Single Particle-Inductively Coupled Plasma Mass ... - ACS Publications

Subscriber access provided by CORNELL UNIVERSITY LIBRARY

Technical Note

SP-ICP-MS Analysis of Metallic Nanoparticles in Environmental Samples with Large Dissolved Analyte Fractions Dina M. Schwertfeger, Jessica R. Velicogna, Alexander H. Jesmer, Richard P. Scroggins, and Juliska I. Princz Anal. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.analchem.6b02716 • Publication Date (Web): 15 Sep 2016 Downloaded from http://pubs.acs.org on September 15, 2016

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.

Analytical Chemistry 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

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

1

Title: SP-ICP-MS Analysis of Metallic Nanoparticles in Environmental Samples

2

with Large Dissolved Analyte Fractions

3

Authors: D. M. Schwertfeger, Jessica R. Velicogna, Alexander H. Jesmer, Richard P. Scroggins,

4

Juliska I. Princz*

5

Affiliations: Biological Assessment and Standardization, Environment and Climate Change

6

Canada, 335 River Road South, Ottawa, Ontario K1V 1C7 Canada

7

* To whom correspondence may be addressed ([email protected])

8 9

Juliska Princz, Ph.D.

10

Soil Toxicology Laboratory, Biological Assessment and Standardization

11

Environment and Climate Change Canada

12

335 River Road

13

Ottawa (Ontario) K1V 1C7

14

Email:

[email protected]

15

Telephone:

613-990-9544

ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 23

2 16

Abstract

17

There is an increasing interest to use SP-ICP-MS to help quantify exposure to engineered

18

nanoparticles, and their transformation products, released into the environment. Hindering the

19

use of this analytical technique for environmental samples is the presence of high levels of

20

dissolved analyte which impedes resolution of the particle signal from the dissolved. While

21

sample dilution is often necessary to achieve the low analyte concentrations necessary for SP-

22

ICP-MS analysis, and to reduce the occurrence of matrix effects on the analyte signal, it is used

23

here to also reduce the dissolved signal relative to the particulate, while maintaining a matrix

24

chemistry that promotes particle stability. We propose a simple, systematic dilution series

25

approach where by the first dilution is used to quantify the dissolved analyte, the second is used

26

to optimize the particle signal, and the third is used as an analytical quality control. Using simple

27

suspensions of well characterized Au and Ag nanoparticles spiked with the dissolved analyte

28

form, as well as suspensions of complex environmental media (i.e., extracts from soils

29

previously contaminated with engineered silver nanoparticles), we show how this dilution series

30

technique improves resolution of the particle signal which in turn improves the accuracy of

31

particle counts, quantification of particulate mass and determination of particle size. The

32

technique proposed here is meant to offer a systematic and reproducible approach to the SP-ICP-

33

MS analysis of environmental samples and improve the quality and consistency of data

34

generated from this relatively new analytical tool.

35

36

INTRODUCTION

37

In order to evaluate the risk of engineered nanoparticles (ENPs) released into the environment, it

38

is necessary to perform quantitative assessments of both the exposure and effects to non-target


Page 3 of 23

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


3 39

organisms in environmental media.1,2 Single-particle inductively-coupled plasma mass

40

spectroscopy (SP-ICP-MS) has been identified as a valuable tool in this regard, given its ability

41

to detect and quantify the elemental composition of nanoparticles in suspensions at

42

environmentally relevant concentrations (i.e., ng L-1).3 However, SP-ICP-MS is not without its

43

shortcomings. One difficulty that arises in SP-ICP-MS, particularly for the analysis of

44

environmental media, occurs when dissolved analyte concentrations are large relative to the

45

analyte mass in the nanoparticles, which can lead to erroneous particle information.4

46

For SP-ICP-MS, samples must be sufficiently dilute such that, ideally, one particle at a time is

47

ionized in the plasma, which generates a burst of ions that enter the mass spectrometer and result

48

in a spike of the intensity signal, often referred to as pulse intensity. In contrast, the dissolved

49

analyte will produce a constant, continuous signal because of the steady stream of ions entering

50

the plasma. The intensity of a particle is essentially the pulse intensity after subtracting the

51

intensity of the dissolved or background signal, which can then be related to the mass of the

52

particle via calibration with well characterized nanoparticles, or via dissolved analyte calibration.

53

5

54

fraction of the analyte within the particle are known. Particle concentration (i.e., particles per

55

mL) is determined by counting the number of pulses detected and accounting for operational

56

factors that affect the rate of ions reaching the detector, such as sample flow rate and transport

57

efficiency (i.e., the fraction of nebulized particles that reach the plasma). An overview of SP-

58

ICP-MS theory and related calculations are provided in Laborda et al.,6 while an evaluation of

59

transport efficiency calculations is provided in Pace et al.5

Particle size is calculated from its mass if the analyte density, particle geometry and mass

60

One of the key aspects in processing SP-ICP-MS data is distinguishing the pulse intensity of

61

the particle from the dissolved signal. The intensity threshold used to discriminate the two



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 23

4 62

analyte forms is often determined using an iterative computation.4 An approximate threshold

63

limit is established, above which, events will be removed (e.g., events with intensities above the

64

mean intensity + 3δ). This initial calculation usually removes the majority of particle events. Of

65

the remaining events, the mean intensity is re-calculated and the procedure is repeated until

66

convergence (i.e., no more events can be removed). The final intensity threshold can be

67

converted to analyte mass, or particle volume, to determine the detection limits for particulate

68

mass or particle size, respectively. Some suppliers of SP-ICP-MS instruments are now supplying

69

software which performs these computations automatically for each sample.

70

A difficulty in the computation used to resolve particle intensity from the dissolved signal

71

occurs when there is a high level of dissolved analyte. When this occurs, the dissolved signal

72

overlaps the particle signal, resulting in the computation of a threshold intensity that is larger

73

than the intensities of some (or potentially all) the particles in a sample. For example, when an

74

aqueous suspension of Ag nanoparticles (60 nm nominal size) was spiked with increasing

75

concentrations of dissolved Ag, the number of particles detected decreased and the particle size

76

detection limit increased.4

77

There are many instances where environmental samples are expected to contain high levels of

78

the dissolved analyte. This may occur if the analyte of interest naturally occurs in the

79

environment in the form of dissolved ionic or complexed species (i.e., geologic or pedogenic in

80

origin), or if ENPs released into the environment undergo weathering and transformation to

81

dissolved species during their translocation to, or residence in, the media being tested.

82

Few researchers have developed reliable techniques to address the problem of analyzing

83

environmental samples with large dissolved analyte fractions. Hadioui and colleagues4

84

developed a technique whereby samples were first passed through an exchange column to


Page 5 of 23

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


5 85

remove much of the dissolved Ag+ species from samples prior to introduction into the ICP-MS.

86

This technique resulted in more accurate sizing and counting of the nanoparticles in their

87

suspensions and was further applied to the analysis of ZnO nanoparticles in wastewater.7 While

88

effective, the procedure adds a considerable amount of time to the analysis. This may be a

89

reasonable trade-off for suspensions containing NPs that are highly unstable and readily dissolve,

90

but may not be practical for all scenarios.

91

In this study, we investigate a sample dilution technique, which is ideally suited to the

92

analysis of suspensions that are relatively stable, such as soil and biosolids extracts. The

93

technique involves diluting samples in order to reduce the dissolved signal, thereby increasing

94

the ability to resolve the particle signal. While sample dilution for SP-ICP-MS is not a novel

95

concept, it is often done haphazardly from one sample to the next with the focus being to reduce

96

the occurrence of analytical coincidence (i.e., two or more particles entering the plasma at once).

97

Unlike standard ICP-MS analysis for which well-established technical guidance have been

98

published (e.g., EPA 6020A, 200.8), SP-ICP-MS is still in its infancy and lacks supportive

99

technical guidance and implementation of consistent quality controls, particularly for the analysis

100

of complex environmental matrices. We propose the use of a simple dilution series wherein each

101

successful dilution achieves a specific goal: (1) the capture of a dissolved analyte reading above

102

the limit of quantitation (LoQ) for the dissolved signal, (2) the capture of an optimized particle

103

signal, and (3) a quality control check that the particle analysis is truly optimized and free of

104

coincidence and matrix effects.

105

For this study, Au and Ag nanomaterials were selected. Citrate-stabilized Au nanoparticles

106

are highly stable and thus allowed us to explore the dilution technique without any interfering

107

effects of particle dissolution. Silver nanoparticles are one of the most widely used engineered



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 23

6 108

nanomaterials and their release from consumer products into the wastewater stream and

109

subsequent accumulation in sewage sludge represent a major translocation pathway.8 In North

110

America, processed sewage sludge (i.e., biosolids) is applied to agricultural lands representing an

111

important ecotoxicological exposure scenario. As such, governing agencies have initiated

112

assessments on the fate and effects of ENPs released into the terrestrial environment.9 The

113

overall goal of this study is to help further the development of a systematic, standardized

114

approach to SP-ICP-MS analysis of environmental samples (e.g., soil and biosolids) which is

115

expected to improve the quality of data generated by this relatively new analytical technique.

116 117

EXPERIMENTAL SECTION

118

Chemical. All Au and Ag nanoparticles used in this study were commercially purchased and

119

are listed in Table 1, along with suppliers’ specifications. Verification of these specifications was

120

carried out using dynamic light scattering (DLS), transmission electron microscopy (TEM) with

121

energy-dispersive X-ray spectroscopy (EDS), and SP-ICP-MS. All purchased suspensions were

122

placed in an ultrasonic bath (Branson 3510, 40 kHz, 130 watts, 6 L capacity) for at least 20 min

123

prior to analysis. In the case of the Ag nano-powder, a 15 g L-1 dispersion was prepared using

124

NanopureTM water (Barnstead 2-cartridge water filtration system, > 18.3 MΩ·cm, TOC < 10 µg

125

L-1) and placed in the ultrasonic bath for 40 min. The DLS analysis was conducted on ca. 20 mg

126

L-1 dilutions of initial suspensions using a MalvernTM Nano ZS Zetasizer with NIBS® optics

127

arranged at 175° for backscatter detection. Data was analyzed using cumulants analysis.10

128

Imaging of the same dispersion was conducted using an FEITM Tecnai G2 F20 TEM (FEI,

129

Hillsboro OR) at the Carleton University Nano Imaging Facility (Ottawa, Ontario). Details of the


Page 7 of 23

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


7 130

SP-ICP-MS analysis are provided in the relevant section below. The Au and Ag nanomaterials

131

are hereafter referred to by their nominal sizes.

132

Test Samples. Two types of samples were analyzed in this study: (1) suspensions of gold

133

nanoparticles (AuNP) or silver nanoparticles (AgNP) prepared in NanopureTM water, and (2) soil

134

extracts prepared from AgNP-contaminated soils. The nanoparticle suspensions were prepared

135

with 30 nm AuNP (NIST 8012), 60 nm AuNP (8013) or 30 nm AgNP (Ted Pella). For each

136

material used, a set of suspensions were prepared for which the particle concentration was kept

137

constant at approximately 105 particles mL-1, while increasing the dissolved Au or Ag

138

concentrations. Stock solutions of dissolved Au and Ag were prepared with an Au standard

139

solution (High-Purity Standards, 1000 µg mL-1 in 2% HCl) and silver nitrate (Sigma Aldrich

140

Reagent Plus®), respectively. Dissolved analyte concentrations ranged from 0.1 to 10 µg L-1 for

141

the suspensions containing the 30 nm particles, and 0.1 to 40 µg L-1 for suspensions containing

142

the 60 nm particles. Dilutions of test samples were always prepared fresh using NanopureTM

143

water and analyzed immediately. The soil extracts analyzed in this study were used to

144

demonstrate the dilution series approach to SP-ICP-MS analysis on real environmental samples

145

and were taken from a soil previously amended with AgNP-contaminated biosolids. Details of

146

the method development of the soil extraction technique are not critical to the understanding of

147

how to optimize the SP-ICP-MS analysis presented here in, and are outside the scope of this

148

study; however, further details have been included in a manuscript being prepared by the authors.

149

Soil preparation and the soil extraction procedures are briefly described as follows: An aliquot of

150

a 15 g L-1 suspension of the 40 nm AgNP (Nano-composix) was mixed into a slurry of municipal

151

biosolids, aged 3 d in sealed containers, then incorporated into an agricultural field soil (sandy

152

loam, pH 5.8) at a rate of 1 g biosolids per 100 g soil (dry weights) to obtain various Ag



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 23

8 153

concentrations. The soil samples used in this study included a soil sample spiked with 833 µg g-1

154

Ag taken from an earthworm assay (nAgEA), a soil sample spiked with 833 µg g-1 Ag taken

155

from a microbial assay (nAgMicr), and a soil sample spiked with 145 µg g-1 Ag taken from a

156

collembolan assay (nAgFC). Soils were extracted with tetrasodiumpyrophosphate (2.5 mM

157

TSPP) using a soil-to-reagent ratio of 1:10 (g:mL) and mixed on a rotary mixer for 30 min

158

followed by ultrasonication. Suspensions were allowed to settle overnight to allow large particles

159

to settle out. Sample suspensions were initially diluted 100x and all subsequent dilutions were

160

maintained with a 0.025 mM TSPP matrix.

161

SP-ICP-MS Instrumentation. SP-ICP-MS analysis was performed using a PerkinElmer

162

NexION 300D ICP-MS operating in single particle mode using the SyngistixTM Nano-

163

Application module (PerkinElmer Inc., Waltham, MA). Instrumental and data acquisition

164

parameters of the analysis are provided in Table 2. Time resolved signals were acquired using

165

fast-scanning mode, for which settling times have been reduced to negligible. Due to the fast

166

data acquisition rate, dwell times as low as 10 µs can be achieved. Previous research has shown

167

that dwell times of 50 to 100 µs are optimal for our instrumental set-up and for the analysis of

168

nanoparticles similar to ours.11 Transport efficiency was determined by the particle size method5

169

using the 60 nm NIST Au standard for reference; all calculations were performed within the

170

SyngistixTM module. Dissolved and particle calibrations were performed for

171

analytes, with dissolved calibrations ranging from 0.1 to 10 µg L-1. Dissolved Au and Ag

172

calibration solutions were prepared in 0.5% HCl and 1% HNO3 matrices, respectively. Gold

173

particle calibrations were performed using NIST Au 30 and 60 nm suspensions while for the Ag

174

particle calibrations, Ted Pella Ag 30 and 60 nm suspensions were used. For particle

175

calibrations, the intensity histograms were fitted to a log normal curve from which the most


197

Au and

107

Ag

Page 9 of 23

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


9 176

frequent area intensity was derived and related to particle mass (i.e., size). The dissolved and

177

particle calibrations produced comparable particle size data with mean particle intensities

178

derived from the dissolved calibration being on average 9% lower. The particle calibration

179

method was selected to process the results of this study, as the un-acidified matrix was more

180

consistent with that of the samples. SP-ICP-MS analysis requires that samples are sufficiently

181

dilute such that analytical coincidence is avoided. Based on our operating parameters, we

182

estimated that coincidence could be avoided if sample particle concentrations were ≤ 1.2 x 105

183

mL-1; therefore target particle concentrations for calibration suspensions were approximately 105

184

particles mL-1. The method detection limit (MDL) and LoQ for the dissolved analysis were

185

calculated as the mean reading of six method blanks (i.e., 0.025 mM TSPP solution) plus 3δ and

186

10δ, respectively.

187

Data Processing. Initial data processing was performed within the SyngistixTM Nano-

188

application module. Assuming a particle event in the plasma is approximately 350 µs,11 the 50 µs

189

dwell time used allows for multiple readings per particle. An algorithm is used to identify when a

190

succession of readings comprise a particle event from which peak area is calculated. In this way,

191

particle intensity is related to peak area. The determination of the intensity threshold used to

192

discriminate whether a signal is part of a particle event, or the dissolved (i.e., background) signal

193

is also performed within the SyngistixTM module using an iterative “mean + 3δ” computation,

194

similar to that described by Hadioui et al.4 The software records the number of particle events

195

and produces a frequency distribution based on particle intensity (i.e., peak area). Particle

196

intensity was related to particle mass (and thus size) using a particle calibration curve derived

197

from analyzing the 30 and 60 nm nanoparticle suspensions (Au or Ag) based on the assumption

198

that Au and Ag particles were spherical and pure Au0 or Ag0 with a density of 19.32 and 10.49 g



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 23

10 199

cm-3, respectively. Particle size distributions (PSD) were fit to a Gaussian smoothing function

200

from which mean and mode particle sizes were derived. It must be noted that while the particle

201

geometry and mass fraction may be known for well characterized particles (i.e., test samples

202

prepared from purchased suspensions), this is usually not the case for particles in environmental

203

media, for which a plasma event may include transformation products of various composition, or

204

naturally occurring particles. Therefore, for particles detected in environmental media, only the

205

mass of the analyte detected in a discrete particle (i.e., particulate Ag) is detectable, the size of

206

the actual particle is unknown. However, for the analysis of particulate Ag in our soil extracts,

207

we processed the data using a default spherical shape and mass fraction of 1, in order to calculate

208

an equivalent mono-metal particle size, denoted with the subscript “eqv” (e.g., PSDeqv), indicative,

209

not of actual particle sizes, but reference sizes used to aid in the understanding of the amount of

210

mass detected.

211

RESULTS AND DISCUSSION

212

Particle sizes of the Au and Ag nanomaterials were verified by DLS (Table S1, Figure S1,

213

Supporting Information) as well as by TEM/EDS and SP-ICP-MS (selected data, Figures S2 and

214

S3, Supporting Information). Measured particle sizes generally fell within the suppliers’ size

215

specifications, except for the Nanocomposix 40 nm Ag particles which showed a minor fraction

216

of particles smaller than 19 nm.

217

To investigate the effects of the SP-ICP-MS particle signal when samples contain large

218

concentrations of the dissolved metal, nanoparticle suspensions containing ca. 105 particles mL-1

219

were spiked with increasing amounts of dissolved metal and analyzed. Figure S4 (Supporting

220

Information) shows a portion of the time-resolved Au signal during analysis of the 30 nm


Page 11 of 23

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


11 221

suspensions with and without dissolved Au added. When analyzing the suspension spiked with

222

the dissolved Au, the steady dissolved signal drowns out most of the particle signals, resulting in

223

fewer particles detected of relatively large intensities, likely agglomerates or artifacts of

224

coincidence. Particle intensities were converted to particle size for which frequency histograms

225

were derived. Figure 1 shows the histograms generated from analyzing the 30 nm Au suspension

226

with increasing amounts of dissolved Au. Adding 0.5 µg L-1 dissolved Au had little effect on the

227

size distribution, while the addition of 1, 2 and 3 µg L-1 dissolved Au significantly lowered the

228

number of detectable particles while skewing particle sizes to increasingly larger sizes. This

229

trend is clearly illustrated with both 30 nm and 60 nm Au suspensions, in Figure 2, which shows

230

that at low concentrations of dissolved Au, the detected particle sizes and counts were accurate

231

and constant, but above a certain threshold, increasing concentrations of dissolved Au led to

232

erroneous increases to particle size and decreases in the number of particles detected. For the 30

233

nm Au suspensions, this threshold of dissolved Au is between 0.5 and 1.0 µg L-1, while for the

234

60 nm Au suspensions, it was between 10 and 20 µg L-1. When suspensions containing the

235

highest dissolved Au additions tested, i.e., 30 nm + 3 µg L-1 dissolved Au and 60 nm + 40 µg L-1

236

dissolved Au, were diluted such that the dissolved concentrations were below these thresholds,

237

particle size and particle counts were measured more accurately (Figure 2). The reason for this

238

improvement is demonstrated by comparing the particle histograms of the 30 nm + 3 µg L-1

239

dissolved Au sample before, and after dilution (Figures 1e and 1f). For the diluted sample, the

240

dissolved signal was sufficiently reduced to allow for the detection of the particles, thereby

241

resolving the sample’s particle size histogram. A similar trend was observed for suspensions

242

made from the 30 nm Ag nanoparticles for which a dissolved Ag threshold between 0.5 and 1.0

243

µg L-1 was also observed. (Figure S5, Supporting Information). These data show that the



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 23

12 244

dissolved analyte had a similar effect in magnitude on particle data for these similarly-sized

245

particles, despite being different elements.

246

When analyzing environmental samples, the particle sizes as well as particulate and dissolved

247

concentrations are generally unknown. Because the dissolved analyte fraction is an important

248

parameter for environmental assessments, we began with obtaining dissolved analyte readings

249

within the calibration range and above the LoQ of 0.1 μg L-1, diluting the samples if necessary.

250

For the three environmental samples analyzed in this study, only the nAgMicr extract required an

251

initial dilution (i.e., 100x) to achieve a valid dissolved reading (Table 3). With knowledge of the

252

dissolved fraction, the next analysis involved diluting the sample to achieve a dissolved analyte

253

concentration near the MDL of 0.05 μg L-1, thereby maximizing the ratio of particle signal /

254

background (dissolved) signal. This “optimal” dilution also needed to provide a particle

255

concentration below the (coincidence) criteria of 105 particles mL-1. When applying this dilution

256

technique to the analysis of soil extracts taken from ecotoxicological assays, a third dilution that

257

was 2x the optimal dilution proved advantageous as a quality control. If the dilution selected for

258

the “optimal” analysis was truly optimal, then diluting it 2x should yield a similar particle

259

intensity (i.e., size) with half the number of particles counted. For example, analysis of the soil

260

extract from the earthworm assay (nAgEA) showed 5.05 µg L-1 dissolved Ag and generated an

261

equivalent particle size distribution (PSDeqv) with a mode of 68.2 nm (Table 3). The optimal

262

dilution for the analysis of this sample was 100x, which reduced the dissolved signal to near the

263

MDL and provided a particle count < 105 particles mL-1, resulting in a PSDeqv with a mode of

264

37.4 nm (Figure 3a). When this optimal dilution was further diluted 2x, the resulting PSDeqv

265

showed a similar mode, but approximately half the number of particles, as expected, thus


Page 13 of 23

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


13 266

providing confidence that the analysis performed on the “optimal” dilution selected was indeed

267

not significantly impacted by excess dissolved analyte, analytical coincidence, nor matrix effects.

268

For the analysis of the soil extract from the microbial assay, the initial 100x dilution captured

269

a dissolved reading of 0.48 µg L-1 (Table 3). While we would not expect this dissolved

270

concentration to impact the particle signal (see Figure S5, Supporting Information), the particle

271

concentration detected at this dilution exceeded 105 particles mL-1, thus resulting in a high

272

probability for analytical coincidence. An optimal dilution of 2500x was determined which

273

shifted the PSDeqv left (Figure 3), resulting in the smaller mode of 32.7 nm and a greater overall

274

particle concentration, evidence that analytical coincidence likely affected the analysis of the

275

100x dilution. A further 2x dilution reduced the number of detectable particles to ca. 700 (data

276

not shown). However, we have found that when the event sample size drops much below

277

approximately 1000 events, there is greater variability among SP-ICP-MS data from analytical

278

replicates. Extending the sample time to 120 s allowed for a larger sample size of ca. 1400

279

events, which allowed for more reproducible data, and yielded a PSDeqv which generally

280

overlapped that from the optimal dilution, as would be expected given the operational

281

parameters. This technique of increasing sample time to increase the number of particle events

282

can be particularly useful when analyzing a sample with a large ratio of dissolved analyte /

283

particle counts, for which the technique of diluting the sample to reduce the dissolved signal

284

yields a sample with a very low particle count.

285

Similar to the previous samples, dilution of nAgFC shifted the initial PSDeqv left (Figure 3c),

286

resulting in a lower mode of 40.6 nm. However, a further 2x dilution (with the extended sample

287

time of 120 s), revealed a bimodal PSDeqv. This sample was re-run several times to verify this

288

bimodal pattern and data from duplicate analyses were combined. Given that over 30% more



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 23

14 289

particulate Ag mass was detected from the analysis of the most dilute sample compared to the

290

next more concentrated sample (Table 3), it is likely that the detection of the bimodal distribution

291

was due to an improved signal optimization (i.e., lower particle DL), rather than from dissolution

292

of the 40 nm particles. Each of the two distributions was fit to a log-normal curve separately to

293

determine their equivalent mode particle sizes which were 28.7 and 43.4 nm. While other

294

analytical techniques are required to further identify/characterize particles in these two distinct

295

size classes (e.g., Ag2S, AgNP with surface precipitate), the example here illustrates the

296

usefulness of checking the data acquired from what was believed to be the “optimal” dilution by

297

analyzing the sample with a further 2x dilution.

298

The most obvious drawback to analyzing environmental samples using a dilution series is the

299

potential for particle dissolution. While research has shown that AgNPs in distilled water likely

300

undergo fairly rapid dissolution in the short-term (i.e., < 24 h),12 AgNPs contained in more

301

complex matrices, particularly in the presence of natural organic matter, are more stable,13,14 with

302

humic substances likely providing both steric and charge stabilization. Mitrano et al.14 showed

303

that the presence of ligands, such as chloride and sulfide, also had a stabilizing effect on AgNPs,

304

and suggested the formation of relatively insoluble halide-Ag or sulfide-Ag precipitates on the

305

particle surface. Thus environmental samples containing ENPs, and/or their transformation

306

products may very well be stabilized by their solution chemistry enough to carry out dilution and

307

analysis in the very short-term (i.e., minutes) without significant changes to analyte mass of the

308

particulates. It would be useful to perform dilution-stability tests on samples of various complex

309

matrices to gain information regarding the rate of change that dilution may have on particle

310

stability. The dilution approach shown here may be more suited for soil extracts or wastewaters

311

which generally have a large humic component. In the case of soil analysis, all extractions (e.g.,


Page 15 of 23

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


15 312

soil pore water) and preparatory procedures are by necessity operationally defined, thus, it would

313

be advantageous to utilize an extraction technique that is amenable to SP-ICP-MS analysis. We

314

have found that the alkaline and dispersing attributes of TSPP result in large particle/dissolved

315

concentration ratios and stable extracts. For some environmental samples, it may be possible to

316

inhibit, or retard particle dissolution by maintaining the ionic strength of the environmental

317

sample during the dilution, (i.e., using a dilute Na solution) and/or adding a dispersant or

318

surfactant which may offer some steric stability. In our case, we maintained the TSPP

319

concentration of 0.025 mM in all dilutions.

320

Conclusions

321

For SP-ICP-MS analysis of environmental samples affected by engineered nanoparticle

322

contamination, quantification of both the dissolved and nano-particulate forms of the metal

323

analyte are of interest; however, it is doubtful that one dilution/analysis can optimally capture

324

both signals. As we have shown, a large dissolved signal can drown-out the particle signal

325

leading to erroneous sizing and particle counts. The degree to which an analysis will be

326

negatively impacted by presence of the dissolved analyte depends on both the particle size and

327

dissolved analyte concentration. The approach demonstrated here relies on analyzing samples in

328

a dilution series such that (1) an analysis captures the dissolved signal within the calibrated range

329

and above the LoQ, (2) an analysis captures an optimal particle signal, and (3) a third analysis is

330

used to verify the optimal particle data. An optimal particle signal can be achieved by diluting

331

the sample such that the dissolved signal is at or below the MDL and the particle concentration is

332

< 105 particles mL-1, to minimize the probability of analytical coincidence. Sampling time may

333

be extended to ensure sufficient number of events (i.e., > 1000) to achieve a suitable level of

334

statistical confidence when generating the particle frequency histograms and for ensuring



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 16 of 23

16 335

sufficient reproducibility. For the verification analysis, a 2x dilution of the optimal sample

336

dilution is recommended. The verification analysis ensures that interferences from the dissolved

337

signal, sample matrix effects and analytical coincidence have been sufficiently minimized. While

338

humic substances contained in environmental samples are likely to help stabilize particles and

339

buffer any rapid changes to particles that may be caused by the dilution process, further research

340

on how to preserve and stabilize environmental samples for SP-ICP-MS analysis is required. The

341

dilution technique demonstrated here offers a systematic and reproducible approach to the

342

analysis of environmental samples which is anticipated to improve the quality of data generated

343

by this relatively new analytical tool.

344 345 346 347

Associated Content

348

Additional information as noted in the text.

349

Acknowledgements

350

The research was funded by Environment and Climate Change Canada through the Chemicals

351

Management Plan.

352

References

353 354 355 356 357 358 359 360 361 362 363

(1) von der Kammer, F.; Ferguson, P. L.; Holden, P. A.; Masion, A.; Rogers, K. R.; Klaine, S. J.; Koelmans, A. A.; Horne, N.; Unrine, J. M. Environ. Toxicol. Chem. 2012, 31, 32-49. (2) Montano, M. D.; Ranville, J. F.; Lowry, G. V.; Lowry, G. V.; Von Der Kammer, F.; Blue, J. Environ. Chem. 2014, 11, 351-366. (3) Hassellöv, M.; Readman, J. W.; Ranville, J. F.; Tiede, K. Ecotoxicol. 2008, 17, 344-361. (4) Hadioui, M.; Peyrot, C.; Wilkinson, K. J. Anal. Chem. 2014, 86, 4668-4674. (5) Pace, H. E.; Rogers, N. J.; Jarolimek, C.; Coleman, V. A.; Higgins, C. P.; Ranville, J. F. Anal. Chem. 2011, 83, 9361-9369. (6) Laborda, F.; Jimâenez-Lamana, J.; Bolea, E.; Castillo, J. R. J. Anal. At. Spectrom. 2011, 26, 1362-1371. (7) Hadioui, M.; Merdzan, V.; Wilkinson, K. J. Environ. Sci. Technol. 2015, 49, 6141-6148.

Supporting Information


Page 17 of 23

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


17 364 365 366 367 368 369 370 371 372 373 374 375 376 377

(8) Gottschalk, F.; Nowack, B. JEM 2011, 13, 1145-1155. (9) Regulatory Cooperation Council (RCC) Nanotechnology Initiative, Final Report on Work Element 2, Priority Setting: Development of a joint nanomaterial classification scheme.: Available at: Government of Canada NanoPortal (http://nanoportal.gc.ca), Accessed May 28, 2012. (10) Koppel, D. E. J. Chem. Phys. 1972, 57, 4814. (11) Hineman, A.; Stephan, C. J. Anal. At. Spectrom. 2014, 29, 1252. (12) Liu, J.; Hurt, R. H. Environ. Sci. Technol. 2010, 44, 2169-2175. (13) Cumberland, S. A.; Lead, J. R. J.Chrom. A 2009, 1216, 9099-9105. (14) Mitrano, D. M.; Ranville, J. F.; Bednar, A.; Kazor, K.; Hering, A. S.; Higgins, C. P. Environ. Sci.: Nano 2014, 1.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 18 of 23

18 378 379 380 381

382 383 384 385 386 387

TABLES & GRAPHS Table 1. Supplier Information for the Nanomaterials Used Metal

Nominal size

Supplier

Product Name

Surface Coating

Form

Supplier’s size specification

Au

30 nm

NIST*

Citrate

60 nm

NIST

Ag

30 nm

Ted Pella

Ag

60 nm

Ted Pella

Ag

40 nm

Nano-composix

Aqueous dispersion Aqueous dispersion Aqueous dispersion Aqueous dispersion Powder

27.6 ± 2.1 nm

Au

Reference Material 8012 Reference Material 8013 PELCO NanoXact Silver PELCO NanoXact Silver Redispersible silver nanospheres

Citrate Citrate Citrate PVP, 88% by mass

* National Institute of Standards and Technology

Table 2. SP-ICP-MS Instrumental and Analytical Details Parameter Group ICP-MS component

Parameter

Parameter Value

ICP-MS model

Spray chamber

Perkin Elmer NexION 300D Low pressure PFA nebulizer Cyclonic, glass

Neb gas flow

1.04 L/min

Plasma RF power

1600 W

Sample flow rate

0.270 mL/min

Analyte monitored

107

Sample introduction Transport efficiency Instrument dwell time Sample analysis time Readings per sample

Manual, with a 60 s rinse and 20 s read delay 9.2% to 9.8%

Nebulizer

ICP-MS settings

Single Particle Parameter

Ag or 197Au

50 µs 60 s (or more) ~ 1.2 x 106 (or more)

388 389


56.0 ± 0.5 nm 32.7 ± 4.8 nm 62 ± 7 nm 38.6 ± 9.8 nm

Page 19 of 23

19 390 350

(a)

300

Frequency

Frequency

250 200 150 100 50 0 24

391

37

44

49

54

35

58

400 350 300 250 200 150 100 50 0

(b)

24

61

37

44

49

54

10

(c)

30

58

(d)

8

Frequency

Frequency

25 20 15 10 5

4 2

24 37 44 49 54 58 61 64 67

24 37 44 49 54 58 61 64 67 69

392

6

0

0

3

140

(e)

(f)

120 100

2

Frequency

Frequency

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


1

80 60 40 20

0

0 24 37 44 49 54 58 61 64 67 69 72 74 76

393 394 395 396 397

Diameter (nm)

24

37

44

49

54

58

Diameter (nm)

Figure 1. Particle size histograms for the 30 nm Au particle suspension with (a) no dissolved Au added, (b) 0.5 µg L-1 dissolved Au, (c) 1 µg L-1 dissolved Au, (d) 2 µg L-1 dissolved Au, (e) 3 µg L-1 dissolved Au, and (f) 3 µg L-1 dissolved Au following 5x dilution and extending sample analysis time to 120 s.



20 30 nm 30 nm (5x dilute)

90

60 nm 60 nm (10x dilute)

mean size (nm)

80 70 60 50 40 30 20 1.0

2.0

3.0

4.0

5.0

log [dissolved Au] (ng L-1)

398 1000000 100000

particles mL-1

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 20 of 23

10000 1000 100 1.0

399 400 401 402 403 404 405 406 407 408 409

2.0

3.0

4.0

5.0

log [dissolved Au] (ng L-1)

Figure 2. Mean particle size (nm) and number of particles (mL-1) measured in 30 nm and 60 nm Au particle suspensions (nominal particle concentrations = 105 particles mL-1) spiked with increasing concentrations of dissolved Au, shown as log concentration (ng L-1) along the x-axis. Included are the results for the 30 nm + 3 µg L-1 dissolved Au sample which was diluted 5x (×), and results for the 60 nm + 40 µg L-1 dissolved Au sample which was diluted 10x (+). Note: The dilution factors have been applied to the particle counts so that values shown represent the final particle concentrations for the original (undiluted) samples.


Page 21 of 23

21 nAgEA (1x) nAgEA (50x) nAgEA (100x)

60 50

Frequency

40 30 20 10 0 0

20

40

60

80

100

120

140

410 nAgMicro (100x) nAgMicro (2500x) nAgMicro (5000x)

140 120

Frequency

100 80 60 40 20 0 0

20

40

60

80

100

120

411 nAgFC (1x) nAgFC (100x) nAgFC (200x) - 1st curve nAgFC (200x) - 2nd curve

60 50 40

Frequency

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


30 20 10 0 0

412 413 414 415 416 417 418 419

20

40

60

80

100

120

140

Diameter (nm)

Figure 3. Equivalent particle size distributions (PSDeqv) for particulate Ag detected in soil extracts using a dilution series technique. Soil samples were taken from: an earthworm assay (nAgEA), a microbial assay (nAgMicro), and a collembolan assay (nAgFC). Extract dilutions are shown in brackets. Note: analysis time for nAgMicro (5000x) and nAgFC (200x) was extended from 60 s to 120 s.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 22 of 23

22 420 421 422 423

Table 3. Analytical information and SP-ICP-MS results for soil extracts analyzed for 107Ag using a dilution series. Sample time for analyses was 60 s, except for nAgMicr (5000x) and nAgFC (200x) which were analyzed for 120 s. Note: DF = dilution factor; part. = particle

Sample

Dissolv ed Ag (µg L-1)

Mode sizeeqv (nm)

# peaks detecte d

Final* part. counts log [part.] (mL-1)

1

5.05

68.2

8763

5.5

1.03

100

0.06

37.4

2542

6.9

6.77

200

< 0.05

36.6

1227

6.9

6.08

100

0.48

39.8

13055

7.7

41.5

2500

Single Particle-Inductively Coupled Plasma Mass ... - ACS Publications

Recommend Documents