Surfactant-coated carbon nanotubes for the liquid–liquid extraction of phthalates and other...

PAPER IN FOREFRONT

Surfactant-coated carbon nanotubes for the liquid–liquidextraction of phthalates and other migrants in virginolive oils

Silvia López-Feria & Rafael Lucena &

Soledad Cárdenas & Miguel Valcárcel

Received: 12 May 2009 /Revised: 9 August 2009 /Accepted: 11 August 2009 /Published online: 4 September 2009# Springer-Verlag 2009

Abstract This article presents a novel application of thesurfactant-coated carbon nanotubes for the joint extractionof phthalates, xylene isomers and styrene from virgin oliveoils. For this purpose, two carbon nanotubes, multi-walledand single-walled carbon nanotubes, were evaluated, theformer providing the better results. The target migrantswere liquid–liquid extracted from the virgin olive oilsamples by means of the carbon nanotube pseudophase,and the aqueous extracts were analysed by headspace-gaschromatography-mass spectrometry. Sodium chloride wasadded to favour the release of the analytes to the gaseousphase of the vial. The comparison of the chromatogramsobtained with the direct analysis reveals a sensitivityenhancement with the use of the surfactant-coated carbonnanotubes, especially for the phthalates, which can beascribed to the π–π interaction between the aromatic ringand the nanotube surface. Limits of detection were between6 and 30 µgL−1 for 1 mL of sample. The precision of themethod was better than 7.6% and the recoveries higher than92%. Finally, the influence of the package material and thestorage conditions on the concentration of the analytes wasaccomplished, and it was found that the concentration ofphthalates increases with time in all the packages studied,while the more volatile compounds were found to decreaseunder the same conditions.

Keywords Phthalates . Xylene isomers . Styrene .

Liquid–liquid extraction . Carbon nanotubes .

Headspace-gas chromatography-mass spectrometry

Introduction

Phthalic acid esters, commonly called phthalates, arewidely used (about 80% of these compounds) asplastizicers in the polymer industry to improve flexibil-ity, workability and general handling properties. Sincethese compounds do not form stable bonds with polymerto which they are added, these substances tend to migratefrom the package to the product contained in it,especially in the case of fatty and oily foodstuffs, sobecoming contaminants. Because of their extensive use,they are potential hazards for human health as aconsequence of their endocrine disruptive effects [1, 2].Regarding xylene isomers (ortho, meta and para) andstyrene, all are volatile aromatic hydrocarbons that canappear either naturally (viz. superior plant wax, natural oilseepage, amino acid fermentation) or by anthropogenicsources (viz. combustion products, industrial paints,adhesives, packages). Their lipophilic nature makes themalso especially harmful by accumulation in fatty foodstuffsand concretely edible oils and fats, and their effectscomprise changes in the liver to damages in heart andnervous system, among others [3, 4].

With the aim of preserving health safety, a substantialamount of research effort is being undertaken in order tomonitor the level of contamination in different types offoods directly in contact with materials made up of thesecompounds, mainly plasticised packages. Most of theminvolve gas chromatography coupled to several detectors.For matrices with high fat contents, it is necessary to

S. López-FeriaCIFA Venta del Llano, IFAPA,Ctra. Bailén-Motril km 18.5 Mengíbar,23620 Jaén, Spain

R. Lucena : S. Cárdenas :M. Valcárcel (*)Department of Analytical Chemistry, University of Córdoba,Marie Curie Building (Annex), Campus de Rabanales,Cordoba 14071, Spaine-mail: [email protected]

Anal Bioanal Chem (2009) 395:737–746DOI 10.1007/s00216-009-3066-4

completely remove the high molecular mass fat prior to theinstrumental analysis to preserve the performance of thechromatographic column. These procedures usually requiremany cleanup steps to: (a) remove any remaining lipid andother matrix components and (b) preconcentrate the targetcompounds from the sample, favouring the decrease ofdetection limits for these substances. Liquid–liquid extrac-tion (LLE) and solid phase extraction processes have beentraditionally used for the determination of a wide range ofanalytes such as pesticides, aromatic hydrocarbons andphthalates in environmental and food samples. Concerningthe LLE, two different approaches have been recentlydriven for the detection and quantification of phthalates inwaters, both in conjunction with gas chromatography-massspectrometry (GC-MS). One uses a simple and efficienthollow fibre membrane protected liquid phase microex-traction, while the other develops a dispersive liquid–liquidmicroextraction, thus minimising the solvent volume usedfor the procedure [5, 6]. Regarding the solid phase

extraction process, there is a trend for the reduction ofsample preparation steps and the automation of the systemby the joint use of solid phase microextraction (SPME) andGC-MS detection. It has been applied for the determinationof benzene, toluene, ethyl benzene, xylene isomer andstyrene (BTEXS) and phthalate contaminants in a widevariety of environmental (waters) and food (wine or milk)matrices [7–11]. Some of them use an additional cleanupstep based on gel permeation chromatography wheremacromolecules are separated according to their size. Thismethodology is also effective for removing fats and oils andhas been utilised for determining several phthalates in dairyproducts [12]. In the case of the determination ofcontaminants in vegetable oils, this system has beenrecently applied for the detection of phthalates in olive oilsamples by gas chromatography-tandem mass spectrometry,

Table 1 Figures of merit of the optimised analytical procedure

Analyte Rt (min) Quantification ionsb,c Qualifier ionsc Linear range (mg L−1) R2 LOD (µg L−1) RSD (%) (n=5) Recoveryd

(m+p)-xylenea 5.81 91 77, 106 0.02–10 0.98 6 6.2 97±4

o-xylene 6.32 91 77, 106 0.05–10 0.97 10 5.8 99±5

styrene 6.32 104 51, 78 0.05–10 0.99 10 3.8 100±4

dimethyl phthalate 13.59 163 77, 194 0.5–50 0.99 50 3.7 103±4

diethyl phthalate 14.74 149 177, 222 0.1–50 0.98 25 4.1 92±5

dibutyl phthalate 17.39 149 57, 278 0.1–50 0.99 30 7.6 104±7

di(2-ethylhexyl)phthalate

20.76 149 167, 390 0.1–50 0.99 30 5.1 101±7

a The provided features correspond to the mixture of both isomersb Base peakscMolecular peaks are indicated in italicsd Recovery values expressed in percentages (mean ± SD, n=3) for the analysis of deodorised olive oil spiked with 1 mgL−1 of the analytes

Table 2 Selection of variables for liquid–liquid extraction andinstrumental procedure

Variable Range studied Optimum

Extractant MWCNTs MWCNTsc-SWCNTs

Surfactant CTAB SDSTriton X-100

SDS

Sample/PSP ratio Sample: 1–5 mL 1 mL

PSP: 1–5 mL 3 mL

Additives MeOH NaClNaCl

NaCl concentration [% (w/v)] 10–40 10

HS temperature (°C) 60–90 90

HS time (min) 50–100 90 Fig. 1 Extraction capability of surfactant-coated c-SWCNTs andMWCNTs for the studied analytes

738 S. López-Feria et al.

where two ionisation methodologies were compared toachieve better analytical performances [13]. Previously, asolvent-free analytical procedure based on SPME was usedwith the aim of automating the determination of thesemigrants from rape oil coming to the conclusion that themanual SPME method was more adequate than theautomated one due to the better efficiency of the sampleshaking versus sample stirring [14]. Finally, benzene,toluene, ethylbenzene and xylene isomers and styrene havebeen successfully determined in olive oil samples by usingsurfactant-coated carbon nanotubes as extractant [15]. Inthis study, the sorption properties of multi-walled carbonnanotubes have been exploited, demonstrating the capabilityof the system for the extraction of volatile organic com-pounds from an oily matrix with low limits of detection.

The purpose of the present work is to evaluate theusefulness of the dispersed carbon nanotubes for the jointextraction of semivolatile compounds as phthalates and other

migrants such as xylene isomers and styrene usually presentin olive oil samples. The use of these carbon nanoparticlesdispersed in a surfactant medium is proposed prior toan automatic headspace-gas chromatographic-massspectrometric method, enhancing significantly the limitsof detection achieved compared with the direct analysis.In addition, the evolution of the concentration of theselected compounds with the time under differentstorage conditions has also been afforded in this article.

Experimental

Reagent and samples

The following phthalates were purchased from Sigma-Aldrich (Madrid, Spain): dimethyl phthalate (DMP, 99+%), diethyl phthalate (DEP, 99.5%), dibutyl phthalate

Fig. 2 a Effect of the addition of a chemical modifier in the PSP on the analytical response of the migrant and b influence of the concentration ofNaCl on the abundance of the signals obtained

5000001

Abundance

400000

51

200000

300000

6

2+3

4a

2+3

4 5

1000006 7

6 b

c

4 5

0

0 5 10 15 20 25 30 Time (min)

Fig. 3 Chromatograms provid-ed by analysis of a a migrants-free olive oil sample, b an oliveoil sample spiked with the ana-lytes and c an olive oil spikedwith the analytes and analysedby the optimised LLE system.1 (m + p)-xylene, 2, o-xylene, 3styrene, 4 DMP, 5, DEP, 6 DBP,and 7, DEHP

Liquid–liquid extraction of phthalates in virgin olive oils 739

(DBP, 99+%) and di(2-ethylhexyl) phthalate (DEHP,99%). In addition, m-xylene, p-xylene, o-xylene andstyrene were provided by Fluka (Barcelona, Spain) andwere all of GC grade. Sodium chloride, sodium dodecylsulphate (SDS) and methanol (Sigma-Aldrich) were alsoemployed. Individual stock standard solutions of eachanalyte were prepared in methanol (99.8%, Panreac,Barcelona, Spain), at a concentration of 1,000 mg L−1

and stored in dark at 4 °C until analyses.Working standard solutions used for calibration purposes

were obtained by dilution of the stocks in a deodorisedolive oil coming from a Spanish supplier.

Multi-walled carbon nanotubes (MWCNTs), withpurity over 95%, and single-walled carbon nanotubes(SWCNTs) with purity of 90% were supplied by Sigma-Aldrich (Madrid, Spain). The outside diameter, internaldiameters and length ranged between 20 and 30 nm, 5and 10 nm, and 0.5 and 200 µm, respectively forMWCNTs. The SWCNT outside diameter varied be-tween 1 and 2 nm, while the length ranged from 0.5 to2 µm. SWCNTs were carboxylated following theprocedure proposed by Wang et al. [16].

In order to study the migration's effect due to the bottlingof olive oils in different package materials, six extra-virginolive oils purchased at a local market and bottled inceramic, glass, plastic, metal can, tetrabrik and single-dose20-mL plastic container were selected and stored at severalroom conditions until analysis.

In order to evaluate the usefulness of the proposedmethodology in the routine determination of these migrants,another set of commercial samples (containing differentextra-virgin olive oils to those previously commented) wereanalysed.

Pseudophase preparation

For the liquid–liquid extraction of the analytes from oliveoil samples, 5.0 mg of MWCNTs were dispersed in 50 mLof distilled water with 130 mg of solid sodium dodecylsulphate according to a previously optimised procedure[17]. The mixture, placed in a 50-mL volumetric flask, wassonicated for 30 min in an ultrasonic bath (50 W, 60 Hz),changing the bath water every 15 min in order to preventthe overheating of the pseudophase (PSP). The finaldispersion contained the MWCNTs and SDS at concen-trations of 0.1 mg mL−1 and 9 mM, respectively.

Apparatus

The analytical method was carried out using an HP6890 gaschromatograph (GC) equipped with an HP5973 massspectrometric (MS) detector based on a quadrupole analyserand a photomultiplier detector (Agilent Technologies, Palo

Alto, USA). The system was coupled to a MPS2 headspace(HS) autosampler made up of a 32-space tray for vials of10/20 mL, a robotic arm with a temperature-controlledsyringe and a heating/headspace generation module. Gaschromatographic separation was achieved on a SLB-5 ms(Supelco, Madrid, Spain) fused silica capillary column(30 m×0.25 mm i.d. and 0.25 µm of film thickness).Chromatographic conditions were as follows: 250 °C,injection temperature; split mode operation with a splitratio 1:10; initial temperature 50 °C (kept 5 min), raised upto 280 °C at 15 °C/min (kept 5 min). Helium (5.0 gradepurity, Air Liquide, Sevilla, Spain), regulated by a digitalpressure and flow controlled, was used as carrier gas(constant flow 1.5 mL min−1). A split ratio 1:25 and asolvent delay of 4 min was employed in the optimisedmethod. The temperatures of the mass spectrometer sourceand quadrupole were maintained at 230 and 150 °C,respectively. The detector operated in full scan mode, byscanning a mass range from m/z 40 to 400, and the electronimpact ionisation was used for analyte fragmentation withionisation energy of 70 eV. Total ion current chromato-grams were acquired and processed using G1701BAStandalone Data Analysis software (Agilent Technologies,Palo Alto, USA) on a Pentium IV computer that alsocontrolled the whole system.

Liquid–liquid extraction

In the optimised procedure, aliquots of 1 mL of virginolive oil with the analytes at appropriate concentrationwere placed in a 10-mL vial, and 3 mL of thesurfactant-coated MWCNTs was added as extractant.The vial was sealed and shaken for 1 min to facilitate

Table 3 Variables affecting the concentration of xylenes, styrene andphthalates in virgin olive oils

Variable Conditions studied

Packing material Ceramic

Glass

Plastic

Metal can

Tetrabrik

Single-dose 20-mL plastic container

Light Darkness

Fluorescent lamp

Time (weeks) 1

2

3

7

19


the contact between the two phases. After that, it wasleft to stand for 15 min for phases' separation. Then, thepseudophase (bottom phase) containing the extractedanalytes was transferred to a 10-mL headspace glass vialby means of a Pasteur micropipette. In order to facilitatethe release of the analytes from the PSP to the headspaceof the vial, 0.3 g of NaCl was added. Finally, the vial washermetically sealed with a polytetrafluorethylene/siliconeseptum and placed into the HS tray for the chromato-graphic analysis.

Instrumental procedure

The vial containing the PSP enriched with the phthalates,xylene isomers and styrene was heated at 90 °C for 90 minunder continuous mechanical stirring into the HS heatingmodule. After this time, the 2.5 mL gastight HS syringeintroduced this volume of the homogenised headspace fromthe vial into the injector (fixed at 250 °C) of the GC, wherethe helium stream drove the compounds through thecapillary column to the mass spectrometer detector. Theanalytes were identified by their retention time, and theirmass spectra were compared with those provided by thecorresponding standards analysed using similar procedure. Inall cases, quantification was achieved by extracting a specificm/z fragment, namely: 91 for (m + p)-xylene and o-xylene,104 for styrene, 163 for DMP and 149 for DEP, DBP andDEHP. Table 1 shows the two additional m/z fragment ions(including base peak and molecular ion) employed for theidentification of the target analytes (qualifier ions).

Carbon nanotube recovery

Once used, the carbon nanotubes pseudophase was mixedwith methanol and water under magnetic stirring for disper-sion breakage. Then, the mixture was filtered, and the carbonnanotubes retained on the filter were sequentially washed withn-hexane, water and methanol and further dried at 50 °C inan oven. Following this procedure, carbon nanotubes werereused during the whole study since no irreversible inter-actions with the matrix were observed.

Results and discussion

Carbon nanotubes (CNTs) have been used for the design ofpreconcentration tools aimed at enhancing the performance ofthe analytical procedure involved. They have strong adsorp-tion affinity to a wide variety of organic compounds and arealso characterised by their high surface area. The dispersion ofCNTs in a surfactant medium is usually employed to minimisethe presence of aggregates, which reduces the active surfacearea of this material and therefore their effectiveness. In the T

able4

Initialandmaxim

umconcentrations

inmgL−1

(CiandCm,respectively)

foun

din

thecommercialextra-virgin

oliveoilsbo

ttled

into

differenttypesof

packages,w

hich

wereanalysed

and

mon

itoredby

theprop

osed

metho

d

Analyte

Ceram

icDarkglass

Plastic

Metal

can

Tetra

Brik

Singledo

se

Ci±SD

Cm±SD

Ci±SD

Cm±SD

Ci±SD

Cm±SD

Ci±SD

Cm±SD

Ci±SD

Cm±SD

Ci±SD

Cm±SD

(m+p)-X

ylene

0.06

±0.03

0.08

±0.04

0.01

±0.02

0.02

±0.02

nd0.08

±0.04

0.02

±0.01

0.02

±0.01

0.03

±0.02

0.03

±0.02

0.02

±0.01

0.02

±0.01

o-Xylene

0.11

±0.05

0.12

±0.05

0.01

±0.01

0.02

±0.02

0.01

±0.01

0.01

±0.01

0.12

±0.07

0.16

±0.09

0.11

±0.06

0.12

±0.07

0.10

±0.08

0.17

±0.08

Styrene

1.9±0.2

1.9±0.2

1.6±0.1

1.6±0.1

0.05

±0.03

0.05

±0.03

0.04

±0.02

0.05

±0.02

1.6±0.1

1.6±0.1

1.4±0.1

1.4±0.1

Dim

ethy

lph

thalate

nd0.21

±0.03

nd2.7±0.2

nd10

.9±0.9

nd4.8±0.6

0.15

±0.03

0.20

±0.04

nd11

±2

Diethyl

phthalate

0.43

±0.02

0.82

±0.05

0.52

±0.03

1.2±0.6

0.25

±0.02

0.73

±0.04

nd1.6±0.3

0.41

±0.04

1.3±0.4

0.31

±0.03

1.8±0.8

Dibutyl

phthalate

0.56

±0.06

1.9±0.7

nd1.1±0.5

nd2.9±0.8

ndnd

nd3.2±0.2

0.87

±0.06

4.3±0.3

Di(2-ethy

lhexyl)

phthalate

2.5±0.4

4.5±0.6

1.2±0.3

6.4±0.8

2.1±0.1

5.2±0.9

1.6±0.5

7.7±0.9

2.7±0.2

6.4±0.9

3.2±0.8

12±1

ndno

nedetected


present work, surfactant-coatedMWCNTs aqueous solution isevaluated as pseudophase for the liquid–liquid extraction ofthe migrant compounds selected in virgin olive oil samples.

Selection of variables

Instrumental working conditions were previously studied inorder to obtain a chromatographic signal for the phthalates,xylene isomers and styrene by analysing an aqueous standardcontaining each analyte at a concentration of 1 mg L−1. Forthis purpose, the headspace generation procedure wasevaluated through the incubation temperature and timeranging from 60 to 90 °C (10 °C intervals) and from 50 to100 min (10 min intervals), respectively. For the firstparameter, the highest temperature provided the highestanalytical signal. In the case of incubation time, 90 minwas necessary to obtain an appropriate transfer of theanalytes to the HS, mainly for phthalate compounds. Nosignificant differences were found over this incubation time.

Concerning the selection of variables, Table 2 lists theanalytical parameters taken into account in this study. Themost critical variables on the sensitivity and selectivity ofthe system are those related with the liquid–liquid extractionprocess. In a previous publication [15], the usefulness ofsurfactant-coated MWCNTs for the extraction of BTEXSfrom olive oil was pointed out. In this article, we initiallystudied the best carbon nanoparticles for the joint extractionof phthalates, xylene isomers and styrene from olive oils. Forthis purpose, aqueous dispersions of MWCNTs andcarboxylated-single-walled carbon nanotubes (c-SWCNTs)in 9 mM SDS were evaluated as extractant (PSP). In bothcases, the carbon nanoparticles were at a concentration of

0.1 mg mL−1. An olive oil sample spiked with the analytes ata concentration of 50 µg L−1 (for xylenes and styrene) and500 µg L−1 (phthalates) was used in this study. Aliquots of2 mL of sample were placed in a 10-mL glass vial and 2 mLof PSP was added. The vial was then sealed and manuallyshaken for 1 min. After phases' separation, 1.5 mL of thePSP was transferred to a 10-mL headspace glass vial andanalysed by HS-GC-MS. As can be seen in Fig. 1,MWCNTs provided the best results in terms of extractioncapability for all the selected analytes. Next, three differentsurfactants namely: cetyl trimethyl ammonium bromide(CTAB), Triton X-100 and sodium dodecyl sulphate (SDS)were evaluated for the dispersion of MWCNTs. SDS wasfinally chosen at a final concentration of 9 mM as it providesthe best PSP stability and extractability. In addition, the ratiobetween the sample and extractant volumes was also studiedbetween 1:5 and 5:1, obtaining the best results for 1 mL ofsample extracted with 3 mL of PSP. Afterwards, theextraction time was studied within the interval 0.5–5.0 min(five points). Slight increases of the analyte concentrationwith the extraction time were found, despite which 1 minwas selected in order to enhance the sample throughput ofthe method. Finally and regarding the addition of an organicmodifier, in a previous article it was described that the use ofmethanol favoured the extraction of BTEXS from olive oilsamples [15]. However, in the present work the key point inthe liquid–liquid extraction process was found to be theextraction of the phthalate compounds from the virgin oliveoil sample. In this way, the addition of methanol resultedcounterproductive despite its positive effect for the rest ofstudied migrants. For this reason, no modifier was employedduring the extraction procedure.

Fig. 4 Comparison of the abundance of the analytes in the analysed samples stored in darkness and those obtained after 2 weeks under light condition


Once the liquid–liquid extraction of the analytes fromthe sample was established, chemical and instrumentalvariables affecting the headspace generation were studied.For this aim, the PSPs obtained after the extraction of theanalytes from olive oil samples were analysed underdifferent conditions. First, the presence of organic modifiers(methanol, 100 µL) or salting out agents [NaCl, 10% (w/v)]in the PSP was evaluated. As can be seen in Fig. 2a, NaCl wasmore efficient in favouring the analytes' release to theheadspace of the vial, and therefore, its influence was deeplystudied by adding increasing concentrations [10–40% (w/v)]in the extracted PSP solution before the instrumentalanalysis. A concentration of 10% of NaCl was sufficient toimprove the transfer of analytes to HS fraction (see Fig. 2b).

Finally, the headspace generation procedure was againstudied once the composition of the extractant wasestablished. It was observed that the optimum values

selected for the aqueous standards of the analytes (viz.90 °C of temperature and 90 min of incubation time)provided the best results in terms of peak area whenworking with the dispersed carbon nanotubes as extractant.

Analytical performance

Linearity of the proposed method for the selected analyteswas demonstrated using a deodorised olive oil as blankwhich was spiked with nominal concentrations of theanalytes from 20 µg L−1 to 50 mg L−1. The figures ofmerits of the method are listed in Table 1. The sampleswere analysed in triplicate, and calibration curves weregenerated by plotting peak areas of the extracted target ionsfor each analyte versus their concentration. Linearity withR2 values higher than 0.97 was observed in all calibrationplots within the evaluated concentration range. The limit of

4000

5000

6000

single-dose

0

1000

2000

3000

1 3 7 19 2 4 9 1 3 7 19 2 4 9 1 3 7 19 2 4 9 1 3 7 19

1 3 7 19 2 4 9 1 3 7 19 2 4 9 1 3 7 19 2 4 9 1 3 7 19

1 3 7 19 2 4 9 1 3 7 19 2 4 9 1 3 7 19 2 4 9 1 3 7 19

abun

danc

e

5000

6000

ceramic(m+p)-xylene styrene DMP DEP DBP DEHP

(m+p)-XYLENE o-XYLENE STYRENE DMP DEP DBP DEHP



weeks

0

1000

2000

3000

4000

abun

danc

e

5000

glass

o-xylene

(m+p)-xylene styrene DMP DEP DBP DEHPo-xylene


weeks

10001500200025003000350040004500

abun

danc

e

0500

weeks

aFig. 5 Influence of the time andtype of package in the concentra-tion of xylene isomers, styreneand phthalates in commercialvirgin olive oil samples


detection (LOD) for the seven migrant compounds wasobtained as the concentration providing a peak area threetimes higher than the noise (S/N=3), and the values werefound within the range of 6–30 µg L−1. In light ofthe results, the proposed method is sensitive enough forthe detection of the selected analytes, and although themaximum residue levels are not well defined by theRegulation for olive oils, the LOD resulted to be belowthose provided by the bibliography when SPME is used[14]. In addition, the precision of the method expressed interms of relative standard deviation (RSD) was estimatedby analysing five aliquots of the blank olive oil samplespiked with the studied analytes at a concentration of5 mg L−1. The results obtained were found below 7.6%,and they can also be seen in the table.

On the other hand, recovery studies were carried out byspiking a deodorised olive oil with 1.0 mg L−1 of the target

compounds. Analytes were extracted by the LLE procedurefrom the sample (n=3) and subsequently analysed by theoptimised HS-GC-MS method. As it can be seen in Table 1,the average recovery values were found to be higher than92%, which demonstrates the accuracy and feasibility of theprocedure also expressed by the standard deviation (SD) ofthe measurements that were, in all cases, lower than 7%.

Finally, in order to show the extractive capacity of thepseudophase based on the use of MWCNTs, Fig. 3 depictsthe chromatograms corresponding to a migrants-free oliveoil sample (a), an olive oil sample spiked with 10 mg L−1 ofthe standard solution, analysed by the direct HS-GC-MSmethod (b) and analysed by the proposed procedure usingsurfactant-coated carbon nanotubes and HS-GC-MS (c). Asit can be seen in the figure, the use of the PSP markedlyenhance the sensitivity of the method for all the analytes.Moreover, it allows the detection of the phthalates, which

25003000350040004500

metal can

0500

100015002000

abun

dace

3000

3500

plastic

0

500

1000

1500

2000

2500

abun

danc

e

5000

tetrabrik

50010001500200025003000350040004500

abun

danc

e

0




1 3 7 19 2 4 9 1 3 7 19 2 4 9 1 3 7 19 2 4 9 1 3 7 19

1 3 7 19 2 4 9 1 3 7 19 2 4 9 1 3 7 19 2 4 9 1 3 7 19

(m+p)-XYLENE o-XYLENE STYRENE DMP DEP DBP DEHP weeks


1 3 7 19 2 4 9 1 3 7 19 2 4 9 1 3 7 19 2 4 9 1 3 7 19


bFig. 5 (continued)


was only achieved by means of the surfactant-coatedMWCNTs pseudophase.

Application to commercial extra-virgin olive oils. Influenceof the package material and the storage conditions

The analytical strategy developed was applied to determinethe concentration of xylene isomers, styrene and phthalatesin extra-virgin olive oils commercialised under differentpackages. Moreover, the influence of the storage conditionsin them, mainly the light/darkness and time, has also beenevaluated. Table 3 summarises the parameters taken intoaccount in this study. The objective was to evaluate themigration process for the selected compounds from thepacking materials used by the olive oil industry, monitoringthe influence of the storage conditions usually utilised inmarkets and kitchens. For this reason, a headspace in eachbottle was also generated, which was increasing during thestudy's time similar to the bottle-emptying in a commonhome (single-dose package excepted). The initial concen-trations found for each compound in the different commer-cial extra-virgin olive oil samples analysed are listed inTable 4, as well as the maximum value found in the courseof the study. In addition, Figs. 4 and 5 represent thevariations in the concentrations found during the studyconcerning the presence/absence of light and the time.

As can be seen in Fig. 4, no clear conclusions can befollowed for the influence of light during the 2-week periodevaluated. We consider that a possible explanation for thisis that two opposite effects occurred in the analysedsamples. The increasing headspace volume generated inthe samples will favour the release of the compounds to thegaseous phase. This process will be more important for thesamples stored under light conditions due to highesttemperature they suffered during these 2 weeks. However,the light can also favour the migration of the compoundsfrom the package and/or degrade the analytes, andtherefore, the final concentration will depend on themagnitude of these processes in each case.

Concerning the influence of the storage time, twodifferent behaviours were observed within the group ofselected analytes. Xylene isomers and styrene were foundto be present in all of the virgin olive oil samples analysed,regardless the type of container. (m + p)-Xylenes were thesole exception, as they were not initially detected in theplastic bottle (see Table 4). The maximum concentration ofthe analytes was found in the initial analyses of thecommercial samples. While the xylene isomers were inthe low micrograms per litre range, for styrene, theconcentrations reached highest values, comparable in somecases with those of the phthalates. The decrease of theconcentration of these four compounds with the time (seeFig. 5) can be directly related to their high volatility which T

able

5Resultsforthe12

extra-virgin

oliveoilsamples

analysed

bytheproposed

method

Ceram

icDarkglass

Plastic

Metal

can

Tetrabrik

Singledo

se

Analyte

Sam

ple1

Sam

ple2

Sam

ple1

Sam

ple1

Sam

ple2

Sam

ple3

Sam

ple4

Sam

ple1

Sam

ple2

Sam

ple1

Sam

ple1

Sam

ple2

Ca±SDb

Ca±SDb

Ca±SDb

Ca±SDb

Ca±SDb

Ca±SDb

Ca±SDb

Ca±SDb

Ca±SDb

Ca±SDb

Ca±SDb

Ca±SDb

(m+p)-xylene

0.06

±0.03

0.09

±0.02

0.01

±0.02

nd*

0.01

±0.01

nd*

nd*

0.02

±0.01

0.03

±0.01

0.03

±0.02

0.02

±0.01

0.03

±0.01

o-xy

lene

0.11

±0.05

0.15

±0.05

0.01

±0.01

0.01

±0.01

nd*

0.02

±0.01

0.01

±0.01

0.12

±0.07

0.08

±0.06

0.11

±0.06

0.10

±0.08

0.20

±0.15

styrene

1.9±0.2

2.3±0.2

1.6±0.1

0.05

±0.03

0.1±0.1

0.04

±0.01

0.05

±0.01

0.04

±0.02

0.10

±0.07

1.6±0.1

1.4±0.1

2.7±0.2

dimethy

lph

thalate

nd*

nd*

nd*

nd*

nd*

0.8±0.01

nd*

nd*

nd*

0.15

±0.03

nd*

nd*

diethy

lph

thalate

0.43

±0.02

0.32

±0.03

0.52

±0.03

0.25

±0.02

0.31

±0.02

0.41

±0.03

nd*

nd*

nd*

0.41

±0.04

0.31

±0.03

0.90

±0.09

dibu

tylph

thalate

0.56

±0.06

0.40

±0.05

nd*

nd*

nd*

nd*

nd*

nd*

nd*

nd*

0.87

±0.06

2.0±0.15

di(2-ethylhexy

l)ph

thalate

2.5±0.4

1.8±0.5

1.2±0.3

2.1±0.1

2.6±0.1

5.0±0.9

4.8±0.8

1.6±0.5

2.8±0.8

2.7±0.2

3.2±0.8

5.1±1.1

aCon

centratio

nexpressedin

mgL−1

bSD

standard

deviation

*ndno

nedetected


increases their concentration in the headspace of thecontainer and, thus, decreasing the concentration in theliquid sample.

As far as phthalates are concerned, at least one of theselected four compounds was found in all the commercialvirgin olive oil samples assayed. This can be ascribed to theunsecure but usual presence of plastic material through thewhole process behind the olive oil production, frommalaxation to bottling. Moreover, the metal can wasinternally laminated with plastic, the ceramic package wasa glazed one and the glass bottle presented a plastic elementin the screw-on cap. It can explain the increasingconcentration of phthalates with the time in all the samplesanalysed as the likely result of a migration from thecontainer to the olive oil sample (Fig. 5). DEHP wasthe most abundant phthalate, which is in accordance withthe data reported in the literature. DMP was initiallydetected in the tetrabrik package, although it was presentafter 2 months in all the samples. DBP was not detected inthe metal can during the study (see Table 4). The highestconcentration of all the phthalates was reached after2 months, and it decreased in the last part of the study. Itcan be explained by a combined effect of the headspace ofthe bottles and the highest room temperature as these lastmonths corresponded to July–September.

Finally, a new set of commercial samples were analysedby the proposed method in order to evaluate the usefulnessof the proposal. The obtained results, as well as thoseobtained for the initial analysis of the first six samples, aresummarised in Table 5.

Conclusions

This article presents a novel alternative for the jointdetermination of volatile and semivolatile organic com-pounds from virgin olive oils. The methodology provides amarked sensitivity improvement compared with the directanalysis of the sample. This effect is especially relevant forphthalates, which cannot be detected by the analysis of theheadspace generated after the direct heating of the sample. Itcan be explained by the favourable interaction of thecompounds with the nanotubes and the absence of interfer-ence from endogenous matrix components in the finalaqueous extract. The additional study developed concerningthe influence of the packing material, light and time on theconcentration of the analytes in commercial samples pointsout that the material used for the package can modify theconcentration of the migrants in the commercial samples.

Acknowledgements This work was supported by grant CTQ2007-60426 of the DGI of the SpanishMinistry of Science and Innovation. SLFwants to thank the IFAPA for the financial support through a Junta deAndalusia (Regional Government of Andalusia, Spain) fellowship.

References

1. VanWezel AP, VanVlaardingen P, Posthumus R, Crommentuijn GH,Sijm DTHM (2000) Environmental risk limits for two phthalates,with special emphasis on endocrine disruptive properties. EcotoxicolEnviron Safety 46:305–321

2. Heudorf U, Mersch-Sundermann V, Angerer J (2007) Phthalates:toxicology and exposure. Int J Environ Health 210:623–634

3. Vichi S, Pizzale L, Conte LS, Buxaderas S, López-Tamames E(2007) The occurrence of volatile and semi-volatile aromatichydrocarbons in virgin olive oils from north-eastern Italy. FoodControl 18:1204–1210

4. Irwin RJ, VanMouwerik M, Stevens L, Seese MD, Basham W(1997) Environmental Contaminants Encyclopedia. National ParkService, Fort Collins

5. Psillakis E, Kalogerakis N (2003) Hollow-fibre liquid-phasemicroextraction of phthalate esters from water. J Chromatogr A999:145–153

6. Farahani H, Norouzi P, Dinarvand R, Reza Ganjali M (2007)Development of dispersive liquid-liquid microextraction com-bined with gas chromatography-mass spectrometry as a simple,rapid and highly sensitive method for the determination ofphthalate esters in water samples. J Chromatogr A 1172:105–112

7. Peñalver A, Pocurull E, Borrull F, Marcé RM (2000) Determina-tion of phthalate esters in water samples by solid-phase micro-extraction and gas chromatography with mass spectrometricdetection. J Chromatogr A 872:191–201

8. Cao XL (2008) Determination of phthalates and adipate in bottledwater by hadspace solid-phase microextraction and gas chroma-tography/mass spectrometry. J Chromatogr A 1178:231–238

9. Gaujac A, Soares Emídio E, Navickiene S, Costa Ferreira SL,Silveira Dórea H (2008) Multivariate optimization of a solid phasemicroextraction-headspace procedure for the determination ofbenzene, toluene, ethylbenzene and xylenes in effluent samplesfrom a waste treatment plant. J Chromatogr A 1203:99–104

10. Carrillo JD, Martínez MP, Tena MT (2008) Determination ofphthalates in wine by headspace solid-phase microextractionfollowed by gas chromatography-mass spectrometry. Used ofdeuterated phthalates as internal standards. J Chromatogr A1181:125–130

11. Feng YL, Zhu J, Sensenstein R (2005) Development for aheadspace solid-phase microextraction method combined withgas chromatography mass spectrometry for the determination ofphthalate esters in cow milk. Anal Chim Acta 538:41–48

12. Sharman M, Read WA, Castle L, Gilbert J (1994) Levels ofdi-(2-ethylhexyl) phthalate and total phthalate esters in milk,cream, butter and cheese. Food Addit Contam 11:375–385

13. Cavaliere B, Macchione B, Sindona G, Tagarelli A (2008) Tandemmass spectrometry in food safety assessment: the determination ofphthalates in olive oil. J Chromatogr A 1205:137–143

14. Holadová K, Prokůpková G, Hajšlová J, Poustka J (2007)Headspace solid-phase microextraction of phtalic acid esters fromvegetable oil employing solvent based matrix modification. AnalChim Acta 582:24–33

15. Carrillo-Carrión C, Lucena R, Cárdenas S, Valcárcel M (2007)Liquid–liquid extraction/headspace/gas chromatographic/massspectrometric determination of benzene, toluene, ethylbenzene,(o-, m- and p-)xylene and styrene in olive oil using surfactant-coated carbón nanotubes as extractant. J Chromatogr A 1171:1–7

16. Wang Y, Iqbal Z, Malhotra SV (2005) Functionalization of carbonnanotubes with amines and enzymes. Chem Phys Letters 402:96–101

17. Carrillo-Carrión C, Lucena R, Cárdenas S, Valcárcel M (2007)Surfactant-coated carbón nanotubes as pseudophases in liquid–liquid extraction. Analyst 132:551–559


Surfactant-coated carbon nanotubes for the liquid–liquid extraction of phthalates and other...

Documents

Transcript of Surfactant-coated carbon nanotubes for the liquid–liquid extraction of phthalates and other...