Electronic Supplementary Material (ESI) for Environmental Science: Water Research & Technology.

This journal is © The Royal Society of Chemistry 2022

1 Supporting Information for

2

3 Destruction of Per/poly-fluorinated Alkyl Substances by Magnetite

4 Nanoparticle-Catalyzed UV-Fenton reaction
5 Danielle. R. Schlesinger, Collin McDermott, Nam Q. Le, Jesse S. Ko, James K. Johnson, Plamen
6 A. Demirev, Zhiyong Xia*
7 The Johns Hopkins University Applied Physics Laboratory, Laurel, MD 20723, USA
8 *Corresponding author: zhiyong.xia@jhuapl.edu

10 This file includes the following content:

11 S1. High Resolution Mass Spectrometry (HRMS) Experimental Details.

12 S2. XANES Experimental Details.

13 S3. Results of HRMS, Fluorescence Probe and XANES Analyses.

14 Supplementary Figures SF1 – SF5.

15 Supplementary Tables ST1 – ST2.

16 Supplementary References

17

18

19

20

21

22

23

S1
24 ST1. Percent destruction of PFOA and PFOS with variable UV exposure time, hydrogen peroxide

25 (H2O2) concentration, and Fe3O4 concentration.

Fe3O4 conc. % Destruction % Destruction

UV time (min) H2O2 conc. (M)
(ppm) PFOAa PFOSa
5 0.10 100 13.3 20.0
5 2.55 100 26.7 38.7
5 2.55 1050 68.0 88.0
5 2.55 2000 66.7 84.7
5 5.00 1050 56.0 75.3
5 5.00 2000 68.7 83.3
32.5 0.10 1050 6.7 36.0
32.5 0.10 2000 0.0 33.3
32.5 2.55 100 67.3 76.0
32.5 2.55 1050 61.3 86.0
32.5 2.55 2000 66.0 80.7
32.5 5.00 100 74.7 74.0
32.5 5.00 2000 69.3 82.7
60 0.10 1050 0.0 33.3
60 0.10 2000 6.7 34.0
60 2.55 100 65.3 78.7
60 2.55 2000 63.3 74.0
60 5.00 100 39.3 64.7
60 5.00 1050 13.3 47.3
60 5.00 2000 54.0 72.7
26 a Percent error for samples was calculated to be ±7.0% for PFOA and ±4.9% for PFOS.

27

28 S1. High Resolution Mass Spectrometry (HRMS) Experimental Details.

29 HRMS analysis was conducted on a Thermo Scientific Orbitrap Exploris 240 instrument. Control

30 PFOA and PFOS samples (100 ppm) and samples after reaction at specific conditions (pH 7 or pH

31 9, 2.5M or 5M) were analyzed under the same instrumental conditions. Samples were prepared by

32 filtration through 0.2 µm PTFE syringe filters followed by dilution in methanol in a 1:10 ratio of

33 sample to methanol. Sample solutions were introduced by continuous direct infusion (10 µl/min)

S2
34 into the electrospray ion source of the instrument and were analyzed in both negative and positive

35 ion modes. Optimized source conditions were 3600V and 2600V spray voltage for positive and

36 negative ion generation, respectively. The instrument was autotuned (rf lens rel. value 70%) with

37 automatic gain control and custom injection time functions switched on. The system was mass

38 calibrated in both polarities each day before each analysis by following the manufacturer's

39 protocols with calibration compound mixtures purchased from the manufacturer. The resolving

40 power was set at 240,000 FWHM. For each analyzed sample, 150 individual ion spectra were

41 collected in a range from 50 to 500 m/z (50 to 700 m/z for PFOS) in MS or tandem MS mode for

42 either polarity, respectively. The estimated experimental mass accuracy was better than 5 ppm in

43 the entire range. Tandem mass spectrometry was performed by isolating monoisotopic precursor

44 ions (window 1 m/z) and subsequent manual optimization of the relative collision energy value for

45 efficient collision-induced dissociation (CID). Elemental composition assignments from accurate

46 mass determination of select ions in the averaged negative ion spectra were performed using the

47 Web-based software Chem Calc (https://www.chemcalc.org/mf-finder). The mass of a single

48 electron (0.00054 Da) was subtracted from each experimentally determined anion mass, i.e., the

49 mass of the neutral species was input into the program. The calculated masses of the species with

50 tentative elemental composition assignment were restricted to be less than 5 ppm than the

51 measured neutral species mass.

52 S2. XANES Experimental Details.

53 Fe3O4 nanoparticles were filtered from solutions following Fenton reactions in the UV oven and

54 collected on P5 grade filter paper. A section of filter paper approximately 1 cm  2 cm with a

55 uniform layer of magnetite was cut from the whole filter paper. The filtered Fe3O4 section was

56 mounted between Kapton tape and X-ray clean polyfilm. Fe XANES were collected for each

S3
57 sample from -100 to -20 eV below the Fe k-edge (5.0 eV step size), from -20 eV below to 50 eV

58 above the Fe k-edge (0.2 eV step size), and 50 eV to 150 eV above the Fe k-edge (5.0 eV step size)

59 with a 0.5 sec point–1 acquisition time. Spectra were collected in both transmission and

60 fluorescence modes, and a standard ionization chamber and a four element Silicon drift detector

61 were used for each of these measurement types, respectively. Fe XANES standard spectra of

62 Fe3O4, Fe foil, hematite (Fe2O3), and wustite (FeO) were collected and used to calibrate all samples

63 to the same Fe K-edge energy and to conduct linear combination fitting of each spectrum.

64 S3. Results of HRMS and XANES Analyses.

65 S3.1. HRMS.

66 High resolution negative ion mass spectra of PFOS control and PFOS reacted under varying

67 conditions are plotted on Figure SF1. For the control, the most intense ion peak corresponds to

68 intact PFOS anion [C8F17O3S]-. at m/z 498.93015 (molecular ion). The absolute intensity of the

69 intact PFOS peak in spectra from reacted samples is decreasing considerably. Similar to PFOA,

70 the initial PFOS sample has been completely degraded at 5M H2O2 and pH9, and no molecular ion

71 is observed in the spectrum (Figure SF2.C). Additional peaks observed in spectra from reacted

72 PFOA and PFOS are interpreted as the respectve degradation products (Supplementary Tables ST2

73 and ST3).

74

S4
75

[NO3]-. Mth = 61.98834 [C7F15]-. [C8F15O2]-. A

Mth = 368.97654 Mth = 412.96636

[CO3]-. Mth = 59.98527 B

76
C

77
78 Figure SF1. Comparison between negative ion spectra of PFOA control (A) and PFOA reacted at
79 different conditions: 2.55M H2O2 at pH9 (B), and 5M H2O2 at pH9 (C). Arrows indicate peaks
80 corresponding to degradation products, asterisks indicate peaks reported in the literature.
81

S5
 A
[C8F17O3S]-.
Mth = 498.93019

[C7F15O3S]-.
Mth = 448.93339
[C6F13O3S]-
Mth = 398.93658
* *

82
C

83
84
85 Figure SF2. Comparison between negative ion spectra of PFOS control (A), and PFOS reacted at
86 different conditions - 5M H2O2 at pH7 (B), and 5M H2O2 at pH9 (C). Arrows indicate peaks
87 corresponding to degradation products, asterisks indicate peaks reported in the literature.2 Strong
88 peaks in spectrum C (at m/z 293, 333, 405, 451, 608) are attributed to metalloorganic cluster
89 anions, that are also observed in spectra of blank (non-PFOS-containing) samples at the same
90 experimental conditions.

S6
 91

92 S3.2. FLUORESCENCE PROBE MEASUREMENTS.

93

94 Figure SF3. Fluorescence intensity of detected ROS with increasing concentration of Fe3O4
95 nanoparticles.

96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112

S7
113 S3.3. MAGNETITE ADSORPTION CONTROLS
114

115
116 Figure SF4. PFOA concentrations in solutions with magnetite with and without added H2O2 and
117 exposed to UV radiation. There is no significant change in PFOA concentration with magnetite
118 particles suspended in the solution, compared to the control. Therefore, adsorption to magnetite is
119 not a contributing factor to PFAS removal. In contrast, ROS generated when H2O2 is present in
120 the solution result in a significant reduction in PFAS concentration.
121
122

123 S3.3. XANES.

124 All samples were prepared with 1000 ppm Fe3O4 and underwent 30 min of UV-C exposure before

125 analysis. Linear combination fitting results for these spectra are given in Table ST4. Analysis

126 indicates no change in the peak energy of the K-edge (generally associated with oxidation state),

127 disappearance or appearance of pre-edge or K-edge features, or intensity of these features. In all

128 cases, the collected spectrum matches well with the Fe3O4 standard spectrum collected and shown

129 in Figure SF3A. Further analysis by linear combination fitting (Table ST4) of 4 common Fe

130 mineral spectra containing varied oxidation states of Fe emphasizes there is no change in the Fe3O4

131 after multiple reaction cycles. All samples are shown to be between 94-100% (estimated fit error

132 ± 3%) Fe3O4 in their Fe bonding characteristics. There is one outlier in the 2.55M, pH9 sample

S8
133 after one cycle, which shows only about 89% Fe3O4, and this outlier is likely due to difficulty in

134 normalizing this specific spectrum, which may cause the fit to have a larger error.

135
136 Figure SF5. Fe K-edge XANES spectra for Fe standards (A) and Fe3O4 NPs after variable number
137 of UV-Fenton cycles and solution conditions: (B) 2.55 M H2O2, pH 7, (C) 2.55 M H2O2, pH 9, (D)
138 5 M H2O2, pH 7, (E) 5 M H2O2, pH 9.
139

S9
140 Table ST2. Linear combination fitting of Fe K-edge XANES using four common Fe standards
141 with varied Fe speciation and local bonding states. The reported R-factor represents the goodness-
142 of-fit.
Samples Fe foil Fe2O3 FeO Fe3O4 R-factor
2.55 M H2O2, pH 7
1 cycle 0.016 0 0.014 0.970 0.000055
2 cycles 0.049 0 0.015 0.930 0.000325
3 cycles 0.022 0 0.004 0.974 0.000068
2.55 M H2O2, pH 9
1 cycle 0.075 0 0.032 0.889 0.000633
2 cycles 0.020 0 0.007 0.972 0.000054
3 cycles 0.017 0 0.010 0.974 0.000044
5 M H2O2, pH 7
1 cycle 0.000 0 0.005 0.997 0.000007
2 cycles 0.010 0 0.004 0.987 0.000025
3 cycles 0.045 0 0.013 0.941 0.000242
5 M H2O2, pH 9
1 cycle 0.028 0 0.017 0.955 0.000094
2 cycles 0.028 0 0.016 0.956 0.000112
3 cycles 0.022 0 0.006 0.971 0.000059

143

144

145

S10
146 Supplementary References

147 (1) Trojanowicz, M.; Bobrowski, K.; Szostek, B.; Bojanowska-Czajka, A.; Szreder, T.;

148 Bartoszewicz, I.; Kulisa, K. A Survey of Analytical Methods Employed for Monitoring of

149 Advanced Oxidation/Reduction Processes for Decomposition of Selected Perfluorinated

150 Environmental Pollutants. Talanta 2018, 177 (September 2017), 122–141.

151 https://doi.org/10.1016/j.talanta.2017.09.002.

152 (2) Yamamoto, T.; Noma, Y.; Sakai, S. I.; Shibata, Y. Photodegradation of Perfluorooctane

153 Sulfonate by UV Irradiation in Water and Alkaline 2-Propanol. Environ. Sci. Technol.

154 2007, 41 (16), 5660–5665. https://doi.org/10.1021/es0706504.

155

S11