Enhanced Uranium Immobilization by Phosphate Amendment under

Apr 12, 2018 - A dual-domain reactive transport model is developed here with constraints from batch and column experimental data to understand the mec...
0 downloads 6 Views 1MB Size
Subscriber access provided by Universitaetsbibliothek | Johann Christian Senckenberg

Remediation and Control Technologies

Enhanced Uranium Immobilization by Phosphate Amendment Under Variable Geochemical and Flow Conditions: Insights from Reactive Transport Modeling Hang Wen, Zezhen Pan, Daniel E. Giammar, and Li Li Environ. Sci. Technol., Just Accepted Manuscript • DOI: 10.1021/acs.est.7b05662 • Publication Date (Web): 12 Apr 2018 Downloaded from http://pubs.acs.org on April 12, 2018

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 22

Environmental Science & Technology

1

Enhanced Uranium Immobilization by Phosphate Amendment Under Variable

2

Geochemical and Flow Conditions: Insights from Reactive Transport Modeling

3 4 5 6 7 8 9 10 11 12 13 14 15

Hang Wen1, Zezhen Pan2, Daniel Giammar2, Li Li1* 1.

Dept. of Civil and Environmental Engineering, The Pennsylvania State University, University Park, PA

16802, United States 2.

Dept. of Energy, Environmental and Chemical Engineering, Washington University in St. Louis, St.

Louis, MO 63130, United States

*Corresponding Author Address: Dept. of Civil and Environmental Engineering, The Pennsylvania State University, University Park, PA 16802, United States Phone: (814) 867-0151 E-mail: [email protected]

16 17

Abstract. Phosphate amendment has shown promise for enhancing uranium immobilization. The

18

mechanism of the enhancement however has remained unclear with contrasting observations

19

under different geochemical conditions. A dual-domain reactive transport model (RTM) is

20

developed here based on batch and column experimental data to understand the mechanisms and

21

to explore the effectiveness of enhanced U(VI) immobilization under variable geochemical and

22

flow conditions. Modeling results indicate that under low U(VI) conditions in natural waters,

23

phosphate addition promotes immobilization through the formation of U(VI)-phosphate ternary

24

surface complexes and the precipitation of calcium phosphate, both decreasing the concentrations

25

of mobile U(VI)-Ca-CO3 aqueous complexes. This contrasts with previously proposed

26

hypotheses attributing the enhancement to U(VI)-phosphate precipitation under high U(VI)

27

experimental conditions. Sensitivity analysis shows that phosphate is effective under relatively

28

low Ca (< 0.1 mM) and total inorganic carbon (TIC) (< 0.5 mM) conditions, where > 60% of

29

U(VI) still remains on sediments after 113 residence times of flushing with low phosphate

30

solutions (< 0.1 mM). Under high Ca or TIC conditions, a similar level of U(VI) immobilization

31

can be achieved only when the phosphate concentration is higher than Ca or TIC concentrations.

32

Compared to the strong geochemical effects, flow conditions have limited impacts on U(VI)

33

immobilization. These results explain contrasting observations on the effectiveness of phosphate

34

amendment and offer capabilities to extrapolate observations to other environmental conditions.

1 ACS Paragon Plus Environment

Environmental Science & Technology

35

Page 2 of 22

TOC art

36 37

2 ACS Paragon Plus Environment

Page 3 of 22

38

Environmental Science & Technology

INTRODUCTION

39

Uranium commonly exists in the environment either as sparingly soluble U(IV) or as

40

soluble U(VI) species.1 The U(VI) (UO22+) can form aqueous complexes with hydroxide and

41

carbonate,2 absorb on minerals (e.g., goethite and silica),3-6 precipitates,7 and transform to U(IV)

42

through microbe-mediated redox reactions.8,9 A fundamental understanding of uranium reactive

43

transport in natural systems is important for predicting natural attenuation and designing

44

remediation strategies for uranium-contaminated sites.

45

Phosphate addition has been observed to lower U(VI) mobility by forming U(VI)

46

phosphate precipitates and therefore has been considered as a remediation strategy for uranium

47

immobilization.10,11 Recent studies observed contrasting effects of phosphate arising from

48

different processes instead of precipitation alone.4,12 Some studies have demonstrated that

49

phosphate enhances uranium adsorption onto mineral surfaces (e.g., ferrihydrite, goethite and

50

natural sediments) under low U(VI) concentrations (< 5 µM) and pH around 7-8,13 owing to the

51

formation of ternary surface complexes with phosphate.14,15 Recent work conducted in well-

52

mixed batch experiments indicated that U(VI) immobilization can occur via precipitation of

53

U(VI) phosphate, adsorption to freshly precipitated calcium phosphate solids, or incorporation

54

into calcium phosphate solids depending on geochemical conditions (pH and relative

55

concentrations of U(VI), Ca, and phosphate).16 Other experimental studies, however, documented

56

that the effects of phosphate addition are limited for sediments from Rifle, Colorado.17 These

57

contrasting observations have not been resolved. Overall, there has been a lack of mechanistic

58

understanding of which phosphate-related processes control uranium immobilization and the

59

conditions under which phosphate addition is effective.

60

Most existing studies have been carried out in well-mixed batch reactors with U(VI)

61

concentrations (10~100 µM) that are much higher than the level of ~ 1 µM or less typically found

62

in contaminated groundwater. In addition, in natural systems where flow, transport and multiple

63

reactions occur simultaneously and are often coupled, identifying the dominant contributing

64

process presents a challenge. Reactive transport modeling (RTM) is known for its capabilities of

65

integrating multiple processes (e.g., sorption process in terms of surface complexation modeling

66

(SCM)) while at the same time differentiating the role of individual processes and providing

67

mechanistic understanding of complex process coupling.18 RTM has been developed to explore

68

the effects of geochemical condition, flow rates and mineral properties on U(VI) transport and

69

bioremediation.8,19-25 Such approaches however have not been used to examine the competing

70

effects of aqueous complexation, adsorption, mineral precipitation and transport on U(VI)

3 ACS Paragon Plus Environment

Environmental Science & Technology

Page 4 of 22

71

immobilization in the presence of phosphate, although SCM has been used to explore the reaction

72

mechanisms.4,26,27

73

This work aims to 1) develop a reactive transport model to understand and quantify

74

enhanced uranium immobilization by phosphate based on batch reactors and column experiments

75

and 2) quantitatively explore the extent of uranium immobilization under an array of geochemical

76

and flow conditions in natural environments that are not done in the laboratory. The results

77

provide important insights into reaction mechanism and the conditions under which phosphate

78

addition can significantly promote U(VI) immobilization.

79 80 81

METHODS 1. Batch and Column Experiments

82

The sediments used in the batch and column experiments are from the Hanford 300 Area

83

in Washington and are in the size fraction with grains smaller than 2 mm as obtained through

84

sieving. Previous work revealed that the sediments are mostly quartz and plagioclase feldspar

85

with smaller amounts of pyroxene and clays.28,29 Adsorbed and precipitated U(VI) species were

86

observed to be dominant in the subsurface,11,30 with a wide concentration range of 10-6 ~ 10-3

87

mmol/g sediment.20,31 Batch experiments were carried out using 250 g/L sediment (pre-

88

equilibrated with Synthetic Hanford groundwater (SHGW) containing no U(VI) and no

89

phosphate) in freshly prepared SHGW with U(VI) varying from 1.0×10-4 to 3.0×10-2 mM and

90

phosphate from 0 to 1.0 mM to examine U(VI) adsorption under a wide range of conditions.

91

SHGW mimicked the groundwater composition at the Hanford site with a pH of 8.05, Ca (1.0

92

mM), Na (2.0 mM), Mg (0.5 mM), Cl (1.0 mM), SO42- (1.5 mM), TIC (1.0 mM), U(VI) (0 mM)

93

and PO43- (0 mM).13

94

Four glass columns (2.5 cm diameter, 15.0 cm length) packed with Hanford sediments

95

were flushed with SHGW at 4.4×10-6 m/s (ColS-U and ColS-U-P) and 8.9×10-6 m/s (ColF-U and

96

ColF-U-P) using a peristaltic pump. The porosity of the four columns was 0.29-0.30. The

97

columns were run through a series of phases (Figure 1). A conditioning phase with injected

98

SHGW containing no U(VI) was performed to remove labile U(VI) from sediments. In the U(VI)

99

uptake phase that followed, SHGW with 3.5 µM U(VI) but no phosphate was used with bromide

100

as a conservative tracer. After the U(VI) uptake phase, columns ColS-U and ColF-U were

101

stopped while ColS-U-P and ColF-U-P continued to run for an additional U(VI) release phase by

102

introducing SHGW with 1.0 mM phosphate but no U(VI) to explore the effects of phosphate on

103

U(VI) immobilization. Effluent aqueous samples were collected by a fraction collector

4 ACS Paragon Plus Environment

Page 5 of 22

Environmental Science & Technology

104

(Spectrum/Chrom CF-1) to 15 mL tubes. All aqueous samples were acidified to 1.0% HNO3 for

105

measurement using inductively coupled plasma-mass spectrometry (ICP-MS, Perkin Elmer).

106

After the conclusion of each experiment, sediments were divided into inlet, middle and

107

outlet sections (roughly 5.0 cm each). Sediments in each section were homogenized and two

108

sediment samples of 2.0 g from each section were used for the five-step sequential extraction

109

following published procedures in literature.13,32 Detailed extraction information is in Table S1.

110

The background U(VI) concentration is estimated to be around 6.3×10-6 mmol/g using sediments

111

after the conditioning phase.13 The difference in extracted U(VI) concentrations between column

112

sediments and background sediments is the adsorbed U(VI) concentration. Conditioning phase (SHGW with no U(VI), no PO43-) u = 4.4×10 -6 m/s 1 RT = 2.75 hrs

ColS-U

1

ColS-U-P

2

U(VI) uptake phase (SHGW with U(VI), no PO43-)

42 RT

68 RT

42 RT

68 RT

U(VI) release phase (SHGW with no U(VI), with PO43-)

58 RT PO43-, no U(VI)

u = 8.9×10 -6 m/s 1 RT = 1.38 hrs

113 114 115 116 117 118 119 120 121 122 123 124

ColF-U

3

ColF-U-P

4

48 RT

74 RT

48 RT

74 RT

113 RT PO43-, no U(VI)

Figure 1. Operation phases for four columns. Each phase has one influent composition for the specified numbers of residence times (RT). In the conditioning phase, SHGW with no U(VI) was introduced to flush out labile U(VI) from the sediments. In the U(VI) uptake phase (sorption stage), SHGW with U(VI) but no phosphate was injected with the addition of bromide as a conservative tracer. After the U(VI) uptake phase, ColS-U and ColF-U were stopped. ColS-U-P and ColF-U-P were run for an additional U(VI) release phase by introducing SHGW with 1.0 mM phosphate but no U(VI) to further explore the effects of phosphate on U(VI) immobilization. Black dots represent the end of the operation for each column. Reactive transport modeling starts from U(VI) uptake phase. 2. Reactive Transport Modeling

125

Governing equations. A dual-domain reactive transport model is used to simulate column

126

experiments using the code PHREEQC.33 The column was conceptualized as having two

127

domains: an advection-dominated mobile domain and a diffusion-dominated immobile domain.

128

Solute exchange between the two domains is described by first-order kinetics driven by the

129

concentration gradients between the two domains. Both mobile and immobile domains have

130

uranium sorption sites. The dual-domain representation in PHREEQC has been used previously

131

for reactive transport.34,35 The code solves the mass conservation equations (1 and 2) integrating

5 ACS Paragon Plus Environment

Environmental Science & Technology

Page 6 of 22

132

flow, transport, and geochemical reactions.33,36 A representative equation for species i is as

133

follows:

θm

134

(

∂ Ci,m + M i,m



∂t

θ im

135

) =  θ

(

∂ Ci,im + M i,im

∂t

D m L

∂ 2 Ci,m

− θ mu m

∂x 2

) = α (C

i,m

∂Ci,m   − α Ci,m − Ci,im + θ m Ri,m ∂x 

(

)

− Ci,im + θ im Ri,im

)

(1)

i = 1,..., N

(2)

136

Where the subscripts m and im represent mobile and immobile domains; θ is porosity; Ci and

137

Mi is the aqueous and adsorbed concentration (mol/m3), respectively, with the latter calculated by

138

normalizing sorbed mass by the mobile and immobile pore volumes; DL is the hydrodynamic

139

dispersion (m2/s); u is the flow velocity (m/s); Ri is the overall rate of kinetically-controlled

140

reactions that species i is involved in (mol/m3/s); N is the total number of aqueous species. The

141

first-order mass transfer coefficient ( α , s-1) is assumed to be related to specific geometries of the

142

immobile zones as a reflection of spatial heterogeneity and follows the relation

Deθim (rf s→1 )2

α=

143

(3)

144

where De is the diffusion coefficient in porous media (m2/s), r is the radius of the assumed sphere

145

shape for the immobile zones (m), and f s →1 is a shape factor for sphere-to-first-order-model

146

conversion. The parameter α quantifies the solute transfer rate between mobile and immobile

147

zones. Its value is estimated using the breakthrough curves of a conservative tracer, as will be

148

discussed later.

149

Reactions. The reactions include uranium sorption through surface complexation reactions,

150

kinetically controlled calcium phosphate precipitation and thermodynamically controlled aqueous

151

complexation. Table 1 lists the major U(VI) aqueous and sorption reactions together with their

152

thermodynamics and kinetic parameters. For other aqueous reactions, we used the database

153

EQ3/6 and PSI/Nagra Chemical Thermodynamic Database version 12/07.37 Uranium surface

154

complexes

155

≡ SOUO2 CO3 HCO3

156

is newly proposed in this work. Calcium-phosphate precipitation R (mol/m3/s) follows the

157

transition state theory (TST) rate law38 with the following form:

158

(

include

(

≡ SOUO2 CO3 HCO3

)2− is based on literature

)2−

28

and

≡ SOUO2 PO42−

.

Note

that

in phosphate-free systems while ≡ SOUO2 PO4

 IAP  R = kA  1 −   K eq  

2−

(4)

6 ACS Paragon Plus Environment

Page 7 of 22

Environmental Science & Technology

159

Where k is the kinetic rate constant (mol/m2/s) in Table 1, A is the mineral surface area per unit

160

volume (m2/m3), IAP is the ion activity product for the reaction, and Keq is the equilibrium

161

constant. The term IAP/Keq quantifies the extent of disequilibrium. Three potential calcium-

162

phosphate precipitates were considered: amorphous calcium phosphate (ACP, Ca:P = 1.50),

163

octacalcium phosphate (OCP, Ca:P = 1.35), and hydroxylapatite (HAP, Ca:P = 1.67).

164

Note that the equilibrium constants of surface complex reactions (1)-(5) and kinetic rate

165

constant of calcium-phosphate precipitates ACP are the only parameters calibrated in this work.

166

All other geochemical parameters are taken directly from previous literature so we don’t have a

167

large number of parameters. The relatively abundant experimental data from well-mixed batch

168

reactors and columns help constrain the model parameters. For example, equilibrium constants of

169

surface complex reactions calibrated independently from batch reactors were directly applied to

170

reproduce column data without further calibration, as will be discussed later. Although not shown

171

here, we have done sensitivity analysis on a range of parameters. The numbers listed in Table 1

172

were identified to provide the best fit.

173 174

Table 1. Reactions and reaction parameters Reaction

Log Ka

Log k (mol/m2/s)

9.94

-

16.61

-

21.84

-

54.00

-

-0.86

-

0.65

-

36.41

-

25.02

-

30.70

-

27.18

-

13.23

-

19.59

-

22.82

-

U(VI) aqueous complexation reaction

7 ACS Paragon Plus Environment

Environmental Science & Technology

Page 8 of 22

22.46

-

44.04

-

45.05

-

-27.60b

-

-1.60b

-

20.70c

-

10.10b

-

15.50d

-

Surface complexation reaction e (1) (2) (3) (4) (5) Mineral dissolution and precipitation f

175 176 177 178 179 180 181 182 183 184 185 186 187

(6) ACP:

28.90g

-8.30h

(7) OCP:

13.10g

-7.58i

(8) HAP:

44.33g

-8.60i

a. From PSI/Nagra Chemical Thermodynamic Database 12/07 unless otherwise noted. b. Calculated from the experimental data from well-mixed batch reactors in this work. c. From Stoliker, et al.28 d. Calibrated in this work, in the range of 1015.0-1019.0 for phosphate sorption on solids.39,40 e. The surface area used is 13.0 m2/g, within the reported range of 11.5 to 14.1 m2/g for Hanford sediment.28 The site density is 3.84 µmol/m2, a typical value for Hanford sediments.28

(

) ( s) : amorphous calcium phosphate (ACP); Ca ( PO ) (OCP); Ca ( PO ) OH ( s ) : hydroxylapatite (HAP). f. Ca3 PO4 5

4 0.74

2

()

H 0.22 s : octacalcium phosphate

4 3

g. Calculated from Christoffersen, et al.41 h. Calibrated in this work, in the range of 10-9.0 – 10-6.0 mol/m2/s for typical calcium-phosphate solids.42 i. From Tang, et al.42

The U(VI) immobilization ratio β quantifies U(VI) immobilization as the ratio of the

188

immobilized U(VI) mass after the release phase (MU(VI),re) compared to the total U(VI) sorbed

189

mass during the uptake phase (MU(VI),up):

190

β=

MU (VI ),re MU (VI ),up

= 1−

(C Q∫ (C

Q∫

trelease

0 tuptake 0

out

) ) dt

− Cin dt

− Cout in

(5)

191

where Q is the flow rate (m3/s); Cin and Cout are the influent and effluent U(VI) concentrations

192

(mM); tuptake and trelease are the total U(VI) uptake and release time (s), respectively. Different

193

geochemical and flow conditions may lead to different extent of U(VI) immobilization in the

194

presence of phosphate (i.e., the immobilization ratio β). Values close to one mean most

195

immobilized U(VI) still remains on sediments after the release phase, and therefore phosphate

196

amendment is effective for U(VI) immobilization. Values close to zero indicate that most U(VI)

8 ACS Paragon Plus Environment

Page 9 of 22

Environmental Science & Technology

197

desorbs from sediments and that phosphate amendment is not effective. Note that MU (VI ),up is

198

always above zero in this work due to the U(VI) sorption onto sediments during the uptake phase.

199 200

RESULTS AND DISCUSSION

201

Reactions in batch reactors. The model with the proposed U(VI)-phosphate surface

202

complex (SC, Reaction 4 in Table 1) and Ca3 ( PO4 )2 ( s ) (ACP, Reaction 6) reproduces the data

203

well (Figures 2A-B). Including only surface complex (SC only) or ACP precipitation (ACP only)

204

underestimates sorbed U(VI) (Figure 2A), indicating that both surface complexation and

205

precipitation are important in enhancing U(VI) immobilization. At lower phosphate conditions

206

(e.g., 10-2 mM), the form of calcium phosphate precipitate does not make a big difference (< 10

207

%). At 1.0 mM phosphate, the model underestimates adsorbed U(VI) while overestimates

208

phosphate precipitates (compared to experimental data) (Figure 2B) if calcium phosphate

209

precipitate is assumed to be in the form of octacalcium phosphate Ca ( PO4 )0.74 H 0.22 ( s ) (OCP) or

210

hydroxylapatite Ca5 ( PO4 )3 OH ( s ) (HAP) (Reactions 7-8 in Table 1). No uranium-phosphate

211

solids (e.g., autunite and schoepite) form under these conditions as indicated by saturation indices

212

that are less than 0.

213

Under low initial U(VI) loadings (e.g., 2.6×10-3 mM), U(VI)-Ca-CO3 aqueous complexes

214

(CaUO2(CO3)32- and Ca2UO2(CO3)3(aq) in Figure 2C1) are the dominant U(VI) aqueous species,

215

which is consistent with other studies under similar geochemical conditions.9,43 The model

216

indicates that more U(VI) can sorb on the sediments if assuming no formation of U(VI)-calcium-

217

carbonates (Figure S1). This indicates that ACP precipitation enhances U(VI) adsorption by

218

decreasing Ca concentration and lowering the formation of U(VI)-Ca-CO3 aqueous complexes

219

2− (Figure 2C1). The surface complexes formed include both ≡ SOUO2 (CO3HCO3 ) (S-U-C) and

220

≡ SOUO2PO42− (S-U-P) (Figure 2C2). The S-U-C dominates until the initial phosphate reaches

221

0.62 mM, after which S-U-P dominates. That is, under low phosphate conditions, U(VI)

222

immobilization is enhanced primarily by forming Ca2(PO4)3(s) that lowers Ca and U(VI)-Ca-CO3

223

aqueous complexes. At high phosphate concentrations (> 0.62 mM), U(VI) immobilization is

224

enhanced directly by the formation of S-U-P. Under high initial U(VI) loading (~ 2.5×10-2 mM),

225

aqueous and surface speciation follows a similar trend as those under low initial U(VI) loading

226

(Figure S2).

9 ACS Paragon Plus Environment

Environmental Science & Technology

A1

A2 Batch data ACP + SC SC only ACP only 3-

1.0 mM PO 4

-5

0 mM PO34

10

-2

10 -1

10-7 -6 10

10

mM PO43-

mM PO43-

10-7 -6 10

Final U(VI)(aq) (mM)

10-1 ACP + SC -2

10

ACP only

1:1

Final phospahte (mM)

ACP + SC

10-3 OCP +SC

1 1: HAP + SC

10-4 10-2 Final U(VI)(aq) (mM)

Total U(VI)(aq)

10-4 U-Ca-CO3(aq)

0.0

10-1

10-5 -5 10

C1. Initial U(VI)(aq) = 2.6×10-3 mM

10-3

10-2 10-1 100 Initial phosphate (mM)

B2

10-5

10-5

SC only

B1 ACP + SC OCP + SC HAP + SC

10-2

100

10-3 -3 10

10-4 10-2 Final U(VI)(aq) (mM)

0.2 0.4 0.6 0.8 Initial phosphate (mM)

12.0 Sorbed U(VI) (×10-6 mmol/g)

Sorbed U(VI) (mmol/g)

10-3

227 228 229 230 231 232 233 234 235 236

Final phospahte (mM)

Sorbed U(VI) (mmol/g)

10-3

Page 10 of 22

1.0

10-3 10-1 Initial phosphate (mM)

C2. Initial U(VI)(aq) = 2.6×10-3 mM Total (S-U-P + S-U-C)

8.0

SOUO2(CO3HCO3)2- (S-U-C)

4.0 SOUO2PO42- (S-U-P)

0.0

0.0

0.2 0.4 0.6 0.8 Initial phosphate (mM)

1.0

Figure 2. Experimental data (points) and model output (lines) for well-mixed batch reactors when including uranium phosphate surface complex reactions (SC) and different forms of calcium phosphate precipitates: Ca3 ( PO4 )2 ( s ) (ACP), Ca ( PO4 )0.74 H0.22 ( s ) (OCP), or Ca5 ( PO4 )3 OH ( s ) (HAP). Note that under phosphate-free conditions (red color), no uranium phosphate surface complexation (SC) or calcium phosphate precipitation (ACP, OCP or HAP) form so all model output lines overlap. (C) Simulated U(VI) aqueous and surface species under initial U(VI)(aq) of 2.6×10-3 mM. Filled symbols in Figure 2C represent the total U(VI) aqueous and sorbed species. Characterization of column physical properties. The bromide breakthrough curves in

237

all columns have early starts and long tails compared to the corresponding homogeneous case

238

during the U(VI) uptake phase (Figure S3), indicating some extent of heterogeneity although

239

sediments are packed to be homogeneous to the extent possible. This may be caused by the

10 ACS Paragon Plus Environment

Page 11 of 22

Environmental Science & Technology

240

internal porosity of sediments. Although it is challenging to fully characterize, several studies of

241

the Hanford sediments have shown the internal porosity (e.g., unconnected pores) through

242

backscattered scanning electron microscopy of grains.44,45 This means water preferentially flows

243

through well-connected pores, leading to the formation of non-symmetric breakthrough curves.

244

The dual-domain model reproduces the observed breakthrough curves of all columns in both

245

U(VI) uptake and release phases with estimated physical transport parameters (Table 2). The

246

ratios of immobile to mobile domain porosity in all columns are around 0.75, indicating similarly

247

distributed pore water (43%/57%) between immobile and mobile domains. The ColF columns

248

(i.e., ColF-U and ColF-U-P) with relatively fast flow (8.9×10-6 m/s) have relatively high

249

dispersivity αL and small mass transfer coefficient α compared to the ColS columns, indicating

250

less accessible immobile zones at short residence times in ColF columns. Similar values of α

251

(e.g., 1.1×10-5 s-1 for the < 2.0 mm sediments) have been previously reported for Hanford

252

sediments.21 The physical parameters of the columns are directly used in the reactive transport

253

simulations for U(VI).

254

Table 2. Physical Parameters in the Dual-Domain Model for Column Experiments Columns

Porositymobile (ϕm) 0.17 0.17

ColS-U & ColS-U-P ColF-U & ColF-U-P

Porosityimmobile (ϕim) 0.13 0.12

Dispersivity (αL, cm) 0.5 1.0

Transfer coefficient (α, s-1) 2.0×10-5 1.5×10-5

Flow velocity (m/s) 4.4×10-6 8.9×10-6

255 256

Uranium reactive transport in columns. The calibrated model reproduces breakthrough

257

curves of U(VI), Ca, and phosphate during the uptake and release phases (Figure 3). This

258

indicates that the dual-domain reactive transport model has captured the major dynamics of the

259

system by coupling geochemical parameters from batch reactors and physical parameters from

260

tracer breakthrough curves. During the uptake phase, effluent U(VI) starts out as low as the

261

background and then approaches the influent after 68 residence times in ColS-U and ColS-U-P,

262

and 74 residence times in ColF-U and ColF-U-P, respectively. At the end of the uptake phase, the

263

total sorbed U(VI) (= Q ∫0

264

mmol,

265

Q∫

266

flow promotes uranium sorption (reaction (3) in Table 1). Such transport-controlled adsorption

267

induced by physical heterogeneity (e.g., heterogeneous permeability and porosity distribution)

268

have been observed in other column experiments46 and field studies.47,48

tuptake

tuptake

0

respectively.

( Cin − Cout )dt

( Cin − Cout )dt ) in ColS and ColF are around 1.75×10-3 and 1.52×10-3

The

calculated

corresponding

average

sorption

rates

(=

tuptake ) are 8.7×10-6 and 1.5×10-5 mmol/hr, respectively, indicating that fast

11 ACS Paragon Plus Environment

Environmental Science & Technology

Page 12 of 22

269

During the release phase, effluent U(VI) drops to less than 0.5 µM within 5 residence

270

times. After 58 and 113 residence times, 83% and 80% of adsorbed U(VI) still remain in columns

271

ColS-U-P and ColF-U-P, respectively. The effluent Ca and phosphate concentrations are lower

272

than influent concentrations, indicating precipitation of calcium phosphate. Based on the

273

difference between influent and effluent Ca and phosphate, ColS-U-P and ColF-U-P has a Ca:P

274

molar ratio of 1.41 and 1.62, respectively, which is close to the ratio of 1.50 in ACP.

275

To further understand the role of phosphate, we used the reactions inferred from batch

276

reactors and compared 4 scenarios under (i) phosphate free (blue line in Figure 3A1-A3) and (ii)

277

phosphate amendment conditions however with different processes (red lines): (ii-1) no ACP

278

precipitation (light red); (ii-2) ACP precipitation controlled by kinetics (bright red); and (ii-3)

279

ACP precipitation controlled by thermodynamics (i.e., extremely fast precipitation, dark red).

280

Note that the simulation results under the scenario (i) capture the trend of U(VI) release data in

281

experimental columns with the same uptake and release phase operation from our previous work13

282

(Figure S4). This also validates our model in predicting U(VI) reactive transport under phosphate

283

free conditions. The comparison confirms the conclusions that phosphate addition enhances

284

U(VI) immobilization by 1) forming S-U-C and S-U-P surface complexes and 2) enhancing

285

U(VI) adsorption via ACP precipitation thus decreasing Ca concentration. Compared to the

286

scenario of ii-2, fast ACP precipitation (ii-3) lowers Ca rapidly and underestimates U(VI) and

287

phosphate concentrations, indicating that the importance of ACP precipitation kinetics in

288

capturing the system dynamics.

289

Several studies have suggested that U(VI) can be sorbed onto or be incorporated into

290

amorphous calcium phosphate solids.16,49 The volume fraction (v/v) of the ACP precipitate in

291

ColS-U-P after 126 RT and in ColF-U-P after 187RT is 4.7×10-4 and 7.5×10-4, respectively, much

292

smaller than the sediments (~ 1.0 v/v). Its capacity to adsorb or incorporate U(VI) is estimated to

293

be ~6.5 µmol/g ACP,16 which is also smaller compared to the sediment sorption capacity of ~ 50

294

µmol/g. Therefore, the U(VI) immobilization through adsorption onto or incorporation into ACP

295

is considered negligible here. At longer time scales, however, if phosphate is injected

296

continuously, the uptake of U(VI) by continued precipitation of calcium phosphate can become

297

important.

12 ACS Paragon Plus Environment

Page 13 of 22

Environmental Science & Technology

A2

ColS-U-P

2.0 1.0 0.0

298 299 300 301 302 303 304 305 306 307 308

With PO34& thermodynamic prep.

0.3

No PO43-

0.0

Influent Ca

0.9 0.6 0.3 0.0

B2 Influent U(VI)

3.0 ColF-U-P

2.0 1.0

50

100 150 Residence time

B3

1.2

ColF-U

0.0 0

3-

With PO4 & no prep.

With PO34 & kinetic prep.

0.6

B1 Effluent phosphate (mM)

Effluent U(VI) (×10-3 mM)

4.0

Influent PO43-

0.9

Effluent Ca (mM)

ColS-U

1.2

200

1.2 3-

Influent PO4

0.9

Effluent Ca (mM)

3.0

A3

1.2

Influent U(VI)

Effluent phosphate (mM)

Effluent U(VI) (×10-3 mM)

4.0 A1

0.6 0.3 0.0 0

50

100 150 Residence time

200

Influent Ca

0.9 0.6 0.3 0.0 0

50

100 150 Residence time

200

Figure 3. Experimental column data (points) and model output (lines) of aqueous uranium, phosphate, and calcium in (A) ColS-U and ColS-U-P, and (B) ColF-U and ColF-U-P. Four scenarios were run to explicitly explore the role of calcium phosphate precipitation in Figure 3A1-A3: (i) phosphate free (blue line) and (ii) phosphate amendment conditions however with different processes (red lines): (ii-1) no ACP precipitation (light red); (ii-2) ACP precipitation controlled by kinetics (bright red); and (ii-3) ACP precipitation controlled by thermodynamics (dark red). Light gray, blue and light red lines overlap in Figure 3A3. For ColF-U and ColF-U-P in Figure 3B1-B3, only the best-fit scenario (ACP precipitation controlled by kinetics) is shown. The model reproduces the magnitude of the sorbed uranium concentration, as shown in 2−

309

Figure 4A. The adsorbed U(VI) ( ≡ SOUO2 (CO3 HCO3 )

310

similarly between the mobile and immobile domain (Figure S5A). The model however predicts a

311

spatial distribution of adsorbed U(VI) that is different from observations at the end of the release

312

phase. The modeled U(VI) on solids increases while the observed concentration decreases along

313

the flow path (Figure 4A) largely due to decreasing U(VI)(aq) concentrations as the sediments

314

continue to adsorb U(VI) from the inlet. This is probably due to the use of thermodynamically

315

controlled U(VI) desorption in the model. When phosphate is injected, the model predicts that the

316

sorbed U(VI) (S-U-C) during the uptake phase is quickly released back to the solution due to the

317

thermodynamics control and then resorb onto the sediments along the flow path, leading to high

318

2− sorbed U(VI) concentrations in a form of ≡ SOUO2PO4 (S-U-P) on sediments toward the outlet

319

of the column (126 residence times in Figure 4B). In reality local U(VI) desorption can happen at

320

a much slower rate due to the local spatial heterogeneity and transport control,14 therefore leading

321

to less aqueous U(VI) and lower resorbed U(VI) at the outlet. The data-model discrepancy is

322

much smaller in ColF-U-P because the longer residence time (187 residence times) diminishes

, S-U-C in Figure 4B) is distributed

13 ACS Paragon Plus Environment

Environmental Science & Technology

Page 14 of 22

323

the difference caused by transport limitation between the immobile and mobile zones (Figure

324

S5B). This is confirmed by the reproduction of effluent U(VI) data (i.e., U(VI) dynamics at the

325

domain scale) where the water-solid contact time is sufficiently long (Figure 3).

326 327 328 329 330 331 332 333

Figure 4. (A) Data and model output of sorbed U(VI) at 68, 74, 126, and 187 residence times in ColS-U, ColF-U, ColS-U-P, and ColF-U-P, respectively. (B) Predicted U(VI) speciation on the solid phase in ColS-U-P at 68, 87, 126, and 181 residence times. The sorbed U(VI) data in Figure 4A are the difference between the total extracted concentrations and the background concentration (Figure S6). Note that 68 residence times (68RT) is the end of U(VI) uptake phase and the start of the release phase in ColS-U-P.

334

Sensitivity Analysis

335

Effects of calcium and phosphate. To further evaluate uranium immobilization in the presence

336

of phosphate, we carried out a series of sensitivity analysis after the U(VI) uptake phase with the

337

same adsorbed content and specifications however under different geochemical conditions. More

338

U(VI) is immobilized under high phosphate concentrations (Figure 5A), as expected from

339

pervious experimental work.13 On the other hand, high calcium concentrations mobilize sorbed

340

U(VI) (Figure 5B) because of the formation of Ca2UO2(CO3)3(aq) and CaUO2(CO3)32- aqueous

341

2− complexes, which inhibits the formation of surface complexes ≡ SOUO2 (CO3HCO3 ) and

342

≡ SOUO2PO42− .

343

The U(VI) immobilization ratio β varies by more than one order of magnitude when

344

calcium and phosphate concentrations vary by about two orders of magnitude. Under low calcium

345

concentrations (< 0.1 mM), β values are high (> 0.6) even after 181 residence times (498 hrs).

346

The β values are close to one when phosphate exceeds 1.0 mM. Under high calcium

347

concentrations (> 1.0 mM) and calcium-to-phosphate ratios (Ca:P > 1.5), phosphate addition has

348

negligible impact on U(VI) immobilization. More than 80% of U(VI) is flushed out (β < 0.2) and

349

most phosphate precipitates as ACP. When Ca:P is lower than 1.5, β values are higher than 0.5.

350

This indicates that the Ca:P determines the effectiveness of phosphate amendment on U(VI)

14 ACS Paragon Plus Environment

Page 15 of 22

Environmental Science & Technology

351

immobilization. This explains the observed differences among experimental studies in

352

literature.13,17 For example, at the Rifle site where the groundwater has high calcium

353

concentration (~5.0 mM), phosphate (~ 1.0 mM) was observed to be much less effective in

354

immobilizing U(VI)17 than at the Hanford 300 Area.

355 356 357 358 359 360 361 362

Figure 5. (A) Simulated effluent U(VI) concentrations as a function of residence times under different influent (A) phosphate and (B) calcium concentrations. (C) Simulated U(VI) immobilization ratio β after 181 residence times (= 498 hrs) as a function of influent calcium and phosphate concentrations. U(VI) is immobilized under high phosphate and low Ca conditions. The dark and light red bar line represents the typical range of Ca concentration in Hanford and Rifle groundwater, respectively.50,51

363

Effects of pH and TIC. The pH and total inorganic carbonate (TIC) vary significantly in natural

364

subsurface conditions and can largely influence the effectiveness of phosphate amendment.

365

Figure 6A-B shows that high pH and TIC concentrations minimize U(VI) retention on the solid

366

phase. In contrast, low pH and TIC concentrations lower effluent U(VI) concentrations and

367

enhance U(VI) immobilization. Note that all release cases here are after the U(VI) uptake phase

368

with the same adsorbed content. Under low TIC concentrations (< 0.3 mM), the immobilization

369

ratios β approach 0.9 and are relativley constant under different pH conditions. At high TIC

370

concentrations (> 3.0 mM), pH effects are also negligible; however, β values are lower than 0.1

371

because of the formation of U(VI)-carbonate aqueous complexes. In the intermediate TIC

372

concentration range (0.3 ~ 3.0 mM), the immobilization ratios depend most strongly on TIC

373

concentrations and pH values because of the relative concentrations between calcium and

374

carbonate. At this range, high TIC concentrations and pH significantly increase CO32-

375

concentrations so that more mobile U(VI) aqueous complexes form, leading to decreasing

376

adsorption. At low or high TIC concentrations, the formation of U(VI)-Ca-CO3 complexes

377

however is limited and minimally influence U(VI) immobilization.

378

Note that calcite precipitation occurs at high pH (≥ 8.5) and TIC concentrations (≥ 2.0

379

mM), where β values are lower than 0.1. Although the production of calcite reduces Ca2+ and

380

CO32- concentrations, this has little impact on U(VI) immobilization as Ca2+ and CO32-

15 ACS Paragon Plus Environment

Environmental Science & Technology

Page 16 of 22

381

concentrations are still about 2 orders of manitude higher than the U(VI) aqueous concentration.

382

On the other hand, several studies documented that U(VI) can coprecipitate with calcite.52

383

However, the volume fraction of calcite after 113 residence times in this work is relatively small

384

(~ 10-5 v/v) compared to the calcium phosphate precipitates (~ 10-4 v/v) and sediments (~ 1.0 v/v).

385 386 387 388 389 390 391 392

Figure 6. (A) Simulated U(VI) breakthrough curves under different influent (A) pH and (B) TIC concentrations. (C) U(VI) immobilization ratios after 113 residence times from simulations as a function of influent pH and TIC concentrations. Note that all cases here have the same initial U(VI) content and speciation for the release phase. The residence time only reflects the duration of the U(VI) release phase. The dark and light red line in Figure 6C represents the typical ranges of TIC concentrations and pH in Hanford and Rifle groundwater, respectively.50,51,53

393

Effects of Flow Conditions. Figure 7A-B shows U(VI) release under different flow velocity and

394

α conditions. Note that all cases here have the same uptake stage and therefore the same initial

395

U(VI) content. In general, under the same geochemical conditions, low flow enhances U(VI)

396

immobilization because of the long residence time that allows more U(VI) sorption in the

397

immobile zone. Compared to flow velocity, the impact of α is much smaller, because most U(VI)

398

preferentially desorbs from the advection-dominant mobile domain during the release phase of

399

113 residence times (Figure S5). The immobilization ratios are smaller under small α conditions

400

because the effects of enhanced U(VI) immobilization induced by calcium-phosphate

401

precipitation in the immobile domain are not significant and phosphate diffuses relatively slowly

402

into the immobile domain under small α conditions. This means that spatial heterogeneity has a

403

relatively small effect on U(VI) immobilization. In other words, the low flow velocity leads to

404

negligible impact of spatial heterogeneity on reactions. This is consistent with previous

405

observations for mineral dissolution.54 With the flow velocity varying by 2 orders of magnitude,

406

the changes of β values are relatively small (from 0.72 to 0.40).

16 ACS Paragon Plus Environment

Page 17 of 22

Environmental Science & Technology

407 408 409 410 411 412 413

Figure 7. Simulated U(VI) breakthrough curves under different (A) flow velocity u and (B) mass transfer coefficient α conditions. (C) U(VI) immobilization ratio β after 113 residence times from simulations as a function of u and α. The residence time only reflects the duration of the U(VI) release phase. The dark and light red line in Figure 7C represents the typical range of groundwater velocity in Hanford and Rifle subsurface, respectively.55-57

414

Environmental Implications. This work used a limited set of batch and flow-through

415

experimental data to develop a reactive transport model that enables the process-based

416

understanding of U(VI)-Ca-phosphate interactions under a range of geochemical (e.g., pH, Ca,

417

TIC and phosphate concentrations) and flow conditions, therefore unifying contrasting

418

observations under different geochemical conditions in the literature.

419

Previous studies have observed uranium phosphate precipitates as the major reaction

420

mechanism contributing to enhanced U(VI) immobilization with high U(VI) levels (10 – 100 µM)

421

under laboratory conditions.16,26 This study reveals that under simulated natural water conditions

422

where uranium is often at ~ 1 µM, phosphate addition enhances immobilization of U(VI) without

423

the formation of U(VI)-phosphate precipitates. Instead, the combination of experiments and

424

models suggests that U(VI) immobilization is enhanced by: 1) the formation of U(VI)-phosphate

425

ternary surface complex (S-U-P) with strong sediment surface binding; and 2) the lowering of

426

mobile Ca2(UO2)(CO3)3(aq) and Ca(UO2)(CO3)32- through the formation of Ca3(PO4)2(s), which

427

allows more U(VI) adsorption through surface complex reactions.

428

Sensitivity analysis provides a comprehensive picture of conditions under which

429

phosphate addition promotes U(VI) immobilization. Phosphate addition is more effective under

430

relatively low calcium, pH and TIC conditions where U(VI)-phosphate surface complexes can

431

easily form (e.g. the Harford site). In natural water with high calcium or TIC concentrations (e.g.

432

the Rifle site), the phosphate addition needs to be sufficiently high to exceed those of calcium or

433

TIC, so that the aqueous complexes of U(VI)-Ca-CO3 do not dominate the partitioning of U(VI)

434

between aqueous and sorbed species. Compared with geochemical conditions, flow conditions

435

have relatively limited effects on U(VI) immobilization: two orders of magnitude increase in flow

436

velocity only leads to a 29.5% decrease in U(VI) immobilization after 113 residence times of

17 ACS Paragon Plus Environment

Environmental Science & Technology

Page 18 of 22

437

flushing with U(VI)-free water. On the other hand, the role of transfer coefficient α, a quantitative

438

measure of spatial heterogeneity, is time dependent. Large α values increase the immobilization

439

in early times. However, at relatively long time scales (> 100 residence times), it does not affect

440

U(VI) immobilization.

441

Here surface complexation parameters independently characterized from well-mixed

442

batch reactors were used to simulate U(VI) reactive transport in natural sediments under variable

443

flow conditions. With constraints of a geochemical reaction network from batch reactors and

444

physical transport from non-reactive tracer breakthrough curves, the dual-domain reactive

445

transport model with surface complexation thermodynamics reproduces observations for both

446

aqueous and solid phases, indicating the potential of directly using parameters estimated from

447

batch reactors in columns with coupled reactive transport, as long as the characteristics of

448

heterogeneity (here dual domain parameters) are sufficiently represented. It potentially provides

449

an effective approach to predict contaminant reactive transport at large scales where geochemical

450

data are expensive to obtain. This message echoes conclusions from other studies that reaction

451

parameters from well-mixed batch reactors can often be used directly, if the physical and

452

geochemical heterogeneities of the systems are represented at sufficient levels of details.54,58

453 454

Supporting Information

455

Supporting information is available online, which include model outputs under a variety of

456

different geochemical conditions.

457 458

Acknowledgements

459

This work was supported by the U.S. Department of Energy (DOE) Subsurface Biogeochemical

460

Research program (No. DE-SC0006857). We appreciate the editor Dr. David Waite for handling

461

the manuscript and three anonymous reviewers whose critical comments helped us significantly

462

improve the work.

463 464 465 466 467 468 469 470 471 472

References 1. Campbell, K. M.; Gallegos, T. J.; Landa, E. R., Biogeochemical aspects of uranium mineralization, mining, milling, and remediation. Appl Geochem 2015, 57, 206-235. 2. Guillaumont, R.; Fanghanel, T.; Fuger, J.; Grenthe, I.; Neck, V.; Palmer, D. A.; Rand, M. H., Update on the Chemical Thermodynamics of Uranium, Neptunium, Plutonium, Americium and Technetium. Elsevier: Amsterdam: 2003. 3. Troyer, L. D.; Maillot, F.; Wang, Z. M.; Wang, Z. M.; Mehta, V. S.; Giammar, D. E.; Catalano, J. G., Effect of phosphate on U(VI) sorption to montmorillonite: Ternary complexation and precipitation barriers. Geochim Cosmochim Ac 2016, 175, 86-99.

18 ACS Paragon Plus Environment

Page 19 of 22

473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521

Environmental Science & Technology

4. Singh, A.; Ulrich, K. U.; Giammar, D. E., Impact of phosphate on U(VI) immobilization in the presence of goethite. Geochim Cosmochim Ac 2010, 74, (22), 6324-6343. 5. Waite, T. D.; Davis, J. A.; Payne, T. E.; Waychunas, G. A.; Xu, N., Uranium(VI) adsorption to ferrihydrite: Application of a surface complexation model. Geochim Cosmochim Ac 1994, 58, (24), 5465-5478. 6. Barnett, M. O.; Jardine, P. M.; Brooks, S. C., U(VI) adsorption to heterogeneous subsurface media: Application of a surface complexation model. Environ Sci Technol 2002, 36, (5), 937-942. 7. Buck, E. C.; Brown, N. R.; Dietz, N. L., Contaminant uranium phases and leaching at the Fernald site in Ohio. Environ Sci Technol 1996, 30, (1), 81-88. 8. Li, L.; Steefel, C.; Kowalsky, M.; Englert, A.; Hubbard, S., Effects of physical and geochemical heterogeneities on mineral transformation and biomass accumulation during a biostimulation experiment at Rifle, Colorado. J Contamin Hydrol 2010, 112, 45 - 63. 9. Bao, C.; Wu, H. F.; Li, L.; Newcomer, D.; Long, P. E.; Williams, K. H., Uranium Bioreduction Rates across Scales: Biogeochemical Hot Moments and Hot Spots during a Biostimulation Experiment at Rifle, Colorado. Environ Sci Technol 2014, 48, (17), 1011610127. 10. Jerden, J. L.; Sinha, A. K.; Zelazny, L., Natural immobilization of uranium by phosphate mineralization in an oxidizing saprolite-soil profile: chemical weathering of the Coles Hill uranium deposit, Virginia. Chem Geol 2003, 199, (1-2), 129-157. 11. Singer, D. M.; Zachara, J. M.; Brown, G. E., Uranium Speciation As a Function of Depth in Contaminated Hanford Sediments - A Micro-XRF, Micro-XRD, and Micro- And Bulk-XAFS Study. Environ Sci Technol 2009, 43, (3), 630-636. 12. Munasinghe, P. S.; Madden, M. E. E.; Brooks, S. C.; Madden, A. S. E., Dynamic interplay between uranyl phosphate precipitation, sorption, and phase evolution. Appl Geochem 2015, 58, 147-160. 13. Pan, Z. Z.; Giammar, D. E.; Mehta, V.; Troyer, L. D.; Catalano, J. G.; Wang, Z. M., Phosphate-Induced Immobilization of Uranium in Hanford Sediments. Environ Sci Technol 2016, 50, (24), 13486-13494. 14. Cheng, T.; Barnett, M. O.; Roden, E. E.; Zhunag, J. L., Reactive transport of uranium(VI) and phosphate in a goethite-coated sand column: An experimental study. Chemosphere 2007, 68, (7), 1218-1223. 15. Singh, A.; Catalano, J. G.; Ulrich, K. U.; Giammar, D. E., Molecular-Scale Structure of Uranium(VI) Immobilized with Goethite and Phosphate. Environ Sci Technol 2012, 46, (12), 6594-6603. 16. Mehta, V. S.; Maillot, F.; Wang, Z. M.; Catalano, J. G.; Giammar, D. E., Effect of Reaction Pathway on the Extent and Mechanism of Uranium(VI) Immobilization with Calcium and Phosphate. Environ Sci Technol 2016, 50, (6), 3128-3136. 17. Mehta, V. S.; Maillot, F.; Wang, Z. M.; Catalano, J. G.; Giammar, D. E., Transport of U(VI) through sediments amended with phosphate to induce in situ uranium immobilization. Water Research 2015, 69, 307-317. 18. Li, L.; Maher, K.; Navarre-Sitchler, A.; Druhan, J.; Meile, C.; Lawrence, C.; Moore, J.; Perdrial, J.; Sullivan, P.; Thompson, A.; Jin, L.; Bolton, E. W.; Brantley, S. L.; Dietrich, W. E.; Mayer, K. U.; Steefel, C. I.; Valocchi, A.; Zachara, J.; Kocar, B.; McIntosh, J.; Tutolo, B. M.; Kumar, M.; Sonnenthal, E.; Bao, C.; Beisman, J., Expanding the role of reactive transport models in critical zone processes. Earth-Sci. Rev. 2017, 165, 280-301. 19. Liu, C. X.; Shi, Z. Q.; Zachara, J. M., Kinetics of Uranium(VI) Desorption from Contaminated Sediments: Effect of Geochemical Conditions and Model Evaluation. Environ Sci Technol 2009, 43, (17), 6560-6566.

19 ACS Paragon Plus Environment

Environmental Science & Technology

522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570

Page 20 of 22

20. Liu, C. X.; Shang, J. Y.; Kerisit, S.; Zachara, J. M.; Zhu, W. H., Scale-dependent rates of uranyl surface complexation reaction in sediments. Geochim Cosmochim Ac 2013, 105, 326341. 21. Shang, J. Y.; Liu, C. X.; Wang, Z. M.; Zachara, J., Long-term kinetics of uranyl desorption from sediments under advective conditions. Water Resour Res 2014, 50, (2), 855-870. 22. Li, L.; Steefel, C. I.; Williams, K. H.; Wilkins, M. J.; Hubbard, S. S., Mineral Transformation and Biomass Accumulation Associated With Uranium Bioremediation at Rifle, Colorado. Environ Sci Technol 2009, 43, (14), 5429-5435. 23. Curtis, G. P.; Davis, J. A.; Naftz, D. L., Simulation of reactive transport of uranium(VI) in groundwater with variable chemical conditions. Water Resour Res 2006, 42, (4), W04404. 24. Greskowiak, J.; Hay, M. B.; Prommer, H.; Liu, C. X.; Post, V. E. A.; Ma, R.; Davis, J. A.; Zheng, C. M.; Zachara, J. M., Simulating adsorption of U(VI) under transient groundwater flow and hydrochemistry: Physical versus chemical nonequilibrium model. Water Resour Res 2011, 47, W08501. 25. Yabusaki, S. B.; Fang, Y.; Long, P. E.; Resch, C. T.; Peacock, A. D.; Komlos, J.; Jaffe, P. R.; Morrison, S. J.; Dayvault, R. D.; White, D. C.; Anderson, R. T., Uranium removal from groundwater via in situ biostimulation: Field-scale modeling of transport and biological processes. J Contam Hydrol 2007, 93, (1-4), 216-235. 26. Comarmond, M. J.; Steudtner, R.; Stockrnann, M.; Heim, K.; Muller, K.; Brendler, V.; Payne, T. E.; Foerstendorf, H., The Sorption Processes of U(VI) onto SiO2 in the Presence of Phosphate: from Binary Surface Species to Precipitation. Environ Sci Technol 2016, 50, (21), 11610-11618. 27. Cheng, T.; Barnett, M. O.; Roden, E. E.; Zhuang, J. L., Effects of phosphate on uranium(VI) adsorption to goethite-coated sand. Environ Sci Technol 2004, 38, (22), 60596065. 28. Stoliker, D. L.; Kent, D. B.; Zachara, J. M., Quantifying Differences in the Impact of Variable Chemistry on Equilibrium Uranium(VI) Adsorption Properties of Aquifer Sediments. Environ Sci Technol 2011, 45, (20), 8733-8740. 29. Stoliker, D. L.; Liu, C. X.; Kent, D. B.; Zachara, J. M., Characterizing particle-scale equilibrium adsorption and kinetics of uranium(VI) desorption from U-contaminated sediments. Water Resour Res 2013, 49, (2), 1163-1177. 30. Arai, Y.; Marcus, M. K.; Tamura, N.; Davis, J. A.; Zachara, J. M., Spectroscopic evidence for uranium bearing precipitates in vadose zone sediments at the Hanford 300-area site. Environ Sci Technol 2007, 41, (13), 4633-4639. 31. McKinley, J. P.; Zachara, J. M.; Wan, J.; McCready, D. E.; Heald, S. M., Geochemical controls on contaminant uranium in vadose Hanford formation sediments at the 200 area and 300 area, Hanford Site, Washington. Vadose Zone J 2007, 6, (4), 1004-1017. 32. Tessier, A.; Campbell, P. G. C.; Bisson, M., Sequential extraction procedure for the speciation of particulate trace-metals. Analytical Chemistry 1979, 51, (7), 844-851. 33. Parkhurst, D. L.; Appelo, C. A. J. Description of input and examples for PHREEQC version 3-A computer program for speciation, batch-reaction, one-dimensional transport, and inverse geochemical calculations; U.S. Geological Survey Techniques and Methods: 2013, book 6, chap. A43, 497 p. 34. Rotter, B. E.; Barry, D. A.; Gerhard, J. I.; Small, J. S., Modeling the effectiveness of U(VI) biomineralization in dual-porosity porous media. J Hydrol 2011, 402, (1-2), 14-24. 35. Greskowiak, J.; Gwo, J.; Jacques, D.; Yin, J.; Mayer, K. U., A benchmark for multi-rate surface complexation and 1D dual-domain multi-component reactive transport of U(VI). Computational Geosciences 2015, 19, (3), 585-597.

20 ACS Paragon Plus Environment

Page 21 of 22

571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620

Environmental Science & Technology

36. Vangenuchten, M. T.; Wagenet, R. J., Two-site/two-region models for pesticide transport and degradation: theoretical development and analytical solutions. Soil Science Society of America Journal 1989, 53, (5), 1303-1310. 37. Thoenen, T.; Hummel, W.; Berner, U.; Curti, E., The PSI/Nagra Chemical Thermodynamic Database 12/07. 2014. 38. Lasaga, A. C., Kinetic theory in the earth sciences. Princeton University Press: Princeton, 1998; p 811. 39. Romero-Gonzalez, M. R.; Cheng, T.; Barnett, M. O.; Roden, E. E., Surface complexation modeling of the effects of phosphate on uranium(VI) adsorption. Radiochimica Acta 2007, 95, (5), 251-259. 40. Gao, Y.; Mucci, A., Acid base reactions, phosphate and arsenate complexation, and their competitive adsorption at the surface of goethite in 0.7 M NaCl solution. Geochim Cosmochim Ac 2001, 65, (14), 2361-2378. 41. Christoffersen, M. R.; Christoffersen, J.; Kibalczyc, W., Apparent solubilities of two amorphous calcium phosphates and of octacalcium phosphate in the temperature range 30– 42 C. Journal of Crystal Growth 1990, 106, (2-3), 349-354. 42. Tang, R. K.; Hass, M.; Wu, W. J.; Gulde, S.; Nancollas, G. H., Constant composition dissolution of mixed phases II. Selective dissolution of calcium phosphates. Journal of Colloid and Interface Science 2003, 260, (2), 379-384. 43. Cumberland, S. A.; Douglas, G.; Grice, K.; Moreau, J. W., Uranium mobility in organic matter-rich sediments: A review of geological and geochemical processes. Earth-Sci Rev 2016, 159, 160-185. 44. Zachara, J.; Brantley, S.; Chorover, J.; Ewing, R.; Kerisit, S.; Liu, C. X.; Perfect, E.; Rother, G.; Stack, A. G., Internal Domains of Natural Porous Media Revealed: Critical Locations for Transport, Storage, and Chemical Reaction. Environ Sci Technol 2016, 50, (6), 2811-2829. 45. McKinley, J. P.; Zachara, J. M.; Liu, C. X.; Heald, S. C.; Prenitzer, B. I.; Kempshall, B. W., Microscale controls on the fate of contaminant uranium in the vadose zone, Hanford Site, Washington. Geochim Cosmochim Ac 2006, 70, (8), 1873-1887. 46. Wang, L.; Li, L., Illite Spatial Distribution Patterns Dictate Cr(VI) Sorption Macrocapacity and Macrokinetics. Environ Sci Technol 2015, 49, (3), 1374-1383. 47. Fox, P. M.; Davis, J. A.; Hay, M. B.; Conrad, M. E.; Campbell, K. M.; Williams, K. H.; Long, P. E., Rate-limited U(VI) desorption during a small-scale tracer test in a heterogeneous uranium-contaminated aquifer. Water Resour Res 2012, 48, W05512. 48. Greskowiak, J.; Prommer, H.; Liu, C.; Post, V. E. A.; Ma, R.; Zheng, C.; Zachara, J. M., Comparison of parameter sensitivities between a laboratory and field-scale model of uranium transport in a dual domain, distributed rate reactive system. Water Resour Res 2010, 46, W09509. 49. Finch, R.; Murakami, T., Systematics and Paragenesis of Uranium Minerals. Reviews in Mineralogy 1999, 38, 91-179. 50. Hartman, M. J.; Morasch, L. F.; Webber, W. D. Hanford Site groundwater monitoring for fiscal year 2005; Pacific Northwest National Laboratory, Richland, WA: 2006. 51. DOE. Final Site Observational Work Plan for the UMTRA Project Old Rifle Site; U.S Department of Energy, Grand Junction Office, Grand Junction, CO: 1999. 52. Reeder, R. J.; Nugent, M.; Tait, C. D.; Morris, D. E.; Heald, S. M.; Beck, K. M.; Hess, W. P.; Lanzirotti, A., Coprecipitation of uranium(VI) with calcite: XAFS, micro-XAS, and luminescence characterization. Geochim Cosmochim Ac 2001, 65, (20), 3491-3503. 53. Zachara, J. M.; Davis, J. A.; Mckinley, J.; Wellman, D.; Liu, C.; Qafoku, N.; Yabusaki, S. B. Uranium geochemistry in vadose zone and aquifer sediments from the 300 Area uranium plume; Pacific Northwest National Laboratory: Richland, WA: 2005.

21 ACS Paragon Plus Environment

Environmental Science & Technology

621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638

Page 22 of 22

54. Salehikhoo, F.; Li, L., The role of magnesite spatial distribution patterns in determining dissolution rates: When do they matter? Geochim Cosmochim Ac 2015, 155, 107-121. 55. Williams, M. D.; Rockhold, M. L.; Thorne, P. D.; Chen, Y. Three-dimensional groundwater models of the 300 Area at the Hanford Site, Washington State; Pacific Northwest National Laboratory, Richland, WA: 2008. 56. DOE. Remedial Actions at the Former Union Carbide Corporation Uranium Mill Sites, Rifle, Garfield County, Colorado; U.S. Department of Energy, UMTRA Project Office, Albuquerque Operations Office, Albuquerque, NM: 1990. 57. Peterson, R. E. Groundwater Monitoring and Assessment Plan for the 100-K Area Fuel Storage Basin; Pacific Northwest National Laboratory, Richland, WA: 2002. 58. Wen, H.; Li, L., An upscaled rate law for magnesite dissolution in heterogeneous porous media. Geochim Cosmochim Ac 2017, 210, 289-305.

22 ACS Paragon Plus Environment