Qsae024 METODOS

Page 1 of 33 The Journal of AOAC INTERNATIONAL
1
2
3 1 Validation of a method for surveillance of nanoparticles in mussels using single particle inductively
4
5
2 coupled plasma mass spectrometry
6
7
8 3
9
10
11 4 Are Bruvold*a,b, Stig Valdersnesa,b, Katrin Loeschnerc, André Marcel Bienfaita
12
Downloaded from https://academic.oup.com/jaoac/advance-article/doi/10.1093/jaoacint/qsae024/7632740 by guest on 26 March 2024

13
14 5 *E-mail: arebru@gmail.com
15
16
17 6 a Institute of Marine Research (IMR), P.O. Box 1870 Nordnes, N-5817 Bergen, Norway
18
19
20 7 b Department of Chemistry, University of Bergen, P.O. Box 7803, N-5020 Bergen, Norway
21
F
22
O
23 8 c National Food Institute, Technical University of Denmark, Kemitorvet 201, DK-2800 Kgs.
24
O
25
9 Lyngby, Denmark
26
27
PR
28 10
29
D
30
TE
31
32
11 Abstract
EC
33
34
35 12 Background: Determining the concentration of nanoparticles in marine organisms is
R
36
37
R
13 important for evaluating their environmental impact and to assess potential food safety risks
38
O
39
40 14 to human health.
C
41
N
42
15 Objective: The current work aimed at developing an in-house method based on single
U
43
44
45 16 particle inductively coupled plasma mass spectrometry suitable for surveillance of
46
47
48 17 nanoparticles in mussels.
49
50
51 18 Method: A new low-cost and simple protease mixture was utilized for sample digestion, and
52
53 19 a novel open-source data processing was used, establishing detection limits on a statistical
54
55
56 20 basis using false positive and false negative probabilities. The method was validated for 30
57
58 21 and 60 nm gold nanoparticles spiked to mussels as a proxy for seafood.
59
60
1
© The Author(s) 2024. Published by Oxford University Press on behalf of AOAC INTERNATIONAL.
This is an Open Access article distributed under the terms of the Creative Commons Attribution License
(http://creativecommons.org/licenses/by/4.0/), which permits unrestricted reuse, distribution, and reproduction in
any medium, provided the original work is properly cited.
The Journal of AOAC INTERNATIONAL Page 2 of 33
1
2
3 22 Results: Recoveries were 76-77% for particle mass concentration and 94-101 % for particle
4
5
6 23 number concentration. Intermediate precision was 8-9% for particle mass concentration and
7
8 24 7-8% for particle number concentration. Detection limits for size was 18 nm and for
9
10
11 25 concentration 1.7 ng/g and 4.2 x 105 particles/g mussel tissue.
12

13
14 26 Conclusion: The performance characteristics of the method were satisfying compared with
15
16 27 numeric Codex criteria. Further, the method was applied to titanium-, chromium- and
17
18
19 28 copper-based particles in mussels.
20
21
F
22 29 Highlights: The method demonstrates a new practical and cost-effective sample treatment
O
23
24 30 and streamlined, transparent and reproducible data treatment for the routine surveillance
O
25
26
27 31 of NPs in mussels. PR
28
29
D
30 32 Keywords: nanoparticles, SP-ICP-MS, seafood, mussels, validation

TE
31
32
33 Introduction
EC
33
34
35
R
36 34 Whereas organisms have evolved to coexist with natural nanoparticles (NPs) in the environment, the
37
R
38
35 effects of exposure to anthropogenic incidental and engineered NPs are unknown (1). Identifying the
O
39
40
C
41 36 number percentage of particles in the nanoscale has become a point of interest due to the unique
N
42
U
43 37 properties of compounds in the nanoscale. This topic has only been studied rigorously the last two
44
45 38 decades (2), scientists just recently having the tools to study and distinguish natural and
46
47
39 anthropogenic NPs (3). For this reason, few studies have been performed on marine organisms (1)
48
49
50 40 and coastal waters (4), less being known about NPs in the marine environment than any other
51
52 41 principal earth compartment (2). This despite the ocean being a sink for contaminants and subject to
53
54 42 e.g. disposal of mining waste and deep sea mining containing NPs. Blue mussels are widely used
55
56 43 bioindicators for monitoring anthropogenic pollution trends (5), capturing aggregated NPs in the
57
58
59
60
2
ScholarOne Support phone: 434-964-4100 email: ts.mcsupport@thomson.com
1
2
3 44 marine environment (6) and having lower biotransformation rate then e.g. fish (5). They are further a
4
5
45 popular food source, the European market estimated at near 600 000 tons per year (7).
6
7
8 46 Detecting and quantifying NPs in complex biological matrices remains a challenge. However, single
9
10
11 47 particle inductively coupled plasma mass spectrometry (SP-ICP-MS) has shown to be an invaluable
12

13 48 tool holding further promise given instrumental improvements (8). Sample preparation for matrix
14
15 49 degradation is generally required using alkaline, acidic or enzymatic digestion prior to analysis (9).
16
17 50 Stability of NPs and matrix effects may influence results requiring careful sample treatment and
18
19
20
51 method development (8). NPs exhibit an extrinsic or media-dependent nature (10), demanding extra
21
F
22 52 efforts toward standardization and validation (11). A number of extraction protocols have been
O
23
24 53 employed (9), enzymatic (mainly Proteinase K) and alkaline digestion (mainly tetramethylammonium
O
25
26
27
54 PR
hydroxide) generally showing the greatest promise for biological matrices (12). Enzymatic digestion
28
55 has been most extensively researched (13), while alkaline digestion has recently gained popularity
29
D
30
TE
31 56 due to its wide applicability (12) and has been found preferential for some applications (14), though
32
57 inferior for others (15). However, alkaline digestion may be problematic for certain NPs such as Ag
EC
33
34
35 58 (16) due to the potential formation of NPs from ionic forms of the element. Further,
R
36
37
R
59 tetramethylammonium hydroxide is avoided in many laboratories due to its hazards. Protamex®

38
O
39
40 60 (protease from Bacillus sp.) has previously been applied for the digestion of Atlantic salmon and
C
41
N
42 61 yellowfin tuna (17, 18). Protamex is active at mild conditions with neutral pH and only slightly
U
43
44 62 elevated temperatures, offering a quick, simple, robust and safe alternative for matrix digestion, at a
45
46 63 cost up to orders of magnitudes lower than for Proteinase K (19). This reduces the chances of the
47
48
49 64 nanoparticles undergoing transformations such as dissolution or agglomeration, and additionally
50
51 65 provides applicability and practicability for routine use.
52
53
54 66
55
56
57 67 To collect data, conduct surveillance and official controls it is a requirement to use validated
58
59 68 methods with proven performance fit for the intended purpose (20). The Food and Agriculture
60
3
1
2
3 69 Organization of the United Nations has stated that one of the most urgent challenges identified in
4
5
70 relation to enforcing a regulatory framework is the lack of routine methods for NPs in food (21).
6
7
8 71 Currently, relatively few NP types have been investigated and there are inconsistencies both in
9
10 72 reported metrics and their determination (12). Furthermore, signal processing is often performed
11
12 73 with untransparent commercial algorithms and inaccurate statistical models. Few validated methods

13
14 74 have been published for NPs using ICP-MS (TiO2 in various matrices (22, 23)) or SP-ICP-MS (Au and Ag
15
16
17 75 in mammalian tissue and blood (24, 25), Ag in chicken meat (26, 27), Ag in confectionary (28) and
18
19 76 TiO2 in crab sticks (29). To the best of our knowledge, no validated method exists for quantifying NPs
20
21 77 in blue mussels (Mytilus edulis). A few in-house validated methods for other types of edible mussels
F
22
O
23 78 exist. However, these generally do not assess intermediate precision (day-to-day variability) and
24
O
25
79 overall uncertainty of the method. As no certified reference materials for NPs in food or biological
26
27
PR
28 80 tissue exist (12), spiking with NPs of known properties, ideally of same type as the analyte, is
29
D
30 81 required (30). While not generally considered a contaminant or a food safety issue, gold has been
TE
31
32 82 used as a model nanomaterial (31). Gold NPs are ideal for method development and validation
EC
33
34
35
83 studies to be used for spiking since they are available with well-defined monodisperse size
R
36
37 84 distributions. Further, gold has low environmental levels, high stability and sensitivity, and no
R
38
O
39 85 abundant isobaric or polyatomic interferences using ICP-MS (32).

40
C
41
N
42 86 The objective of this study was to develop and validate a method for surveillance of nanoparticles in
U
43
44 87 blue mussels. The method was required to allow the cost-efficient analysis of larger sample sets for
45
46 88 observing relative changes in particle concentrations of at least an order of magnitude and over
47
48
49 89 longer time periods in relation to potential environmental and food safety issues, e.g. arising in
50
51 90 connection to the initiation of mining activities and dumping of mineral containing mining waste.
52
53 91 Reproducibility and control of false positives and negative rates were ensured by employing an open-
54
55 92 source signal processing developed in-house. By evaluating the method performance for particle
56
57
58 93 diameter, particle number, and mass concentrations versus Codex Alimentarius general and numeric
59
60 94 method performance criteria, it can be established whether the method is fit for the intended
4
1
2
3 95 purpose (33). As a starting point for assessing the method performance, blue mussels were spiked
4
5
96 with gold NPs of two sizes (30 and 60 nm) and analyzed on five different days. We assessed method
6
7
8 97 quality parameters such as selectivity, limit of detection, working range, trueness, precision, and
9
10 98 robustness for gold. The method was further applied to other types of NPs by analyzing three
11
12 99 samples from different locations for the presence of NPs containing titanium, chromium and copper.

13
14 100 The analysis was repeated on two days, providing novel precision data on NPs in seafood. The
15
16
17 101 method is intended to serve as a basis for subsequent method development toward surveillance of
18
19 102 different NPs in marine samples.
20
21
F
22 103 Materials and methods
O
23
24
O
25 104 Sample preparation
26
27
PR
28 105 Samples of blue mussel (Mytilus edulis) were obtained from the regular program for monitoring of
29
D
30
106 seafood at the Institute of Marine Research (IMR, Bergen/Norway) (34). An aggregate sample of 25
TE
31
32
107 random blue mussels was prepared by following IMR’s internal standard procedure for sample
EC
33
34
35 108 preparation of mussels for trace metals. First, the mussels were removed from the freezer and
R
36
37 109 allowed to thaw overnight. The next day, the mussels were opened by cutting the sphincter muscle
R
38
O
39 110 with a knife and the mussels were left standing upright for five minutes to allow any water inside the
40
C
41
N
42 111 mussel to drain. The tissue inside was then removed with a blunt knife and transferred to a sieve for
U
43
44 112 further collective drainage of water for five minutes following rinsing with UPW. The aggregate
45
46 113 sample was subsequently homogenized with a food processor and further homogenized by a high-
47
48 114 speed homogenizer (Polytron Pt-2100, Kinematica AG, Switzerland) to increase homogeneity.
49
50
51 115 Following homogenization, the test sample was transferred to polypropylene (PP) cups with screw
52
53 116 caps for storage in the freezer until further sample preparation and analysis. Prior to analysis, a
54
55 117 container of homogenized mussel sample was allowed to thaw overnight at 4 °C. For the method
56
57 118 validation, blue mussel tissue test portion (1 g) was weighed into a polypropylene (PP) tube. Method
58
59
119 blanks (without mussel tissue) were instead added 1 mL of UPW. For test portions spiked prior to
60
5
1
2
3 120 digestion, 100 µL of a gold NP dispersion with nominal concentration of 500 µg/L of 30 or 60 nm size
4
5
121 was added the mussel tissue or UPW blank, resulting in a nominal concentration of 50 ng/g
6
7
8 122 particles/mussel tissue. Unspiked matrix blanks (only digested mussel tissue, no gold NP) were added
9
10 123 100 µL UPW instead. For the test portions to be spiked after digestion, nothing was added. Then 3
11
12 124 mL of the enzyme solution was added to each of the tubes prior to the PP tubes being placed

13
14 125 horizontally on an incubator heat-shaker (Heidolph, Schwabach, Germany; Unimax @ 300rpm, 50 °C)
15
16
17 126 for 1 hour. Matrix blank test solutions were then split, and one parallel spiked appropriately followed
18
19 127 by dilution with UPW to the final concentrations corresponding to 500 times dilution of the mussel
20
21 128 tissue. For every dilution step, tubes were thoroughly mixed with a vortex-shaker (MS1 Minishaker,
F
22
O
23 129 IKA, Germany). For the method applicability demonstration, the enzyme solution was diluted by an
24
O
25
130 extra factor of five and allowed to incubate overnight. In addition, UPW added to the method blanks
26
27
PR
28 131 in place of mussel tissue was homogenized similarly to the mussel tissue. To properly simulate the
29
D
30 132 mechanical abrasion of the mussel tissue, blanks containing homogenized polyethylene beads were
TE
31
32 133 also investigated (supplementary figure 2). Another 12-fold dilution was performed for the method
EC
33
34
35
134 application on environmental samples to decrease the background, resulting in a total dilution factor
R
36
37 135 of 6000.
R
38
O
39
40 136
C
41
N
42
137 Measurement standards and NPs
U
43
44
45
138 Ionic standard for gold was purchased from Spectrascan, Ski, NO (SS-1118N) and used for calibrating
46
47
48 139 the ICP-MS and for determination of transport efficiency. Spherical gold NPs of 60 and 30 nm
49
50 140 nominal diameters with nominal concentrations of 50 mg/L were bought from Perkin Elmer,
51
52 141 Waltham, MA (N8142303, lot #E2840N and N8142300, lot #SPD543N, respectively) and used as both
53
54
142 reference materials for determination of transport efficiency and for the spiking experiment. Ionic
55
56
57 143 standards were stored at room temperature and NP standards were stored in the refrigerator at 4 °C
58
59 144 for approximately 18 months. The reference samples for the ACEnano proficiency test were used to
60
6
1
2
3 145 determine trueness as z-score for Au NPs in UPW (RIKILT test 2018-02) (35). For the method
4
5
146 demonstration on NPs of titanium, chromium and copper, ICP standards of ionic titanium from
6
7
8 147 Spectrascan (SS-1164), chromium from Fluka Analytical (68131) and copper from Sigma-Aldrich
9
10 148 (68921) were used. NanoXact 60 nm Gold 50 mg/L, lot TJC0086 were used for calibration, and TiO2
11
12 149 NPs with a primary particle size 115 nm from the Joint Research Centre ́s Nanomaterial Repository

13
14 150 (JRCNM10200a, ID: 010118) and copper(II) oxide nanopowder from Sigma-Aldrich (544868) were
15
16
17 151 used to estimate peak widths or recovery.
18
19
20
152 Reagents
21
F
22
153 Ultrapure water (UPW) with resistivity of 18.2 MΩ·cm at 25 °C (Elix Progard TNP and Milli-Q
O
23
24
O
25 154 Advantage A10, Merck Millipore, MA, USA) was used for aqueous dilutions. Nitric acid (65% EMSURE
26
27 155
PR
for analysis, Merck) was used for standards and rinsing. Protamex® (Merck), protease from Bacillius
28
29
D
156 sp. was purchased from Sigma Aldrich, Saint Louise, MO. A 200 g/L enzyme solution was prepared in
30
TE
31 157 UPW. Ionic and gold NPs used as standards and for spiking were prepared by dilution of the stock
32
EC
33
34 158 suspension in UPW. Final concentrations for gold NPs were 100 ng/L and 150 ng/L for reference
35
R
36 159 materials and 500 µg/L for spiking. Ionic gold standards were prepared to concentrations of 0.5, 1
37
R
38 160 and 5 µg/L. Ionic standards for titanium, chromium and copper were prepared in 0.1% HNO3 to the
O
39
40
C
161 same concentrations as gold.

41
N
42
U
43 162 Data acquisition, processing and visualization

44
45
46 163 Aliquots were analyzed using an Agilent 7900 ICP-MS fitted with a peristaltic pump and SPS-4
47
48 164 autosampler (Agilent Technologies, California, USA). Acquisition times of 2 and 3 minutes were used
49
50
51 165 for the validation and method applicability demonstration, with nominal injection rate of 0.346 mL
52
53 166 per minute. Additional instrumental parameters are given in supplementary tables 6 and 7.
54
55
56 167 All data processing, calculations and visualization was carried out in R version 4.2.1, whereas
57
58 168 graphical elements were generated or edited using Affinity Designer 1.10.5, Vectormagic by Cedar
59
60 169 Lake Ventures and DALL-E 2 by OpenAI. For discriminating particle signals from background noise, a
7
1
2
3 170 maximum peak intensity threshold was determined. This threshold was calculated based on the
4
5
171 assumption that the background noise followed a Poisson distribution, with a constraint of a 95%
6
7
8 172 probability of observing no more than one false positive per minute. This was done using the inverse
9
10 173 Poisson cumulative density function in R using the local filtered rolling mean as the Poisson mean
11
12 174 and a rate parameter of 1 - 0.05/600000.Peaks above this threshold were integrated, subtracting the

13
14 175 background corresponding to a rolling median with a correction for low count backgrounds. For the
15
16
17 176 multielement comparison, the lowest mass-normalized intensity threshold was used for all
18
19 177 measurements to ensure similar false negative rates. Transport efficiency was calculated in
20
21 178 accordance with the particle size method (36). More specifically, a kernel density mode of the signal
F
22
O
23 179 distribution resulting from the 60 nm Au NP standards were utilized to establish the signal per
24
O
25
180 analyte mass. The mass flow rate at the detector was calculated from ionic standards and used in
26
27
PR
28 181 combination with the intake rate to establish the transport efficiency. Particle diameters were
29
D
30 182 calculated assuming spherical shape and composition of pure gold with density 19.32 g/cm3, TiO2
TE
31
32 183 with density 4.17 g/cm3, CuO with density of 6.31 g/cm3 and Cr2O3 with density of 5.22 g/cm3. Code
EC
33
34
35
184 and raw data is available through an online repository (37).
R
36
37
R
185
38
O
39
40
C
186 Validation setup and calculation of validation results

41
N
42
U
43 187 Validation was carried out based on general in-house method validation guidelines from Eurachem
44
45
188 and NMKL (38, 39). Numeric values obtained for different validation parameters were compared to
46
47
48 189 Codex Alimentarius numeric method criteria (33). The method parameters evaluated were particle
49
50 190 mean diameter, particle mass concentration and particle number concentration. The daily validation
51
52 191 sequence was set up with ionic standards at the end to minimize contamination and memory effects.
53
54
192 No outliers were removed. An overview of the treatment is presented in figure 1.
55
56
57 193
58
59
60 194 FIGURE1
8
1
2
3 195 Selectivity was evaluated by assessing interferences in UPW versus in the blue mussel matrix. This
4
5
196 was achieved by comparing the parameters under investigation in test solutions both spiked and
6
7
8 197 unspiked.
9
10
11 198 The working range of the method in terms of particle mass/size, particle mass and number
12

13 199 concentration were determined by analyzing ionic concentrations of gold and by approximating the
14
15 200 probability of particle coincidence from Gaussian and Poisson statistics, respectively. For the mass or
16
17 201 size per particle detection limit, a triangular peak shape was assumed using an estimated peak width
18
19
20
202 of 500 µs for gold and 600 µs for the other elements, and a height corresponding to the peak
21
F
22 203 intensity threshold. The mass per particle detection threshold could then be approximated by
O
23
24 204 relating the signal area to the mass per signal from the calibration. For the method applicability
O
25
26
27
205 PR
demonstration, the highest mass-normalized threshold was utilized across all days and
28
206 measurements to ensure an unbiased comparison with similar false negative rates. For
29
D
30
TE
31 207 determination of the particle mass concentration detection limit, the mean pooled value of the mass
32
208 concentrations of all blanks per day plus three times the standard deviation was used. The detection
EC
33
34
35 209 limit for the particle number concentration was similarly determined by a 99.7% Poisson confidence
R
36
37
R
210 interval from the mean pooled values of procedural blanks on each day. The upper range of particle
38
O
39
40 211 mass/size was not experimentally assessed.
C
41
N
42
212 Trueness of size as well as particle mass and number concentration were evaluated by recovery
U
43
44
45 213 experiments and analyzing a proficiency test. Recovery was assessed by comparing the measured
46
47 214 values of gold NPs spiked in mussel tissues and NPs spiked in UPW to reveal matrix specific bias,
48
49 215 whereas laboratory bias was investigated by analyzing a proficiency test (PT) sample consisting of
50
51
52
216 gold nanoparticles in UPW and calculating z-scores (equations in the supplementary information).
53
54 217 For laboratory bias, particle mass concentration was not assessed.
55
56
57 218 Precision as repeatability and intermediate precision was determined from replicates on each day
58
59 219 over the five days using a one-way ANOVA test as described by EURACHEM (39)
60
9
1
2
3 220 To account for random and systematic errors, measurement uncertainty was estimated for each
4
5
221 measurand using a type B bottom-up-approach. This was achieved through a Monte Carlo simulation
6
7
8 222 using R version 4.2.1 and the metRology package (40) and is detailed in the supplementary
9
10 223 information. A coverage factor of two was used to calculate the expanded measurement uncertainty
11
12 224 corresponding to a 95% confidence interval.

13
14
15 225 Stability of NP standards and spiked mussels were investigated by analyzing spiked method blanks
16
17 226 and spiked mussels . Robustness was evaluated by using different batches of reagents and solutions
18
19
20
227 on different days. Different personnel were also involved on different days during the five separate
21
F
22 228 days of this validation study.
O
23
24
O
25 229 Results and discussion
26
27
PR
28 230 Selectivity
29
D
30
TE
31 231 Selectivity as determined by evaluating instrument blanks and matrix blanks showed low
32
EC
33 232 concentrations in the unspiked aliquots as displayed in table 1. The particles found were attributed
34
35
233 to carryover and contributed to higher detection limit, and could be further reduced by prolonging
R
36
37
R
38 234 the rinsing procedure or changing its composition. This is consistent with the absence of isobaric
O
39
40 235 interferences for particle signals at 197Au, with the only reported polyatomic interference being
C
41
N
42 236 181Ta16O (41). Studies have reported that the morphology of particles can impact the shape of their
U
43
44
237 peaks (42). However, the current method lacks the ability to differentiate between gold NPs of
45
46
47 238 various forms and size is determined based on assumptions regarding shape and density. It is
48
49 239 possible to differentiate ionic from particulate species, however, it has been speculated that ionic
50
51 240 species that are adsorbed onto particles could be falsely detected as particles (43). For other
52
53
241 elements, interferents may be more abundant, yet interferents not causing particle signals will only
54
55
56 242 contribute to a higher background and thus higher mass/size detection limits.
57
58
59 243
60
10
1
2
3 244 FIGURE2
4
5
6 245 Matrix effects were evaluated by comparing NPs spiked to enzymatically digested mussel matrix
7
8 246 (matrix blank) with NPs suspended in UPW (instrument blank). Mussel matrix resulted in an 8%
9
10
11 247 decrease in mean diameter for 30 and 60 nm particles and a 19% and 21% decrease in terms of
12

13 248 particle mass concentration in comparison to UPW (figure 2; table 1). For the particle number
14
15 249 concentration, no significant difference could be found as determined by a t-test. From this, the
16
17 250 particles appear stable in mussel tissue, as no aggregation or change in the size distribution can be
18
19
20
251 observed. However, the blue mussel matrix was linked with a signal suppression resulting in lower
21
F
22 252 size and mass concentration, yet did not have an impact on either the transport efficiency or the
O
23
24 253 particle number concentration. Consequently, bias may be observed for particle diameters and
O
25
26
27
254 PR
particle mass concentrations depending on the matrix concentration and element measured.
28
255 However, either by matrix matching particle standards, or diluting the matrix, this difference could
29
D
30
TE
31 256 be mitigated or corrected for as discussed elsewhere (16, 44–46). However, correction was not
32
257 performed due to being unfeasible for other elements for which no reference materials are available,
EC
33
34
35 258 and as the method performance was fit for purpose and acceptable versus Codex criteria. Whereas
R
36
37
R
259 not attributed to the matrix, for 60 nm particles a signal artifact or tailing effect can be observed
38
O
39
40 260 appearing as a second population around 20 nm (figure 2). Investigations of this phenomena
C
41
N
42 261 inspecting both individual peak shapes and signal distributions from monodisperse gold NPs of eight
U
43
44 262 different sizes in UPW found a clear dependence on particle size (supplementary figure 3). Evidence
45
46 263 signifies that this artifact is not an effect of the background or the signal processing, and may be a
47
48
49 264 source for size bias and non-linear particle signal to particle mass response for larger particles using
50
51 265 SP-ICP-MS (supplementary figure 4). This could lead to an overestimation in particle number
52
53 266 concentration and an underestimation of particle sizes, whereas the mass concentration would
54
55 267 remain the same. This artifact has also been described for Ag NPs in fruit juice (47) and UPW (48), the
56
57
58 268 latter attributing this to the expansion of ionic cloud as detailed in earlier works using microdroplet
59
60 269 generators (49). However, the cause remains unclear. The artifact may often be eliminated due to
11
1
2
3 270 many vendor-software's reliance on subjective manual adjustments for particle discrimination, and
4
5
271 be dependent on instrumental setup and parameters. Widening of the particle size distribution using
6
7
8 272 high nebulizer gas flow and sampling depth (50) may be related and underlines the importance of
9
10 273 parameter optimization and reporting. This effect could explain seemingly contradictory reports of
11
12 274 quantitative transfer for up to 1 µm SiO2 particles versus substantial non-linear relationships at

13
14 275 several hundred nm (51), and as low as around 100 nm (52, 53).
15
16
17 276 Working range
18
19
20 277 Working range for particle size or mass, particle mass concentration and particle number
21
F
22
278 concentration is limited on the lower end by the limit of detection, and in the upper range by a non-
O
23
24
O
25 279 linear response of the measured quantity.
26
27
280
PR
The size or mass detection limit per particle corresponds to a Currie critical limit with a false positive
28
29
D
30 281 probability (α) of 8.33 x 10-6% per dwell under the assumption of Poisson noise, as detailed
TE
31
32 282 previously. This resulted on each day in a size of 18 nm or particle masses between 55 and 63
EC
33
34 283 attograms with a false negative probability of 50%.
35
R
36
37
R
284 The upper detection limit per particle was not experimentally assessed in the validation setup.
38
O
39
285 However, for larger particles, aforementioned particle signal artifact, decreased transmission
40
C
41
N
42 286 efficiency (54), detector saturation and decreased transport efficiency can cause a bias. We therefore
U
43
44 287 expect the upper limit of our method to be at least 100 nm for gold NPs, in line with previous reports
45
46 288 (8, 50, 52, 53, 55) and separate experiments shown in supplementary figure 4. Above the working
47
48 289 range, bias may be substantial. For monitoring nanomaterials e.g. in food for legislative purposes,
49
50
51 290 nanoparticles below 100 nm are of particular relevance (56). Non-detection of particles below the
52
53 291 detection limit may be consequential since they may be numerous. In terms of particle mass
54
55 292 concentration and particle number concentration, detection limits were determined to 1.7 ng/g
56
57 293 mussel tissue and 4.2 x 105 particles/g mussel tissue, respectively.
58
59
60
12
1
2
3 294 The upper working range for determination of particle mass concentration may be assumed to be
4
5
295 equivalent to the working range for ionic forms of gold. Generally for ICP-MS, linear range spans
6
7
8 296 many orders of magnitudes (57), and for the current work found to be linear at least in the range of
9
10 297 0.25 to 5 ng/g: ionic standards used to determine transport efficiency were linear in this
11
12 298 concentration range with R2 always 0.99997 or better (supplementary table 4). The upper working

13
14 299 range for particle number concentration was defined to 5% particle coincidence and approximated
15
16
17 300 from peak widths of 500 µs under the assumption of interarrival times following an exponential
18
19 301 distribution and occurrence per time following a Poisson distribution. This resulted in a working
20
21 302 range of up to 7.76 x 107 particles/g mussel tissue for 30 nm and 5.96 x 107 particles/g for 60 nm for
F
22
O
23 303 the employed dilution factor of 500. This is predominantly in the range of reported environmental
24
O
25
304 concentrations of metallic NPs or higher. For this reason, the working range is sufficient for
26
27
PR
28 305 determining expected environmentally relevant concentrations. If necessary, the method working
29
D
30 306 range could be extended by increasing or decreasing the dilution factor. Transport efficiency was in
TE
31
32 307 the range 5.4 - 6.4%, corresponding to values reported with comparable setup (58–60).
EC
33
34
35 308
R
36
37
R
38 309
O
39
40
C
41 310 Trueness as recovery

N
42
U
43
44 311 Trueness as recovery was evaluated by comparison of gold NPs in UPW with gold NPs spiked to
45
46 312 method blanks (enzymatic digestion in absence of mussel tissue) and mussel tissue to assess the
47
48 313 matrix specific bias. Mussel tissue was spiked both before (referred to as “blue mussel”) and after
49
50
51 314 enzymatic digestion (referred to as “matrix blank”). As demonstrated in table 2, mean particle
52
53 315 diameters were comparable between gold NPs spiked to mussel tissue before and after enzymatic
54
55 316 digestion and the method blank with recoveries in the range of 86 to 92% for 60 nm particles and 92
56
57 317 to 93% for 30 nm particles. The previously described signal suppression (smaller diameters in
58
59
60 318 comparison to gold NPs in UPW) may be caused by the enzyme as it is observed already in the
13
1
2
3 319 method blank. It is also reflected by mussel tissue showing particle mass concentration recoveries
4
5
320 at 77% and 84% for 60 nm and 76% and 81% for 30 nm spiked prior to and after digestion,
6
7
8 321 respectively. Corresponding particle number concentration recoveries were 94%, 103%, 101% and
9
10 322 103%. Thus, recoveries were similar whether spiked to the matrix before or after digestion or in
11
12 323 ultrapure water. This indicates that particles in the mussel matrix were stable during incubation and

13
14 324 resulted in no change in transport efficiency.
15
16
17 325 Recoveries were within Codex Alimentarius' numeric method criteria interval of 40 to 120% (33).
18
19
20
326 Analytical recoveries for enzymatic digestion of bivalves have been reported at 109, 92 and 85% for
21
F
22 327 Ag (15, 61) and 95% for Ti (62) in terms of number concentrations, and 104 and 93% (15, 63) for mass
O
23
24 328 concentrations, whereas recoveries for enzymatic digestion in general have been reported as low as
O
25
26
27
329 PR
near 20% due to incomplete digestion or high detection limits (12).
28
29
D
330 Trueness may also be calculated from the nominal sizes and concentrations of the reference material
30
TE
31 331 used for spiking, presented in supplementary table 5. However, the availability of certified reference
32
EC
33
34 332 materials is sparse and there is inherent uncertainty in the reference values, which may also change
35
R
36 333 over time due to surface adsorption, dissolution, and aggregation. In the present work, recoveries
37
R
38 334 versus nominal gold mass concentrations were found as low as approximately 40%. The low
O
39
40
C
335 concentration recoveries were attributed to the stability of the reference material used, as further
41
N
42
336 laboratory experience with the method using other NP standards consistently produced recoveries
U
43
44
45 337 closer to 100%. Nonetheless, while trueness is crucial for comparison with e.g. regulatory limits,
46
47 338 meaningful interpretations of trends and changes of environmental and food safety relevance can be
48
49 339 derived from precision trends. It is worth noting that partly due to these limitations in establishment
50
51
52
340 of trueness, sp-ICP-MS is characterized as a screening technique and that unequivocal confirmation
53
54 341 will have to be done by other techniques such as e.g. transmission or scanning electron microscopy
55
56 342 (64). Observed recoveries were considered satisfactory for the purpose. However, spiked engineered
57
58 343 NPs may behave different than NPs found in the environment.
59
60
14
1
2
3 344 Trueness from proficiency test
4
5
6 345 Trueness as z-score was evaluated by analyzing the ACEnano proficiency test samples in parallel for
7
8 346 each experimental run to evaluate the laboratory bias (35). Relative trueness of particle number
9
10
11 347 concentration and mean size was determined to be 112% and 94%, respectively, corresponding to
12

13 348 satisfactory z-scores of 0.3 and 0.6. However, it should be noted that this proficiency test was an
14
15 349 ideal sample without complex matrix since there is also a lack of available proficiency tests in
16
17 350 matrices.
18
19
20 351 Repeatability and intermediate precision
21
F
22
O
23 352 Repeatability and intermediate precision for size determination is generally more robust than
24
O
25 353 concentration parameters due to being the mass equivalent diameter with a cubic dependence. This
26
27
354
PR
is reflected in the precision parameters obtained of below 2% (table 3). For particles of different
28
29
D
30 355 elements in various matrices, e.g. gold and silver NPs in fruit juices (47), silver NPs in chicken meat
TE
31
32 356 (65), silver NPs in confectionery (28) and titanium dioxide NPs in human urine (66) and gold NPs in
EC
33
34 357 UPW (67), repeatability standard deviations for size have mostly been reported below 10% and
35
R
36
37
358 intermediate precision below 15%. Hence, the present method compares favorably to previous
R
38
O
39 359 studies.
40
C
41
N
42 360 The precision of particle mass and number concentrations are also acceptable when compared to
U
43
44 361 numeric criteria in Codex Alimentarius (33), with repeatability and intermediate precision up to 5%
45
46 362 and 9%, respectively. These are lower in comparison to precision parameters reported for biological
47
48 363 matrices, e.g. with repeatability of 8% for number concentration in mussels (61), a repeatability of
49
50
51 364 14% and reproducibility of 16% for mass concentration in urine (66), repeatabilities of 16-29% for
52
53 365 number and mass concentrations in confectionary (28), repeatability and reproducibility in
54
55 366 confectionary and pristine solutions for number concentrations of 9-21% and 8-97%, respectively
56
57 367 (59). Interlaboratory studies have reported repeatabilities between 7 and 18 % and reproducibilities
58
59
60 368 between 70 and 90% (8).
15
1
2
3 369 Measurement uncertainty
4
5
6 370 Although most uncertainties in SP-ICP-MS measurements are top-down estimates using the standard
7
8 371 deviation of replicates or from precision parameters, some recent efforts have been made to more
9
10
11 372 rigorously quantify their underlying sources (28, 68). Here, measurement uncertainties were
12

13 373 determined by a conservative bottom-up Monte Carlo approach as detailed in the supplementary
14
15 374 information along with quantity inputs. The expanded uncertainties using a coverage factor of two
16
17 375 were determined to 8.8% for mean particle diameter, 37% for particle mass concentration and 45%
18
19
20
376 for particle number concentrations. While dependent on particle sizes, polydispersity, and
21
F
22 377 calculation of uncertainty, these values fall in similar range as reported in the aforementioned
O
23
24 378 studies. The greater uncertainty for particle mass and number concentration is primarily due to the
O
25
26
27
379 PR
uncertainty in the counting statistics at the lower end of the working range. Particle number
28
380 concentrations are additionally dependent on both ionic and particulate reference materials. The
29
D
30
TE
31 381 uncertainty can be decreased by replicate measurements or measuring a larger number of particles.
32
382 E.g. for 1000 measured particles, uncertainties decrease to 8.7%, 16% and 29%, for size, mass and
EC
33
34
35 383 number concentrations, respectively.
R
36
37
R
38 384
O
39
40
C
41 385
N
42
U
43
44 386 Stability
45
46
47 387 Particle stability was assessed by the difference in particle number and mass concentrations in spiked
48
49 388 method blanks versus matrix blanks spiked before and after digestion. As shown in table 2,
50
51 389 recoveries were highest for mussels spiked after digestion, followed by mussels spiked before
52
53
54
390 digestion and spiked method blanks. Method blanks deviated from the other matrices with
55
56 391 recoveries in the range of only 13-20% in terms of particle mass and number concentrations for 30
57
58 392 and 60 nm. Matrices of UPW and mussel matrix spiked before and after were found to be similar for
59
60 393 particle number concentration recoveries, whereas particle mass concentration differences were
16
1
2
3 394 attributed to matrix effects as discussed under trueness. These findings imply that the particles were
4
5
395 stable in UPW and in mussel matrices, with high recoveries and no apparent aggregation, but not
6
7
8 396 when incubated with the protease mixture in isolation. Due to no change in the particle sizes
9
10 397 measured, it is reasonable to assume surface adsorption to be the main mechanism, which has been
11
12 398 reported as a potential source of bias in dilute particle suspensions (45, 69). This has implications for

13
14 399 the selection of blanks and demonstrates that caution should be exercised when blank subtracting
15
16
17 400 and establishing detection limits based on blanks with different matrix. Stability of spiked solutions
18
19 401 has been observed to be poor, however matrices such as mussel tissue have a stabilizing effect.
20
21
F
22 402
O
23
24
O
25 403 Method application to other NPs in mussels
26
27
404
PR
28
29
D
30
405 FIGURE3
TE
31
32
EC
33 406
34
35
R
36 407 The application of the method to environmentally relevant particles of chromium, copper and
37
R
38
408 titanium-containing particles was demonstrated using aggregate samples from three different
O
39
40
C
41 409 locations, analyzed on two days. The aggregate samples represented a mussel farm, a harbor central
N
42
U
43 410 in Bergen and an unspecified location in the surveillance program. As demonstrated in figure 3, the
44
45 411 urban location, harbor, exhibited significantly higher mass concentrations of Cr and Cu and had
46
47
412 higher concentrations of Ti particles in comparison to the surveillance program sample. Particle
48
49
50 413 number concentrations exhibited similar trends, shown in supplementary figure 1. Concentrations
51
52 414 found for metal NPs can be orders of magnitude lower than for total metals (46, 70, 71), for which
53
54 415 there are presently no regulatory limits. The between-days variability was in many cases smaller than
55
56
416 the within-day variability, as found in other studies (28), indicating the method's robustness. The
57
58
59 417 larger variance for the harbor sample may be attributed to sample heterogeneity due to differences
60
17
1
2
3 418 in physiology. The relative standard deviations were generally higher than observed for gold, yet
4
5
419 acceptable as per the Codex criteria except for particle mass concentrations for titanium and
6
7
8 420 chromium in some samples (table 4). This is to be expected due to the added uncertainty of sampling
9
10 421 in combination with the polydispersity of environmental particles for which a single large particle
11
12 422 may account for a substantial part of the total mass concentration. This was observed for titanium in

13
14 423 one parallel of harbor (figure 2). Similarly, detection limits in terms of size are higher due to the
15
16
17 424 presence of dissolved and interfering species causing elevated background noise. For 48Ti, spectral
18
19 425 interference from 48Ca is possible, although a natural abundance of below 0.2% for this isotope limits
20
21 426 this to a minor effect. However, the primary limitation of particle measurements of environmental
F
22
O
23 427 samples is determining the trueness. Due to the absence of representative certified reference
24
O
25
428 materials and proficiency tests, both pure and in matrix, and the difficulty of determining
26
27
PR
28 429 environmentally relevant concentrations using alternative techniques, trueness generally cannot be
29
D
30 430 established. However, mass recoveries based on spiking titanium dioxide NPs into mussels prior to
TE
31
32 431 incubation resulted in recoveries of 81 - 93 % in comparison to a freshly prepared UPW solution
EC
33
34
35
432 (supplementary table 10). A second limitation is that determining the composition of particles is not
R
36
37 433 achievable, although time-of-flight instrumentation can offer more information. However, SP-ICP-MS
R
38
O
39 434 may anyway serve as fit for purpose standard method provided full validation is performed in a
40
C
41 435 collaborative study. As environmental particle concentrations may increase logarithmically with
N
42
U
43
436 decreasing size (72), especially comparisons on particle number basis are generally unsubstantiated
44
45
46 437 unless detection limits are similar in terms of both false negatives and false positives. However,
47
48 438 ensuring consistent sample and data treatment, different samples may be compared on a
49
50 439 quantitative basis. This provides valuable insights into relative concentrations and allows surveillance
51
52
440 of concentration of nanoparticles in the marine environment.
53
54
55
56 441
57
58
59 442 Conclusion
60
18
1
2
3 443 The study validated a method for determination of the size, mass and number concentration of
4
5
444 nanoparticles in mussels using a cost-effective enzyme mixture suitable for routine applications. The
6
7
8 445 method was highly selective for gold, yet interference as signal suppression may cause an
9
10 446 underestimation unless corrected for. The working range in terms of particle size ranged from 18 nm
11
12 447 to at least 100 nm. The working range of particle number and mass concentrations from 1.7 ng/g to

13
14 448 at least 5 ng/g and in the 108s to 1011s, respectively. Trueness as determined from proficiency test
15
16
17 449 samples without matrix and in matrix versus UPW was between 76 and 112% for all parameters.
18
19 450 Repeatability and intermediate precision were below 2% for size and at most 9% for mass and
20
21 451 number concentrations. Measurement uncertainties were 9%, 16% and 29% for size, mass and
F
22
O
23 452 number concentrations, respectively. Thus, the overall method performance is acceptable in
24
O
25
453 comparison to Codex recommendations and working ranges mostly span expected environmental
26
27
PR
28 454 levels and could be extended by dilution. As such, the method is fit for purpose in determining metal
29
D
30 455 containing NPs such as gold in seafood samples with similar matrix composition as mussels. Whereas
TE
31
32 456 as demonstrated the high precision data is suitable to monitor environmental trends, a remaining
EC
33
34
35
457 challenge is the difficulty of ascertaining the trueness, especially given the unavailability of certified
R
36
37 458 reference materials and proficiency tests containing metal containing NPs. However, by employing
R
38
O
39 459 similar sample treatment, analysis and detection limits through future method standardization,
40
C
41 460 comparable quantitative data may be obtained for surveillance and monitoring purposes.
N
42
U
43
44 461
45
46
47 462
48
49
50
463
51
52
53
54
464
55
56
57 465
58
59
60 466 CRediT Author Statement
19
1
2
3 467 AB: Conceptualization; Data curation; Formal analysis; Investigation; Methodology;
4
5
6 468 Software; Validation; Visualization; Writing—original draft; Writing—review & editing. SV:
7
8 469 Conceptualization; Data curation; Formal analysis; Funding acquisition; Investigation;
9
10
11 470 Methodology; Project administration; Supervision; Validation; Roles/Writing—original draft;
12

13 471 Writing—review & editing. KL: Conceptualization; Funding acquisition; Methodology;
14
15
16
472 Supervision; Validation; Writing—original draft; Writing—review & editing. AMB:
17
18 473 Conceptualization; Data curation; Formal analysis; Funding acquisition; Investigation;
19
20 474 Methodology; Validation; Roles/Writing—original draft; Writing—review & editing.
21
F
22
O
23
24 475 Acknowledgements
O
25
26
27 476
PR
We kindly thank Berit Solli for her contributions to the nano-project. JRC Nanomaterials Repository is
28
29
D
477 thanked for supplying the TiO2 nanoparticles.
30
TE
31
32 478 Funding
EC
33
34
35
479 The work was funded by the Institute of Marine Research project Marine nanoparticles (15318).
R
36
37
R
38
Conflicts of Interest
O
39 480
40
C
41
N
42 481 The authors declare that we have no known competing financial or personal interests or
U
43
44
45
482 relationships or personal relationships that could influence the work.
46
47
48 483 Data availability
49
50
51 484 All data and code to reproduce all results and visualizations reported herein has been made
52
53
54 485 publicly available through online repositories (37).
55
56
57 486
58
59
60 487 References
20
1
2
3 488 1. Baker, T.J., Tyler, C.R., & Galloway, T.S. (2014) Environ. Pollut. 186, 257–271.
4
5
489 doi:10.1016/j.envpol.2013.11.014
6
7 490 2. Hochella, M.F., Mogk, D.W., Ranville, J., Allen, I.C., Luther, G.W., Marr, L.C., McGrail, B.P.,
8 491 Murayama, M., Qafoku, N.P., Rosso, K.M., Sahai, N., Schroeder, P.A., Vikesland, P.,
9 492 Westerhoff, P., & Yang, Y. (2019) Science 363, eaau8299. doi:10.1126/science.aau8299
10
11 493 3. Montaño, M.D., von der Kammer, F., Cuss, C.W., & Ranville, J.F. (2019) J. Anal. At.
12
494 Spectrom. 34, 1768–1772. doi:10.1039/c9ja00168a

13
14
15 495 4. Canesi, L. & Corsi, I. (2016) Sci. Total Environ. 565, 933–940.
16 496 doi:10.1016/j.scitotenv.2016.01.085
17
18 497 5. Beyer, J., Green, N.W., Brooks, S., Allan, I.J., Ruus, A., Gomes, T., Bråte, I.L.N., & Schøyen,
19
20
498 M. (2017) Mar. Environ. Res. 130, 338–365. doi:10.1016/j.marenvres.2017.07.024
21
F
22 499 6. Ward, J.E. & Kach, D.J. (2009) Mar. Environ. Res. 68, 137–142.
O
23 500 doi:10.1016/j.marenvres.2009.05.002
24
O
25 501 7. FAO. Available at: https://www.fao.org/in-action/globefish/fishery-information/resource-
26
27
502 detail/en/c/338588/. Accessed on 7 June 2023 PR
28
29 503 8. Mozhayeva, D. & Engelhard, C. (2020) J. Anal. At. Spectrom. 10.1039.C9JA00206E.
D
30 504 doi:10.1039/c9ja00206e
TE
31
32 505 9. Laborda, F., Bolea, E., Cepriá, G., Gómez, M.T., Jiménez, M.S., Pérez-Arantegui, J., &
EC
33 506 Castillo, J.R. (2016) Anal. Chim. Acta 904, 10–32. doi:10.1016/j.aca.2015.11.008
34
35
507 10. Giusti, A., Atluri, R., Tsekovska, R., Gajewicz, A., Apostolova, M.D., Battistelli, C.L.,
R
36
37 508 Bleeker, E.A.J., Bossa, C., Bouillard, J., Dusinska, M., Gómez-Fernández, P., Grafström, R.,
R
38 509 Gromelski, M., Handzhiyski, Y., Jacobsen, N.R., Jantunen, P., Jensen, K.A., Mech, A., Navas,
O
39 510 J.M., Nymark, P., Oomen, A.G., Puzyn, T., Rasmussen, K., Riebeling, C., Rodriguez-Llopis, I.,
40
C
511 Sabella, S., Sintes, J.R., Suarez-Merino, B., Tanasescu, S., Wallin, H., & Haase, A. (2019)
41
NanoImpact 16, 100182. doi:10.1016/j.impact.2019.100182
N
42 512
U
43
44 513 11. Gao, X. & Lowry, G.V. (2018) NanoImpact 9, 14–30. doi:10.1016/j.impact.2017.09.002
45
46 514 12. Laycock, A., Clark, N.J., Clough, R., Smith, R., & Handy, R.D. (2022) Environ. Sci. Nano
47 515 10.1039.D1EN00680K. doi:10/gr5fgv
48
49
516 13. Correia, M., Verleysen, E., & Loeschner, K. (2019) in Nanomaterials for Food Applications,
50
51 517 Elsevier, pp 273–311. doi:10.1016/B978-0-12-814130-4.00010-5
52
53 518 14. Sun, Y., Yang, Y., Tou, F.-Y., Niu, Z.-S., Guo, X.-P., Liu, C., Yan, J., Wu, J.-Y., Xu, M., Hou, L.-
54 519 J., & Liu, M. (2022) J. Hazard. Mater. 424, 127383. doi:10.1016/j.jhazmat.2021.127383
55
56 520 15. Suzuki, Y., Harimoto, M., Takahashi, M., Akiyama, H., Hirose, A., & Tsutsumi, T. (2024) J.
57
58
521 Environ. Chem. 34, 9–20. doi:10.5985/jec.34.9
59
60
21
1
2
3 522 16. Vidmar, J., Buerki-Thurnherr, T., & Loeschner, K. (2018) J. Anal. At. Spectrom. 33, 752–
4
5
523 761. doi:10/ggkvjv
6
7 524 17. Nguyen, H.T.M., Sylla, K.S.B., Randriamahatody, Z., Donnay-Moreno, C., Moreau, J., Tran,
8 525 L.T., & Bergé, J.P. (2011)
9
10 526 18. Liaset, B., Nortvedt, R., Lied, E., & Espe, M. (2002) Process Biochem. 37, 1263–1269.
11 527 doi:10.1016/s0032-9592(02)00003-1
12

13
14 528 19. Loeschner, K., Navratilova, J., Købler, C., Mølhave, K., Wagner, S., Von Der Kammer, F., &
15 529 Larsen, E.H. (2013) Anal. Bioanal. Chem. 405, 8185–8195. doi:10.1007/s00216-013-7228-z
16
17 530 20. European Commission (2022)
18
19 531 21. FAO (2010) in International Conference on Food and Agriculture, FAO, Sao Pedro, Brazil
20
21
532 22. Watkins, P.S., Castellon, B.T., Tseng, C., Wright, M.V., Matson, C.W., & Cobb, G.P. (2018)
F
22
533 Bull. Environ. Contam. Toxicol. 100, 809–814. doi:10.1007/s00128-018-2336-2
O
23
24
O
25 534 23. Sandrine Millour, Y.N. (2015) J. Nanomedicine Nanotechnol. 06. doi:10.4172/2157-
26
27
535 7439.1000269 PR
28 536 24. Witzler, M., Küllmer, F., & Günther, K. (2018) Anal. Lett. 51, 587–599.
29
D
30 537 doi:10.1080/00032719.2017.1327538
TE
31
32 538 25. Gray, E.P., Coleman, J.G., Bednar, A.J., Kennedy, A.J., Ranville, J.F., & Higgins, C.P. (2013)
Environ. Sci. Technol. 47, 14315–14323. doi:10.1021/es403558c
EC
33 539
34
35 540 26. Loeschner, K., Navratilova, J., Grombe, R., Linsinger, T.P.J., Købler, C., Mølhave, K., &
R
36
541 Larsen, E.H. (2015) Food Chem. 181, 78–84. doi:10.1016/j.foodchem.2015.02.033
37
R
38
O
39 542 27. Peters, R.J.B., Rivera, Z.H., van Bemmel, G., Marvin, H.J.P., Weigel, S., & Bouwmeester, H.
40 543 (2014) Anal. Bioanal. Chem. 406, 3875–3885. doi:10.1007/s00216-013-7571-0
C
41
N
42 544 28. Waegeneers, N., De Vos, S., Verleysen, E., Ruttens, A., & Mast, J. (2019) Materials 12,
U
43 545 2677. doi:10.3390/ma12172677

44
45
46 546 29. Taboada-López, M.V., Herbello-Hermelo, P., Domínguez-González, R., Bermejo-Barrera,
47 547 P., & Moreda-Piñeiro, A. (2019) Talanta 195, 23–32. doi:10.1016/j.talanta.2018.11.023
48
49 548 30. Linsinger, T.P.J., Chaudhry, Q., Dehalu, V., Delahaut, P., Dudkiewicz, A., Grombe, R., von
50 549 der Kammer, F., Larsen, E.H., Legros, S., Loeschner, K., Peters, R., Ramsch, R., Roebben, G.,
51
550 Tiede, K., & Weigel, S. (2013) Food Chem. 138, 1959–1966.
52
53 551 doi:10.1016/j.foodchem.2012.11.074
54
55 552 31. Gilroy, K.D., Neretina, S., & Sanders, R.W. (2014) J Nanopart Res 8. doi:10/gr5fgj
56
57 553 32. Kammer, F. von der, Ferguson, P.L., Holden, P.A., Masion, A., Rogers, K.R., Klaine, S.J.,
58 554 Koelmans, A.A., Horne, N., & Unrine, J.M. (2012) Environ. Toxicol. Chem. 31, 32–49.
59
60
555 doi:10.1002/etc.723
22
1
2
3 556 33. FAO and WHO (2019) Codex Alimentarius Commission – Procedural Manual twenty-
4
5
557 seventh edition, Rome
6
7 558 34. Duinker, A., Storesund, J., Lundestad, B.T., & Sanden, M. (2020) National monitoring
8 559 program for bivalves and other molluscs, Institute of Marine Research
9
10 560 35. Peters, R., Elbers, I., Undas, A., Sijtsma, E., Briffa, S., Carnell-Morris, P., Siupa, A., Yoon, T.-
11 561 H., Burr, L., Schmid, D., Tentschert, J., Hachenberger, Y., Jungnickel, H., Luch, A., Meier, F.,
12
562 Radnik, J., Hodoroaba, V.-D., Lynch, I., & Valsami-Jones, E. (2021) Molecules 26, 5315.

13
14 563 doi:10/gr5fgn
15
16 564 36. Pace, H.E., Rogers, N.J., Jarolimek, C., Coleman, V.A., Higgins, C.P., & Ranville, J.F. (2011)
17 565 Anal. Chem. 83, 9361–9369. doi:10.1021/ac201952t
18
19
20
566 37. Bruvold, A. (2023). doi:https://doi.org/10.5281/zenodo.8301248
21
F
22 567 38. Julshamn, K., Lea, P., & Lindeberg, J. (2009) Validation of Chemical Analytical Methods,
O
23 568 Nordic Committee on Food Analysis
24
O
25 569 39. B. Magnusson & U. Örnemark (2014) Eurachem Guide: The Fitness for Purpose of
26
27
570 PR
Analytical Methods – A Laboratory Guide to Method Validation and Related Topics, 2nd ed.
28
29 571 40. Ellison, S., L.R. (2017) metRology: Support for Metrological Applications v0.9-28-1,
D
30 572 cran.rproject.org/package=metRology
TE
31
32 573 41. May, Thomas W. & Wiedmeyer, Ray H. (1998) At. Spectrosc. 19, 150–155
EC
33
34 574 42. Kálomista, I., Kéri, A., Ungor, D., Csapó, E., Dékány, I., Prohaska, T., & Galbács, G. (2017) J.
35
575 Anal. At. Spectrom. 32, 2455–2462. doi:10.1039/c7ja00306d
R
36
37
R
38 576 43. Gonzalez de Vega, R., Lockwood, T.E., Xu, X., Gonzalez de Vega, C., Scholz, J., Horstmann,
O
39 577 M., Doble, P.A., & Clases, D. (2022) Anal. Bioanal. Chem. 414, 5671–5681. doi:10/gr5fgh
40
C
41 578 44. Torregrosa, D., Gómez-Pertusa, C., Grindlay, G., Gras, L., & Mora, J. (2023) J. Anal. At.
N
42 579 Spectrom. 38, 403–413. doi:10.1039/D2JA00342B

U
43
44
45 580 45. Liu, J., Murphy, K.E., Winchester, M.R., & Hackley, V.A. (2017) Anal. Bioanal. Chem. 409,
46 581 6027–6039. doi:10.1007/s00216-017-0530-4
47
48 582 46. Bruvold, A.S., Bienfait, A.M., Ervik, T.K., Loeschner, K., & Valdersnes, S. (2023) Mar.
49 583 Environ. Res. 188, 105975. doi:10.1016/j.marenvres.2023.105975
50
51
584 47. Witzler, M., Küllmer, F., Hirtz, A., & Günther, K. (2016) J. Agric. Food Chem. 64, 4165–
52
53 585 4170. doi:10/f8p9k9
54
55 586 48. Tuoriniemi, J., Cornelis, G., & Hassellöv, M. (2014) J Anal Spectrom 29, 743–752.
56 587 doi:10.1039/C3JA50367D
57
58 588 49. Stewart, I.I. & Olesik, J.W. (1999) J. Am. Soc. Mass Spectrom. 10, 159–174.
59
60
589 doi:10.1016/S1044-0305(98)00136-6
23
1
2
3 590 50. Torregrosa, D., Grindlay, G., Gras, L., & Mora, J. (2023) J. Anal. At. Spectrom. 38, 1874–
4
5
591 1884. doi:10.1039/D3JA00134B
6
7 592 51. Laborda, F., Bolea, E., & Jiménez-Lamana, J. (2016) Trends Environ. Anal. Chem. 9, 15–23.
8 593 doi:10.1016/j.teac.2016.02.001
9
10 594 52. Schardt, A., Schmitt, J., & Engelhard, C. (2022) Single Particle Inductively Coupled Plasma
11 595 Mass Spectrometry with Nanosecond Time Resolution, Chemistry. doi:10.26434/chemrxiv-
12
596 2022-rc0c2

13
14
15 597 53. Ho, K.-S., Lui, K.-O., Lee, K.-H., & Chan, W.-T. (2013) Spectrochim. Acta Part B At.
16 598 Spectrosc. 89, 30–39. doi:10/gfb2v9
17
18 599 54. Olesik, J.W. & Gray, P.J. (2012) J. Anal. At. Spectrom. 27, 1143–1155. doi:10/gfb2v5
19
20
21
600 55. Lee, W.-W. & Chan, W.-T. (2015) J. Anal. At. Spectrom. 30, 1245–1254. doi:10/gfb2vd
F
22
601 56. The European Commission (2022) Off. J. Eur. Union 65
O
23
24
O
25 602 57. Olesik, J.W. (2014) in Treatise on Geochemistry, Elsevier, pp 309–336. doi:10.1016/B978-
26
27
603 0-08-095975-7.01426-1 PR
28 604 58. Gomez-Gomez, B., Perez-Corona, M.T., & Madrid, Y. (2020) Anal. Chim. Acta 1100, 12–
29
D
30 605 21. doi:10.1016/j.aca.2019.11.063

TE
31
32 606 59. Geiss, O., Bianchi, I., Senaldi, C., Bucher, G., Verleysen, E., Waegeneers, N., Brassinne, F.,
Mast, J., Loeschner, K., Vidmar, J., Aureli, F., Cubadda, F., Raggi, A., Iacoponi, F., Peters, R.,
EC
33 607
34 608 Undas, A., Müller, A., Meinhardt, A.-K., Walz, E., Gräf, V., & Barrero-Moreno, J. (2021) Food
35
609 Control 120, 107550. doi:10.1016/j.foodcont.2020.107550
R
36
37
R
38 610 60. Jiménez-Lamana, J., Marigliano, L., Allouche, J., Grassl, B., Szpunar, J., & Reynaud, S.
O
39 611 (2020) Anal. Chem. doi:10.1021/acs.analchem.0c01536

40
C
41 612 61. Taboada-López, M.V., Alonso-Seijo, N., Herbello-Hermelo, P., Bermejo-Barrera, P., &
N
42 613 Moreda-Piñeiro, A. (2019) Microchem. J. 148, 652–660. doi:10.1016/j.microc.2019.05.023

U
43
44
45 614 62. Taboada-López, M.V., Iglesias-López, S., Herbello-Hermelo, P., Bermejo-Barrera, P., &
46 615 Moreda-Piñeiro, A. (2018) Anal. Chim. Acta 1018, 16–25. doi:10.1016/j.aca.2018.02.075
47
48 616 63. Zhou, Q., Liu, L., Liu, N., He, B., Hu, L., & Wang, L. (2020) Ecotoxicol. Environ. Saf. 198,
49 617 110670. doi:10.1016/j.ecoenv.2020.110670
50
51
618 64. Mech, A., Wohlleben, W., Ghanem, A., Hodoroaba, V., Weigel, S., Babick, F., Brüngel, R.,
52
53 619 Friedrich, C.M., Rasmussen, K., & Rauscher, H. (2020) Small 16, 2002228.
54 620 doi:10.1002/smll.202002228
55
56 621 65. Weigel, S., Peters, R., Loeschner, K., Grombe, R., & Linsinger, T.P.J. (2017) Anal. Bioanal.
57 622 Chem. 409, 4839–4848. doi:10/gbq58z
58
59
60
24
1
2
3 623 66. Salou, S., Larivière, D., Cirtiu, C.-M., & Fleury, N. (2021) Anal. Bioanal. Chem. 413, 171–
4
5
624 181. doi:10/gr5fgm
6
7 625 67. De La Calle, I., Menta, M., Klein, M., & Séby, F. (2018) Food Chem. 266, 133–145.
8 626 doi:10.1016/j.foodchem.2018.05.107
9
10 627 68. Montoro Bustos, A.R., Murphy, K.E., & Winchester, M.R. (2022) Anal. Chem. 94, 3091–
11 628 3102. doi:10/gr5ff3
12

13
14 629 69. Malysheva, A., Ivask, A., Hager, C., Brunetti, G., Marzouk, E.R., Lombi, E., & Voelcker, N.H.
15 630 (2016) Nanotoxicology 10, 385–390. doi:10.3109/17435390.2015.1084059
16
17 631 70. Suzuki, Y., Kondo, M., Akiyama, H., & Ogra, Y. (2022) Environ. Pollut. 307, 119555.
18 632 doi:10.1016/j.envpol.2022.119555
19
20
21
633 71. Xu, L., Wang, Z., Zhao, J., Lin, M., & Xing, B. (2020) Environ. Pollut. 260, 114043.
F
22 634 doi:10.1016/j.envpol.2020.114043
O
23
24 635 72. Wilkinson, K.J. & Lead, J.R. (2007) Environmental Colloids and Particles: Behaviour,
O
25 636 Separation and Characterisation, John Wiley & Sons, Ltd
26
27
PR
28
29
D
30
TE
31
32
EC
33
34
35
R
36
37
R
38
O
39
40
C
41
N
42
U
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
25
1
2
3
4
5
6
7
8
9
10

11
12
13
14
15
16 Schematic illustration of the sample treatment used in the validation study, spiking with gold NPs with
17 nominal sizes of 30 or 60 nm.
18
19
F
392x110mm (300 x 300 DPI)
20
O
21
O
22
23
PR
24
25
26
D
27
28
TE
29
30
EC
31
32
33
R
34
35
R
36
O
37
38
C
39
N
40
41
U
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60 ScholarOne Support phone: 434-964-4100 email: ts.mcsupport@thomson.com
1
2
3
4
5
6
7
8
9
10

11
12
13
14
15
16
17
18 Figure 2: Size distributions of 30 nm (left) and 60 nm gold NPs (right) spiked into enzymatically digested
19 blue mussel tissue (matrix blank) and UPW (instrument blank), illustrating the signal suppression by the
F
blue mussel matrix. Dashed red line indicates the pooled mean diameters across all days
20
O
21 1143x381mm (72 x 72 DPI)
O
22
23
PR
24
25
26
D
27
28
TE
29
30
EC
31
32
33
R
34
35
R
36
O
37
38
C
39
N
40
41
U
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
1
2
3
4
5
6
7
8
9
10

11
12
13
14
15
16
17
18
19
F
20
O
21 Particle mass concentrations in mussel tissue, the red line denoting the detection limit and error bars
O
22 indicating the 95% confidence interval of the mean of 6 replicates over two days.
23
PR
24 1143x508mm (72 x 72 DPI)
25
26
D
27
28
TE
29
30
EC
31
32
33
R
34
35
R
36
O
37
38
C
39
N
40
41
U
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
1
2
3 1 Validation of a method for surveillance of nanoparticles in mussels using single particle inductively
4
5
2 coupled plasma mass spectrometry
6
7
8 3
9
10
11 4
12

13
14 5 Table 1: Mean values and their corresponding standard deviations for each parameter in instrument and matrix blanks for
15
16 6 all days, normalized to per kg mussel tissue.
17
18
19 Particle diameter Particle mass concentration Particle number concentration
20
21 [nm] [ng/g] [#/g]
F
22
O
23 Instrument blank (UPW)
24
O
25
26
27
unspiked 28.4 ± 5.6 0.1 ± 0.1 PR 1.7 x 10^5 ± 1.4 x 10^5
28 spiked with 30 nm gold 35.7 ± 0.1 34.4 ± 1.3 7.1 x 10^7 ± 3.0 x 10^6
29
D
30 NPs
TE
31
32
spiked with 60 nm gold 59.1 ± 1.0 23.2 ± 2.0 1.0 x 10^7 ± 7.1 x 10^5
EC
33
34
NPs
35
R
36
37 Matrix blank (enzymatically digested mussel tissue)
R
38
O
39
unspiked 30.1 ± 3.1 0.8 ± 1.3 4.9 x 10^5 ± 4.3 x 10^5
40
C
41
N
U
43
44 NPs
45
47
48 NPs
49
50
7
51
52
53 8
54
55
56 9
57
58
59 10
60
1
1
2
3 11
4
5
6 12
7
8
9 13 Table 2: Recoveries of particle diameters, mass and number concentrations in mussel tissue and method blanks versus UPW.
10
11
12 Sample Particle diameter Particle mass concentration Particle number concentration

13
14 30 nm gold NPs spiked to
15
16
17 Blue mussel 92% 77% 94%
18
19 Matrix blank 92% 84% 103%
20
21 Method blank 93% 17% 20%
F
22
O
23
24 60 nm gold NPs spiked to
O
25
26
27
Blue mussel 90% 76% PR 101%
28
29 Matrix blank 92% 81% 103%
D
30
TE
31 Method blank 86% 13% 18%

32
EC
33 14
34
35
R
36 15
37
R
38
O
39
40
C
41
N
42
U
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2
1
2
3 16
4
5
6 17 Table 3: Repeatability and intermediate precision for the particle diameter, particle mass concentration and particle number
7
8 18 concentration in terms of percentage relative standard deviation (RSD).
9
10
11 RSDrepeatability RSDintermediate precision
12

13 Mean size
14
15 Blue mussel +
16 1.2% 1.6%
17
30 nm
18
19
Blue mussel +
20
1.1% 1.4%
21
F
60 nm
22
O
23
24 Particle mass concentration
O
25
26
27
Blue mussel +
3.4%
PR 7.9%
28 30 nm
29
D
30 Blue mussel +
TE
31 5.0% 8.8%
32 60 nm
EC
33
34 Particle number concentration
35
R
36 Blue mussel +
37
R
2.1% 7.1%
38 30 nm
O
39
40
C
Blue mussel +
41
N
3.4% 7.9%
42
60 nm
U
43
44
45 19
46
47
48 20
49
50
21
51
52
53 22
54
55
56 23
57
58
59 24
60
3
1
2
3 25 Table 4: Precision parameters and detection limits across the two days for mass concentration in ng/g mussel tissue.
4
5
6 Mean [ng/g] Detection limit [ng/g] RSDRepeatability RSDIntermediate precision
7
8
9 Ti
10
11 Farm 80 3.4 11% 13%
12

13
14 Harbor 122 3.4 92% *90%
15
16 Surveillance 14 3.4 31% 46%
17
18
Cr
19
20
21 Farm 3 1.5 25% *25%
F
22
O
23 Harbor 8 1.5 25% 26%
24
O
25
Surveillance 1 1.5 40% 57%
26
27
PR
28 Cu
29
D
30
TE
Farm 17 0.4 11% 12%

31
32
Harbor 29 0.4 20% 26%
EC
33
34
35 Surveillance 17 0.4 12% *12%
R
36
37
R
*RSD is presented due to the variance being lower between days than within days.
38
O
39
40 26
C
41
N
42
U
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
4

Qsae024 METODOS

Uploaded by

Copyright:

Available Formats

You might also like

Qsae024 METODOS

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Qsae024 METODOS

Uploaded by

Copyright:

Available Formats

Page 1 of 33 The Journal of AOAC INTERNATIONAL

Downloaded from https://academic.oup.com/jaoac/advance-article/doi/10.1093/jaoacint/qsae024/7632740 by guest on 26 March 2024

Downloaded from https://academic.oup.com/jaoac/advance-article/doi/10.1093/jaoacint/qsae024/7632740 by guest on 26 March 2024

30 32 Keywords: nanoparticles, SP-ICP-MS, seafood, mussels, validation

Downloaded from https://academic.oup.com/jaoac/advance-article/doi/10.1093/jaoacint/qsae024/7632740 by guest on 26 March 2024

59 tetramethylammonium hydroxide is avoided in many laboratories due to its hazards. Protamex®

Downloaded from https://academic.oup.com/jaoac/advance-article/doi/10.1093/jaoacint/qsae024/7632740 by guest on 26 March 2024

39 85 abundant isobaric or polyatomic interferences using ICP-MS (32).

Downloaded from https://academic.oup.com/jaoac/advance-article/doi/10.1093/jaoacint/qsae024/7632740 by guest on 26 March 2024

Downloaded from https://academic.oup.com/jaoac/advance-article/doi/10.1093/jaoacint/qsae024/7632740 by guest on 26 March 2024

Downloaded from https://academic.oup.com/jaoac/advance-article/doi/10.1093/jaoacint/qsae024/7632740 by guest on 26 March 2024

161 same concentrations as gold.

43 162 Data acquisition, processing and visualization

Downloaded from https://academic.oup.com/jaoac/advance-article/doi/10.1093/jaoacint/qsae024/7632740 by guest on 26 March 2024

186 Validation setup and calculation of validation results

Downloaded from https://academic.oup.com/jaoac/advance-article/doi/10.1093/jaoacint/qsae024/7632740 by guest on 26 March 2024

Downloaded from https://academic.oup.com/jaoac/advance-article/doi/10.1093/jaoacint/qsae024/7632740 by guest on 26 March 2024

Downloaded from https://academic.oup.com/jaoac/advance-article/doi/10.1093/jaoacint/qsae024/7632740 by guest on 26 March 2024

Downloaded from https://academic.oup.com/jaoac/advance-article/doi/10.1093/jaoacint/qsae024/7632740 by guest on 26 March 2024

Downloaded from https://academic.oup.com/jaoac/advance-article/doi/10.1093/jaoacint/qsae024/7632740 by guest on 26 March 2024

41 310 Trueness as recovery

Downloaded from https://academic.oup.com/jaoac/advance-article/doi/10.1093/jaoacint/qsae024/7632740 by guest on 26 March 2024

Downloaded from https://academic.oup.com/jaoac/advance-article/doi/10.1093/jaoacint/qsae024/7632740 by guest on 26 March 2024

Downloaded from https://academic.oup.com/jaoac/advance-article/doi/10.1093/jaoacint/qsae024/7632740 by guest on 26 March 2024

Downloaded from https://academic.oup.com/jaoac/advance-article/doi/10.1093/jaoacint/qsae024/7632740 by guest on 26 March 2024

Downloaded from https://academic.oup.com/jaoac/advance-article/doi/10.1093/jaoacint/qsae024/7632740 by guest on 26 March 2024

Downloaded from https://academic.oup.com/jaoac/advance-article/doi/10.1093/jaoacint/qsae024/7632740 by guest on 26 March 2024

Downloaded from https://academic.oup.com/jaoac/advance-article/doi/10.1093/jaoacint/qsae024/7632740 by guest on 26 March 2024

Downloaded from https://academic.oup.com/jaoac/advance-article/doi/10.1093/jaoacint/qsae024/7632740 by guest on 26 March 2024

Downloaded from https://academic.oup.com/jaoac/advance-article/doi/10.1093/jaoacint/qsae024/7632740 by guest on 26 March 2024

43 545 2677. doi:10.3390/ma12172677

Downloaded from https://academic.oup.com/jaoac/advance-article/doi/10.1093/jaoacint/qsae024/7632740 by guest on 26 March 2024

42 579 Spectrom. 38, 403–413. doi:10.1039/D2JA00342B

Downloaded from https://academic.oup.com/jaoac/advance-article/doi/10.1093/jaoacint/qsae024/7632740 by guest on 26 March 2024

30 605 21. doi:10.1016/j.aca.2019.11.063

39 611 (2020) Anal. Chem. doi:10.1021/acs.analchem.0c01536

42 613 Moreda-Piñeiro, A. (2019) Microchem. J. 148, 652–660. doi:10.1016/j.microc.2019.05.023

Downloaded from https://academic.oup.com/jaoac/advance-article/doi/10.1093/jaoacint/qsae024/7632740 by guest on 26 March 2024

Downloaded from https://academic.oup.com/jaoac/advance-article/doi/10.1093/jaoacint/qsae024/7632740 by guest on 26 March 2024

Downloaded from https://academic.oup.com/jaoac/advance-article/doi/10.1093/jaoacint/qsae024/7632740 by guest on 26 March 2024

Downloaded from https://academic.oup.com/jaoac/advance-article/doi/10.1093/jaoacint/qsae024/7632740 by guest on 26 March 2024

Downloaded from https://academic.oup.com/jaoac/advance-article/doi/10.1093/jaoacint/qsae024/7632740 by guest on 26 March 2024

Downloaded from https://academic.oup.com/jaoac/advance-article/doi/10.1093/jaoacint/qsae024/7632740 by guest on 26 March 2024

31 Method blank 86% 13% 18%

Downloaded from https://academic.oup.com/jaoac/advance-article/doi/10.1093/jaoacint/qsae024/7632740 by guest on 26 March 2024

Downloaded from https://academic.oup.com/jaoac/advance-article/doi/10.1093/jaoacint/qsae024/7632740 by guest on 26 March 2024

Farm 17 0.4 11% 12%

You might also like