Anal Bioanal Chem DOI 10.1007/s00216-015-8646-x
NOTE
An innovative arrangement for in-vial membrane-assisted liquid-liquid microextraction: application to the determination of esters of phthalic acid in alcoholic beverages by gas chromatography-mass spectrometry Juan Gabriel March 1 & Victor Cerdà 1
Received: 25 January 2015 / Revised: 14 March 2015 / Accepted: 16 March 2015 # Springer-Verlag Berlin Heidelberg 2015
Abstract A new arrangement for membrane-assisted liquidliquid microextraction is presented. The extracting organic solvent was placed into a chromatographic microvial, compatible with the chromatograph autosampler, whose septum was replaced by a disc of porous hydrophobic membrane. This extraction device was completely immersed into the analytical sample contained in a cylindrical container subjected to rotary and basculant movement. Then, the extraction of analytes took place from the sample to the organic solvent contained in the vial through the membrane. Esters of the phthalic acid have been selected as model analytes to determine the performance characteristics of the extraction system. The limits of detection, limits of quantification and relative standard deviations (%) were in the range 0.1–0.4, 0.3–1 and 4–7, respectively. Esters of phthalic acid have been successfully analysed in alcoholic beverages. The main operational advantages of this arrangement consisted of minimal required handling, minimal risk of cross contamination and its simplicity.
Keywords Liquid-liquid microextraction . Membrane-assisted extraction . Phthalates . Alcoholic beverages . Gas chromatography . Mass spectrometry
Electronic supplementary material The online version of this article (doi:10.1007/s00216-015-8646-x) contains supplementary material, which is available to authorized users. * Juan Gabriel March [email protected] 1
Department of Chemistry, University of Balearic Islands, Carretera de Valldemossa, Km 7,5, 07122 Palma de Mallorca, Spain
Introduction The use of membranes to assist liquid-liquid extraction has been extensively used for analytical purposes. In membraneassisted extraction, the extracting solvent (acceptor phase) and the aqueous sample (donor phase) are physically separated by means of a porous hydrophobic membrane. The organic solvent penetrates the pores of the membrane and remains static by capillary forces [1]. The kinetics of the extraction strongly depends on the area of the membrane and agitation conditions that affect the diffusive transport of analytes in both phases and diminishes the thickness of the donor-phase diffusion layer surrounding the membrane. In fact, these two variables (membrane area and agitation conditions) very significantly affect the characteristics of the different devices for membrane-assisted extraction. Tubular and flat sheet membranes have been extensively and successfully used. Typically, tubular membranes (hollow fibres [1] or membrane bags [2]) containing a microvolume of the organic solvent were immersed into the aqueous sample and the aqueous sample was magnetically stirred while the microvolume of the organic phase remained (relatively) static in the lumen of the tubular membrane. In order to improve the kinetics of the extraction processes, alternative arrangements that allowed the agitation of both phases based on either tubular or flat sheet membranes have been proposed. Thus, using tubular membranes, solvent bar microextraction (SBME) [3] and dual-solvent-stir bar microextraction (DSSBME) [4] techniques are worth mentioning. In SBME, a short hollow fibre sealed at both ends was immersed into the stirred donor phase where the fibre tumbled freely. In DSSBME, a pair of hollow fibres was fixed on a stainless steel stir bar with four sites for fixing each end of the two hollow fibres. In both configurations, the extraction devices moved, but it was not evident if the organic solvent
J.G. March, V. Cerdà
was effectively agitated. Arrangements aiming for the agitation of the organic acceptor phase based on flat sheet membranes have been proposed more recently, namely, in-vial membrane liquid-liquid extraction [5], proposed by our group, and stir membrane liquid-liquid microextraction (SM-LLME) [6]. In in-vial membrane liquid-liquid extraction, the acceptor phase was in a microvial, compatible with a chromatograph autosampler, where the septum of the stopper was replaced by the extraction membrane, and the stopper assembled to a volumetric flask containing the aqueous donor phase; the extraction was accomplished by means of orbital agitation. SMLLME used an extraction unit consisting of a cylindrical chamber, with an incorporated iron bar, whose upper base is the flat extraction membrane. For extraction, the unit filled with the organic solvent is immersed into the donor phase and magnetically stirred. In general, for the above techniques, good analytical characteristics have been demonstrated. From an operational point of view, it is worth mentioning the minor handling required in the in-vial membrane liquid-liquid extraction, but its slower kinetic of extraction is a considerable drawback. In this context, the aim of this paper is to further develop and improve the in-vial membrane liquid-liquid extraction; a new arrangement is presented with less handling requirements where the extraction microvial is immersed into the donor phase (as in SM-LLME), thus avoiding the coupling to the volumetric flask.
chromatography grade) purchased from Scharlau Chemie (Barcelona, Spain). Besides, phthalic acid esters (PAE) standards in toluene were also prepared for enrichment factor (EF) and extraction efficiency (EE) determination. Purified water obtained using a Millipore Milli-Q system (Bedford, MA, USA) and stored in a glass bottle with polytetrafluoroethylene (PTFE) stopper containing activated charcoal (Scharlau Chemie), was used to prepare aqueous phases. The extraction device The extraction device consisted of a 200-μL chromatographic glass microvial (bought from Teknokroma, Barcelona, Spain) compatible with conventional autosamplers. The septum of the screw cap of the vial was replaced by a sized piece of membrane obtained by portioning a PTFE membrane disc (150 μm wall thickness, porosity 85 %, 0.45 μm pore size) purchased from Millipore, Billerica, MA, USA. The microvial, containing the organic acceptor phase, was placed into a 50-mL glass tube (with a screw stopper) containing the hydro-ethanolic donor phase. The whole was maintained at a rotary movement using a Movil-Rod roller and tilt mixer (Selecta, Barcelona, Spain) with seven rollers of 24 cm long, which allowed 12 devices be simultaneously shaken. The constant speed of the mixer was 45 rpm. A schematic drawing is shown in Fig. 1. As can be glimpsed, both the donor and the acceptor phases were agitated at the same time and contacted the extraction membrane; the membrane surface available for extraction was 0.79 cm2.
Experimental Procedure Reagents Dimethyl phthalate 99.9 % (DMP), diethyl phthalate 99.2 % (DEP), di-n-butyl phthalate 99.0 % (DBP), di-n-pentyl phthalate 99.0 % (DPP), bis(2-methoxyethyl) phthalate 99.4 % (BMEP), benzyl benzoate 99.0 % (internal standard, IS) and toluene 99.9 % were purchased from Sigma-Aldrich (Madrid, Spain). Standards were prepared in ethanol 99.9 % (liquid
Forty-eight millilitres of the sample and the corresponding volume of ethanol were added to the 50-mL tube. Aliquots (from 0 to 0.5 mL) of a 20 mg L−1 composite standard and 0.1 mL of the IS solution (40 mg L−1) were then added, resulting a total added volume of ethanol of 0.6 mL. A microvial containing 190 μL of toluene with the extraction membrane incorporated was then placed into the 50-mL tube.
Fig. 1 Set-up used for liquidliquid extraction
Chromatographic micro-vial
Donor phase
Rotary and basculant movement
Acceptor phase PTFE membrane
An innovative arrangement for in-vial membrane-assisted liquid-liquid microextraction: application to the...
The whole was rolled and rocked during 24 h using a roller mixer. After, the microvial was recovered and dried by simple contact with absorbent paper, the membrane was changed by a conventional septum, and its content was homogenised and placed into the autosampler for chromatographic analysis. Alternatively, when practicable, the same commercial container (i.e. the same glass bottle) can be used instead of the 50-mL tube. Equipment and chromatographic analysis The chromatographic analysis of the organic acceptor phases was performed using a Shimadzu GC-2010 gas chromatograph equipped with a Shimadzu GCMS-QP20105 mass selective detector, a Shimadzu AOC-20i autosampler and a TRB-624 fused silica capillary column (30 m, 0.25 mm i.d., 1.40 μm film thickness) purchased from Teknokroma. Conditions on chromatographic analysis are in the Electronic Supplementary Material (ESM).
Results and discussion Remarkable differences between the proposed set-up for membrane-assisted liquid-liquid extraction and that for already reported devices are the agitation conditions, the required handling and the area of the extraction membrane. Typically, magnetic stirring of the donor phase is used to favour the extraction. The device reported here by means of a rotary and basculant movement of the tube containing the microvial allowed both the donor and the acceptor phases be agitated simultaneously. Also, the cautious handling of a microvolume with microsyringes or micropipettes after extraction is avoided if the microvial is compatible with a conventional autosampler. In opposition to commented operational advantages, a drawback of the new device is the reduced area of the extraction membrane which ineludibly slows the extraction rate and, also, the use of an autosampler (normally) requires a higher volume of organic solvent than the manual handling. Moreover, compared with the commented operational conditions (mode of agitation, membrane area and mode of injection), a key factor that determines the EE of the process is the organic solvent. It was found that polar organic solvents, as ethyl acetate and n-octanol, due to its considerable interaction with water, after a prolonged extraction time (as is the case of the present work) a considerable amount of water can permeate the membrane, thus causing the presence of an aqueous phase into the vial which affected negatively the precision of the measurements. Recommended volatile organic solvents in gas chromatography, as hexane, cannot be used as an extracting solvent because after extraction for several hours, most of the solvent
was evaporated/dissolved into the acceptor phase and the remaining volume in the vial was reduced to few microlitres, thus making impracticable the use of the autosampler. Satisfactory results were obtained using toluene, carbon tetrachloride and decane. On the basis of EF values and repeatability, toluene was selected to continue the study (ESM Fig. S1). In fact, aromatic solvents, as toluene and xylene, have been reported as optimal solvents for PAEs in membrane-based extraction techniques [7]. Using solvents with high vapour pressure, as toluene or carbon tetrachloride, for storage of the organic extracts after extraction, the membrane must be replaced by an appropriate septum; otherwise, evaporation of the organic solvent took place through the pores of the membrane. For short intervals between extraction and analysis, the uncontrolled evaporation of the solvent through the membrane was satisfactorily corrected using the corresponding internal standard. Using decane, the organic extracts can be stored for 24 h being not necessary the replacement of the membrane by the septum. The other studied variables were the ethanol concentration in the donor phase and the donor and acceptor phase volumes. The presence of ethanol increases the solubility of the analytes in the donor phase, and consequently, the EFs diminished, which provoked a decrease in the sensitivity. Nevertheless, even at ethanol concentration of the order of 15 %, EFs higher than 10 were obtained (ESM Fig. S2); this made the extraction method suitable for analysis of samples whose matrix contains a considerable amount of ethanol, as alcoholic beverages. In an extraction process, when stationary conditions are reached, the EF increases when the volume of the donor phase increases and the volume of the acceptor phase decreases. If stationary conditions are not reached, as is the case of the present work, the kinetics of the extraction can provoke deviations from this theoretical behaviour. So, volumes of the acceptor phase and donor phase in the ranges 0.150–200 μL and 30–50 mL, respectively, were studied. In such ranges, averaged EFs followed the expected tendency, but the differences were of low significance.
Table 1 Validation data of the analytical determination and extraction efficiency (EE) of the extraction procedure Analyte R2
LOD (μg L−1) LOQ (μg L−1) RSD % (n=5) EEa
DMP DEP DBP BMEP DPP
0.2 0.2 0.2 0.1 0.4
a
0.998 0.990 0.994 0.991 0.995
0.6 0.6 0.5 0.3 1
7 5 6 5 7
0.18 0.23 0.10 0.14 0.08
EE was estimated as the percentage of the total analyte that was extracted from the donor phase, assuming constant (and equal to the initial value) the volume of the acceptor phase
J.G. March, V. Cerdà Table 2
Qualitative and quantitative analysis of alcoholic beverages
Sample
PAEs Identified
Found μg L−1 (SDa)
Brandy
DEP DBP BOP DBP
5.0 (0.9) 65 (7) Not determined 25 (4)
DIBP DBP DPP BBP DEP DHP DUP DBP DPP DBP DUP DEP DBP DIBP DEP DBP DIBP
Not determined 16 (4) 21 (4) Not determined 4.2 (0.6) Not determined Not determined 30 (4) 32 (4) 9 (3) Not determined 1.0 (0.5) 4.3 (0.4) Not determined 0.4 (0.2) 2.2 (0.4) Not determined
Red wine A Red wine B
Withe wine
Sangriab Beer Ab Beer Bb
Low-alcohol beerb
DIBP diisobutyl phthalate, BOP butyloctyl phthalate, BBP benzylbutyl phthalate, DHP diheptyl phthalate, DUP diundecyl phthalate a
planning of the experimental work. The eventual disadvantages of this prolonged time can be compensated by parallel processing of many samples simultaneously (using a conventional agitator, 12 samples can be extracted at the same time), thus reducing the time of analysis by sample and makes the device potentially attractive for routine analysis.
Performance and evaluation for quantitative analysis Under selected conditions (indicated in the procedure), the performance of the method was evaluated in terms of linearity, precision, limit of detection (LOD) and limit of quantification (LOQ) (Table 1). A calibration graph for each analyte was established in the concentration range of LOQ–300 μg L−1 (eight levels in triplicate). The LOD and LOQ were estimated from a signal-to-noise ratio of 3 and 10, respectively. A satisfactory linearity was obtained (coefficient of determination >0.99). The repeatability, as the relative standard deviation (RSD, %), calculated from five replicates at a concentration level of 100 μg L−1 for each analyte was lower than 8 %. The LODs and LOQs were in the range 0.1–0.4 and 0.3–1 μg L−1, respectively. Such satisfactory limits are a consequence of the conjunction of two opposite factors. By the one hand, the reduced membrane area caused a poor EE; on the other hand, the presented set-up allows the use of a relatively big sample size (if available) which had a beneficial influence on the finally accomplished LODs.
Standard deviation obtained from the regression line corresponding to the method of standard additions (number of additions=5)
b Samples were decarbonated before extraction; for this, the samples were gently shaken at atmospheric pressure until elimination of most of the carbon dioxide and then sonicated until no gas longer escapes
Finally, the extraction kinetics was studied. The relatively reduced area of the extraction membrane, as expected, slowdowns the kinetics of the extraction, and consequently, using this device, a much more prolonged extraction time than other similar dispositives based on the same principles was required (ESM Fig. S3). In fact, stationary conditions were not observed even after 30 h of extraction. A 24-h extraction time was selected to validate the method because it facilitated the Fig. 2 Chromatograms from a sangria real sample. a Scan mode. b Selected ion mass monitoring; m/z=149 for DBP and DPP, and m/z=105 for benzyl benzoate (internal standard)
Analysis of samples A variety of commercial alcoholic beverages purchased from the local market have been analysed, i.e. brandy, wine, sangria (drink based on red wine and carbonated beverages) and beer. First of all, an in-vial extraction from 48 mL of the sample was carried out for qualitative analysis. Mass spectra of chromatographic peaks (recorded in scan mode) were compared with standard spectra in the NIST library. Identified PAEs are shown in Table 2. When identified PAE coincided with a PAE selected to carry out this study, the identification was
An innovative arrangement for in-vial membrane-assisted liquid-liquid microextraction: application to the...
confirmed by the tR from the corresponding standard and the quantitative analysis was done. For quantitative analysis, due to the unpredictable variability of the matrix, the method of standard additions was applied, as recommended for such complex matrices [8]. The results are shown in Table 2. As extensively reported, most of the analysed alcoholic beverages were PAE contaminated, being DBP a fairly frequent contaminant, as already reported for wine samples [9]. For illustrative purposes, chromatograms from a sangria real sample obtained in scan mode and selected ion monitoring mode are depicted in Fig. 2. As can be seen, clean chromatograms were obtained using the proposed extraction technique.
membrane surface wasted for extraction reduces the cost of the determination. It requires a long extraction time, but a reasonable extraction time per sample is achieved, running several extractions in parallel. It makes the device attractive for routine analysis. Acknowledgments Grant CTQ2013-47461-R (‘Ministerio de Economía y Competitividad’ Spanish Government) is acknowledged. Informatics assistance from D. Moreno is strongly appreciated.
References 1. 2.
Conclusions In this paper, an innovative set-up for membrane-assisted liquid-liquid extraction that allowed effective agitation of both donor and acceptor phase, thus favouring the mass transfer of target analytes, is described. Operational advantages of the presented arrangement are a minimal handling and a low risk of accidental contamination or losses of analytes and sample carryover. The extraction scheme does not need any dedicated equipment than a simple roller mixer. Moreover, the reduced
3. 4. 5. 6. 7. 8. 9.
Pedersen-Bjergaard S, Rasmussen KE (2008) J Chromatogr A 1184: 132–142 Moeder M, Bauer C, Popp P, van Pinxteren M, Reemtsma T (2012) Anal Bioanal Chem 403:1731–1741 Jiang X, Lee HK (2004) Anal Chem 76:5591–5596 Yu C, Liu Q, Lan L, Hu B (2008) J Chromatogr A 1188:124–131 March JG, Moukhchan F, Cerdà V (2011) Anal Chim Acta 685:132– 137 Alcudia-León MC, Lucena R, Cárdenas S, Valcárcel M (2011) J Chromatogr A 1218:869–874 Huang G, Li HF, Zhang BT, Ma Y, Lin JM (2012) Talanta 100:64–70 Carrillo JD, Salazar C, Moreta C, Tena MT (2007) J Chromatogr A 1164:248–261 Cinelli G, Avino P, Notardonato I, Centola A, Russo MV (2013) Anal Chim Acta 769:72–78
NOTE
An innovative arrangement for in-vial membrane-assisted liquid-liquid microextraction: application to the determination of esters of phthalic acid in alcoholic beverages by gas chromatography-mass spectrometry Juan Gabriel March 1 & Victor Cerdà 1
Received: 25 January 2015 / Revised: 14 March 2015 / Accepted: 16 March 2015 # Springer-Verlag Berlin Heidelberg 2015
Abstract A new arrangement for membrane-assisted liquidliquid microextraction is presented. The extracting organic solvent was placed into a chromatographic microvial, compatible with the chromatograph autosampler, whose septum was replaced by a disc of porous hydrophobic membrane. This extraction device was completely immersed into the analytical sample contained in a cylindrical container subjected to rotary and basculant movement. Then, the extraction of analytes took place from the sample to the organic solvent contained in the vial through the membrane. Esters of the phthalic acid have been selected as model analytes to determine the performance characteristics of the extraction system. The limits of detection, limits of quantification and relative standard deviations (%) were in the range 0.1–0.4, 0.3–1 and 4–7, respectively. Esters of phthalic acid have been successfully analysed in alcoholic beverages. The main operational advantages of this arrangement consisted of minimal required handling, minimal risk of cross contamination and its simplicity.
Keywords Liquid-liquid microextraction . Membrane-assisted extraction . Phthalates . Alcoholic beverages . Gas chromatography . Mass spectrometry
Electronic supplementary material The online version of this article (doi:10.1007/s00216-015-8646-x) contains supplementary material, which is available to authorized users. * Juan Gabriel March [email protected] 1
Department of Chemistry, University of Balearic Islands, Carretera de Valldemossa, Km 7,5, 07122 Palma de Mallorca, Spain
Introduction The use of membranes to assist liquid-liquid extraction has been extensively used for analytical purposes. In membraneassisted extraction, the extracting solvent (acceptor phase) and the aqueous sample (donor phase) are physically separated by means of a porous hydrophobic membrane. The organic solvent penetrates the pores of the membrane and remains static by capillary forces [1]. The kinetics of the extraction strongly depends on the area of the membrane and agitation conditions that affect the diffusive transport of analytes in both phases and diminishes the thickness of the donor-phase diffusion layer surrounding the membrane. In fact, these two variables (membrane area and agitation conditions) very significantly affect the characteristics of the different devices for membrane-assisted extraction. Tubular and flat sheet membranes have been extensively and successfully used. Typically, tubular membranes (hollow fibres [1] or membrane bags [2]) containing a microvolume of the organic solvent were immersed into the aqueous sample and the aqueous sample was magnetically stirred while the microvolume of the organic phase remained (relatively) static in the lumen of the tubular membrane. In order to improve the kinetics of the extraction processes, alternative arrangements that allowed the agitation of both phases based on either tubular or flat sheet membranes have been proposed. Thus, using tubular membranes, solvent bar microextraction (SBME) [3] and dual-solvent-stir bar microextraction (DSSBME) [4] techniques are worth mentioning. In SBME, a short hollow fibre sealed at both ends was immersed into the stirred donor phase where the fibre tumbled freely. In DSSBME, a pair of hollow fibres was fixed on a stainless steel stir bar with four sites for fixing each end of the two hollow fibres. In both configurations, the extraction devices moved, but it was not evident if the organic solvent
J.G. March, V. Cerdà
was effectively agitated. Arrangements aiming for the agitation of the organic acceptor phase based on flat sheet membranes have been proposed more recently, namely, in-vial membrane liquid-liquid extraction [5], proposed by our group, and stir membrane liquid-liquid microextraction (SM-LLME) [6]. In in-vial membrane liquid-liquid extraction, the acceptor phase was in a microvial, compatible with a chromatograph autosampler, where the septum of the stopper was replaced by the extraction membrane, and the stopper assembled to a volumetric flask containing the aqueous donor phase; the extraction was accomplished by means of orbital agitation. SMLLME used an extraction unit consisting of a cylindrical chamber, with an incorporated iron bar, whose upper base is the flat extraction membrane. For extraction, the unit filled with the organic solvent is immersed into the donor phase and magnetically stirred. In general, for the above techniques, good analytical characteristics have been demonstrated. From an operational point of view, it is worth mentioning the minor handling required in the in-vial membrane liquid-liquid extraction, but its slower kinetic of extraction is a considerable drawback. In this context, the aim of this paper is to further develop and improve the in-vial membrane liquid-liquid extraction; a new arrangement is presented with less handling requirements where the extraction microvial is immersed into the donor phase (as in SM-LLME), thus avoiding the coupling to the volumetric flask.
chromatography grade) purchased from Scharlau Chemie (Barcelona, Spain). Besides, phthalic acid esters (PAE) standards in toluene were also prepared for enrichment factor (EF) and extraction efficiency (EE) determination. Purified water obtained using a Millipore Milli-Q system (Bedford, MA, USA) and stored in a glass bottle with polytetrafluoroethylene (PTFE) stopper containing activated charcoal (Scharlau Chemie), was used to prepare aqueous phases. The extraction device The extraction device consisted of a 200-μL chromatographic glass microvial (bought from Teknokroma, Barcelona, Spain) compatible with conventional autosamplers. The septum of the screw cap of the vial was replaced by a sized piece of membrane obtained by portioning a PTFE membrane disc (150 μm wall thickness, porosity 85 %, 0.45 μm pore size) purchased from Millipore, Billerica, MA, USA. The microvial, containing the organic acceptor phase, was placed into a 50-mL glass tube (with a screw stopper) containing the hydro-ethanolic donor phase. The whole was maintained at a rotary movement using a Movil-Rod roller and tilt mixer (Selecta, Barcelona, Spain) with seven rollers of 24 cm long, which allowed 12 devices be simultaneously shaken. The constant speed of the mixer was 45 rpm. A schematic drawing is shown in Fig. 1. As can be glimpsed, both the donor and the acceptor phases were agitated at the same time and contacted the extraction membrane; the membrane surface available for extraction was 0.79 cm2.
Experimental Procedure Reagents Dimethyl phthalate 99.9 % (DMP), diethyl phthalate 99.2 % (DEP), di-n-butyl phthalate 99.0 % (DBP), di-n-pentyl phthalate 99.0 % (DPP), bis(2-methoxyethyl) phthalate 99.4 % (BMEP), benzyl benzoate 99.0 % (internal standard, IS) and toluene 99.9 % were purchased from Sigma-Aldrich (Madrid, Spain). Standards were prepared in ethanol 99.9 % (liquid
Forty-eight millilitres of the sample and the corresponding volume of ethanol were added to the 50-mL tube. Aliquots (from 0 to 0.5 mL) of a 20 mg L−1 composite standard and 0.1 mL of the IS solution (40 mg L−1) were then added, resulting a total added volume of ethanol of 0.6 mL. A microvial containing 190 μL of toluene with the extraction membrane incorporated was then placed into the 50-mL tube.
Fig. 1 Set-up used for liquidliquid extraction
Chromatographic micro-vial
Donor phase
Rotary and basculant movement
Acceptor phase PTFE membrane
An innovative arrangement for in-vial membrane-assisted liquid-liquid microextraction: application to the...
The whole was rolled and rocked during 24 h using a roller mixer. After, the microvial was recovered and dried by simple contact with absorbent paper, the membrane was changed by a conventional septum, and its content was homogenised and placed into the autosampler for chromatographic analysis. Alternatively, when practicable, the same commercial container (i.e. the same glass bottle) can be used instead of the 50-mL tube. Equipment and chromatographic analysis The chromatographic analysis of the organic acceptor phases was performed using a Shimadzu GC-2010 gas chromatograph equipped with a Shimadzu GCMS-QP20105 mass selective detector, a Shimadzu AOC-20i autosampler and a TRB-624 fused silica capillary column (30 m, 0.25 mm i.d., 1.40 μm film thickness) purchased from Teknokroma. Conditions on chromatographic analysis are in the Electronic Supplementary Material (ESM).
Results and discussion Remarkable differences between the proposed set-up for membrane-assisted liquid-liquid extraction and that for already reported devices are the agitation conditions, the required handling and the area of the extraction membrane. Typically, magnetic stirring of the donor phase is used to favour the extraction. The device reported here by means of a rotary and basculant movement of the tube containing the microvial allowed both the donor and the acceptor phases be agitated simultaneously. Also, the cautious handling of a microvolume with microsyringes or micropipettes after extraction is avoided if the microvial is compatible with a conventional autosampler. In opposition to commented operational advantages, a drawback of the new device is the reduced area of the extraction membrane which ineludibly slows the extraction rate and, also, the use of an autosampler (normally) requires a higher volume of organic solvent than the manual handling. Moreover, compared with the commented operational conditions (mode of agitation, membrane area and mode of injection), a key factor that determines the EE of the process is the organic solvent. It was found that polar organic solvents, as ethyl acetate and n-octanol, due to its considerable interaction with water, after a prolonged extraction time (as is the case of the present work) a considerable amount of water can permeate the membrane, thus causing the presence of an aqueous phase into the vial which affected negatively the precision of the measurements. Recommended volatile organic solvents in gas chromatography, as hexane, cannot be used as an extracting solvent because after extraction for several hours, most of the solvent
was evaporated/dissolved into the acceptor phase and the remaining volume in the vial was reduced to few microlitres, thus making impracticable the use of the autosampler. Satisfactory results were obtained using toluene, carbon tetrachloride and decane. On the basis of EF values and repeatability, toluene was selected to continue the study (ESM Fig. S1). In fact, aromatic solvents, as toluene and xylene, have been reported as optimal solvents for PAEs in membrane-based extraction techniques [7]. Using solvents with high vapour pressure, as toluene or carbon tetrachloride, for storage of the organic extracts after extraction, the membrane must be replaced by an appropriate septum; otherwise, evaporation of the organic solvent took place through the pores of the membrane. For short intervals between extraction and analysis, the uncontrolled evaporation of the solvent through the membrane was satisfactorily corrected using the corresponding internal standard. Using decane, the organic extracts can be stored for 24 h being not necessary the replacement of the membrane by the septum. The other studied variables were the ethanol concentration in the donor phase and the donor and acceptor phase volumes. The presence of ethanol increases the solubility of the analytes in the donor phase, and consequently, the EFs diminished, which provoked a decrease in the sensitivity. Nevertheless, even at ethanol concentration of the order of 15 %, EFs higher than 10 were obtained (ESM Fig. S2); this made the extraction method suitable for analysis of samples whose matrix contains a considerable amount of ethanol, as alcoholic beverages. In an extraction process, when stationary conditions are reached, the EF increases when the volume of the donor phase increases and the volume of the acceptor phase decreases. If stationary conditions are not reached, as is the case of the present work, the kinetics of the extraction can provoke deviations from this theoretical behaviour. So, volumes of the acceptor phase and donor phase in the ranges 0.150–200 μL and 30–50 mL, respectively, were studied. In such ranges, averaged EFs followed the expected tendency, but the differences were of low significance.
Table 1 Validation data of the analytical determination and extraction efficiency (EE) of the extraction procedure Analyte R2
LOD (μg L−1) LOQ (μg L−1) RSD % (n=5) EEa
DMP DEP DBP BMEP DPP
0.2 0.2 0.2 0.1 0.4
a
0.998 0.990 0.994 0.991 0.995
0.6 0.6 0.5 0.3 1
7 5 6 5 7
0.18 0.23 0.10 0.14 0.08
EE was estimated as the percentage of the total analyte that was extracted from the donor phase, assuming constant (and equal to the initial value) the volume of the acceptor phase
J.G. March, V. Cerdà Table 2
Qualitative and quantitative analysis of alcoholic beverages
Sample
PAEs Identified
Found μg L−1 (SDa)
Brandy
DEP DBP BOP DBP
5.0 (0.9) 65 (7) Not determined 25 (4)
DIBP DBP DPP BBP DEP DHP DUP DBP DPP DBP DUP DEP DBP DIBP DEP DBP DIBP
Not determined 16 (4) 21 (4) Not determined 4.2 (0.6) Not determined Not determined 30 (4) 32 (4) 9 (3) Not determined 1.0 (0.5) 4.3 (0.4) Not determined 0.4 (0.2) 2.2 (0.4) Not determined
Red wine A Red wine B
Withe wine
Sangriab Beer Ab Beer Bb
Low-alcohol beerb
DIBP diisobutyl phthalate, BOP butyloctyl phthalate, BBP benzylbutyl phthalate, DHP diheptyl phthalate, DUP diundecyl phthalate a
planning of the experimental work. The eventual disadvantages of this prolonged time can be compensated by parallel processing of many samples simultaneously (using a conventional agitator, 12 samples can be extracted at the same time), thus reducing the time of analysis by sample and makes the device potentially attractive for routine analysis.
Performance and evaluation for quantitative analysis Under selected conditions (indicated in the procedure), the performance of the method was evaluated in terms of linearity, precision, limit of detection (LOD) and limit of quantification (LOQ) (Table 1). A calibration graph for each analyte was established in the concentration range of LOQ–300 μg L−1 (eight levels in triplicate). The LOD and LOQ were estimated from a signal-to-noise ratio of 3 and 10, respectively. A satisfactory linearity was obtained (coefficient of determination >0.99). The repeatability, as the relative standard deviation (RSD, %), calculated from five replicates at a concentration level of 100 μg L−1 for each analyte was lower than 8 %. The LODs and LOQs were in the range 0.1–0.4 and 0.3–1 μg L−1, respectively. Such satisfactory limits are a consequence of the conjunction of two opposite factors. By the one hand, the reduced membrane area caused a poor EE; on the other hand, the presented set-up allows the use of a relatively big sample size (if available) which had a beneficial influence on the finally accomplished LODs.
Standard deviation obtained from the regression line corresponding to the method of standard additions (number of additions=5)
b Samples were decarbonated before extraction; for this, the samples were gently shaken at atmospheric pressure until elimination of most of the carbon dioxide and then sonicated until no gas longer escapes
Finally, the extraction kinetics was studied. The relatively reduced area of the extraction membrane, as expected, slowdowns the kinetics of the extraction, and consequently, using this device, a much more prolonged extraction time than other similar dispositives based on the same principles was required (ESM Fig. S3). In fact, stationary conditions were not observed even after 30 h of extraction. A 24-h extraction time was selected to validate the method because it facilitated the Fig. 2 Chromatograms from a sangria real sample. a Scan mode. b Selected ion mass monitoring; m/z=149 for DBP and DPP, and m/z=105 for benzyl benzoate (internal standard)
Analysis of samples A variety of commercial alcoholic beverages purchased from the local market have been analysed, i.e. brandy, wine, sangria (drink based on red wine and carbonated beverages) and beer. First of all, an in-vial extraction from 48 mL of the sample was carried out for qualitative analysis. Mass spectra of chromatographic peaks (recorded in scan mode) were compared with standard spectra in the NIST library. Identified PAEs are shown in Table 2. When identified PAE coincided with a PAE selected to carry out this study, the identification was
An innovative arrangement for in-vial membrane-assisted liquid-liquid microextraction: application to the...
confirmed by the tR from the corresponding standard and the quantitative analysis was done. For quantitative analysis, due to the unpredictable variability of the matrix, the method of standard additions was applied, as recommended for such complex matrices [8]. The results are shown in Table 2. As extensively reported, most of the analysed alcoholic beverages were PAE contaminated, being DBP a fairly frequent contaminant, as already reported for wine samples [9]. For illustrative purposes, chromatograms from a sangria real sample obtained in scan mode and selected ion monitoring mode are depicted in Fig. 2. As can be seen, clean chromatograms were obtained using the proposed extraction technique.
membrane surface wasted for extraction reduces the cost of the determination. It requires a long extraction time, but a reasonable extraction time per sample is achieved, running several extractions in parallel. It makes the device attractive for routine analysis. Acknowledgments Grant CTQ2013-47461-R (‘Ministerio de Economía y Competitividad’ Spanish Government) is acknowledged. Informatics assistance from D. Moreno is strongly appreciated.
References 1. 2.
Conclusions In this paper, an innovative set-up for membrane-assisted liquid-liquid extraction that allowed effective agitation of both donor and acceptor phase, thus favouring the mass transfer of target analytes, is described. Operational advantages of the presented arrangement are a minimal handling and a low risk of accidental contamination or losses of analytes and sample carryover. The extraction scheme does not need any dedicated equipment than a simple roller mixer. Moreover, the reduced
3. 4. 5. 6. 7. 8. 9.
Pedersen-Bjergaard S, Rasmussen KE (2008) J Chromatogr A 1184: 132–142 Moeder M, Bauer C, Popp P, van Pinxteren M, Reemtsma T (2012) Anal Bioanal Chem 403:1731–1741 Jiang X, Lee HK (2004) Anal Chem 76:5591–5596 Yu C, Liu Q, Lan L, Hu B (2008) J Chromatogr A 1188:124–131 March JG, Moukhchan F, Cerdà V (2011) Anal Chim Acta 685:132– 137 Alcudia-León MC, Lucena R, Cárdenas S, Valcárcel M (2011) J Chromatogr A 1218:869–874 Huang G, Li HF, Zhang BT, Ma Y, Lin JM (2012) Talanta 100:64–70 Carrillo JD, Salazar C, Moreta C, Tena MT (2007) J Chromatogr A 1164:248–261 Cinelli G, Avino P, Notardonato I, Centola A, Russo MV (2013) Anal Chim Acta 769:72–78
