Accepted Manuscript

Title: Single-step extraction and cleanup of bisphenol A insoft drinks by hemimicellar magnetic SPE prior to liquidchromatography/tandem mass spectrometry

Author: <ce:author id="aut0005"> Samaneh RaoufYazdinezhad<ce:author id="aut0010"> AnaBallesteros-Gomez<ce:author id="aut0015"> LoretoLunar<ce:author id="aut0020"> Soledad Rubio

PII: S0003-2670(13)00365-6DOI: http://dx.doi.org/doi:10.1016/j.aca.2013.03.025Reference: ACA 232454

To appear in: Analytica Chimica Acta

Received date: 16-1-2013Revised date: 6-3-2013Accepted date: 9-3-2013

Please cite this article as: S.R. Yazdinezhad, A. Ballesteros-Gomez, L. Lunarb, S.Rubio, Single-step extraction and cleanup of bisphenol A in soft drinks by hemimicellarmagnetic SPE prior to liquid chromatography/tandem mass spectrometry, AnalyticaChimica Acta (2013), http://dx.doi.org/10.1016/j.aca.2013.03.025

This is a PDF file of an unedited manuscript that has been accepted for publication.As a service to our customers we are providing this early version of the manuscript.The manuscript will undergo copyediting, typesetting, and review of the resulting proofbefore it is published in its final form. Please note that during the production processerrors may be discovered which could affect the content, and all legal disclaimers thatapply to the journal pertain.

Page 1 of 24

Accep

ted

Man

uscr

ipt

1

1

2

3

4

5

6

7

8

Title: Single-step extraction and cleanup of bisphenol A in soft drinks by hemimicellar magnetic 9

SPE prior to liquid chromatography/tandem mass spectrometry10

11

12

13

Authors: Samaneh Raouf Yazdinezhada, Ana Ballesteros-Gómezb, Loreto Lunarb and Soledad 14

Rubiob*15

16

17

18

Address: a Department of Chemistry, Faculty of Sciences, Ferdowsi University of Mashhad, 19

Mashhad, Iran 20b Department of Analytical Chemistry. Institute of Fine Chemistry and Nanochemistry. 21

Edificio Anexo Marie Curie Campus de Rabanales. University of Córdoba. Spain 22

23

24

25

*corresponding author26

E-mail: qa1rubrs@uco.es ; Phone: 34-957-218644; Fax: 34-957-21864427

28

29

30

31

Page 2 of 24

Accep

ted

Man

uscr

ipt

2

Abstract32

33

Hemimicelles of tetradecanoate chemisorbed onto magnetic nanoparticles (MNPs) are here 34

proposed as a sorbent for the single-step extraction and cleanup of bisphenol A (BPA) in soft 35

drinks. The purpose of this work was to develop a simple, rapid and low-cost sample treatment 36

suitable to assess the human exposure to BPA from this type of high consumption food. The 37

nanoparticles were easily coated by mixing commercially available magnetite of 20-30 nm mean 38

particle diameter with tetradecanoate at 85 ºC for 30 min. The extraction/cleanup procedure 39

involved stirring the samples (3 mL) with 200 mg of tetradecanoate-coated MNPs for 20 min, 40

isolating the sorbent with a Nd-Fe-B magnet and eluting BPA with methanol. The extraction 41

efficiency was not influenced by salt concentrations up to 1 M and pH values over the range 4-9. 42

No cleanup of the extracts was needed, and the method proved matrix-independent. The extracts 43

were analyzed by liquid chromatography, electrospray ionization tandem mass spectrometry. 44

Quantitation was performed by internal standard calibration using BPA-13C12. The limit of 45

quantitation obtained for the method, 0.03 ng mL-1, was below the usual range of concentrations 46

reported for BPA in soft drinks (0.1-3.4 ng mL-1). The proposed method was successfully 47

applied to the determination of BPA in different samples acquired from various supermarkets in 48

southern Spain; the concentrations found ranged from 0.066 to 1.08 ng mL-1. Recoveries from 49

samples spiked with 0.33 ng mL-1 of BPA ranged from 91 to 105% with relative standard 50

deviations from 3 to 8%. 51

52

53

54

55

Keywords: bisphenol A, soft drinks, magnetic SPE, chemisorbed hemimicelles, liquid 56

chromatography, tandem mass spectrometry57

58

59

60

61

62

Page 3 of 24

Accep

ted

Man

uscr

ipt

3

1. Introduction63

64

Bisphenol A (BPA) is a high production volume chemical [1] widely used in the manufacture of 65

polycarbonate polymers and epoxy resins [2]. The food industry uses the former for food 66

packaging and plastic bottles, and the latter as internal protective lining for food and beverage 67

cans, coating on metal lids for glass jars and bottles and surface-coating on drinking water 68

storage tanks and wine vats [3]. Migration of BPA from plastics to food is mainly due to the 69

presence of non-polymerized residues or the hydrolysis of ester bonds, which is accelerated by 70

heat and contact with either acid or basic contents in cans [4]. 71

72

Intake of BPA-contaminated foods constitutes the primary route of human exposure to this 73

chemical [5]. Concentrations reported for BPA range in the interval 0.1-3.4 ng mL-1 and 0.3-458 74

ng g-1 for drink and food, respectively [6]. Estimated intakes of BPA for adults, infants and 75

children are all well below the Tolerable Daily Intake (TDI) of 0.05 mg/kg body weight/day set 76

by the European Food Safety Authority (EFSA) for this substance in 2006. However, new 77

findings from ongoing studies on low dose effects observed in rodents [7-10] have urged EFSA 78

to launch a full re-evaluation on BPA, focusing on human exposure, how and how much is 79

absorbed by the human body and possible low dose effects [11]. Reassessment of human 80

exposure to BPA makes it necessary to take into account all the possible dietary sources and 81

consequently, to analyze a vast array of products and brands. In this context, it is recognized the 82

need for development of simple, high-throughput analytical methods for the determination of 83

BPA in high consumption, low BPA-content food such as drinks. 84

85

Only a few methods sensitive and selective enough to give accurate data of BPA levels in soft 86

drinks and mineral water have been reported [reviewed in 6]. Sample preparation still constitutes 87

the key-step for the determination of BPA in drinks and it is the origin of the main drawbacks in 88

the available methodologies. Solvent extraction and solid phase extraction (SPE) are by far the 89

most used extraction techniques for both isolation of BPA and clean-up of matrix components. 90

Ethyl acetate [12], chloroform [13] and dichloromethane [14] have been used as solvents, the 91

overall consumption ranging from 60 mL [14] to 300 mL [12]. Oasis HLB has been proposed as 92

a sorbent for the SPE of BPA using sample sizes between 50 mL and 2 L [15]. In both cases, 93

Page 4 of 24

Accep

ted

Man

uscr

ipt

4

evaporation of the solvent for further treatment or BPA concentration is necessary, that rendering 94

the respective sample treatments laborious and time-consuming. On-line SPE [16,17] and a 95

number of microextraction techniques have been proposed as good alternatives for beverage96

analysis, namely, stir bar sorptive extraction (SBSE) [18], supramolecular solvent 97

microextraction [6], solid-phase microextraction (SPME) [19-21] and ultrasound-assisted 98

emulsification-microextraction (USAEME) [22]. The need for matrix matched calibration 99

[16,17] and long extraction times (e.g. 3 h for SBSE[18], 40 min for SPME [21]) are some 100

disadvantages of these methodologies.101

102

Magnetic nanoparticles have attracted increasing attention in SPE owing to their unique 103

properties, namely high surface area, small diffusion resistance, easiness of surface modification 104

and easy and fast separation from sample solution by applying an external magnetic field. 105

Surface modification is essential to avoid nanoparticle aggregation and alteration of their 106

magnetic properties, as well as to improve sorbent capacity and selectivity for target analytes. 107

Reported coatings for MNPs include a variety of materials such as carbon, carbon nanotubes, 108

polymers, surfactants, biological active compounds, ionic liquids [23-32], etc. In this context, 109

molecularly imprinted polymer (MIP) coated-MNPs have been proposed for the selective 110

extraction of BPA from packed food [33] and different types of water [34]. 111

112

Among MNPs coatings, surfactant physisorption results in the formation of different surfactant 113

concentration-dependent types of aggregates (e.g. admicelles, hemimicelles and their mixtures) 114

that offer a variety of mechanisms for interaction with solutes. Although outstanding applications 115

have been reported using surfactant-coated MNPs [25,26,30,32], the easy disruption of the 116

physisorbed surfactant during analyte elution produces surfactant-rich extracts. This surfactant 117

can co-elute with the target compounds in LC, that causing interferences in UV, fluorescent and 118

MS detection or incompatibility with the analysis of the sample by gas chromatography or 119

capillary electrophoresis. These shortcomings have been recently overcome by coating MNPs 120

with chemisorbed surfactants [35,36].121

122

In the present study, alkyl carboxylates chemisorbed onto MNPs [35] are proved to serve as a 123

simple-handling, efficient and low time-consuming method for extraction and cleanup of BPA. 124

Page 5 of 24

Accep

ted

Man

uscr

ipt

5

This sample treatment, combined with LC separation and electrospray ionization tandem mass 125

spectrometric detection, provides a sensitive and reliable method for the determination of BPA in 126

drinks. 127

128

2. Experimental129

2.1. Chemical and apparatus130

All chemicals were of analytical reagent-grade and were used as supplied. Bisphenol A and iron 131

(II, III) oxide nanopowder with mean diameter of 20-30 nm were purchased from Aldrich (St. 132

Louis, MO, USA). Sodium n-tetradecanoate and n-octadecanoate were obtained from Sigma (St. 133

Louis, MO, USA). Sodium chloride, and LC grade methanol and acetonitrile were supplied by 134

Sigma-Aldrich (France), while sodium hydroxide was purchased from Merck (Darmstadt, 135

Germany). Bisphenol A (ring-13C12), used as internal standard (IS) was supplied by IS 136

Cambridge Isotope Laboratories (Andover, USA) as a solution of 100 mg L-1 in acetonitrile. BPA 137

and BPA-13C12 standard stock solutions (100 ng mL-1) were prepared in methanol and stored at 4 138

°C. Ultrahigh-quality water was obtained from a Milli-Q water purification system (Millipore, 139

Madrid, Spain).140

A PrecisTerm water bath from P Selecta (Barcelona, Spain) was used for synthesizing magnetic 141

nanoparticles. Glass vials (i.d. 25 mm) with metal caps were employed for extraction of samples 142

that were stirred in a multi-position vortex VorTemp 1550 from Labnet International (Edison, 143

NJ, USA). An Nd-Fe-B magnet (viz. a disk 2 cm in diameter and 1 cm thick from Farplas, S.L., 144

Barcelona, Spain) was used for isolation of MNPs after extraction. A high speed brushless 145

centrifuge MPW-350 R from MPW Med-Instruments (Warschow, Poland) was used for 146

separation of fruit pulps in samples containing lemon or orange extract. Some precautionary 147

measures were established to prevent BPA background contamination coming LC equipment 148

polymers, solvents or labware. Milli-Q water was filtered through styrene DVB (SDB-XC) disks 149

of 47 mm from Empore (3M, St Paul, Minnesota, USA) and glassware and eppendorf microtubes 150

were rinsed with methanol several times before their use. Contamination blanks were routinely 151

run with each batch of samples and were always below the quantification limit.152

153

Page 6 of 24

Accep

ted

Man

uscr

ipt

6

2.2. Synthesis of tetradecanoate-coated magnetic nanoparticles154

The tetradecanoate-coated MNPs (C14-MNPs) were synthesized according to our previous report 155

[35]. Briefly, 200 mg of MNPs were mixed with 50 mg of tetradecanoic acid sodium salt in 25 156

mL deionized Millipore water. The mixture was manually homogenized and placed in a water 157

bath at 85 ºC for 30 min. Following chemisorption, C14-MNPs were separated from the 158

equilibrium solution using a Nd-Fe-B magnet and washed three times with 4 mL of methanol to 159

remove physisorbed surfactant. The SPE material thus obtained was stored in a glass container at 160

4ºC until use. For synthesizing more amount of adsorbent, the reagents ratio can be increased at 161

will. The synthesis of octadecanoate-coated MNPs, also tested as a sorbent, was performed 162

similarly, but an amount of 62 mg of octadecanoic acid sodium salt was used instead.163

2.3. Determination of bisphenol A in soft drinks and drinking water164

2.3.1. Sample collection and preparation165

Nine beverages including different types of canned soft drinks were purchased from several166

supermarkets in Cordoba (Spain). They were stored unopened at room temperature until analysis. 167

Carbonated samples were degassed by sonication for 30 min. Samples containing orange or 168

lemon, were centrifuged for 15min at 15000 rpm to separate the fruit pulps. The pH of samples 169

ranged in the interval 2.5-3.5 and it was adjusted in the interval 4-9 with some drops of sodium 170

hydroxide (50%, w/v). 171

2.3.2. Magnetic solid phase extraction of BPA172

A volume of sample of 3 mL, at pH in the interval 4-9 and containing 1 ng of BPA-13C12 (10 µL 173

of the 100 ng mL-1 stock solution), and 200 mg of C14-MNPs were mixed in a glass vial and 174

vortex stirred for 20 min at 800 rpm. Then, the C14-MNPs were separated from the sample 175

solution by placing an Nd-Fe-B magnet in the bottom of the vial and pouring the solution away. 176

Next, the MNPs were washed with about 1 ml deionized water and the solution discarded. A177

mild nitrogen stream (~1bar for 2 min) was used for complete removal of the washing solution178

remaining on the MNPs. Finally, the analyte was eluted by vortex stirring (800 rpm) the dry179

Page 7 of 24

Accep

ted

Man

uscr

ipt

7

MNPs with 2 mL of methanol for 10 min. The magnet was again attached to the vial and a 180

desired portion of supernatant was transferred to LC vials for further analysis.181

2.3.3. Quantification of BPA by LC(ESI-) QQQ-MS2182

The separation and quantitation of BPA was accomplished by using a hybrid triple 183

quadrupole/linear ion trap (Applied Biosystems MSD Sciex 4000QTRAP, Foster City, CA, 184

USA) coupled to a liquid chromatograph (Agilent HP 1200 series, Palo Alto, CA, USA) with a 185

TurboIonSpray (TIS) interface. All data were acquired and processed using the Analyst 1.5.1 186

Software. A Symmetry Shield RP18 column (50 mm×2.1 mm ID, 3.5 µm, Waters, Ireland) was 187

used for LC separation. It was preceded by a C18 Guard Cartridge (Ascentis C18 Supelguard, 20 188

mm × 4 mm ID, 3 μm; Sigma-Aldrich). A Symmetry C18 Column (75 mm × 4.6 mm ID, 3.5 189

µm, Waters) was placed in the line of the mobile phase before the injection valve for retention of 190

the potential BPA leached from the plastic components of the LC system, so that it eluted later in 191

the chromatogram. The retention time for BPA from plastic components and the sample were 192

15.7 min and 12.6 min, respectively. The column oven was set at 40 ◦C, the flow rate was 400 193

µLmin−1, and the injection volume was 15 µL. The mobile phase consisted of methanol and 194

water. The gradient elution program used was 30% methanol during the first 2 min, ramp from 195

30% to 88.5% methanol over the next 13 min and then 100% methanol for 5 min. 196

Reconditioning the column took about 5 min. The eluates from the analytical column were 197

diverted to waste using the switching valve in the intervals 0-9 min and 14.50-25 min in order to 198

avoid the entrance of matrix components in the mass spectrometer. The mass spectrometer was 199

operated in ESI negative polarity and multiple-reaction monitoring (MRM) mode. Identification 200

and quantitation of BPA were performed by recording the transitions (m/z) 227→212 (quantifier 201

ion) and 227→133 (qualifier ion). BPA-13C12 used as an internal standard was monitored at (m/z) 202

239→224 and 239→139 transitions. The dwell time was set up at 150 ms. The TIS source values 203

were adjusted as follows: probe vertical y-axis position, 2 mm; probe horizontal y-axis position, 204

6 mm; curtain gas (N2), 50 psig; ion source gas 1 (nebulizer gas), 50 psig; ion source gas 2 (turbo 205

gas), 30 psig; temperature of the turbo gas, 650ºC; ion spray voltage: -4500 V; entrance 206

potential, -10 V; and declustering potential, -100 and -90 V for BPA and BPA-13C12, 207

respectively. Parameter values for the analyzer were as follows: 1.0 unit resolution for the first 208

and third quadrupoles; collision gas, 3.0x10-5 Torr; collision energy, -30, -36, -26 and -38 V; and 209

Page 8 of 24

Accep

ted

Man

uscr

ipt

8

collision cell exit potential, -15, -7, -11 and -15 V for the transitions 227→212, 227→133, 210

239→224 and 239→139, respectively. 211

212

3. Results and discussion213

3.1. Optimization of the extraction of BPA with alkylcarboxylate-coated MNPs214

Hemimicellar magnetic SPE was selected as a convenient sorbent and format to efficiently 215

extract BPA and speed up sample treatment, respectively. Coating of MNPs with alkyl 216

carboxylates allowed the formation of stable hemimicellar aggregates, via the formation of 217

bidentate mononuclear complexes between carboxylate groups and iron atoms [37], that avoiding 218

their disruption during BPA elution and permitting the obtainment of surfactant free extracts. 219

Optimization of the magnetic SPE was accomplished by conducting triplicate tests that involved 220

extracting 2 mL of BPA (10 ng mL-1) aqueous standard solutions with coated MNPs under a 221

variety of experimental conditions. Because MS afforded instrumental quantitation limits low 222

enough to determine BPA in soft drinks without the need for preconcentration, the aim was to 223

develop an extraction/cleanup method efficient, simple and rapid, able to deal with a high 224

number of samples. Adsorption was investigated by measuring the remaining concentration of 225

BPA under equilibrium conditions. 226

Two alkyl carboxylates, namely tetradecanoate and octadecanoate, were assessed as coatings of 227

MNPs. After sorbent synthesis, magnetite was coated with both chemisorbed (~75%) and 228

physisorbed (~25%) surfactant. Removal of the latter was carried out by washing with methanol. 229

The maximum coverage of chemisorbed surfactant on MNPs was previously found to be ~0.35 230

mmol g-1, irrespective of the chain length of the surfactant [35]. It was calculated from the value 231

of the surface area of the magnetite (~60 m2 g−1) and the maximum amount of chemisorbed 232

surfactant obtained from the chemisorption isotherm. The chemisorbed surfactant was calculated 233

by measuring the remaining amount of surfactant in the equilibrium solution, after synthesis and 234

the subsequent washing steps, by LC-UV and subtracting it from the total amount of surfactant 235

used for synthesis. Maximal coverage required the use of at least 1 mmol of surfactant per gram 236

of magnetite in the synthetic procedure.237

Page 9 of 24

Accep

ted

Man

uscr

ipt

9

238

The coating of magnetite with single layers of alkyl carboxylates occurs via the formation of 239

bidentate mononuclear complexes between carboxylate groups and iron atoms. So dispersion 240

interactions between the aromatic rings of BPA and the alkyl chain of the surfactant were 241

expected to be the main forces driving extraction. Adsorption of BPA on both tetradecanoate-242

and octadecanoate-coated MNPs was similar at the different amounts of sorbents assayed. The 243

former was selected on the basis of its higher solubility in methanol which made easier the 244

removal of the physisorbed surfactant fraction. Figure 1A shows the dependence of the 245

percentage of adsorption of BPA on tetradecanoate- and octadecanoate-coated MNPs as a 246

function of the amount of sorbent. Percentages of adsorption above 90% with relative standard 247

deviations below 3% (n=3) were obtained from 200 mg of sorbent, this being the amount 248

recommended for further studies. 249

Adsorption of BPA exhibited no change with the concentration of NaCl in the range investigated 250

(0-1M), that making the method robust against changes in the ionic strength of the samples. The 251

pH, examined over the range 4-9, had no influence on the adsorption of BPA. More acid or 252

alkaline pHs were avoided to prevent MNPs from oxidation and BPA from ionization, 253

respectively. Regarding the extraction time, the extraction efficiency increased rapidly from 5 to 254

15 min (Figure 1B) and then reached a plateau. The precision raised as the extraction time did 255

(e.g. from RSD = 7.4% at 5 min to 0.09% at 60 min). Percentages of BPA adsorption above 90% 256

and RSDs below 3% were obtained at extraction times of 20 min and this value was selected as 257

optimal.258

Desorption of BPA from MNPs was investigated with MeOH and ACN using various elution 259

programs that involved solvent volumes from 1 to 3 mL, desorption times from 5 to 30 min and 260

1-3 elution steps. Complete removal of the water adsorbed onto the MNPs was required to avoid 261

a decrease in solvent desorption efficiency and changes in the final volume. Water was displaced 262

from MNPs by passing a nitrogen stream at ~1 bar for ~2 min while they were retained on the 263

magnet. 264

Extraction efficiencies for BPA were 81±4% and 74 ± 6.6% using 2 mL of MeOH and ACN, 265

respectively. The former was selected for further experiments, because of its slightly higher 266

Page 10 of 24

Accep

ted

Man

uscr

ipt

10

desorption ability and precision and the better chromatographic peak shape obtained. Extraction 267

efficiencies decreased by decreasing the volume of MeOH (e.g. 46.4±4.3% for 1 mL of solvent) 268

and they kept constant for higher volumes (e.g. 82.4 ±4.3% for 3mL of MeOH), so 2 ml were 269

selected as optimal. 270

271

The number of steps used in the elution (e.g. 1x2 mL; 2x1mL; 3x0.7 mL) did not affect BPA 272

extraction efficiency but the precision decreased as the number of steps increased (e.g. 273

81.3±3.8%, 80.7±5.5% and 84.2±7.0% for 1, 2 and 3 steps, respectively). BPA desorption 274

reached equilibrium after vortex mixing the MNPs at 800 rpm for 10 min and that was the 275

selected elution time. 276

277

Under the chosen experimental conditions (see section 2.3.2), extraction efficiencies kept 278

constant in the whole range of BPA amount tested (2.5-50 ng mL-1). Because of the low 279

concentration of BPA in drinks (0.1-3.4 ng mL-1) and its relatively high solubility in water (89 280

μg mL-1), the volume of sample was expected to greatly influence extraction efficiency. Figure 281

1C shows the results obtained as the volume of sample increased from 2 to 10 mL; extraction 282

efficiency progressively decreased from 3 mL of sample and it was about 56% at the maximal 283

sample volume tested. A volume of 3 mL was chosen as optimal for extraction and cleanup of 284

BPA. 285

286

We assessed the possibility to concentrate the 2 mL of MeOH to 200 μL under evaporation under 287

a gentle stream of nitrogen. No losses were found to result from this drying operation and this 288

allowed the sensitivity to be increased 10-fold. However, the instrumental quantitation limit 289

afforded by MS (0.04 ng mL-1) is low enough to determine BPA in drinks without prior 290

preconcentration, so we propose dispensing with the evaporation step in order to make the 291

procedure simpler and faster for routine analysis. 292

293

No leaching of tetradecanoate from the MNPs was observed under the recommended extraction 294

conditions, as checked by chromatographic analysis of the extracts. 295

296

Page 11 of 24

Accep

ted

Man

uscr

ipt

11

3.2. Analytical performance297

Internal standard calibration (n=16) was employed for quantification (BPA concentrations 298

calculated from the calibration curve obtained by plotting the ratio of analyte peak area to IS 299

peak area against the analyte concentration). The calibration equation was y=1.17x+0.0487 with 300

correlation coefficient (r) of 0.9993. Calibration standard solutions were performed in methanol 301

in the range 0.04-30 ng mL-1 of BPA, each solution containing 0.5 ng mL-1 of BPA-13C12. Linear 302

regression with a weighing 1/x was selected for quantification. The instrumental limits of 303

detection (LOD) and quantification (LOQ), calculated from blank determinations by using a 304

signal-to-noise ratio of 3 and 10 were 0.01 and 0.04 ng mL-1, respectively, for 15 µL injection 305

volumes. The method LOQ (estimated from the actual method concentration factor) was 0.03 ng 306

mL-1. Consequently, the proposed method affords the quantification of BPA at concentrations 307

below the usual range found for BPA in soft drinks (0.1-3.4 ng mL-1). Figure 2 shows the 308

chromatograms obtained, under optimal magnetic SPE conditions, for 3 mL of aqueous standard 309

solutions containing (A) 0.67 ng mL-1 of BPA and (B) 0.33 ng mL- 1 of BPA-13C12. Under these 310

conditions (section 2.3.2), extraction efficiencies for BPA and BPA-13C12 were similar and kept 311

constant at 81 ± 3 %.312

Because the instrumental sensitivity was high enough to quantify BPA in soft drinks without 313

preconcentration, the need for sample clean-up was investigated. For this purpose, the slope of 314

the calibration curve (n=10) obtained from fortified aqueous standard solutions was compared 315

with those obtained from fortified soft drinks. Concentrations of BPA were in the range 0.04-30 316

ng mL-1. The concentration of BPA-13C12 in all the solutions was 0.5 ng mL-1. Removal of 317

carbonic acid and/or separation of fruit pulps (section 2.3.1) were the only treatment applied to 318

the samples. The difference between the slopes was found to be statistically significant by 319

applying an appropriate Student´s test [38] for the three randomly selected soft drinks. The 320

calculated t-values for samples of lemon tea (4.90), orange carbonated drink (2.84) and energy 321

drink (2.18) were above the critical t-value (2.14) being significance established at the 0.05 level. 322

So, clean-up of samples will be necessary to get accurate results. Using clean-up with 323

hemimicellar MNPs, the difference between the slopes of calibration curves obtained from 324

fortified blank samples and the solvent were not significantly different. 325

Page 12 of 24

Accep

ted

Man

uscr

ipt

12

326

3.3. Analysis of soft drinks327

The suitability of the proposed method for use with real-world samples was established by 328

applying it to the determination of BPA in a variety of soft drinks belonging to different 329

trademarks. Table 1 shows the concentrations found as well as the recoveries obtained after 330

spiking the samples with 0.33 ng mL-1 of BPA. Both the concentration of analyte and recoveries 331

were expressed as the mean value of three independent determinations, besides their 332

corresponding standard deviations. Both blank and BPA-spiked samples were fortified with 0.33 333

ng mL-1 of BPA-13C12 before extraction. As can be seen, relative recoveries changed from 91 to 334

105% with relative standard deviations from 3 to 8%. BPA was present at quantifiable levels in 335

all the analyzed samples, the concentrations ranging from 0.066 to 1.08 ng mL-1, which is in 336

accordance with the values previously reported in the literature [6]. Figure 3 shows the 337

chromatograms obtained from a sport drink (A,B) and a lemon tea (C,D) before (A,C) and after 338

spiking with BPA at 0.33 ng mL- 1 of concentration level (B,D) and applying the proposed 339

hemimicellar magnetic SPE method. As can be seen, no interference from matrix components 340

was detected in the samples. 341

342

4. Conclusion343

A simple, rapid and inexpensive dispersive magnetic SPE has been developed for the extraction 344

and cleanup of BPA in soft drinks. Hemimicelles of tetradecanoate chemisorbed onto MNPs, 345

provide surfactant- and matrix components-free extracts for LC-MS BPA analysis. The 346

extraction procedure is independent of the pH and ionic strength in a wide range, as well as 347

matrix-independent, that making it robust. 348

349

Compared to sample treatments based on liquid-liquid extraction [12, 14, 41,42], the present 350

method is faster and much less solvent consumer (see Table 2). On the other hand, the use of 351

dispersive magnetic SPE simplifies the general procedure of conventional SPE, and as it is 352

Page 13 of 24

Accep

ted

Man

uscr

ipt

13

applied to BPA, it also prevents the contamination arising from the SPE plastic cartridges353

[15,16,39,45-48]. 354

355

The detection system used is essential to get detection limits low enough to quantify the usual 356

low BPA concentrations found in beverages (viz 0.1-3.4 ng mL-1, [6]). Thus, ultraviolet [45,47],357

fluorescence [46] or electrochemical [12] detection are not sensitive enough to cover the whole 358

range of BPA concentrations present in these matrices and the use of MS is mandatory for this 359

purpose (Table 2). Both LC/MS and GC/MS meet the requirements of sensitivity to quantify 360

BPA in beverages, the latter with the disadvantage of requiring derivatization. These techniques 361

feature similar LODs, and solvent evaporation [14] or a high sample size [15] will be always 362

required to achieve LODs below 1 ng L-1. 363

364

365

Acknowledgments366

The authors gratefully acknowledge financial support from Spanish MICINN (Project CTQ2011-367

23849) and from the Andalusian Government (Junta de Andalucía, Spain, Project P09-FQM-368

5151). FEDER also provided additional funding. S. Raouf Yazdinezhad thanks the Ministry of 369

Science, Research and Technology of the Islamic republic of Iran for its financial support for her 370

stay in the University of Córdoba. 371

References372

373

[1] E. Burridge, Eur. Chem. News 17 (2003) 14-20.374

[2] O.C. Dermer, in: J.J. McKelta, G.E.Weismantel (Eds.), Encyclopedia of Chemical Processing 375

and Design, Marcel Dekker, New York, 1999, p. 406.376

[3] Opinion of the Scientific Panel on Food Additives, Flavourings, Processing Aids and 377

Materials in Contact with Food (ACF) on request from the Commission related to 2,2-bis(4-378

hydroxyphenyl)propane (Bisphenol A). Question number EFSA-Q-2005-100, adopted on 29 379

November 2006. The EFSA Journal 428 (2006) 1-75.380

Page 14 of 24

Accep

ted

Man

uscr

ipt

14

[4] A.V. Krishnan, P. Stathis, S.F. Permuth, L. Tokes, D. Feldman, Endocrinology 132 (1993)381

2279-2286.382

[5] Opinion of the Scientific Committee on Food on Bisphenol A, expressed on 17 April 2002, 383

SCF/CS/PM/3936 final, European Commission, Brussels, 2002, available online at 384

http://ec.europa.eu/food/fs/sc/scf/out128_en.pdf385

[6] A. Ballesteros-Gomez, S. Rubio, D. Perez-Bendito, J. Chromatogr. A 1216 (2009) 449-469.386

[7] F.S. Vom Saal, C. Hughes, Environ. Health Perspect. 113 (2005) 926-933.387

[8] B.T. Akingbemi, C.M. Sottas, A.I. Koulova,; G.R. Klinefelter, M.P. Hardy, Endocrinology 388

145 (2004) 592-603. 389

[9] R.T. Zoeller, R. Bansal, C. Parris, Endocrinology 146 (2005) 607-612.390

[10] J.T. Wolstenholme, E.F. Rissman, J.J. Connelly, Horm. Behav. 59 (2011) 296-305.391

[11] Bisphenol A: EFSA launches full re-evaluation focussing on exposure and possible low 392

dose effects. Available at: http://www.efsa.europa.eu/en/press/news/120424.htm393

[12] T. Toyo’oka, Y. Oshige, Anal. Sci. 16 (2000)1071-1076.394

[13] J.A. Brotons, M.F. Olea-Serrano, M. Villalobos, V. Pedraza, N. Olea, Environ. Health 395

Perspect. 103 (1995) 608-612.396

[14] P. Varelis, D. Balafas, J. Chromatogr. A 883 (2000)163-170. 397

[15] B. Shao, H. Han, J. Hu, J. Zhao, G. Wu, Y. Xue, Y. Ma, S. Zhang, Anal. Chim. Acta 530 398

(2005) 245-252.399

[16] H. Gallart-Ayala, E. Moyano, M.T. Galceran, Anal. Chim. Acta 683 (2011) 227-233.400

[17] R. Braunrath, M. Cichna, J. Chromatogr. A 1062 (2005) 189-198.401

[18] J.I. Cacho, N. Campillo, P. Viñas, M. Hernández-Córdoba, J. Chromatogr. A 1247 (2012) 402

146-153.403

[19] J. Salafranca, R. Batlle, C. Nerín, J. Chromatogr. A 864 (1999) 137-144.404

Page 15 of 24

Accep

ted

Man

uscr

ipt

15

[20] C. Chang, C. Chou, M. Lee, Anal. Chim. Acta 539 (2005) 41-47.405

[21] P. Viñas, N. Campillo, N. Martínez-Castillo, M. Hernández-Córdoba, Anal. Bioanal. Chem. 406

397 (2010) 115-125.407

[22] A.R. Fontana, M. Munoz de Toro, J.C. Altamirano, J. Agric. Food Chem. 59 (2011) 3559-408

3565.409

[23] M. Safarikova, I. Safarik, J. Magn. Magn. Mater. 194 (1999)108-112.410

[24] M. Safarikova, P. Lunackova, K. Komarek, T. Hubka, I. Safarik, J. Magn. Magn. Mater.411

311 (2007) 405-408.412

[25] X. Zhao, Y.Shi, T. Wang, Y. Cai, G. Jiang, J. Chromatogr. A 1188 (2008)140-147.413

[26] L. Sun, L. Chen, X. Sun, X. Du, Y, Yue, D.He, H. Xu, Q. Zeng, H. Wang, L. Ding, 414

Chemosphere, 77 (2009)1306-1312.415

[27] Y. Jiao, L. Ding, S. Fu, S. Zhu, H. Li, L. Wang, Anal. Methods 4 (2012) 291-298.416

[28] D. Horák, M. Babic, H. Macková, M.J. Benes, J. Sep. Sci. 30 (2007)1751-1772.417

[29] A.S. de Dios, M.E. Díaz-García, Anal. Chim. Acta 666 (2010) 1-22.418

[30] Q. Cheng, F. Qu, N. Bing Li, H. Qun Luo, Anal. Chim. Acta 715 (2012) 113-119.419

[31] X. Zhao, Y. Cai, F. Wu, Y. Pan, H. Liao, B. Xu, Microchem. J. 98 (2011) 207-214.420

[32] J. Li, X. Zhao, Y. Shi, Y. Cai, S. Mou, G. Jiang, J. Chromatogr. A 1180 (2008) 24-31.421

[33] Z. Xu , L. Ding , Y. Long , L.G. Xu , L. Wang, C. Xu, Anal. Methods 3 (2011) 1737-1744.422

[34] Z. Lin, W. Cheng, Y. Li, Z. Liu, X. Chen, C. Huang, Anal. Chim. Acta 720 (2012) 71-76.423

[35] A. Ballesteros-Gómez, S. Rubio, Anal. Chem. 81 (2009) 9012-9020.424

[36] I. P. Román, A, Chisvert, A. Canals, J. Chromatogr. A 1218 (2011) 2467-2475.425

[37] P. Roonasi, A. Holmgren, Appl. Surf. Sci. 255 (2009) 5891-5895.426

Page 16 of 24

Accep

ted

Man

uscr

ipt

16

[38] L. Cuadros, A.M. García, F. Alés, C. Jiménez, M. Román, J. AOAC Int. 78 (1995) 471-476.427

[39] X.-L. Cao, J. Corriveau, S. Popovic, J. Agric. Food Chem. 57 (2009) 1307-1311.428

[40] X.-L. Cao, J. Corriveau, S. Popovic, J. Food Prot. 73 (2010)1548-1551.429

[41] B. M. Thomson, P.R. Grounds, Food Addit. Contam. 22 (2005) 65-72.430

[42] A. Goodson, W. Summerfield, I. Cooper, Food Addit. Contam. 19 (2002) 796-802.431

[43] R. Braunrath, D. Podlipna, S. Padlesak, M. Cichna, J. Agric. Food Chem. 53 (2005) 8911-432

8917.433

[44] S.C. Cunha, C. Almeida, E. Mendes, J.O. Fernandes, Food Addit. Contam. 28 (2011) 513-434

526.435

[45] S-Y. Wu, Q. Xu, T-S. Chen, M. Wang, X-Y. Yin, N-P. Zhang, Y-Y. Shen, Z-Y. Wen, Z-Z. 436

Gu, Chin. J. Anal. Chem. 38 (2010) 503-507.437

[46] X.D. Wang, X.H. Zhao, X.B. Zhang, H.X. Li, Y. Yue, Chin. J. Analysis Lab. 25 (2006) 27-438

29.439

[47] R.S. Zhao, X. Wang, J.P. Yuan, L.L. Zhang, Microchim Acta, 165 (2009) 443-447.440

441

442

443

444

445

446

447

Page 17 of 24

Accep

ted

Man

uscr

ipt

17

Figure Captions448

449

Figure 1. (A,B) Dependence of the adsorption of BPA onto (1) tetradecanoate- and (2) 450

octadecanoate-coated MNPs on (A) the amount of sorbent and (B) the extraction time. (C) 451

Dependence of the extraction efficiency on the volume of sample. Error bars mean the standard 452

deviation of triplicate measurements.453

454

Figure 2. Chromatograms obtained from the analysis of 3 mL of deionized water spiked with 455

(A) 0.67 ng mL-1 of BPA and (B) 0.33 ng mL- 1 of BPA-13C12 after applying the proposed 456

hemimicellar magnetic SPE method457

458

Figure 3. Chromatograms obtained from the analysis of 3 mL of a sport drink (A,B) and a lemon 459

tea (C,D) before (A,C) and after spiking with 0.33 ng mL- 1 of BPA (B,D) and applying the 460

proposed hemimicellar magnetic SPE method.461

462

463

464

465

466

467

468

469

470

471

Page 18 of 24

Accep

ted

Man

uscr

ipt

18

A

0

20

40

60

80

100

0 20 40 60 80

BPA

ads

orpt

ion

(%)

Extraction time (min)

B

2030405060708090

0 5 10 15

BPA

reco

very

(%)

Sample volume (mL)

C

•

0

2040

60

80100

120

0 100 200 300 400

BPA

ads

orpt

ion

(%)

Alkyl carboxylic - coated MNPS (mg)

1

2

Figure 1472

473

Page 19 of 24

Accep

ted

Man

uscr

ipt

19

%

100

11.5 12.5 13.5 14.5

BPA

Time (min)

0

%

100

11.5 12.5 13.5 14.5

Time (min)

0

2.00e3

BPA-13C12

2.00e3A B

%

100

11.5 12.5 13.5 14.5

BPA

Time (min)

0

%

100

11.5 12.5 13.5 14.5

Time (min)

0

2.00e3

BPA-13C12

2.00e3A B

%

100

11.5 12.5 13.5 14.5

BPA

Time (min)

0

%

100

11.5 12.5 13.5 14.5

Time (min)

0

2.00e3

BPA-13C12

2.00e3A B

%

100

11.5 12.5 13.5 14.5

BPA

Time (min)

0

%

100

11.5 12.5 13.5 14.5

Time (min)

0

2.00e3

BPA-13C12

2.00e3A B

Figure 2473

474

Page 20 of 24

Accep

ted

Man

uscr

ipt

20

%

100

11.5 12.5 13.5 14.5

Time (min)

0

%

100

11.5 12.5 13.5 14.5

Time (min)

0

BPA1.49e3

BPA

1.49e3

%

100

11.5 12.5 13.5 14.5

Time (min)

0

%

100

11.5 12.5 13.5 14.5

Time (min)

0

BPA1.49e3

BPA

1.49e3

%

100

11.5 12.5 13.5 14.50

%

100

11.5 12.5 13.5 14.50

BPA1.05e3

BPA

1.05e3

A B

C

%

100

11.5 12.5 13.5 14.50

%

100

11.5 12.5 13.5 14.50

BPA1.05e3

BPA

1.05e3

D

100

11.5 12.5 13.5 14.5

Time (min)

0

%

100

11.5 12.5 13.5 14.5

Time (min)

0

BPA1.49e3

BPA

1.49e3

%

100

11.5 12.5 13.5 14.5

Time (min)

0

%

100

11.5 12.5 13.5 14.5

Time (min)

0

BPA1.49e3

BPA

1.49e3

100

11.5 12.5 13.5 14.50

%

100

11.5 12.5 13.5 14.50

BPA1.05e3

BPA

1.05e3

A B

C

%

100

11.5 12.5 13.5 14.50

%

100

11.5 12.5 13.5 14.50

BPA1.05e3

BPA

1.05e3

D

%

11.5 12.5 13.5 14.5

Time (min)

0

%

11.5 12.5 13.5 14.5

Time (min)

0

BPA1.49e3

BPA

1.49e3

%

11.5 12.5 13.5 14.5

Time (min)

0

%

11.5 12.5 13.5 14.5

Time (min)

0

BPA1.49e3

BPA

1.49e3

%

100

11.5 12.5 13.5 14.5

Time (min)

0

%

100

11.5 12.5 13.5 14.5

Time (min)

0

BPA1.05e3

BPA

1.05e3

A B

C

%

100

11.5 12.5 13.5 14.5

Time (min)

0

%

100

11.5 12.5 13.5 14.5

Time (min)

0

BPA1.05e3

BPA

1.05e3

D

%

11.5 12.5 13.5 14.5

Time (min)

0

%

11.5 12.5 13.5 14.5

Time (min)

0

BPA1.49e3

BPA

1.49e3

%

11.5 12.5 13.5 14.5

Time (min)

0

%

11.5 12.5 13.5 14.5

Time (min)

0

BPA1.49e3

BPA

1.49e3

%

100

11.5 12.5 13.5 14.5

Time (min)

0

%

100

11.5 12.5 13.5 14.5

Time (min)

0

BPA1.05e3

BPA

1.05e3

A B

C

%

100

11.5 12.5 13.5 14.5

Time (min)

0

%

100

11.5 12.5 13.5 14.5

Time (min)

0

BPA1.05e3

BPA

1.05e3

D

474

Figure 3475

476

Page 21 of 24

Accep

ted

Man

uscr

ipt

21

476477

Table 1. Mean concentrations and recoveries, along with their respective standard deviations (SD), found 478for BPA in the analysis of soft drinks 479

480Canned soft drink Mean concentration±SDa

(ng mL-1)

Mean recoveryb±SDa

(%)

Cola drink 0.16 ± 0.02 98 ± 8Light Cola drink 0.066 ± 0.005 98 ± 5Sports drink, brand 1 0.36 ± 0.02 102 ± 5Sports drink, brand 2 1.08 ± 0.04 96 ± 8Lemon (6%) carbonated drink 0.09 ± 0.01 99 ± 3Orange (8%) carbonated drink 0.09 ± 0.01 91 ± 4Energy drink 0.16 ± 0.01 105 ± 3Non-alcoholic bitter drink 0.091 ± 0.003 96 ± 5Lemon tea 0.44 ± 0.02 105 ±6an=3, b Fortification level: 0.33 ng mL-1481

482

Page 22 of 24

Accep

ted

Man

uscr

ipt

22

Table 2. Comparison of the analytical features of current methodologies for the determination of BPA in beverages483

aMethod detection limit; bOverall solvent consumption per sample during the sample preparation steps484

485

Sample size(mL)

Sample treatment aMLD/ (ngL-1)

bOverall solvent consumption(mL)

Solvent evaporation

Recovery (%) Separation/Detection technique

Reference

50 Solvent extraction (acetonitrile)Derivatization

-- 40 Yes 81-103 GC/MS 42

50 Solvent extraction (acetonitrile)Derivatization

-- 40 Yes 42-112 GC/MS 41

10 Solvent extraction(dichloromethane)Two times aqueous back-extraction for clean upDerivatization

0.17 60 Yes 99.5-105.7 GC/MS 14

500 Solvent extraction(ethylacetate)

5700 300 Yes 63 LC/ED 12

10 Dispersive liquid-liquidmicroextraction (tetrachloroethylene/acetonitrile) Derivatization

10 -- No 68-114 GC/MS 44

10 SPE (nylon6 nanofibers membrane) 150 1.15 Yes 95 LC/UV 45

10 SPE ( 45 -- -- 99.9-101 GC/MS 39

50 SPE (Oasis HLB) 0.6 13 Yes 82.1-96.5 LC/MS/MS 15

1 On-line SPE (C18) 15-25 -- No 85-100 LC/MS 16

- SPE (C18) 1200 21 95.8 LC/F 46

200 SPE (bamboo-charcoal) 170 10 No 90.7 LC/UV 47

3 Magnetic SPE 7 2 No 91-105 LC/MS/MS Present method

Page 23 of 24

Accep

ted

Man

uscr

ipt

FeO4

FeO4

FeO 4

FeO 4

FeO4FeO4

FeO4FeO4

FeO4

FeO4

FeO4

FeO4

FeO 4

FeO 4

FeO4FeO4

FeO4FeO4

FeO4

FeO4

HEMIMICELLAR MAGNETIC SPE OF BPA

BPA

LC-MS/MS QUANTITATION

*Graphical Abstract

Page 24 of 24

Accep

ted

Man

uscr

ipt

Analytica Chimica ActaHighlights

●A single-step extraction and cleanup process is proposed for bisphenol A in soft drinks

●Hemimicelles of tetradecanoate adsorbed onto magnetic nanoparticles are used as a sorbent

●Magnetic SPE/LC-MS/MS provides a reliable method for evaluation of human BPA intake from drinks