." (NCP p. 168). Continued release of
contamination often results in expansion of a plume. If the plume can still expand, clearly the plume is
not stable and the rate of natural attenuation is less than the rate of contaminant loading to the water
table beneath the source area. Source control measures at this site have been inadequate, inefficient or
in some areas have not been implemented. EPA policy requires effective containment of the plume.
EPA policy regarding MNA specifically requires plumes to be stable or shrinking for MNA to be an
appropriate remedy at that plume.
The 2nd paragraph on page 4 of the conceptual site model report states the reason for improved source
control measures:
"Diethylene oxide is the most widespread and most persistent chemical in the site groundwater. Process
knowledge and its distribution in the groundwater indicate that its presence is generally related to
GRUB disposal in the former GRUB disposal area. Due to its miscibility and persistence in the
environment, its distribution reveals the groundwater migration pathways from the former disposal
area and also indicates that it has the potential to migrate in very low concentrations off site
where the groundwater eventually discharges to receiving streams. "
Persistent contaminants should not be allowed to migrate from the site, contaminating groundwater which
is currently drinkable precisely because they are persistent. The existing Inner Tier pump & treat system
Page 7 Printed: June 6, 2008 (12:01 PM) -DRAFT search for -and resolve issues before final print
C:\Documents and Settings\dmattison\My Documents\Documents\2008\NPL Site Documents\Celanese
Corporation\0804 Conceptual Site Model 0709 DNJ Comments.doc
• •
is in the vicinity of the GRUB disposal area: EPA should consider the need to restart the pump and treat
system, then enhance and optimize the system effectively to capture the plume.
COMMENT REGARDING SOURCE CONTROL AND CONTAMINANT MIGRATION #2:
This report ignores the need for remedial measures in groundwater near monitoring wells F-55, TD-3 and
TD-4. The report states on page 3: " ... localized areas of impact were identified at well F-55 (ethylene
glycol and DOWTHERM A and at wells TD-3/TD-4 (TCE). These are interpreted as localized issues not
related to the source area identified for remediation under OU-1 and OU-2'. In response, EPA must
state:
1.) The monitoring well data are inadequate to defend a conclusion that these source areas are
small and localized. Contamination may be moving vertically downward from these areas.
The vertical extent of contamination has not been determined.
2.) The groundwater contamination near monitoring wells F-55, TD-3 and TD-4 may not be
related to OU-1 and OU-2, but this contamination is on Celanese property and is related to
operations at Celanese. The concept model report should describe what this contamination
is related to, not what it isn't related to.
3.) EPA has not conceded that the TCE source area at TD-3 & 4 can not be related to the TCE
contamination in off-site wells HH48 and HH78. Data in this report show that a TCE plume
has been present beneath the site buildings since before 1994 at concentrations and depths
similar to those in the off-site HH wells.
4.) The sources for contamination in new areas not included in OU-1 or OU-2 have not been
identified or controlled. The extent of contamination has not been determined.
COMMENT REGARDING SOURCE CONTROL AND CONTAMINANT MIGRATION #3:
The section titled "Contaminant Identification and Extent" (page 3) states "The metals are expected to be
oxidized and precipitated as the plumes migrate away from the reducing zones." Expectations are not
sufficient evidence that human health and the environment are protected or that attenuation processes
are occurring according to expectations. The expected contaminant behavior must be documented by
site-specific observations obtained from the groundwater monitoring program. EPA has published
guidance documents regarding MNA for inorganic substances. These guidelines have not been followed.
REGARDING TCE IN WELLS TD-3, TD-4, HH-48, HH-78 and others:
The geologic cross-section shown in Appendix D Figure 6 shows there is a downward vertical hydraulic
gradient at the TD-3 and TD-4 well pair. TD-3 is 45 feet deep and TD-4 is 80 feet deep (Table 2-1 ). Both
wells are located on a ridge top between two streams, where a downward hydraulic gradient should be
expected simply because of the topography of the site.
EVIDENCE FOR INCREASING TCE CONCENTRATIONS: The MCL for TCE is 0.005 mg/L. The NC
ARAR for TCE is 0.0028µg/L. The TCE concentration in the sample collected from TD-3 during 1998 was
1.2 mg/L. TCE is heavier than water, so if a pathway downward is present, vertical migration of TCE
should be anticipated. The presence of a downward hydraulic gradient at these wells would enhance the
potential for vertical migration. Well-TD-4 was sampled in 2006 (Figure 5-1) when the TCE concentration
was determined to be about 2.7mg/L. TCE concentrations in TD-4 are increasing (Figure 5-1 ). The
plume appears to be migrating downward. Neither the source for the TCE contamination nor the extent of
the TCE plume has been determined.
EVIDENCE FOR TCE MIGRATION AND THE ADEQUACY OF THE MONITORING WELL NETWORK:
Appendix D Figure 12 nicely illustrates some aspects of vertical contaminant migration which may be
working at this site. The groundwater model cross-section shown in Appendix D Figure 12 demonstrates
the groundwater flow direction which would result from contamination particles placed at the model's
location for TD-3 does not pass through TD-4. This figure from the groundwater model shows that the
hydraulic gradient probably may not be straight downward, so the center of a contaminant plume from the
Page 8 Printed: June 6, 2008 (12:01 PM) -DRAFT search for -and resolve issues before final print
C:\Documents and Settingsldmattison\My Documents\Documents\2008\NPL Site Documents\Celanese
Corporation\0804 Conceptual Site Model 0709 DNJ Comments.doc
• •
vicinity of TD-3 may not move directly down to TD-4. Instead, well TD-4 in the model appears to be on
l!llr::JIOI
T0-3/TCM
fARTH@T ■CH
A tqco rvtfq,-,AfJON41 ao COl-'NW"
Flgu,. U
Slmulatnd Transport By Partlclo Tracking
from T0-3 and T0-4 -Wast to East Soctlon
CNA Hold1r,,gs, Inc I neon,
ShelOy, NC
In this model, TD-4 is not along the flow path from TD-3. The penciled markings I made on the figure
show that contamination predicted in the model at TD-4 would be less that the concentration present at
the same elevation about 450 feet farther down gradient in the cross-section. This is because the flow
paths in the model are not perfectly, vertically straight down. In this model, TD-4 is not ideally located to
monitor contaminate flow paths from the vicinity of TD-3.
Clearly this model response is different than the history of contamination determined from samples from
well TD-4 shown on Figure 5-1. It is important to note that the model represents a hydrogeologist's best
interpretation of the groundwater flow system at this site based on the information available. The model
provides an excellent illustration of some of the groundwater flow system, but may not accurately portray
specific flow paths to specific well screens.
•I
Appendix D Figure 12 (cross-section) also shows that few of the monitoring wells on-site in the vicinity of
the ponds and the Inner Tier {IT} wells are likely to detect TCE contamination from the TD-3 area is it is
present. The predicted flow paths are deeper than many of the monitoring wells. Appendix D Figure 11
(map view not reproduced here) shows the particle tracks from the model location for TD-3 pass beneath
well M-28, which is located between the 1st two ponds. M-28 is 28 feet deep {Table 2-1 }, but Appendix D
Page 9 Printed: June 6, 2008 (12:01PM) -DRAFT search for-and resolve issues before final print
C:\Documents and Settingsldmattison\My Documents\Documents\2008\NPL Site Documents\Celanese
Corporation\0804 Conceptual Site Model 0709 DNJ Comments.doc
• •
Figure 12 shows the particle tracks from TD-3 are at 11 O feet below the land surface (BLS), at least 80
feet below the bottom of M-28. If the model accurately simulates the real flow paths at the site, then M-28
is much too shallow to detect TCE contamination from the vicinity of TD-3. If this is correct, pathways
from on-site TCE sources may be present which are not detected by the existing monitoring well network.
Appendix D Figure 11 (map view not reproduced here) shows the particle tracks from the model location
for TD-3 pass beneath well M-28. Figure 5-1 shows that TCE has been detected in the T-35 area, which
is about 500 feet south of the particle tracks shown on Appendix D Figure 11. The TCE source has not
been located and may not be precisely at TD-3 as simulated in the model. The flow paths from the real
TCE source may pass closer to well T-35 in the real world than to well M-28 in the model. Again, it is
important to note that the groundwater model presented in Appendix D is a good model, but it is just a
model, not a perfect simulation of the real world.
Four monitoring well clusters on the site have been drilled to the depths of the flow paths shown on
Appendix D Figure 12. The locations of the LL, MM, NN and 00 well clusters are shown on Figure 3-4.
Of these well clusters, only the NN cluster located southeast of the plant buildings are close to a possible
pathway between the TD-3 area and the HH wells. In map view (Appendix D Figure 11 ), the flow paths
shown in Appendix Figure 12 are about 400 feet north of T-35 (Figure 5-1) where low-level detections of
TCE have been observed. Table 5-4 does not show sample results for wells in the NN well cluster.
These wells do not appear to have been sampled in 2006. Table 5-4 does show wells TD-3 and TD-4
were the only wells with samples analyzed for TCE in 2006. The existing monitoring program is not
adequate. There is not enough data to evaluate the extent of TCE in groundwater beneath this site and
to compare the extent of TCE contamination with the predictions of the groundwater model.
EVIDENCE FOR THE EFFECTIVENESS OF NATURAL ATTENUATION: Appendix D Table 1 shows the
concentration of cis-DCE was 0.067 mg/L in TD-3 in the 1998 sample. DCE is a degradation product of
TCE, so some degradation is occurring. But DCE concentration was 2 orders of magnitude less than the
TCE concentration suggesting that conditions may not be very favorable for reductive dechlorination of
TCE. Vinyl chloride is the next degradation product of TCE after DCE. Appendix D Table 1 shows that
vinyl chloride was non-detected in any of the four samples from TD-3 and TD-4. The absence of vinyl
chloride also suggests that conditions may not be favorable for reductive dechlorination of TCE.
The reductive dechlorination of TCE is normally expected to be an anaerobic process. Dissolved oxygen
(DO) data is presented graphically in Appendix C. Unfortunately, all of the wells which were monitored
are down gradient from TD-3, so conditions closer to the TCE source area have not been defined.
Clearly the monitoring wells were selected to monitor the wells of the Inner Tier area and not the TCE
plume. Well F-55 is the closest well to TD-3 for which DO data was collected. The graph in Appendix C
shows that groundwater in well F-55 is typically aerobic, not anaerobic. The reductive dechlorination of
TCE is normally expected not to be effective under aerobic conditions.
Dissolved oxygen is sometimes a difficult parameter to measure in the field. The results are easily
contaminated by atmospheric oxygen. Consequently, the Oxidation/Reduction Potential (ORP) may be a
better indicator of the potential for effective reductive dechlorination of TCE. The optimal ORP range for
reductive dechlorination is between -200 and -350 milliVolls (Figure B3-3, p.B3-6 in Wiedemeier, T.H .. M.A.
Swanson, D.E. Moutoux, E.K. Gordon, J.T. Wilson, B.H. Wilson, O.H. Kampbell, J.E. Hansen, P. Haas, 1996, Technical Protocol
for Evaluating Natural Attenuation of Chlorinated Solvents in Groundwater, USAF Center for Environmental Excellence, Brooks
AFB, San Antonio, TX, November, 1996.). The second graph of ORP results in Appendix Con the Concept
Model report shows that the ORP levels in groundwater from wells at this site are rarely in the optimal
range for the reductive dechlorination of TCE. The graph of ORP trends versus lime in Appendix C
suggests that conditions throughout most of this aquifer are not favorable for reductive dechlorination of
CVOCs.
Page 10 Printed: June 6, 2008 (12:01 PM) -DRAFT search for -and resolve issues before final print
C:\Documents and Settings\dmattison\My Documents\Documents\2008\NPL Site Documents\Celanese
Corporation\0804 Conceptual Site Model 0709 DNJ Comments.doc
•
COMMENT:
Please correct the units on the Y axis in the graph of ORP trends versus time in Appendix C. This
parameter typically is measured in millivolts (mV), not mg/L. An ORP value in negative mg/L can not be
correct.
COMMENT:
Please correct the legend for the top graph in Appendix D Figure 5-18. The 2nd "TD-3 Measured: Linear''
trend line probably should be labeled as "TD-4 Measured: Linear''.
COMMENT REGARDING THE USE OF MODELS VERSUS OBSERVATIONS:
Appendix D Figure 5-18 is a typical graphical presentation of the effectiveness of a groundwater model
calibration. The lower graph in this figure suggests that a reasonable correlation between the model
results and the TCE values observed in PEW-1 is been obtained in the model. There are some spikes in
the sample results which are not predicted by the model, but overall, the model fits the observed data
reasonably well during the period of record.
However, the upper graph on Appendix D Figure 5-18 is distinctly different at well TD-4. The TCE
concentration trend in TD-4 shows a rapid rise in TCE concentrations between 2001 and 2006, but the
well has been sampled only 3 times. The last sample result showed the TCE concentration to be
2,700µg/L. The curve labeled TD-4 Simulated in the upper graph of Figure 5-18 shows the model
prediction of future concentrations at this well. This prediction is consistent with the assumptions built into
the model, but no samples have been collected to show that the concentrations in the well are doing what
the model predicts they should do. This is an example of why the EPA guidelines for MNA specify that
predictions of contaminant migration and cleanup times should be based on field observations, not solely
on model results and calculations. A groundwater monitoring program which follows EPA guidelines
should be implemented at this site.
RECOMMENDATION FOR FUTHER INVESTIGATION: The extent of TCE contamination near TD-3
should be investigated. Vertical hydraulic gradients in this area should be determined. The potential for
vertical migration of TCE near TD-3 should be investigated. The potential for a relationship between the
release of ethylene glycol detected in F-55, a short distance away from TD-3, should be investigated.
The EPA should be able to show the extent of contamination, show incomplete pathways and show that
the chemistry of the plume is understood before making a determining who is not responsible for
contamination observed off-site and down gradient from a potential source.
One important element in addressing the off-site TCE is the identification of the source area for the off-
site TCE plume. If the source isn't the Celanese property, who released this contamination, where and
how? This site is in a relatively rural, agricultural area, with no major industrial land-uses other than
Celanese. The area does not appear to have been the target of illegal "midnight dumping". There appear
to be only 2 properties and 3 owners (Celanese, Mr. Elliot and the current property owner Mr. Causby).
Before EPA clears any property owner of responsibility, it seems that direct or indirect proof of at least
one or more of the following should be provided:
1.) that there is no complete pathway between TCE on the Celanese property and TCE in the off
site wells;
2.) that there is evidence of another likely source of TCE contamination not on Celanese
property;
Page 11 Printed: June 6, 2008 (12:01 PM) -DRAFT search for -and resolve issues before final print
C:\Documents and Settings\dmattison\My Documents\Documents\2008\NPL Site Documents\Celanese
Corporation\0804 Conceptual Site Model 0709 DNJ Comments.doc
•
3.) that the contaminant chemistry off-site is distinctly different and therefore not related to the
contaminant chemistry on the Celanese property.
Portions of these lines of evidence can be accomplished during delineation of the extent of TCE
contamination of the TCE plume on Celanese property. Portion so of these lines of evidence can be
accomplished through establishment of a monitoring program on Celanese property plus the HH wells.
These activities should be implemented soon.
Page 12 Printed: June 6, 2008 (12:01PM) -DRAFT search for-and resolve issues before final print
C:\Documents and Settings\dmattison\My Documents\Documents\2008\NPL Site Documents\Celanese
Corporation\0804 Conceptual Site Model 0709 DNJ Comments.doc
• .=&;;;:;.:'iih •
NCDENR
North Carolina Department of Environment and Natural Resources ,,,.\
Dexter R. Matthews, Director Division of Waste Management
Ms. Beth Walden ·
Remedial Project Manager
December 5, 2007
Superfund Remedial & Site Evaluation Branch
U. S. Environmental Protection Agency, Region 4
Sam Nunn -Atlanta Federal Center ·
61 Forsyth Street, S.W.
Atlanta, GA 30303
RE: Conceptual Site Model Report
Celanese Corporation NPL Site
Shelby, Cleveland County, NC
Dear Ms. Walden:
Michael F. Easley: Governor
William G. Ross Ji., Secretary
The North Carolina Department of Environment and Natural Resources (NC DENR) Superfund
Section has received the Conceptual Site Model Report for the Celanese Corporation National
Priorities List (NPL) Site. The Superfund Section has reviewed this document and offers the
following attached comments.
The NC DENR Superfund Section appreciates the opportunity to comment on this document. If
you have any questions or comments, please feel free to contact me at [phone].
Sincerely, ·
vYi~
David B. Mattison
Environmental Engineer
NC DENR Superfund Section
Attachment
Cc: David Lilley, NC DENR Superfund Section
1646 Mail Service Center, Raleigh, North Carolina 27699-1646
Phone [phone] I FAX [phone]\ Internet http://wastenotnc.org
An Equal Opportunity I Affirmative Action Employer-Printed on Dual Purpose Recycled Paper
•
Ms. Beth Walden
Conceptual Site Model Report
Celanese Corporation NPL Site
December 5, 2007
Page I
Conceptual Site Model Report
CELANESE CORPORATION NPL SITE
General
•
1. Pages 5, 38, and 47: It is stated on these pages that the risk assessments done on this site
showed no human health or ecological concerns. In comments dated September 19,
2007, NC DENR came to the conclusion that both risk assessments were in need of major
revisions and should not be used in their current form to make any decisions, Please
_make the necessary modifications prior to continuing with this process.
2. The Conceptual Site Model Report (Report) indicates that the diethylene oxide
contaminated groundwater plume has migrated off the site property boundaries. The
Report further indicates that a strong destructive degradation mechanism for di ethylene
oxide does not appear to be present at this site, and the attenuation of di ethylene oxide
largely relies on the physical mechanisms such as dilution, advection, and dispersion.
Lastly, the Report indicates that the diethylene oxide plume is expected to discharge to
surface water where the diethylene oxide can potentially photodegrade. Section 7.2 of the
Report provides the US EPA criteria for evaluating monitored natural attenuation (MNA)
and its effectiveness as a remedial technology. The NC DENR likewise has rules (North
Carolina Administrative Code (NCAC) Title 15A, Chapter 2L, Section .0106) regarding
the use of a MNA remedial alternative. In accordance with these US EPA criteria, the
NC DENR's criteria and the aforementioned statements from the Report, MNA does not
appear to be a viable remedial technology for diethylene oxide. For MNA to be the
accepted remedy for the Site, all contaminants of concern must be addressed: Please
revise the Conceptual Site Model Report accordingly.
3. If this document is intended to act as a supplemental feasibility study, it should be
renamed and revised accordingly. This task will require completion before the issuance
of any Amendment to the Record of Decision (ROD) can be made. In accordance with
the National Contingency Plan (NCP), the relative performance of each alternative should
be evaluated using the nine criteria (40 CFR § 300.430(e)(9)(iii)) as a basis for
comparison. The purpose of the evaluation process is to determine which alternative: (a)
meets the threshold criteria of overall protection of human health and the environment
and attainment of Applicable or Relevant and Appropriate Requirements (ARARs), (b)
provides the "best balance" with respect to the five balancing criteria of 40 CFR §
300.430(e)(9)(iii)(C)-(G), and (c) takes into consideration the acceptance of the support
agency (here, the NC DENR) and the community. Please revise Section 7 and Section 8
of the Conceptual Site Model Report accordingly.
•
Ms. Beth Walden
Conceptual Site Model Report
Celanese Corporation NPL Site
December 5, 2007
Page 2
Table of Contents
•
4. Please correct the Table of Contents to indicate that the title of Section 5.0 is
"Distribution, Migration and Fate of Site Contaminants".
Executive Summary
5. Please revise the last sentence of the second paragraph of this section in accordance with
the NC DENR comments submitted September 19, 2007 on the Draft Risk Assessment.
6. Please revise the first sentence of the last paragraph of this section in accordance with the
NC DENR comments submitted September 19; 2007 on the Draft Risk Assessment.
Section 2.0 Conceptual Site Model
Hydrogeology
7. Please revise the last sentence of this section to reference Section 4.2 of the Conceptual
Site Model Report.
Public Health and Ecological RiskConsiderations
8. Please revise this section in accordance with the NC DENR comments submitted
September 19, 2007 on the Draft Risk Assessment.
Summary
9. Please revise the last bullet item in this section in accordance with the NC DENR
· comments submitted September 19, 2007 on the Draft Risk Assessment.
Section 4.4 Site Geochemistry
.I!!!
I 0. · Please correct the first sentence of this section to state "The pH data show that pH values
were frequently less than 5 at wells IT-6, N-29, U-38, and V-23."
Section 5.0 Distribution, Migration and Fate of Site Contaminants
11. Please correct Section 5.0 to indicate that the title is "Distribution, Migration and Fate of
Site Contaminants".
Ms. Beth Walden
Conceptual Site Model Report
Celanese Corporation NPL Site
December 5, 2007
Page 3
•
Section 5.2 Ethylene Glycol Distribution, Migration and Fate
12. Please correct the last sentence of the fifth paragraph of Section 5.2 to. state "This well
has shown some detections of ethylene giycol in the range of 13 mg/L to 22 mg/L. .. "
Section 6.0 Site Risk
13. Please revise Section 6.0 and its subsections in accordance with the NC DENR comments
submitted September 19, 2007 on the Draft Risk Assessment.
Section 7.1 Groundwater Recovery and Treatment
14. Please revise the last sentence of the fourth paragraph of Section 7.1 in accordance with
the NC DENR comments submitted September 19, 2007 on the Draft Risk Assessment.
Section 7.2 Natural Attenuation Effectiveness
15. Please revise the last sentence of the third paragraph of Section 7.2 in accordance with the
NC DENR comments submitted September 19, 2007 on the Draft Risk Assessment.
16. Please revise the last sentence of the sixth paragraph of Section 7.2 in accordance with
the NC DENR comments submitted September 19, 2007 on the Draft Risk Assessment.
Section 8.0 Summary of Conclusions and Recommendations
17. Please revise the first and thirteenth bullet items in Section 8.0 in accordance with the NC
DENR comments submitted September 19, 2007 on the Draft Risk Assessment.
Table 3-2 Site Chronology
18. Please correct the entry in Table 3-2 for October 21, 1988 to state "Consent decree for
OUl".
•
UNITED STATES ENVIRONMENTAL PROTECTION AGENCY REGION 4
61 Forsyth Street
Atlanta, Georgia 30303-3104
4WD-TSS
SUBJECT:
MEMORANDUM August 9, 2007
Status Update and Conceptual Path Forward dated 3JUL07
FROM: David N. Jenkins, Environmental Scientist
Technical Support Section, Superfund Division
THROUTH: Scott Sudweeks, Chief, Technical Support Section,
TO:
Superfund Division
Beth Walden, Remedial Project Manager
Beth,
I have read the letter regarding the Status Update and Conceptual Path Forward for the Celanese site
dated July 3, 2007 as you requested. Here my comments. Please call me at 404-562-8462 if you have
any questions.
The document reviewed is titled:
EarthTech, 2007, Status Update and Conceptual Path Forward, CNA Holdings/TICONA, Celanese
Fiber Operations, Shelby, NC, EarthTech, Inc., 1455 Old Alabama Road, Suite 170, Roswell,
Georgia 30076, Earth Tech letter report from Everett W. Glover, Jr. to EPAR4 RPM Beth Walden,
dated July 3, 2007.
GENERAL COMMENT:
The intent of this letter appears to be accomplishment of two goals:
1.) Get approval for the Site-Wide Monitoring program;
2.) Prevent re-starting the Inner Tier Well contaminant recovery system.
The monitoring proposal appears to be acceptable and appropriate as applied to the existing wells. The
monitoring well network was designed to monitor the distribution of contamination under pumping
conditions. The monitoring well network was not designed to monitor contaminant migration under non-
pumping conditions. There are very few wells down gradient from the Inner Tier wells, but close enough
to detect the migration of contamination since the pumps were turned off. Additional monitoring points
should be installed.
The monitoring wells primarily are located either near the Inner Tier pumping wells, or are the Outer Tier
wells themselves. The Outer Tier wells have not been pumped in many years, and these are located
hundreds of feet from the Inner Tier wells. In short, the existing monitoring well network does not provide
data which can support the claim that MNA is as effective at containing the plume as pumping the Inner
Tier Wells would be. This claim is made in the 3rd paragraph on page 1 of the Earth Tech letter: "Based
on the data produced during the demonstration period, natural attenuation has been shown to be as
effective as continued operation of the Inner Tier (IT) extraction system for remediation of ethylene glycol
associated with the former source area."
Two points in the memo demonstrate that the reasons offered in this memo for not restarting the Inner
Tier Pumping ~stem are not valid:
The 2 bullet at the top of page 2 refers to a need for" ... evaluating the need for additional hot-
spot remediation in the former source area, ... "
The 2nd last paragraph on page 2 states that "Contaminant concentrations have remained fairly
stable for the last few years in the area immediately downgradient of the waste disposal
area remediated in 1992 and 1993 as part of Operable unit (OU) 2." The letter does not
say that the ethylene glycol concentrations the concentrations in monitoring well V23
have been fairly stable at a level approximately 650 times greater than the North Carolina
ARAR.
Page 1 June 6, 2008 (1 :01 PM) C:\Documents and Settingsldmattison\My Documents\Documents\2008\NPL Site
Documents\Celanese Corporation\Status Update and Conceptual Path Forward OJ Comments.doc
• •
The North Carolina 2L requirements as excerpted in the EarthTech letter from NCAC Title 15A
Subchapter 2L.0106(I) state that all sources of contamination and free product should have been
removed or controlled. Stable contaminant concentrations ai the observed levels over a period of more
than 3 years indicate that a source of groundwater contamination remains on site. Statements in the
EarthTech letter regarding the effectiveness of MNA are premature while effective source control
measures have not been implemented. This letter states in the 3"' paragraph that MNA has been shown
to be as effective as continued operation of the Inner Tier extraction system for remediation of ethylene
glycol. The letter ignores the demonstration presented in my June 4th 2007 memo showing that 300 years
from now, the ethylene glycol concentration in well V23 may still be 360 times greater than the North
Carolina ARAR. A remedy which requires centuries to return groundwater to unrestricted beneficial use
does not constitute a demonstration that Natural Attenuation is working, effective or adequate as a
remedial measure.
RECOMMENDATIONS:
We don't know where the down gradient edge of lhe plume is. It may be stable or it may be moving
toward the Outer Tier wells, re-contaminating portions of the aquifer which had been cleaned by previous
pumping of the Inner Tier and Outer Tier wells. I recommend that the Inner Tier pumping system be
restarted until effective source control measures are implemented and there is reason to believe that
MNA will be effective in returning groundwater to unreslricled beneficial use in a reasonable period of
time. A 300 year clean up time probably is not reasonable from anyone's point of view. However,
implementation of this recommendation could be delayed for a while because the groundwater
contamination appears to be inside the potential influence of the Outer Tier wells. These wells could be
refurbished if necessary and pumping could begin again to contain contaminant migration. Unfortunately,
the existing monitoring well network within the Outer Tier of wells is too sparse and is inadequate for MNA
purposes. This is a common problem: a monitoring well network for a pump & treat system will be
different from a monitoring well network for an MNA remedy because different forces are working on the
plume.
If re-starting the pumps is delayed, the delay should be short and should be on the condition that the
monitoring well network will be supplemented by installation of additional monitoring wells or direct-push
holes which can be sampled. These points should be localed down gradient from the most contaminated
Inner Tier wells and should create a series of monitoring points along the same flow paths from the
contamination to the natural discharge area at the local surface water streams.
The issue of getting a series of monitoring points along the same flow path is critical. This is the basis for
degradation versus distance and degradation versus time calculations. Water passing through points A,
B, & C, etc., must be the same water on the same flow path or the calculations are not valid.
Groundwater flow directions must be evaluated based on water levels in the monitoring wells and the
observed distribution of contamination. This evaluation must consider potential mounding on the water
table caused by leakage from the unlined ponds. This request has been made numerous times in the
pasl.
Regarding the groundwater monitoring plan, Table 1 proposes a 3-year Quarterly sampling schedule.
think a more flexible plan and schedule would be appropriate. Given the length of the existing sampling
record, one year of quarterly sampling data might be beneficial. If seasonal variations prove to be
relatively limited, one year of quarterly sampling may be sufficient. But as described above, additional
sampling points should be installed. Quarterly sampling should focus primarily on contaminated wells
and supplemental monitoring points on flow paths down gradient from these wells.
The sampling plan should be flexible in that more distant sampling points which are uncontaminated may
be sampled less frequently. Initially, uncontaminated wells not believed to be on flow paths from
contaminated areas may be sampled annually. Later, uncontaminated wells with a long sampling history
Page 2 June 6, 2008 (1 :01 PM) C:\Documents and Settingsldmattison\My Documents\Documents\2008\NPL Site
Documents\Celanese Corporation\Status Update and Conceptual Path Forward DJ Comments.doc
f •
may be sampled less frequently. But water levels should be collected at all wells during all sample events
so groundwater flow directions can be defined each time a sample is collected. This site has been under
investigation for many years. Annual sampling at a larger number of wells may be more informative than
the extensive quarterly sampling proposed in Table 1, especially at wells not believed to be on
groundwater flow paths from contaminated areas.
Page 3 June 6, 2008 (1 :01 PM) C:\Documents and Settingsldmattison\My Documents\Documents\2008\NPL Site
Documents\Celanese Corporation\Status Update and Conceptual Path Forward DJ Comments.doc
• •
UNITED STATES ENVIRONMENTAL PROTECTION AGENCY REGION 4
61 Forsyth Street
Atlanta, Georgia 30303-3104
4WD-TSS
SUBJECT:
MEMORANDUM August 9, 2007
Status Update and Conceptual Path Forward dated 3JUL07
FROM: David N. Jenkins, Environmental Scientist
Technical Support Section, Superfund Division
THROUTH: Scott Sudweeks, Chief, Technical Support Section,
TO:
Superfund Division
Beth Walden, Remedial Project Manager
Beth,
I have read the letter regarding the Status Update and Conceptual Path Forward for the Celanese site
dated July 3, 2007 as you requested. Here my comments. Please call me at 404-562-8462 if you have
any questions.
The document reviewed is titled:
Earth Tech, 2007, Status Update and Conceptual Path Forward, CNA Holdings/TICONA, Celanese
Fiber Operations, Shelby, NC, EarthTech, Inc., 1455 Old Alabama Road, Suite 170, Roswell,
Georgia 30076, Earth Tech letter report from Everett W. Glover, Jr. to EPAR4 RPM Beth Walden,
dated July 3, 2007.
GENERAL COMMENT:
The intent of this letter appears to be accomplishment of two goals:
1.) Get approval for the Site-Wide Monitoring program;
2.) Prevent re-starting the Inner Tier Well contaminant recovery system.
The monitoring proposal appears to be acceptable and appropriate as applied to the existing wells. The
monitoring well network was designed to monitor the distribution of contamination under pumping
conditions. The monitoring well network was not designed to monitor contaminant migration under non-
pumping conditions. There are very few wells down gradient from the Inner Tier wells, but close enough
to detect the migration of contamination since the pumps were turned off. Additional monitoring points
should be installed.
The monitoring wells primarily are located either near the Inner Tier pumping wells, or are the Outer Tier
wells themselves. The Outer Tier wells have not been pumped in many years, and these are located
hundreds of feet from the Inner Tier wells. In short, the existing monitoring well network does not provide
data which can support the claim that MNA is as effective at containing the plume as pumping the Inner
Tier Wells would be. This claim is made in the 3'• paragraph on page 1 of the Earth Tech letter: "Based
on the data produced during the demonstration period, natural attenuation has been shown to be as
effective as continued operation of the Inner Tier (IT) extraction system for remediation of ethylene glycol
associated with the former source area."
Two points in the memo demonstrate that the reasons offered in this memo for not restarting the Inner
Tier Pumping sc7stem are not valid:
The 2" bullet at the top of page 2 refers to a need for " ... evaluating the need for additional hot-
spot remediation in the former source area, ... "
The 2nd last paragraph on page 2 states that "Contaminant concentrations have remained fairly
stable for the last few years in the area immediately downgradient of the waste disposal
area remediated in 1992 and 1993 as part of Operable unit (OU) 2." The letter does not
say that the ethylene glycol concentrations the concentrations in monitoring well V23
have been fairly stable at a level approximately 650 times greater than the North Carolina
ARAR. .
Page 1 August 27, 2007 (9:13AM) C:\Documents and Settingsldmattison\My Documents\Documents\2006-
2007\NPL Site Documents\Celanese Corporation\Status Update and Conceptual Path Forward DJ Comments.doc
r • •
The North Carolina 2L requirements as excerpted in the EarthTech letter from NCAC Title 15A
Subchapter 2L.0106(I) state that all sources of contamination and free product should have been
removed or controlled. Stable contaminant concentrations at the observed levels over a period of more
than 3 years indicate that a source of groundwater contamination remains on site. Statements in the
EarthTech letter regarding the effectiveness of MNA are premature while effective source control
measures have not been implemented. This letter states.in the 3"' paragraph that MNA has been shown
to be as effective as conti_nued operation of the Inner Tier extraction system for remediation of ethylene
glycol. The letter ignores the demonstration presented in my June 4th 2007 memo showing that 300 years
from now, the ethylene glycol concentration in well V23 may still be 360 times greater than the North
Carolina ARAR. A remedy which requires centuries to return groundwater to unrestricted beneficial use
does not constitute a demonstration that Natural Attenuation is working, effective or adequate as a
remedial measure.
RECOMMENDATIONS:
We don't know where the down gradient edge of the plume is. It may be stable or it may be moving
toward the Outer Tier wells, re-contaminating portions of the aquifer which had been cleaned by previous
pumping of the Inner Tier and Outer Tier wells. I recommend that the Inner Tier pumping system be
restarted until effective source control measures are implemented and there is reason to believe that
MNA will be effective in returning groundwater to unrestricted beneficial use in a reasonable period of
time. A 300 year clean up time probably is not reasonable from anyone's point of view. However,
implementation of this recommendation could be delayed for a while because the groundwater
contamination appears to be inside the potential influence of the Outer Tier wells. These wells could be
refurbished if necessary and pumping could begin again to contain contaminant migration. Unfortunately,
the existing monitoring well network within the Outer Tier of wells is too sparse and is inadequate for MNA
purposes. This is a common problem: a monitoring well network for a pump & treat system will be
different from a monitoring well network for an MNA remedy because different forces are working on the
plume.
If re-starting the pumps is delayed, the delay should be short and should be on the condition that the
monitoring well network will be supplemented by installation of additional monitoring wells or direct-push
holes which can be sampled. These points should be located down gradient from the most contaminated
Inner Tier wells and should create a series of monitoring points along the same flow paths from the
contamination to the natural discharge area at the local surface water streams.
The issue of getting a series of monitoring points along the same flow path is critical. This is the basis for
degradation versus distance and degradation versus time calculations. Water passing through points A,
8, & C, etc., must be the same water on the same flow path or the calculations are not valid.
Groundwater flow directions must be evaluated based on water levels in the monitoring wells and the
observed distribution of contamination. This evaluation must consider potential mounding on the water
table caused by leakage from the unlined ponds. This request has been made numerous times in the
past.
Regarding the groundwater monitoring plan, Table 1 proposes a 3-year Quarterly sampling schedule.
think a more flexible plan and schedule would be appropriate. Given the length of the existing sampling
record, one year of quarterly sampling data might be beneficial. If seasonal variations prove to be
relatively limited, one year of quarterly sampling may be sufficient. But as described above, additional
sampling points should be installed. Quarterly sampling should focus primarily on contaminated wells
and supplemental monitoring points on flow paths down gradient from these wells.
The sampling plan should be flexible in that more distant sampling points which are uncontaminated may
be sampled less frequently. Initially, uncontaminated wells not believed to be on flow paths from
contaminated areas may be sampled annually. Later, uncontaminated wells with a long sampling history
Page 2 August 27, 2007 (9:13AM) C:\Documents and Settingsldmattison\My Documents\Documents\2006-
2007\NPL Site Documents\Celanese Corporation\Status Update and Conceptual Path Forward DJ Comments.doc
r • •
may be sampled less frequently. But water levels should be collected at all wells during all sample events
so groundwater flow directions can be defined each time a sample is collected. This site has been under
investigation for many years. Annual sampling at a larger number of wells may be more informative than
the extensive quarterly sampling proposed in Table 1, especially at wells not believed to be on
groundwater flow paths from contaminated areas.
Page 3 August 27, 2007 (9:13AM) C:\Documents and Settingsldmattison\My Documents\Documents\2006-
2007\NPL Site Documents\Celanese Corporation\Status Update and Conceptual Path Foiward DJ Comments.doc
@EarthTech
A "tf./CD International Ltd. Company
July 3, 2007
Ms. Beth Walden
U.S. EPA Region 4
Atlanta Federal Center
61 Forsyth Street
•
Atlanta, Georgia 30303-3104
Subject: Status Update and Conceptual Path Forward
1455 Old Alabama Road P 770.990.1400
Suite 170 F 770.649.8721
Roswell, GA 30076 earthtech.com
~~~L ~;,:,@1~
SUPERHJND SECTION
CNA Holdings, Inc. (F.K.A. Celanese Fibers Operations) Facility
Shelby, Cleveland County, North Carolina
NCD003446721
Consent Decree, Civil Action No. SH-C
Earth Tech Project No. 79750
Document Control Number -026SB-252
Dear Ms. Walden:
The purpose of this letter is to present a status of the project and a summary of the proposed plan
forward for the Celanese Fibers Operations site in Shelby, North Carolina. Details of this plan will
be provided as described in the subsequent paragraphs.
We are in the process of developing a Site Strategy Report. This strategy report will present a plan
forward based on the results of the monitored natural attenuation (MNA) demonstration, as well as
other data collected and assessments completed. The assessments include development and
updating of the conceptual site model, the groundwater flow and transport model, and the
site-specific risk assessment. The report will include an overall strategy for the site and site
chemicals of concern. We plan to submit this report in the third quarter of 2007. The groundwater
modeling report will be included as an appendix of this report. The site-specific risk assessment
will be submitted as an independent deliverable prior to the submission of the Site Strategy report,
and relevant information and conclusions from the risk assessment will be included in the Site
Strategy Report as appropriate.
The MNA demonstration period has been in effect for approximately 3 years. Based on the data
produced during the demonstration period, natural attenuation has shown to be as effective as
continued operation of the Inner Tier (IT) extraction system for remediation of ethylene glycol
associated with the former source area. The Site Strategy Report will present this conclusion and
recommend that the IT extraction system not be reactivated. This determination will conclude the
MNA demonstration.
The site risk assessment concludes that there are no unacceptable risks to human health or the
environment. The groundwater transport model focused on ethylene glycol and diethylene oxide
and provided input for the risk assessment. The modeling indicates that the plumes for site-related
parameters have generally stabilized and that additional pumping would have little or no positive
influence on the life of the plume.
@EarthTech
A 1:qca International ltd. Company
• I) Status Update and Path Forward
CNA Holdings, Inc. (F.K.A. Celanese Fibers Operations) Facility
Shelby, Cleveland County, North Carolina
NCD003446721, Consent Decree, Civil Action No. SH-C
Document Control Number -026SB-252
Pa e 2
The updated conceptual site model, technical analyses, and data collection described in the
preceding paragraphs will provide the basis for developing the path forward. Conceptually, the
path forward will focus on:
• completing a site-wide monitoring program that is relevant to the status of the site today
and that provides a statistically valid basis for developing the longer-term site-wide
monitoring program,
• evaluating the need for additional hot-spot remediation in the former source area, and
• compiling the technical information for modifying the current Record of Decision (ROD) to
reflect the improved conditions at the site and advancements in remediation technologies
that have occurred since the current ROD was developed in 1988.
The first proposal for this strategy is a site-wide monitoring program relevant to the current state of
knowledge and status of the site. A subset of this program is the collection of a statistically
significant population of groundwater quality data capable of defining changes in groundwater
quality with respect to both location and time as suggested in our meeting of February 1, 2007.
The proposed sampling plan is attached. The first table presents the quarterly sampling schedule
which will be implemented for a period of three years. This plan focuses on the primary site
parameters {diethylene oxide, ethylene glycol, Dow Therm A™, and TCE). The collection of these
parameters will provide the desired statistical dataset. The second table presents the expanded
sampling program which will be completed annually. This annual plan includes the complete
quarterly sampling with the addition of analysis for volatile organic compounds (VOCs) and natural
attenuation indicator parameters at select locations. The annual addition of these parameters will
provide updates to the site characterization and continued confirmation of the results of the MNA
demonstration period. Lastly, the third table presents the annual sampling plan with the addition of
analysis for selected metals, phenols and chlordan_e. These parameters are identified as site
chemicals of concern (COC) but are not considered primary site parameters based on frequency
and/or magnitude of detections. This expanded sampling plan will replace the annual plan one
time during the three year period with the intention of repeating this list every five years or until the
site COC list is revised. All samples will also be analyzed for routine field parameters including
groundwater elevation, pH, ORP, dissolved oxygen, specific conductance, temperature and
turbidity. The proposed sampling locations are presented in the attached figure.
Contaminant concentrations have remained fairly stable for the last few years in the area
immediately downgradient of the waste disposal area remediated in 1992 and 1993 as part of
Operable Unit (OU) 2. Since the OU-2 objective was waste removal and not clean closure, some
residual impact to the soils between the base of the waste and the water table was expected. It is
interpreted that this residual impact to the soil is resulting in the relative stability seen today in the
groundwater contaminant levels in wells immediately downgradient of the former source area. A
focused feasibility study work plan will be submitted to evaluate whether further remedial activity is
warranted to address the residual contamination in the former source area and whether additional
active remediation would significantly shorten the projected length of time for the site to attain the
ARARs.
The 3 years of data collected during the MNA demonstration project have shown that the ethylene
glycol plume has remained relatively unchanged during the time that the extraction system has
been shutdown and that the natural attenuation processes continue to degrade the ethylene glycol
and its degradation products. Furthermore, the risk.assessment has shown that the site does not
2
® EarthTech
A 1:f./CD International Ltd. Company
• • Status Update and Path Forward
CNA Holdings, Inc. (F.K.A. Celanese Fibers Operations) Facility
Shelby, Cleveland County, North Carolina
NCD003446721, Consent Decree, Civil Action No. SH-C
Document Control Number -026SB-252
Pa e 3
present an unacceptable risk as it currently exists. Because of the changes in the subsurface
conditions and geochemistry induced by pumping, reactivation of the IT system is expected to
adversely impact the data collection in the proposed sampling plan. Therefore, reactivation of the
IT system would not provide any significant improvement in the risk associated with the site, but
would delay the ability to complete characterization of the site and develop an appropriate
long-term response. Consequently, it is recommended that the IT extraction system remain
inactive, and it is requested that EPA approve this recommendation. A recommendation for a
revised remedial approach will be the primary deliverable of the work to be proposed and will
address the potential need for additional active remediation at the former source area. The
culmination of these analyses and data collection will be a recommendation for modifying the ROD
to make it relevant to the current site situation.
We are requesting that the proposed sampling plan be approved at this time so that it may be
initiated in time for the data assessment to be included in the next 5-year review. The findings of
this study, along with the results of the former source area focused feasibility study and resulting
actions will be included in the information supporting the recommendation for a modification to the
existing ROD.
We look forward to your responses on the proposed path forward. Please contact us if you need
further information.
Sincerely,
Earth ch, Inc.
B on ahlgren
Project Engineer
~~w.~qi-
Everett W. Glover, Jr. PE
Project Manager
cc: Dave Mattison, NCDENR ./
PEM Carter, Ticona Shelby
Steven Simpson, Celanese Representative
file
3
•
Location Ethylene Glycol
AA-54 X
B-34
C-49 X
CC-33 X
CC-64
DD-58R
EE-58 X
F-55 X
FF-34 X
FF-62 X
G-50
GG-61 X
H-59
HH-48
HH-77
1-57
11-65
IT-1 X
IT-5 X
IT-6 X
IT-BR X
IT-9 X
J-29 X
J-59
K-28 X
K-58
KK-55
N-29 X
NN-105
0-25 X
OT-1R
OT-2R
OT-6
OT-6A
OT-7
P-58 X
PEW-1 X
PEW-3 X
PEW-4 X
Q-33 X
R-42
SW-4
SW-7
T-17
T-35 X
T-58
TD-3 X
TD-4 X
Tl-1 X
Tl-2
U-38 X
V-23 X
V-65 X
W-23 X
X-32 X
Table 1
Proposed Three Year Quarterly Sampling Plan
Ticona Shelby Site
Earth Tech Project 49750
TOC Alkalinity DowTherm A Diethelyene Oxide
X X X
X
X X X X
X X X
X
X
X X X
X X X X
X X X
X X X
X
X X X
X
X
X
X X X
X X X
X X X
X X X
X X X
X X X X
X
X X X X
X
X
X X X X
X
X X X X
X
X
X
X
X
X X X
X X X X
X X X X
X X X X
X X X
X
X
X
X
X X X X
X
X X X X
X X X
X X X X
X
X X X X
X X X X
X X X X
X X X X
X X X
TCE
X
X
X
X
X
X
X
X
X
X
X
X
X
Location Ethylene Glycol
AA-54 X
8-34
C-49 X
CC-33 X
CC-64
DD-58R .
EE-58 X
F-55 X
FF-34 X
FF-62 X
G-50
GG-61 X
H-59
HH-48
HH-77
1-57
11-65
IT-1 X
IT-5 X
IT-6 X
IT-8R X
IT-9 X
J-29 X
J-59
K-28 X
K-58
KK-55
N-29 ... X
NN-105
0-25 X
OT-1R
OT-2R
OT-6
OT-6A
OT-7
P-58 X
PEW-1 X
PEW-3 X
PEW-4 X
Q-33 X
R-42
SW-4
SW-7
T-17
T-35 X
T-58
TD-3 X
TD-4 X
Tl-1 X
Tl-2
U-38 X
V-23 X
V-65 X
W-23 X
X-32 X
• Table 2
Proposed Three Year Annual Sampling Plan
Ticona Shelby Site
Earth Tech Project 49750
TOC Alkalinity DowTherm A Diethelyene Oxide
X X X
X
X X X X
X X X
X
X
X X X
X X X X
X X X
X X X
X
X X X
X
X
X
X X X
X X X
X X X
X X X
X X X
X X X X
X
X X X X
X
X
X X X X
X
X X X X
X
X
X
X
X
X X X
X X X X
X X X X
X X X X
X X X
X
X
X
X
X X X X
X
X X X X
X X X
X X X X
X
X X X X
X X X X
X X X X
X X X X
X X X
voes NA Indicator
X X
X X
X X
X X
X X
X X
X X
X X
X
X
X X
X X
X X
X X
X X
X X
X X
X X
X X
X X
X X
X X
X X
X X
X X
X X
X X
X X
X X
X X
X X
X X
X X
NA Indicators Parameters: Nitrate-Nitrogen, Sulfate, Total Iron, Total Manganese, Methane, Acetate, Ethyl Alcohol
location Ethylene Glycol TOC
M-54 X X
B-34
C-49 X X
CC-33 X X
CC-64
D0-58R
EE-58 X X
F-55 X X
FF-34 X X
FF-62 X X
G-50
GG-61 X X
H-59
HH-48
HH-77
1-57
11-65
IT-1 X X
IT-5 X X
IT-6 X X
IT-BR X X
ff-9 X X
J-29 X X
J-59
K-28 X X
K-58
KK-55
N-29 X X
NN-105
0-25 X X
OT-1R
OT-2R
OT-6
OT-BA
OT-7
P-58 X X
PEW-1 X X
PEW-3 X X
PEW-4 X X
Q-33 X X
R-42
SW-4
SW-7
T-17
T-35 X X
T-58
TD-3 X X
T0-4 X X
Tl-1 X X
Tl-2
U-38 X X
V-23 X X
V-65 X X
W-23 X X
X-32 X X
• Table 3
Proposed Three Year Annual Sampling Plan
With Five Year Parameters
Ticona Shelby Site
Earth Tech Project 49750
Alkalinity -Dow Therm A Diethelyene Oxide voes
X X X
X
X X X X
X X X
X
X
X X X
X X X X
X X X
X X X
X
X X X
X
X
X
X
X
X X X
X X X
X X X
X X X
X X X
X X X X
X
X X X X
X
X
X X X X
X
X X X X
X
X
X
X
X
X X X
X X X X
X X X X
X X X X
X X X
X
X
X
X
X X X X
X
X X X X
X X X
X X X X
X
X X X X
X X X X
X X X X
X X X X
X X X
•
NA Indicator
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
NA Indicators Parameters: Nitrate-Nitrogen, Sulfate, Total Iron, Total Manganese, Methane, Acetate, Ethyl Alcohol
Metals: Antimony, Chromium, Cobalt, lead, Barium, Nickle, Selenium
Metals Phenols Chlordane
X X X
X X X
X X X
X X X
X X X
X X X
X X X
X X X
X X X
X X X
X X X
X X X
X X X
X X X
X X X
X X X
X X X
X X X
X X X
X X X
X X X
X X X
X X X
X X X
X X X
X X X
X X X
X X X
X X X
X X X
X X X
"
B-34
®
......
KK-55
®
-.........
i
i
! ! i I
!
!
i · ..... .. .....................
OT-7 M -S4 ®
HH-48
·••• HH-7-,@
@) EarthTech
A Tyco International Ltd. Company
Legend
® MONITORING WELL
G INNER TIER EXTRACTION WELL
OUTER TIER EXTRACTION WELL
PEW EXTRACTION WELL
• SURFACE WATER
--CREEK
~ BUILDING
D INDUSTRIAL POND
-RECREATION POND ,------, l_ ____ _! PROPERTY BOUNDARY
N
500 250 0
FIGURE 1
500
Feet
PROPOSED SAMPLE LOCATIONS
JULY 2007
CNA Holdings, lnc./Ticona
Shelby, North Carolina
79750