Sulfidation of Iron-Based Materials: A Review of Processes and

Subscriber access provided by Eastern Michigan University | Bruce T. Halle Library

Critical Review

Sulfidation of Iron-Based Materials: A Review of Processes and Implications for Water Treatment and Remediation Dimin Fan, Ying Lan, Paul G. Tratnyek, Richard L. Johnson, Jan Filip, Denis Michael O'Carroll, Ariel Nunez Garcia, and Abinash Agrawal Environ. Sci. Technol., Just Accepted Manuscript • DOI: 10.1021/acs.est.7b04177 • Publication Date (Web): 16 Oct 2017 Downloaded from http://pubs.acs.org on October 16, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Environmental Science & Technology is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 44

Environmental Science & Technology

Sulfidation of Iron-Based Materials: A Review of Processes and Implications for Water Treatment and Remediation

1 2 3

Dimin Fan1, Ying Lan2, Paul G. Tratnyek2*, Richard L. Johnson2, Jan Filip3, Denis M. O’Carroll4, Ariel Nunez Garcia5, Abinash Agrawal6

4 5 6

1

Oak Ridge Institute for Science and Education (ORISE) Fellow, Office of Superfund Remediation and Technology Innovation, U.S. Environmental Protection Agency, 2777 Crystal Drive, Arlington, VA, 22202

7 8 9 10 11 12

2

OHSU/PSU School of Public Health, Oregon Health & Science University, 3181 SW Sam Jackson Park Road, Portland, OR 97239 3

13

4

14 15 16 17 18 19

Regional Centre of Advanced Technologies and Materials, Palacký University Olomouc, Šlechtitelů 27, CZ-78371 Olomouc, Czech Republic

5

School of Civil and Environmental Engineering, Connected Water Initiative, University of New South Wales, Manly Vale, NSW, 2093, Australia.

Department of Civil and Environmental Engineering, Western University, 1151 Richmond St., London, ON, Canada. 6

Department of Earth and Environmental Sciences, Wright State University, Wright State University, 3640 Colonel Glenn Highway, Dayton, OH 45435

20 21

*Corresponding author: Email: [email protected], Phone: 503-346-3431

22 23 24

Keywords: Sulfidation, Zerovalent iron, ZVI, Iron sulfide, Mackinawite, Dechlorination, Groundwater, Remediation, Water treatment

25 26 27

Critical Review for Environmental Science & Technology Revised Version (13-Oct-2017)

ACS Paragon Plus Environment


Page 2 of 44 Fan et al.

28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51

ABSTRACT Iron-based materials used in water treatment and groundwater remediation—especially micro- and nano-sized zerovalent iron (nZVI)—can be more effective when modified with lower-valent forms of sulfur (i.e., “sulfidated”). Controlled sulfidation for this purpose (using sulfide, dithionite, etc.) is the main topic of this review, but insights are derived by comparison with related and comparatively well-characterized processes such as corrosion of iron in sulfidic waters and abiotic natural attenuation by iron sulfide minerals. Material characterization shows that varying sulfidation protocols (e.g., concerted or sequential) and key operational variables (e.g., S/Fe ratio and sulfidation duration) result in materials with structures and morphologies ranging from core-shell to multiphase. A meta-analysis of available kinetic data for dechlorination under anoxic conditions, shows that sulfidation usually increases dechlorination rates, and simultaneously hydrogen production is suppressed. Therefore, sulfidation can greatly improve the efficiency of utilization of reducing equivalents for contaminant removal. This benefit is most likely due to inhibited corrosion as a result of sulfidation. Sulfidation may also favor desirable pathways of contaminant removal, such as (i) dechlorination by reductive elimination rather than hydrogenolysis and (ii) sequestration of metals as sulfides that could be resistant to reoxidation. Under oxic conditions, sulfidation is shown to enhance heterogeneous catalytic oxidation of contaminants. These net effects of sulfidation on contaminant removal by iron-based materials may substantially improve its practical utility for water treatment and remediation of contaminated groundwater.

52

2017-10-13


2

Page 3 of 44

Environmental Science & Technology Fan et al.

53

1. INTRODUCTION

54

Just like “contaminants of emerging concern”, most emerging technologies or areas of research

55

are not entirely new, but rather are of renewed interest because of other changes in science,

56

technology, or social or market factors. Sulfidation exemplifies most aspects of this dynamic,

57

ranging from increased appreciation of previously neglect research to heightened competition

58

among simultaneously-initiated studies on this topic. The overall goal of this review is to show-

59

case these dynamics, while at the same time advancing the field through meta-analysis of

60

existing data, cataloging the current state-of-the-art, identifying knowledge gaps, and

61

highlighting opportunities for future research and development.

62

In general, “sulfidation” (or sulfidization) can refer to any modification or transformation

63

of a metal-based material by exposure to sulfur compounds of various oxidation states, and is

64

most commonly used in geochemistry, corrosion science, and material science for metals with

65

strong affinities for sulfide. There are many such metals of environmental significance (Fe, Cu,

66

Zn, Hg, Ag, Tc, etc.), but iron is by far the most prominent among these. The earliest

67

(electronically indexed) example of this is the work by Rickard et al.1, 2 on the diagenesis of

68

anoxic marine sediments. In that work, they explicitly defined sulfidation as the complex

69

sequence of reactions leading from goethite (a-FeOOH), via metastable iron sulfides, to pyrite

70

(FeS2). Since then, an extensive literature has been developed on Fe and S interactions in

71

biogeochemical processes, mainly involving saturated soils and sediments, including

72

groundwater, but also extending to hyporheic zones, wetlands, and stratified surface waters

73

(e.g.3-6) A mostly separate, but fundamentally related and extensive body of literature concerns

74

sulfidation in the context of oil, gas, and nuclear energy production, which has been motivated

75

by the need to protect iron-based structures, such as pipelines and storage tanks, from both biotic

76

and abiotic sulfur-induced corrosion.7-11

77

In the last few years, interest in sulfidation for environmental applications has increased

78

sharply because of the potential benefits in water treatment and remediation processes—

79

including both reduction and oxidation reactions—resulting from controlled sulfidation of iron,

80

largely nano-scale zerovalent iron (nZVI)12-23 and recently extended to include micro-scale

81

zerovalent iron (ZVI)24, 25 and iron oxides (FeOx)26 (collectively referred to as “iron-based

82

materials” in this review). However, the majority of these studies introduce the sulfidation

2017-10-13


3



83

concept from the perspective of material synthesis and improvement, often with inadequate

84

appreciation and recognition of the fundamental similarities in chemistry between controlled and

85

natural sulfidation processes. In fact, the emerging interest in controlled sulfidation of iron-based

86

materials for water treatment stems from the convergence of two developments: (i) maturation

87

and diversification of the use of ZVI/iron oxides for water treatment and (ii) the role of iron

88

sulfide minerals in abiotic contaminant transformation process. Relevant literature in these two

89

areas is scattered across thirty years of research on reactivity of iron and sulfur in

90

biogeochemistry and environmental engineering. Therefore, it is informative to first summarize

91

those areas of research that are—in the context of this review—precursors to recent

92

developments of sulfidated iron-based materials for environmental applications.

93

2. PRECURSORS TO IRON-BASED MATERIALS SULFIDATION

94

Starting in the mid 1990s, the prospect of making permeable reactive barriers (PRBs) with ZVI

95

for passive in situ remediation of contaminated groundwater plumes triggered a surge in research

96

into the chemistry of this technology. Most of this work focused on the reduction of organic

97

contaminants coupled to oxidation of the Fe0 with charge-transfer through the partially-

98

passivating surface layer dominated by iron oxides.27 However, several groups recognized that S

99

incorporated in the oxide layer could play a significant role in the overall reactivity of ZVI with

100

contaminants, and reported preliminary results in favor of this hypothesis during the first

101

symposium on contaminant degradation by ZVI at the 1995 ACS National meeting in Anaheim,

102

CA, as well as in several peer-reviewed publications.28, 29 At the time, this work received little

103

attention, even though it soon became evident that sulfide was abundant in ZVI PRBs due to

104

stimulated microbial reduction of sulfate.30-39 While some other studies used dithionite to treat

105

sediment and surface soil both in laboratory and in situ—and resulting in sulfidation effects on

106

contaminant degradation—earlier work used dithionite mainly to reduce structural iron, and

107

emphasized the reactivity of reduced iron with contaminants.40-42 Later, sulfidation of ZVI and

108

its effects on dechlorination was investigated in a series of studies that included extensive

109

characterization of the sulfidated iron phases, by spectroscopy and electrochemistry,43, 44 but

110

even this work received little recognition at the time.

111 112

Although ZVI has become the model system used in most studies of abiotic reduction of organic contaminants,45 the largely-independent, parallel literature on contaminant removal by

2017-10-13


4

Page 5 of 44


113

reducing S(-II) species starts earlier and is nearly as extensive. The earliest work in this area was

114

dominated by the dissertations of Barbash,46 Roberts,47 and Kriegmen-King,48, 49 which covered

115

the mechanisms and kinetics of abiotic dehalogenation of alkanes and alkenes by aqueous sulfide

116

and pyrite. Subsequent laboratory work focusing on freshly precipitated nanocrystalline

117

mackinawite (FeS) has confirmed repeatedly that the high surface area FeS is reactive with

118

halogenated organic compounds, such as carbon tetrachloride (CT), trichloroethene (TCE),

119

tetrachloroethylene (PCE), and ethylene dibromide (EDB).50-62 Similarly designed laboratory

120

studies have found that FeS can also be effective at sequestration of various metals, metalloids,

121

and oxyanions, which occurs through a combination of reduction, adsorption and coprecipitation

122

processes.63-72 These processes certainly apply to ZVI PRBs where sulfate reduction is often

123

significant, but they also are relevant in scenarios where FeS forms under natural conditions,

124

such as hyporheic zones6 and the rhizosphere in wetlands73. The full scope of potential

125

applications of iron sulfides to groundwater remediation—including organics, metals, and

126

radionuclides—was recently reviewed, with emphasis on the applications of FeS nanoparticles

127

for optimal remedial performance.74

128

The favorable effects of Fe/S combinations on contaminant degradation/sequestration

129

identified from laboratory studies have recently been utilized in commercial technologies for in

130

situ remediation of contaminated groundwater. One example of this is the process termed “in situ

131

biogeochemical transformation” (ISBGT),75-77 which involves creating fresh, high surface area

132

biogenic iron sulfides that react directly with chlorinated solvents. Another commercial

133

application is a microscale ZVI (mZVI) product that contains a slow release sulfate source,

134

which is designed primarily for in situ sequestration of metals. Upon injection of this product,

135

the ZVI initially creates strongly reducing conditions, and microbial sulfate reduction is

136

stimulated in the longer-term. This causes in situ sulfidation of ZVI, which favors precipitation

137

of stable low solubility metal sulfide phases.78, 79

138

The prior work on contaminant degradation with ZVI and FeS, which is summarized

139

above, provides the foundation for recent work on contaminant transformations by sulfidated

140

iron-based materials. In early work, sulfidation was considered only as an in situ and/or natural

141

biogeochemical process; whereas interest in sulfidation is now also focused on formulating

142

engineered materials that are composed of multiple phases—especially those based on nZVI—

143

with properties that are optimized for environmental applications. To provide a rational basis for 2017-10-13


5



144

this process of material development, it is necessary to develop a fundamental understanding of

145

the various sulfidation methods and their mechanisms.

146

3. FUNDAMENTALS OF SULFIDATION

147

3.1. Methods. Sulfidation has been performed using a variety of sulfidation agents, iron-based

148

materials, and sulfidation process. The combinations that have been (or are soon to be) published

149

are summarized in Table 1, arranged in order of decreasing average valence state on S in the

150

sulfidation agent. Another important organizational theme in Table 1 is the sulfidation process

151

used for sulfidation, which includes aqueous-solid, aqueous-aqueous, solid-solid, and gas-solid.

152

The former two encompass the majority of the sulfidation studies conducted so far and thus will

153

be the main focus of this review. The latter two methods represent recent and ongoing

154

developments, which are included for the sake of completeness.

155

Table 1. Methods of sulfidating iron-based materials for (ground)water treatment. Sulfidation Agent (S valence) a

Material Sulfidated

Sulfidation Conditions

References b

Sulfate SO42− (+VI)

Granular ZVI

34

FeOx

Aqueous-Solid (microbial)

Sulfur Dioxide SO2 (+IV)

nZVI

Gas-Solid

Dithionite S2O42− (+III)

nZVI

Thiosulfate S2O32− (+II) c Elemental Sulfur S0 (0)

156 157 158 159

75 80

Aqueous-Aqueous,

14-16, 19-21, 81-85

Aqueous-Solid,

86, 87

nZVI

Aqueous-Solid and Aqueous-Aqueous

18

FeOx

Aqueous-Aqueous

22, 26, 88

mZVI

Solid-Solid

Sulfide, bisulfide, and polysulfides S2− (−II)

nZVI mZVI

Thioacetamide H3CC(=S)NH2 (−II)

nZVI

25 12, 13, 17, 89

Aqueous-Solid

FeOx

24, 43 90

Aqueous-Solid

91

a

Average valence of sulfur in parenthesis. Precursory work, only representative references are included. c Average sulfur oxidation state does not represent the oxidation state of individual sulfur atoms in thiosulfate. b

2017-10-13


6

Page 7 of 44


160 161

Aqueous-solid phase sulfidation processes involve reactions between dissolved sulfur

162

species and solid iron materials, like ZVI. Commonly used aqueous sulfur compounds include

163

sulfide (either biogenic or abiotic)12, 13, 17, 23, 34, 75, dithionite86, 87, and thiosulfate18. This process

164

has been described as “post-synthesis” 18, and the resulting material is referred to as S-(n)ZVI

165

hereafter. In contrast, aqueous-aqueous sulfidation processes start with soluble sulfur and iron

166

species, in the presence of a strong reducing agent (e.g., NaBH4) if Fe(0) is desired. Both

167

dissolved Fe(II) and newly formed Fe(0) may react concurrently with sulfur to form FeS. This

168

process has been described as “one-pot synthesis”14 or “pre-synthesis sulfidation”18, and the

169

resulting material is referred to as Fe/FeS hereafter. Most of the published studies that fall into

170

this group used dithionite as the sulfidation agent source. The solution chemistry of this process

171

is highly complex due to multiple reaction intermediates of dithionite; however, due to the ease

172

of material preparation, this method has been frequently selected. In the absence of a strongly

173

reducing agent and under high pH, sulfidated iron oxides are synthesized. 26, 88 The new and

174

novel method of sulfidation by ball-milling ZVI and S(0) is a “two-step” process, but could form

175

relatively uniform materials as the aqueous-aqueous method, depending on the degree of

176

mechanochemical mixing.25

177

3.2. Mechanisms. Most of what is known about sulfidation mechanisms applies to aqueous-

178

solid sulfidation methods because of its analogy to sulfur induced corrosion, which has been

179

extensively studied in the literature on corrosion of ferrous-metals in sulfidic environments.8, 9, 11

180

Regardless of the sources of sulfur or iron, the reaction between the two is essentially a redox

181

process at the aqueous-solid interface, governed by their thermodynamic equilibrium (Figure 1).

182

Taking sulfide, the most common sulfur source, as an example, it serves as a reductant during

183

sulfidation of Fe(III) oxides (e.g., goethite) by first reducing Fe(III) oxides to Fe(II) via

184

interfacial electron transfer.4, 92 Sorbed sulfide is oxidized to form S(0). If excess aqueous sulfide

185

is available, it reacts with the produced Fe(II) to form FeS,2 and FeS may continue to transform

186

to greigite (Fe3S4) and/or pyrite (FeS2).93 As for sulfidation of ZVI, the process resembles sulfur-

187

induced corrosion under alkaline conditions,44 which proceeds via two steps.44, 94-96 The initial

188

step involves rapid adsorption of HS- onto the Fe(0) surface, replacing previously adsorbed OH-

189

(eq. 1), which then forms a thin layer of FeS film on the surface of Fe(0) (eq. 2 and 3).

190

) ) Fe − OH&'( + HS ) ↔ Fe − HS&'( + OH )

2017-10-13


(1) 7



191

) Fe − HS&'( ↔ Fe − HS&'( + 2𝑒 )

(2)

192

Fe − HS&'( → Fe II − S2)3 + x HS ) + 1 − x H -

(3)

193

The enhanced localized acidity (eq. 3) can induce localized corrosion of Fe(0), which produces

194

dissolved Fe(II) that then reacts with HS- to form additional FeS. It has been suggested that the

195

morphological characteristics and protectiveness of the FeS film resulting from these two

196

processes of FeS formation are very different.94, 95 Depending on the sulfidation conditions and

197

duration, the FeS resulting from aqueous precipitation can be either highly corrosive or

198

protective.95-97 The implication of these dynamics to contaminant transformation by sulfidated

199

iron-based material will be discussed in more details later (Section 5 and 6).

200 201 202 203 204

205

Figure 1. Eh-pH diagram for the Fe-S-H2O system at 25°C. Hematite and pyrite are suppressed. Figures were drawn with fixed-concentration species (A: [ST] = 10−6 M, B: [FeT] = 10−6 M), and contours representing total ions (A: [FeT], B: [ST]) were showed in different colors: yellow = 10-6 M, red = 10-5 M, and blue = 10-4 M.

Compared to sulfide, the sulfidation mechanisms by dithionite are more complicated due

206

to its complex aqueous chemistry and possibly additional interactions between Fe(0) and

207

dithionite or its sulfur intermediates. It is well documented that dithionite at low pH mainly

208

undergoes disproportionation to form thiosulfate (eq. 4) and sulfite, whereas it is relatively stable

209

at high pH and slowly decomposes to sulfite and sulfide (eq. 5). Either scenario could be relevant

210

depending on whether dithionite is used in aqueous-aqueous or aqueous-solid sulfidation. The

211

former scenario is more relevant during aqueous-aqueous sulfidation as the initial solution pH is 2017-10-13


8

Page 9 of 44


212

acidic and gradually rises to alkaline as reaction proceeds, which may result in different sulfur

213

species at different stages of particle formation. For aqueous-solid sulfidation, dithionite is added

214

to the nZVI suspension that is already alkaline after synthesis (due to synthesis reaction), where

215

Fe(0) may accelerate the decomposition reaction of dithionite by precipitating aqueous sulfide to

216

form FeS. Given the difficulty in tracking and quantifying various sulfur containing

217

intermediates, explanations of the mechanism of sulfidation remained speculative and the

218

reaction stoichiometry is unclear.

219

7) ) 2 S7 O7) 8 + H7 O → S7 O9 + 2 HSO9

(4)

220

7) ) 7) 3 S7 O7) + 3 H7 O 8 + 6 OH → 5 SO9 + S

(5)

221

Thiosulfate (S2O32−), a strong corrosion promotor under acidic conditions,8 is relatively

222

stable at high pH, which is more relevant to the aqueous-solid sulfidation scenario. However, in

223

the presence of Fe(0), thiosulfate is readily reduced to HS-, which is more stable according to the

224

thermodynamic speciation modeling.8 Experimental evidence for this was presented in an early

225

study on corrosion of Fe-17Cr alloy by thiosulfate at (circum)neutral pHs.98 In that study, it was

226

found that thiosulfate was only corrosive when reduced to sulfide, and that the reduction reaction

227

only occurred on the bare surface of the metal, not on the passive film, suggesting that the metal

228

might be the source of the reductant. If these results apply under environmental conditions, Fe(0)

229

should reduce thiosulfate to sulfide before sulfidation takes place. Similar redox reactions may

230

also occur for gas-solid sulfidation (e.g., between pre-formed Fe(0) and sulfur dioxide at elevated

231

temperature80) and solid-solid sulfidation (e.g., between Fe(0) and elemental sulfur (S(0)) by

232

mechanochemical processes such as ball milling25).

233

4. CHARACTERIZATIONS OF SULFIDATED IRON-BASED MATERIALS

234

Compared to the material characterizations conducted in the precursory work on sulfidation of

235

iron-based materials which was mostly limited to identifying FeS minerals in ZVI-based

236

PRBs,30, 33, 34 more recent studies on sulfidation of iron-based materials (mostly nZVI) have

237

placed greater emphasis on qualitative and quantitative characterizations of the properties of

238

sulfidated materials. These materials were prepared under a wide range of conditions—including

239

processes, operational and experimental variables—some of which are summarized in Table 1.

240

A holistic examination of these characterization results is presented below to identify common

2017-10-13


9



241

features and differences to determine the major factors that control properties of sulfidated

242

materials. The following section focuses on two most relevant aspects of characterizations

243

featuring nZVI as an example: mineralogy and morphology. Other characterization methods

244

have been used—such as electrochemistry15, 81, 99—but are not included.

245

4.1. Bulk Mineralogy and Surface Composition. Bulk characterizations of mineralogy and

246

composition of sulfidated nZVI mainly concern two parts: the characteristics of FeS and the

247

amount of Fe(0) remaining. The FeS phase formed in sulfidation studies are mainly amorphous

248

or nanocrystalline mackinawite, according to X-ray diffraction (XRD) data.12, 20, 83 It is usually

249

the first metastable iron sulfide phase formed in sulfidic environments.2 Further transformation

250

to crystalline mackinawite and more stable pyrite requires extended period of reaction time that

251

is often more relevant in corrosion studies.90, 97, 100 Given the limited utility of XRD in

252

characterizing amorphous phases and phases present in low content (e.g., as a thin surface layer),

253

X-ray photoelectron spectroscopy (XPS) is usually used to probe the changes of surface species

254

and chemical states that take place on the surface of sulfidated materials. Comparing the XPS

255

spectra collected in various studies shows consistency in the S2p XPS spectra of sulfidated nZVI,

256

generally indicating the presence of monosulfide (S2−), disulfide (S22−), and other polysulfide

257

species, regardless of sulfur sources or processes.12, 14, 17, 18, 20, 81, 86 In contrast, their Fe 2p spectra

258

counterparts can be clearly divided into two groups (Figure S1). One group shows much lower

259

FeS peak than the other, suggesting lower surface abundance of FeS.12, 14, 82, 86 It was found that

260

the spectra with weak FeS signal mainly came from studies that either used aqueous-aqueous

261

sulfidation or used aqueous-solid sulfidation but with short sulfidation duration; whereas a larger

262

FeS peak was observed only in studies that utilized high S/Fe ratio and long sulfidation duration.

263

This is expected because aqueous-aqueous sulfidation results in simultaneous precipitation of

264

FeS and Fe(0) and hence more uniform distribution of S. For aqueous-solid sulfidation, longer

265

sulfidation duration may accumulate more FeS on the surface, leading to a larger peak from FeS

266

and shielding of the spectral band of Fe(0).

267

The remaining Fe(0) content after sulfidation is the other important parameter, as Fe(0)

268

accounts for the majority of the reducing capacity of nZVI. Although it is expected that the

269

remaining Fe(0) would decrease with increasing S/Fe ratio, several studies have shown that the

270

degree of sulfidation plateaued beyond a given S/Fe ratio. For example, for laboratory prepared

2017-10-13


10

Page 11 of 44


271

nZVI, Mössbauer spectra of S-nZVI showed that a S/Fe ratio of 1.12 after 24 h sulfidation

272

resulted in only 10% of Fe(0) transforming to FeS, which is the same as the S/Fe ratio of 0.112,

273

likely due to inhibited corrosion at high S/Fe ratio.12 A similar observation was also reported for

274

nZVI synthesized on carboxylmethylcelluose (CMC-nZVI)—using hydrogen measurements

275

after acidification—that the low and high S/Fe ratio (0.33 and 1) recovered similar amount of

276

Fe(0) when sulfide was used for sulfidation.13 However, when dithionite was used for

277

sulfidation, substantially more Fe(0) is converted to FeS at both low and high S/Fe ratios,

278

suggesting that dithionite is a more aggressive sulfidation reagent than sulfide.13

279

4.2. Microscopic Structure and Morphology. Microscopic characterizations from existing

280

sulfidation studies further support the bulk characterization results by showing that the type of

281

sulfidation process and two main operational variables (S/Fe ratio and sulfidation duration) play

282

major roles in affecting the structure and morphology of sulfidated materials. This is illustrated

283

in Figure 2, which shows conceptual models of S-nZVI prepared using aqueous-solid methods

284

(Figures 2C, 2E and 2F) and Fe/FeS prepared using aqueous-aqueous methods (Figure 2G),

285

arranged as a function of sulfidation duration (y-axis) and S/Fe ratio (x-axis). Each conceptual

286

cartoon describes the key features of the material, and is accompanied by a published

287

transmission electron microscopy (TEM) image, when available, of sulfidated material prepared

288

under similar conditions. To highlight the effects of operational variables while minimizing other

289

variables, only nZVI synthesized by borohydride reduction is included among the TEMs shown.

290

For unmodified nZVI, it has been shown repeatedly that exposure in water results in

291

shrinking of the Fe(0) core and increasing thickness of the oxide shell,101-104 which can be seen in

292

Figures 2A, 2B. When sulfide was introduced to react with the nZVI (i.e., aqueous-solid

293

sulfidation), the resulting materials appear to fall into two groups. The short sulfidation duration

294

appears to preserve the chain-like and the core-shell structure of the original nZVI (Figures 2C

295

and 2E). Comparison between Figures 2C and 2E shows no significant morphological change

296

between S/Fe ratios of 0.04 and 0.08. However, residual aqueous sulfur measurements at the end

297

of sulfidation did show increasing sulfur sequestration by nZVI at higher S/Fe ratio, likely

298

suggesting increasing FeS coverage or thickness.17 With increasing sulfidation duration,

299

sulfidation is expected to continue, resulting in precipitation of secondary FeS, which has a

2017-10-13


11



300

distinct flake-like structure and is associated with the nZVI-like structure (Figure 2F: S/Fe =

301

0.112 and t = 24 h).

302 303 304 305 306 307 308 309 310

311

Figure 2. Conceptual model of sulfidated nZVI prepared using aqueous-solid sulfidation (C-F) and aqueous-aqueous sulfidation (G), with corresponding TEM images. (A) nZVI core with thin FeOx shell (TEM105: freshly prepared nZVI) (B) nZVI core with thick FeOx shell (TEM12: nZVI in synthetic groundwater buffered at pH 7.9, t = 24 h); (C) S-nZVI with low S/Fe ratio and short sulfidation duration (TEM17: S/Fe = 0.04, t = 10 min); (D) S-nZVI with low S/Fe ratio and long sulfidation duration; (E) S-nZVI with high S/Fe ratio and short sulfidation duration (TEM17: S/Fe = 0.08, t = 10 min); (F) S-nZVI with high S/Fe ratio and long sulfidation duration (TEM12: S/Fe = 0.112, t = 24 h) (G) Fe/FeS, formed via aqueous-aqueous synthesis (TEM20: S/Fe = 0.28).

The morphological transition over time is similar to the process described during sulfur

312

induced corrosion.44, 95, 100 In this process, a thin layer of FeS film forms initially upon brief

313

exposure of steel to sulfide by chemisorption.44, 95 A secondary FeS layer may then form,

314

provided extended exposure in sulfidic environment. A similar process has also been described

315

for the morphological changes of silver nanoparticles (AgNP) during sulfidation with increasing

316

S/Ag ratio.106 The formation of a thin sulfur layer on S-nZVI was also observed in our

317

preliminary characterizations of oxide-free nZVI (prepared by thermal reduction) treated with

318

1% sulfide for 5 min (Figure S2). However, elemental mapping indicates that the sulfur layer

2017-10-13


12

Page 13 of 44


319

was underneath an oxygen-rich layer. This is inconsistent with the XPS results of S-nZVI

320

(synthesized by borohydride reduction), which show sulfur abundance decreased with increasing

321

depth from the surface17. The reason for this discrepancy is unclear, but it could arise because the

322

oxide-free nZVI undergoes simultaneous sulfidation and dissolution/reprecipitation.

323

For Fe/FeS formed via aqueous-aqueous methods, microscopic characterization revealed

324

that the resulting material contains both spherical and flake-like structures (Figure 2G), which

325

appear to be analogous to the original Fe(0) and the secondary FeS, respectively, shown in

326

Figure 2F (S/Fe = 0.112, t = 24 h). However, energy dispersive X-ray spectroscopy (EDX)

327

showed that the core actually has slightly higher S content than the flakes,20 which cannot be

328

explained by the Fe(0)-FeS core-shell model used to describe materials made by the aqueous-

329

solid method. The authors proposed an alternative hypothesis that as sulfidation proceeds, it

330

results in a core that is mainly composed of Fe(0) and S(0), and flakes that are a mixture of FeS

331

and iron (hydr)oxides.20 To differentiate from the core-shell structure formed during aqueous-

332

solid sulfidation, this structure is referred to as Fe/FeS multiphase. The distinctly different

333

patterns of S distribution observed with the different sulfidation processes agrees with the

334

reported XPS Fe 2p spectra (Figure S1) showing that aqueous-aqueous sulfidation results in

335

lower abundance of surface FeS despite comparable or even a higher S/Fe ratio than aqueous-

336

solid sulfidation. Surface layers are also less prominent in materials sulfidatedd by

337

mechanochemical mixing (i.e., ball milling).25 Together, they confirm that different sulfidation

338

processes result in particles with distinct properties, which may have important implications to

339

their reactivity with contaminants.

340

5. EFFECTS OF SULFIDATION ON CONTAMINANT TRANSFORMATION

341

5.1. Metals and Metalloids. Mechanism, Product, and Kinetics. The currently available studies

342

on removal of metals and metalloids by sulfidated iron-based materials12, 20, 83, 107, 108 suggest that

343

sulfidation can have profound effects on mechanisms, products, and kinetics in these systems.

344

An early and prominent example investigated the sequestration of pertechnetate (TcO4–, a

345

Tc(VII) metal oxyanion and long-lived radionuclide of concern) by nZVI with and without

346

sulfidation (using sulfide).12 In both systems, Tc(VII) was reduced to Tc(IV), but the speciation

347

of the Tc(IV) shifted from TcO2 to TcS2 with increasing S/Fe ratio. TcS2 became the dominant

348

product at S/Fe as low as 0.05, suggesting that the reduction of TcO4– by FeS is favorable 2017-10-13


13



349

enough to be the preferred sequestration pathway even in the presence of nZVI. This resulted in

350

increased rates of TcO4– removal up to S/Fe = 0.112, but TcO4– removal rates decreased at higher

351

S/Fe, which was shown to be due to inhibition by excess aqueous sulfide.12 A similar peak in

352

removal rates at intermediate S/Fe was observed for the metal oxyanion chromate (CrO4–), also

353

at S/Fe ≈ 0.11, although Cr(III) oxyhydroxides were the only products across the whole range of

354

S/Fe.23

355

Among metal cations of environmental concern (Cu, Pb, Cd, etc.), many interact strongly

356

with S due to the low solubility of metal sulfide phases,109 but only cadmium (Cd) has been used

357

to investigate the effect of sulfidation on sequestration of metals by nZVI.20 Using dithionite as a

358

sulfidation agent, the quantity of Cd removed after 2 hours of contact with Fe/FeS was found to

359

decrease and then increase with S/Fe, resulting in a minimum at S/Fe ≈ 0.07. This trend, while

360

opposite to what has been seen with the removal kinetics of metal oxyanions (e.g., Tc), is

361

suggested to be also caused by change of sequestration mechanisms. In this case, it was proposed

362

Cd sequestration at low S/Fe is dominated by adsorption and surface complexation on nZVI,

363

whereas displacement of Fe(II) in FeS becomes dominant at high S/Fe ratios, resulting in the

364

formation of (Fe, Cd)S phases, which were detected by XPS.20 Although the metals discussed above only represent a small subset of metals and

365 366

metalloids of environmental concern, the reported effects of sulfidation should also apply to

367

other metals and metalloids because of similar underlying mechanisms that control their removal

368

by iron-based materials. For example, it is well established that the removal mechanisms of

369

metals/metalloids by unsulfidated nZVI are primarily controlled by their reduction potentials.103,

370

110

Metals/metalloids with significantly higher standard reduction potentials than Fe(0), such as

371

Cu(II), Ag(I), and SeO42–, are removed predominantly by reductive immobilization,111, 112

372

whereas those with similar or lower standard reduction potential than Fe(0) (e.g., Zn(II)), are

373

removed primarily via non-redox based adsorption or precipitation.103 In contrast, for FeS, the

374

solubility of the metal sulfide phases is shown to govern the removal mechanisms.109 Thus it is

375

expected that sulfidation of iron-based material may shift the removal of the metals and

376

metalloids (especially those have strong affinity to sulfide, such as Hg) from a redox potential

377

controlled process to a solubility controlled process (Figure 3). For metalloids that have multiple

378

oxidation states, such as arsenic or selenium, more complexity could arise as different

2017-10-13


14

Page 15 of 44


379

experimental variables (e.g., pH, contaminant concentrations, redox condition, iron types or

380

sulfur levels) could all result in a diverse range of sequestration pathways and products for both

381

sulfidated and unsulfidated materials, as shown by multiple studies.68, 113, 114 This diversity will

382

become a challenge as new studies on contaminant metal sequestration by sulfidated ZVI

383

become available, but general principles (e.g., structure-activity relationships for metals

384

sequestration, such as in Kim et al.15) should eventually emerge.

-0.5

Zn2+/Zn0

-1.0 2+

Mn /Mn

0

Non-reductive ↔ Reductive

0.0

Cu2+/Cu0 UO2(OH)2/UO2 TcO4–/TcO2 HAsO42+/H3AsO3 Pb2+/Pb0 Ni2+/Ni0 Cd2+/Cd0 Fe2+/Fe0

0 MnS As2S3 FeS NiS ZnS CdS PbS CuS

-20

-40

Ag2S HgS

-60

-80

Log (Solubility of Metal Sulfide)

0.5

Hg2+/Hg0 Ag+/Ag0 Sb(V)/Sb(III) Cr(VI)O4/Cr(OH)3

Original Species ↔ Metal Sulfide

Reduction Potential at pH 7 (mV)

1.0

Sb2S3

-1.5

-100

Increasing Degree of Sulfidation

385

390

Figure 3. Sequestration mechanisms for representative metal and metalloid contaminants by ZVI and FeS that indicates likely shift of sequestration mechanism upon sulfidation (Left axis: the redox couples are positioned based on their formal reduction potential at pH 7; Right axis: the metal sulfide phases are positioned based on their solubility constant; The color gradient represents gradual shift from one sequestration mechanism to another).

391

Stability of Sequestration Products. Sequestration of metals and metalloids by sulfidated iron-

392

based materials is favorable under anoxic, reducing conditions. However, subsurface redox

393

conditions can fluctuate between reducing and oxidizing as a result of natural (e.g., flooding,

394

surface water intrusion, seasonal water-table fluctuations) or anthropogenic (e.g., dredging)

395

events, and these changes may affect the long-term stability of sequestrated metals and

396

metalloids.107, 115, 116 Under non-sulfidic conditions, oxidative remobilization of metals—e.g., U

397

and Tc—has been demonstrated by common natural oxidants, including dissolved oxygen,

386 387 388 389

2017-10-13


15



398

manganese dioxide (MnO2), and nitrate that oxidize these metals, either abiotically or

399

biotically.117-123

400

Among the metals for which sequestration by sulfidated ZVI is well characterized,

401

reoxidation of the sequestration products has been studied only for Tc.107 Starting with the

402

multiphase material formed from TcO4– exposed to S-nZVI it was shown that Tc reoxidation

403

became slower with increasing Fe/S, which is consistent with other evidence that TcS2 is less

404

susceptible to oxidation than TcO2. It was shown that oxidation of TcS2 occurs via TcO2, which

405

may contribute to the slower oxidation rates of the former. It was also found that FeS acts as a

406

redox buffer by scavenging dissolved oxygen, which further inhibits Tc reoxidation. The latter

407

was evidenced by a lag phase, during which no aqueous Tc was measured, followed by a sharp

408

increase in aqueous Tc that was concurrent with the rapid increase of solution oxidation-

409

reduction potential (ORP). Similar observations have been reported for UO2, Cr(VI), and As(III)

410

in pure FeS systems.118, 121, 124, 125 For As(III), the resulting As(V) sorbs strongly to the iron

411

oxyhydroxides that form at high pH, thereby prevented remobilization.125 However, for Cr(VI)

412

reduction by FeS, where Fe-Cr solids were formed, upon exposure to the oxidant, significant

413

dissolved Cr(VI) was released;121 for UO2, depletion of FeS also results in complete

414

remobilization of uranium.118 This range of results suggest that the long-term stability of metals

415

and metalloids sequestered by sulfidated iron-based materials may vary significantly depending

416

on the characteristic interactions between metals/metalloids and iron mineral phases.

417

5.2. Organic Contaminants. To date, most of the interest in sulfidation of iron-based materials

418

arises from its beneficial effects on the degradation of organic contaminants, where the collective

419

benefits may contribute substantially to overcoming the issues that have limited application of

420

existing iron based remediation and water treatment technologies.126 The development of this

421

technology has followed two parallel paths, focusing on reduction and oxidation, respectively.

422

Reduction is mostly relevant in groundwater treatment under anoxic conditions, where sulfidated

423

(n)ZVI is used as the primary reductant to degrade halogenated compounds (mainly chlorinated

424

aliphatic compounds but recently including a couple of brominated flame retardants as well as

425

one antibiotic).13-18, 81, 82, 127 Oxidation mostly concerns surface water and wastewater treatment

426

under oxic conditions, where sulfidated iron oxides or (n)ZVI serve as catalysts in oxidation

427

process for treating a variety of organic contaminants.22, 26, 85, 88, 89, 128 This section discusses both

2017-10-13


16

Page 17 of 44


428

processes but with the main focus on reduction as sulfidated iron-based materials are directly

429

involved in contaminant degradation. In addition, literature on reduction is more abundant to

430

allow a more in-depth analysis.

431

Reduction Mechanisms. The products, pathways, and mechanisms of abiotic dechlorination of

432

chlorinated aliphatic compounds have been characterized extensively using model systems based

433

on (n)ZVI and various iron oxides and sulfides.129, 130 In all of these systems, reductive β-

434

elimination pathways generally predominate over hydrogenolysis pathways for chlorinated

435

ethenes like TCE 130-133 so a major change in dechlorination pathway is not expected upon

436

sulfidation of (n)ZVI. This is consistent with results reported to date on TCE reduction by

437

sulfidated nZVI, which identified mainly the β-elimination products, such as acetylene, ethene,

438

and ethane.18, 25 Abiotic dechlorination of CT generally occurs mainly by hydrogenolysis for

439

both ZVI134 and FeS130, and thus is expected to apply to sulfidated iron-based reductants as well.

440

Both reductive elimination and hydrogenolysis have also been reported for abiotic

441

debromination, and a series of studies on brominated flame retardants has shown that the

442

branching between these pathways is not greatly affected by sulfidation, although it is strongly

443

influenced by the degree of halogenation of the specific contaminant.81, 82

444

From the reductant perspective, there has been a great deal of work on identification of

445

the specific reducing species (Fe(0) or Fe(II) surface species, surface bound atomic hydrogen,

446

S(-II) species, etc.) that are responsible for abiotic dehalogenation by ZVI, iron oxides, and

447

sulfides.132, 135-137 This background should also apply to sulfidated iron-based reductants, but

448

some clear differences are expected depending on material composition. For example, there is

449

evidence for both H-transfer and electron-transfer from (n)ZVI,136, 138 whereas FeS can donate

450

electrons but does not produce atomic H. For (n)ZVI system, there is evidence that the relative

451

significance of contaminant reduction by direct electron transfer versus the indirect route via

452

activated hydrogen varies with a variety of factors, especially contaminant type. In particular,

453

early studies with (n)ZVI concluded that the primary TCE degradation mechanism was via

454

atomic hydrogen whereas degradation of CT proceeds via electron transfer.137-140 For FeS (in the

455

absence of (n)ZVI), dechlorination must be by electron transfer, although it is not entirely clear if

456

the donor sites are Fe or S.51, 62, 81 For sulfidated (n)ZVI, the relative contribution of atomic

457

hydrogen versus Fe (or S) electron-donor surface sites to contaminant reduction is an active area

2017-10-13


17



458

of research and debate, as elaborated below. New work on this topic is expected, in part because

459

a better understanding of the beneficial effects of sulfidation on contaminant degradation will aid

460

in its optimization for practical applications in remediation.

461

Reduction Kinetics. Recent work on sulfidation, like the many studies using various ways to

462

enhance the utility of ZVI for remediation (bimetallic amendments, nano-sized particles,

463

oxygenation, etc.), have emphasized increased rates of contaminant removal as a major benefit.14,

464

17, 18, 25

However, most of these conclusions are based on comparisons between control and

465

sulfidation effects under a narrow range of controlled conditions, so the generality of such results

466

merits further consideration. To address this, we compiled most of the kinetic data that are

467

currently available for reduction of chlorinated organics by iron-based reductants, and subjected

468

them to the meta-analysis shown in Figure 4. By far, the majority of data available are for CT

469

and TCE, which are summarized in the format of surface area normalized rate constant (kSA) vs.

470

mass normalized rate constant (kM) in Figures 4A and 4B, respectively. For both contaminants,

471

the available data obtained with sulfidated reductants are shown with open markers and solid

472

circles are used for non-sulfidated iron-based materials (including nZVI, microscale ZVI of

473

various purities, FeS, and iron oxides). For CT (Figure 4A), the data for sulfidated materials

474

form clusters roughly coincident with data obtained with non-sulfidated forms of the same

475

material, suggesting that sulfidation does not generally increase rates of CT degradation. In

476

contrast, the TCE kinetic data (Figure 4B) show at least one order of magnitude increase in kM

477

and/or kSA for sulfidated materials compared with their non-sulfidated analogs, and there is

478

sufficient data to conclude that this enhancement probably applies over most conditions. To

479

further illustrate and extend this analysis, Figure 4C shows the ratio of kM for sulfidated versus

480

non-sulfidated nZVI for CT and TCE and the other contaminants for which a limited amount of

481

kinetic data are available. In this representation, the diagonal contours represent the enhancement

482

factor, R (defined in Ref24). Figure 4C clearly shows that R for TCE is significantly greater than

483

1 (~50), CT and HBCD are roughly 1 (i.e., no enhancement), and the data on other contaminants

484

are too sparse to generalize, except that the overall range in R is roughly from 50 to 0.1,

485

depending on the S/Fe ratio (discussed in detail later).

486 487

The contrasting effect of sulfidation on reduction rates of TCE and CT was rationalized as a reflection of differences in dechlorinating species for these two contaminants. It was

2017-10-13


18

Page 19 of 44


488

hypothesized that surface sulfur poisons the recombination of atomic hydrogen to form H2,

489

thereby leaving more atomic hydrogen at the surface for TCE degradation.18 This hypothesis is

490

supported by data showing (i) no enhancement of CT reduction (which occurs predominantly by

491

electron transfer) at S/Fe of 0.05, (ii) a similar enhancement of TCE degradation rates by adding

492

another known poison for atomic hydrogen recombination (i.e., arsenic141), and (iii) the decrease

493

in H2 formation.18 However, this hypothesis may be inconsistent with other data, such as the

494

increasing TCE reduction rate by S-nZVI with increasing pH, which has been observed with

495

both S-nZVI17 and Fe/FeS16. This pH dependence is opposite that for H2 formation by ZVI

496

corrosion (and therefore, presumably, the availability of surface H, which is an intermediate in

497

the formation of H2). Furthermore, the observed pH dependence of TCE degradation by

498

sulfidated nZVI is consistent with the pH effect observed with pure nanocrystalline FeS51, but

499

opposite to the trend with unmodified nZVI140. Prior studies on TCE degradation by pure FeS

500

have noted that higher pH increases the negative charge of FeS surface, thereby providing

501

increased driving force for electron transfer.51, 54 These results favor an alternative hypothesis

502

proposed by Rajajayavel et al.,17 that TCE reduction by S-nZVI occurs predominantly via FeS,

503

which serves as a mediator of electron transfer from Fe(0) to FeS-sorbed contaminants. This also

504

is consistent with the high electron conductivity of FeS142, and the inhibited H2 production from

505

S-nZVI may simply be due to FeS blocking direct access of water to Fe(0) sites (see conceptual

506

model presented below).

507

The operational variable that seems to have the greatest overall effect on the kinetics of

508

contaminant reduction by sulfidated nZVI is the S/Fe ratio. For all contaminant types, sulfur

509

sources, and sulfidation methods, contaminant reduction rates increase from low to moderate

510

S/Fe ratios and then plateau or decrease at high S/Fe ratios. For most contaminants shown in

511

Figure 4C, this trend is hard to track in the figure due to the closeness of data points, but can be

512

observed by comparing the kM in Table S1. One distinct example is CT (Figure 4C), the data

513

suggest a small enhancement at the lowest S/Fe ratios, and inhibition at increased S/Fe ratios (i.e.,

514

R ≤ 1).143 The consistency of the trend in R with S/Fe ratio suggests that there is a set of

515

controlling factors that is common to all conditions, most likely the formation of FeS and its

516

subsequent effects on corrosion of Fe(0). Nevertheless, the range of R among different

517

contaminants may indicate different sensitivities to various operational variables.

2017-10-13


19



518

519

520 521 522 523 524 525 526 527 528 529

Figure 4. Comparison of rate constants for contaminant reduction by reducing materials with and without sulfidation: (A) kSA vs kM plot for CT (331 k’s from 21 refs43, 53, 134, 135, 138, 143-158); (B) kSA vs kM plot for TCE (189 k’s from 24 refs14-17, 52, 54, 56, 60, 132, 133, 148, 152, 159-170); (C) kM with sulfidation vs kM without (95 k’s from 8 refs14, 16-18, 43, 82, 86, 143). The diagonal contours in (A) and (B) represent specific surface areas used to calculate kSA and the contours in (C) represent the relative effect of treatment (R). Elaboration on the format of these graphs can be found where they were first used: (A,B)155 and (C)24. All of the data plotted for (C) are given in Supporting Information (Table S1).

2017-10-13


20

Page 21 of 44


530

Reduction Pathway Selectivity. While most of the kinetic studies have focused on enhanced

531

contaminant degradation by nZVI due to sulfidation, they did not fully emphasize/appreciate the

532

fact that the enhanced contaminant degradation may be caused, at least in part, by improved

533

selectivity of nZVI between contaminant degradation and H2 production. The selectivity

534

quantifies the ratio of electrons used for contaminant degradation and electrons consumed by

535

natural reductant demand (NRD).126 The partial inhibition of hydrogen production of sulfidated

536

nZVI noted in two kinetics studies imply that more Fe(0) could become available for

537

contaminant degradation.17, 18, 25 A comprehensive evaluation on the effects of sulfidation on

538

selectivity showed complete inhibition of H2 production achieved for CMC-nZVI sulfidated for

539

24 h at high S/Fe ratios.13 In contrast, non-sulfidated CMC-nZVI was completely oxidized by

540

H2O in

Sulfidation of Iron-Based Materials: A Review of Processes and

Recommend Documents