Supporting Information

Sorption and Desorption of Anionic, Cationic and Zwitterionic

Polyfluoroalkyl Substances by Soil Organic Matter and Pyrogenic

Carbonaceous Materials

Yue Zhia, Jinxia Liua,*

a Department of Civil Engineering, McGill University, Montréal, Québec, H3A 0C3, Canada

*Corresponding author. Tel: +1 514 398 7938; fax: +1 514 398 7361

e-mail address: jinxia.liu@mcgill.ca

1

Table of content1. LC-MS/MS analytical method................................................................................................................32. Analysis of the oil residue extracted from the soot................................................................................4

3. Pore size distribution of biochar and oil-free soot.................................................................................64. Results of transmission electron microscopy (TEM).............................................................................7

5. Sorption kinetics....................................................................................................................................86. Mass balance test...............................................................................................................................12

7. Sorption isotherms of PFASs to the soot and oil-free soot..................................................................148. Speciation of PrePFAAs......................................................................................................................16

9. Pair-wise comparison of solid-water partitioning coefficients (Kd, L g-1) of different sorbents.............1910. A Hypothesis test to confirm the difference between sorption and desorption isotherm branches...20

11. References........................................................................................................................................22

2

1. LC-MS/MS analytical method

Details on the LC-MS/MS analytical method are described in Tables S1

Table S1 Quantitative LC-MS/MS analytical method

Instrument Shimadzu Nexera UHPLC coupled to an AB Sciex 5500 QTrap mass spectrometer

Ionization Positive and negative electrospray

Acquisition mode Multiple reaction monitoring (scheduled MRM)

Analytical column Agilent Zorbax SB-C8, 3.5 μm, 2.1 x 100 mm

Delay column Kinetex EVO C18, 5 μm, 50 x 3.1 mm

Column Temperature 40°C

Mobile PhasesA: 0.15% acetic acid in LC-MS water

B: 0.15% acetic acid in acetonitrile

Gradient Profile Time (min) Percentage B Flow Rate (mL/min)

1.0 5 0.5

7.0 95 0.5

9.0 95 0.5

9.1 5 0.5

11.5 Stop

Injection Volume 5 μL

Calibration Linear calibration curve, 1/x weighted

Table S2 Monitored quantitation transitions for quantitative analytical methods

Analyte Transition Monitored Internal standard IS transition monitored

PFOA 413>369 MPFOA 417>372

PFOS 499>80 MPFOS 503>80

PFOAB 555>378 MPFOA 417>372

6:2 FTAB 569>120 M6:2FTSA 429>81

PFOAAmS 513>454

3

2. Analysis of the oil residue extracted from the soot

The oil residue of the acid-washed soot was extracted in an automatic Soxtherm extraction

apparatus (Gerhardt Soxtherm, UK) and the extracted oil was quantified using the Quebec

method (MA400HYD11) for total petroleum hydrocarbon (TPH, C10-C50) analysis. Briefly, a

soot sample (3 g) was placed in a cellulose extraction thimble. In the automatic Soxtherm

extraction apparatus (Gerhardt Soxtherm, UK), 150 mL of the extraction solvent comprised of

50:50 (vol:vol) hexane: acetone was added to a set of extraction beakers containing boiling

stones. The extraction proceeded with an optimized extraction program, for which the recovery

for an oil reference standard (weathered diesel, purchased from Restek Corporation, USA)

was 98.5 ±1.5%. The solvent extracts were concentrated by nitrogen blow-down to 10 mL and

then cleaned up with 0.8 g of silica gel. The supernatant was pipetted out for analysis. TPH

was analyzed by a gas chromatograph equipped with flame ionization detector. The TPH

concentration was determined based on the integration of C10-C50 peaks after subtraction of

a blank solvent run.

4

Table S3 Quantitative analytical method for total petroleum hydrocarbons (TPH, C10-C50)

Instrument Agilent 7820A GC-FID

Column Model Agilent 125-1011, DB1 capillary column

Carrier gas Helium, 5mL/min

Oven Rate(°C /min) Temperature (°C) Hold (min)40 0.25

30 300 010 340 7

ALS Injection volume 1 μL

Solvent wash (PreInj) 3

Solvent wash (PostInj) 3

Sample wash 1

Sample pump 4

Viscosity delay 1 sec

Inlet COC inlet, oven track

Detector Heater 360°C

H2 flow 45mL/min

Airflow 450mL/min

Makeup flow 45mL/min

5

Figure S1. GC chromatographs of a diesel standard, a motor oil standard and the oil residue extracted from the acid-washed soot.

3. Pore size distribution of biochar and oil-free soot

Figure S2 Pore size distribution of the biochar and oil-free soot samples

6

4. Results of transmission electron microscopy (TEM)

All the adsorbents were imaged by Transmission electron microscopy (TEM) (Tecnai G2 F20,

OR, USA) under a high-resolution mode using a carbon coated copper micro-grid as substrate.

All three adsorbents have both glassy and amorphous domains. The glassy domain has

distinct characters compared with the amorphous domain: 1) Glassy domains feature clear

pattern of carbon atom columns (e.g., left side on Figure c); 2) As shown in Figure b, with

lower magnification, the black particle has a clear interface with surrounding the flocculent

amorphous area. Besides, the degree of contrast of the black particles significant changed with

rotation of the sample, while no change was observed in the amorphous area; 3) There are

bright visible diffraction spots in the micro-diffraction imaging at the glassy domain, which is

evident of the crystallographic characteristics; 4) The glassy domains (e.g., the black particle in

Figure b) showed lattice contrast under the proper diffraction conditions. It is noted that the

sorbents are not homogenous, and, therefore, we scanned the material first and picked out the

spots that can best illustrate the physical state of the matrix. More technique would be needed

to quantify the actual percentage of glassy vs. amorphous domains.

Figure S3 Morphology of peat, biochar soot and oil soot imaged by TEM.

(c) Biochar(a) Peat

Amorphous domain

Glassy

(b) Peat (d) Biochar

(e) Soot

7

5. Sorption kinetics

Sorption and desorption kinetics were determined using single-solute batch experiments on an

orbital shaker (Multitron Pro, Infors HT) at 20 °C and 250 RPM. Details of the kinetics

experiment were given in the main text. Figure S4 1-2 shows the results of individual sorption

and desorption kinetics of five PFASs on peat and biochar. The equilibration time for the

kinetic study was chosen to be 60 h, and the aqueous phase was replaced with sorbate-free

mobile phase 60 h. All sorbent concentrations experienced dramatic increase at the beginning

and then followed by a slow increase.

The adsorption uptake data were modeled with a pseudo-second-order kinetic model shown in

Equation (S1):

tCS

= 1kC∞

2 + tC∞

= 1v0

+ tC∞

(Eq. S1)

where, C s and C∞ are the PFAS adsorbent concentration (µg g-1 ) at time t (h) and equilibrium,

respectively, v0 is the initial adsorption rate (µg g-1 h-1), and k the rate constant (g µg-1 h-1) . The

parameters from the best fit obtained by non-linear least squares regression to the model were

listed in the Table S4. According to the model, more than >95.7 % of the final equilibrium

uptake would be achieved within 60 h, and 99.99% of the equilibrium uptake would reach in 7

days, which is the time for sorption isotherm study.

Table S4 Parameters of the pseudo-second-order kinetic model

vo C∞ k r2 C s/C∞,t = 60 hrs

µg g-1 h-1 µg g-1 g µg-1 h-1

PeatPFOS 65.4 65.4 0.02 0.99

9 98.4%

PFOA 39.2 31.0 0.04 0.999 98.7%

PFOAB 35.0 66.7 0.01 0.999 96.9%

6:2 FTAB 111 62.5 0.03 0.999 99.1%

PFOAAmS 24.1 64.9 0.01 0.994 95.7%

8

Biochar

PFOS 2500 106 0.22 0.999 99.9%

PFOA 256 97.1 0.03 0.999 99.4%

PFOAB 10000 125 0.64 0.999 100.0%

6:2 FTAB 241 114 0.02 0.999 99.2%

PFOAAmS 2000 89.3 0.25 0.999 99.9%

Overall, biochar exhibits faster sorption kinetics and higher equilibrium solid phase

concentration than peat. Both sorbents were ground and sieved through a 250-µm sieve by

following the same protocol, the fraction less than 250-µm was collected for sorption

experiments. Therefore, both peat and biochar possess similar bulk particle size and

geometry. The difference in sorption rate is attributed to the nature of the sorbent.

9

Figure S4-1. Sorption and desorption kinetics profiles of PFOS, PFOA and PFOAB for peat and biochar; the aqueous phase was changed at 60 h.

10

Figure S5-2. Sorption and desorption kinetics profiles of 6:2 FTAB and PFOAAmS for peat and biochar; the aqueous phase was changed at 60 h.

11

6. Mass balance test

After desorption, tubes with an initial concentration of 1000 µg L -1 were checked for mass

balance for five PFASs individually. The supernatant of each tube was discarded after 20 min

of centrifugation at 3000 g. Peat and biochar were subjected to extraction by 15 ml of

acetonitrile (ACN) with 25 mM NaOH (25 mM). Sonication (30 min) was followed by 2 hrs of

vigorous stirring. Extraction was repeated three times: in the second and third times, only 3 ml

of ACN with 25 mM NaOH was added to each tube. Recovery of PFASs with peat and biochar

is illustrated in Figure S5.

Figure S5. Mass balance test of PFASs from biochar and peat

As shown in Figure S5, the recovery fractions of PFASs are between 0.76-1.34 for peat and

0.70-1.13 for biochar. For PrePFAAs, possible transformation compounds (e.g., PFOS, PFOA

or other short-chain PFAAs) were also checked in solvent extracts, but none was found,

indicating that biotic and abiotic transformations were negligible. Good recovery is the

precondition to precisely quantify concentrations of PFAS and PrePFAAs in the various

matrices. Among PFASs investigated, PFOS and PFOAB have the best recovery (among

0.92-1.10) while 6:2 FTAB has the lowest recovery (0.76 to peat, and 0.73 to biochar). Mass

recovery of PFASs from the peat is higher than that of biochar except for PFOS. According to

US EPA method, a recovery fraction of 0.70-1.30 is considered satisfactory.

12

One factor that may contribute to the deviation of recovery from unity is the accumulative

errors during multiple handling steps in the sorption-desorption loop. The discrepancy of

recovery could also be attributed to sorbent properties. Sorption of PFASs to biochar is

described as adsorption, whereas partitioning or dissolution better describes the sorption to

peat [1]. Extraction of PFASs from peat is hypothesized as a re-partitioning process between

rubbery organic matters and organic solvent while extraction of PFASs from biochar could be

more difficult. PFASs that have travelled to the deep micropores in the biochar by diffusion

may be resistant to solvent extraction. Furthermore, the recovery test was performed after the

whole sorption-desorption loop 14 days later. The sorbent materials could be conditioned by

PFAS molecules to lead to irreversible deformation of the sorbent matrix. Pores may be closed

afterward and PFAS thus entrapped are unavailable for extraction [2].

13

7. Sorption isotherms of PFASs to the soot and oil-free soot

Figure S6 Sorption isotherms (at 20 ºC) of PFAS to the soot and the oil-free soot.

14

Table S5 Summary of literature logKd and logKoc values for sorption of PFOS and PFOA

by natural sorbents.

Sorbent Sorbate logKd

(L kg-1)

foc of sorbents (%)

log Koc

(L kgoc-1 ) Reference

Soil with black carbon amendment PFOS 1.78 – 3.11 [3]

Freshwater sediments 1 PFOS 3.94 – 4.26a 0.75 2.9 – 3.2 [4]Freshwater sediments 2 PFOS 1.11b 0.56 – 9.66 2.7 [5]Freshwater sediments 3 PFOS 0.27 – 1.37c 0.16 – 1.49 3.0 – 3.2 [6]Freshwater sediments 4 PFOS 0.03 – 1.6 3.4 – 3.7 [7]Freshwater sediments 5 PFOS 2.40 4.80 3.7 [8]Loam PFOS 2.62 – 15.6b 0.80 – 1.70 2.8 [9]

Soil with black carbon amendment PFOA ~ 1.56 [3]

Freshwater sediments 2 PFOA 0.56 – 9.66 2.1 [5]Freshwater sediments 4 PFOA 0.03 – 1.60 2.3 – 2.5 [7]Loam PFOA 0.63 – 3.50b 0.80 – 1.70 1.9 [9]

8. Speciation of PrePFAAs

6:2 FTAB speciationSMILES: O=S(CCC(F)(F)C(F)(F)C(F)(F)C(F)(F)C(F)(F)C(F)(F)F)(NCCC[N+](C)(C)CC(O)=O)=O

Species:

S1 S

F F

FF

F F

FF

F F

FF

FHN N+

O

OO

OH

S2 S

F F

FF

F F

FF

F F

FF

FHN N+

O

OO

O-

S3 S

F F

FF

F F

FF

F F

FF

F N- N+

O

OO

O-

S4 S

F F

FF

F F

FF

F F

FF

F N- N+

O

OO

OH

S

F F

FF

F F

FF

F F

FF

FHN N+

O

OO

OH ⇌ S

F F

FF

F F

FF

F F

FF

FHN N+

O

OO

O-

pKa1 = 2.26

S

F F

FF

F F

FF

F F

FF

FHN N+

O

OO

O- ⇌ S

F F

FF

F F

FF

F F

FF

F N- N+

O

OO

O-

pKa2 = 11.12

Figure S7-1. Speciation plot of 6:2 FTAB predicted by SPARC (ARCHem, GA, USA)

PFOAB speciationSMILES: FC(F)(F)C(F)(F)C(F)(F)C(F)(F)C(F)(F)C(F)(F)C(F)(F)C(NCCC[N+](C)(C)CC(O)=O)=O

Species:

S1 F

F F

FF

F F

FF

F F

FF

F FHN N+

O

O

OH

S2 F

F F

FF

F F

FF

F F

FF

F FHN N+

O

O

O-

S3 F

F F

FF

F F

FF

F F

FF

F F

N- N+

O

O

O-

S4 F

F F

FF

F F

FF

F F

FF

F F

N- N+

O

O

OH

F

F F

FF

F F

FF

F F

FF

F FHN N+

O

O

OH ⇌ F

F F

FF

F F

FF

F F

FF

F FHN N+

O

O

O-

pKa1 = 2.25

F

F F

FF

F F

FF

F F

FF

F FHN N+

O

O

O- ⇌ F

F F

FF

F F

FF

F F

FF

F F

N- N+

O

O

O-

16

pKa2 = 7.79

Figure S7-2. Speciation plot of PFOAB predicted by SPARC (ARCHem, GA, USA)

PFOAAmS speciationSpecies

S1

S2

⇌pKa = 7.71.

17

Figure S7-3. Speciation plot of PFOAAmS predicted by SPARC (ARCHem, GA, USA)

18

9. Pair-wise comparison of solid-water partitioning coefficients (Kd, L g-1) of different sorbents

Figure S8. Pair-wise comparison of solid-water distribution coefficients (Kd, L g-1) of different sorbents, where the values of Kd are calculated at Cw = 0.5, 2, 10 and 100 µg L-1.

Pearson's r is a measure of the linear dependence, the closer the Pearson's r is to ± 1,

the stronger correlation there is. A Linear dependence was noticed between PCM

sorbents. Correlations were found between Kd of soot and oil-free soot with Pearson's r

ranging from 0.893 - 0.812 at Cw = 0.5 - 10 µg·L-1 and 0.654 at Cw = 100 µg·L-1 (with 95

% confidence level), suggesting oil residues did not fundamentally alter the nature of

sorption to soot. Likewise, r for biochar and oil-free soot is 0.743 - 0.919 at Cw = 2 - 100

µg·L-1, indicating sorption behaviors of two different types of PCMs are correlated and

similar. Adversely, no correlation was concluded with peat to any PCMs. In fact, for

some PFOA, PFOAB and 6:2 FTAB, Kd values show roughly negative correlations,

which is due to high sorption nonlinearity of the PCMs.

(d)

(c)

(a)

19

10. A Hypothesis test to confirm the difference between sorption and desorption isotherm branches

Difference between sorption and desorption isotherm branches is represented by:d i=C s

d−C ss|T ,Cw❑ (Eq. S2)

Where C sd and C s

s are are solid-phase solute concentrations for the single-cycle sorption and desorption experiments, respectively, at Cw and temperature of T. If di = 0, C s

d=C ss,

sorption and desorption isotherm branches are identical and overlap, the sorption is reversible, without hysteresis effect. Null hypothesis is di = 0, and the analysis was performed with IBM SPSS Statistics 24.

Table S6. Results of the one-sample statistics analysis

Compounds Number of data point

Mean of di

Std. Deviation of

di

Significance, P (95% confident

interval)

Peat

PFOS 21 -0.0037 0.00409 0.000PFOA 21 -0.10 0.005 0.000

PFOAB 21 -0.10 0.019 0.0006:2 FTAB 21 0.00 0.015 0.040

PFOAAmS 21 -0.02 0.012 0.000

Biochar

PFOS 21 -1.01 0.284 0.000PFOA 21 -0.04 0.009 0.000

PFOAB 21 -2.54 1.13 0.0006:2 FTAB 21 -11.25 19.3 0.015

PFOAAmS 21 -2.57 0.292 0.000

Note that when P < 0.05, reject the null hypothesis, di ≠ 0, sorption and desorption isotherms are different at 95% confident interval. Otherwise, accept the null hypothesis.

20

PFOS

PFOA

PFOAB

6:2 FTAB

PFOAAmS

PFOS

PFOA

PFOAB

6:2 FTAB

PFOAAmS

0.00

0.01

0.10

1.00

10.00

100.00

-di

(µg

g-1

)

Figure S9. Mean Difference of di, which denotes the mean distance between sorption and desorption isotherm branches ¿¿ ) at the same aqueous solute concentration (Cw).

Peat Biochar

21

11. References

[1] B. Xing, N. Senesi, P.M. Huang, Biophysico-Chemical Processes of Anthropogenic Organic Compounds in Environmental Systems, 2011. doi:10.1002/9780470944479.

[2] W.J. Braida, J.J. Pignatello, Y. Lu, P.I. Ravikovitch, A. V. Neimark, B. Xing, Sorption hysteresis of benzene in charcoal particles, Environ. Sci. Technol. 37 (2003) 409–417. doi:10.1021/es020660z.

[3] D. Kupryianchyk, S.E. Hale, G.D. Breedveld, G. Cornelissen, Treatment of sites contaminated with perfluorinated compounds using biochar amendment, Chemosphere. 142 (2016) 35–40. doi:10.1016/j.chemosphere.2015.04.085.

[4] G. Pan, C. Jia, D. Zhao, C. You, H. Chen, G. Jiang, Effect of cationic and anionic surfactants on the sorption and desorption of perfluorooctane sulfonate (PFOS) on natural sediments, Environ. Pollut. 157 (2009) 325–330. doi:10.1016/j.envpol.2008.06.035.

[5] C.P. Higgins, R.G. Luthy, Sorption of perfluorinated surfactants on sediments, Environ. Sci. Technol. 40 (2006) 7251–7256. doi:10.1021/es061000n.

[6] C. You, C. Jia, G. Pan, Effect of salinity and sediment characteristics on the sorption and desorption of perfluorooctane sulfonate at sediment-water interface, Environ. Pollut. 158 (2010) 1343–1347. doi:10.1016/j.envpol.2010.01.009.

[7] L. Ahrens, Polyfluoroalkyl compounds in the aquatic environment: a review of their occurrence and fate, J. Environ. Monit. 13 (2011) 20–31. doi:10.1039/C0EM00373E.

[8] P. Labadie, M. Chevreuil, Partitioning behaviour of perfluorinated alkyl contaminants between water, sediment and fish in the Orge River (nearby Paris, France), Environ. Pollut. 159 (2011) 391–397. doi:10.1016/j.envpol.2010.10.039.

[9] J.L. Guelfo, C.P. Higgins, Subsurface transport potential of perfluoroalkyl acids at aqueous film-forming foam (AFFF)-impacted sites, Environ. Sci. Technol. 47 (2013) 4164–4171. doi:10.1021/es3048043.

22