Abiotic Treatment Technologies for In-Site Remediation of Persistent

Perfluoroalkyl Acids

Linda S. Lee1,2*, Saerom Park2,3, and Jenny Zenobio2,3

1Professor of Environmental Chemistry, Agronomy2Interdisciplinary Ecological Sci. & Engineering Graduate Program, Program Head

3PhD Students

June 26, 2015NICOLE Workshop

Manchester, UK

ER-2426Office of the

Vice President

US Military big user of AFFFs, which are a primary source of PFAAs and likely present at > 550 sites.

Additional, many other military sites due to use/releases of AFFFs to extinguish hydrocarbon fires or industrial activities, etc.

PFAAs are persistent, bioaccumulative, and relatively mobile.

● Currently, no viable in-situ technology has been successfully demonstrated for the entire suite of PFASs, particularly linear PFOS, or PFAA precursors.

● Provisional health advisory levels are well below what is expected at military sites

*Moody and Field, 1999; Moody et al., 2003; Schultz et al., 2004

** *

Aqueous Film-Forming Foam (AFFF) Sources

Telomer-based fluorinated surfactants

Electrofluorination-based fluorinated surfactants

Perfluorocarboxylic acidsPFCAs (PFOA pKa < 4)

Perfluorosulfonic acidsPFSAs (PFOS pKa < 0)

Terminal microbial end products

Σ = PFAAs = perfluoro alkyl acids

PFASs in Aqueous Film Forming Foams (AFFFs)(Place & Field, EST, 2012)

B, C & D

Fenton’s Reagent◦ Thermodynamic studies indicate not effective◦ PFOS not degraded with an aggressive treatment.

Ultraviolet reactions◦ Successful for only PFOA◦ Conditions not applicable in situ

Permanganate ◦ some success with PFOS at ≥ 65 ºC & low pH (Liu et al, 2012)

Sonolysis◦ Good results for PFOA and PFOS◦ Not applicable for in situ

Persulfate◦ UV and heat activation established successful for PFOA (Hori et

al., 2005; Lee et al, 2012; Liu et al., 2013)

Initial Approach: Chemical Oxidative Treatment

*Choice of activation mode dependent on effectiveness and efficiency for a given environment, e.g., in-situ versus ex-situ.

-↔

PFOA/PFOA-

PFOS-

6:2 FTS (6:2 Fluorotelomer sulfonate)

Temperature (20 ~ 60 °C)

Unbuffered pH

Persulfate (Na2S2O8) Conc. (0 ~ 20,000 mg/L, 84.00 mM)

PFACs conc. (50 ~ 2,500 µg/L, 0.12 µM ~ 6 µM)

Fuel co-contaminants (BTEXs up to 1,000 µg/L)

Aquifer

Experimental Variables

Heat-activated Persulfate for 3 Representative PFASs

Varying Temperature[Na2S2O8 ] =- 10,000 mg/L Varying [Na2S2O8 ], T = 50 ºC

Heat-Activated Persulfate(100 µg/L PFOA)

Removal rate ↑ with ↑ temperature and ↑ [Na2S2O8] Linear Arrhenius Plots (right graph) Unbuffered pH (pH decreases with reaction) Same for PFCA range tested: 50 ~ 2,500 µg/L (0.12 µM ~ 6 µM)

0.0

0.2

0.4

0.6

0.8

1.0

0 50 100 150 200 250 300 350

Ct/C

o

Time (h)

20 C 30 C 40 C 50 C

-8

-6

-4

-2

0.0028 0.003 0.0032 0.0034 0.0036

ln (k

1, h-

1 )

1/(T, K)

This Study (unadjusted pH)Lee et al. (2012) pHo=2.5Lee et al. (2012) pHo=7.1Lee et al. (2012) pHo=11Liu et al. (2012) pHo=7.1

0.00

0.05

0.10

0.15

0.20

0.25

0 10 20 30 40 50 60 70 80 90

-k1

(h-1

)

Na2S2O8 (mM)

~100 µg/L M) (0.241 µM) PFOA ~100 µg/L (0.215 µM) 6:2 FTS

Heat (50 ºC) - Activated Persulfate[Na2S2O8] – 10,000 mg/L, 50 ºC

PFOA and 6:2 FTS unzipped to smaller PFCAs Eventually complete abiotic mineralization (F- and CO2) Co-contaminants (@ 1000 mg/L) - PFAS degradation NOT

affected

• PFOS unaltered at 90 C with 60.5 mM Na2S2O8 as well as at 90 C with 84 mM Na2S2O8

• PFOA & PFOS will coexist in groundwater, thus technologies that can remediate both are needed

Heat-Activated PersulfatePFOS (100 µg/L)

0.0

0.2

0.4

0.6

0.8

1.0

0 20 40 60 80 100

PFO

S (µ

M)

Time (h)

60.5 mM NaS2O8No Oxidant Contorl

Persulfate Oxidation - Yang et al. (2013)100,000 µg/L PFOS (0.2 mM)

• High PFOS concentration (0.2 mM PFOS = 3.4 mM Fluoride)• High persulfate concentration and no controls presented• Hydrothermal (HT) – high temperature and pressure• Other literature: Kingshott (2008) Thesis – probing using reactant

concentrations approaching solubility limits

doi:10.1371/journal.pone.0074877.g006

Accurately quantify target parent PFAS (and isomers)◦ High extraction efficiencies of all phases◦ Reliable chromatographic separation and detection

Accurately quantify fluoride and sulfonate metabolites◦ Above background noise◦ Optimize nanocomposite synthesis to minimize background ◦ Minimize analytical matrix effects ◦ High extraction efficiencies and appropriate analytical methods

Identify organic metabolites and quantify if possible◦ TOF, MS/MS probing of aqueous and extractants◦ Headspace sampling (requires airtight reaction vessel)

Analytical methods are treatment technique dependent

Method DevelopmentApproach for successful exploration requires:

UV (254 nm) photolysis- generated aquated electrons (eaq-) from

iodide (Eo= -2.9 V) (Park et al., 2009) ◦ PFOS degraded faster than PFOA (under argon atmosphere)◦ Not applicable for in situ

Vitamin B12 (Ochoa-Herrera et al., 2008)

◦ Successful for only branched PFOS isomers, but not linear chain Boron Doped Diamond Electrode (Carter and Ferrell, 2008)

◦ Some success for PFOS◦ Very expensive ◦ Not applicable for in-situ

Zero Valent Iron (ZVI) (Eo = -0.447 V) (Hori et al., 2005; 2008)

◦ Degraded PFOS at high T & high pressure (→ near supercritical) Electrochemical (Schaefer et al., 2015, J Hax. Materials)

Theoretical and experimental evidence: PFOS is amenable to reduction. Need: Enhance reactivity and delivery

Reductive Approaches: Literature

What extent of reductive transformation especially of linear PFOS can be achieved with the primary reductive techniques proposed?

What are the intermediates of PFCAs and PFSAs defluorination using the proposed reductive techniques?

How is reductive transformation affected by various solution parameters (pH, electrolyte composition, redox potential) and co-contaminants likely to be associated with AFFF-contaminated groundwater at military sites?

How does the presence of porous media affect defluorination potential and reductant reactivity?

How amenable are intermediates to other remediation processes (treatment train potential)?

Our Current Research Questions

Metals◦ nZVI – not successful by itself◦ nZVI/Pd as a function of % Pd, initial pH, temperature◦ nZVI/Pd/Al◦ nZVI/Ni with GO or GAC◦ Mg/Pd

Enhanced delivery◦ Fe3O4 composites◦ Deposit reactive metals within clay layers◦ Organo-bentonite

Vitamin B12 with◦ Ti-citrate◦ Cu◦ n-ZVI◦ n-ZVI/Pd

Reductive Approaches Probed to Date

→F- + SO3

2- + per/poly-fluoroalkanes +Poly-fluoroalkyl sulfonates

Any sulfite generated expected to rapidly convert to sulfate:

Consistent with Literature: Sulfite oxidation rate = 0.009 /h (Shizuo Tsunogai , 1971)

Expected Inorganic Metabolites

PFOS Isomers in Technical Grade PFOS & Formuations

Linear 65-75% 4-PFOS

5-PFOS

6-PFOS

dm-PFOS3-PFOS

1-PFOS

2-PFOS

HPLCZorbac C8

Isomer Separation and Response Optimization

• Standard analyses – all isomers elute in one peak• Ascentis F5 allows for isomer separation• Selexion allows for enhanced response to each isomer

ZVI is a potent reductant (Eo = -0.447 eV) Nano-sized ZVI (nZVI) - high surface area (40~60 m2/g) Cost-effective (low installation and operation cost) Low environmental impact and low toxicity nZVI can be doped with a catalyst such as Pd, Pt, Cu, Ni and Ag to

improve the reactivity. Successful for dehalogenation of chlorinated alkenes, alkanes, and

PCBs

Mechanisms of contaminants (e.g. chlorinated solvent) removal

a) Direct electron transfer b) Hydro-defluorinationonto the catalyst surface

c) Hydro-defluorinationonto the nZVI surfaceGunawardana et al, 2011

Why nZVI Systems?

Effect of %Pd and Mass of Bimetal

• For Pd/Fe ≥ 0.043, no significant difference on PFOS removal.

• [PFOSo] = 1,487 µg/L, unbuffered pH, T=25°C for 6 d (% Pd/Fe g).

2331 32 29

0

5

10

15

20

25

30

35

40

1 2 3 4

PFO

S re

mov

al %

0.014(0.43%/0.3 g)

0.043(0.43%/0.1 g)

0.05(1.0%/0.2 g)

0.175(1.75%/0.1 g)

a a a

b

• [PFOS]o = 1,570 µg/L• No pH adjustment• Initial pH = 5.57• Final pH after reaction:

→ 6.53, 6.45, 4.99, 6.26• T= 22o C for 6 d

Pd/nZVI wt/wt ratio

Effect of Initial pH with Pd/nZVI

• Acidic and basic initial pH had highest and similar PFOS removal.• Removal lowest at near neutral pH condition.

• [PFOS]o = 1,570 µg/L • 1% Pd/0.2 g nZVI • pH adjusted with HCl or NaOH• T= 22 °C for 6 d37

3239

05

1015202530354045

1 2 3

PFO

S R

emov

al %

[pH]o=3.6[pH]f=4.26

[pH]o=6.3[pH]f=4.99

[pH]o=12.5[pH]f=12.4

aa

b

Temperature Effect with Pd(1%)/nZVI (0.2 g)

Role of temperature not clear

22 °C 50 °C 70 °C

T-P

FOS

rem

oval

%

0

10

20

30

40

50

a

a*a

32 37 35

T (°C) [PFOS]o(µg/L)

Initial pH

Final pH

22 1,570 6.29 4.99

50 1,570 6.29 4.15

70 3, 460 5.62 7.29

~ 0.4 mol F- generated per mole PFOS lost~ 0.3 mol SO4

2- per mole PFOS lost

• [PFOS]o = 880 µg/L • 1% Pd/0.5g of Fe/Al • pH adjusted with HCl/NaOH• Base added at 0, 24, and 48 h• T= 22 °C• Reacted for 6 d

pH Effect with Pd/Fe/Al Trimetal

36 32

14

28 27

0

5

10

15

20

25

30

35

40

45

Acid-PFOS Neutral-PFOS Base-0-PFOS Base-24-PFOS Base-48-PFOS

Rem

oval

%

aa,b

bb

c

% P

FOS

Rem

oved

• Initial high pH slowed/reduced PFOS removal likely due to Al(OH)4‒ formation

• Reaction appeared complete by 24 h

• Ni content over Fe0 was 1.8 wt%. • Batch experiments were carried out in

triplicates for: Ni/FeNi/Fe-Graphene (NiFe-GO)Ni/Fe-Activated Carbon (NiFe-AC)

• 0.2 gr was added to each replicate and matrix. Both were mixed for 3 days.

• A blank containing 10 mL of PFOS (2 ppm) was carrying out together with all the samples.

• 4 extractions using ethyl acetate.

• pH shifts during reaction:

Method

ID NiFe NiFe-GO NiFe-ACMatrix Day 0 7.21 7.71 7.70Matrix Day 3 8.62 8.90 8.65

Day 0 7.39 7.17 7.95Day 3 10.18 10.61 10.28

Ni-nZVI

ID-Extract#% relative to

Sum recoveredNiFe-AC-E1 86NiFe-AC-E2 9.7NiFe-AC-E3 2.8NiFe-AC-E4 1.6

0%

10%

20%

30%

40%

50%

60%

NiFe NiFe-GO NiFe-AC

%R

emov

al

• 51% to 57% PFOS not recovered so assume ‘transformed’• Four extractions was found sufficient for PFOS from nZVI

particles• Apparent higher loss for Ni-nZVI on activated carbon may be

extraction efficiency issue – now using 5 extractions

baa

Ni-nZVI

• More associated with NiFe-AC particles (~85%).• Significant differences were observed on NiFe-AC compared

to NiFe (P < 0.0001) and NiFe-GO (P < 0.0001)

% Extracted from Ni/nZVI particlesversus in Aqueous Phase

0%

20%

40%

60%

80%

100%

NiFe NiFe-GO NiFe-AC

Rel

ativ

e %

on

nano

-par

ticle

sb

aa

9.2 9.3 9.4 9.5 9.6 9.7 9.8 9.9 10.0 10.1 10.2 10.3 10.4 10.5 10.6 10.7 10.8 10.9 11.0 11.1 11.2 11.3 11.4 11.5 11.6

Time, min

0.0e0

1.0e4

2.0e4

3.0e4

4.0e4

5.0e4

6.0e4

7.0e4

8.0e4

9.0e4

1.0e5

1.1e5

1.2e5

1.3e5

1.4e5

Inte

nsi

ty

11.12

10.70

10.42

10.149.98

TIC from NiFe-AC-R1.wiff (sample 1) - NiFe-AC-R1, Experiment 3, -TOF MS^2 of 499.0 (50 - 600)

TIC from NiFe-Stock.wiff (sample 1) - NiFe-Stock, Experiment 3, -TOF MS^2 of 499.0 (50 - 600)

TIC from NiFe-R1.wiff (sample 1) - NiFe-R1, Experiment 3, -TOF MS^2 of 499.0 (50 - 600)

TIC from NiFe-GO-R1.wiff (sample 1) - NiFe-GO-R1, Experiment 3, -TOF MS^2 of 499.0 (50 - 600) L-PFOS

6-PFOS

1-PFOS

0%

25%

50%

75%

NiFe NiFe-GO NiFe-AC

% R

emov

al

Branched PFOS

0%

20%

40%

60%

NiFe NiFe-GO NiFe-AC

% R

emov

al L-PFOS

Isomer-specific Removal

• Vitamin B12 (cyanocobalamin) is produced by anaerobic bacteria contains a corrin ring that coordinates a cobalt atom• Cobalt center of V-B12 can exist in 3 oxidation states (B12a, B12r, B12s) • Electrons transferred from reduced metal (Ti exemplified below but also assessed with other metals) to the contaminant

Reduced Ti (III) Citrate

Oxidized Ti (IV) Citrate

2e–

B12a↓

B12r↓

B12S

2e–

Some degradation of branched PFOS, not linear PFOS

70 C, pH=9 (Ochoa et al., 2008)

Unknown intermediates +

FluorideMost nucleophilic

Why Vitamin B12 Systems?

Vitamin B12 (0.4 mM) + n-ZVI (0.2 g)

• Similar PFOS removal in both systems• Inorganic metabolite production varies• Organic metabolites also appear different

16.12

0.921.05 0.3102468

1012141618

1 2m

ol/m

ol

B12-FePd-Fe

F– or SO42- / PFOS removed

(Comparing 2 systems)

Fluoride SO42-

Effect of Temperature and Initial pH for B12+ Ti System

• Heat enhanced only PFOA removal• At T=70 °C, PFOS removal increased with increase in initial pH

17

43

16 16

32

0

5

10

15

20

25

30

35

40

45

50

1 2 3

Rem

oval

%

PFOA PFOS

At 22 °C, pH not adjusted

At 70 °C, pH not adjusted

At 70 °C, at pH 9.54

• [PFOS]o= 4,655 µg/L • [B12]= 0.4 mM• [Ti-citrate]= 45 mM• pH adjusted with NaOH• T= 70 °C for 5 d

• [PFOA]o= 609 µg/L• [PFOS]o= 358 µg/L • [B12]= 0.5 mM• [Ti-citrate]= 35 mM• Unadjusted pH• T= 22 or 70 °C for 5 d

B12-Fe B12-Fe-Pd B12-Ti-Cu B12-Ti-Fe B12-Ti

T-P

FOS

rem

oval

%

0

10

20

30

40

50

22

9

2334 32

Effect of Cu°, Pd ° or Fe° in B12 - Ti system[PFOS]o = 4,655 µg/L, T = 70 °C, Reaction = 5 d

Treatment [pH]o [pH]f

B12-Fe 10.41 8.03B12-Pd/Fe 10.41 7.61B12-Ti-Cu 9.54 6.86B12-Ti-Fe 9.54 9.23

B12-Ti 9.54 6.03

• Addition of Fe° (as an electron donor) to B12+Ti systemt did not enhance PFOS removal.• Addition of Cu° or Pd°(as a catalyst) to B12-Ti-citrate system decreased PFOS removal.• In Ti-system, high Cl– concentration interfered with F– quantitation. • B12+Fe system, F- and SO4

2- were detected indicating partial reduction.

10-14 mol F- generated per mole PFOS lost~ 0.7 mol SO4

2- per mole PFOS lost

B12 + Fe Fe/Pd Ti-Cu Ti-Fe Ti

22°C at 21 day70°C at 21 day

70°C at 5 day

T-P

FOS

rem

oval

%

0

10

20

30

40

Col 5: 14.7 Col 5: 18.27

B12-Fe

B12-PFOS (NO Fe)

Effect of Temp. in B12 (0.4 mM) + nZVI (0.2 g) System

• PFOS removal increased with temperature .• At 70 °C, PFOS removal continued to increase from 5 to 21 days.• Most degraded PFOS is branch-PFOS.• PFOS removal in only Vitamin B12 (no metals to enhance electron shuttle) unexpected.

PFOS isomers Chromatographs in B12-Fe at 70 °C (in Aqueous phase)→ Stock, D-14, and D-21

Sample ID [pH]o [pH]f

22 °C at 21 d 10.51 9.8970 °C at 21 d 10.51 6.6770 °C at 5 d 10.41 8.03B12-PFOS

(no Fe)10.51 9.14 (22 °C)

7.52 (70 °C)

20

15

32

1622

Isomer-Specific Peak Comparison for B12-nZVI

L-PFOS 22C70C

D-5 D-7 D-11 D-14 D-21Stock D-1 D-2 D-31 2 3 4 5 6 7 8 9

6-PFOS 22C70C

D-5 D-7 D-11 D-14 D-21Stock D-1 D-2 D-3

1 2 3 4 5 6 7 8 9

5-PFOS 22C70C

D-5 D-7 D-11 D-14 D-21Stock D-1 D-2 D-31 2 3 4 5 6 7 8 9

3&4-PFOS 22C

70C

D-5 D-7 D-11 D-14 D-21Stock D-1 D-2 D-3

1 2 3 4 5 6 7 8 9

Other isomers 22C

70C

D-5 D-7 D-11 D-14 D-21Stock D-1 D-2 D-3

TIME (d)

REL

ATIV

E to

STO

CK

• LC/MS-ESI/TOF (and GC/MS Orbitrap forthcoming)• Several metabolites produced not in controls• Expect some perfluoroalkyl alkane/alkenes• Currently using a mass defect approach to identify metabolite• Example: Peaks in Ni/nZVI Treatment Systems:

• Positive Mode• 4 similar peaks observed in all systems• One additional peak in graphene oxide (GO) system

• Negative Mode• 1 peak observed in all• Peak was different in activated carbon (AC) system

Organic Metabolitess

Many combinations have been used in field applications◦ ZVI-type treatment zone coupled to natural or enhanced anaerobic

treatment zone◦ ISCO treatment zone coupled to enhanced or natural aerobic

degradation◦ Aerobic degradation treatment zone coupled to an anaerobic

abiotic (ZVI) or bioremediation (biowall) treatment zone One treatment can ‘soften’ a compound to be amenable to natural

anaerobic or aerobic degradation and/or to a subsequent chemical treatment.

At many of the sites where PFASs exist, redox conditions may already be depleted in O2 (redox range-100 to -250 mV) due to degradation of fuel and other easily biodegradable contaminants.

Order and placement of treatment train for PFASs depends on proximity to source and contaminants present – precursor oxidation will lead to increased concentration of PFAAs.

NRC (2013); IRTC (2011)

Treatment Train Potential

Used for Cl-methanes, ethanes, and ethenes, chlorinated phenols, PCBs, hexachlorocyclohexanes, munitions, metals, nitrates, perchlorate

Permeable Reactive Barriers (PRBs) Reductions in the target zone of typically 50 to 90 percent ◦ Treatment Trains◦ In-situ Injection

Barrier Media Superfund Industrial US Gov. International TotalZVI 5 36 26 16 83Non-Fe reactive materials 1 6 1 1 9Bio-barrier 0 6 5 4 15Combo/Sequenced (* ZVI) 0 4 2 0 6 (5*)

NRC (2013); IRTC (2011)

Enhancing ZVI-based techniques Nano-ZVI (nZVI) - pilot & full-scale tests with nZVI1 shows 50-90%

conc. reduction Surface coated/encapsulated nZVI-type particles (enhanced for in-situ

injection) Carbon/ZVI, C-rich organic/iron combinations (PRBs) Emulsified ZVI, ZVI-clays, organophilic-ZVI clays (PRBs, in-situ

injection)

Current Use of Reductive Technologies in the Field

Take Home Messages

No magic bullet! PFOS is the big challenge - more about steric hindrances Reductive approaches showing some promise for ‘softening’

PFOS (partial transformation) Linear PFOS remains the primary challenge – energetically

the most stable and ~75% of what is in the environment Many published studies lack sufficient recovery efficiencies

and background issues We are not ready for any field-scale attempts for in-situ

remediation due to◦ Increasing PFAAs due to aerobic transformation of

precursors – increased mobility◦ Unknown metabolites◦ Toxicity potential unknown or unclear for metabolites

Co-investigatorsDr. Linda S. Lee (PI)

Purdue UniversityEnvironmental Chemist; per/polyfluoro alkyl chemical fate expertise

Dr. Loring NiesPurdue UniversityEnvironmental engineering, biodegradation, nanoparticle

Dr. Victor MedinaU.S. Army Engineering Research & Development CenterChemical/metal remediation experience with military

Drs. Marc Mills & Kavitha Dasu*U.S. EPA, NRMRLFate/remediation expertise; * PFAS fate/ analytical expertise (with PI)

Dr. Hongtao YuJackson State UniversityChemistry/biochemistry; NSF-funded NMR Center

ER-2426

Qualityof Life

Our challenge