Evaluation of Different Parameters in the Extraction of Incurred

Subscriber access provided by NEW YORK UNIV

Article

Evaluation of different parameters in the extraction of incurred pesticides and environmental contaminants in fish Yelena Sapozhnikova, and Steven J. Lehotay J. Agric. Food Chem., Just Accepted Manuscript • DOI: 10.1021/jf506256q • Publication Date (Web): 16 Feb 2015 Downloaded from http://pubs.acs.org on February 18, 2015

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.

Journal of Agricultural and Food 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 16

Journal of Agricultural and Food Chemistry

1 Evaluation of different parameters in the extraction of incurred pesticides and environmental contaminants in fish Yelena Sapozhnikova and Steven J. Lehotay*

US Department of Agriculture, Agricultural Research Service, Eastern Regional Research Center, 600 East Mermaid Lane, Wyndmoor, PA 19038, USA. *

Corresponding author: phone: 215-233-6433, fax: 215-233-6642 E-mail address: [email protected]

Disclaimer Mention of brand or firm names does not constitute an endorsement by the U.S. Department of Agriculture above others of a similar nature not mentioned.

ACS Paragon Plus Environment


Page 2 of 16

2 Abstract 1

Sample processing is often ignored during analytical method development and

2

validation, but accurate results for real samples depend on all aspects of the analytical

3

process. Also, validation is often conducted only using spiked samples, but extraction yields

4

may be lower in incurred samples. In this study, different variables in extraction for incurred

5

pesticides and environmental contaminants in fish were investigated. Among 207 analytes

6

screened using low-pressure gas chromatography–tandem mass spectrometry, consisting of

7

150 pesticides, 15 polycyclic aromatic hydrocarbons (PAHs), 14 polychlorinated biphenyls

8

(PCBs), 6 polybrominated diphenyl ethers (PBDEs), and 22 other flame retardants (FRs), 35

9

(16 pesticides, 9 PCBs, 5 PBDEs, and 5 PAHs) were identified for quantification in samples

10

of salmon, croaker, and NIST Standard Reference Material 1947 (Lake Michigan Fish

11

Tissue). Extraction efficiencies using different extraction devices (blending, vortexing, and

12

vibrating) vs. time, sample size, and sample/solvent ratio were determined. In comparison to

13

blending results, use of a pulsed-vortexer for 1 min with 1/1 (g/mL) sample/acetonitrile ratio

14

was generally sufficient to extract the incurred contaminants in the homogenized fish tissues.

15

Conversely, extraction with a prototype vibration shaker often took 60 min to achieve 100%

16

extraction efficiency. A main conclusion from this study is that accurate results for real

17

samples can be obtained using batch extraction with a pulsed-vortexer in a simple and

18

efficient method that achieves high sample throughput.

19 20

Keywords: extractability, sample preparation, pesticides, environmental contaminants, fish,

21

residues analysis

22 23 24 25

Introduction Prior to obtaining regulatory approval of a new pesticide application on crops, the

26

registrant must usually demonstrate that the proposed regulatory method of analysis achieves

27

total extractability of the defined pesticide residues from the crop and soil.1 Liquid

28

scintillation counting of sample extracts containing radioactive isotopes of the pesticide

29

applied in the field is typically conducted to show if the method meets this total extractability

30

requirement.2 However, the approved regulatory method is rarely used in practice for

31

monitoring of pesticide residues because it is typically designed only for a single pesticide

32

(including defined metabolites), and it is often unwieldy to perform. Instead, multi-class

33

multiresidue monitoring methods, such as QuEChERS (quick, easy, cheap, effective, rugged,


Page 3 of 16


3 34

and safe),3 are employed to provide wide analytical scope with high sample throughput at an

35

affordable cost. In addition, official methods tested for total extractability exist only for

36

registered pesticide/crop combinations, not for unapproved pesticides or environmental

37

contaminants that must also be routinely monitored for food safety, risk assessment, and a

38

range of other purposes.

39

However, radiotracing to determine total extractability of residues from a diverse

40

range of samples is not feasible for many chemicals of interest. Method development and

41

validation studies involving multiresidue analysis of contaminants in food and environmental

42

matrices often only involve laboratory-fortified (spiked) samples to assess suitable method

43

performance. Yet, it is common knowledge that analytes are more easily extracted from

44

laboratory-spiked tissues than from real-world incurred samples, which may have stronger

45

analyte-matrix inter-actions to be overcome. Furthermore, normalization of incurred analytes

46

to a spiked internal standard does not compensate for lower extraction efficiency if incurred

47

recoveries are different from those for spiked samples.

48

In a search of “extractability” in Scopus, hundreds of papers in the scientific literature

49

were found on the topic, but nearly all of them concerned metals, soils, and food composition

50

studies. Only one paper addressed extractability of chemical contaminants from fish,

51

describing the effect of freeze drying and how re-hydration of samples improves extraction of

52

incurred residues.4 It is a well-known phenomenon and standard practice to add water to dry

53

samples in pesticide residue analysis.5 For example, Cajka et al. (2012) demonstrated that

54

underestimation of various pesticides concentrations in incurred tea samples were 25-74%

55

when comparing two methods with different degree of extractability.6 Also, Anastassiades et

56

al. (2013) extensively studied extractability of multiple pesticides from fruits and vegetables

57

by assessing extraction time, temperature, and shaking intensity in QuEChERS.7

58

After method implementation, quality control samples to check ongoing method

59

performance are also typically spiked with the analytes, not incurred. 8 Even in the case of

60

proficiency testing, spiked samples are usually sent to the laboratories for evaluation, and

61

even if incurred samples are used, the true analyte concentration is seldom known for

62

comparison with the experimental results. Thus, it is conceivable that reported concentrations

63

in real samples are lower than actual because total extractabilities of analyte-matrix

64

combinations are rarely tested for common multiresidue monitoring methods.

65

Ideally, standard reference materials (SRMs) 9, 10 would be used to determine total

66

extractability, but none are available for many common contaminants and food sample types.

67

Despite the known concerns related to spiked-only samples, few chemists validate their



Page 4 of 16

4 68

methods using incurred samples 6, 11, 12 because it is often difficult to obtain appropriate

69

samples, and the true concentrations still remain unknown. When possible, a bridging study

70

should be conducted in which results from new methods of extraction should be compared

71

with reference methods in side-by-side analysis of the same sample.13 However, this is not

72

common practice due to the extra time, effort, and cost involved. Thus, many chemists risk

73

inaccurate results in their analyses, even if they follow standard practices for method

74

validation, quality control, and laboratory accreditation (including proficiency testing).

75

The goal of this study was to investigate variables impacting QuEChERS-based

76

extraction yields of incurred pesticides and environmental contaminants in fish samples

77

(white croaker, salmon, and NIST SRM 1947 – Lake Michigan fish tissue). A diverse range

78

of pesticides, polycyclic aromatic hydrocarbons (PAHs), polychlorinated biphenyls (PCBs),

79

polybrominated diphenyl ethers (PBDEs) and flame retardants (FRs) were chosen for analysis

80

using low-pressure gas chromatography – tandem mass spectrometry (LPGC-MS/MS). We

81

sought to compare extraction efficiencies of different sample sizes, sample/solvent ratio, and

82

extraction time using different shakers and a probe blender as extraction devices.

83 84

Materials and Methods

85 86 87

Samples, reagents, and extraction devices Fish samples of salmon and Pacific white croaker (Genyonemus lineatus) were from

88

the National Institute of Standards and Technology (NIST) Marine Environmental Specimen

89

Bank in Charleston, SC, USA. The croaker was collected in 2002 from Palos Verdes Shelf

90

Los Angeles Sanitation District Zone 1 EPA Superfund site in Palos Verdes, CA, USA.

91

Croaker samples were received as a frozen, homogenized powder. The salmon of

92

undetermined type collected from an unknown location was contributed to the Specimen bank

93

by Axys Analytical, Sidney, BC, Canada. The salmon was received frozen as a whole body 81

94

cm fish weighing 5,096 g. It was thawed, filleted into skinless and boneless chunks, refrozen

95

and then homogenized with dry ice using a Robot Coupe (Ridgeland, MS, USA) RSI 2Y1

96

chopper, and stored in glass containers in a -18°C freezer until portions were thawed for

97

analysis. Measured moisture content was 63% and 79% for the salmon and croaker,

98

respectively. Reported average lipid content is 3-4 and 5-10% for croaker and salmon,

99

respectively.14 SRM 1947 (Lake Michigan Fish Tissue) was purchased from NIST

100 101

(Gaithersburg, MD, USA), and contained 73% moisture and 10% fat content. Pesticide standards were from the Environmental Protection Agency’s National


Page 5 of 16


5 102

Pesticide Repository (Fort Meade, MD, USA), Dr. Ehrenstorfer GmbH (Augsburg, Germany),

103

Chemservice (West Chester, PA, USA), and Accustandard, Inc. (New Haven, CT, USA).

104

PCB and PBDE congeners and PAHs were purchased from AccuStandard, and standards of

105

FRs were from AccuStandard and Sigma-Aldrich (St. Louis, MO, USA). For use as internal

106

standards (ISTDs), 13C12-2,2',4,4',5,5'-hexachlorobenzene (13C12 - PCB 153), and 13C12-p,p’-

107

DDE, PAH surrogate cocktail containing acenaphthylene-d8, benzo(a)pyrene-d12,

108

benzo(g,h,i)perylene-d12, fluoranthene-d10, naphthalene-d8, phenanthrene-d10 and pyrene-d10,

109

were purchased from Cambridge Isotope Laboratories (Tewksbury, MA, USA). 5’-fluoro-

110

2,3’,4,4’,5-pentabromodiphenyl ether (FBDE 126) was purchased from AccuStandard, and

111

atrazine-d5 (ethyl-d5) and fenthion-d6 (o,o-dimethyl-d6), were from C/D/N Isotopes (Pointe-

112

Claire, Quebec, Canada). All standards were ≥98% purity.

113

All solvents were HPLC-grade. Acetonitrile (MeCN) was from Fisher Scientific

114

(Pittsburgh, PA, USA) and methanol (MeOH) was from VWR International (Radnor, PA,

115

USA). Anhydrous magnesium sulfate (anh. MgSO4) and sodium chloride (NaCl) reagent-

116

grade salts, and primary secondary amine (PSA) and C18 sorbents were purchased from UCT

117

(Bristol, PA, USA). Deionized water of 18.2Ω was prepared with an E-Pure Model D4641

118

from Barnstead/Thermolyne (Dubuque, IA, USA). Zirconium dioxide based sorbent (Z-Sep)

119

was from Supelco (Bellefonte, PA, USA). Ammonium formate was from Sigma Aldrich.

120

PVDF 0.2 µm filter-vials were from Thomson Instrument Co. (Oceanside, CA, USA).

121

As shown in Figure 1, the extraction devices used in the study consisted of an Ultra-

122

Turrax T25 probe-blending homogenizer (Janke & Kunkel GMBH & Co., IKA Labortechnik,

123

Wilmington, NC, USA), pulsed-vortexer (Glas-Col, Terre Haute, IN, USA), and a prototype

124

vibration shaker (GL Sciences, Eindhoven, The Netherlands).

125 126 127

Sample preparation The initial experiments to determine incurred contaminants in fish samples were

128

performed using a QuEChERS sample preparation method described earlier.9 After addition

129

of internal standards, 10 g of homogenized fish subsamples were extracted by shaking for 1

130

min with 10 mL of MeCN in 3 replicates each of salmon and croaker, and one replicate for

131

SRM 1947. Anh. MgSO4 (4 g) and NaCl (1 g) were added to the initial extracts to induce

132

phase separation, and after centrifugation, a 1 mL portion of the extract was transferred to 2

133

mL mini-centrifuge tubes containing 150 mg MgSO4 and 50 mg Z-Sep for dispersive solid-

134

phase extraction (d-SPE) cleanup. The tubes were centrifuged, and extracts were subjected to

135

LPGC-MS/MS analysis. The final sample preparation method used in experiments described



Page 6 of 16

6 136

below was updated in accordance with subsequent streamlining improvements already

137

reported.15

138

To determine the optimum extraction time and compare extraction devices, croaker

139

and salmon homogenized tissues (10 g) were weighed into 50 mL polypropylene tubes (5 g

140

was used for SRM 1947) and spiked with the internal standards mixture to achieve 100 ng/g

141

for isotopically-labeled pesticides, 50 ng/g for FBDE 126, 20 ng/g for isotopically-labeled

142

PAHs, and 2 ng/g for 13C12 PCB 153. After adding 10 mL of MeCN, the tubes were placed

143

into the pulsed-vortexer and shaken at 80% setting with maximum pulsing for 1, 10, 30, and

144

60 min. Another set of samples was extracted for the same time intervals using the vibration

145

shaker. Four replicates of salmon and croaker, and one replicate of SRM 1947 were

146

conducted for each extraction/time parameter. Another 4 replicates each of croaker, salmon,

147

and SRM 1947 subsamples were extracted using the probe blender for 1 min. After extraction,

148

5 g of ammonium formate was added to induce the phase separation, 16 the samples were

149

shaken for 1 min on the pulsed-vortexer, and centrifuged for 2 min at 3700 rcf. Extracts (0.5

150

mL) were transferred into the shells (bottom half) of 0.2 µm PVDF filter-vials containing 75

151

mg anh. MgSO4, 25 mg PSA, 25 mg C18, and 25 mg Z-Sep sorbents. The filter-vial plungers

152

(top half) were partially depressed, the extracts were shaken by hand for 30 s in an

153

autosampler vial tray, and then the plungers were fully depressed into the filter-vial shells. A

154

commercial product of this filter-vial d-SPE approach was not yet available, thus we had to

155

transfer 200 µL of the filtered extracts to glass autosampler vials containing inserts prior to

156

LPGC-MS/MS analysis.

157

Sample size and solvent ratio were studied in additional experiments using the pulsed-

158

vortexer at 80% setting with maximum pulsing for 10 min and the streamlined filter-vial d-

159

SPE cleanup method. Sample size of 2, 4, and 10 g of each fish were extracted with 2, 4, and

160

10 mL of MeCN for a 1 g/mL sample/solvent ratio, as well as 4 g of each fish was extracted

161

with 8 mL of MeCN to evaluate 0.5 g/mL sample/solvent ratio. These experiments were also

162

conducted using 4 replicates at each set of conditions for salmon and croaker, and 1 replicate

163

for the limited amount of SRM 1947.

164 165 166

Rapid LPGC-MS/MS analysis An Agilent 7890 GC system with multimode inlet (MMI) interfaced to a 7000B triple

167

quadrupole MS/MS was used for analysis. The GC instrument was upgraded with a 220 V

168

rapid heating oven device. A restriction capillary of 5 m × 0.18 mm i.d., HydroGuard (Restek,

169

Bellefonte, PA, USA) coupled to 15 m × 0.53 mm i.d. × 1 µm film thickness Rti-5ms column


Page 7 of 16


7 170

(Restek) was used for low-pressure (vacuum outlet) GC. High purity He was the carrier gas at

171

a constant flow rate of 2 mL/min. Injection volume was 5 µL using the MMI initially set at

172

80°C for 0.31 min with vent at 50 mL/min, then ramped at 420°C/min to 320°C, and held

173

until 9.5 min. The initial oven temperature was at 70°C for 1.5 min, which was ramped at

174

80°C/min to 180°C, and then at 40°C/min to 250°C, followed by 70°C/min to 290°C; where it

175

was held until 9.5 min. The transfer line and ion source were set at 280°C and 320°C,

176

respectively, and the quadrupole was at 150°C. Two MRMs per each analyte were used with a

177

dwell time of 2.5 ms, similar to settings used in previous studies.9, 15, 17

178

The initial MRM method was divided into 28 time segments totaling 207 analytes and

179

12 ISTDs (see Supplemental information Table S1 for the full list, retention times, and MRM

180

transitions). This screen included 150 pesticides, 14 PCB congeners, 15 PAHs, 6 PBDE

181

congeners, and 22 FRs. After non-detected analytes were dropped, the modified method

182

included 35 analytes (16 pesticides, 5 PBDE and 9 PCB congeners, and 5 PAHs) and 9

183

isotopically-labeled ISTDs (atrazine-d5, fenthion-d6, 13C12-p,p’-DDE, pyrene-d10,

184

acenaphthylene-d8, fluoranthene-d10, phenanthrene-d10, 13C12 - PCB 153, and FBDE 126)

185

divided into 8 time segments. The ISTDs were not used to normalize the results for the

186

analytes, but only as quality control spikes to demonstrate ISTD recoveries.

187

Due to matrix effects in GC-MS analyses of certain pesticides,17 we intended to use

188

matrix-matched calibration standards, but we could not find analyte-free salmon or white fish

189

tissues. However, previous studies9,15,17-19 demonstrated that matrix effects in the GC-MS

190

method were insignificant for the nonpolar analytes determined in the samples, which enabled

191

our use of solvent-based calibration standards to obtain accurate quantification of the final list

192

of monitored contaminants. Calibration points were typically set at 0.1, 1, 5, 10, 50, 100, 250

193

and 500 ng/g for the analytes, and 10 times lower levels for PCB congeners. In the case of

194

croaker, additional points were added for PCB congeners at 100 and 150 ng/g. To quantify the

195

high concentrations of DDE in the croaker, a 1000 ng/g calibration standard of o,p’-DDE was

196

added along with 500, 1000, 2500, 5000, 7500, 10,000 and 15,000 ng/g of p,p’-DDE. To

197

avoid detector saturation and still achieve linear calibration in the same injected extract as the

198

ultratrace level contaminants, the least intense MRM transitions were selected for DDE,

199

which avoided re-injection of diluted extracts just for this high concentration analyte.

200 201

Results and Discussion

202 203

Extraction devices



Page 8 of 16

8 204

Figure 1 displays the 3 shaking devices employed in the study. A probe blender, or

205

tissue homogenizer, works by inserting the metal probe mixer into a container containing the

206

extraction solvent surrounding the fish tissue, which is ripped to tiny pieces, thus providing

207

effective access of the solvent to the chemicals within the tissue. The pulsed-vortexer, capable

208

of holding two standard trays with 25 tubes of 50 mL each, operates by intensive whirling

209

action in pulsating fashion to thoroughly mix the pre-homogenized samples with the

210

extraction solvent. The tubes were placed vertically into the vortexer, albeit horizontal

211

placement is also possible. The prototype vibration shaker permitted only horizontal

212

placement of the tubes, and entailed rapid vibrational motion to conduct extraction. This

213

device was provided by the manufacturer to be evaluated for possible commercialization.

214

In terms of sample throughput, the pulsed-vortexer was able to process 50 samples at a

215

time, whereas the prototype vibration shaker fit 6 tubes at a time. The probe blender, however,

216

is not well-suited for use to extract a batch of samples because it requires cleaning of the

217

probe between each sample to avoid cross-contamination. Thorough cleaning involved

218

disassembling the probe, scrubbing with soap and water, rinsing with solvent, drying, and re-

219

assembly. The probe blender procedure has higher potential for introducing chemical

220

interferences or excess liquid (if drying was not complete) into the sample, and always causes

221

partial loss of extract onto the probe surfaces. The probe is also a safety risk and is very loud,

222

and requires an open vessel that lets possible laboratory contaminants into the tube and sprays

223

extract out of the tube. The sequential operation of the probe blender to extract samples is

224

inconvenient, slow, laborious, and potentially hazardous.

225

We previously showed that the use of MeCN as an extraction solvent allows for

226

simultaneous extraction of liphophilic nonpolar and relatively polar analytes (such as PAHs

227

and some FRs) in one extraction step, with added benefit of decreased amount co-extracted fat

228

compared to use of nonpolar solvents such as hexane or ethyl acetate.9 However, even the

229

extensively validated QuEChERS sample preparation method is not exhaustive, and therefore

230

it is not known if total extractability of incurred contaminants is achieved, particularly for

231

lipophilic analytes in fat-containing matrices.

232

The probe blender extraction method serves as a traditional standard procedure, which

233

entails thorough homogenization of tissues with the solvent to yield total extraction. Table S2

234

(Supplemental information) lists average determined concentrations of incurred contaminants

235

in croaker and salmon using 1 min extraction with the probe blender, and concentrations

236

measured in SRM 1947 also using the probe blender vs. the NIST-certified/reference

237

concentrations. %Accuracy of results for SRM 1947 was 69-144% with probe blending


Page 9 of 16


9 238

extraction; thus demonstrating the degree of total extractability of the analytes in fish tissue

239

with 10% lipid content. The determinations were 100% within typical expected experimental

240

error for ultratrace level analyses. This supports the assertion that the probe blender provided

241

100% yield of incurred contaminants in the incurred fish samples. Extraction efficiency of

242

contaminants in croaker (Figure 2) and salmon (Table 1) was expressed as relative recovery

243

(%) for concentrations determined in samples extracted using the pulsed-vortexing and

244

vibration-based shakers normalized to the probe blender results.

245

In terms of extraction efficiency, probe blender extraction for 1 min yielded the same

246

concentrations of nearly all contaminants as maximum concentrations obtained for both types

247

of shakers, which further supports the assertion that total extraction was achieved and true

248

analyte concentrations were determined. As Figure 2 demonstrates, 1 min shaking with the

249

pulsed-vortexer was enough to yield 80-100% relative recovery of incurred contaminants

250

from all classes in croaker tissues. Meanwhile, the vibration shaker yielded 30-50% extraction

251

efficiency in 1 min, and in most cases, required 60 min extraction time to achieve ≈100%

252

relative recoveries. This trend also held true for pesticides, PCBs and PBDEs. With respect to

253

the detected PAHs (anthracene and fluorene), Figure 2 shows that 10 min extraction with the

254

vibration shaker was enough to yield 100% relative recovery, although 100% relative

282

recoveries (high bias) and worse precision in the results were probably due to greater

283

measurement uncertainty at low analyte concentrations, especially in fatty samples.

284

In our previous study,18 we observed negative correlations between recoveries of

285

PBDE congeners and their log Kow values. Pearson correlation coefficients were from (-0.95)

286

to (-0.99), indicating that increased log Kow caused decreased recoveries from lipid-containing

287

samples. PBDE 153 has the highest log Kow value of 7.9 compared to the other congeners

288

determined in SRM 1947. PBDE 153 required 60 min extraction time with the vibration

289

shaker to reach the same yield as 1 min extraction with the pulsed-vortexer or probe blender

290

(see Supplemental Figure S5). Clearly, the pulsed-vortexer performed well to fully extract

291

even the most lipophilic analytes in the fish samples containing up to 10% fat, and the

292

prototype vibrational shaker did not adequately meet the need for rapid total extraction.

293 294 295

Sample size and solvent ratio Another aspect of the study was to determine the effect of sample size and

296

sample/solvent ratio on the results for the incurred fish tissues. Smaller subsample size often

297

translates into faster, easier, and less wasteful methods, provided that the subsample

298

adequately represents the original sample. We used the pulsed-vortexer with 10 min

299

extraction time in the experiments to assess subsample sizes of 2, 4, 10 g extracted with 2, 4,

300

and 10 mL of MeCN, respectively (1 g/mL sample/MeCN ratio), and subsamples of 4 g were

301

extracted with 8 mL of MeCN to evaluate 0.5 g/mL ratios. Due to limited supply and high

302

cost of SRM 1947, only 5/5 g/mL (n = 4) was tested in that case.

303

For salmon and croaker, the determined concentrations from extractions of 2/2, 4/4,

304

and 4/8 g/mL were compared to concentrations measured using 10/10 g/mL, which was

305

already shown to yield total extractability. As Figure 3 demonstrates, there was no difference


Page 11 of 16


11 306

in yields for any combination of sample/solvent ratio. This finding suggests that homogenous

307

subsample size can be minimized as long as it remains representative,21 thereby requiring less

308

organic solvent and leading to less waste. The same conclusions were drawn for SRM 1947 in

309

the experiment, in which %accuracy of determined vs. certified concentrations did not depend

310

on the subsample amount or sample/solvent ratio. This finding exposes another disadvantage

311

of extraction using the probe blender, which needs at least 5 mL of solvent to work properly.

312

Fussell et al. (2007) evaluated homogeneity of analytical samples homogenized

313

cryogenically vs. room temperature, and they concluded that smaller portions of analytical

314

sample can be representative when using cryogenic homogenization for certain matrices.2 In

315

the case of sample/solvent ratios in QuEChERS, Chamkasem et al. (2013) investigated

316

sample/solvent ratios of 3/10, 3/15, 3/20, 3/25 and 3/30 g/mL MeCN for the analysis of

317

lipophilic pesticides in avocado (≈15% fat),22 and they concluded that 3/25 g/mL was

318

sufficient to yield 100% recoveries from fortified samples. In our study, we did not observe a

319

difference in extraction efficiencies for the same nonpolar pesticides using 1/1 or 1/2 g/mL

320

sample/MeCN ratios in incurred fish. Perhaps the higher fat content in avocado and fat

321

composition differences (salmon contains ≈3.5% of monounsaturated fat vs. ≈9.5% in

322

avocado) affected these findings.

323

In conclusion, we compared extractabilities of 35 incurred contaminants in 3 different

324

types of fish tissues using 3 extraction devices (vibration shaker, pulsed-vortexer, and probe

325

blender) for different lengths time, subsample size, and solvent ratio. Our results showed that

326

1 min extraction with the pulsed-vortexing shaker was sufficient for extraction of many

327

incurred contaminants in the pre-homogenized fish tissues, but we chose to use 10 min

328

extractions as a precaution, which still attains 0.2 min/sample equivalent for a batch of 50

329

samples. The probe blender yielded similar results in 1 min, but cannot come close to

330

approaching the high sample throughput, ease, convenience, safety, versatility, and other

331

practical benefits of the pulsed-vortexer. Extraction with the prototype vibration shaker often

332

took 60 min to achieve total extractability, and the device is not likely to be commercialized.

333

In other respects, 1 g/mL sample/MeCN was sufficient to achieve full extraction, and 2 g

334

homogenized subsample gave equivalent results as 4 and 10 g subsamples of the

335

cryogenically-homogenized fish tissues. This study demonstrated for real-world fish samples

336

up to 10% fat content that a simple QuEChERS method using the pulsed-vortexer for batch

337

extraction of 2 g subsamples with 2 mL MeCN, followed by a convenient filter-vial d-SPE

338

cleanup, provides complete extraction and high sample throughput. The use of LPGC-MS/MS

339

permitted 10 min quantitative and qualitative analysis of 207 targeted contaminants



Page 12 of 16

12 340

(pesticides, PCBs, PAHs, PBDEs, and other FRs) at ultratrace concentrations in fish and

341

similar sample types.

342 343

Acknowledgments

344

The authors thank Rebecca Pugh of NIST Marine Environmental Specimen Bank for

345

providing the croaker and salmon samples, Sam Ellis of Thomson Instrument Co. for d-SPE

346

filter-vials, and Mitsuhiro Kurano and Nikolay Gribov of GL Sciences for the prototype

347

vibration shaker. Technical assistance of Tawana Simons and Alan Lightfield is greatly

348

appreciated.

349 References (1) Federal Register 40 CFR Part 158 - Data Requirements for Pesticides, 860. 1340 Residue Analytical Method. (2) Fussell, R. J.; Hetmanski, M. T.; Macarthur, R.; Findlay, D.; Smith, F.; Ambrus, A.; Brodesser, P. J. Measurement uncertainty associated with sample processing of oranges and tomatoes for pesticide residue analysis. J. Agr. Food Chem. 2007, 55, 1062-1070. (3) Anastassiades, M.; Lehotay, S. J.; Stajnbaher, D.; Schenck, F. J. Fast and easy multiresidue method employing acetonitrile extraction/partitioning and "dispersive solidphase extraction" for the determination of pesticide residues in produce. J. of AOAC Int. 2003, 86, 412-431. (4) Zhang, Y. Y.; Lin, N.; Su, S.; Shen, G. F.; Chen, Y. C.; Yang, C. L.; Li, W.; Shen, H. Z.; Huang, Y.; Chen, H.; Wang, X. L.; Liu, W. X.; Tao, S. Freeze drying reduces the extractability of organochlorine pesticides in fish muscle tissue by microwave-assisted method. Environ. Pollut. 2014, 191, 250-252. (5) Pesticide Analytical Manual. 3rd ed.; FDA, U.S. Department of Health and Human Services, Public Health Service: 1994; Vol. I, Multiresidue Methods. (6) Cajka, T.; Sandy, C.; Bachanova, V.; Drabova, L.; Kalachova, K.; Pulkrabova, J.; Hajslova, J. Streamlining sample preparation and gas chromatography-tandem mass spectrometry analysis of multiple pesticide residues in tea. Anal. Chim. Acta 2012, 743, 5160. (7) Anastassiades, M.; Barth, A.; Dork, D.; Steffens, T., Studies on the extractability of incurred pesticide residues. In North America Chemical Residue Workshop, St Pete Beach, FL, USA, 2013. (8) SANCO/12571/2013. www.eurlpesticides.eu/library/docs/allcrl/AqcGuidance_Sanco_2013_12571.pdf. (9) Sapozhnikova, Y.; Lehotay, S. J. Multi-class, multi-residue analysis of pesticides, polychlorinated biphenyls, polycyclic aromatic hydrocarbons, polybrominated diphenyl ethers and novel flame retardants in fish using fast, low-pressure gas chromatography–tandem mass spectrometry. Anal. Chim. Acta 2013, 758, 80-92. (10) Kalachova, K.; Pulkrabova, J.; Drabova, L.; Cajka, T.; Kocourek, V.; Hajslova, J. Simplified and rapid determination of polychlorinated biphenyls, polybrominated diphenyl ethers, and polycyclic aromatic hydrocarbons in fish and shrimps integrated into a single method. Anal Chim Acta 2011, 707, 84-91. (11) Li, H.; Kijak, P. J.; Turnipseed, S. B.; Cui, W. Analysis of veterinary drug residues in


Page 13 of 16


13 shrimp: A multi-class method by liquid chromatography-quadrupole ion trap mass spectrometry. J. Chromatogr. B 2006, 836, 22-38. (12) Guillet, V.; Fave, C.; Montury, M. Microwave/SPME method to quantify pesticide residues in tomato fruits. J. Environ. Sci. Health B 2009, 44, 415-22. (13) Feng, S. X.; Chiesa, O. A.; Kijak, P.; Chattopadhaya, C.; Lancaster, V.; Smith, E. A.; Girard, L.; Sklenka, S.; Li, H. Determination of ceftiofur metabolite desfuroylceftiofur cysteine disulfide in bovine tissues using liquid chromatography-tandem mass spectrometry as a surrogate marker residue for ceftiofur. J. Agr. Food Chem. 2014, 62, 5011-5019. (14) USDA, National Nutrient Database for Standard Reference. http://ndb.nal.usda.gov. (15) Han, L. J.; Sapozhnikova, Y.; Lehotay, S. J. Streamlined sample cleanup using combined dispersive solid-phase extraction and in-vial filtration for analysis of pesticides and environmental pollutants in shrimp. Anal. Chim. Acta 2014, 827, 40-46. (16) González-Curbelo, M. A.; Lehotay, S. J.; Hernández-Borges, J.; Rodríguez-Delgado, M. A. Use of ammonium formate in QuEChERS for high-throughput analysis of pesticides in food by fast, low-pressure gas chromatography and liquid chromatography tandem mass spectrometry. J. Chromatogr. A 2014, 1358, 75-84. (17) Kwon, H.; Lehotay, S. J.; Geis-Asteggiante, L. Variability of matrix effects in liquid and gas chromatography-mass spectrometry analysis of pesticide residues after QuEChERS sample preparation of different food crops. J. Chromatogr. A 2012, 1270, 235-245. (18) Sapozhnikova, Y. Rapid sample preparation and fast GC-MS/MS for the analysis of pesticides and environmental contaminants in fish. LCGC North America, 2014, 32, 878-886. (19) Sapozhnikova, Y. Evaluation of low-pressure gas chromatography-tandem mass spectrometry method for the analysis of >140 pesticides in fish. J. Agric. Food Chem. 2014 62, 3684-3689. (20) Venkatesan, M. I.; Merino, O.; Baek, J.; Northrup, T.; Sheng, Y.; Shisko, J. Trace organic contaminants and their sources in surface sediments of Santa Monica Bay, California, USA. Mar. Environ. Res. 2010, 69, 350-362. (21) Omeroglu, P. Y.; Ambrus, A.; Boyacioglu, D. Estimation of sample processing uncertainty of large-size crops in pesticide residue analysis. Food Anal. Method 2013, 6, 238247. (22) Chamkasem, N.; Ollis, L. W.; Harmon, T.; Lee, S.; Mercer, G. Analysis of 136 pesticides in avocado using a modified QuEChERS method with LC-MS/MS and GCMS/MS. J. Agr. Food Chem. 2013, 61, 2315-2329.



Page 14 of 16

14 Table 1. Extraction efficiency (relative recovery ± standard deviation, %) for PCBs, pesticides, and PAHs in salmon vs. concentration determined in samples extracted for 1 min with the probe blender (100%).

Analytes a 3 PCBs

7 Pesticides

5 PAHs

a

Time (min) 1 10 30 60 1 10 30 60 1 10 30 60

Vortexing (%) 74 ± 11 153 ± 31 129 ± 32 132 ± 18 79 ± 9 99 ± 14 136 ± 19 84 ± 10 100 ± 9 128 ± 12 125 ± 11 127 ± 11

Vibrating (%) 75 ± 13 64 ± 14 72 ± 15 88 ± 14 62 ± 7 67 ± 10 83 ± 10 70 ± 9 75 ± 6 94 ± 16 133 ± 12 145 ± 13

See Table S2 for the specific analytes determined.

Figure 1. The 3 types of extraction devices compared in the study.


Blending (ng/g) 6.68 ± 0.82

25.5 ± 2.7

3.71 ± 0.27

Page 15 of 16


15 Figure 2. Extraction yields of contaminants in croaker: sum of 9 PCB congeners, sum of 5 PBDE congeners, sum of 5 DDT pesticides, and sum of PAHs (fluorene and anthracene); the pulsed-vortexer and vibrating shaker results are normalized to the probe blender results. Error bars represent standard deviation (n=4).



Page 16 of 16

16 Figure 3. Average recoveries of incurred contaminants from croaker (23 analytes) and salmon (13 analytes), normalized to concentrations measured in samples using 10 g/10 mL (100%) and accuracy of measurements for SRM 1947 vs. certified concentrations (27 analytes) for different sample/solvent ratios. See Table S2 for the specific analytes determined. Error bars represent standard deviation (n=4). 140%

2 g + 2 mL 4 g + 8 mL

120%

4 g + 4 mL 10 g + 10 mL

100% 80% 60% 40% 20% 0% Croaker

Salmon

SRM 1947

Probe Blender

Vortexing

TOC Graphic:

120

Vibrating

100

% Accuracy

80 60 40 20 0 1 min

10 min

30 min

60 min

%Accuracy of PBDEs (sum of 5 congeners) measurements in SRM 1947


Evaluation of Different Parameters in the Extraction of Incurred

Recommend Documents