Sulfidated nano zerovalent iron (S-nZVI) for in situ treatment of chlorinated solvents: A field study

Journal Pre-proof Sulfidated nano zerovalent iron (S-nZVI) for in situ treatment of chlorinated solvents: A field study Ariel Nunez Garcia, Hardiljeet K. Boparai, Ahmed I.A. Chowdhury, Cjestmir V. de Boer, Chris Kocur, Elodie Passeport, Barbara Sherwood Lollar, Leanne M. Austrins, Jose Herrera, Denis M. O'Carroll PII:

S0043-1354(20)30130-5

DOI:

https://doi.org/10.1016/j.watres.2020.115594

Reference:

WR 115594

To appear in:

Water Research

Received Date: 2 December 2019 Revised Date:

3 February 2020

Accepted Date: 6 February 2020

Please cite this article as: Garcia, A.N., Boparai, H.K., Chowdhury, A.I.A., de Boer, C.V., Kocur, C., Passeport, E., Lollar, B.S., Austrins, L.M., Herrera, J., O'Carroll, D.M., Sulfidated nano zerovalent iron (S-nZVI) for in situ treatment of chlorinated solvents: A field study, Water Research (2020), doi: https:// doi.org/10.1016/j.watres.2020.115594. This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. © 2020 Published by Elsevier Ltd.

2

Sulfidated Nano Zerovalent Iron (S-nZVI) for In Situ Treatment of Chlorinated Solvents: A Field Study

3 4 5

Ariel Nunez Garcia,1Hardiljeet K. Boparai,1,2 Ahmed I. A. Chowdhury,1,3 Cjestmir V. de Boer,1,4 Chris M.D. Kocur,1,5 Elodie Passeport,2,6 Barbara Sherwood Lollar,7 Leanne M. Austrins,8 Jose Herrera,9 Denis M. O’Carroll*1,10

1

6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40

1

Department of Civil and Environmental Engineering, Western University, 1151 Richmond Rd., London, Ontario, N6A 5B8, Canada 2

Department of Civil and Mineral Engineering, University of Toronto, 35 St. George Street, Toronto, Ontario, M5S 1A4, Canada 3

Institute of Water and Flood Management, Bangladesh University of Engineering and Technology, Dhaka, Bangladesh 4

Netherlands Organization for Applied Research, TNO, Princetonlaan 6, 3584 CB, Utrecht, The Netherlands 5

OHSU-PSU School of Public Health, Oregon Health & Science University, 3181 SW Sam Jackson Park Road, Portland, OR 97239, United States 6

Department of Chemical Engineering and Applied Chemistry, University of Toronto, 200 College Street, Toronto, Ontario, M5S 3E5, Canada 7

Department of Earth Sciences, University of Toronto, 22 Russell Street, Toronto, Ontario, M5S 3B1, Canada 8

Arcadis, 28550 Cabot Dr #500, Novi, 48377, MI, US

9

Department of Chemical and Biochemical Engineering, Western University, 1151 Richmond Rd., London, Ontario, N6A 5B8, Canada 10

School of Civil and Environmental Engineering, Water Research Centre, University of New South Wales, Sydney, NSW 2052 Australia

*Corresponding author School of Civil & Environmental Engineering Water Research Centre Kensington Campus, University of New South Wales Sydney, NSW 2052 Australia Email: [email protected]

1

41

Abstract

42

Sulfidated nano zerovalent iron (S-nZVI), stabilized with carboxymethyl cellulose (CMC), was

43

successfully synthesized on site and injected into the subsurface at a site contaminated with a

44

broad range of chlorinated volatile organic compounds (cVOCs). Transport of CMC-S-nZVI to

45

the monitoring wells, both downgradient and upgradient, resulted in a significant decrease in

46

concentrations of aqueous-phase cVOCs. Short-term (0 to 17 days) total boron and chloride

47

measurements indicated dilution and displacement in these wells. Importantly however,

48

compound specific isotope analysis (CSIA), changes in concentrations of intermediates, and

49

increase in ethene concentrations confirmed dechlorination of cVOCs. Dissolution from the

50

DNAPL pool into the aqueous phase at the deepest levels (4.0 - 4.5 m bgs) was identifiable from

51

the increased cVOCs concentrations during long-term monitoring. However, at the uppermost

52

levels (~1.5 m above the source zone) a contrasting trend was observed indicating successful

53

dechlorination. Changes in cVOCs concentrations and CSIA data suggest both sequential

54

hydrogenolysis as well as reductive β-elimination as the possible transformation mechanisms

55

during the short-term abiotic and long-term biotic dechlorination. One of the most positive

56

outcomes of this CMC-S-nZVI field treatment is the non-accumulation of lower chlorinated

57

VOCs, particularly vinyl chloride. Post-treatment soil cores also revealed significant decreases in

58

cVOCs concentrations throughout the targeted treatment zones. Results from this field study

59

show that sulfidation is a suitable amendment for developing more efficient nZVI-based in situ

60

remediation technologies.

61 62

Keywords: sulfidation, nano zerovalent iron, dithionite, groundwater, chlorinated VOCs, in situ

63

remediation

2

64

1.

Introduction

65

Sulfidation is a recent development related to the use of zerovalent iron (ZVI) based

66

materials for groundwater remediation (Fan et al. 2017, Li et al. 2017). Though most of the focus

67

in recent years has been on engineered sulfidation of nano ZVI (nZVI), biogenic sulfidation of

68

ZVI has been extensively investigated since the 1990s (Benner et al. 2002, Phillips et al. 2000,

69

Puls et al. 1999, Wilkin et al. 2003). Earlier studies focused on the identification of authigenic

70

mineral phases formed during the application of ZVI permeable reactive barriers (PRBs).

71

Formation of iron sulfides (FeSs) was attributed to the concurrent oxidation of ZVI and

72

generation of sulfides via biogeochemical processes (e.g., microbial reduction of SO42-). These

73

mineral phases were studied in the context of the hydraulic performance of PRBs, noting that the

74

accumulation of FeS precipitates on ZVI surface could contribute to pore clogging, decreased

75

permeability, and slower groundwater flow. However, parallel work on the dechlorination of

76

chlorinated volatile organic compounds (cVOCs) by FeSs (Butler and Hayes 1999, 2000) led to

77

the recognition of these mineral phases as an additional remediant during the operation of PRBs

78

(He et al. 2008, Shen and Wilson 2007). Similar field studies utilized the in situ formation of

79

FeSs to promote abiotic reduction of cVOCs in the Biogeochemical Reductive Dechlorination

80

technology (Kennedy et al. 2006a, 2006b). Investigations on reactive FeSs for remediation

81

purposes is now a thriving field as demonstrated by recent advances on their synthesis,

82

stabilization, and applicability for the removal of contaminants (Gong et al. 2016).

83

In contrast to the biogenic processes described above, abiotic sulfidation can be achieved

84

by modifying the nZVI particles with sulfur compounds; mainly sulfate (Cumbal et al. 2015),

85

dithionite (Cao et al. 2017, Fan et al. 2016, Li et al. 2016, Nunez Garcia et al. 2016, Song et al.

86

2017, Lv et al. 2019), thiosulfate (Han and Yan 2016), and sulfide (Fan et al. 2016, Rajajayavel

3

87

and Ghoshal 2015, Zhao et al. 2019). The resultant sulfidated nZVI (S-nZVI) is more reactive

88

than sulfur-free nZVI for dechlorination of cVOCs (Han and Yan 2016, Jin et al. 2018, Nunez

89

Garcia et al. 2016, Rajajayavel and Ghoshal 2015), adsorption of heavy metals (Cumbal et al.

90

2015, Lv et al. 2019, Zhao et al. 2019), and transformation of organic contaminants (Cao et al.

91

2017, Li et al. 2016, Song et al. 2017). Increased longevity (Nunez Garcia et al. 2016) and higher

92

colloidal stability in suspension (Cao et al. 2017, Song et al. 2017) has also been reported for S-

93

nZVI. To the best of our knowledge, published studies on S-nZVI have been performed solely at

94

the laboratory scale. As such, the field performance of S-nZVI for dechlorination of cVOCs is

95

yet to be evaluated.

96

Multiple pilot- and field-scale studies have been conducted to evaluate the efficacy of

97

nZVI for in situ treatment of contaminated soil and groundwater (Elsner et al. 2010, He et al.

98

2010, Henn and Waddill 2006, Kocur et al. 2014, Qian et al. 2020, Sheu et al. 2016). However,

99

application of nZVI has often faced limitations related to colloidal instability and side oxidation

100

reactions with natural in situ oxidants causing rapid passivation (Fang et al. 2018, Stefaniuk et al.

101

2016). While a significant amount of research has been directed to improve colloidal stability,

102

fewer studies have been dedicated to the minimization of undesirable oxidation reactions.

103

Controlled abiotic sulfidation of nZVI has shown potential to minimize such reactions and

104

improve selectivity towards targeted pollutants (Fan et al. 2016). Such functionality makes S-

105

nZVI more advantageous than nZVI for large-scale applications, as more electron equivalents

106

would hypothetically be directed to the reduction of contaminants, resulting in a more cost-

107

effective treatment.

108

In this study, we have reported results from a field synthesis and injection of

109

carboxymethyl cellulose (CMC) stabilized S-nZVI to remediate groundwater and soil

4

110

contaminated with cVOCs. To assess the effectiveness of CMC-S-nZVI for the in situ

111

transformation of contaminants, Compound Specific Isotope Analysis (CSIA) was used to

112

differentiate

113

transformations (Hunkeler et al. 2009). The specific objectives of the current study were to (1)

114

assess the short- and long-term spatial and temporal variability of cVOCs concentrations in

115

groundwater and soil after injection of CMC-S-nZVI, (2) utilize CSIA as an advanced diagnostic

116

tool to distinguish chemical transformation from physical processes, and (3) to monitor changes

117

in chloride and total boron concentrations to assess dilution and displacement.

between

physical

processes

(dilution

and

displacement)

and

chemical

118 119

2.

Materials and Methods

120

2.1. Site History and Description

121

Located in Sarnia, Ontario, the site was home to cVOCs production facilities, resulting in

122

the accumulation of a multicomponent dense non-aqueous phase liquid (DNAPL) source zone. A

123

description of the site can be found in our previous publication (Nunez Garcia et al. 2020). In

124

short, the study area is composed of a porous, non-native sandy material emplaced along a utility

125

corridor within the native clay. A DNAPL pool is located directly below the treatment zone. Due

126

to the differences in permeability between the backfill and the surrounding clay, DNAPL

127

primarily migrated and accumulated between 4 and 5 m below ground surface (bgs). This was

128

consistent with the appearance of the grey clay, as revealed by the borehole logs, and visual

129

observations of DNAPL in the form of staining or sheening of soil cores. DNAPL was further

130

confirmed by organic vapor monitoring measurements using a photoionization detector (Fig. S1).

131

The wide range of cVOCs production processes on this site contributed to the formation of

132

a complex source zone, with major compounds previously reported as tetrachloroethene (PCE),

5

133

trichloroethene (TCE), and chloroform (Kocur et al. 2015). The distribution and concentrations

134

of the cVOCs in the source zone could have been impacted by past remedial activities as well as

135

natural attenuation. The abundance of typical daughter products from PCE (i.e., dichloroethenes

136

(DCEs) and ethene) and carbon tetrachloride (CCl4) (i.e., chloroform and dichloromethane

137

(DCM)), present in background samples (Fig. S2), supports the hypothesis that transformation of

138

parent compounds has occurred over time. The present study was conducted at the fringes of a

139

previous field trial that took place four years prior to this study, when a total of 620 L of 1 g L-1

140

CMC-nZVI was introduced into four wells (Kocur et al. 2014, 2015, 2016). A plan view of the

141

study area with the sets of wells from both studies is presented in Fig. S3. Evidence of natural

142

attenuation at the site has been previously reported and attributed to the abundance of

143

Dehalococcoides spp. (Dhc) in the background samples prior to the CMC-nZVI injection (Kocur

144

et al. 2015).

145 146

2.2. Monitoring Network

147

Eight multilevel bundle piezometers were installed, six downstream (NA1, NB1, NC1,

148

NA2, NB2, and NA4) and two upstream (NA3 and NB3) of the injection well (Figs. S3-S4). The

149

injection well consisted of a conventional 5 cm well with 0.61 m screen, advanced using hollow

150

stem augers. Bundled piezometers were made up of seven color coded ¼” teflon tubes, mounted

151

on a ¼” steel threaded rod (McMaster-Carr, USA) for stability. The stainless steel screen length

152

of each teflon tube was 0.127 m (100-mesh, McMaster-Carr, USA) and placed 0.305 m vertically

153

apart with fabric mesh pockets holding ¼” coated bentonite pellets (Canpipe, CA) in between in

154

order to target different sampling depths (Figs. S4-S5). Unless otherwise specified, each color

155

denotes the following depths (m bgs) for all wells: Black - 2.90 m, Yellow - 3.20 m, Green - 3.51

6

156

m, Clear - 3.81 m, Blue - 4.12 m, and White - 4.42 m. The Red level (4.73 m bgs) was emplaced

157

within the DNAPL pool and therefore not sampled for cVOCs analysis. Information on bundle

158

piezometers for multilevel sampling can be found in the supplementary material.

159 160

2.3. CMC-S-nZVI Synthesis and Characterization

161

Details on the synthesis procedure and characterization of the CMC-S-nZVI particles were

162

described previously (Nunez Garcia et al. 2020). Briefly, nZVI was synthesized on site by first

163

mixing ferrous sulfate heptahydrate with CMC (90K) and then reducing the mixture using

164

sodium borohydride. Aqueous-solid sulfidation was then carried out by treating the freshly-

165

synthesized CMC-nZVI with sodium dithionite to produce a suspension of 1 g L-1 CMC-S-nZVI,

166

stabilized in 0.77% weight/volume CMC and doped with 22 mM dithionite. A total of 620 L of

167

the suspension was prepared in four distinct batches, 155 L each, and introduced under gravity-

168

feed conditions via the injection well for 16 hours.

169

Transmission Electron Microscopy (TEM) coupled with Energy Dispersive X-ray

170

Spectroscopy (EDS) of CMC-S-nZVI, from synthesis barrels, confirmed the presence of two

171

different types of particles after sulfidation (Nunez Garcia et al. 2020 and Fig. S6). The first type

172

consisted of discrete spherical nZVI-like particles with an average size of ~90±13 nm and iron as

173

their major constituent. Some particles also showed the presence of oxygen and sulfur, indicating

174

the formation of a thin iron oxide/sulfide coating. The second type of particles were larger flake-

175

like structures, with an average particle size of ~505±81 nm. These were relatively fewer in

176

number and were composed of iron and sulfur, suggesting the formation of larger iron sulfide

177

particles. Dynamic light scattering (DLS) also showed a bimodal particle size distribution,

178

further confirming the presence of two types of particles (Nunez Garcia et al. 2020). The size of

7

179

smaller particles in DLS analysis ranged from 357.4 to 438.7 nm that was close to the

180

hydrodynamic diameter of unsulfidated CMC-nZVI particles. The size of larger particles ranged

181

between 881 and 1038 nm. The Fe0 content of CMC-S-nZVI suspension could not be quantified

182

by acid digestion with hydrochloric acid, possibly due to its reaction with the sulfur compounds

183

(e.g., thiosulfate) present in the suspension (Nunez Garcia et al. 2020).

184 185

2.4. Sampling and Analytical Methods

186

Groundwater samples were collected using 40 mL VOA (volatile organic analysis) glass

187

vials, leaving no headspace and preserved with 0.2 grams of sodium bisulfate (NaHSO4).

188

Background samples were collected ~28.5 hours before the injection and are referred as ‘0 day’.

189

cVOCs (PCE, CCl4, tetrachloroethanes (1,1,1,2-TeCA & 1,1,2,2-TeCA), TCE, chloroform,

190

trichloroethanes (1,1,1-TCA & 1,1,2-TCA), and 1,2-dichloroethane (1,2-DCA)) were extracted

191

by transferring 250 µL aliquot to 1 mL hexane and analyzed with a modified EPA 8021 method

192

using an Agilent 7890 Gas Chromatograph (GC) equipped with an Electron Capture Detector

193

(ECD), a DB-624 capillary column, and an autosampler. For hydrocarbons (ethane and ethene)

194

and lower chlorinated VOCs (DCEs, 1,1-DCA, vinyl chloride (VC), chloroethane, and DCM),

195

aliquots of 1 mL were transferred to 2-mL GC vials and allowed to equilibrate for a minimum of

196

one hour before manually sampling 250 µL of the headspace and injecting into the GC. Analysis

197

was carried out using a Flame Ionization Detector (FID) and a GS-Gaspro column. External

198

standards were used for preparing calibration curves for all the cVOCs and hydrocarbons.

199

For cVOCs in soil, background samples were collected during the installation of the wells

200

(25-28 days before CMC-S-nZVI injection), followed by post-injection sampling at 94 and 554

201

days. The soil cores were logged and sub-sampled at either pre-determined depths or targeted

8

202

locations considered to be highly impacted by cVOCs. Post-injection boreholes were located

203

between the locations of the monitoring wells, 0.3-0.6 meters apart, to sample along the CMC-S-

204

nZVI flow path. Bulk soil samples were collected and stored in 60-mL jars, filling the container

205

to the brim and leaving no headspace to minimize losses, in accordance with EPA Method

206

5035A. Jars were stored on ice, transported to the laboratory, and kept in a cold room at 4 °C. In

207

the laboratory, 10 g of soil sample was quickly transferred into pre-weighted vials containing 10

208

mL methanol and the vials were kept on shaker for thirty minutes for the cVOCs extraction. The

209

extractant solution was then diluted with water. Analysis of the cVOCs was performed with a

210

GC-ECD and a GC-FID, as described above.

211

Chloride was analyzed using a high-performance liquid chromatograph equipped with a

212

conductivity detector (Model 432, Waters, Milford, MA), a 4.6 × 50 cm IC-Pak Anion column

213

(#Wat007355) using a 12% water-acetonitrile eluent as mobile phase. For elemental analysis,

214

soil samples were digested using U.S. EPA Method 3051A. Digested samples, as well as total

215

iron and total boron in water, were analyzed as reported previously (Nunez Garcia et al. 2020).

216

Sulfide in monitoring well samples was measured by iodometric titration (APHA 1999).

217 218

2.5. Compound Specific Isotope Analysis

219

Background samples were collected from NB1-White and NB2-White 28.5 hours before

220

injection and preserved in 40 mL VOA vials using NaHSO4. Post-injection samples were also

221

collected from the same wells 17 days after CMC-S-nZVI injection, preserved in 1 mL

222

concentrated hydrochloric acid in 40 mL VOA vials with 5 mL headspace following the method

223

of Elsner et al. (2006). Vials were then covered with aluminum foil and frozen upside down to

224

allow for a gradual freezing process and minimize losses. Headspace sampling and analysis was

9

225

carried out using Gas Chromatograph - Combustion - Isotope Ratio Mass Spectrometer

226

(Finnigan 252 IRMS). Stable carbon isotope values are reported in the δ-notation (‰), relative to

227

the international Vienna Pee Dee Belemnite standard, as follows (Eq. 1):

‰ =

228

.



.

.

−1

.

$

$

229

where "

230

the sample and standard, respectively. All stable carbon isotope values are reported with a 0.5-‰

231

error encompassing both accuracy and reproducibility (Sherwood Lollar et al. 2007). A

232

minimum of a 1 to 2‰ difference between two δ13C values is considered significant (Hunkeler et

233

al. 2009). Background information on CSIA and details on the GC method can be found in the

234

supplementary material.

.

# .

%&'()*

and "

(1)

.

# .

%+&,-&.-

are the ratios of carbon-13 and carbon-12 in

235 236

3.

Results and Discussion

237

3.1. Fate and Transport of CMC-S-nZVI

238

Detailed results for the fate and transport of CMC-S-nZVI suspension at this site were

239

discussed previously (Nunez Garcia et al. 2020). Briefly, the suspension was quite mobile with

240

significant transport to the downgradient wells NB1-White, NC1-White, and NA4-Blue at a

241

distance of 0.86 m, 0.91 m, and 2.7 m, respectively, from the injection well (Fig. 1a-b). Notable

242

migration of the suspension was also found in NA3-White, 1.71 m upgradient from the injection

243

well. In these wells, concentrations of sulfate, sulfur, and total boron often followed a similar 10

244

trend as total iron. CMC-S-nZVI also travelled vertically up to the Black (2.90 m bgs) level of

245

NB1 and Green level (3.51 m bgs) of NB2, as was shown by significant increase in total iron and

246

total boron concentrations, predominantly during the injection period (Nunez Garcia et al. 2020

247

and Fig. 2a-b). A noticeable increase in sulfide concentrations in the monitoring wells also

248

indicated the lateral and vertical transport of the suspension (Fig. S7). TEM analysis of the

249

monitoring well samples, collected during injection and on day 3, confirmed the presence of both

250

nZVI-like particles and larger flaky structures (possibly FeSX), similar to those observed in the

251

synthesis barrels (Nunez Garcia et al. 2020). Moreover, the presence of CMC-S-nZVI

252

suspension was clearly visible from the black color of the monitoring well samples. Significant

253

amounts of total iron, total boron, and sulfide were found in most of these wells up to 17 days

254

(Figs. 1b, 2a-b, and S7) but the concentrations decreased thereafter. Suspended black particles

255

remained in the injection well for several months (>196 days) but visible particles were not

256

found in the monitoring wells during the long-term sampling events (Nunez Garcia et al. 2020).

257 258

3.2. Changes in cVOCs Concentrations Due to Physical versus Chemical Processes

259

The distribution and concentrations of cVOCs in groundwater samples could have changed

260

due to both chemical transformations as well as physical processes such as dilution and

261

displacement. Contaminant transformation can be assessed by CSIA (Elsner et al. 2010),

262

chloride ion generation, and formation of intermediates and end products while dilution can be

263

assessed by investigating changes in the concentrations of conservative species (He et al. 2010).

264 265

3.2.1. Compound Specific Isotope Analysis

11

266

Stable carbon isotope values were measured for PCE, TCE, cis-1,2-DCE, and VC for NB1-

267

White and NB2-White groundwater samples before (day 0) and after (day 17) of CMC-S-nZVI

268

injection (Table 1). These sampling locations were chosen because of their different CMC-S-

269

nZVI breakthroughs though they were roughly on the same flow path, 0.86 (NB1-White) and

270

1.78 m (NB2-White) downgradient from the injection well (Fig. S3). CMC-S-nZVI breakthrough

271

(Nunez Garcia et al. 2020) was greater at NB1-White with a maximum total iron concentration

272

of 1309 µM whereas the maximum total iron concentration detected at NB2-White was 219 µM

273

(Fig. 1b).

274

Before injection, the δ13C values for PCE were very similar (i.e., within ± 0.5 ‰) in NB1-

275

White and NB2-White, suggesting that -26.0 ‰ was a relatively homogenous initial isotope

276

signature for PCE at the site at the time of this study. Compared to NB1-White, TCE was

277

enriched in 13C (-1.2 ‰) whereas cis-1,2-DCE and VC were depleted in 13C by -2.9 and -1.2 ‰,

278

respectively, in NB2-White. These results indicate the occurrence of TCE transformation and

279

generation of cis-1,2-DCE and VC prior to CMC-S-nZVI injection, as previously reported for

280

the adjoining area (Kocur et al. 2015, 2016). These isotopic changes prior to the CMC-S-nZVI

281

injection, however, would not impact the results of the current study as the CSIA method

282

involves determination of absolute changes in isotopic composition before and after injection.

283

After injection, PCE, TCE, and cis-1,2-DCE concentrations decreased at NB1-White on

284

day 17 whereas VC remained constant (Table 1 and Fig. 1f-i). PCE concentration declined from

285

392 to 73.6 µM while its δ13C value increased from -26.0 to -24.6 ‰ (-1.4 ‰ enrichment in 13C)

286

indicating in situ transformation of PCE between days 0 and 17. For TCE, the δ13C value became

287

more depleted in

288

62.6 µM. This suggests that TCE generation from PCE transformation was likely more

13

C (-22.9 to -25.0 ‰) even though its concentration decreased from 91.9 to

12

289

significant than the TCE transformation. In addition, δ13C for TCE was more negative than that

290

for PCE on day 17 indicating incorporation of

291

transformation product of PCE. Past literature has reported that molecules containing exclusively

292

light isotopes (12C) are preferentially transformed leading to an accumulation of

293

molecules (Elsner et al. 2008, 2010). In both NB1-White and NB2-White, cis-1,2-DCE

294

concentrations decreased by approximately 50-60% while getting enriched in

295

22.8 to -20.2 ‰ in NB1-White and from -25.7 to -24.1 ‰ in NB2-White). Such enrichment

296

trends are consistent with the breaking of bonds during transformation (Lojkasek-Lima et al.

297

2012). The δ13C value for cis-1,2-DCE in NB1-White was less negative than that for PCE and

298

TCE on day 17 suggesting that cis-1,2-DCE transformation exceeded its generation as a

299

PCE/TCE transformation product. Contrary to NB1-White, the concentrations and δ13C values

300

for PCE and TCE were relatively constant at NB2-White between days 0 and 17, indicating

301

limited transformation likely due to the very limited CMC-S-nZVI breakthrough at this location.

302

In both wells, VC concentrations remained constant and VC stable carbon isotope signatures did

303

not change significantly. Overall, the CSIA results provided strong evidence for in situ

304

transformation of PCE, TCE, and cis-1,2-DCE in the well with significant CMC-S-nZVI

305

breakthrough (i.e., NB1-White) but limited transformation in the well with limited CMC-S-nZVI

306

breakthrough (i.e., NB2-White).

12

C in TCE, further pointing towards TCE as a

13

12

C in product

C (i.e., from -

307 308 309 310 311

3.2.2. Chloride Analysis The extent of cVOCs transformation is also explored through chloride analysis. Chloride ions are generated via reductive dechlorination of cVOCs (Eq. 2). /0 1 + 3 4 + 5 6 → /0 #6 + 35 + 4 8

(2)

13

312

where RCl represents a generic chlorinated aliphatic compound.

313

The background chloride concentrations in the monitoring wells were in the range of 6597

314

to 32120 µM (Fig. S8) that are much higher than the chloride concentrations (1290-7300 µM) to

315

be generated from the complete dechlorination of all cVOCs in the background samples of these

316

wells. Thus, the changes in chloride concentrations after CMC-S-nZVI injection would not be

317

able to depict a clear picture of the cVOCs dechlorination due to CMC-S-nZVI. For example, the

318

chloride concentrations, calculated based on the generation of daughter products from the

319

dechlorination of parent compounds, account for only ~12% (540 µM) and ~15% (1069 µM) of

320

the total measured chloride in the Black and Yellow levels of NB1 on day 3, respectively (Table

321

S1). The difference between the predicted and the measured chloride could be due to

322

displacement/dilution as well as generation of unmonitored/unidentified dechlorination products.

323

However, some interesting changes in chloride concentrations were observed at various NB1

324

levels that are worth mentioning. For example, higher chloride concentrations were observed at

325

the lower levels of NB1 (3.51 - 4.42 m bgs) before CMC-S-nZVI injection (Fig. 2c). The trend

326

reversed after the injection, with greater chloride concentrations detected at shallower depths (2.9

327

- 3.2 m bgs). Specifically, the concentrations at the Black and Yellow levels increased by ~4550

328

µM and ~7000 µM, respectively, on day 3. On the other hand, concentrations decreased from

329

14621 to 9594 µM at the White level. Chloride concentrations also decreased significantly for

330

Blue, Clear, and Green levels on day 3. During the injection, CMC-S-nZVI suspension first

331

reached the lower levels (White and Blue) and at greater concentrations. This might have pushed

332

the pre-existing well water vertically to the upper levels, resulting in upward displacement of

333

chloride at NB1. On day 17, chloride concentrations further decreased at White and Blue levels

334

of NB1, increased for Green, but remained constant at the uppermost levels. Along with

14

335

displacement, dilution by the CMC-S-nZVI suspension would also have contributed to these

336

changes in chloride concentrations, especially at the lower levels.

337 338

3.2.3. Depth Profiles of cVOCs and Ethene

339

Similar to chloride data, the depth profiles for cVOCs show that the concentrations of all

340

the cVOCs decreased at NB1-White on day 3 (Fig. 2d-j). However, the trend was not the same

341

for other levels. For example, the concentrations of parent compounds PCE and CCl4 decreased

342

noticeably whereas the concentrations of intermediates (e.g. cis-1,2-DCE, VC, chloroform, and

343

DCM) increased significantly at NB1-Black. This clearly indicates the occurrence of

344

dechlorination even if displacement/dilution was happening. Similarly, the increasing-decreasing

345

trend for the cVOCs was not consistent for the various NB1 levels on day 17.

346

Though significant concentrations of ethene were present in the background samples,

347

considerable changes in the ethene concentrations were noticed after the CMC-S-nZVI injection.

348

On day 3, the trend for ethene was also similar to that of the intermediates, with decreased

349

concentrations at the lower levels and increased concentrations at the upper levels of NB1 (Fig.

350

2k). In contrast to chloride data, ethene concentrations increased at all the levels, except Black,

351

from day 3 to day 17, further indicating dechlorination at NB1.

352 353

3.2.4. Boron Analysis

354

To evaluate the extent of dilution, the inorganic conservative constituent boron was

355

analyzed. Total boron at NB1-White, where chloride concentrations decreased after injection,

356

increased approximately seven-fold in comparison to total boron at the Black level (Fig. 2b). To

15

357

make this observation more quantitative, the following relationship (Eq. 3) is used (He et al.

358

2010):

359

9 = 1 −

360

where D is “dilution” factor, Ct is the total boron concentration in the groundwater sample on t =

361

3 or 17 days, and C0 is the total boron concentration in the injected suspension (37.7 mM).

362

Values approaching unity mean little to no dilution of the groundwater by the injected

363

suspension. On day 3, dilution is most noticeable at NB1-White (D = 0.87, Fig. S9a), followed

364

by NC1-White (0.92) and NA4-Blue (D = 0.93, Fig. S9b). All other wells, including upper levels

365

of NB1, had D ≥0.95 indicating lesser dilution. On day 17, D values increased or remained

366

constant for all the wells, except for NC1-White that decreased to 0.90. The presence of total

367

boron, above background concentrations, indicates that the injected suspension was still present

368

in the targeted area on day 3 and to a lesser extent on day 17. This suggests that dilution has also

369

contributed to changes in cVOCs concentrations.

(;)

=(;)

(3)

370 371

3.3. Dechlorination of cVOCs in Groundwater

372

A previous field study in the adjoining area showed significant cVOCs transformation in a

373

three-week period after CMC-nZVI injection, indicating the occurrence of short-term abiotic

374

transformation that was then followed by long-term enhanced biotic transformation (Kocur et al.

375

2015). Thus, the changes in cVOCs concentrations after CMC-S-nZVI injection are presented as

376

short- and long-term changes in the current study.

377 378

3.3.1. Short-Term Changes in Aqueous cVOCs

16

379

Significant CMC-S-nZVI transport was found at 0.86 m (NB1-White), 0.91 m (NC1-

380

White), and 2.7 m (NA4-Blue) downgradient as well as 1.71 m (NA3-White) upgradient of the

381

injection well (Fig. 1b). NB2-White, located at 1.78 m downgradient, observed limited CMC-S-

382

nZVI breakthrough. Of note is that some of these wells retained high total iron concentrations

383

even 17 days after injection. Coincident with the transport of CMC-S-nZVI and associated

384

geochemical changes (Nunez Garcia et al. 2020), considerable changes in concentrations of

385

aqueous cVOCs in these wells were observed (Figs. 1 and S10). Total iron concentrations did not

386

change much at NA2-Blue and NB3-White and the results for these wells are not discussed here.

387

NB1-White showed a noticeable decrease in all the cVOCs on day 3 (Figs. 1 and S10),

388

concurrent with the high concentration of total iron (763 µM). A simultaneous decrease in ethene

389

concentration (244 to 72.8 µM) on day 3 suggests the occurrence of dilution/displacement due to

390

CMC-S-nZVI injection. However, even when total iron concentration decreased to 96.7 µM on

391

day 17, the decrease in cVOCs (except TCE and VC) concentrations continued, indicating

392

dechlorination. Furthermore, increase in TCE concentration from 28.7 µM on day 3 to 62.6 µM

393

on day 17 had resulted from its generation as a dechlorination product of PCE, as indicated by

394

CSIA results (Section 3.2.1.). A significant increase in ethene concentration from 72.8 µM on

395

day 3 to 173 µM on day 17 further confirms dechlorination.

396

CMC-S-nZVI migrated to NA3-White during injection, with total iron concentrations

397

remaining relatively high for an extended period (i.e., 220 µM total iron on day 17). Like NB1-

398

White, NA3-White showed a noticeable decrease in almost all cVOCs concentrations on day 3

399

that continued until day 17 (Figs. 1 and S10). Although this decrease can be partly attributed to

400

dilution/displacement, increase in ethene concentration on day 17 indicates that dechlorination

17

401

also took place. There was also an increase in TCE concentration from 47.7 µM on day 3 to 57.5

402

µM on day 17 that might have generated from PCE dechlorination.

403

NA4-Blue was another well with good CMC-S-nZVI breakthrough during injection. In this

404

well, concentrations of PCE, cis-1,2-DCE, and chlorinated ethanes did not change considerably

405

on day 3 (Figs. 1 and S10). However, TCE concentration decreased from 87.8 to 37.7 µM (~50

406

µM), concurrent with a proportional increase in VC (~11 µM) and ethene (~36 µM)

407

concentrations on day 3, suggesting TCE dechlorination. A significant decrease in concentrations

408

of chlorinated methanes was also observed on day 3. In contrast to NB1-White and NA3-White,

409

the concentrations of all cVOCs rebounded on day 17 at NA4-Blue. Interestingly, the total iron

410

concentration also increased simultaneously in this well. Transport data shows that NA4-Blue is

411

connected to the injection well via preferential flow paths (Nunez Garcia et al. 2020) which

412

might have contributed to cVOCs and iron mobilization to NA4-Blue on day 17. Like NA4-Blue,

413

cVOCs and ethene concentrations at NC1-White also decreased on day 3 but the concentrations

414

rebounded on day 17 with a concurrent increase in total iron concentration.

415

Although limited CMC-S-nZVI migrated to NB2-White during injection, a significant

416

amount of total iron (219 µM) was retained in this well up to day 17. At this location, limited

417

change (<10%) in concentrations of parent compounds PCE, CCl4, and TCE was observed but

418

concentrations of lower chlorinated VOCs continued to decrease noticeably up to day 17. The

419

decrease in ethene concentration on day 3 suggests dilution/displacement, however, its increase

420

on day 17 indicates occurrence of dechlorination.

421

Proximity of the Blue and White levels (4 - 4.5 m bgs) to the DNAPL pool along with the

422

dilution/displacement effects by CMC-S-nZVI injection make it challenging to distinguish

423

between the various processes that govern the changes in cVOCs concentrations. A clearer

18

424

picture of potential dechlorination can be deduced from the changes in cVOCs concentrations at

425

the uppermost level of NB1 (Black level - 2.90 m bgs), positioned approximately 1.5 m above

426

the source zone (Fig. S11). Vertical transport of the CMC-S-nZVI suspension to this location

427

was observed by an increase in the total iron and total boron concentrations in the upper levels of

428

NB1 on day 3 (Fig. 2a-b). PCE concentration decreased from 374 to 272 µM on day 3 but did

429

not change much on day 17 (Fig. S11). Concurrently, cis-1,2-DCE and ethene increased from

430

70.6 to 169 µM and from 96.9 to 145 µM, respectively, on day 3 but decreased to 110 and 101

431

µM on day 17. It is important to note that there was no appreciable accumulation of VC, with

432

cis-1,2-DCE and ethene observed as the main dechlorination products. Similarly, the decrease in

433

CCl4 concentration (from 255 to 166 µM) was accompanied by an increase in chloroform and

434

DCM on day 3. Unlike PCE, CCl4 concentration decreased further to 124 µM on day 17.

435

Decreases in concentrations were also observed for other parent compounds (1,1,1,2-TeCA and

436

1,1,2,2-TeCA) matched by increases in daughter products (i.e., trans 1,2-DCE and 1,1-DCA).

437

As stated above, the short-term changes in cVOC concentrations were influenced

438

simultaneously by dechlorination, dilution, and displacement. Thus, due to the complexity of the

439

system, changes in cVOC concentrations would not yield a straightforward correlation with

440

changes in total iron, total boron, or sulfate (CMC-S-nZVI constituents) concentrations. For

441

example, the concentrations of intermediates (e.g., cis-1,2-DCE) increased in some wells due to

442

their generation from dechlorination of parent compounds (e.g., PCE) but concurrently decreased

443

in other wells due to their own dechlorination. In contrast, ethene would yield a more

444

straightforward correlation since it is an end product of the reductive dechlorination of

445

chlorinated ethenes present at the site. From day 0 to day 3, changes in ethene concentration

446

were negatively correlated with the changes in total iron concentration (Fig. S12a). The total iron

19

447

concentrations increased as a result of CMC-S-nZVI transport to the monitoring wells. Ethene

448

concentrations were also supposed to increase due to its generation as a dechlorination product.

449

However, the simultaneous occurrence of dilution and displacement resulted in decreased ethene

450

concentrations, especially in the lower levels of the monitoring wells. This trend was reversed at

451

the later stage (Fig. S12d). Changes in ethene concentrations from day 3 to day 17 were

452

positively correlated to changes in total iron concentrations from 0 to 3 days. These results show

453

that increase in ethene concentration was relatively much higher in wells with higher total iron

454

breakthrough, presumably due to abiotic dechlorination of cVOCs by the CMC-S-nZVI. Total

455

boron (Fig. S12b & e) and sulfate (Fig. S12c & f) followed the same trend as the total iron.

456 457

3.3.2. Long-Term Changes in Aqueous cVOCs

458

NB1-White and NB1-Black were selected for analysis of long-term dechlorination. NB1-

459

White lies just above the DNAPL pool whereas NB1-Black, the uppermost level located 1.5 m

460

above, is expected to be least affected by the source zone. Short-term monitoring indicated a

461

significant decrease in concentrations of parent compounds (PCE, CCl4, and TeCAs) at both

462

White and Black levels of NB1 (Fig. 3), which was partly due to dechlorination as discussed in

463

sections 3.2.1., 3.2.3., and 3.3.1. There was also a noticeable change in the concentrations of the

464

daughter products on day 17 at both the levels. Overall, the concentration of total cVOCs

465

decreased from 922 µM to 692 µM at NB1-Black (Fig. 3d) and from 1620 µM to 443 µM at

466

NB1-White on day 17 (Fig. 3h). However, the long-term trend was opposite at these locations.

467

For NB1-Black, the concentrations of PCE, CCl4, and TeCAs continued to decline (e.g. ~70% on

468

day 157), showing further dechlorination of these cVOCs (Fig. 3a-c). Concurrently, the

469

concentrations of daughter products (e.g., DCE isomers, chloroform, and 1,2-DCA) increased on

20

470

day 157 but then decreased noticeably for the next sampling rounds. Production of ethene

471

throughout this period confirms the occurrence of dechlorination (Fig. 3d). Total cVOCs

472

concentration decreased to as low as 245 µM at NB1-Black on day 561. In contrast, total cVOCs

473

concentration rebounded at NB1-White on day 157 and continuously increased to 2585 µM on

474

day 561 (Fig. 3h). There was a significant and continued increase in the concentrations of parent

475

compounds as well as the daughter products at this level (Fig. 3e-g). The constant generation of

476

ethene indicates dechlorination (Fig. 3h) although continued dissolution of the DNAPL pool

477

below seems to be the dominant process. Similar trends were observed for long-term cVOCs

478

data at NB2 where a continuous decrease in cVOC concentrations was observed at NB2-Black

479

but cVOCs concentrations at NB2-White increased with time (Fig. S13).

480

Lack of visible black particles and noticeable decrease in total iron and sulfide

481

concentrations in the monitoring well samples, during the long-term sampling events, indicated

482

the absence of injected CMC-S-nZVI suspension. This suggests that CMC-S-nZVI did not play a

483

direct role in the long-term dechlorination of cVOCs. Previous field studies have also shown that

484

abiotic dechlorination (caused by the initial ZVI corrosion) diminishes during long-term

485

monitoring, while biotic dechlorination becomes the primary degradation pathway (He et al.

486

2010, Kocur et al. 2015, 2016). The injected CMC-S-nZVI might have created favorable

487

conditions for the long-term biotransformation of cVOCs, as reported earlier for a CMC-nZVI

488

injection study (Kocur et al. 2016). Moreover, excess dithionite in the injected suspension would

489

also have indirectly contributed to the long-term dechlorination of cVOCs. Dithionite employed

490

in the in situ redox manipulation (ISRM) technology successfully reduces native Fe(III) from

491

aquifer sediments/soils to reactive Fe(II) species, resulting in the transformation of chlorinated

492

organic compounds (Boparai et al. 2006, Szecsody et al. 2004). It even maintains the subsurface

21

493

conditions favorable for reductive degradation up to >3 years (Fruchter et al. 2000). As the soil

494

in the current study area is rich in iron (Fig. S14), the excess dithionite in the suspension is

495

expected to have enriched the subsurface zone with the reactive Fe(II) species, resulting in the

496

abiotic transformation of cVOCs.

497 498

3.4. Possible Reaction Mechanisms

499

It is challenging to determine if changes in cVOCs concentrations were due to abiotic or

500

biotic transformation. While the CSIA results for NB1-White confirmed the chemical

501

transformation of chlorinated ethenes, distinguishing between biotic and abiotic (i.e., via CMC-

502

S-nZVI) processes typically requires comparison of the isotope signatures of the dechlorination

503

products cis-1,2-DCE, VC, acetylene, ethene, and ethane (Elsner et al. 2008). Ethene and ethane

504

were not analyzed by CSIA. Acetylene, which may serve as an indicator for abiotic

505

dechlorination (Butler and Hayes 1999, Elsner et al. 2010), was not detected. The rapid and

506

efficient removal of cVOCs within 17 days suggests that dechlorination was mainly abiotic

507

during this short-term period. A past field study in the adjoining area also showed noticeable

508

reduction of cVOCs within three weeks after CMC-nZVI injection, indicating the occurrence of

509

short-term abiotic dechlorination (Kocur et al. 2015). As the injected CMC-S-nZVI suspension

510

contained excess dithionite, it might have resulted in some short-term abiotic dechlorination.

511

Moreover, the reactive Fe(II) species generated via ISRM might also have contributed to both

512

short- as well as long-term abiotic dechlorination. However, biotransformation is expected to be

513

the major contributor for long-term dechlorination, as reported earlier for the CMC-nZVI

514

injection study (Kocur et al. 2016).

22

515

For short-term dechlorination, CSIA results confirm the generation of TCE as a product of

516

PCE transformation in NB1-White (Table 1). Other wells (e.g. NB1-Black) also showed a

517

temporary increase in the concentrations of intermediates cis-1,2-DCE and VC with concurrent

518

dechlorination of PCE. This indicates hydrogenolysis as the dechlorination mechanism for

519

chlorinated ethenes. However, PCE dechlorination generally exceeded the generation of

520

intermediates and ethene. Similarly, dechlorination of cis-1,2-DCE (e.g. NB1-White and NB2-

521

White) exceeded the formation of VC and ethene. These results suggest that reductive β-

522

elimination was also happening simultaneously. Past research has reported both reductive β-

523

elimination and hydrogenolysis as the dechlorination mechanisms for TCE, treated by S-nZVI,

524

where ethene or acetylene were found as the major dechlorination products (Han and Yan 2016,

525

Rajajayavel and Ghoshal 2015). Experimental conditions, particularly the method of nZVI

526

sulfidation, determine which mechanism would dominate. For example, Han and Yan (2016)

527

reported ethene as the major product, with ethane and acetylene as the minor products, while

528

treating TCE with S-nZVI developed by post-synthesis addition of dithionite (method similar to

529

the one used in this study). In the current study, no acetylene was detected and chloroacetylenes

530

were not analyzed but this does not completely rule out their formation. The quantification of

531

acetylene and chloroacetylenes is challenging in the field as they can quickly volatilize in air.

532

This may be the reason that reductive β-elimination products are rarely reported for the iron-

533

treated field studies. For chlorinated methanes, the formation of chloroform and DCM indicates

534

the transformation of CT via sequential hydrogenolysis, as also reported by Jin et al. (2018).

535

For long-term dechlorination, the formation of intermediates (e.g., DCE isomers,

536

chloroform, and DCM) at NB1-Black and NB2-Black indicate hydrogenolysis as the

537

dechlorination pathway. Significant decrease in cis-1,2-DCE concentrations without equivalent

23

538

production of VC suggest that reductive β-elimination was also happening. Microbial

539

transformation of higher chlorinated ethanes and ethenes often results in partial dechlorination,

540

leading to the accumulation of intermediates such as DCE isomers and VC (He et al. 2010,

541

Kocur et al. 2016). The accumulation of VC, a highly toxic and confirmed human carcinogen

542

(ATSDR 2006), is usually of particular concern due to its poor biodegradability. Past study in the

543

adjacent area reported the generation and accumulation of VC during long-term microbial

544

transformation of PCE and TCE following a CMC-nZVI injection (Kocur et al. 2015, 2016).

545

However, one of the most positive outcomes of this CMC-S-nZVI field treatment is the non-

546

accumulation of lower chlorinated VOCs, particularly VC. CMC-S-nZVI (dithionite sulfidated)

547

injection would result in decreased H2 evolution, formation of FeSs, and significant reduction of

548

native Fe(III) to reactive Fe(II) species whereas CMC-nZVI injection would generate higher

549

amounts of H2 and may not significantly impact the latter two conditions. Thus, there can be two

550

possible reasons for the non-accumulation of VC in this CMC-S-nZVI field study as opposed to

551

the CMC-nZVI field trial. Firstly, differences in geochemical changes, along with the direct

552

interactions between nZVI/S-nZVI and microbes, are expected to result in different

553

inhibitory/stimulatory effects on the microbial communities for the two treatments. Certain

554

classes of bacteria have the ability to intrinsically biodegrade VC in anaerobic aquifers (Bradley

555

et al. 1998, Lorah and Voytek 2004). Although not yet investigated, the geochemical conditions

556

in the subsurface created by CMC-S-nZVI injection might be favorable for the enrichment of

557

these bacteria. Secondly, reactive Fe(II) species from dithionite-reduced sediments would

558

dechlorinate chlorinated ethenes via reductive β-elimination without noticeable production of

559

VC, as reported for TCE (Szecsody et al. 2004).

560

24

561

3.5. Changes in Soil cVOCs

562

Visual observations in the form of staining/sheening of soil cores as well as OVM

563

measurements (Fig. S1) indicated an appreciable amount of cVOCs present as DNAPL and

564

sorbed mass. Soil cVOCs concentrations are considered as a better metric for determining

565

contaminant mass reduction in comparison to aqueous-phase concentrations (Henn and Waddill

566

2006). Thus, changes in soil cVOCs were quantified by analyzing the soil samples collected

567

before and after CMC-S-nZVI injection. Locations for pre- and post-injection soil cores are

568

shown in Fig. S15. In Table S2, the summarized data shows a significant decrease in

569

concentrations of most of the cVOCs in soil samples collected on 94 and 554 days after CMC-S-

570

nZVI injection, with some of them not even detected at many locations. Fig. S16 shows a

571

continuous decline in the total cVOCs concentrations where the background average of 1496

572

µmol/kg decreased to 653 and 125 µmol/kg, respectively, on day 94 and 554 after CMC-S-nZVI

573

injection.

574

Table S2 and Fig. S16 also show that soil cVOCs concentrations were highly variable. For

575

example, PCE concentrations varied between 35.5 and 1759 µmol/kg for the three zones on day

576

94. Thus, the results are also presented as ‘box and whisker’ plots by grouping the data for each

577

sampling event (Fig. 4). Median concentrations of PCE and CCl4 showed some increase on day

578

94 but noticeably decreased on day 554. For all the other cVOCs, except 1,1,2-TCA, there was a

579

continuous downward trend with time. For example, DCM median concentrations decreased

580

from 278 to 8.62 µmol/kg on day 94 and further decreased to 1.55 µmol/kg on day 554. Median

581

concentrations for 1,1,2-TCA remained relatively constant throughout the monitoring period.

582

This shows that the trends among the ten cVOCs analyzed were not always consistent.

583

Quantification of the extent of remediation using soil cores is challenging, in part due to spatial

25

584

variations and highly stratified distribution of contaminants within aquifers. This complexity is

585

exacerbated by the varying sampling depths for different locations and sampling times, not

586

allowing for a systematic depth-by-depth comparison. A correlation analysis, depicted in Table

587

S3, was performed with the purpose of evaluating the overall effectiveness of the CMC-S-nZVI

588

injection on altering the cVOCs concentrations in soil. The ‘r’ values indicate that the cVOCs

589

concentrations correlate better with each other over time, suggesting the treatment was effective

590

in causing a change in the concentrations and distribution of contaminants in the soil. This long-

591

term decrease in soil cVOCs might have occurred partly due to enhanced biological activity after

592

CMC-S-nZVI injection. However, more work needs to be done to investigate the effect of this

593

CMC-S-nZVI formulation on the microbial communities in the treatment zone. Dithionite-

594

reduced structural Fe(II) in the aquifer sediments/soil might also have played an important role in

595

the long-term transformation of soil cVOCs (Paul et al. 2003). These results suggest CMC-S-

596

nZVI is an effective strategy for cVOCs dechlorination in soil.

597 598

4.

Conclusions

599

Results reported herein demonstrate the suitability of CMC-S-nZVI as an effective

600

technology for soil and groundwater remediation at existing contaminated sites. A rapid decrease

601

in cVOCs concentrations was observed in groundwater samples immediately after injection,

602

followed by sustained long-term dechlorination. Although CMC-S-nZVI injection resulted in

603

some dilution and displacement of cVOCs, the changes in intermediate concentrations and an

604

increase in ethene concentrations clearly indicate dechlorination. CSIA serves as another line of

605

evidence, confirming the direct impact of chemical transformation as shown by the changes in

606

stable isotope values of key chlorinated compounds. Proximity to the DNAPL pool resulted in

26

607

mass transfer of non-aqueous constituents into the aqueous phase at the deeper Blue and White

608

levels (4-4.5 m bgs), although, significant ethene generation indicated concurrent dechlorination.

609

In contrast, the uppermost level (Black), which is expected to be least affected by the source

610

zone, observed a continuous decline in cVOCs concentrations accompanied with the generation

611

of ethene, confirming dechlorination. Transformation was not limited to the aqueous phase as

612

concentrations of soil cVOCs also decreased significantly at 94 and 554 days after injection.

613

Presence of dithionite might additionally have resulted in reducing the native Fe(III) to the

614

reactive Fe(II) species which can degrade cVOCs.

615

Long-term success of in situ emplacement of nZVI often relies on the biotransformation

616

that follows the short-term abiotic dechlorination (Kocur et al. 2016) and care must be taken not

617

to inhibit the growth of healthy microbial communities. In this study, the sustained long-term

618

dechlorination points towards biotransformation suggesting that growth of microbial

619

communities was not inhibited. However, the extent to which biotic processes contributed to the

620

transformation of cVOCs is unknown and further characterization of the microbial communities

621

after emplacement of CMC-S-nZVI is required.

622

As fundamental work about synthesis, characterization, and overall mechanisms of CMC-

623

S-nZVI reactivity is ongoing, this study is the first pilot test to upscale the application of

624

abiotically sulfidated CMC-nZVI from laboratory to field. With the growing interest from all

625

sectors of the remediation community, as evidenced by the introduction of new commercial

626

products of sulfidated (n)ZVI (REGENESIS 2018), S-nZVI is likely to become an important In

627

Situ Chemical Reduction technology.

628 629

Acknowledgements

27

630

Financial support for this project was provided by CH2M Canada Limited, Dow Chemical, the

631

Natural Sciences and Engineering Research Council of Canada (NSERC) Remediation

632

Education Network (RENEW) Program, the NSERC Industrial Postgraduate Scholarship to Ariel

633

Nunez Garcia, a NSERC Collaborative Research and Development (CRD) Grant (CRDPJ

634

530665 - 18) and NSERC Discovery Grant to Barbara Sherwood Lollar. The authors thank Ka

635

Yee Lam and Georges Lacrampe-Couloume for help with the stable isotope analyses.

636 637

References

638 639 640

Agency for Toxic Substances and Disease Registry (ATSDR). 2006. Toxicological profile for Vinyl Chloride. Atlanta, GA: U.S. Department of Health and Human Services, Public Health Service.

641 642 643

American Public Health Association (APHA). 1999. Standard Methods for the Examination of Water and Wastewater, 20th ed., American Public Health Association, American Water Works Association and Water Environment Federation, Washington DC, pp. 845-860.

644 645 646

Benner, S.G., Blowes, D.W., Ptacek, C.J., Mayer, K.U., 2002. Rates of sulfate reduction and metal sulfide precipitation in a permeable reactive barrier. Appl. Geochem. 17 (3), 301320.

647 648 649

Boparai, H.K., Shea, P.J., Comfort, S.D., Snow, D.D., 2006. Dechlorinating chloroacetanilide herbicides by dithionite-treated aquifer sediment and surface soil. Environ. Sci. Technol. 40 (9), 3043-3049.

650 651 652

Bradley, P.M., Chapelle, F.H., Wilson, J.T., 1998. Field and laboratory evidence for intrinsic biodegradation of vinyl chloride contamination in a Fe(III)-reducing aquifer. J. Contam. Hydrol. 31 (1-2), 111-127.

653 654

Butler, E.C., Hayes, K.F., 1999. Kinetics of the transformation of trichloroethylene and tetrachloroethylene by iron sulfide. Environ. Sci. Technol. 33 (12), 2021-2027.

655 656

Butler, E.C., Hayes, K.F., 2000. Kinetics of the transformation of halogenated aliphatic compounds by iron sulfide. Environ. Sci. Technol. 34 (3), 422-429.

657 658 659

Cao, Z., Liu, X., Xu, J., Zhang, J., Yang, Y., Zhou, J.L., Xu, X.H., Lowry, G.V., 2017. Removal of Antibiotic Florfenicol by Sulfide-Modified Nanoscale Zero-Valent Iron. Environ. Sci. Technol. 51 (19), 11269-11277.

660 661 662

Cumbal, L.H., Debut, A., Delgado, D.A., Jurado, C.B., Stael, C., 2015. Synthesis of Multicomponent Nanoparticles for Immobilization of Heavy Metals in Aqueous Phase. NanoWorld J. 1 (4), 103-109.

28

663 664 665 666

Elsner, M., Chartrand, M., VanStone, N., Lacrampe Couloume, G., Sherwood Lollar, B., 2008. Identifying Abiotic Chlorinated Ethene Degradation: Characteristic Isotope Patterns in Reaction Products with Nanoscale Zero-Valent Iron. Environ. Sci. Technol. 42 (16), 5963-5970.

667 668 669

Elsner, M., Couloume, G.L., Mancini, S., Burns, L., Sherwood Lollar, B., 2010. Carbon Isotope Analysis to Evaluate Nanoscale Fe(O) Treatment at a Chlorohydrocarbon Contaminated Site. Ground Water Monit. Remediat. 30 (3), 79-95.

670 671 672

Elsner, M., Couloume, G.L., Sherwood Lollar, B., 2006. Freezing to preserve groundwater samples and improve headspace quantification limits of water-soluble organic contaminants for carbon isotope analysis. Anal. Chem. 78 (21), 7528-7534.

673 674 675 676

Fan, D., Lan, Y., Tratnyek, P.G., Johnson, R.L., Filip, J., O’Carroll, D.M., Nunez Garcia, A., Agrawal, A., 2017. Sulfidation of Iron-Based Materials: A Review of Processes and Implications for Water Treatment and Remediation. Environ. Sci. Technol. 51 (22), 13070-13085.

677 678 679

Fan, D., O’Brien Johnson, G., Tratnyek, P.G., Johnson, R.L., 2016. Sulfidation of Nano Zerovalent Iron (nZVI) for Improved Selectivity During In-Situ Chemical Reduction (ISCR). Environ. Sci. Technol. 50 (17), 9558-9565.

680 681 682

Fang, Y., Wen, J., Zeng, G., Shen, M., Cao, W., Gong, J., Zhang, Y., 2018. From nZVI to SNCs: development of a better material for pollutant removal in water. Environ. Sci. Pollut. Res. 25 (7), 6175-6195.

683 684 685 686

Fruchter, J.S., Cole, C.R., Williams, M.D., Vermeul, V.R., Amonette, J.E., Szecsody, J.E., Istok, J.D., Humphrey, M.D., 2000. Creation of a subsurface permeable treatment zone for aqueous chromate contamination using in situ redox manipulation. Ground Water Monit. Remediat. 20 (2), 66-77.

687 688

Gong, Y., Tang, J., Zhao, D., 2016. Application of iron sulfide particles for groundwater and soil remediation: A review. Water Res. 89, 309-320.

689 690 691

Han, Y., Yan, W., 2016. Reductive Dechlorination of Trichloroethene by Zero-valent Iron Nanoparticles: Reactivity Enhancement through Sulfidation Treatment. Environ. Sci. Technol. 50 (23), 12992-13001.

692 693 694

He, F., Zhao, D., Paul, C., 2010. Field assessment of carboxymethyl cellulose stabilized iron nanoparticles for in situ destruction of chlorinated solvents in source zones. Water Res. 44 (7), 2360-2370.

695 696 697

He, Y.T., Wilson, J.T., Wilkin, R.T., 2008. Transformation of reactive iron minerals in a permeable reactive barrier (biowall) used to treat TCE in groundwater. Environ. Sci. Technol. 42 (17), 6690-6696.

698 699

Henn, K.W., Waddill, D.W., 2006. Utilization of nanoscale zero-valent iron for source remediation—A case study. Remediat. J. 16 (2), 57-77.

700 701 702

Hunkeler, D., Meckenstock, R.U., Sherwood Lollar, B., Schmidt, T.C., Wilson, J.T., 2009. A Guide for Assessing Biodegradation and Source Identification of Organic Ground Water Contaminants Using Compound Specific Isotope Analysis (CSIA), pp. 1 - 82, U.S. EPA.

29

703 704

Jin, X., Chen, H., Yang, Q., Hu, Y.A., Yang, Z.L., 2018. Dechlorination of Carbon Tetrachloride by Sulfide-Modified Nanoscale Zerovalent Iron. Environ. Eng. Sci. 35 (6), 560-567.

705 706 707

Kennedy, L.G., Everett, J.W., Becvar, E., Defeo, D., 2006a. Field-scale demonstration of induced biogeochemical reductive dechlorination at Dover Air Force Base, Dover, Delaware. J. Contam. Hydrol. 88 (1-2), 119-136.

708 709 710

Kennedy, L.G., Everett, J.W., Gonzales, J., 2006b. Assessment of biogeochemical natural attenuation and treatment of chlorinated solvents, Altus Air Force Base, Altus, Oklahoma. J. Contam. Hydrol. 83 (3-4), 221-236.

711 712 713

Kocur, C.M., Chowdhury, A.I., Sakulchaicharoen, N., Boparai, H.K., Weber, K.P., Sharma, P., Krol, M.M., Austrins, L., Peace, C., Sleep, B.E., O'Carroll, D.M., 2014. Characterization of nZVI Mobility in a Field Scale Test. Environ. Sci. Technol. 48 (5), 2862-2869.

714 715 716 717

Kocur, C.M.D., Lomheim, L., Boparai, H.K., Chowdhury, A.I.A., Weber, K.P., Austrins, L.M., Edwards, E.A., Sleep, B.E., O'Carroll, D.M., 2015. Contributions of Abiotic and Biotic Dechlorination Following Carboxymethyl Cellulose Stabilized Nanoscale Zero Valent Iron Injection. Environ. Sci. Technol. 49 (14), 8648-8656.

718 719 720 721

Kocur, C.M.D., Lomheim, L., Molenda, O., Weber, K.P., Austrins, L.M., Sleep, B.E., Boparai, H.K., Edwards, E.A., O'Carroll, D.M., 2016. Long-Term Field Study of Microbial Community and Dechlorinating Activity Following Carboxymethyl Cellulose-Stabilized Nanoscale Zero-Valent Iron Injection. Environ. Sci. Technol. 50 (14), 7658-7670.

722 723

Li, D., Mao, Z., Zhong, Y., Huang, W., Wu, Y., Peng, P.a., 2016. Reductive transformation of tetrabromobisphenol A by sulfidated nano zerovalent iron. Water Res. 103, 1-9.

724 725 726

Li, J., Zhang, X., Sun, Y., Liang, L., Pan, B., Zhang, W., Guan, X., 2017. Advances in Sulfidation of Zerovalent Iron for Water Decontamination. Environ. Sci. Technol. 51 (23), 13533–13544.

727 728 729 730

Lojkasek-Lima, P., Aravena, R., Shouakar-Stash, O., Frape, S.K., Marchesi, M., Fiorenza, S., Vogan, J., 2012. Evaluating TCE Abiotic and Biotic Degradation Pathways in a Permeable Reactive Barrier Using Compound Specific Isotope Analysis. Ground Water Monit. Remediat. 32 (4), 53-62.

731 732 733 734

Lorah, M.M., Voytek, M.A., 2004. Degradation of 1,1,2,2-tetrachloroethane and accumulation of vinyl chloride in wetland sediment microcosms and in situ porewater: biogeochemical controls and associations with microbial communities. J. Contam. Hydrol. 70 (1), 117145.

735 736 737

Lv, D., Zhou, J., Cao, Z., Xu, J., Liu, Y., Li, Y., Yang, K., Lou, Z., Lou, L., Xu, X. 2019. Mechanism and influence factors of chromium(VI) removal by sulfide-modified nanoscale zerovalent iron. Chemosphere 224, 306-315.

738 739 740

Nunez Garcia, A., Boparai, H.K., de Boer, C.V., Chowdhury, A.I.A., Kocur, C.M.D., Austrins, L.M., Herrera, J., O'Carroll, D.M., 2020. Fate and Transport of Sulfidated Nano Zerovalent Iron (S-nZVI): A Field Study. Water Res 170, 115319.

741 742 743

Nunez Garcia, A., Boparai, H.K., O’Carroll, D.M., 2016. Enhanced Dechlorination of 1,2Dichloroethane by Coupled Nano Iron-Dithionite Treatment. Environ. Sci. Technol. 50 (10), 5243-5251. 30

744 745 746 747

Paul, C.J., Khan, F.A., Puls, R.W., 2003. In Situ Reduction of Chromium-Contaminated Groundwater, Soils, and Sediments by Sodium Dithionite. Chapter 16, Handbook of Groundwater Remediation using Permeable Reactive Barriers. Naftz, D.L., Morrison, S.J., Fuller, C.C., Davis, J.A. (eds), pp. 465-493, Academic Press, San Diego.

748 749 750

Phillips, D.H., Gu, B., Watson, D.B., Roh, Y., Liang, L., Lee, S.Y., 2000. Performance evaluation of a zerovalent iron reactive barrier: Mineralogical characteristics. Environ. Sci. Technol. 34 (19), 4169-4176.

751 752 753

Puls, R.W., Paul, C.J., Powell, R.M., 1999. The application of in situ permeable reactive (zerovalent iron) barrier technology for the remediation of chromate-contaminated groundwater: a field test. Appl. Geochem. 14 (8), 989-1000.

754 755 756

Qian, L., Chen, Y., Ouyang, D., Zhang, W., Han, L., Yan, J., Kvapil, P., Chen, M., 2020. Field demonstration of enhanced removal of chlorinated solvents in groundwater using biochar-supported nanoscale zero-valent iron. Sci. Total Environ. 698, 134215.

757 758

Rajajayavel, S.R.C., Ghoshal, S., 2015. Enhanced reductive dechlorination of trichloroethylene by sulfidated nanoscale zerovalent iron. Water Res. 78, 144-153.

759 760

REGENESIS, 2018. AquaZVI Specification Sheet. https://regenesis.com/wpcontent/uploads/2018/04/AquaZVI_SpecSheet-8-1.pdf (November 19, 2018).

761 762 763

Shen, H., Wilson, J.T., 2007. Trichloroethylene removal from groundwater in flow-through columns simulating a permeable reactive barrier constructed with plant mulch. Environ. Sci. Technol. 41 (11), 4077-4083.

764 765 766 767

Sherwood Lollar, B., Hirschorn, S.K., Chartrand, M.M.G., Lacrampe-Couloume, G., 2007. An Approach for Assessing Total Instrumental Uncertainty in Compound-Specific Carbon Isotope Analysis:  Implications for Environmental Remediation Studies. Anal. Chem. 79 (9), 3469-3475.

768 769 770

Sheu, Y.T., Lien, P.J., Chen, K.F., Ou, J.H., Kao, C.M., 2016. Application of NZVI-contained emulsified substrate to bioremediate PCE-contaminated groundwater – A pilot-scale study. Chem. Eng. J. 304, 714-727.

771 772 773

Song, S., Su, Y., Adeleye, A.S., Zhang, Y., Zhou, X., 2017. Optimal design and characterization of sulfide-modified nanoscale zerovalent iron for diclofenac removal. Appl. Catal. B Environ. 201, 211-220.

774 775

Stefaniuk, M., Oleszczuk, P., Ok, Y.S., 2016. Review on nano zerovalent iron (nZVI): From synthesis to environmental applications. Chem. Eng. J. 287, 618-632.

776 777 778

Szecsody, J.E., Fruchter, J.S., Williams, M.D., Vermeul, V.R., Sklarew, D., 2004. In situ chemical reduction of aquifer sediments: Enhancement of reactive iron phases and TCE dechlorination. Environ. Sci. Technol. 38 (17), 4656-4663.

779 780 781

Wilkin, R.T., Puls, R.W., Sewell, G.W., 2003. Long-term performance of permeable reactive barriers using zero-valent iron: Geochemical and microbiological effects. Ground Water 41 (4), 493-503.

782 783 784

Zhao, L., Zhao, Y., Yang, B., Teng, H. 2019. Application of Carboxymethyl Cellulose– Stabilized Sulfidated Nano Zerovalent Iron for Removal of Cr(VI) in Simulated Groundwater. Water, Air, & Soil Pollution 230(6), 113. 31

Table 1. δ13C values for chlorinated ethenes from NB1-White and NB2-White (0.86 and 1.78 m downgradient of the injection well, respectively) before (0 days) and after (17 days) CMC-SnZVI injection.

Location

Time (days)

PCE

TCE

cis-1,2-DCE

VC

[µM]

δ13C (‰)

[µM]

δ13C (‰)

[µM]

δ13C (‰)

[µM]

δ13C (‰)

0

392

-26.0

91.9

-22.9

252

-22.8

39.3

-23.0

17

73.6

-24.6

62.6

-25.0

99.8

-20.2

39.2

-22.0

0

394

-26.3

121

-21.7

461

-25.7

29.1

-24.2

17

416

-26.1

125

-22.6

217

-24.1

33.3

-25.0

NB1-White

NB2-White

Distance (m)

0.8

a) Plan View Interval range: 4 - 4.5 m bgs

500

North

Conc. (µM)

1.0 0.6

NA3-White

0.4 NB1-White

0.2 NA4-Blue

0.0 -0.2 2.8

1500 Conc. (µM)

1200

NB2-White

NC1-White

2.1

0.7 0.0 Distance (m)

1.4

-1.4

17 d

200 100 NA3-W 200

b) Total iron

NB1-W

NC1-W

NB2-W

NA4-B

NB1-W

NC1-W

NB2-W

NA4-B

NB1-W

NC1-W

NB2-W

NA4-B

NB1-W

NC1-W

NB2-W

NA4-B

NB1-W

NC1-W

NB2-W

NA4-B

g) TCE

150 100

600

50

300

0 NA3-W

Conc. (µM)

3d

300

-2.1

0 NB1-W

NC1-W

NB2-W

NA4-B

NA3-W 500

c) CCl4

400

300

h) cis-1,2-DCE

300

200

200 100

100

0

0 NA3-W

300 Conc. (µM)

1d

0 -0.7

900

400

400

0d

f) PCE

NB1-W

NC1-W

NB2-W

NA4-B

NA3-W 50

d) Chloroform

40

200

i) VC

30 20

100

10 0

0 NA3-W

Conc. (µM)

100

NB1-W

NC1-W

NB2-W

NA4-B

NA3-W 400

e) DCM

75

300

50

200

25

100

j) Ethene

0

0 NA3-W

NB1-W

NC1-W

NB2-W

NA4-B

NA3-W

Fig. 1. Short - term changes in cVOC concentrations at five locations representing upstream and downstream conditions. a) Plan View is shown for reference and only relevant wells are presented. The origin (0, 0) represents the location of the injection well. b) Total iron concentrations are shown for the same sampling times, including the peak concentrations measured during active injection (0 – 16 hours).

Depth (m bgs)

2.8

b)

2.8

3.0

3.0

3.2

3.2

3.2

3.4

3.4

3.4

3.6

3.6

3.6

3.8

3.8

3.8

4.0

4.0

4.0

4.2

4.2

4.2

4.4

4.4

4.4

0 150 300 450 600 750 Total iron (µM)

d)

0 2.8

c)

Black (2.90 m)

Green (3.51 m) Clear (3.81 m) Blue (4.12 m) White (4.42 m)

2000 4000 6000 0 Total boron (µM) 2.8

5000 10000 15000 Chloride (µM)

e)

f)

2.8

3.0

3.0

3.0

3.2

3.2

3.2

3.2

3.4

3.4

3.4

3.4

3.6

3.6

3.6

3.6

3.8

3.8

3.8

3.8

4.0

4.0

4.0

4.0

4.2

4.2

4.2

4.2

4.4

4.4

4.4

4.4

150 300 PCE (µM)

2.8

450

h)

0

2.8

25 50 75 100 125 0 TCE (µM) 2.8

i)

3.0

3.0

3.2

3.2

3.2

3.2

3.4

3.4

3.4

3.4

3.6

3.6

3.6

3.6

3.8

3.8

3.8

3.8

4.0

4.0

4.0

4.0

4.2

4.2

4.2

4.2

4.4

4.4

4.4

4.4

300

0

50 100 Chloroform (µM)

150

0

10

20 30 40 VC (µM)

j)

3.0

100 200 CCl4 (µM)

g)

100 200 300 400 0 cis-1,2-DCE (µM) 2.8

3.0

0

0d 3d 17 d

Yellow (3.20 m)

3.0

0

Depth (m bgs)

2.8

3.0

2.8

Depth (m bgs)

a)

20 40 DCM (µM)

60

0

50

k)

100 200 Ethene (µM)

Fig. 2. Depth profiles for a) total iron, b) total boron, c) chloride, d-j) cVOCs, and k) ethene at NB1.

300

NB1- Black 500

e) 1000

400 300

100

200 0 0

3

17 157 247 365 561 TeCAs 1,1,1-TCA 1,1,2-TCA 1,2-DCA trans-1,2-DCE/1,1-DCA

b)

125 100

Conc. (µM)

Conc. (µM)

600 400

0

75 50 25 0

300

0

3

17 157 247 365 561

c)

250

900 800 700 600 500 400 300 200 100 0

1000

CCl4 Chloroform DCM

0

3

17 157 247 365 561

3

17 157 247 365 561

3

17 157 247 365 561

f)

0

g) Conc. (µM)

350 Conc. (µM)

800

200

150

NB1-White

1200

PCE TCE cis-1,2-DCE VC

a)

Conc. (µM)

Conc. (µM)

600

200 150

800 600 400

100 200

50

0

0 Total cVOCs Ethane Ethene

200

1000 800 600

100

400

50

200 0 0

3

17 157 247 365 561 Time (days)

0

3000

h)

150

0

1000

2500

800

2000

600

1500 400

1000

200

500

0

0 0

3

17 157 247 365 561 Time (days)

Fig. 3. Background and long - term post-injection concentrations for a – d) NB1-Black and e – h) NB1-White.

Total cVOCs (µM)

d)

17 157 247 365 561

Ethane/Ethene (µM)

3

Total cVOCs (µM)

Ethane/Ethene (µM)

250

0

Conc. (µmol/kg)

10000

1000

a) PCE

100

100

100

10

10

10

1

1

1

i) 1,1,2,2-TeCA

1

0 Background

1000 Conc. (µmol/kg)

1000

g) 1,1,1,2-TeCA

100

0 94 days

94 days

554 days

10000

Background 1000

b) TCE

e) Chloroform

100

0

0 Background

554 days

100

10

94 days

Background

554 days 1000

h) 1,1,1-TCA

100

100

10

10

1

1

94 days

554 days

j) 1,1,2-TCA

1 1 0

0 Background

94 days

0

0 Background

554 days

94 days

1000

1000 Conc. (µmol/kg)

1000

d) CCl4

554 days

Background

94 days

554 days

Background

94 days

554 days

f) DCM

c) cis-1,2-DCE 100

100 10 10

1 0

1 Background

94 days

554 days

Background

94 days

554 days

Fig. 4. cVOCs concentrations in soil for background and post-injection (94 and 554 days) samples collected between 2.5 and 4.5 m bgs. The box-and-whisker plot shows the median ( ), the interquartile range (box), and the extrema (whiskers). Dash lines are straight connectors between the medians.

Highlights •

CMC-S-nZVI was applied for the first time at a site contaminated with a mixture of cVOCs.

•

Compound specific isotope analysis (CSIA) confirmed the transformation of aqueous cVOCs.

•

Unlike nZVI, there was no accumulation of lower cVOCs, particularly VC, in the longterm degradation.

•

CMC-S-nZVI treatment resulted in significant reduction of sorbed phase cVOCs.

Declaration of interests ☒ The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. ☐The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: