Subscriber access provided by UNIV OF CALIFORNIA SAN DIEGO LIBRARIES
Article
Time and nanoparticle concentration affect the extractability of Cu from CuO NP amended soil Xiaoyu Gao, Eleanor Spielman-Sun, Sónia Morais Rodrigues, Elizabeth A. Casman, and Gregory V. Lowry Environ. Sci. Technol., Just Accepted Manuscript • DOI: 10.1021/acs.est.6b04705 • Publication Date (Web): 20 Jan 2017 Downloaded from http://pubs.acs.org on January 22, 2017
Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Environmental Science & Technology is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.
Page 1 of 23
Environmental Science & Technology
Time and nanoparticle concentration affect the extractability of Cu from CuO NP amended soil
1 2 3 4 5 6
Xiaoyu Gao†, §, Eleanor Spielman-Sun†, §, Sónia M. Rodrigues‡, Elizabeth A. Casman§, #, and Gregory V. Lowry†, §, *.
†
Department of Civil and Environmental Engineering, Carnegie Mellon University, Pittsburgh, Pennsylvania 15213, United States
7
‡
Centre for Environmental and Marine Studies (CESAM), Department of Chemistry, Universidade de Aveiro, 3810-193 Aveiro, Portugal
8
§
Center for Environmental Implications of NanoTechnology (CEINT), Carnegie Mellon University, Pittsburgh, Pennsylvania 15213, United
9
States
10
#
11
*Address correspondence to
[email protected] 12 13
Abstract
14
We assess the effect of CuO nanoparticle (NP) concentration and soil aging time on the
15
extractability of Cu from a standard sandy soil (Lufa 2.1). The soil was dosed with CuO NP or
16
Cu(NO3)2 at 10 mg Cu kg-1 soil (mg/kg) or 100 mg/kg total copper, then extracted using either
17
0.01M CaCl2 or 0.005M DTPA (pH 7.6) extraction fluids at selected times over 31 days. For 100
18
mg/kg CuO NP, the amount of DTPA-extractable Cu in soil increased from 3 wt% immediately
19
after mixing to 38 wt% after 31 days. In contrast, the extractability of Cu(NO3)2 was highest
20
initially, decreasing with time. The increase in extractability was attributed to CuO NP
21
dissolution in soil. This was confirmed with synchrotron X-ray absorption near edge structure
22
(XANES) measurements. The CuO NP dissolution kinetics were modeled by a first-order
23
dissolution model. Our findings indicate that dissolution, concentration, and aging time are
24
important factors influencing Cu extractability in CuO NP-amended soil, and suggest that a time
25
dependent series of extractions could be developed as a functional assay to determine the
26
dissolution rate constant.
Department of Engineering and Public Policy, Carnegie Mellon University, Pittsburgh, Pennsylvania 15213, United States
27 28
Introduction
29
Copper based nanoparticles (NP) including metallic copper (Cu NP), copper oxides (Cu2O NP
30
and CuO NP), and copper hydroxides (Cu(OH)2 NP) are manufactured nanomaterials that have
31
been used as pesticides and fungicides because of their antimicrobial properties1, 2 .They can also
32
be used as fertilizers to deliver micronutrient-Cu to plants, which can improve fertilizer
1
ACS Paragon Plus Environment
Environmental Science & Technology
33
efficiency and crop yield3,4. Copper salt (mainly as Cu(NO3)2 or CuSO4) based micronutrients
34
and pesticides have historically been widely used. Excessive use of Cu containing fertilizers and
35
pesticides may lead to negative impacts on ecosystems, soil microorganisms, microbial
36
processes5, plants6 and soil invertebrates7.
37
In the U.S., Cu containing fertilizers and pesticides are regulated, with the maximum
38
application rate of 75 kg/ha/year (USEPA, 1993). However, these regulations were determined
39
using highly soluble Cu salts (e.g. Cu(NO3)2 and CuSO4) in soil. Dynamic processes including
40
aggregation, oxidation, and dissolution will likely make the available pool of Cu derived from
41
Cu based NP time-dependent8, 9. While the importance of time on the fate and bioavailability of
42
Cu salts is documented10-12, aging effects for Cu based NP has not been elucidated. In order to
43
assess the impact of Cu based NP to agroecosystems, it is important to determine the factors
44
controlling their bioavailability in soils.
45
Chemical extraction methods are used to predict the bioavailability of metal in soil13.
46
Several single extraction methods, originally developed to determine the fraction of metals in
47
soil involved in geochemical equilibrium processes including sorption and precipitation, can
48
predict the leaching of soil metals to groundwater, their impact on ecosystems, and their
49
bioavailability for soil organisms or plants11-22. Two extraction methods, 0.01M CaCl2 extraction
50
and 0.005M diethylenetriaminepentaacetic acid (DTPA) extraction (pH 7.3~7.6) are commonly
51
used for predicting the bioavailability or lability of metals such as Cu, Zn and Cd, in soil13-20.
52
CaCl2 extraction (0.01M) predicts metal bioavailability by mimicking the chemistry of soil pore
53
water and targets the exchangeable metal ions in soil pore water which are ‘readily available’ to
54
plants in soils14, 20, 21. DTPA is a strong chelating agent that mimics the chelating effect of root
55
exudates to enhance the nutrient availability from soil for subsequent uptake 15. The DTPA
56
extraction not only targets the free ions in soil pore water, but also the carbonate-bound and the
57
organic-bound fractions of metal in soil, which could be ‘potentially available’ to plants14, 22.
58
While these extraction methods for assessing the lability of Cu in Cu salt (CuSO4 and Cu(NO3)2)
59
amended soil or for metal contaminated soils are well-developed, there are only a few reports
60
using such methods with Cu-based NP or other metal/metal oxide nanoparticles in soil23-25.
61
Recently, a few studies have used single time point CaCl2 extraction and DTPA
62
extractions to predict the lability of metal/metal oxide nanoparticles in soil. Judy et al.25 used
63
CaCl2 and DTPA extractions to estimate the bioavailability of ZnO-NP, TiO2-NP and Ag-NP in
2
ACS Paragon Plus Environment
Page 2 of 23
Page 3 of 23
Environmental Science & Technology
64
soil and concluded that these extraction methods could not predict their bioavailability to plants
65
(Medicago truncatula) in Woburn sandy soil. Pradas del Real et al.23 used DTPA and CaCl2
66
extractions to assess the labile pool of Ag in Ag NP amended soil, and concluded that the low
67
extractability of Ag in soil was consistent with the low bioavailability of Ag to plants (wheat and
68
rape) in a loamy soil. Xu et al.24 used CaCl2, EDTA and DTPA extractions to estimate the
69
bioavailability of CuO NP and TiO2 NP to soil microbes and their community structures in a
70
typical paddy soil. They observed that DTPA and EDTA extractable Cu in CuO NP amended
71
soil correlated well with microbial activity (microbial biomass, soil enzyme activity, and total
72
phospholipid fatty acids) in a CuO NP amended soil. So far, results from studies on the use of
73
chemical extraction methods to predict the bioavailability of metals from nanoparticles in soils
74
are contradictory and often inconclusive. One reason for this may be the fact that these studies
75
did not assess the rates of transformations of NP in those soils and the corresponding effect on
76
metal extractability. We hypothesize that aging time and concentration will be important factors
77
influencing these particles’ transformation and bioavailability in soil, which may explain the
78
absence of a correlation between extractability and bioavailability using a single time-point
79
extractions23-25.
80
The dissolution and transformation of some metal and metal oxide NP in soil have been
81
determined. The dissolution of copper oxide nanoparticles over time in three soils was reported
82
by McShane et al26. In their study, they measured an increase in free Cu2+ activity in soil pore
83
water over time and concluded that CuO NPs were dissolving. However, the rate of dissolution
84
of CuO NP was not modeled or reported. The present study extends this work by McShane et al.
85
by measuring pore water and SOM-associated Cu species using well-established extraction
86
methods designed to assess bioavailable fractions of Cu, by synchrotron X-ray analysis to
87
confirm changes in copper speciation, and by determining the effect of NP concentration on the
88
dissolution behavior. The transformations of metal and metal oxide nanoparticles in soil have
89
been monitored using synchrotron X-ray absorption spectroscopy (XAS)27-32 to measure changes
90
in metal speciation over time. Recently, Sekine et al.27 used XAS to monitor the change of
91
speciation of Ag-NP, AgCl-NP and Ag2S-NP in soil over time. They observed that an increase in
92
S-bound Ag species, including Ag2S-NP, Ag-cysteine and Ag-cysteine, correlates with the
93
decrease in labile Ag determined using diffusive gradients thin films. However, the Ag NP
94
transformation kinetics were not studied.
3
ACS Paragon Plus Environment
Environmental Science & Technology
95
The dissolution of a number of metal and metal oxide nanoparticles in water has been
96
reported33-39. Most studies use empirical first-order dissolution models to describe their
97
dissolution 35-38, and evidence suggests that the measured dissolution rate constants are
98
concentration dependent36. However, the dissolution rate of metal and metal oxide nanoparticles
99
in soils, where water content and SOM can greatly affect the dissolution, is less well-understood.
100
These rates are needed to understand the dynamic nature of nanoparticulate metals relative to
101
soluble metals added to soils and to parameterize fate and transport models for engineered
102
nanomaterials. 40
103
The objectives of the present study are to (a) compare the extractability of CuO NP with
104
the extractability of Cu(NO3)2 in soil, (b) quantify the extractability of CuO NP as a function of
105
time and nanoparticle concentration in a sandy (Lufa 2.1) soil (c) determine the fate processes
106
influencing the extractability of CuO NP in soil and (d) to model the dissolution kinetics of CuO
107
NP in soil from extraction experiments. We used 0.01 M CaCl2 and 0.005M DTPA (pH=7.6)
108
extraction methods to study the extractability of Cu(NO3)2 and CuO NP in aerated soils over a
109
one-month period at two different total added Cu concentrations (10 and 100 mg Cu kg-1 (mg/kg)
110
dried soil). Changes in speciation of Cu in soil were monitored using XAS to infer the
111
dissolution of CuO NP.
112
Method and Materials
113
Chemicals. CuO NP (50 nm), DTPA, (>99% (titration)) and triethanolamine (TEA, ≥99.0%
114
(GC)) were purchased from Sigma-Aldrich. Cu(NO3)2 (>98% ACS grade), calcium chloride
115
(≥99.0%, (ACS grade)) and sodium bicarbonate (≥99.7%, (ACS grade)) were purchased from
116
Fisher Scientific.
117 118
Nanoparticle Characterization. Primary particle size distribution of the CuO NP was
119
characterized by transmission electron microscopy (TEM, Hitachi H-9000 TEM microscope
120
operating at 300 kV). The hydrodynamic diameter and zeta potential of CuO NP in suspension
121
(80 mg/kg as Cu in 5mM pH=7 NaHCO3 buffer) were determined by dynamic light scattering
122
(Zetasizer Nano, Malvern). The isoelectric points of 80mM CuO NP in 5mM NaHCO3 buffer
123
and in 5mM NaNO3 were calculated from measurements of the zeta potential of the particles in
4
ACS Paragon Plus Environment
Page 4 of 23
Page 5 of 23
Environmental Science & Technology
124
suspension over a range of pH. The crystal structure of CuO NP was determined by X-ray
125
powder diffraction (XRD, Panalytical X’Pert Pro MPD X-Ray Diffractometer).
126 127
Soils and Characterization of Soil Properties. Standard soil (2.1-sandy soil) was purchased
128
from Lufa, Germany. The standard soil (Lufa 2.1) was used because it is commonly used in
129
bioavailability studies and therefore can enhance comparison from different studies. Lufa 2.1 soil
130
also contains very little extractable Cu and total Cu (as discussed later in ‘Soil and nanoparticle
131
characterization’ section), making background interference minimal. Lufa soil was air dried and
132
sieved < 2mm before shipping. The soil was further air-dried for 12 hours before all experiments.
133
Soil pH was determined according to the standard procedures recommended by the USDA41.
134
Specifically, 5 g of air-dried soil was mixed by hand for 10s with 5ml of deionized water. The
135
pH of the solution was measured after allowing the mixture to settle for 10 minutes. To
136
determine the soil moisture content, 2 g of the air dried soil were dried in an oven at 105 ºC for
137
24 h 42. The moisture content was then determined gravimetrically. Soil field moisture capacity
138
was determined using a modified cylinder method in which air-dried soil was added to a 15ml-
139
graduated cylinder. Deionized water was then added into the cylinder to wet the top 2 cm of soil.
140
After 24h, the wetting front in the soil moved downward. After removing the top 2 cm of soil,
141
the moisture content of soil above the wetting front (which was assumed to be at soil’s field
142
capacity) was determined.
143 144
Soil amendment and incubation. Two doses of CuO NP and Cu(NO3)2 were used in our study:
145
10mg Cu/kg dry soil for the low dose amendment, and 100 Cu mg/kg dry soil for the high dose
146
amendment. These two doses were selected to investigate the influence of concentration on
147
extractability of CuO NP in soil. While the low does is more realistic, the high dose provided
148
sufficient Cu concentration for XAS study. Soils were amended with CuO NPs or Cu(NO3)2. All
149
amended soil samples were incubated in 50ml centrifuge tubes under aerobic conditions between
150
0 and 31 days before being extracted and digested. Additional details of the amendment
151
procedure can be found in SI.
152 153
Total Metal Concentration. Soil total metal concentration was determined using acid digestion
154
according to USEPA Method 3050B (1996). According to the procedure, 1g of air-dried soil was
5
ACS Paragon Plus Environment
Environmental Science & Technology
155
digested with concentrated nitric acid and 30% hydrogen peroxide at 95 ºC using a hot plate.
156
After digestion, the samples were centrifuged at 3000 rpm for 10 min, followed by filtration
157
using 0.45um filter to remove fine particles in the supernatant. The filtered supernatant was
158
diluted with Milli-Q water and acidified with 20% HNO3 (final HNO3 concentration was 2%) for
159
analysis by ICP-MS (Agilent 7700x). The instrument was calibrated with a mixed calibration
160
standard (purchased from Agilent Technologies) every time before measurement. The calibration
161
ranges used for different samples can be found in table S1.
162 163
Extractions to assess the labile Cu in soil samples. After different incubation periods, 2.0 g of
164
air-dried soils or 2.3 g of wet soils were extracted with two standard extractants: The first one
165
(termed DTPA) uses a 4 mL mixture of 0.01M CaCl2, 0.005M DTPA and 0.1M triethanolamine
166
(TEA) (pH=7.6). The second one (termed CaCl2) (pH=5) uses 20 mL of 0.01M CaCl2. All
167
extractions were done using a reciprocal shaker at 180 rpm for 2 hours. Sample bottles were laid
168
horizontally in the shaker. Both wet soil and air dried soil were used to study the effect of air
169
drying. After extraction, all samples were centrifuged at 3000 rpm for 10 min, and the
170
supernatants were filtered with using a 0.2 um PTFE filter. In order to monitor the impact of
171
CuO NP suspension or Cu(NO3)2 solution on pH of soil, the pH of CaCl2 extracts for air-dried
172
amended soil and a unamended soil (no nanoparticle or Cu(NO3)2 added) were also measured to
173
estimate the soil pore water pH. The samples collected were further filtered with a 3kda filter to
174
separate the dissolved and nanoparticulate fraction of Cu in extracts. All samples were acidified
175
with 20% HNO3 (final HNO3 concentration was 2%) and Milli-Q-water and analyzed by ICP-
176
MS. Due to the large difference between Cu concentrations from CaCl2 extracts and Cu
177
concentration from DTPA extracts, different calibration ranges were used. The different
178
calibration ranges used for different samples can be found in Table S1 in supporting information.
179 180
Determination of Cu speciation in soils. Cu speciation in soils (Lufa 2.1) on 1, 4, 7 and 19 days
181
after amendment was analyzed by Cu K-edge
182
XAS at the Stanford Synchrotron Radiation Lightsource (SSRL) on Beamline 11-2. Spectra for
183
both 100mg/kg and 10mg/kg amended soils were collected. However, the signal-to-noise ratio
184
for the 10mg/kg amended soils was too poor for adequate speciation. Specifically, samples were
185
lyophilized, ground with a mortar and pestle to achieve uniformity, pressed into pellets, and
6
ACS Paragon Plus Environment
Page 6 of 23
Page 7 of 23
Environmental Science & Technology
186
placed between Kapton tape. A double crystal Si (220) monochromator was calibrated by setting
187
the first inflection of the K-edge of a metallic Cu foil to 8979 eV. Harmonic rejection was
188
achieved by detuning the monochromator crystal by 25%. Spectra of soil samples were recorded
189
in fluorescence mode at room temperature using a 100-element germanium detector. The scans
190
were averaged, energy corrected using a metallic Cu foil standard, deadtime-corrected,
191
background subtracted with E0 defined at 8988 eV, and de-glitched using SIXPack data analysis
192
software43. Spectra were analyzed by linear combination fitting (LCF) using the following
193
reference spectra: CuO NP, metallic Cu, CuSO4, Cu(NO3)2, CuPO4, Cu-cysteine, Cu2S
194
(chalcocite mineral sample) ,CuS (covellite mineral sample), Cu- iron oxide, Cu+ sorbed to
195
humic acid (Cu(I)-HA) and Cu2+ sorbed to humic acid (Cu(II)-HA). Inclusion of a reference
196
spectrum into the combination fit required at least a 10% decrease in the R-value, indicating a
197
significant change to the quality of the fit.
198 199
Dissolution kinetics. For both extraction methods, the extractable Cu (either in pore water
200
(CaCl2), or pore water plus soil bound Cu (DTPA)) is assumed to increase proportionally as the
201
CuO NPs dissolve.
202
The increase in the extractability of Cu over time is modeled using equation 1,
203 ௗா
204
ௗ௧
= ݇൫ܧ − ܧ൯
(1)
205 206
where E is the concentration of extractable Cu at time t, k is the empirical 1st order extraction rate
207
constant, and Efinal is the concentration of extractable Cu at the end of experiment. If the
208
dissolution of the CuO NP is the rate limiting step, i.e. the Cu-soil organic matter interaction is
209
much faster than the dissolution of CuO NP in soil, then the measured extraction rate constants
210
from both extractions should be similar, and equal to the CuO NP dissolution rate constant.
211 212 213
Results and Discussion
214
Soil and nanoparticle characterization. Lufa 2.1 soil is a sandy soil, containing 3 wt% clay, 11
215
wt% silt and 86 wt% sand (as provided by Lufa). It has low organic matter content (organic
216
carbon content is 0.7 wt% as provided by Lufa). After air-drying, Lufa soil had 1.2 wt% moisture
7
ACS Paragon Plus Environment
Environmental Science & Technology
Page 8 of 23
217
content. The soil pH was 5.6 and the field capacity was 16 wt%. The total Cu concentration of
218
the unamended soil was 2.95±0.11mg/kg. Total Cu concentration measured in each of the
219
amended soils is presented in Table S2. The DTPA extractable Cu in unamended soils ranged
220
from 0.37 to 0.53 mg/kg dried soil while the CaCl2 extractable Cu in unamended soils ranged
221
from 0.005 to 0.024mg/kg (Table S3).
222
The primary particle size of CuO NP (measured from TEM) was 38nm (s.d. =14nm, 278
223
particles were counted). The hydrodynamic diameter and zeta potential of 80mg/kg CuO NP in
224
pH=7, 5mM NaHCO3 buffer were 557nm (s.d. =56nm, 3 replicates, polydispersity index 10 d (p > 0.05, Kolmogorov-Smirnov test). This suggests that
277
the Cu may be fully dissolved and “aging” similarly to the Cu(NO3)2. However, the slight
9
ACS Paragon Plus Environment
Environmental Science & Technology
Page 10 of 23
278
downward trend in extractability for t > 10d is not statistically significant (P > 0.05, one-way
279
ANOVA test). Although extraction procedures generally used air dried soils21,46 , we used both the air
280
dried soils (after incubation) and wet soils for extractions to investigate the influence of air
282
drying on extractability of CuO NP in soil. Our results indicated that air drying has no significant
283
effect (P>0.05, Kolmogorov-Smirnov test) on extractability of Cu in both CuO NP amended and
284
Cu(NO3)2 amended soils. Thus, only the results from air dried soil is shown in Figure 1 for
285
clarity. Additional discussion on the effect of air drying can be found in the supporting
286
information.
Cu(NO3)2
6 4 CuO NP
2 0
287 (c)
Extractable Cu (mg /kg dried soil)
8
0
Cu(NO 3)2
50 CuO NP
25 0 0
Cu(NO 3)2
0.1
CuO NP
0.0 0
100mg/kg DTPA extraction
75
10mg/kg, CaCl 2 extraction
0.2
10 20 30 40 Incubation time (days)
100
Extractable Cu (mg /kg dried soil)
(b) 0.3
10mg/kg, DTPA extraction
(a) 10
10 20 30 40 Incubation time (days)
10 20 30 40 Incubation time (days)
100mg/kg, CaCl 2 extraction
(d) 12
Extractable Cu (mg /kg dried soil)
Extractable Cu (mg /kg dried soil)
281
Cu(NO 3)2
8 4
CuO NP 0 0
10 20 30 40 Incubation time (days)
288 289
Figure 1. Extractable Cu and in CuO NP and Cu(NO3)2 amended soils as a function of time and
290
the first order dissolution fit for CuO NP in soil: (a) DTPA extraction for 10 mg/kg amendment,
291
(b) CaCl2 extraction for 10 mg/kg amendment, (c) DTPA extraction for 100 mg/kg amendment
292
and (d) CaCl2 extraction for 100 mg/kg amendment. Error bars indicate ± 1 standard error.
293
Dashed lines indicate model fits using equation 1. For the low dose amendment, because CuO
294
NPs were fully dissolved after the 7-day sampling time, we modeled only the first 7 days.
295
represents extractable Cu in CuO NP amended soils air dried after incubation and
296
extractable Cu in Cu(NO3)2 amended soils air dried after incubation.
10
ACS Paragon Plus Environment
represents
Page 11 of 23
Environmental Science & Technology
297
298 299
Figure 2. Fraction of small particles and dissolved ions (those passing 3kDa filter) in (a) DTPA
300
extracts and (b) CaCl2 extracts. D1, D2, D31 stand for 1 day, 2 days and 31 days after dosing.
301
Error bars indicate ± 1 standard error.
302 303
Fractions of dissolved Cu and particulate Cu in extracts. Bioavailability of Cu depends on its
304
speciation, e.g. free ions, complexed ions and particulate species47. We used filtration (first a
305
0.2-micron filter followed with a 3kDa filter) to distinguish between dissolved and particulate
306
species of Cu in each of the extracts. Figure 2 shows the fraction of Cu that passes the 3 kDa
307
filter (considered dissolved) in CaCl2 and DTPA extracts. For DTPA extraction, nearly all
308
extractable Cu (from 90% to 100%) was dissolved. This is because most Cu in DTPA extracts
309
bound with the chelating agent (DTPA) and the Cu-DTPA complex can pass through the 3kDa
310
filter. In contrast, filtration of the CaCl2 extract indicated the presence of Cu-containing particles
311
compared to the DTPA extracts (P0.05, one way ANOVA test).
316 317
Effect of CuO NP concentration on its extractability in soil. The concentration of added Cu
318
influences the extraction behavior for CuO NP compared to Cu(NO3)2. For the low Cu dose, the
319
extractability of Cu in CuO NP amended soil was the same as for the Cu(NO3)2 amended soil
320
after ~10 days. No statistically significant difference (p>0.05, Kolmogorov-Smirnov test) is 11
ACS Paragon Plus Environment
Environmental Science & Technology
321
found for extractable Cu for both CaCl2 extractions and DTPA extractions between CuO NP
322
amended soil and Cu(NO3)2 amended soil on day 13, 19 and 31, suggesting that the CuO NP
323
were fully dissolved before 13 days in soil at the lower dose. The behavior was quite different at
324
the high Cu dose. For the high dose of added Cu, extractable Cu in Cu(NO3)2 amended soils was
325
always higher than the extractable Cu in CuO NP amended soil. The extractability of Cu from
326
the CuO NP amended soil increased over the entire 31day period, suggesting that CuO NP was
327
dissolving over 31 days, but the dissolution of CuO NP in soil was not complete. One possible
328
explanation on the persistence of CuO NP and the slower dissolution rate after ~7 days in the
329
high dose soil (100mg/kg Cu) is that the free Cu2+ in soil pore water approached saturation with
330
respect to CuO(s). Conversely, the lower dose system (10 mg/kg) was not oversaturated with
331
respect to the CuO(s) phase. While CaCl2 extraction is a well-established method to assess the
332
pore water concentration of dissolved Cu, the potential for artefact during the extraction and
333
uncertainty in the complexation constants for Cu and the NOM in our system prevents an
334
accurate determination of the degree of saturation in the pore water. .
335 336
Dissolution rate of CuO NP in soil. For the high dose of CuO NP (100 mg/kg), the first-order
337
extraction model describes the change of extractable Cu over time well (R2>0.995) (dashed lines
338
in Figure 1). However, we should note that Cu2+ ions dissolved from CuO NP can become
339
irreversibly bound with soil organic matter, making it unextractable by DTPA, as indicated in
340
former sections. This irreversible interaction is about 20% for our soils, and has a minimal effect
341
on the calculated dissolution rate constant. This is in part because it is a small fraction of the
342
total, and in part because the time scale for partitioning into this irreversible fraction is short, i.e.
343
less than 1d compared to the dissolution processes being investigated, i.e. many days to weeks.
344
For CaCl2 extraction, the fraction of extractable ionic Cu was significantly less, with only 2% to
345
10% of the ionic Cu being extractable because it targeted only Cu in soil pore water. Despite the
346
differences in the extractable amount of Cu, the modeled dissolution rate constants for DTPA
347
extractable Cu and CaCl2 extractable Cu are similar (Table 1). This indicates that the extractable
348
amount of Cu by either the DTPA or CaCl2 extraction can be used to monitor the CuO NP
349
dissolution in the soils. This is a natural consequence of a first-order dissolution process, which
350
scale with the ratio of the final and initial concentration (C/Co) so any process that reduced C and
351
Co by the same constant fraction will not affect the calculated rate. Moreover, it suggests that
12
ACS Paragon Plus Environment
Page 12 of 23
Page 13 of 23
Environmental Science & Technology
352
Cu2+ binding to SOM is rapid enough, such that dissolution of the CuO NP is the rate-limiting
353
process controlling both DTPA extractable Cu and CaCl2 extractable Cu in soil.
354 355
Effect of aging on speciation of Cu in Cu(NO3)2 and CuO NP amended soil. Speciation of Cu
356
in the 100mg/kg CuO NP and 100mg/kg Cu(NO3)2 amended soils were determined at selected
357
time points using XANES (Figure 3). Details regarding the spectra for model compounds and
358
fitting result can be found in the supporting information (Figure S7 and Table S4). The
359
speciation of Cu in Cu(NO3)2-amended soils can be adequately modeled using only the Cu(II)-
360
HA model compound, indicating that the Cu has predominantly Cu-O character, i.e. associated
361
with humic acids or potentially (but less likely) with clay or metal oxide surfaces of the solids.
362
This is consistent with prior speciation studies indicating that the main species of Cu in soil is
363
Cu(II)-HA using experimental approachs48-49 and with results of equilibrium partitioning
364
modeling50. This also suggests that Lufa 2.1 soil has the capacity to sorb up to 100mg/kg of
365
added Cu, because our data showed that all Cu in the 100mg/kg Cu(NO3)2 amended soil was
366
Cu(II)-HA. In contrast, the Cu speciation in CuO NP amended soil required both Cu(II)-HA and
367
CuO NP model compounds. In the high dose CuO NP amended soil, linear combination fitting
368
indicates that the presence of CuO decreases over time, with a subsequent increase in the Cu(II)-
369
HA. This suggests that the CuO NPs were dissolving relatively fast in the first 7 days and then
370
more slowly after that as the pore water becomes saturated with respect to CuO(s). The rapid
371
dissolution in the first 7 days in consistent with the DTPA and CaCl2 extractability data, which
372
increased most rapidly in the first 7 days, followed by a slower increase. The dissolution of CuO
373
NP slowed down after 7 days even though the soil has not reached its capacity to adsorb Cu,
374
which confirms our former assumption that dissolution of CuO NP is the limiting factor
375
controlling the extractability of Cu from soil. Note that we also analyzed the 10mg/kg soils and
376
the unamended soil samples, but the signal-to-noise ratio was too poor for adequate speciation.
377
13
ACS Paragon Plus Environment
Environmental Science & Technology
378
Page 14 of 23
Table 1. Modeled first-order dissolution parameters for CuO NP amended soil. Extraction type
k (day-1)
95% confidence
Half-life
E0 a
Efinal
intervals for k (day-1)
(days)
(mg/kg)
(mg/kg)
R2
High dose amendment DTPA extraction (dry soil) CaCl2 extraction (dry soil)
0.15
0.11-0.19
4.6
3.35
37.4
0.995
0.13
0.12-0.18
5.2
0.05
1.0
0.998
Low dose amendment DTPA extraction (dry soil) CaCl2 extraction (dry soil)
0.16
0.06-0.25
4.5
0.36
6.71
0.936
0.11
0.07-0.14
6.6
0.03
0.16
0.975
379 380
a: E0= initial extractable Cu at day 0 (intercept at y axis)
381 382
383 384 385
Figure 3. Change of Cu speciation in amended soils as inferred by XANES: in (a) Cu(NO3)2
386
amended soil and (b) CuO NP amended soil dosed at 100 mg/kg total Cu. The red dash lines are
387
fitted data while the black lines are experimental data. Model compounds used for the fits are
388
below the experimental spectra. The pie charts represent linear combination fits of the various
389
model compounds.
14
ACS Paragon Plus Environment
Page 15 of 23
Environmental Science & Technology
390 391
Environmental Implications
392
The extractability of Cu from CuO NP-amended soils is different from that in soils dosed with
393
Cu ions as Cu(NO3)2, suggesting that the lability of CuO NP may be different from the lability of
394
the highly soluble Cu salts used as pesticides in soils. CuO NP was much less labile than
395
Cu(NO3)2 in soil immediately after they were added to the soil, but its lability increased over
396
time. The differences in lability between CuO NP and Cu(NO3)2 became negligible at low Cu
397
doses (10 mg/kg) after about 7 days, but differences in lability remained over 31 days for the
398
high dose. The increase of the labile pool of CuO NP over time was a result of their slow
399
dissolution. Thus, our research shows that dissolution is an important process controlling the
400
extractability of CuO NP in soil, but the dissolution rate and CuO NP persistence will be
401
concentration dependent. Moreover, the aging time in soil must be considered when assessing
402
the lability or bioavailability of CuO NP in soils as was also previously suggested by Sekine et al
403
for Ag NP, Ag2S NP and AgCl NP and McShane et al. for CuO NP (dosed at 500 mg/kg)26, 27,
404
along with the total applied dose. If toxicity is purely the result of the release of copper ion, the
405
regulatory limit for applying nano CuO in agriculture could be adjusted to consider its “slow
406
release” behavior and concentration-dependent persistence. Because of the relatively slow
407
dissolution behavior of CuO NP, the regulatory limit for CuO NP could be higher than that set
408
for Cu salts. This is especially true if, with some additional surface modification, the dissolution
409
rate of Cu-based nanoparticles could be further reduced. Compared with a direct spray
410
application of Cu salt, a slow sustained release of ions from CuO NP may have lower
411
environmental impact to groundwater and rivers because particles have lower leachability and
412
mobility. On the other hand, if CuO NPs exhibit nanoparticle specific toxicity51,52, for higher
413
doses where CuO NPs persist, regulations will need to consider this persistence if CuO NPs
414
show greater toxicity than the Cu salts. Overall, the regulation of nano enhanced particles might
415
be better based on their dissolution rate at the applied dose, which could be easily determined
416
with the methods used in this study.
417
This work advances our understanding of the fate of CuO NP in several important ways.
418
First of all, we found CuO NP dissolution is the rate limiting step in controlling the increase of
419
CaCl2 extractable Cu and DTPA extractable Cu in CuO NP amended soil, indicating the
420
dissolution process of CuO NP in soil is much slower than the Cu-SOM interaction. Thus, we
15
ACS Paragon Plus Environment
Environmental Science & Technology
421
can monitor the dissolution of CuO NP in soil from either the increase in dissolved Cu in soil
422
pore water (as indicated by CaCl2 extraction) or increase in extractable Cu by DTPA extraction
423
(dissolved Cu plus Cu bound to SOM and carbonates). While McShane et al.26 suggested that soil
424
pH is an important factor controlling the dissolution of CuO NP in soil, we also suggest that the
425
amount of SOM in soil may be as or more important because it provided the sink for the released
426
Cu in the soils used here. Secondly, our research indicates that the concentration of soluble
427
nanoparticles added to the soils can affect temporal changes in Cu speciation, which in turn can
428
affect the interpretation of exposure or toxicity testing. At a low dose (10 mg/kg dried soil),
429
CuO NPs became fully dissolved within 10 days. Thus, at low doses, exposures to nanoparticles
430
after ~10 days are not occurring and exposures and toxicity testing would be expected to be
431
consistent with a dissolved Cu species. Moreover, the Cu species present was similar to Cu(II)-
432
HA found in the natural soil so responses to CuO NP amended soils at these low doses would
433
likely be similar to exposures to native soils with the same Cu concentration. However, using a
434
higher CuO NP dose (100 mg/kg dried soil), about 40% of CuO NPs remained undissolved after
435
31 days, potentially because the dissolution was limited by the solubility with respect to CuO(s).
436
In experiments using this high concentration, exposures and effects may be a result of
437
interactions with CuO NP and therefore different than for added ions or native soils.
438
Our research suggests that a single time point extraction after dosing soil may not be
439
adequate for predicting bioavailability unless that extraction is made at the same time as the end
440
point of interest (e.g. plant uptake). Rather, a time series of extractions after dosing may be more
441
appropriate for predicting the bioavailability of metal/metal oxide nanoparticles in soil. The time
442
series of extractions used here could be developed as a functional assay for studying the
443
dissolution kinetics of metal/metal oxide nanoparticles in soil. The functional assay approach has
444
recently been proposed as a means to empirically predict nanomaterial behaviors in complex
445
media53. The method that we developed is simple, and highly reproducible among the three
446
replicates in our experiments. The dissolution rate constant could be used for nanomaterial risk
447
forecasting in soil system, as suggested by Hendren et al53. Further studies need to confirm this
448
method using different metal/metal oxide nanoparticles in different soil systems. For example,
449
several well-known limitations of soil extractions methods, e.g. dilution effects, and the presence
450
of an “irreversibly bound” fraction of metal, exist. In the current study, the irreversibly bound
451
fraction was relatively low (