Electrochemical Characterization of Magnetite with Agarose-Stabilized

Subscriber access provided by Washington University | Libraries

Article

Electrochemical Characterization of Magnetite with AgaroseStabilized Powder Disk Electrodes and Potentiometric Methods Miranda J. Bradley, and Paul G. Tratnyek ACS Earth Space Chem., Just Accepted Manuscript • DOI: 10.1021/ acsearthspacechem.8b00200 • Publication Date (Web): 26 Feb 2019 Downloaded from http://pubs.acs.org on February 28, 2019

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 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 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.

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 30 1 2 3 4 5 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 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Earth and Space Chemistry

1 2 3 4

Electrochemical Characterization of Magnetite with Agarose-Stabilized Powder Disk Electrodes and Potentiometric Methods

5 6

Miranda J. Bradley1, and Paul G. Tratnyek1*

7 8 9 10

1

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

11 12 13 14

*Corresponding author:

15

Email: [email protected], Phone: 503-346-3431, Fax: 503-346-3427

16 17 18

Main Text file for Publication in ACS Earth and Space Chemistry

19

Accompanying Supporting Information file includes 32 pages, 17 Figures, 4

20

Tables, 18 References

21 22

Revised Version (1/19/19 8:40 AM)

23

ACS Paragon Plus Environment

ACS Earth and Space Chemistry 1 2 3 4 5 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 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 30

Bradley | Tratnyek (2019)

24

Abstract

25

The mixed and variable valence of iron in magnetite (Fe(III)tet[Fe(II),Fe(III)]octO42− ) give this

26

mineral unique properties that make it an important participant in redox reactions in

27

environmental systems. However, the variability in its stoichiometry and other physical

28

properties complicates the determination of its effective redox potential. To address this

29

challenge, a robust method was developed to prepare working electrodes with mineral powders

30

of diverse characteristics and agarose-stabilized pore waters of controlled composition. This

31

second-generation powder-disk electrode (PDEv2) methodology was used to characterize the

32

electrochemical properties of magnetite samples from a wide variety of sources (lab-synthesized,

33

commercial, and magnetically separated from environmental samples) using a sequence of

34

complementary potentiometric methods: chronopotentiometry (CP), linear polarization resistance

35

(LPR), and then linear sweep voltammetry (LSV). The passive method CP gave open-circuit

36

potentials (EOC) and the active method LPR gave corrosion potentials (E0,LPR) that agree closely

37

with each other, but vary over a wide range for the magnetite samples tested (ca. 520 mV, from

38

−267 to +253 mV vs SHE). The active method LSV gave values of E0,LSV that become

39

increasingly more negative than EOC for the samples with more positive potentials (by up to 189

40

mV). This effect is consistent with the cathodic polarization applied at the beginning of the LSV

41

scan and suggests there is convergence of substoichiometric magnetites to the potential of

42

stoichiometric magnetite after polarization. By all methods, lab-synthesized magnetites gave

43

more negative potentials and smaller polarization resistances (Rp) than magnetite from

44

commercial sources or magnetic-separation of environmental samples. This is consistent with the

45

common notion that freshly synthesized minerals are more reactive, but clear correlations were

46

not found between the measured redox potentials and surface area, iron stoichiometry, or

47

magnetic susceptibility. All the measured potentials for magnetite fall in a range between

48

calculated thermodynamic values for redox couples involving relevant iron species, which is

49

consistent with the measured values being mixed potentials. The wide range in effective redox

50

potential of magnetite is likely to influence its role in biogeochemistry and contaminant fate.

51 52

Keywords: Magnetite, Spinel Iron Oxide, Potentiometry, Voltammetry, Porous Powder Disk

53

Electrode, Effective Redox Potential

54

1/19/19


2

Page 3 of 30 1 2 3 4 5 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 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Bradley | Tratnyek (2019)

55

Introduction

56

While Fe(III) oxides are the most ubiquitous iron minerals in critical zone environments,1, 2

57

mixed-valent iron oxides containing Fe(II) and Fe(III)—including magnetite and green rust—are

58

of special interest.3 Magnetite (FeIIFeIII2O4) is the most stable mixed-valent iron oxide, and it is

59

abundant in ferrigenous rocks, soils, and sediments.4 Magnetite also plays important roles in

60

corrosion of ferrous metals (making up much of the passive film on metallic iron surfaces),

61

magnetic storage/recording media,5-7 biology (e.g., in magnetotaxis),8 medicine as a contrast

62

agent in MRI imaging9 and in cationic liposomes for drug delivery,10, 11 nanotechnology (e.g., for

63

functionalization of carbon nanotubes),12, 13 and catalysis (e.g., synthesis of ammonia in the

64

Haber-Bosch process).14, 15 In all of these contexts, magnetite is a characteristically stable phase,

65

but it participates as a conductor and/or electron donor in a variety of environmentally important

66

redox processes such as corrosion of iron,16, 17 metabolism by iron bacteria,18-20 and abiotic

67

reduction of contaminants.21-27 Among common environmental minerals, magnetite has the smallest band gap,28 the

68 69

highest conductivity,29 and among the lowest reduction potentials.3, 4 The conductivity of

70

magnetite stems from the mixed-valence of iron in the mineral, which has an inverse spinel

71

structure with Fe3+ in the tetrahedral sites and Fe2+ and Fe3+ in the octahedral sites. However,

72

Mossbauer spectroscopy has shown that electrons from Fe2+ and Fe3+ in the octahedral sites

73

delocalize, giving each iron atom a net charge of +2.5.30, 31 Stoichiometric magnetite has a

74

Fe2+/Fe3+ ratio (x) equal to 0.5, but magnetite from the environment is commonly

75

substoichiometric (x < 0.5) because Fe2+ is oxidized by dissolved or atmospheric O2 and

76

biogeochemical processes.32-34 Other elements can replace the Fe2+ or Fe3+ in the magnetite

77

structure, including magnesium, zinc, and nickel.29 Among these dopants, TiIV is the most

78

common impurity in natural magnetite, where it preferentially replaces Fe3+, which results in

79

magnetite that is superstoichiometric (x > 0.5).23, 35 These differences in magnetite stoichiometry

80

are expected to influence its properties related to material electronic structure, including

81

magnetic susceptibility, conductivity, and redox potential.21 Determining the appropriate redox

82

potential to describe magnetite in biogeochemical systems is complicated by its irregular particle

83

size, shape, and composition (e.g., doping); surface modifications by adsorption and weathering;

84

and direct and/or mediated intraparticle interactions. Theoretical redox potentials for pure mineral phases can be calculated from

85 86

thermodynamics,36, 37 but the effective redox potential of natural materials is affected by other 1/19/19


3


Page 4 of 30

Bradley | Tratnyek (2019) 87

factors and therefore must be measured.38 The redox potential of mineral samples can be

88

measured by indirect methods, such as chemical reactivity probes (CRPs) with

89

spectrophotometric detection39 and electron-transfer mediators with potentiometric detection36, 40

90

or directly by using working electrodes prepared from the mineral samples.30, 34, 41, 42 Electrodes

91

made from polished bulk natural magnetite have been used to study open circuit potential,30, 34

92

polarization resistance, and dissolution rate,30, 41, 43 oxidation/reduction products,30 and the effect

93

of applied potential and electrolyte composition on surface stoichiometry.34 Electrodes made by

94

depositing (mostly) magnetite passive layers directly onto iron or steel rods have been used to

95

study the effects of redox activity and coating integrity on corrosion processes.44, 45 Composite

96

magnetite electrodes, made with mixtures of magnetite and a conductive binder like graphite

97

paste or carbon black, have been used to study the electrodissolution of magnetite,42, 43, 46 and the

98

capacitive behavior of magnetite in aqueous electrolytes.47 Additional methods have been used to study the electrochemisty of particulate materials

99 100

other than magnetite, ranging from thin films of fine particles spin-coated onto disk electrodes to

101

porous beds of coarse particles packed into columns.48-50 In order to characterize intact nano- to

102

micro-sized zerovalent iron (ZVI)—without grinding, polishing, sintering, etc.—we developed a

103

“powder disk electrode” (PDEv1) wherein ZVI samples could be dry-packed into a millimeter-

104

sized cavity in a cap covering a conventional rotating disk electrode.51, 52 Electrochemical

105

methods have been used with PDEv1 to characterize many aspects of the reactivity of granular

106

ZVI, including the effects of electrode design and operational factors,51 solution chemistry on the

107

passive film,53, nano-scale effects,52, 54 aging and stability of zerovalent iron nanoparticles

108

(nZVI),54, 55 sulfidation of nZVI,56, 57 effect of organic coatings,58 and deposition of nZVI as a

109

surface coating.59 A much more limited amount of work has been done using PDEv1 to

110

characterize iron minerals, mostly for control experiments to compare with ZVI,52, 59 and just one

111

study included PDEv1 measurements to characterize the redox properties of granular

112

magnetite.21 Despite the success of prior work using PDEv1, the method has several limitations,

113 114

including: (i) the material must pack well enough to stay in the cavity by compression alone,

115

which precludes studying coarse materials, and (ii) packing dry particles into the PDEv1 and

116

then immersing them into electrolyte results in unknown and possibly variable amounts of the

117

pore space becoming filled with fluid. While we have not found that these limitations

118

significantly complicate the interpretation of results, filling the pore space with a stabilizing fluid 1/19/19


4

Page 5 of 30 1 2 3 4 5 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 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60



of controlled composition should make the electrode performance more robust and open up

120

additional experimental opportunities, such as studying coarse particles, effects of pore solution

121

chemistry, in situ deposition of secondary phases, etc. For this study, a modified method

122

(PDEv2) was developed that involves packing the electrode with a wet slurry of the sample

123

particles in a fluid composed of aqueous electrolyte and a non-reactive polymer to cause

124

gelation. Agarose was chosen as the stabilizing polymer, based—in part—on a previously

125

published method for making and studying model soil aggregates.60 The PDEv2 method, which

126

is first described here, produces a stable electrode that is able to withstand rotation up to 2000

127

rpm, pore fluid chemistry controlled by the recipe used to prepare the agarose, and an easily

128

retrievable “puck” for further characterization after the completion of electrochemical

129

experiments (Figure S1E). The unique features of PDEv2 should eventually allow electrochemical characterization

130 131

of biogeochemical redox processes at mineral-water interfaces, including the adsorption of Fe2+,

132

precipitation of iron oxides (or sulfides), mediation of electron transfer by shuttle compounds,

133

long-range electron transfer via interparticle interactions, etc. However, the results reported in

134

this study are limited to initial characterization of PDEv2 response using a relatively stable iron

135

oxide (magnetite), simple solution chemistry (borate buffer), and a range of common

136

potentiometric and voltammetric electrochemical methods. The results are used to address two

137

process-level objectives: (i) clarification of the relationship between various definitions of

138

“effective redox potential” and their application to (suspended, porous, or packed) particulate

139

materials, and (ii) the relationship between redox potential of minerals and their impurities,

140

particle size, surface coatings, etc. With respect to (i), this work appears to be the first systematic

141

comparison of direct electrochemical methods for measuring redox potentials of mineral (and

142

other material) powders, and demonstrates the diagnostic value of comparing passive and active

143

potential measurement methods. With respect to (ii), the wide range of magnetite samples

144

characterized, bracketed with data on more reduced and oxidized iron-based materials (i.e., Fe0

145

and Fe2O3, respectively), provides a unique perspective on the sources of variability in these

146

measured values, which allows identification of significant differences (e.g., due to passivating

147

surface coating formed in the environment) and indeterminant variability due to sample

148

heterogeneity, etc. One of the most significant differences observed was between relatively fresh

149

magnetite synthesized in the laboratory and relatively-aged magnetite either purchased as

150

chemical reagents or magnetically separated from sediment. While these differences in electrode

1/19/19


5


Page 6 of 30


potential measurements should be indicative of the material’s reactivity, it was not possible to

152

resolve the measured mixed potentials into the contributions of specific interfacial redox

153

reactions.

154

Experimental Section

155

Electrochemical System. Experiments were performed in a three-electrode cell, using 100 mL

156

of 0.1875 M borate buffer at pH 8.4 (mainly for consistency with our prior work using PDEv1

157

and ZVI51, 61). The cell was sparged with N2 throughout the experiment, at a flow rate of

158

approximately 0.5 L/min. The working electrode was based on a ChangeDisk electrode from

159

Pine Research Instrumentation (Durham), with a customized steel and Teflon-lined disk cavity

160

(Figure S1A), which was filled with sample as described below and further in the supporting

161

information. The Ag/AgCl (4M KCl) reference electrode and Pt coil wire (approximate surface

162

area 4.7 cm2) counter electrode were also from Pine. A steel and aluminum foil Faraday cage

163

was used to minimize electronic noise. (Figure S1B) Experiments were run and data

164

collected/analyzed with an Autolab PGSTAT30 potentiostat from Metrohm, running Nova 1.10

165

software.

166

Custom working electrode. A variety of working electrode configurations and assembly

167

protocols were investigated before settling on the method used in the study. Some of the

168

preliminary results obtained during the development process are included as Supporting

169

Information (SI). The powder disk electrodes (PDEs) were made by filling the cavity with the

170

material of interest by suspending the sample in 1.5 to 2 wt% agarose gel (Fisher Agarose Low

171

Melting, BP1360) made with the same buffer used in the cell. Agarose with a low melting

172

temperature was used to avoid having to expose the samples to temperatures greater than 50ºC.

173

The suspension was pipetted into the well of the ChangeDisk RDE tip (to slightly overfill the 30

174

μL cavity), and the tip was centrifuged for 15 minutes at 3200 rpm to ensure the sample was

175

well-packed and had good connection to the steel shaft at the base of the well. After the primary

176

centrifugation step, the surface was cleaned of debris and resurfaced with a thin layer of agarose

177

gel to create a smooth, slightly convex interface with the electrolyte (Figure S1D).

178

Magnetite Samples. A summary of all the magnetite samples used in this study—with standard

179

materials properties and references to prior work with each material (when available)—is given

180

in Table S1 of the Supporting Information. Three commercially-sourced samples of magnetite

181

were chosen as reference materials because they are readily available and have been 1/19/19


6

Page 7 of 30 1 2 3 4 5 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 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60



characterized in prior studies: micron-sized magnetite powder from Bayferrox/Lanxess

183

(Cologne) and nano and micron magnetite powders from Sigma-Aldrich (St. Louis). Several

184

samples prepared with a (sub)stoichiometric ratio of Fe2+/Fe3+ were obtained from M. Scherer’s

185

laboratory at the University of Iowa. Magnetite samples provided by C. Pearce (Pacific

186

Northwest National Laboratory, Richland, WA) had been magnetically separated from sediment

187

at the Hanford nuclear plant in Washington State, and samples from M. Villalobos lab (National

188

Autonomous University of Mexico) had been isolated from iron ore deposits at the Pena

189

Colorada mine in Mexico. Additional magnetically separated magnetite samples were obtained

190

from a commercial supplier (Prospector’s Choice, Surprise, AZ). We also magnetically separated

191

magnetite from sediment samples collected from several locations in the Willamette river

192

(Oregon) that had been dried and size-fractioned. Maghemite from Alfa Aesar (Heysham),

193

reference iron samples, and electrodes made with agarose alone were included for comparison.

194

Electrochemical Experiments. Each PDE preparation was characterized by a complementary

195

combination of electrochemical techniques that proceed from less-destructive to more-

196

destructive.56 The PDE was placed in the electrochemical cell containing 0.1875 M borate buffer

197

(pH 8.4) as electrolyte and chronopotentiometry (CP) was performed by monitoring the open

198

circuit potential (EOC) while rotating the electrode at 2000 rpm. In early studies, CP was

199

performed for 60 minutes and the EOC at that time was recorded, but this period was extended to

200

14 hr (or more) in order to ensure that the electrodes had reached a consistent degree of

201

equilibration. After CP, linear polarization resistance (LPR) was performed by measuring current

202

(i) while applying a potential sweep—at a scan rate of 1 mV/s—from 10 mV below to 10 mV

203

above the last recorded EOC. LPR was repeated at least once, separated by 10-60 min of CP, to

204

verify sample stability. After CP and LPR measurements, linear sweep voltammetry (LSV) was

205

performed by applying potentials from 200 mV below to 200 mV above the last measured EOC,

206

at a scan rate of 1 mV/s. All potentials were measured versus Ag/AgCl reference electrodes, but

207

are reported versus the standard hydrogen electrode (SHE). All polarization resistance (RP) and

208

corrosion current (iCOR) values were normalized to the geometric surface area of the exposed

209

disk. More discussion on the use of surface area for normalization of data obtained with porous

210

electrodes can be found in SI.

211

Non-Electrochemical Characterization. To determine magnetite stoichiometry,62-65

212

approximately 15 mg of sample was dissolved in 1 mL of HCl; commercial and collaborator-

213

synthesized magnetite was dissolved in 5 M HCl, while magnetically separated samples required 1/19/19


7


Page 8 of 30


12 M HCl for complete dissolution. The samples in acid were stored in an anaerobic glove box

215

and were sampled by diluting 5 μL aliquots in 1 mL of 1 M HCl (Fe2+) or 1 mL of 1.4 M

216

hydroxylamine hydrochloride in DI water (Total Fe). The diluted samples were removed from

217

the glovebox and further diluted into 1.5 mL cuvettes containing DI water, then 0.5 mM

218

ferrozine in 0.5 M sodium acetate buffer was added. The ferrozine binds with the Fe2+ in solution

219

to produce a magenta color with a maximum absorbance at 562 nm. Using a calibration curve,

220

the total Fe and Fe2+ concentrations were determined and the Fe3+ content of the original sample

221

was calculated to determine the sample’s Fe2+:Fe3+ ratio. Sampling was done in triplicate with at

222

least one replicate preparation. To verify the commercial and magnetically separated samples were primarily magnetite,

223 224

X-ray diffraction (XRD) analysis was performed using a Rigaku MiniFlex600 X-ray

225

diffractometer with 40 kW/15 mA radiation (Willamette river samples used 40 kW/44 mA

226

radiation). Diffraction patterns were collected over a range of 10 to 80º of 2Ɵ, using a step size

227

of 0.04º and time per step of 0.5 s. The three samples that had not been previously characterized

228

in other literature, magnetically separated magnetite from Prospector’s Choice and two size

229

fractions of magnetite from Willamette river sediment, were also characterized by scanning

230

electron microscopy with energy dispersive X-ray detection (SEM-EDS) using a Zeiss Sigma VP

231

FE-SEM.

232

Results and Discussion

233

Chronopotentiometry (CP). After each freshly-prepared PDE was first immersed into

234

electrolyte, EOC was recorded for up to 17 hours. Representative examples of this

235

chronopotentiometric data (Figure 1A) show that EOC became relatively stable (drift < 5 mV/hr)

236

after about 3 hr of equilibration. EOC was monitored until drift was < 2 mV/hr. No evidence for

237

other changes—such as Flade potentials due to passive film breakdown53—were seen with any

238

of the materials studied. Averages of the final values of EOC across replicate measurements

239

(subsamples of the same material in independently prepared PDEs) shows the expected overall

240

trend of increasingly positive EOC with magnetites that are less pure and/or more oxidized

241

(Figure 1B). ZVI samples (presumably with a largely magnetite outer coating)52, 54, 66, 67 had the

242

lowest measured EOC, while the laboratory synthesized magnetites that had been stored in an

243

anaerobic chamber gave EOC values that ranged from −267 mV for (nearly)stoichiometric

244

magnetites to +159 mV for substoichiometric magnetites. Reagent grade magnetite that was

1/19/19


8

Page 9 of 30 1 2 3 4 5 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 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60



purchased commercially (and not stored anaerobically) and magnetite that was magnetically

246

separated from river sediments both gave measured values of EOC that were similar to maghemite

247

(Figures 1A,B).

248 249 250 251 252

Figure 1. Representative potential measurements using PDEs: (A) EOC by CP during the first 12 hr, (B) Average EOC (passive), E0,LPR (active), and E0,LSV (active) values with standard deviations, and (C) Polarization by LSV at the end of each experiment shown in (A). Current in (C) was normalized to the geometric surface area of the exposed disk.

The error bars on the average values of EOC (final) in Figure 1B reflect variability in all

253 254

the data for each material type. The full set of data is summarized in Figure S3 (and tabulated in

255

Table S2), where the x-axis represents the material sample used to prepare the PDE, with

256

categories of material distinguished by color and ordered by expected degree of oxidation. For

257

most materials, replicate measurements agree to within about 50 mV, which is similar to the

258

consistancy obtained in prior work using single crystal magnetite electrodes.30 In contrast, about

259

one third of the samples included one or more final EOC values that differ by up to 200 mV. This

260

combination of generally consistant with occassional outlier measurements suggests one or more

261

operational variables in the PDEv2 protocol are not well controlled. One possibility is

262

inconsistancy in the quantity of sample contained in the PDE cavity, but no relationship was seen

263

between EOC and the mass of sample material recovered from the PDE (Figure S4). A second

264

possibility is oxygen intrusion into the electrochemical cell, which is difficult to avoid

265

completely even with continuous sparging with N2.59 While the PDE was prepared in an

266

anaerobic chamber, electrochemical measurements were made on the benchtop and the sample

1/19/19


9


Page 10 of 30


material may have been exposed to varying amounts of oxygen during the transition. A third

268

possible source of variability could be heterogeneity of the samples, both particle to particle and

269

in the surface coating of the particles. To better understand the source and significance of the

270

variability in EOC measured passively by CP, these data are compared to potentials measured

271

using the active methods introduced below. Linear Polarization Resistance (LPR). After the passively measured EOC was judged to

272 273

be sufficiently stable (drift

Electrochemical Characterization of Magnetite with Agarose-Stabilized

Recommend Documents