Subscriber access provided by UNIV OF LOUISIANA
Remediation and Control Technologies
Low Molecular Weight Organic Acid Complexation Affects Antimony(III) Adsorption by Granular Ferric Hydroxide Xiaochen Li, Tatiana Reich, Michael Kersten, and Chuanyong Jing Environ. Sci. Technol., Just Accepted Manuscript • DOI: 10.1021/acs.est.8b06297 • Publication Date (Web): 10 Apr 2019 Downloaded from http://pubs.acs.org on April 16, 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
Environmental Science & Technology
119x137mm (300 x 300 DPI)
ACS Paragon Plus Environment
Environmental Science & Technology
1
Low Molecular Weight Organic Acid Complexation Affects
2
Antimony(III) Adsorption by Granular Ferric Hydroxide
3
Xiaochen Li1,2, Tatiana Reich2, Michael Kersten2,*, Chuanyong Jing1,3**
4
1State
5
Eco-Environmental Sciences, Chinese Academy of Sciences, Beijing 100085, China
6
2Geosciences
7
3University
Key Laboratory of Environmental Chemistry and Ecotoxicology, Research Center for
Institute, Johannes Gutenberg University, Mainz 55099, Germany
of Chinese Academy of Sciences, Beijing 100049, China
8
9
ABSTRACT: Antimony(III) mobility in natural aquatic environments is generally enhanced
10
by dissolved organic matter. Tartaric acid is often used to form complexes with and stabilize
11
dissolved Sb(III) in adsorption studies. However, competition between such low molecular
12
organic acid complexation and adsorption of Sb(III) has received little attention, which
13
prompted us to measure Sb(III) adsorption by iron oxyhydroxide adsorbents commonly used in
14
water treatment plants. Sb K-edge X-ray absorption fine structure (EXAFS) spectra gave Sb–
15
O and Sb–Fe distances and coordinations compatible with a bidentate binuclear inner-sphere
16
complex with trigonal Sb(O,OH)3 polyhedra sharing corners with Fe(O,OH)6 octahedra, and a
17
bidentate mononuclear inner-sphere complex but with Sb(O,OH)4 tetrahedra at alkaline pH.
18
Experimental batch titration data were fitted using the charge-distribution multi-site surface
19
complexation (CD-MUSIC) model, constrained by the EXAFS molecular level information and
20
taking competitive effects by the organic ligand into consideration. The proportion adsorbed at
21
acid–neutral pH decreased as the Sb(III) concentration increased. The CD-MUSIC adsorption
22
model indicates that this was solely caused by strong competition from tartrate complexation ACS Paragon Plus Environment
Page 2 of 30
Page 3 of 30
Environmental Science & Technology
2 23
in solution, which leads to adsorption constants higher than those derived without taking this
24
competition into account. The adsorption model results allow for calculating a pe-pH
25
predominance diagram using the USGS PhreePlot code. The study provides consistent surface
26
complexation stability constants, allowing the new database to be used also to model reliably
27
adsorption of toxic oxyanions in anoxic aqueous environments, for example to accurately
28
simulate competition between Sb(III) and As(III).
29 30
INTRODUCTION
31
Antimony (Sb) and its compounds are important environmental contaminants of emerging
32
importance.1 Like As, Sb is primarily found in the +3 valence state (i.e., Sb(III)) and the +5
33
valence state (i.e., Sb(V)), under natural reducing and oxidizing conditions, respectively. Sb(V)
34
is the dominant species in oxygenated aqueous systems, but multiple studies have found Sb(III)
35
in marine and fresh water, groundwater, and soil porewater.1 In sulfide-free solutions, Sb(III)
36
undergoes hydrolysis to give neutral antimonyl (Sb(OH)30) and antimonite oxyanion
37
(Sb(OH)4−). There is evidence that Sb(III) may be a human skin carcinogen, as is the case for
38
As(III).2 The recommended maximum total Sb concentrations in drinking water (US
39
Environmental Protection Agency 6 g L−1, EU 5 g L−1, Japan 2 g L−1) are even lower than
40
the recommended maximum total As concentrations. Total dissolved Sb is 10 mol L−1 (the solubility of Sb2O3 in pure water at ~pH 7) to be used. Laboratory adsorption
106
studies often use adsorbate concentrations much higher than expected in natural aqueous media.
107
This is because surface complexation models require adsorption–pH curves to be fitted, and
108
there are no such curves to fit if ~100% of adsorbate is adsorbed over a broad pH range at low
109
concentrations. The experiments were therefore performed with higher concentrations of 0.5,
110
1.0 and 1.5 mmol L−1, where a pH dependence of the curves becomes obvious.
111
A 30 mL aliquot of a diluted Na tartrate or Sb(III) tartrate stock solution acidified with
112
0.1 M HNO3 to pH 3 was added to a low-density polyethylene centrifuge tube, then the sorbent
113
was added (1.7 g L−1 fresh GFH based on a pore water content of 40% by weight). Between 10
114
and 20 experiments at different pH values were performed for each background electrolyte
115
concentration (10, 50 and 100 mM NaNO3) by adding increasing amount of 1 M NaOH to a
116
series of tubes. The tubes were shaken for 24 h on a horizontal shaker in a dark room. In
117
preliminary adsorption experiments, an equilibration time of 24 h was found to be sufficient
118
even for the very porous GFH (particle size 4. The precision and accuracy of the
128
analytical procedure were assessed by analysing the NIST 1643e certified reference material
129
(US NIST, Gaithersburg, MD, USA) and the analytical uncertainty was less than 5%. Sb
130
speciation in selected solutions was determined by anodic stripping voltammetry and the results
131
indicated no oxidation to Sb(V) as shown in Figure S3 (Supporting Information). The dissolved
132
organic carbon (tartrate) concentrations in the supernatants were determined using a High TOC
133
II carbon analyser with a combustion tube and CO2 detector (Elementar Analysensysteme,
134
Hanau, Germany) using the German DIN EN 1484 (DEV H3) method.
135
X-ray Absorption Analysis. Incident X-ray energy was scanned across the Sb K-edge (30491
136
eV) using the Swiss Light Source SuperXAS beamline (PSI-Villigen, Switzerland). When the
137
measurements were made, the beamline optics consisted of a water-cooled Si(311) double
138
crystal monochromator between two Rh-coated harmonic rejection mirrors, one for beam
139
collimation and the other for focusing. The monochromator was calibrated using a Sb(0) foil.
140
A sample was mounted on the stage and then cooled to 100 K using a nitrogen cryostream
141
manipulator (Cryojet, Oxford Instruments, High Wycombe, UK) to improve data quality and
142
avoid Sb(III) photo-oxidation. Take care of the air humidity when using a cryojet device.
143
Spectra for the akaganéite samples were collected in fluorescence mode. Spectra of Sb(III) and
144
Sb(V) reference minerals were also acquired but in transmission mode at room temperature.
145
The natural reference minerals schafarzikite (FeSbIII2O4) and tripuhyite (FeSbVO4) were ground
146
to fine powders, diluted with polyethylene powder (Merck) and pressed into pellets using a 13
ACS Paragon Plus Environment
Environmental Science & Technology
7 147
mm die. A number of scans for each sample (three scans for a reference mineral and six scans
148
for an akaganéite sample) were averaged to improve the signal-to-noise ratio. The XAS data
149
were automatically deadtime-corrected. The X-ray absorption near-edge structure (XANES)
150
parts of the first and last spectra were compared to ensure that no photon-induced redox
151
reactions had occurred while a sample was being analysed.
152
The energy scales of the experimental data were recalibrated, then spline fitting was
153
performed using the AUTOBK program.18 The background correction, extraction of the
154
EXAFS parts of the spectra and harmonic analysis were performed using the EXAFSPAK
155
software package.19 The Sb K-edge k3-weighted (k) spectra for the reference minerals had
156
very good signal-to-noise ratios out to k values ~15 Å−1 (Figure 1). The resulting EXAFS (k)
157
spectra were then converted to R-space by taking the Fourier transform of (k). Least-squares
158
refinement to the schafarzikite structure was achieved using the OPT mode in the EXAFSPAK
159
program for (k) spectra of the pH 4 sample in the range 2.4–11.4 Å−1, and of the pH 7 and pH
160
10 samples in the range 2.4–12.4 Å−1. Fitting of the data to the EXAFS equation was performed
161
to determine the degeneracy N, half-path length R, and mean-square displacement 2 of the
162
backscatterers. Theoretical scattering phases and amplitudes were obtained using FEFF8.2.20 A
163
124-atom cluster based on the schafarzikite crystal structure was used to define the Hedin–
164
Lundqvist self-energy potential for FEFF calculations of the theoretical effective pathways. All
165
the interactions were modelled using single scattering (SS) interactions derived from previously
166
published structural refinements.21,22 There is no evidence of MS paths in our data. Sb K-edge
167
data processing was achieved using three paths (a SS Sb–O shell, a SS Sb–Fe shell, and a SS
168
Sb–Sb shell), which were found to be significant contributors to the EXAFS signal. Several
169
measures were taken to prevent the number of degrees of freedom exceeding the number of
170
parameters allowed to vary when fitting the data for the reference minerals. E0 was allowed
171
to vary only as a global parameter for each fitting procedure. The amplitude reduction factor ACS Paragon Plus Environment
Page 8 of 30
Page 9 of 30
Environmental Science & Technology
8 172
S02 was set at 1.0 to assure agreement between the resulting structural parameters of reference
173
samples with their XRD data. The structural results (coordination numbers N, real backscatterer
174
distances r, and Debye-Waller factors 2) of the least-squares-fitting process using the SS shell
175
contributions to the reference mineral spectra are summarized in Tables S2 and S3 (Supporting
176
Information), and for the adsorption samples in Table 1. The fit quality was defined as χ2res =
177
∑[χidata − χimodel(x)] / ∑[χidata]2, where χ is the magnitude of EXAFS oscillation, and x is the set
178
of variables to be refined. Our EXAFS agreed well with previously published XRD data for the
179
reference minerals, as shown in Tables S2 and S3 (Supporting Information). Notably, in
180
addition to the schafarzikite coordination shells from previously published XRD refinements,
181
a weak shell of 0.5 S atoms at a distance of ~2.5 Å was detected, indicative of some stibnite
182
(Sb2S3) impurity of the reference as supported by the results of the XRD phase-purity analysis.
183
RESULTS AND DISCUSSION
184
Sb K-edge XAFS data. The K-edge shape (a more or less pronounced white line) and the first
185
inflection point (30496–30497 and 30491–30493 eV for Sb(V) and Sb(III), respectively) were
186
taken from the XANES spectra. The XANES spectra together with their first derivatives of the
187
Sb(III) reference minerals and samples are shown in Figure 1. The spectra confirm that the
188
added Sb remained as Sb(III) and was not oxidized to Sb(V). The Sb K-edge height and shape
189
were identical in the samples and Sb(III) reference material. These results corroborate the
190
voltammetric aqueous speciation analysis results for the batch equilibration experiment
191
solutions (Figure S3, Supporting Information). The EXAFS spectra fits are shown in Figure 2,
192
and the structural results of the least-squares-fitting process for the akaganéite samples loaded
193
with Sb(III) at the three different pH values are summarized in Table 1. The results indicate that
194
the nearest neighbours of the Sb atoms in the samples from the adsorption experiments were O
195
atoms, each at a distance of 1.96 ± 0.02 Å, as expected for Sb(III).23
ACS Paragon Plus Environment
Environmental Science & Technology
9 196
The measured coordination numbers N were in the range of 3.4 – 4.4. These were a little
197
higher than the theoretical N of 3 expected for Sb coordinated to three O atoms (Table 1), but a
198
similar result was found previously for antimonyl surface complexes.24 On the other hand,
199
Sb(V) is coordinated to six O atoms, as shown for the reference mineral tripuhyite in Table S3
200
(Supporting Information) and in previous EXAFS studies of Sb(V) adsorption by goethite.25 As
201
shown in Table 1, the second coordination shell of Sb(III) sorbed to the pure akaganéite samples
202
had an extra feature at R 3.2 Å (not corrected for phase shift), which fitted satisfactorily to a
203
single Sb–Fe path. The Fe backscatterer was fitted to a real distance of r 3.56–3.60 Å with N of
204
3 (Table 1). We therefore conclude that our EXAFS data indicate only one main inner-sphere
205
surface complex in acidic solution. The Sb(III)–Fe path distance, 3.58 ± 0.02 Å, could be
206
interpreted to indicate a bidentate binuclear corner-sharing (2C) complex bridging two edge-
207
sharing FeO3(OH)3 octahedra with one pyramidal Sb(O,OH)3 molecule, as previously found
208
also for ferrihydrite and goethite.24 Schematics of the structural model of this surface species is
209
presented in Figure 3. Although the GFH adsorber material is comprised of the two different
210
solid phases ferrihydrite and akaganéite,16 this does thus not imply that the Sb(III) surface
211
complexes formed are different, at least not in the acidic to circumneutral pH range.
212
Least-squares fitting indicated that the pH 10 samples had another weak Sb–Fe
213
backscattering path at 3.11 ± 0.02 Å (corrected for phase shift) with a N of only ~0.7. The Sb–
214
Fe path distance, commonly interpreted as indicating an inner-sphere bidentate but
215
mononuclear edge-sharing 2E5 surface complex, was found to be required to describe the second
216
coordination shell features of the akaganétite samples conditioned at pH 10. The first-shell
217
oxygen coordination number of four indicates that tetrahedral antimonite-type Sb(OH)4− was
218
the adsorbing species in this additional binding mode. This is in agreement with the antimonite-
219
type complex dominating the dissolved Sb(III) speciation in the alkaline pH range. A schematic
ACS Paragon Plus Environment
Page 10 of 30
Page 11 of 30
Environmental Science & Technology
10 220
of the structural model of this unique 2E5 surface species is presented in Figure S4 (Supporting
221
Information).
222
Sb(III) Adsorption and Speciation. The adsorption isotherms of Sb(III) on GFH conformed
223
to the Langmuir model (Figure S5, Supporting Information). The maximum adsorption capacity
224
was 120 mg g-1, and the BET surface normalized adsorption capacity was 2.0 molecules per
225
nm2. This adsorption capacity is just half of that of self-assembly {001} TiO2 studied
226
previously,26 and 2/3 of that for pure ferrihydrite of the same nominal (i.e., dried BET) specific
227
surface area.27 Figure 4 shows the effect of pH on Sb(III) adsorption by GFH for different Sb
228
and background electrolyte concentrations. Within the pH range studied, it had a negligible
229
effect on Sb(III) adsorption as evidenced by its consistent top 98% removal in the pH range
230
310 at the lowest Sb concentration series, making it one of the most strongly adsorbing
231
oxyanions on Fe oxyhydroxides. The background electrolyte was found not to markedly affect
232
the adsorption curves for this series. Sb(III) must therefore have been bound predominantly as
233
strong inner-sphere complexes with Fe surface sites, as also suggested by the EXAFS results.
234
The pH does not affect Sb(III) sorption because the sorbate moiety is neutral. Unlike the case
235
with As(III), Sb(III) adsorption even do not decline in the alkaline pH region. This could be
236
attributed to the stronger Lewis base behavior of Sb(III) than As(III), having a higher
237
deprotonation pKa value (pKa1 = 9.17 of H3AsO3, pKa2 = 11.82 of Sb(OH)3). Moreover, as
238
indicated by the EXAFS results, the deprotonated antimonite species Sb(OH)4- occurring at
239
high pH values is adsorbed as well. On the other hand, markedly less adsorption occurred at pH
240
