&EPA
United States
Environmental Protection
Agency
I: PA/600/R 1 U\ 7b June 2017
www.epa.gov/homeland-security-research
T reatment of Perf luorinated Alkyl
Substances in Wash Water Using
Granular Activated Carbon and Mixed-
Media
Office of Research and Development
Homeland Security Research Program
-------
EPA/600/R-17/175
June 2017
Treatment of Perfluorinated Alkyl
Substances in Wash Water Using
Granular Activated Carbon and Mixed-
Media
by
Jeffrey Szabo, John Hall and Matthew Magnuson
U.S. Environmental Protection Agency
Cincinnati, OH 45268
Sri Panguluri and Greg Meiners
CB&I Federal Services, LLC
Cincinnati, OH 45204
Interagency Agreement DW-89-92381801
U.S. Environmental Protection Agency Project Officer: John Hall
Office of Research and Development
Homeland Security Research Program
Cincinnati, OH 45268
Contract EP-C-12-014
U.S. Environmental Protection Agency Contracting Officer's Representative: Ruth Corn
Office of Research and Development
Homeland Security Research Program
Cincinnati, OH 45268
-------
This page left intentionally blank
-------
Disclaimer
The U.S. Environmental Protection Agency (EPA) through its Office of Research and
Development funded and managed the research described herein under contract EP-C-12-014
with CB&I Federal Services, LLC and Interagency Agreement DW-89-92381801 with the
Department of Energy. It has been subjected to the Agency's review and has been approved for
publication. Note that approval does not signify that the contents necessarily reflect the views of
the Agency. Any mention of trade names, products, or services does not imply an endorsement
by the U.S. Government or EPA. The EPA does not endorse any commercial products, services,
or enterprises.
The contractor role did not include establishing Agency policy.
11
-------
Table of Contents
Disclaimer ii
Table of Contents iii
Acknowledgements vi
1.0 Introduction 1
1.1 Background 1
1.2 Proj ect Obj ecti ve 1
1.3 WSTB System Description 1
2.0 Description of Experiment and Apparatus 4
2.1 Treatment Drum Setup 5
2.2 Lagoon Contamination Procedure 5
2.3 Contaminated Lagoon Water Treatment Procedure 7
3.0 Analysis of Test Results 11
4.0 Conclusions and Observations 32
5.0 References 33
List of Figures
Figure 1. Schematic overview of the Water Security Test Bed (WSTB) 2
Figure 2. Aerial view of the Water Security Test Bed (WSTB) 3
Figure 3. Water Security Test Bed (WSTB) discharge lagoon 3
Figure 4. PFAS removal from water treatment train 4
Figure 5. Rembind™ media mixed with sand 5
Figure 6. Idaho National Laboratory fire truck AFFF eductor spray mechanism 5
Figure 7. AFFF sprayed to contaminate lagoon water 6
Figure 8. AFFF contaminated lagoon water 7
Figure 9. Plumbing setup for inlet flow control 8
Figure 10. Inlet flow control during testing 8
Figure 11. Removal of PFBS from lagoon water using GAC and Rembind media 14
Figure 12. Normalized PFBS concentration change with increasing treated water volume 14
Figure 13. Normalized PFBS concentration change with increasing throughput (bed volumes). 15
Figure 14. Removal of PFBA from lagoon water using GAC and Rembind media 16
Figure 15. Normalized PFBA concentration change with increasing treated water volume 17
Figure 16. Normalized PFBA concentration change with increasing throughput (bed volumes). 17
Figure 17. Removal of PFPA from lagoon water using GAC and Rembind media 18
Figure 18. Normalized PFPA concentration change with increasing treated water volume 19
Figure 19. Normalized PFPA concentration change with increasing throughput (bed volumes). 19
Figure 20. Removal of PFHxS from lagoon water using GAC and Rembind media 20
Figure 21. Normalized PFHxS concentration change with increasing treated water volume 21
in
-------
Figure 22. Normalized PFHxS concentration change with increasing throughput (bed volumes).
21
Figure 23. Removal of PFHxA from lagoon water using GAC and Rembind media 22
Figure 24. Normalized PFHxA concentration change with increasing treated water volume 23
Figure 25. Normalized PFHxA concentration change with increasing throughput (bed volumes).
23
Figure 26. Removal of PFHS from lagoon water using GAC and Rembind media 24
Figure 27. Normalized PFHS concentration change with increasing treated water volume 25
Figure 28. Normalized PFHS concentration change with increasing throughput (bed volumes). 25
Figure 29. Removal of PFHA from lagoon water using GAC and Rembind media 26
Figure 30. Normalized PFHA concentration change with increasing treated water volume 27
Figure 31. Normalized PFHA concentration change with increasing throughput (bed volumes).27
Figure 32. Removal of PFOS from lagoon water using GAC and Rembind media 28
Figure 33. Normalized PFOS concentration change with increasing treated water volume 29
Figure 34. Normalized PFOS concentration change with increasing throughput (bed volumes). 29
Figure 35. Removal of PFOS from lagoon water using GAC and Rembind media 30
Figure 36. Normalized PFOA concentration change with increasing treated water volume 31
Figure 37. Normalized PFOA concentration change with increasing throughput (bed volumes).31
List of Tables
Table 1. Flowrates achieved through the media 9
Table 2. PFAS bulk water sampling at different times and locations 9
Table 3. Lagoon water quality sampling results 11
Table 4. List of PFAS compounds analyzed 11
Table 5. Average percent PFAS removal 13
iv
-------
Abbreviations
AFFF aqueous film forming foam
ALS ALS Environmental - Commercial laboratory in Kelso, WA
BWS bulk water sample
CB&I CB&I Federal Services LLC
COD chemical oxygen demand
CSIRO Commonwealth Scientific and Industrial Research Organization
EBCT empty-bed contact time
EPA U.S. Environmental Protection Agency
Ft feet
GAC granular activated carbon
gal gallon
gpm gallons per minute
hr hour
IA Interagency Agreement
INL Idaho National Laboratory
L liter
m meter
min minute
NTU nephelometric turbidity units
PFAS perfluorinated alkyl substances
PFBA perfluorobutanoic acid
PFBS perfluorobutane sulfonate
PFHA perfluoroheptanoic acid
PFHS perfluoroheptane sulfonate
PFHxA perfluorohexanoic acid
PFHxS perfluorohexane sulfonate
PFOS perfluorooctanesulfonic acid
PFOA perfluorooctanoic acid
PFPA perfluoropentanoic acid
psi pounds per square inch
QAPP Quality Assurance Project Plan
TOC total organic carbon
WSTB Water Security Test Bed
v
-------
Acknowledgements
Contributions from the following individuals to the field work described in this report are
acknowledged: Stephen Reese, Trent Armstrong, and Michael Carpenter of the Idaho National
Laboratory. Jim Goodrich of the Environmental Protection Agency, who is the WSTB program
manager.
vi
-------
Executive Summary
The U.S. Environmental Protection Agency's (EPA) National Homeland Security Research
Center partnered with the Idaho National Laboratory (INL) to build the Water Security Test Bed
(WSTB) at the INL test site outside of Idaho Falls, Idaho. The centerpiece of the WSTB is an 8-
inch diameter drinking water pipe formerly in service at INL. The pipe was exhumed from the
INL grounds and oriented in the shape of a small drinking water distribution system. The WSTB
has service connections to simulate water demands, fire hydrants, and removable coupons to
collect samples from the pipe interiors. Water from the WSTB pipe empties into a lined 28,000-
gallon (105,980 L) lagoon which contains dirt, algae and organic matter, and was used for this
study. Water from the lagoon can serve as "wash water," or water that is similar in nature to
water flushed from a distribution system into an impoundment, water used to wash down a
contaminated building, or water used to fight a fire.
This report summarizes the results from testing conducted to evaluate the treatment of large
volumes of water containing perfluorinated alkyl substances (PFAS). Specifically, treatment for
water contaminated by aqueous film forming foam (AFFF), which is used to fight very hot
hydrocarbon based fires, as may arise from the response to petroleum spills and transportation
accidents, was studied. Depending on the manufacturing process for the AFFF, the AFFF
contaminated water can contain emerging contaminants such as perfluorooctanesulfonic acid
(PFOS) and perfluorooctanoic acid (PFOA), which are the subject of recent EPA health
advisories. The AFFF selected for this study was a product widely used historically, and it
contained PFOA and PFOS. The goal of the treatment was to reduce the PFAS concentration
before disposal of the water (for example, in a sewer). The goal was not to reduce PFAS
concentrations to drinking water advisory levels.
The WSTB pipe was not used for these experiments; instead, the WSTB discharge lagoon was
contaminated with AFFF and the contaminated water was pumped through the treatment media
then emptied into the bladder tanks. Treatment of the AFFF contaminated water was investigated
via granular activated carbon (GAC) and mixed-media. Specifically, the Calgon Filtrasorb® 600
GAC and the Ziltek RemBind™ mixed-media were selected for evaluation.
The following is a summary of conclusions and observations about the performance and
implementation of adsorptive treatment of AFFF contaminated water, based on the testing
performed at the INL WSTB:
• The test results show that both GAC and RemBind™ are capable of removing various
short-and-long chain PFAS with an efficiency greater than 99.9%, on average, over a 12-
hour period, when the source water is spiked with firefighting levels of AFFF.
• The removal of shorter chains is of particular importance because newer AFFF products
are formulated to eliminate longer chain PFAS. This suggests that water contaminated
with newer AFFF formulations can also be successfully treated with these adsorbents,
although this should be experimentally verified, especially for the site-specific water,
which may contain substances that interfere or compete with PFAS adsorption.
vii
-------
GAC media can accommodate a higher flowrate than the RemBind™ media. Up to 6.5
gallons per minute (gpm) (41.7 bed volumes) was achieved in two drums of GAC in
series, while a total of 4 gpm (22 bed volumes) was achieved through two RemBind™
drums operated in parallel (2 gpm per drum). The RemBind™ media needs to be mixed
with significant amount of sand to achieve operational flows in the field. Also, the low
flow through and high pressure drop across the RemBind™ media drum may be too
significant for the RemBind™ drums to be operated in series without an intermediate
pump and storage mechanism. These factors impact the logistics of implementation of a
RemBind™ based treatment system, if this adsorbent is chosen based on site-specific
needs.
For some PFAS, the data suggests that the first GAC drum in series (drum 1) was losing
its adsorptive capacity, and breakthrough of PFAS was occurring. However,
breakthrough was not observed in the second drum in series. Because such variations
could impact utilization of the drums, this observation merits further investigation to
enable the appropriate implementation of these drums at a specific site.
-------
1.0 Introduction
1.1 Background
The U.S. Environmental Protection Agency's (EPA) National Homeland Security Research Center
has partnered with Idaho National Laboratory (INL) to build the Water Security Test Bed (WSTB)
at INL in Idaho Falls, Idaho. The centerpiece of the WSTB is an 8-inch diameter drinking water
pipe formerly in service at INL. The pipe was exhumed from the INL grounds and oriented in the
shape of a small drinking water distribution system. The WSTB has service connections, fire
hydrants, and removable coupons to collect samples from the pipe interiors. The WSTB has service
connections to simulate water demands, fire hydrants, and removable coupons to collect samples
from the pipe interiors. Water from the WSTB pipe empties into a 28,000 gallon lagoon that
contains dirt, algae and organic matter. Water from the lagoon can serve as "wash water," or water
that is similar in nature to water flushed from a distribution system into an impoundment, water
used to wash down a contaminated building, water used to fight a fire, etc.
This experiment focused on treatment of large volumes of water contaminated with
perfluorinated alkyl substance (PFAS), specifically aqueous film forming foam (AFFF). The
aqueous film forming foam is used to fight very hot hydrocarbon based fires, as may arise from
petroleum spills and transportation accidents. Water containing residual AFFF may need to be
treated before disposal or discharge into a sewer system. The WSTB pipe was not used for these
experiments; instead, the WSTB discharge lagoon was contaminated with AFFF and the
contaminated water was pumped through the treatment media and emptied into the bladder tanks.
Specifically, the Calgon Filtrasorb 600®1 granular activated carbon (GAC) and the Ziltek
RemBind™2 mixed-media were selected as the treatment media. GAC is commonly used for
perfluoro alkyl substances (PFAS) removal from water. Rembind is often used for PFAS
removal from soil, but it was evaluated in this study to determine its applicability to water
treatment.
1.2 Project Objective
The objective of the project was to simulate the treatment of wash-water contaminated with
AFFF and containing perfluoro alkyl substances (PFAS). The PFAS include emerging
contaminants such as perfluorooctanesulfonic acid (PFOS) and perfluorooctanoic acid (PFOA),
which are the subject of recent EPA health advisories (USEPA, 2016). Treatment options
evaluated included GAC and the RemBind™ mixed-media. The goal of the treatment was to
reduce the PFAS concentration before disposal of the water (for example, in a sewer). The goal
was not to reduce PFAS concentrations to drinking water advisory levels.
1.3 WS TB System Description
The WSTB consists primarily of an 8-inch (20 cm) diameter drinking water pipe oriented in the
shape of a small drinking water distribution system. The WSTB contains ports for service
connections and a 15-foot (5 m) removable coupon section, designed to sample the pipe interior
to examine the results from contamination/decontamination experiments on the pipe wall.
1 Filtrasorb® is a trademark of Calgon Carbon Corporation, 3000 GSK Drive, Moon Township, PA
2 RemBind™ is a trade marked powdered media developed by Ziltek Pty Ltd, Adelaide, Australia
-------
(Coupons are excised samples of materials to be tested.) Figure 1 schematically depicts the main
features of the WSTB.
Lagoon
w
Legend
FM Flaw Metor
pg Pressure Gauge
1P1 Instrument Panel 1 (Upslream, Cellular)
IPS [nslrument Panel 3 (Oownslream, Hadto)
XI Valve, Open
H Valve, Closed
O* Valve, Partly Open
TFire Hydrant
Flushing Hydrant
—3 Bltvl Flange
^-¦j Pressure Reducing Valve
Check Vah*>
> Service Connector (Closed)
Not to Scale
15-11 Coupon Section
Existing Fire Hydrant
Fire Hose
Start Oi WST0 |
©-
rejection Poft -Oct—
]—txt j IX 1 XH
Bulk Water Sample Tap
~ linking Water
from
INI Pumphous©
Parking
Area
Drainage
Ditch
Figure 1. Schematic overview of the Water Security Test Bed (WSTB).
Figure 2 shows the aerial view of the WSTB. The lower right corner shows the upstream and
system inlet; the upper left corner shows the lagoon which receives the water exiting the WSTB
pipe.
2
-------
Lagoon
WSTB End
WSTB Start
Figure 2. Aerial view of the Water Security Test Bed (WSTB).
Drinking water supplied to the WSTB is chlorinated ground water that also supplies the
surrounding INL facilities. Other than chlorination, the groundwater is not treated further. The
lagoon (Figure 3) has a total water storage capacity of 28,000 gallons (105,980 L). Water
contained in this lagoon served as surrogate for "wash water" during this testing. Dirt and
organic matter from the area surrounding the lagoon blows in water, and algae grows at the
bottom. For this experiment, the lagoon was contaminated by spraying AFFF over its surface
using a fire-fighting truck.
North
Figure 3. Water Security Test Bed (WSTB) discharge lagoon.
3
-------
2.0 Description of Experiment and Apparatus
Figure 4 shows the experimental setup that was originally proposed (shown using blue flowlines)
to individually test the performance of GAC and mixed-media (RemBind™) to remove PFAS
from AFFF-contaminated water. However, during the initial setup of the drums and flow testing,
it was determined that the RemBind™ media was not designed to handle the desired
experimental flowrate (5 gallons per minute [gpm]) through the drums in series (see section 2.1
for more detail). Therefore, each RemBind™ drum was operated in parallel (shown using red
flowlines in Figure 4), but the sampling locations and identification remained unchanged.
AFFF foam
O Sampling Locations 1 through 5
Effluent
Bladder Tank
GAC
Dram 2
GAC
Drum 1
Lagoon
(contaminated
with AFFF
foam)
Effluent
Bladder Tank
RemBind
Drum 2
RemBind
Drum 1
Figure 4. PFAS removal from water treatment train.
Filtrasorb 600 GAC is made from select grades of bituminous coal to produce a high activity,
durable, granular product capable of withstanding the abrasion associated with repeated
backwashing, hydraulic transport, and reactivation for reuse. Activation is carefully controlled to
produce a significant volume of different types of pores for effective adsorption of a broad range
of high and low molecular weight organic contaminants (Calgon, 2015).
RemBind™ is a powdered reagent for the chemical fixation of organic and inorganic
contaminants in soil. The product was developed by Ziltek in collaboration with Australian-
based Commonwealth Scientific and Industrial Research Organization (CSIRO) and contains a
proprietary blend of reagents (Ziltek, 2014). The main constituents of RemBind™ are:
• Activated carbon
• Aluminum hydroxide (amorphous)
• Kaolin clay and other proprietary additives
The AFFF used in this study was 3M's Light Water™ 3% concentrate (FC-203CF, St. Paul,
MN). This formulation was used historically, and it contains PFOA and PFOS, which can be
found as ground water contaminants at some sites, but are not in many modern AFFF
formulations. The treatment effectiveness of each media type for this AFFF formulation was
evaluated during this study.
4
-------
2.1 Treatment Drum Setup
During initial setup of the treatment train shown in Figure 4, it was observed that 5 gpm was able
to flow through the GAC drums, but that no flow was exiting through the RemBind™ lead drum
when RemBind™ alone was in the drum. The vendor was immediately contacted and the
ensuing discussions indicated that the media (which was finer than the GAC-media) provided by
the vendor was not well suited for a flow-through setup. The media was designed for batch mode
testing, where the media is mixed with water, and then water is removed after the desired contact
time. As a potential solution, the vendor suggested that the RemBind™ drams be emptied, the
media mixed with 50% sand (by volume) to increase the media porosity and then the
RemBind™/sand mixture be put back in the drums. In addition, to prevent clogging, the bottom
4 inches of the drums were filled with pure sand before it was filled back up with the amended
RemBind™ media. Figure 5 shows the RemBind™ media daim as refilled. The sand used in
Figure 5 was swimming pool filter silica sand with a particle size of 0.43 to 0.85 mm.
Figure 5. Rembind™ media mixed with sand.
2.2 Lagoon Contamination Procedure
The AFFF contained in the 5-gallon tank was connected to an eductor mechanism as shown in
Figure 6 for spraying.
Figure 6. Idaho National Laboratory fire truck AFFF eductor spray mechanism.
The eductor is a venturi jet device that uses pressurized water to entrain, mix and pump other
5
-------
liquids such as the AFFF. The eductor consists of two basic parts: (1) the motive nozzle, which
converts the water pressure energy to kinetic (velocity) energy, and (2) the suction
chamber/diffuser section where the entrainment and mixing of water and AFFF takes place.
The eductor has a knob that is used to set the correct mix ratio based on concentration of the mix
of AFFF (in this case 3%). The eductor also has a minimum pressure requirement (60 psi) which
is necessary to generate sufficient suction force to deliver the proper mix of AFFF and water.
The INL fire department used their fire truck equipment to set up the eductor and spray five
gallons of 3M's Lightwater 3% AFFF into the lagoon, as shown in Figure 7. The spray
contamination of the lagoon water was completed in approximately 5 minutes.
Figure 7. AFFF sprayed to contaminate lagoon water.
The majority of the foam dissipated within two hours, with only some remnants remaining along
the edges of the lagoon, as shown in Figure 8.
6
-------
Figure 8. AFFF contaminated lagoon water.
Then, lagoon water was allowed to sit overnight prior to testing the next day. The lagoon was
not actively mixed, and the remaining foam was no longer observed.
2.3 Contaminated Lagoon Water Treatment Procedure
The AFFF-containing lagoon water was pumped using a submersible pump through flexible
tubing with flow controlled rotameters. Any unused water was bypassed back into the lagoon,
and the flow of this water promoted mixing within the lagoon. The contaminated lagoon water
was pumped from the lagoon and through the treatment drums at 5 gpm. However, as mentioned
previously, the RemBind™ media containing drums were unable to handle the desired
experimental flowrate while operating in series (due to an excessive pressure drop). Therefore,
each of the RemBind™ containing media-drums were operated in parallel (as shown in Figure
4), but the sampling locations and identification remained unchanged. Treated water exiting the
drums flowed into bladder tanks. The inlet flow control setup is shown in Figure 9 and Figure
10.
7
-------
Inlet to Carbon Media
Drum # 1
Inlet to RemBind Media
Drums #1 and # 2
Inlet Flow Control Valves
Sampling Port for Inlet
Water (Location Id. #1)
AFFF Contaminated
Lagoon Water
Bypass for Unused Water
back to the Lagoon
Figure 9. Plumbing setup for inlet flow control,
Figure 10. Inlet flow control during testing,
When the water treatment began, flow through the RemBind™ system was set to 1 gpm, or 0.5
gpm through each RemBind™ daim. The Calgon GAC containing drums in series were operated
at the experimental design flowrate of 5 gpm. In the hours after the treatment began, flow was
8
-------
gradually increased through the RemBind™ media. After each attempted flow adjustment, the
flowrate was recorded. Table 1 contains a summary of flowrates and computed empty-bed
contact times (EBCT) achieved throughout the test.
Table 1. Flowrates achieved through the media
Time
After
Treatment
Start
(hr)
GAC Treatment (Drums in series)
RemBind™ Media (drums in parallel)
Flow
(gpm)
EBCT
(min)
Volume
Treated
(gal)
Bed
Volumes
(#)
Flow
(gpm)
EBCT
(min)
Volume
Treated
(gal)
Bed
Volumes
(#)
0.08
5.0
22.0
25
0.2
1.0
110.0
5
0.0
0.50
5.0
22.0
150
1.4
1.0
110.0
30
0.3
1
5.0
22.0
300
2.7
1.0
110.0
60
0.5
3
6.5
16.9
1080
9.8
2.5
44.0
270
2.5
5
6.5
16.9
1860
16.9
4.0
27.5
750
6.8
7
6.5
16.9
2640
24.0
4.0
27.5
1215
11.0
9
6.5
16.9
3420
31.1
4.0
27.5
1695
15.4
12
6.5
16.9
4590
41.7
4.0
27.5
2415
22.0
EBCT: Empty Bed Contact Time (min) (Empty bed volume in gallons divided by flow rate in gallons per minute)
Bulk water samples (BWS) for PFAS analysis were collected throughout the experiment and are
summarized in Table 2.
Table 2. PFAS bulk water sampling at different times and locations
Location/ID
Time ID
GAC
GAC
RemBind™
RemBind™
Inlet
Drum 1
Outlet
Drum 1
Drum 2
(BWS1)
(BWS2)
(BWS3)
(BWS4)
(BWS5)
1
7:25
9:30
7:35
9:30
7:35
2
7:30
11:30
8:00
11:30
8:00
3
7:35
13:30
8:30
13:30
8:30
4
8:00
15:30
9:30
15:30
9:30
5
8:30
19:30
10:30
19:30
10:30
6
11:30
11:30
7
10:30
12:30
12:30
8
13:30
13:30
9
12:30
14:30
14:30
10
15:30
15:30
11
14:30
16:30
16:30
12
17:30
17:30
13
16:30
18:30
18:30
14
19:30
19:30
15
18:30
16
19:30
BWS: Bulk Water Sample, or a sample of the water flowing into and out of the drums
From Table 2, a PFAS sample label ID of BWS1-7 means the sample was collected at the BWS1
(location ID in "bold" is the Inlet) and at time ID 7, which is 10:30 AM (for BWS1). Any blank
9
-------
values in the table means no bulk water sample was collected at that location at that time
ID/sequence. In addition to PFAS samples, periodic samples from lagoon were collected for
quantifying some of the routine water quality parameters including pH, temperature, free
chlorine, turbidity, specific conductivity, chemical oxygen demand (COD), and total organic
carbon (TOC). The results from the testing are presented in Section 3.0.
10
-------
3.0 Analysis of Test Results
The results from routine water quality samples indicated that the addition of AFFF caused an
increase in turbidity, COD and TOC levels of the lagoon water. These data are summarized in
Table 3. It should be noted that TOC, COD, turbidity and pH increased upon introduction of the
AFFF into the lagoon water. The observed increase in temperature throughout the day is due to
the sun shining on the lagoon surface, which heats the water. The lagoon water cools off at
night.
Table 3. Lagoon water quality sampling results
Date/Activity
Clock
Time
Related
PFAS
Sample ID
PH
Temp.
(C)
Free
Chlo-
rine
(mg/L)
Turb.
(NTU)
Specific
Cond.
(ps/cm)
COD
(mg/L)
TOC
(mg/L)
9/19/2016
Background
7:25
BWS 1-1
8.2
12.8
0.50
507
36
2.92
8:45
NA
8.34
12.3
9:15
NA
0.09
0.75
T6 (4 hours)
11:30
BWS 1-8
1.57
491
269
75.44
14:05
NA
8.72
22.7
T10 (8 hours)
15:30
BWS 1-12
1.49
495
271
81.49
18:45
NA
9.04
19.5
T14 (12 hours)
19:30
BWS 1-16
1.47
484
278
81.37
9/20/2016
Restart
pumping
9:00
NA
8.50
12.0
0.09
TOC: Total Organic Carbon; COD: Chemical Oxygen Demand
PFAS samples were analyzed using EPA method 537, which was modified with an expanded list
of 36 analytes (listed with CAS Registry Number® in Table 4) (Shoemaker, et al. 2009). The
expanded list included some of the degradation PFAS precursors such as fluorotelomers. TOC
was measured via EPA Method 415.3 (Potter and Wimsatt, 2005). COD was measured using
Hach Method 8000 (Hach, 2014a). Free chlorine was measured using Hach Method 10102
(Hach, 2014b). Turbidity measurements were conducted according to Standard Method 2130
(APHA, 1999). Specific conductivity was measured according the Thermo Scientific Orion
Versa Star user's manual (Thermo Scientific, 2014). pH measurements were conducted
according to the Extech 407220 pH meter user's manual (Extech, 2016). Temperature was
measured by immersing a National Institute of Standards and Technology (NIST)-traceable
thermometer in the water sample.
Table 4. List of PFAS compounds analyzed
Analyte
CAS Number/ID
2-(N-ethylperfluoro-l-octanesulfonamido)-ethanol
1691-99-2
2-(N-methylperfluoro-l-octanesulfonamido)-ethanol
24448-09-7
6:2 Fluorotelomer sulfonate
27619-97-2
8:2 Fluorotelomer sulfonate
8:2FTS
11
-------
Analyte
CAS Number/ID
d7-N-MeFOSE (Surr)
d7-N-MeFOSE (Surr)
d9-NEtFOSE (Surr)
d9-NEtFOSE (Surr)
N-ethylperfluoro-l-octanesulfonamide
4151-50-2
N-methylperfluoro-l-octanesulfonamide
31506-32-8
Perfluorobutane Sulfonate
45187-15-3
Perfluorobutanoic Acid
375-22-4
Perfluorodecane Sulfonate
335-77-3
Perfluorodecanoic Acid
335-76-2
Perfluorododecanoic Acid
307-55-1
Perfluoroheptane sulfonate
375-92-8
Perfluoroheptanoic Acid
375-85-9
Perfluorohexane Sulfonate
108427-53-8
Perfluorohexanoic Acid
307-24-4
Perfluoro-n-[l,2,3,4,5-13C5] nonanoic acid
PFNNAC13
Perfluoro-n-[l,2,3,4-13C4] butanoic acid
PFBTAC13
Perfluoro-n-[l,2,3,4-13C4] octanoic acid
PFOCAC13
Perfluoro-n-[l,2-13C2] decanoic acid
PFDCAC13
Perfluoro-n-[l,2-13C2] hexanoic acid
PFHXAC13
Perfluoro-n-[l,2-13C2]dodecanoic Acid
PFDDAC13
Perfluoro-n-[l,2-13C2]undecanoic Acid
PFUDAC13
Perfluorononanoic Acid
375-95-1
Perfluoro-n-tetradecanoic acid
376-06-7
Perfluoro-n-tridecanoic acid
72629-94-8
Perfluorooctane Sulfonate
45298-90-6
Perfluorooctanoic Acid
335-67-1
Perfluorooctylsulfonamide
754-91-6
Perfluoropentanoic Acid
2706-90-3
Perfluoroundecanoic Acid
2058-94-8
S_6:2 Fluorotelomer sulfonate-13C2
13C2-6:2FTS (Surr)
S_N-ethyl-d5-perfluoro-l-octanesulfonamide
d5-NEfFOSA (Surr)
Sodium perfluoro-l-[l,2,3,4-13C4] octanesulfonate
NAPFOcSLFN8C13
Sodium perfluoro-l-hexane[1802]sulfonate
180-PFHS
Overall, 43 samples were analyzed for the suite of PFAS. Only 18 of the 36 analytes listed in
Table 4. List of PFAS compounds analyzed were detected in one (or more) of the samples. The
following nine compounds were the most common PFAS detected in the inlet to the treatment
media.
1. Perfluorobutane Sulfonate (PFBS)
2. Perfluorobutanoic Acid (PFBA)
3. Perfluoropentanoic Acid (PFPA)
12
-------
4. Perfluorohexane Sulfonate (PFHxS)
5. Perfluorohexanoic Acid (PFHxA)
6. Perfluoroheptane sulfonate (PFHS)
7. Perfluoroheptanoic Acid (PFHA)
8. Perfluorooctane Sulfonate (PFOS)
9. Perfluorooctanoic Acid (PFOA)
Of these nine detected compounds, the top four compounds PFOS (82.1%), PFHxS (12.8%),
PFHS (2.6%) and PFOA (0.1%) represent a combined total of 98.2% of the inlet loading. A
summary of the average percent removal observed over the 12-hour treatment study is presented
in Table 5.
"able 5. Average percent
PFAS remova
PFAS
GAC
Average %
Removal
Rembind
Average %
Removal
PFOS
99.986%
99.991%
PFOA
99.997%
99.996%
PFBS
99.999%
99.996%
PFBA
99.991%
99.990%
PFHS
99.999%
99.997%
PFHA
99.983%
99.996%
PFHxS
99.9997%
99.997%
PFHxA
99.985%
99.994%
PFPA
99.989%
99.994%
Figure 11 shows the individual treatment system performance for PFBS. Figure 12 and Figure 13
show the normalized PFBS concentration plotted against the total volume treated and bed
volumes treated. On average, the GAC media achieved a removal efficiency of 99.999%, and
the RemBind™ (one drum) achieved 99.996% removal efficiency. The upward trajectory of the
resulting effluent concentration by an order of magnitude from the GAC Drum 1 effluent may be
an indicator that the GAC media is becoming spent and breakthrough is observed. The black
straight line at the bottom in the following set of figures (Figure 11 through Figure 37) represents
the limit of detection for the specific compound shown in the individual graph.
13
-------
100,000
10,000
£3 1,000
o
'•p
n
+¦>
C
0)
u
c
o
u
to
CO
100
10
»
H
-Influent
- GAC Second
Drum Effluent
• GAC First Drum Effluent
9 Rembind Effluent (One Drum]
!
Detection Limit = 0.41 ng/L
0 123456789 10 11 12
Time After Treatment Start (hr)
Figure 11. Removal of PFBS from lagoon water using GAC and Rem bind media.
1.00000
0.10000
0.01000
O 0.00100
u
00
CO
u.
± 0.00010
E 0.00001
0.00000
• GAC Second Drum Effluent
a r~fc J_ : I «— rri
—•—Rembind Effluent
—•—GAC First Drum Effluent ^
L
Lr
-4—
(•V 11 ¦"
-4^a
4000
5000
1000 2000 3000
Volume Treated (gallons)
Figure 12. Normalized PFBS concentration change with increasing treated water volume.
14
-------
1.00000
0.10000
0.01000
0.00100
0.00010
£ o.ooooi
0.00000
u
g
c
o
n
-------
Figure 14 shows the individual treatment system performance for PFBA. Figure 15 and Figure
16 show the normalized PFBA concentration plotted against the total volume treated and bed
volumes treated. On average, the GAC media achieved a removal efficiency of 99.991%, and
the RemBind™ (one drum) achieved 99.990% removal efficiency. Similar to PFBS, the upward
trajectory of the GAC effluent from Drum 1 may be an indicator that the GAC media is
becoming spent.
100,000
10,000
SP 1,000
c
o
0)
u
c
o
u
<
ea
100
10
0
| m |
- Influent
• GAC First Drum
Effluent
GAC Second
urum tmueni
9 K
emDina tmuent tune
: urum]
—•
•—
\jr
^^ ^ *
* ^
r—
Detection Limit = 0.57 ng/L
0
10 11 12
123456789
Time After Treatment Start (hr)
Figure 14. Removal of PFBA from lagoon water using GAC and Rembind media.
16
-------
1.00000
o
u
U1 0.10000
c
o
0.01000
c
01
u
c
o
(J
2 0.00100
•a
oi
M
1 0.00010
o
z
0.00001
j « I L
i i i i
i i i i |
—•— GAC Second Drum Effluent
—•— Rembind Effluent
—•— GAC First Drum Effluent
—
•
•
irV—•—
1000 2000 3000
Volume Treated (gallons)
4000
5000
Figure 15. Normalized PFBA concentration change with increasing treated water volume.
1.00000
o
(J
tj 0.10000
c
o
c
-------
Figure 17 shows the individual treatment system performance for PFPA. Figure 18 and Figure
19 show the normalized PFPA concentration plotted against the total volume treated and bed
volumes treated. On average, the GAC media achieved a removal efficiency of 99.989% and the
RemBind™ (one drum) achieved 99.994% removal efficiency. Similar to PFBS and PFBA, the
upward trajectory of the GAC effluent from Drum 1 may be an indicator that the GAC media is
becoming spent.
100,000
10,000
^ 1,000
o
TO
V
u
c
o
u
<
Q_
100
10
0
• Influent
GAC Second Drum Effluent
•GAC First Drum Effluent
¦Rembind Effluent (One Drum)
Detection Limit = 0.31 ng/L
0
123 45 678 9 10 11 12
Time After Treatment Start (hr)
Figure 17. Removal of PFPA from lagoon water using GAC and Rembind media.
18
-------
o
'•p
re
1,00000
O
u
tj 0.10000
4)
u
£
o
VJ
<
a.
u.
Cl
•u
Rembind Effluent
—•—GAC First Drum Effluent
*
- 9
•
•-
kc
*
^ *
1000 2000 3000
Volume Treated (gallons)
4000
5000
Figure 18. Normalized PFPA concentration change with increasing treated water volume.
1.00000
o
u
tj 0.10000
c
o
0.01000
0.00100
c
<11
u
c
0
u
<
a.
LL
CL
"O
0)
1 0.00010
0.00001
¦ ¦ ¦ ¦
—e
-•-R
AC Second Drum Effluent
embind Effluent
—G
AC First Drum
Effluent^"'^
I m—•-
"•
-•
20 40 60
Throughput (Bed Volumes)
80
100
Figure 19. Normalized PFPA concentration change with increasing throughput (bed
volumes).
19
-------
Figure 20 shows the individual treatment system performance for PFHxS. Figure 21 and Figure
22 show the normalized PFHxS concentration plotted against the total volume treated and bed
volumes treated. On average, the GAC media achieved a removal efficiency of 99.9997% and
the RemBind™ (one drum) achieved 99.9967% removal efficiency. Similar to the PFBS, PFBA,
and PFPA graphs, the upward trajectory of the GAC effluent from Drum 1 may be an indicator
that the GAC media is becoming spent.
10,000,000
1,000,000
_ 100,000
tiD
c
o
u
£
o
u
I/)
X
10,000
1,000
100
10
1
0
0
~—Influent
— GAC Secon
d Drum Effluent
m uml nrsi
* Rembind
l/iuiii cmueni
Effluent (One
)rum)
l
l
TV
Detection
Limit
= 0.35 ng/L
10 11 12
123456789
Time After Treatment Start (hr)
Figure 20. Removal of PFHxS from lagoon water using GAC and Rembind media.
20
-------
1.00000
o
y 0.10000
U
c
o
c
01
u
C
o
(J
1/5
X
¦o
(U
0.01000
0.00100
0.00010
0.00001
0.00000
—•—GAC Second Drum Effluent
a n L: „ _i rXCI. . „ „ .
> Rembind Effluent
• GAC First Drum Effluent
•
1000 2000 3000
Volume Treated (gallons)
4000
5000
Figure 21. Normalized PFHxS concentration change with increasing treated water volume.
1.00000
y 0.10000
u
c
o
'¦S
n
c
-------
Figure 23 shows the individual treatment system performance for PFHxA. Figure 24 and Figure
25 show the normalized PFHxA concentration plotted against the total volume treated and bed
volumes treated. On average, the GAC media achieved a removal efficiency of 99.985%, and
the RemBind™ (one drum) achieved 99.994% removal efficiency. Similar to the previous
graphs, the upward trajectory of the GAC effluent from Drum 1 may be an indicator that the
GAC media is becoming spent.
100,000
10,000
*£P 1,000
£
o
« ioo
+*
C
O
u
J io
<
x
LL. 1
a. 1
Detection Limit = 0,
0
0 123 45 678 9 10 11 12
Time After Treatment Start (hr)
Figure 23. Removal of PFHxA from lagoon water using GAC and Rembind media.
¦Influent
GAC Second Drum Effluent
•GAC First Drum Effluent
•Rembind Effluent (One Drum)
22
-------
c
01
u
c
o
(J
<
> GAC Second Drum Effluent
•Rembind Effluent
¦GAC First Drum Effluent
1.00000
0.10000
0.01000
0.00100
0.00010
0.00001
0 1000 2000 3000
Volume Treated (gallons)
4000 5000
Figure 24. Normalized PFHxA concentration change with increasing treated water volume.
1.00000
c
o
'¦S
n
f
c
41
u
C
o
u
<
0.10000
0.01000
0.00100
0.00010
0.00001
¦GACSecond Drum Effluent
¦ Rembind Effluent
¦GAC First Drum Effluent
0 20 40 60 80
Throughput (BedVolumes)
Figure 25. Normalized PFHxA concentration change with increasing throughput (bed
volumes).
23
-------
Figure 26 shows the individual treatment system performance for PFHS. Figure 27 and Figure
28 show the normalized PFHS concentration plotted against the total volume treated and bed
volumes treated. On average, the GAC media achieved a removal efficiency of 99.999% and the
RemBind™ (one drum) achieved 99.997% removal efficiency. Similar to the previous graphs,
the upward trajectory of the GAC effluent from Drum 1 may be an indicator that the GAC media
is becoming spent.
1,000,000
100,000
w
O
re
v
u
c
o
u
10,000
1,000
100
10
0
0
• Influent
GAC Second Drum Effluent
•GAC First Drum Effluent
•Rembind Effluent (One Drum)
Detection Limit = 0.49 ng/L
123456789 10 11 12
Time After Treatment Start (hr)
Figure 26. Removal of PFHS from lagoon water using GAC and Rembind media.
24
-------
o
u
o
ffi
i—
<5
c
V
u
1.00000
0.10000
0.01000
O 0.00100
u
{/)
-/-
r
1000 2000 3000
Volume Treated (gallons)
4000
5000
Figure 27. Normalized PFHS concentration change with increasing treated water volume.
o
u
c
o
c
41
u
C
o
u
V)
"O
-------
Figure 29 shows the individual treatment system performance for PFHA. Figure 30 and Figure
31 show the normalized PFHA concentration plotted against the total volume treated and bed
volumes treated. On average, the GAC media achieved a removal efficiency of 99.983%, and
the RemBind™ (one drum) achieved 99.996% removal efficiency. Similar to the previous
graphs, the upward trajectory of the GAC effluent from Drum 1 in may be an indicator that the
GAC media is becoming spent.
100,000
10,000
g) 1,000
¦Influent
GAC Second Drum Effluent
• GAC First Drum Effluent
¦Rembind Effluent (One Drum)
O
re
-------
o
V
re
1.00000
o
u
U" 0.10000
0.01000
0.00100
-------
Figure 32 shows the individual treatment system performance for PFOS. Figure 33 and Figure
34 show the normalized PFOS concentration plotted against the total volume treated and bed
volumes treated. On average, the GAC media achieved a removal efficiency of 99.986%, and
the RemBind™ (one drum) achieved 99.991% removal efficiency. Unlike the shorter chain
compounds, there is no noticeable upward trajectory of the GAC effluent from Drum 1.
10,000,000
1,000,000
_ 100,000
10,000
1,000
100
w
c
o
'¦p
u
o
u
I/)
o
W- • 9
•—
—•—
—•
i
t
fl
• Influent
—•—GAC 1
First Drum Effluent
—•—GAC Second Drum Effluent
* Rembind Effluent (One Drum)
J
I—i
10
1
0
0
Detection Limit = 0.60 ng/L
10 11 12
123456789
Time After Treatment Start (hr)
Figure 32. Removal of PFOS from lagoon water using GAC and Rembind media.
28
-------
o
"¦p
rs
1.00000
o
u
U 0.10000
0.01000
4)
u
c
o
u
o 0.00100
M
^ 0.00010
o
z
0.00001
• GAC Second Drum Effluent
•Rembind Effluent
¦GAC First Drum Effluent
1000 2000 3000
Volume Treated (gallons)
4000
5000
Figure 33. Normalized PFOS concentration change with increasing treated water volume.
1.00000
o
u
t? 0.10000
c
o
n
£ 0.01000
u
C
o
u
o 0.00100
¦o
a>
N
^ 0.00010
o
0.00001
L 1 i 1
1 1 1 1 1 1
—c
—R
5AC Second Drum Effluent
embind Effluent
—•—C
jAC First Drum
Effluent
I
. yT
I
rN
/
20 40 60
Throughput (Bed Volumes)
80
100
Figure 34. Normalized PFOS concentration change with increasing throughput (bed
volumes).
29
-------
Figure 35 shows the individual treatment system performance for PFOA. Figure 36 and Figure
37 show the normalized PFOA concentration plotted against the total volume treated and bed
volumes treated. On average, the GAC media achieved a removal efficiency of 99.997%, and
the RemBind™ (one drum) achieved 99.996% removal efficiency. Similar to the previous graphs
(other than PFOS), the upward trajectory of the GAC effluent from Drum 1 in may be an
indicator that the GAC media is becoming spent.
100'000 V«-« • • # 0
10,000
-—. —•—Influent GAC First Drum Effluent
M 1,000 » GAC Second Drum Effluent • Rembind Effluent (One Drum)
c
o
£ 100*
b.
C
O)
u
O 10
u
<~
O
t i L+ T ' T">-*1 ¦ 1 ¦ '"? i tL-1——^
0 L-U.^/Mg/L.
0123456789 10 11 12
Time After Treatment Start (hr)
Figure 35. Removal of PFOS from lagoon water using GAC and Rembind media.
\ • •-
¦ Influent
GAC Second Drum Effluent
¦GAC First Drum Effluent
¦Rembind Effluent (One Drum)
* 1
1 1
> 1—1
Detection 1
Limit =
: 0.27 ng/L
30
-------
o
1.00000
o
u
U" 0.10000
c
<11
u
c
o
u
<
o
01
N
0.01000
0.00100
« 0.00010
£
o
z
0.00001
¦ 1 1 1
—•—GAC Second Drum Effluent
> Rembind Effluent
—•—GAC First Drum Effluent
•—
- r
~- —
1000 2000 3000
Volume Treated (gallons)
4000
5000
Figure 36. Normalized PFOA concentration change with increasing treated water volume.
1.00000
o
u
5 0.10000
c
o
c
41
u
c
o
u
<
O
"O
01
N
0.01000
0.00100
"5 0.00010
E
E
O
z
0.00001
—c
-~-R
AC Second Drum Effluent
embind Effluent
—G
AC First Drum
Effluent
20 40 60
Throughput (Bed Volumes)
80
100
Figure 37. Normalized PFOA concentration change with increasing throughput (bed
volumes).
31
-------
4.0 Conclusions and Observations
Based on the data presented in Section 3, the following general conclusion can be drawn from
the AFFF treatment experiments. Also presented below are some observations that may be
relevant to field-treatment of AFFF-contaminated water at specific sites:
• The study results show that both GAC and RemBind™ are capable of removing various
short-and-long chain PFAS with an efficiency greater than 99.9%, on average, over a 12-
hour period when the lagoon water is spiked with firefighting levels of AFFF. The
removal of shorter chains PFAS is of particular importance because newer AFFF
products are formulated to eliminate longer chain PFAS. This suggests that water
contaminated with newer AFFF formulations can also be treated with these adsorbents,
although this should be experimentally verified, especially for each site-specific water,
which may contain substances that interfere and may compete with PFAS adsorption.
• GAC media can accommodate a higher flowrate than the RemBind™ media. Up to 6.5
gpm was achieved in two drums of GAC in series. A total flow of 4 gpm was achieved
through two modified RemBind™ drums (Rembind media mix with 50% sand by
volume) operated in parallel with 2 gpm per drum. This impacts the configuration of a
RemBind™ based treatment system, if this adsorbent is chosen based on site-specific
needs.
• The RemBind™ media needs to be mixed with significant amount of sand to achieve
operational flows in the field, as needed in this study. Also, the low flow through and
high pressure drop across the RemBind™ media drum could be too significant for the
RemBind™ drums to be operated in series without an intermediate pump and storage
mechanism. This could also impact design of a RemBind™ based treatment system,
along with requiring on-site availability of suitable sand or other material to mix with the
media.
• For some PFAS, the data suggests that the first GAC drum in series (drum 1) was losing
its adsorptive capacity, and breakthrough of PFAS was occurring. However,
breakthrough was not observed in the second drum in series. Because such variations
could impact utilization of the drums (with specific bed volumes), this observation merits
further investigation to evaluate the number of drums needed to achieve the same
treatment goals described in this study, at a specific site. Additionally, the water quality
and the organic content of each water may impact the breakthrough.
In summary, should AFFF contaminate a water body after the response to a fire, the data show
that either GAC or RemBind™ could be used to adsorb most of the PFAS before disposal of the
treated water, e.g., in a sewer or by other means. (Note: it is important to ensure discharge of
treated water conforms to local regulation and requirements of the wastewater authorities at a
particular discharge site.) Despite similar performance at removing contaminants, from an
implementation standpoint, these findings also indicate that the GAC could be used as received,
while the RemBind™ needs to be mixed with sand, which is a time consuming process when
performed manually.
32
-------
5.0 References
American Public Health Association. (1999). Method 2130: Turbidity. Standard Methods for the
Examination of Water and Wastewater, 20th ed. Washington, DC: American Public
Health Association,
Calgon. (2015). Data Sheet for Filtrasorb® 600 Granular Activated Carbon (DS-FILTRA60015-
EIN-E1), Moon Township, PA: Calgon Carbon Corporation.
Extech. (2016). Extech Instruments Model 407229 pH meter User's Manual. Nashua, NH:
Extech Instruments.
http://www.extech.com/displav/?id= 14215. accessed 05/31/2017.
Hach Company. (2014a). Method 8000: Chemical Oxygen Demand.
https://www.hach.com/asset-uet.download.isa'?id=7639983816. accessed 05/3 1/2007.
Hach Company. (2014b). Method 10102: Free Chlorine, https://www.hach.com/asset-
get.download-en.i sa?code=55578. accessed 05/31/2017.
Potter, BB and Wimsatt, JC. (2005) Method 415.3. Measurement of Total Organic Carbon,
Dissolved Organic Carbon, and Specific UV absorbance at 254 nm in Source Water and
Drinking Water. Washington, DC: U.S. Environmental Protection Agency.
https://cfpub.epa.gov/si/si public record report.cfm?dirEntrvld= 103917. accessed
05/31/2017.
Shoemaker JA, Grimmett PE and Boutin BK. (2009) Method 537. Determination of Selected
Perfluorinated Alkyl Acids in Drinking Water by Solid Phase Extraction and Liquid
Chromatography/Tandem Mass Spectrometry (LC/MS/MS), Version 1.1. Cincinnati,
OH: U.S. Environmental Protection Agency.
https://cfpub.epa.gov/si/si public record report.cfm?dirEntrvld= 198984&simpleSearch= 1 &sear
chA11=EPA%2F600%2FR-08%2F092. accessed 05/31/2017.
Thermo Scientific. (2014). Thermo Scientific Orion Versa Star User's Manual. Waltham, MA:
Thermo Fisher Scientific Corporation.
https://www.thermofisher.com/order/catalog/product/VSTAR90. accessed 05/3 1/2017
USEPA. (2016). Drinking Water Health Advisories for PFOA and PFOS.
https://www.epa.gov/ground-water-and-drinking-water/drinking-water-health-advisories-
pfoa-and-pfos. accessed 05/3 1/2017.
Ziltek. (2014). RemBind™ Product Overview. Immobilising Soil Contaminants. Z070-02 04/14.
Thebarton, Australia: Ziltek corporation.
33
-------
vvEPA
United States
Environmental Protection
Agency
PRESORTED STANDARD
POSTAGE & FEES PAID
EPA
PERMIT NO. G-35
Office of Research and Development (8101R)
Washington, DC 20460
Official Business
Penalty for Private Use
$300
------- |