Ferrous Iron Oxidation under Varying pO2 Levels: The Effect of Fe(III

Dec 1, 2017 - Phone: (01) 706-410-1293. ... sorbents and organic matter (OM) in soils and sediments; however, this tertiary system has rarely been stu...
0 downloads 0 Views 1MB Size
Subscriber access provided by READING UNIV

Article

Ferrous Iron Oxidation under Varying pO2 Levels: the Effect of Fe(III)/Al(III) Oxide Minerals and Organic Matter Chunmei Chen, and Aaron Thompson Environ. Sci. Technol., Just Accepted Manuscript • DOI: 10.1021/acs.est.7b05102 • Publication Date (Web): 01 Dec 2017 Downloaded from http://pubs.acs.org on December 1, 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 31

1

Environmental Science & Technology

To Environ. Sci. & Technol.

2 3 4 5 6 7 8 9

Ferrous Iron Oxidation under Varying pO2 Levels: the Effect of Fe(III)/Al(III) Oxide

10

Minerals and Organic Matter

11 12

Chunmei Chen and Aaron Thompson*

13 14 15

Depart of Crop and Soil Sciences

16

The University of Georgia, Athens, GA, USA 30602

17 18 19 20 21 22 23

Corresponding author *

Aaron Thompson, [email protected], (01) 706-410-1293

1 ACS Paragon Plus Environment

Environmental Science & Technology

24

Abstract

25

Abiotic Fe(II) oxidation by O2 commonly occurs in the presence of mineral sorbents and

26

organic matter (OM) in soils and sediments, however this tertiary system is rarely studied.

27

Therefore, we examined the impacts of mineral surfaces (goethite and γ-Al2O3) and organic

28

matter (Suwannee River Fulvic Acid, SRFA) on Fe(II) oxidation rates and the resulting Fe(III)

29

(oxyhydr)oxides under 21% and 1% pO2 at pH 6. We tracked Fe dynamics by adding 57Fe(II) to

30

56

Fe-labeled goethite and γ-Al2O3, and characterized the resulting solids using 57Fe Mössbauer

31

spectroscopy. We found Fe(II) oxidation was slower at low pO2 and resulted in higher

32

crystallinity Fe(III) phases. Relative to oxidation of Fe(II)(aq) alone, both goethite and γ-Al2O3

33

surfaces increased Fe(II) oxidation rates regardless of pO2 levels, with goethite the stronger

34

catalyst. Goethite surfaces promoted well-crystalline goethite formation, while γ-Al2O3 favored

35

goethite and some lepidocrocite, and oxidation of Fe(II)aq alone favored lepidocrocite. SRFA

36

reduced oxidation rates in all treatments except mineral-free systems at 21% pO2 and decreased

37

Fe(III) phase crystallinity, facilitating low-crystalline ferrihydrite in the absence of mineral

38

sorbents, low-crystalline lepidocrocite in the presence of γ-Al2O3, but either well-crystalline

39

goethite or ferrihydrite when goethite was present. This work highlights that oxidation rate, the

40

types of mineral surfaces, and OM control Fe(III precipitate composition.

41 42 43

Introduction Iron (Fe) plays a major role in biogeochemical processes in soils and sediments.1 The

44

transition between Fe(II) and Fe(III) is of particular interest in dynamic environments that

45

experience alternating reducing and oxidizing conditions. The oxidation of Fe(II) and resultant

46

precipitation of Fe(III) oxhydroxides (hereafter termed as Fe(III) oxides) with varying

2 ACS Paragon Plus Environment

Page 2 of 31

Page 3 of 31

Environmental Science & Technology

47

crystallinity occurs in redox dynamic environments, is a key process determining the fate of

48

nutrients and contaminants.1,2 This is due to the inherent physicochemical properties of freshly

49

formed oxides, which have the capacity to adsorb and/or incorporate a variety of contaminants

50

and nutrients 3, 4. Fe(III) polymerization and precipitation by oxidation of Fe(II) may also trap

51

large amounts of organic matter (OM) within the Fe(III) oxide structure, stabilizing the OM

52

against microbial mineralization.5-9 Such iron/organic carbon associations are often invoked to

53

explain the long-term stabilization of OM.10

54

Fe(II) oxidation can occur through abiotic (chemical) and biotic (microbial) processes10,

55

which are strongly influenced by pH and O2 concentration12-14. At circumneutral pH, abiotic

56

oxidation dominates at high O2 levels, while abiotic and biotic oxidation occurs at comparable

57

rates at low O2 levels.15,16 The abiotic oxidation of Fe(II) can proceed along two parallel

58

pathways, one via homogeneous oxidation in solution and secondly via heterogeneous oxidation

59

in association with mineral surfaces. In natural waters, OM and dissolved Fe(II) often co-exist,

60

which may alter abiotic homogeneous Fe(II) oxidation kinetics by complexing both Fe(II) and

61

Fe(III) species17-19. The impact of OM on O2-mediated abiotic Fe(II) oxidation is well studied

62

and OM has been variably found to increase, decrease, or have no effect on the oxidation rate of

63

Fe(II) in solution17-26. Previous studies of Fe(II) oxidation in the presence of OM have generally

64

used high organic ligand concentrations in order to prevent the precipitation of Fe(III) oxides

65

(focusing on how OM influences the rate of Fe(II) oxidation as opposed to the coupled oxidation

66

and precipitation).17-26 In fact, a very limited number of studies have evaluated the impact of OM

67

on the products of Fe(II) oxidation27,28. OM can interfere with the polymerization of Fe(III)

68

phases, promote the formation of amorphous or less crystalline Fe(III) oxides, and substantially

69

retard the Fe(II)-facilitated transformation into crystalline Fe(III) oxides27-33. Therefore,

3 ACS Paragon Plus Environment

Environmental Science & Technology

70

oxidation of Fe(II) in the presence of OM is likely to promote Fe(III) phases that are less-

71

crystalline (more-reactive) than would occur in the absence of OM. This knowledge gap

72

becomes even larger when we consider that neither Fe(II) oxidation rates nor the oxidation

73

products have been evaluated in the common co-occurrence of Fe(II), OM, and oxide mineral

74

surfaces in natural environments.

75

Mineral surfaces are well known to catalyze the oxidation of Fe(II) (e.g. heterogeneous

76

oxidation34-36) and most studies have used contaminants as the oxidant under O2-free (anoxic)

77

conditions. For instance, the oxidation kinetics of Fe(II) on Fe/Al oxides by inorganic

78

contaminant such as CrVI, AsV, HgII 37,38, or organic contaminants such as nitrobenzene and

79

pertechnetate35,36,39-41. In these heterogeneous, contaminant reduction experiments, the repeated

80

sorption and oxidation of Fe(II) on Fe/Al oxide surface results in the accumulation of new Fe(III)

81

phases that often resemble the underlying oxide mineral surface. Larese-Casanova et al.35 for

82

instance found de novo goethite forms via a templating effect on the initial parent goethite

83

mineral surface and others have found hematite41and goethite36 form on α-Al2O3 surfaces. In

84

natural environments, oxidant levels often fluctuate in response to rainfall, irrigation or as the

85

ground water table rises and falls: in these situations, the primary oxidant is either NO3 or

86

O211,42,43. Fe(II) oxidation by abiotic reaction with O2 or via Fe-oxidizing bacteria transferring

87

electrons to O2 or NO3- produces a variety of secondary Fe(III) minerals, ranging from poorly-

88

crystalline minerals such as ferrihydrite (Fh)42,44, to crystalline minerals including lepidocrocite

89

(Lp)45-49, goethite (Gt)14,45,47,50, magnetite51 and green rust51.

90

O2 concentrations are expected to control homogenous Fe(II) oxidation rates13, and

91

influence the mineral composition of the resulting Fe(III) oxides14,52,53. However, it remains

92

unclear how variations in pO2 levels will alter heterogeneous Fe(II) oxidation rates and the

4 ACS Paragon Plus Environment

Page 4 of 31

Page 5 of 31

Environmental Science & Technology

93

nature of the resulting Fe(III) oxides. There have been a few studies investigating O2-mediated

94

abiotic heterogeneous Fe(II) oxidation kinetics in the presence of Fe(III) oxides at fully

95

oxygenated conditions34,54,55, but these may not be applicable to lower O2-conditions. Although

96

biotic Fe(II) oxidation becomes much more significant at lower O2 levels16, even in those

97

systems the accumulation of biogenic Fe(III) oxides results in substantial abiotic oxidation via

98

surface-catalyzed heterogeneous reactions at the periphery of the biotic reaction centers56,57. The

99

nature of these newly precipitated Fe(III) oxides from heterogeneous oxidation is rarely

100

characterized34, and no data appears to be available for O2-mediated Fe(II) oxidation on Al

101

oxides, which likely involves different reaction kinetics than on Fe oxide surfaces as Al oxides

102

do not participate in electron transfer reactions with the adsorbed Fe(II)36.

103

Accordingly, our goal was to investigate Fe(II) oxidation kinetics and the resulting de

104

novo Fe(III) solids in the presence of Gt, γ-Al2O3, and OM (using SRFA as a model compound)

105

under varying pO2 levels, with homogenous oxidation of mineral-free Fe(II) solution as a control.

106

We hypothesized that mineral surfaces would enhance Fe(II) oxidation rates at all pO2 levels and

107

mediate more-crystalline Fe(III) oxide formation via templating effects, whereas the addition of

108

SRFA would decrease the oxidation rates and result in less-crystalline Fe(III) oxide precipitation.

109

We identify the de novo Fe(III) phases by taking advantage of the isotope specificity of 57Fe-

110

Mössbauer spectroscopy to track the mineralogical evolution of the isotopically-labeled 57Fe(II)

111

oxidation products on 56Fe-Gt and γ-Al2O3 phases.

112

Methods

113 114

Materials

115

All treatments were prepared with O2-free ultrapure (18.2 MΩ) water prepared by

116

purging the water with N2 and then storing it inside an anoxic glove box (95% N2, 5% H2) for at 5 ACS Paragon Plus Environment

Environmental Science & Technology

117

least 3 days prior to preparation of reagents. Acidic solutions of 57Fe(II) and 56Fe(II) were made

118

by dissolving powders of 57Fe(0) (96% pure, Chemgas) and 56Fe(0) (99.9% pure, Isoflex),

119

respectively, in anoxic 2 M HCl solutions and filtering the solutions prior to use58. Reference

120

SRFA was obtained from the International Humic Substance Society and dissolved in anoxic,

121

ultrapure water and stored in the glove box. 13C NMR estimates of carbon distribution of SRFA

122

were included in SI Table S1.

123

56

Fe-Gt was synthesized according to a modified method of Fh conversion59. The 56Fe(II)

124

solution was removed from the anoxic glovebox and oxidized to Fe(III) by addition of hydrogen

125

peroxide while stirring within a polypropylene co-polymer (PPCO) bottle. The solution was then

126

filtered and the precipitates diluted with ultrapure water and the pH adjusted to 12.0 by adding

127

5M KOH. The mineral suspension was then capped and aged at 70 °C for 20 d. The precipitate

128

was freeze-dried following rinsing with anoxic ultrapure water and centrifuging 3 times at 20

129

000 g for 10 min. The resulting precipitate was confirmed to be pure Gt by XRD. The specific

130

surface area and total Fe content of the synthesized Gt is 18 m2 g-1 and 685 mg g-1, respectively.

131

The synthesized 56Fe-Gt had negligible Mössbauer signal (i.e., negligible 57Fe isotope). The γ-

132

alumina (Al2O3, Degussa) was purchased from Sigma-Aldrich and exhibited a SSA of 135 m2 g-1

133

and contained no-detectable Fe based on acid-digestion and subsequent ICP analysis60. The γ-

134

Al2O3 and synthesized 56Fe-Gt were stored in the anoxic glove box prior to use to remove any

135

sorbed O2.

136

Fe(II) oxidation experiments

137

All Fe(II) oxidation experiments were done in triplicate and in the dark to avoid possible

138

photochemical oxidation of Fe(II). Prior to the oxidation event, all the reagents and reaction

139

bottles were prepared and stored in the glove-box. Experiments were conducted in 100 ml brown

6 ACS Paragon Plus Environment

Page 6 of 31

Page 7 of 31

Environmental Science & Technology

140

serum bottles containing 20 ml of 50 mM MES (2-(N-morpholino)-ethanesulfonic acid) buffer

141

adjusted to pH 6 using anoxic NaOH or HCl solutions. Aliquots of the 57Fe(II) solution were

142

added to achieve a 1.55 mM Fe(II) concentration. This concentration of 1.55 mM Fe(II) was

143

chosen in order to form sufficient Fe(III) oxidation products to be detectable by 57Fe Mössbauer

144

spectroscopy. This concentration is at the upper end of Fe(II) concentrations observed in many

145

redox fluctuating systems such wetlands and paddy soils and is therefore justified6, 61. The solid

146

phases were added at 30 mg 56Fe-Gt or γ- Al2O3 per sample for a solids loading of 1.5 g L-1.

147

Anoxic SRFA solution was added to achieve a C/Fe molar ratio of 1.6 in the SRFA-containing

148

reactors. This C/Fe molar ratio was chosen to facilitate Fe(III) precipitation as Fe(III)

149

(oxyhydr)oxides, and it is within the range found across various natural environments6, 10, 61.

150

Experiments were initiated by placing the reaction bottles on a rotary shaker in the glove box for

151

0.5 h to equilibrate the added 57Fe(II) across the aqueous and solid phases under anoxic

152

conditions prior to exposure to O2. The reactors were then sampled and Fe(II) oxidation was

153

initiated by placing the reactors on a rotary shaker in custom-built, sealed-atmospheric chambers

154

(fully contained within the anoxic glove-box) supplied with a continuous flow of either

155

laboratory air (~21% O2) or a gas mixture containing 1% O2 balanced with 99% N2. Dissolved

156

O2 (DO) measured by a Hach (USA) DO meter reached a maximum within 15 min exposure to

157

the O2 source and remained at ~8.1 and ~0.4 mg/L for the 21% and 1% pO2 treatments

158

respectively. Samples were taken from the reactors (under anoxic conditions) at various time

159

intervals until Fe(II) was fully oxidized.

160

Sampling and Analysis

161 162

Aqueous samples were extracted in the glove-box using sterile syringes with 0.22-µm filters and analyzed for Fe(II) concentrations using a modified ferrozine method62. Total Fe(II)

7 ACS Paragon Plus Environment

Environmental Science & Technology

163

concentration was measured by extracting an unfiltered sample with 0.5M HCl for 2 hours63 and

164

analyzing as described above. Sorbed Fe(II) was calculated based on the difference between

165

HCl-extractable Fe(II) and dissolved Fe(II). Solid-phase samples were collected at the end of

166

anoxic reaction, and at the end of Fe(II) oxidation event and analyzed for total C content and 57Fe

167

mineral composition via Mössbauer spectroscopy. Samples for solid phase C analysis were

168

freeze-dried following rinsing 3 times with anoxic ultrapure water and centrifuging at 20 000 g

169

for 10 min. C content was measured with a vario Micro cube CHNS analyzer (Elementar). Solid

170

samples for 57Fe Mössbauer analysis were collected after centrifuging at 20 000 g for 10 min,

171

preserved between two layers of O2-impermeable Kapton tape (this step took 46 T at 5 K for Gt and Fh, and < 46 T for Lp) and QS values (near zero

247

for Fh and Lp, while ~ -0.12 mm/s for Gt).

248

In the OM-free systems, addition of γ-Al2O3 promoted Gt formation (CS ≈ 0.48 mm/s,

249

QS = -0.08 to -0.13 mm/s, and H ≈ 47 T) in addition to Lp (CS ≈ 0.42 mm/s, QS ≈ 0.05 mm/s,

250

and H ≈ 43 T), with a higher proportion of Gt at high O2 levels (Table 2; Figure 3E & 3F). In

251

contrast, adding both γ-Al2O3 and SRFA facilitated the formation of Lp by inhibiting Gt

252

formation relative to OM-free γ-Al2O3 system (Table 2; Figure 3G & 3H).

253

In the absence of OM, Gt was the only product formed following 57Fe(II) oxidation in the

254

presence of the parent 56Fe-Gt at both O2 levels (Table 2; Figure 3I & 3J). The overall parameters

255

of the newly formed 57Fe-Gt (CS = 0.48 mm/s, QS ≈ -0.12 mm/s, and H ≈ 50 T) are similar to 11 ACS Paragon Plus Environment

Environmental Science & Technology

256 257

naturally abundant Gt standard (SI Table S2). When 56Fe-Gt and SRFA co-exists in the system, 57

Fe-Gt (CS = 0.48 mm/s, QS = -0.11 mm/s, and H = 49.7 T) was formed at low O2 levels

258

(Figure 3M and Table 2), whereas high O2 levels led to the formation of 42% Fh with a QS close

259

to 0 and of ~49 T and the balance Gt (Table 2; Figure 3L).

260

Fe(III) oxides order magnetically and thereby transform from a doublet into a sextet in

261

the MB spectrum at a characteristic measurement temperature, based on their relative

262

crystallinity. Fe(III) oxides of lower crystallinity require lower measurement temperatures to

263

magnetically order (and hence form a sextet) than Fe(III) oxides of higher crystallinity. In

264

general, for all systems, 1% O2 led to a greater proportion of Fe(III) oxides magnetically ordered

265

at higher temperature than 21% O2 oxidation, indicating that 1%-O2 oxidation led to more

266

crystalline oxide formation, relative to 21% O2 (Figure 4).

267

In the OM-free systems, γ-Al2O3 promoted the formation of a greater proportion of Fe(III)

268

oxides ordered at or above 77K (largely attributed to Gt) (Figure 4; SI Figure S5 & S6; Table S4

269

& S5) than in the mineral-free controls Figure 4; SI Figure S7 & S8; Table S6 & S7). Thus, γ-

270

Al2O3 mediates the formation of more crystalline Gt relative to homogenous oxidation. The 56Fe-

271

Gt enhanced this effect further as de novo 57Fe-Gt was almost completely ordered at 140K—

272

similar to the parent Gt standard—and displayed the highest crystallinity of all treatments

273

(Figure 4; SI Figure S9 & S10; Table S8 & S9).

274

In all treatments, SRFA led to the formation of Fe(III) oxides with lower-ordering

275

temperatures in the MB spectra, indicating lower crystallinity. In the system containing Fe(II)

276

solution only or containing both Fe(II) and γ-Al2O3, all the resulting Fe(III) oxides in the

277

presence of OM—identified as either Fh or Lp—ordered at or below 35K (Figure 4; SI Figure

278

S11, S12, S13, S14; Table S10, S11, S12, S13): such solids are of extremely low crystallinity68.

12 ACS Paragon Plus Environment

Page 12 of 31

Page 13 of 31

Environmental Science & Technology

279

In the presence of Gt, SRFA resulted in less-crystalline oxide formation than SRFA-free reactors

280

(Figure 4; SI Figure S15 & S16; Table S14 & S15).

281

Discussion

282 283 284

Slow Fe(II) oxidation at low pO2 results in more-crystalline Fe(III) oxide formation A number of studies have demonstrated the relative abundance of Gt and Lp produced

285

from the homogenous oxidation of aqueous Fe(II) by O2 at circumneutral pH (without significant

286

foreign ions) is manipulated by Fe(II) oxidation rates.52,53,69,70 Our study showed oxidizing Fe(II)

287

with low pO2 was slower than with high pO2 and resulted in more crystalline solids in both the

288

absence and presence of minerals. The longer exposure to sorbed Fe(II) in slow-oxidizing

289

incubations (at low O2 levels) may have enhanced the recrystallization of Fe(III), relative to

290

faster oxidations reactions. It is well known that, in the absence of foreign ions or OM, sorbed

291

Fe(II) drives Fe mineral transformation to more crystalline Fe phases like Gt30-32,71 via the

292

electron transfer between aqueous Fe(II) and Fe(III) solid phase 72-74. We found that even in the

293

presence of SRFA, the Fh formed during oxidation of Fe(II) in 1% O2 treatment had a higher

294

crystallinity than the Fh formed in the 21% O2 treatment. This may imply that Fe(II)-catalyzed

295

changes in Fh crystallinity can occur without altering the mineral identity. Alternatively, the

296

higher crystallinity resulting from the 1% O2 treatment may directly reflect faster rates of crystal

297

nuclei formation, which are predicted to generate less crystalline structures 75.

298

Mineral solid surfaces enhance Fe(II) oxidation rates, but drive the formation of more

299

crystalline Fe(III) oxides

300

Our results demonstrated that mineral surfaces strongly increased rates of O2-mediated

301

Fe(II) oxidation relative to mineral-free homogenous conditions, regardless of pO2 levels. Gt was

302

much more effective in catalyzing Fe(II) oxidation than γ-Al2O3, despite having a lower specific 13 ACS Paragon Plus Environment

Environmental Science & Technology

303

surface area. This is similar to O2-mediated Mn(II) oxidation in the presence of mineral surfaces,

304

where Lan et al.76 found Fe(III) oxides had a greater catalytic activity than Al oxides; and during

305

Cr(VI) reduction by Fe(II), where Buerge and Hug37 found Gt accelerated the reaction more than

306

Al2O3, although they attributed this to higher Fe(II) sorption on Gt. In our study, oxidation rate

307

constants are up to 7-fold greater for Gt than γ-Al2O3 (Table 1), whereas Gt sorbed only slightly

308

more Fe(II) (~4% of total Fe(II)) than γ-Al2O3 (time = 0.5 hr, SI Figure S2). Thus we propose

309

Gt’s semiconductor capabilities77, 78 facilitate electron transfer reactions that increase Fe(II)

310

oxidation rates relative to those on γ-Al2O3, which has electrical insulator properties 79. In

311

addition, adsorbed Fe(II) is valence stable at the γ-Al2O3 interface, whereas Fe(III) in Gt is an

312

electron acceptor for sorbed Fe(II). Therefore, the Fe(III) in Gt may be able to transport electrons

313

from Fe(II) to O2 when Fe(II)-Fe(III)-O2 complexes form at the surface, similar to electron

314

transport in As(III)-Fe(II, III)-O2 complexes 80.

315

Mineral surfaces increased Fe(II) oxidation rates relative to homogeneous rates, which

316

should favor the formation of lower crystallinity Fe(III) oxides 75. However, the presence of γ-

317

Al2O3 resulted in the formation of higher crystallinity Gt (e.g., higher MB ordering temperature)

318

at the expense of Lp, while the presence of 56Fe-Gt promoted an even higher crystallinity 57Fe-Gt,

319

which resembled the naturally abundant Gt standard. Templating effects likely play an important

320

role here, though unfortunately they have only been investigated for adsorbed Fe(II) on Fe(III)

321

oxides under anoxic conditions35. Our work suggests the formation of higher crystallinity Gt may

322

be caused by repeated sorption and oxidation of Fe(II) on parent Gt surfaces, which likely

323

provide a template for growth of a new phase.

324 325

The potential for a templating effect is perhaps best evaluated in the presence of γ-Al2O3, which promotes Gt formation, in contrast to Lp formation during homogenous oxidation.

14 ACS Paragon Plus Environment

Page 14 of 31

Page 15 of 31

Environmental Science & Technology

326

Peretyazhko et al.41 found α-Al2O3 provided a template for the formation of α-Fe2O3 (hematite)

327

during the oxidation of adsorbed Fe(II) by either Tc(VII) or O2. They attributed the formation of

328

α-Fe2O3 to the fact that α-Al2O3 and hematite (α-Fe2O3) are isostructural and have similar

329

interatomic distances for surficial oxygen atoms. If this is true, γ-Al2O3 should provide a

330

template for the formation of isostructural γ-Fe2O3 (maghemite), yet this phase is typically

331

formed via dehydration mechanisms or under hydrothermal conditions52. Instead, γ-Al2O3

332

evidently favors linked corner-sharing rows of Fe(III)-hydroxide octahedral, which are

333

precursors to Gt crystals. Subsequent sorption of Fe(II) on the precursor Gt nuclei may favor

334

further Gt growth rather than γ-Fe2O3. A similar observation was made by Larese-Casanova et

335

al.35, during oxidation of Fe(II) by nitrobenzene on α-Fe2O3 (hematite), which resulted in the

336

formation of α-FeOOH (Gt) as opposed to α-Fe2O3. This Gt-favoring effect was also observed

337

during heterogeneous Fe(II) oxidation on α-Al2O3 by nitrobenzene36. Taking together, it suggests

338

Gt may be a common product of heterogeneous oxidation.

339

OM decreases Fe(II) oxidation rates, but promotes less-crystalline Fe(III) oxide formation

340

Previous studies have observed that the addition of SRFA at high C/Fe ratios (> 20)

341

increased Fe(II)aq oxidation rates relative to organic-free Fe(II)aq oxidation rates.17, 21, 81 At these

342

high C/Fe ratios significant Fe(II) complexation via SRFA occurs and the dissolved Fe(II)-SRFA

343

complexes oxidize more rapidly than the free Fe(II)aq aquo-ion.17 Consistent with these previous

344

studies, we found SRFA accelerated Fe(II) oxidation when added at a high C/Fe ratio of 15 (SI

345

Figure S17), and this high C/Fe ratio also completely prevented the Fe(III) solid precipitation

346

that occurs in the SRFA-free control.

347

However, when we conducted the same experiments at lower C/Fe ratios that allowed

348

Fe(III) precipitation in both treatments, we found quite different results. At the low C/Fe ratio

349

(1.6) used for our primary experiments, we found that SRFA had a minor impact on Fe(II) 15 ACS Paragon Plus Environment

Environmental Science & Technology

350

oxidation rates at 21% O2, but substantially retarded Fe(II) oxidation at 1% O2 (Figure 1). At this

351

low C/Fe ratio (1.6), Fe(II)-SRFAaq complexes likely represented only 8% of total aqueous Fe(II)

352

prior to the oxidation (Visual Minteq, SI Table S16), thus Fe(III) solids were precipitated during

353

oxidation in both the SRFA-free as well as the SRFA-addition treatment (Figure 3a-d and Table

354

2), which should facilitate autocatalytic Fe(II) oxidation via surface catalysis13, 34 in both

355

treatments. Evidently, surface adsorbed or coprecipitated SRFA on Fe(III) (oxyhydr)oxides (SI

356

Table S17) decreased Fe(II) oxidation kinetics relative to SRFA-free Fe(III) (oxyhydr)oxides (SI

357

Figure S18). We propose that carboxyl group of SRFA (SI Table S1) can bind to Fe(III) solids

358

via ligand exchange 8, 82, prevent the inner-sphere complex formation between Fe(II) and

359

bridging oxo-groups on oxide surface83 and hence reduce Fe(II) oxidation, since the oxidation

360

rate of oxide surface-complexed Fe(II) species like FeOFe+ and ≡FeOFeOH0 are expected to be

361

several orders of magnitude faster than that of dissolved Fe(II)34, 83, 84. This is consistent with our

362

observation in the 1% O2 treatment that Fe(II) was oxidized much more rapidly in the absence of

363

SRFA than in the SRFA addition treatment. Attenuation of the Fe(II) oxidation rate by the

364

coprecipitated/adsorbed SFRA is still evident in the 21% O2 treatment, but the impact is very

365

small (SI Figure S18A), likely because the reaction rates are overall much higher.

366

When mineral surfaces were present initially, SRFA largely decreased Fe(II) oxidation

367

rates at both pO2 levels. SRFA could lower heterogeneous Fe(II) oxidation rates in three ways: (1)

368

SRFA could block part of the reactive mineral surface, which is supported by our findings that

369

OM readily adsorbed to the Fe/Al oxides during the 0.5 h anoxic period (SI Table S17); (2),

370

SRFA could coprecipitate with or adsorb on the newly formed Fe(III) solids (SI table S17) and in

371

turn reduce the auto-catalytic Fe(II) oxidation; and (3) a portion of the sorbed Fe(II) could be

372

preferentially complexed by carboxyl moieties in SRFA31,32, reducing the extent of Fe(II)

16 ACS Paragon Plus Environment

Page 16 of 31

Page 17 of 31

Environmental Science & Technology

373

adsorption on the mineral surfaces and for Gt, the amount of Fe(II)-Fe(III) electron transfer and

374

therefore decreasing Fe(II) oxidation via this route. This scenario is supported by the appearance

375

of a Fe(II) doublet in Mössbauer spectra of Fe(II)-reacted Gt-SFRA sample prior to oxidation

376

(Figure 2b), which have QS values consistent with Fe(II) sorbed to OM67.

377

Differentiating the types of OM-Fe co-precipitates formed across a matrix of low and

378

high pO2, and homogeneous vs. heterogeneous oxidation fills a key knowledge gap of carbon

379

behavior in environmental systems. Most studies synthesize Fe-OM co-precipitates via

380

hydrolysis of Fe(III)-OM solutions5,7,8. Yet, in the environment there is ample evidence

381

suggesting the main route of Fe-OM formation involves oxidation of Fe(II) and often in the

382

presence of reactive Fe and Al surfaces6,9,85,86. Thus far, our conceptual understanding of Fe-OM

383

phases is largely based on synthetic techniques that poorly approximate the likely natural

384

synthesis routes. We find that despite reducing the oxidation rates—which should favor more

385

crystalline Fe phases 75—OM resulted in the formation of less-crystalline Fh and Lp phases in all

386

treatments, except for Fe(II) oxidation at 1% pO2 with 56Fe-Gt.

387

In the mineral-free case, SRFA (initial C/Fe ratio= 1.6) leads to the co-precipitation of

388

OM and Fh with a C/Fe ratio of ~1.2 in the coprecipitates (SI Table S17). This OM-Fh co-

389

precipitate had a particularly low crystallinity and did not transform to a more stable Fe phase

390

during the experiment. This agrees with earlier studies that OM can completely inhibit Fh

391

conversion to more crystalline phases when the C/Fe ratio is >131, 32. Previous studies also

392

observed the formation of Fh following oxidation of ferrous perchlorate in the presence of citrate,

393

a low molecular weight organic acid at high citrate/Fe ratios.27, 87 These low-crystallinity OM-Fh

394

co-precipitates can potentially stabilize OM and provide a positive feedback that increases

395

carbon storage in natural environments.

17 ACS Paragon Plus Environment

Environmental Science & Technology

396

In the presence of γ-Al2O3, a large proportion of the SRFA was adsorbed onto γ-Al2O3

397

surface prior to oxidation (SI Table S17). This may have eliminated some of the OM from the

398

co-precipitation process as the resulting newly precipitated Fe(III) (oxyhydr)oxides had a low

399

OM content (C/Fe ratio ~0.3, SI Table S17). Prior work has shown that low C/Fe ratios favor the

400

formation of Lp over Fh and Gt32. Using the MB ordering temperature as a measurement of the

401

Fe phase crystallinity67, we find the OM-Lp formed in presence of γ-Al2O3 at 21% O2 had a

402

similarly low crystallinity to OM-Fh co-precipitates formed in homogeneous oxidation (Figure

403

4). Traditionally Lp is considered a more thermodynamic stable phase (i.e., higher crystallinity)

404

than Fh, but it may be that OM can impart similar low crystallinity across a variety of phases,

405

which might be similarly reactive—toward sorption, electron transfer reactions—as Fh.

406

In the presence of Gt, pO2 levels had a strong effect on the resulting OM-Fe precipitates,

407

with 21% O2 resulting in the formation of Fh and some Gt, while at 1% O2, only Gt was formed

408

(C/Fe ~0.4) (Table 2 and SI Table S17). As the low-surface area Gt sorbs little SRFA prior to

409

oxidation (SI Table S17), a large amount of the OM in the co-precipitates formed at 21% O2 is

410

likely being associated with the 42% of the Fe(III) that precipitated as Fh (C/Fe ratio ~0.6, SI

411

Table S17; Table 2). High C/Fe ratios in the OM-Fh co-precipitate coupled with a short exposure

412

to Fe(II) (