Analytica Chimica Acta 804 (2013) 135–142
Contents lists available at ScienceDirect
Analytica Chimica Acta journal homepage: www.elsevier.com/locate/aca
Supramolecular solvent-based hollow fiber liquid phase microextraction of benzodiazepines Fatemeh Rezaei a , Yadollah Yamini a,∗ , Morteza Moradi b , Bahram Daraei c a b c
Department of Chemistry, Tarbiat Modares University, P.O. Box 14115-175, Tehran, Iran Department of Semiconductors, Materials and Energy Research Center, Karaj, Iran Department of Toxicology, Tarbiat Modares University, Tehran, Iran
h i g h l i g h t s
g r a p h i c a l
a b s t r a c t
• Hollow fiber microextraction based • • • •
on supramolecular solvents was developed. Supramolecular solvent was produced from coacervation of decanoic acid and Bu4 N+ . The method was introduced for extraction of benzodiazepines from aqueous solutions. Several parameters affecting extraction efficiency were investigated and optimized. HPLC-DAD was applied for the determination of the drugs in fruit juice and urine samples.
a r t i c l e
i n f o
Article history: Received 7 August 2013 Received in revised form 9 October 2013 Accepted 11 October 2013 Available online 19 October 2013 Keywords: Hollow fiber Vesicles Benzodiazepines High performance liquid chromatography Urine Plasma
a b s t r a c t A new, efficient, and environmental friendly hollow fiber liquid phase microextraction (HF-LPME) method based on supramolecular solvents was developed for extraction of five benzodiazepine drugs. The supramolecular solvent was produced from coacervation of decanoic acid aqueous vesicles in the presence of tetrabutylammonium (Bu4 N+ ). In this work, benzodiazepines were extracted from aqueous samples into a supramolecular solvent impregnated in the wall pores and also filled inside the porous polypropylene hollow fiber membrane. The driving forces for the extraction were hydrophobic, hydrogen bonding, and -cation interactions between the analytes and the vesicular aggregates. High-performance liquid chromatography with photodiode array detection (HPLC-DAD) was applied for separation and determination of the drugs. Several parameters affecting the extraction efficiency including pH, hollow fiber length, ionic strength, stirring rate, and extraction time were investigated and optimized. Under the optimal conditions, the preconcentration factors were obtained in the range of 112–198. Linearity of the method was determined to be in the range of 1.0–200.0 g L−1 for diazepam and 2.0–200.0 g L−1 for other analytes with coefficient of determination (R2 ) ranging from 0.9954 to 0.9993. The limits of detection for the target benzodiazepines were in the range of 0.5–0.7 g L−1 . The method was successfully applied for extraction and determination of the drugs in water, fruit juice, plasma and urine samples and relative recoveries of the compounds studied were in the range of 90.0–98.8%. © 2013 Elsevier B.V. All rights reserved.
1. Introduction
∗ Corresponding author. Tel.: +98 21 82883417; fax: +98 21 88006544. E-mail addresses: [email protected], [email protected] (Y. Yamini). 0003-2670/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.aca.2013.10.026
Benzodiazepines (BZDs) are a large group of drugs with a broad range of therapeutic effects, used as hypnotics, anxiolytics, muscle relaxants, and anticonvulsants [1]. The sedative and amnestic
136
F. Rezaei et al. / Analytica Chimica Acta 804 (2013) 135–142
properties of some BZDs are also considered useful in anesthesia. In addition, the BZDs have a rapid onset of action combined with low acute toxicity [2]. Several major metabolites of BZDs possess pharmacological profiles similar to the parent drugs. BZDs are now among the most commonly prescribed drugs, which increases their potential for addiction and abuse in cases of crime, driving under the influence of drugs, suicide, and drug facilitated sexual assault. For these reasons, simultaneous analysis of BZDs and their metabolites in complicated matrices is of great interest to clinicians and forensic toxicologists [1]. A variety of methods exist in the literature for detection of benzodiazepines in complicated matrices. Gas chromatography–mass spectrometry (GC–MS) determinations of BZDs have been reviewed by Maurer [3] and further studies have also been carried out [4–7]. High-performance liquid chromatography (HPLC) methods have been studied [8–13] and a critical evaluation of the application of capillary electrophoresis (CE) to detection of BZDs in formulations and body materials have been made [14]. Chromatographic techniques usually require complex isolation procedures (sample preparation) to separate BZDs from complicated matrices. Several sample preparation procedures such as liquid-liquid extraction (LLE) [15], solid phase extraction [16,17], and liquid phase microextraction [18,19] have been developed for separation and preconcentration of drugs from different matrices. However, the LLE and SPE methods often require large amounts of organic solvents, some of which are harmful and contaminate the environment due to their high vapor pressure. The pressure to decrease organic solvent usage in laboratories is increasing. Thus, miniaturization and improvement of sample handling using alternatives is a challenge that has been discussed by several researchers. From this perspective, supramolecular solvents (SUPRASs)-based extractions were an appropriate choice [20]. In 2006, Pérez-Bendito’s research group for the first time investigated the potential of the tetrabutylammonium (TBA)induced liquid-phase separation in alkyl carboxylic acid vesicular solutions for extraction of organic compounds prior to HPLC [21]. SUPRASs are water-immiscible liquids made up of supramolecular assemblies dispersed in a continuous phase. Two characteristics give the alkyl carboxylic acid-based coacervates a high potential for analytical extraction processes. First, the polar region of molecular aggregates consists of protonated and deprotonated carboxylic groups and ammonium groups; so, a number of interactions (e.g., electrostatics, -cation, hydrogen bonds, formation of mixed aggregates, etc.) can be established with analytes, in addition to hydrophobic interactions in the hydrocarbon region. Second, vesicles have a number of available solubilization sites; so, high concentrations of polar and apolar molecules can be solubilized in each aggregate. Formation of vesicles in the aqueous solution before adding Bu4 N+ ions was not essential to achieve liquid-liquid-phase separation. Several extraction methods, such as single-drop microextraction (SDME) [22], floating drop microextraction (FDME) [23], dispersive liquid–liquid microextraction (DLLME) [24] have been reported based on SUPRASs. In our previous work, SUPRAS comprised of tetrabutylammonium-induced vesicles of decanoic acid are proposed as valuable solvents for hollow-fiber supported LPME (HF-LPME) of organic pollutants from water samples [25]. HF-LPME uses a porous-walled polypropylene hollow fiber to stabilize and protect the organic phase to enhance the mechanical robustness and the extraction efficiency [26,27]. This technique can provide a high analyte preconcentration and excellent sample cleanup, with the advantage that the fiber is disposable after use because of its low price. However, there is still a loss of the organic solvent with a lower boiling point while samples are stirred vigorously. The use of SUPRAS as a liquid membrane phase could overcome these inconveniences due to their negligible vapor pressure and high viscosity.
The aim of the present study is to develop an HF-LPME method using SUPRAS for microextraction of BZDs to eliminate the use of organic solvents. Several factors that affect the extraction efficiency of hollow fiber vesicular-mediated microextraction (HF-VMME) such as HF length, sample pH, stirring rate, extraction time, and salt content were optimized. The proposed method was successfully applied for the determination of trace levels of BZDs in different matrices. 2. Experimental 2.1. Chemicals and reagents All the reagents used were of analytical grade. Five selected BZDs (alprazolam, nitrazepam, clonazepam, midazolam and diazepam) were kindly donated from the Department of Toxicology, Tarbiat Modares University (Tehran, Iran). Decanoic acid was purchased from Fluka (Buchs, Switzerland). Tetrabutylammonium hydroxide (Bu4 NOH, 40%, w/v in water) was obtained from Sigma–Aldrich (Milwaukee, WI, USA). The ultra-pure water was prepared by a model Aqua Max-Ultra Youngling ultra-pure water purification system (Dongan-gu, South Korea). HPLC grade methanol and acetonitrile were purchased from Caledon (Ontario, Canada). Microliter syringes (25–500 L) were purchased from Hamilton (Bonaduz, Switzerland). The Q3/2 Accurel polypropylene microporous hollow-fiber membrane (200 m wall thickness, 600 m inner diameter, 0.2 m pore size, 75% porosity) was obtained from Membrana (Wuppertal, Germany). Stock standard solutions of 1000 g mL−1 BZDs were prepared by dissolving appropriate amount of the compounds in methanol and stored at 4 ◦ C. Working standard solutions were prepared daily by diluting the stock standard solution with ultra-pure water to the required concentrations. 2.2. Apparatus The HPLC instrument consisted of an Agilent (Wilmington, DE, USA) 1200 Series binary pump, micro vacuum degasser, an injector equipped with a 20-L sample loop, a 1200 Series diode array detector, and B.04.03 ChemStation software. An ODS-3 column (250 mm × 4.6 mm, with 5-m particle size) from MZAnalysentechnik (Mainz, Germany) was applied to separate the BZDs. The mobile phase was acetonitrile and 15 mM sodium dihydrogen phosphate with pH 4.6 (45:55) for 15 min and 100% acetonitrile for 5 min at a flow rate of 1 mL min−1 . The diodearray detector allowed the wavelength range from 200 to 400 nm to be scanned to obtain three-dimensional (absorbance × wave length × time) chromatograms. In the present work, absorption measurements were performed at 220 nm, a wavelength very close to the absorption maxima of the five BZDs. Stirring of the solution was carried out on an IKA multi-position magnetic stirrer (Staufen, Germany) using a 12 mm × 4 mm stirring bar. 2.3. Sample preparation (a) Human urine sample was collected from a healthy volunteer. The sample was filtered through a 0.45-m pore size cellulose acetate filter from Millipore (Madrid, Spain). The filtrate was collected in a glass container, which was carefully cleaned with hydrochloric acid and washed with deionized water, and stored at 4 ◦ C to prevent bacterial growth and proteolysis. Twelve milliliter of the urine sample was spiked with the mixed standard solution to obtain desired concentration and diluted to 24 mL with deionized water. Then, pH of the solution was
F. Rezaei et al. / Analytica Chimica Acta 804 (2013) 135–142
137
adjusted at 9.0 by dropwise addition of 0.1 mol L−1 NaOH. These samples were subsequently submitted to HF-VMME procedure. (b) Drug-free human plasma (blood group A+ and O+ ) was obtained from the Iranian Blood Transfusion Organization (Tehran, Iran). The samples were stored at 4 ◦ C, thawed, and shaken before extraction. Five milliliter of the plasma sample was spiked with the mixed standard solution to obtain desired concentration, prepared by the same procedure as urine sample and extracted using HF-VMME method. (c) Fruit juice samples were filtered to remove any suspended material. Then, 8 mL of the filtrate was diluted at 1:2 ratio with deionized water. For preconcentration, pH values of the diluted samples were adjusted at 9.0 using 0.1 mol L−1 NaOH solution before extracting them by the described procedure. (d) Tap water was collected from our lab (Tehran, Iran) and the HF-VMME method was applied to extract the BZDs. The pH of the water samples were adjusted at 9.0 using 0.1 mol L−1 NaOH before analysis. It was found that HF-VMME could not be implemented in some untreated real samples because the solvent became turbid owing to the precipitation of calcium decanoate. The interference was removed by adding EDTA (2 g L−1 ) to the samples before immersing the hollow fiber into the sample solution.
Before extraction, the syringe was rinsed with acetone followed by SUPRAS to avoid carryover and air bubble formation. About 35 L of the supramolecular solvent was withdrawn into the syringe. A piece of the hollow fiber (10 cm length) was fixed onto the tip of the syringe needle and the assembly was immersed in the coacervate phase for 2 min to impregnate the pores of the fiber wall. The fiber was inserted into ultra-pure water for 10 s to wash extra SUPRAS from the hollow fiber surface. Then, the syringe plunger was depressed to fill the hollow fiber with the vesicular coacervate phase. Finally, the end of the hollow fiber was sealed by a piece of aluminum foil. The assembly was then directly immersed into the sample. For each extraction, a new hollow fiber was used to avoid any possible memory effects. The extraction was carried out during the recommended time. After extraction, the acceptor phase solution was retracted into the syringe and injected into the 20-L loop of the HPLC. It is worthy to note that the viscosity of SUPRAS is located between organic solvent and ionic liquid [25]. Regarding application of ionic liquid with higher viscosity in hollow fiber as liquid membrane [28], SUPRAS with less viscosity can be used in this technique more easily. In this work, 30 L of SUPRAS was introduced into the lumen of the hollow fiber and about 28 L was retracted into the syringe at the end of the extraction.
It is worthy to note that if the samples are stored at −20◦ C, benzodiazepine and their metabolites will be stable for at least 12 months [9,10].
3. Results and discussion
2.4. Vesicular coacervate phase formation Vesicular coacervate was obtained by mixing 10.3 g of decanoic acid and 7.8 g of tetrabutylammonium hydroxide in 400 mL distilled water at pH 7.0 (±0.1). To dissolve the decanoic acid, the mixture was stirred at 1200 rpm for 10 min. Finally, phase separation was obtained by centrifugation of the mixture at 4000 rpm for 5 min and the vesicular coacervate solvent obtained was used for further experiments. 2.5. HF-VMME procedure The extractions were performed according to the following procedure: A 24 mL aqueous sample solution (pH ≈ 9) containing 50 g L−1 of each BZD was placed in a 25-mL vial, and a magnetic bar was placed into the solution to ensure efficient stirring during the extraction. The hollow fiber was sonicated for 5 min in acetone to remove any contaminants in the fiber. It was removed from acetone, and the solvent was allowed to evaporate completely.
3.1. Optimization of HF-VMME The SUPRAS used for BZDs microextraction consisted of unilamellar vesicles of decanoic acid (DeA)/tetrabutylammonium decanoate (Bu4 NDe) in the nano- and micro-scale regimes dispersed in a continuous phase. The solvent was produced through a sequential self-assembly process involving two steps. Firstly, DeA and De− formed vesicles in an aqueous solution. The amount and stability of these aggregates were maximal at DeA/De− molar ratios of ca. 1; since in addition to hydrophobic forces, hydrogen bonds between the polar groups of carboxylic and carboxylate molecules were the major force driving their self-assembly. Secondly, aggregate growth was promoted by reduction of ionic head group repulsion with the counterion Bu4 N+ . Thus Bu4 N+ acts as a coacervating agent. In its presence, large non-water-soluble vesicles are formed as a result of neutralization of De− molecules, which avoids repulsive electrostatic interactions between them in the aggregate. The DeA–Bu4 NDe vesicles formed were separated from the aqueous solution as an immiscible liquid (the supramolecular solvent) with density lower than water [29–31]. The aggregates in the SUPRAS are expected to efficiently extract BZDs according to
Fig. 1. A schematic representation of HF-VMME.
138
F. Rezaei et al. / Analytica Chimica Acta 804 (2013) 135–142
Fig. 2. Effect of sample pH on the peak area of the drugs. Extraction conditions: volume of the aqueous solution, 24.0 mL; hollow fiber length, 8 cm; stirring rate, 800 rpm; extraction time, 40 min.
their structure and the parameters related to their extraction, as schematically shown in Fig. 1. The pH value of the sample solution is one of the most important factors affecting the stability of the coacervate phase and extraction efficiency. Carboxylic acid molecules were strongly hydrogen-bonded to each other and it is proposed that hydrogen bond intermolecular forces strongly increase the cohesion among molecules. Generally, sample solution pH determines the state of DeA in the aqueous solution, which plays an important role in the stability of vesicles in aqueous samples. Experimental results show that the maximum amount of vesicles are formed at pH = pKa,DeA (≈7) [21]. However, since the target compounds are weak bases, the extraction efficiency was studied in a wide range of acidic to basic medium (pH values from 3.0 to 11.0). As can be seen in Fig. 2, at pH values lower than 5.0; especially under acidic conditions (pH 3.0), the signals for all BZDs significantly decreased. Extraction of the neutral analyte into the extraction phase is eligible because of conventional interactions. Furthermore, extraction of positively charged (protonated) species seems as an interesting phenomenon, which occurred in the acidic medium for BZDs. This can be a result of ion pair formation between decanoate and protonated analytes. The peak area increases when the pH value of sample solution is increased from 5.0 to 9.0, but then decreases by further increase of pH. It was thus deduced that pH 9.0 is the most suitable value for the extraction. Addition of salt to the sample may have different effects. It can reduce the amount of water available to dissolve analyte molecules due to formation of hydration spheres around the ionic salt molecules and can improve the extraction efficiency for the target analytes into the extraction phase. On the other hand, addition of salt increases the ionic strength of the sample and changes the physical properties of the Nernst diffusion film; it can reduce the rate of diffusion of the analyte into the extraction phase and reduces the extraction efficiency. In this work, effect of salt on the extraction efficiency of the target compounds was investigated by adding sodium chloride from 0 to 15% (w/v) to the sample. According to Fig. 3, addition of salt in this range has a negative effect on the extraction efficiency. This may be due to the increase in viscosity that in turn decreases the mass transfer of the analyte to the extraction solvent. Hence, further extractions were performed without adding NaCl to the solutions. The acceptor phase volume is another important factor affecting the preconcentration factor of the analytes. By increasing the acceptor phase volume, because of dilution effect, the preconcentration factor is reduced. On the other hand, the acceptor phase volume should be large enough to promote analyte transport to the acceptor phase. Generally, an increase in the acceptor volume enhances
Fig. 3. Effect of salt addition on the peak area of BZDs. Extraction conditions: as in the Fig. 2; except that pH of the aqueous solution is 9.0.
the preconcentration factor as much as it does not lead to dilution of the extractant. The acceptor phase volume is proportional to the length of the hollow fiber. Therefore, the influence of the acceptor phase volume was studied by varying of hollow fiber length in the range of 6–11 cm. As demonstrated in Fig. 4, the relative peak area increases up to 10 cm, because the rate of the analyte transport into a solvent is directly related to the interfacial area between the two liquid phases. The results demonstrated that an increase in the HF length from 10 to 11 cm will result in a decrease in the peak areas of the analytes because of the dilution effect. Therefore, HF length of 10 cm was selected in the subsequent experiments. As is known, stirring speed plays an essential role in increasing the kinetics and efficiency of extraction by increasing the mass transfer and reducing the thickness of double layer around SLM. In HF-LPME, the extraction solvent is protected by the hydrophobic hollow fiber; so, high stirring speed can be applied compared with single-drop LPME. On the other hand, this parameter has to be carefully considered because a high stirring speed may cause the fall of the fiber in the sample and a whirlpool appears in the vial, which reduces the effective extraction surface between the hollow fiber and the aqueous sample. In addition, very high stirring rate may lose solvent that could affect the precision. To attain optimal stirring speeds, different stirring speeds ranging from 400 to 1000 rpm were examined. The results showed that the stirring speed of 900 rpm yielded the highest peak area and it was selected for subsequent experiments. In HF-LPME, mass transfer is a time-dependent process; thus, one of the most important parameters affecting the extraction efficiency is the extraction time. As SPME, HF-LPME is not an exhaustive extraction. It can take from several minutes to hours to reach an equilibrium that ensures an optimal extraction recovery
Fig. 4. Effect of hollow fiber length on the peak area of the drugs. Extraction conditions: volume of the aqueous solution, 24.0 mL; pH of the aqueous solution, 9.0; stirring rate, 800 rpm; extraction time, 40 min; concentration of the drugs, 50 g L−1 .
F. Rezaei et al. / Analytica Chimica Acta 804 (2013) 135–142
139
Table 1 Figures of merit of HF-VMMEa method for analyzing the selected BZDs in water, urine, and plasma samples. Sample
Analyte
R2
LOD (g L−1 )b
Deionized water
Alprazolam Nitrazepam Clonazepam Midazolam Diazepam
0.9956 0.9985 0.9993 0.9954 0.9966
0.7 0.7 0.7 0.7 0.5
Urine
Alprazolam Nitrazepam Clonazepam Midazolam Diazepam
0.9976 0.9907 0.9916 0.9945 0.9989
Plasma
Alprazolam Nitrazepam Clonazepam Midazolam Diazepam
0.9944 0.9937 0.9967 0.9984 0.9946
a b c d e
LDR (g L−1 )c
RSD (%)d
PFe
2.0–200.0 2.0–200.0 2.0–200.0 2.0–200.0 2.0–200.0
4.3 5.8 5.2 6.1 5.7
133 112 141 169 198
2.0 2.0 2.0 2.0 2.0
5.0–200.0 5.0–200.0 5.0–200.0 5.0–200.0 5.0–200.0
7.6 5.2 7.3 8.9 6.7
61 45 69 75 90
3.0 3.0 3.0 3.0 3.0
7.0–200.0 7.0–200.0 7.0–200.0 7.0–200.0 7.0–200.0
6.2 4.8 5.9 7.0 5.3
53 38 57 66 81
Hollow fiber vesicular-mediated microextraction. Limit of detection. Linear dynamic range. Relative standard deviation. Preconcentration factor.
and preconcentration factor. If time is chosen too short, the efficiency of extraction equilibrium may not reach. On the contrary, if the extraction time is chosen too long, solvent loss and formation of air bubbles may occur, which would decrease the extraction efficiency. Since SUPRAS have higher viscosity than conventional organic solvents, the diffusion of target species through the membrane and so the extraction rate will slightly be slower. The effect of extraction times in the range of 20–60 min on the extraction efficiency was investigated. The amounts of extracted analytes were increased dramatically by increasing the exposure time from 20 to 50 min and thereafter the curves became flat. On the basis of the results, 50 min was selected as the extraction time. 3.2. Quantitative analysis To evaluate practical applicability of the proposed HF-VMME technique; linearity, relative standard deviations (RSDs), limits of detection (LODs), and preconcentration factors (PFs) were investigated by extraction of BZDs from water samples under the optimal conditions (Table 1). In the first experiment, three-replicate extractions and determinations of BZDs (10 g L−1 of each analyte in the
water samples) were performed by HF-VMME/HPLC-UV and the RSDs% in the range of 4.3–6.1% were obtained for extraction of the analytes from water samples. In a subsequent experiment, linear ranges of 1.0–200.0 g L−1 for diazepam and 2.0–200.0 g L−1 for other analytes with coefficient of determinations (R2 ) of 0.9956–0.9993 for BZDs were obtained in the water samples. Based on the signal to noise ratio of 3 (S/N = 3), LODs of BZDs varied in the range as 0.5–0.7 g L−1 . For comparison, some figures of merit of previously reported methods such as LODs, LDRs, and RSD% for extraction and determination of BZDs are summarized in Table 2. As can be seen, the proposed HF-VMME method has a good sensitivity and proper precision with a suitable dynamic linear range. Also, the LODs obtained for the drugs by the present method are better or comparable with those obtained by other methods with similar detection instruments. 3.3. Analysis of real samples In order to evaluate applicability of the developed extraction method for analysis of BZDs in real samples with complex matrices,
Fig. 5. HPLC-DAD chromatograms of fruit juice (a) non-spiked and (b) spiked by 10 g L−1 of the target analytes, after HF-VMME; (1) alprazolam, (2) nitrazepam, (3) clonazepam, (4) midazolam, and (5) diazepam.
140
F. Rezaei et al. / Analytica Chimica Acta 804 (2013) 135–142
Table 2 Comparison of proposed method with other methods for analyzing the selected BZDs. Method
Analyte
LOD (g L−1 )
LDR (g L−1 )
RSD (%)
Ref.
LLE-HPLC
Diazepam Clonazepam
30 50
50–1200 75–1200
12.3 9.7
[15]
SPE-LC-MS
Alprazolam Nitrazepam Clonazepam Midazolam Diazepam
0.4 0.2 0.3 1.0 4.0
1–500 1–1000 1–1000 5–2000 13–1000
6 6 6 8 4
a
Alprazolam Clonazepam Diazepam
3.5 8.1 2.8
8.1–5000 18.8–5000 6.4–5000
5.6 12.7 3.3
[12]
SPE-HPLC-UV
Alprazolam Clonazepam Diazepam
1.0 2.0 2.5
5.0–200.0 5.0–200.0 7.5–500.0
7.1 7.1 7.1
[32]
SPME-LC-ESI-MS
Clonazepam Diazepam
2 1
2–500 1–500
4.0 4.0
[33]
SPME-LC
Nitrazepam Diazepam
4 1
5–1000 5–400
9.5 13.0
[34]
SPE-LC-ESI-MS/MS
Alprazolam Clonazepam Midazolam
0.02 0.02 0.02
0.05–50 0.05–50 0.05–50
15 15 15
[35]
b
Alprazolam Diazepam
7.4 6.2
– –
2.4 2.1
[36]
SPE-HPLC-DAD
Alprazolam Diazepam
8.8 10
– –
3.5 5.5
[36]
HF-VMME-HPLC-DAD
Alprazolam Nitrazepam Clonazepam Midazolam Diazepam
2.0–200.0 2.0–200.0 2.0–200.0 2.0–200.0 1.0–200.0
4.3 5.8 5.2 6.1 5.7
Proposed method
a b
UA-DLLME-HPLC
MAE-HPLC-DAD
0.7 0.7 0.7 0.7 0.5
[8]
Ultrasound-assisted dispersive liquid–liquid microextraction. Microwave assisted extraction.
tap water, fruit juice, urine and plasma samples were selected and the drugs were extracted using the proposed method under the optimal conditions. Sample preparation for real samples was performed according to Section 2.3. One of the criminal applications of BZDs is the addition of these drugs into fruit juice to make the consumer sleepy for robbery. Extraction of BZDs in fruit juice which was performed in this paper recommended by Research Center of Antinarcotics Police of Iran (Tehran) to develop a simple and easy method for analysis of these drugs in fruit juice. Different samples were analyzed by HF-VMME followed with HPLC-DAD. The results showed that the samples were free from contamination of the drug. Thus, all of the real aqueous samples were spiked with the BZDs standard and then they were prepared as it was discussed in Section 2.3 to assess the matrix effects. The results of analysis are shown in Table 3. Accuracy was calculated as the relative recoveries for analysis of known amounts of target analytes added to actual water samples using the proposed method. Relative recovery (RR%) was acquired from the following equation: RR% =
Cfound − Creal × 100 Cadded
10 g L−1 -spiked fruit juice sample. Fig. 6 shows chromatograms of the non-spiked and 15 g L−1 -spiked urine sample and Fig. 7 shows chromatograms of non-spiked and spiked plasma sample at the concentration level of 15 g L−1 BZDs standards. As can be seen in Figs. 6 and 7, this method is a good clean-up procedure that can be used for trace analysis of BZDs in complicated matrices such as urine and plasma samples.
(1)
where Cfound , Creal , and Cadded are concentration of analyte after addition of a known amount of standard into the real sample, concentration of analyte in real sample, and concentration of a known amount of standard, which was spiked into the real sample, respectively. The RR% of the BZDs were in the range of 90.0–98.8%. Intraand inter-assay precisions ranged between 4.4–8.2% and 7.6–12.1%, respectively. Fig. 5 shows chromatograms of the non-spiked and
Fig. 6. HPLC-DAD chromatograms of urine (a) non-spiked and (b) spiked by 15 g L−1 of the target analytes, after HF-VMME; (1) alprazolam, (2) nitrazepam, (3) clonazepam, (4) midazolam, and (5) diazepam.
F. Rezaei et al. / Analytica Chimica Acta 804 (2013) 135–142
141
Table 3 Results obtained for analysis of BZDs in real samples. Samplea
Alprazolam
Nitrazepam
Clonazepam
Midazolam
Diazepam
Tap water
Foundb (addedc ) RR% RSD (n = 3) Found (added) RR% RSD (n = 3)
9.64 (10.00) 96.4 6.3 24.2 (25.0) 96.8 5.4
9.88 (10.00) 98.8 5.2 23.5 (25.0) 94.0 4.8
9.54 (10.00) 95.4 5.4 24.5 (25.0) 98.0 5.9
9.16 (10.00) 91.6 6.6 23.8 (25.0) 95.2 6.1
9.40 (10.00) 94.0 5.8 23.2 (25.0) 92.8 4.4
Orange juice
Found (added) RR% RSD (n = 3) Found (added) RR% RSD (n = 3)
9.16 (10.00) 91.6 8.2 22.8 (25.0) 91.2 7.1
9.36 (10.00) 93.6 5.6 23.1 (25.0) 92.4 6.4
9.27 (10.00) 92.7 7.4 23.3 (25.0) 93.2 6.8
9.44 (10.00) 94.4 5.6 24.1 (25.0) 96.4 7.5
9.37 (10.00) 93.7 6.3 22.7 (25.0) 90.8 5.9
Sour cherry juice
Found (added) RR% RSD (n = 3) Found (added) RR% RSD (n = 3)
9.24 (10.00) 92.4 5.7 23.6 (25.0) 94.4 6.1
9.56 (10.00) 95.6 6.2 22.7 (25.0) 90.8 5.6
9.19 (10.00) 91.9 6.4 23.9 (25.0) 95.6 6.1
9.08 (10.00) 90.8 7.3 23.5 (25.0) 94.0 6.0
9.52 (10.00) 95.2 5.5 24.6 (25.0) 98.4 5.3
Delster
Found (added) RR% RSD (n = 3) Found (added) RR% RSD (n = 3)
9.35 (10.00) 93.5 5.9 23.9 (25.0) 95.6 6.0
9.76 (10.00) 97.6 6.1 24.1 (25.0) 96.4 6.3
9.49 (10.00) 94.9 7.2 23.6 (25.0) 94.4 6.5
9.43 (10.00) 94.3 6.5 24.1 (25.0) 96.4 5.9
9.58 (10.00) 95.8 6.8 23.5 (25.0) 94.0 6.4
Urine
Found (added) RR% RSD (n = 3) Found (added) RR% RSD (n = 3)
9.02 (10.00) 90.2 5.3 23.2 (25.0) 92.8 6.4
9.06 (10.00) 90.6 5.6 23.6 (25.0) 94.4 7.2
9.17 (10.00) 91.7 7.1 22.9 (25.0) 91.6 6.9
9.34 (10.00) 93.4 5.6 23.1 (25.0) 92.4 5.7
9.22 (10.00) 92.2 6.5 23.4 (25.0) 93.6 6.1
Plasma
Found (added) RR% RSD (n = 3) Found (added) RR% RSD (n = 3)
14.5 (15.0) 96.7 6.7 27.2 (30.0) 90.7 6.5
13.8 (15.0) 92.0 7.9 28.9 (30.0) 96.3 4.8
13.6 (15.0) 90.7 7.4 27.6 (30.0) 92.0 7.4
14.4 (15.0) 96.0 5.3 29.4 (30.0) 98.0 6.9
13.5 (15.0) 90.0 6.1 28.3(30.0) 94.3 5.8
Fig. 7. HPLC-DAD chromatograms of plasma (a) non-spiked and (b) spiked by 15 g L−1 of the target analytes, after HF-VMME; (1) alprazolam, (2) nitrazepam, (3) clonazepam, (4) midazolam, and (5) diazepam.
4. Conclusions The supramolecular solvent based on DeA–Bu4 NDe vesicles constitutes a suitable alternative to organic molecular ones to extract BZDs from water samples. This solvent is produced from environmental friendly and available reagents in a simple way and
the extraction process is simple, rapid, and cheap. SUPRASs consist of amphiphilic nanostructures that provide multiple binding sites and regions of different polarity. These outstanding properties make them suitable for extraction of a variety of analytes with high efficiency and render them ideal for microextractions. The high concentration of DeA in the extractant phase (∼1 mg L−1 )
142
F. Rezaei et al. / Analytica Chimica Acta 804 (2013) 135–142
and capability of analyte solubilization of the vesicles permit favorable partition of analytes using a quite low volume of SUPRAS (30 L). Extractions were based on hydrophobic, hydrogen bond, and -cation interactions between the analytes and the vesicular aggregates. So, coacervates can be considered as multifunctional solvents that have the capability to efficiently extract a wide variety of compounds. The vesicular coacervate phase possesses relatively high viscosities, low vapor pressure, and high cohesion between the aggregates. These properties make its use very attractive in obtaining stable SLM. In this work, BZDs were extracted from water, urine, and juice samples into the supramolecular solvent impregnated in the pores and also filled inside the porous polypropylene hollow fiber membrane. Moreover, by using SUPRAS as the extraction phase, the need for using toxic organic solvents was eliminated. References [1] M. Nakamura, Biomed. Chromatogr. 25 (2011) 1283–1307. [2] E.N. Sauve, M. Langødegård, D. Ekeberg, Å.M.L. Øiestad, J. Chromatogr. B 883–884 (2012) 177–188. [3] H.H. Maurer, J. Chromatogr. 580 (1992) 3–41. [4] S. Bourcier, Y. Hoppilliard, Eur. J. Mass Spectrom. 7 (2001) 359–371. [5] L. Wang, H. Zhao, Y. Qiu, Z. Zhou, J. Chromatogr. A 1136 (2006) 99–105. [6] R. Cordero, S. Paterson, J. Chromatogr. B 850 (2007) 423–431. [7] K. Arnhard, R. Schmid, U. Kobold, R. Thiele, Anal. Bioanal. Chem. 403 (2012) 755–768. [8] B.E. Smink, J.E. Brandsma, A. Dijkhuizen, K.J. Lusthof, J.J. de Gier, A.C.G. Egberts, D.R.A. Uges, J. Chromatogr. B 811 (2004) 13–20. [9] A. El Mahjoub, C. Staub, J. Pharm. Biomed. Anal. 23 (2000) 1057–1063. [10] A. El Mahjoub, C. Staub, J. Pharm. Biomed. Anal. 23 (2000) 447–458. [11] K.E. Rasmussen, S. Pedersen-Bjergaard, M. Krogh, H.G. Ugland, T. Grønhaug, J. Chromatogr. A 873 (2000) 3–11.
[12] P. Fernández, C. Gonzálezb, M.T. Pena, A.M. Carro, R.A. Lorenzo, Anal. Chim. Acta 767 (2013) 88–96. [13] H.L. Lord, M. Rajabi, S. Safari, J. Pawliszyn, J. Pharm. Biomed. Anal. 44 (2007) 506–519. [14] W.F. Smyth, S. McClean, Electrophoresis 19 (1998) 2870–2882. [15] K.B. Borges, E.F. Freire, I. Martins, M.E.P.B. Siqueira, Talanta 78 (2009) 233–241. [16] R.A. Anderson, M.M. Ariffin, P.A.G. Cormack, E.I. Miller, Forensic Sci. Int. 174 (2008) 40–46. [17] V.F. Samanidou, A.P. Pechlivanidou, I.N. Papadoyannis, J. Sep. Sci. 30 (2007) 679–687. [18] H.G. Ugland, M. Krogh, K.E. Rasmussen, J. Chromatogr. B 749 (2000) 85–92. [19] H.G. Ugland, M. Krogh, L. Reubsaet, J. Chromatogr. B 798 (2003) 127–135. [20] F. Rezaei, Y. Yamini, M. Moradi, B. Ebrahimpour, Talanta 105 (2013) 173–178. [21] F.J. Ruiz, S. Rubio, D. Pérez-Bendito, Anal. Chem. 78 (2006) 7229–7239. [22] F.J. López-Jiménez, S. Rubio, D. Pérez-Bendito, J. Chromatogr. A 1195 (2008) 25–33. [23] M. Moradi, Y. Yamini, J. Chromatogr. A 1229 (2012) 30–37. [24] S. Jafarvand, F. Shemirani, J. Sep. Sci. 34 (2011) 455–461. [25] M. Moradi, Y. Yamini, F. Rezaei, E. Tahmasebi, A. Esrafili, Analyst 137 (2012) 3549–3557. [26] S. Pedersen-Bjergaard, K.E. Rasmussen, Anal. Chem. 71 (1999) 2650–2656. [27] S. Pedersen-Bjergaard, K.E. Rasmussen, J. Chromatogr. A 1184 (2008) 132– 142. [28] R. Fortunato, C.A.M. Afonso, M.A.M. Reis, J.G. Crespo, J. Membr. Sci. 242 (2004) 197–209. [29] F.J. Ruiz, S. Rubio, D. Pérez-Bendito, J. Chromatogr. A 1163 (2007) 269–276. [30] A. Moral, M.D. Sicilia, S. Rubio, Anal. Chim. Acta 650 (2009) 207–213. [31] A. Moral, M.D. Sicilia, S. Rubio, J. Chromatogr. A 1216 (2009) 3740–3745. [32] L. Mercolini, R. Mandrioli, M. Amore, M.A. Raggi, J. Sep. Sci. 31 (2008) 2619–2626. [33] H. Yuan, Z. Mester, H. Lord, J. Pawliszyn, J. Anal. Toxicol. 24 (2000) 718–725. [34] K. Jinno, M. Taniguchi, M. Hayashida, J. Pharm. Biomed. Anal. 17 (1998) 1081–1091. [35] O. Quintela, F.L. Sauvage, F. Charvier, J.M. Gaulier, G. Lachâtre, P. Marquet, Clin. Chem. 52 (2006) 1346–1355. [36] P. Fernández, C. Vázquez, R.A. Lorenzo, A.M. Carro, I. Álvarez, P. Cabarcos, Anal. Bioanal. Chem. 397 (2010) 677–685.
Contents lists available at ScienceDirect
Analytica Chimica Acta journal homepage: www.elsevier.com/locate/aca
Supramolecular solvent-based hollow fiber liquid phase microextraction of benzodiazepines Fatemeh Rezaei a , Yadollah Yamini a,∗ , Morteza Moradi b , Bahram Daraei c a b c
Department of Chemistry, Tarbiat Modares University, P.O. Box 14115-175, Tehran, Iran Department of Semiconductors, Materials and Energy Research Center, Karaj, Iran Department of Toxicology, Tarbiat Modares University, Tehran, Iran
h i g h l i g h t s
g r a p h i c a l
a b s t r a c t
• Hollow fiber microextraction based • • • •
on supramolecular solvents was developed. Supramolecular solvent was produced from coacervation of decanoic acid and Bu4 N+ . The method was introduced for extraction of benzodiazepines from aqueous solutions. Several parameters affecting extraction efficiency were investigated and optimized. HPLC-DAD was applied for the determination of the drugs in fruit juice and urine samples.
a r t i c l e
i n f o
Article history: Received 7 August 2013 Received in revised form 9 October 2013 Accepted 11 October 2013 Available online 19 October 2013 Keywords: Hollow fiber Vesicles Benzodiazepines High performance liquid chromatography Urine Plasma
a b s t r a c t A new, efficient, and environmental friendly hollow fiber liquid phase microextraction (HF-LPME) method based on supramolecular solvents was developed for extraction of five benzodiazepine drugs. The supramolecular solvent was produced from coacervation of decanoic acid aqueous vesicles in the presence of tetrabutylammonium (Bu4 N+ ). In this work, benzodiazepines were extracted from aqueous samples into a supramolecular solvent impregnated in the wall pores and also filled inside the porous polypropylene hollow fiber membrane. The driving forces for the extraction were hydrophobic, hydrogen bonding, and -cation interactions between the analytes and the vesicular aggregates. High-performance liquid chromatography with photodiode array detection (HPLC-DAD) was applied for separation and determination of the drugs. Several parameters affecting the extraction efficiency including pH, hollow fiber length, ionic strength, stirring rate, and extraction time were investigated and optimized. Under the optimal conditions, the preconcentration factors were obtained in the range of 112–198. Linearity of the method was determined to be in the range of 1.0–200.0 g L−1 for diazepam and 2.0–200.0 g L−1 for other analytes with coefficient of determination (R2 ) ranging from 0.9954 to 0.9993. The limits of detection for the target benzodiazepines were in the range of 0.5–0.7 g L−1 . The method was successfully applied for extraction and determination of the drugs in water, fruit juice, plasma and urine samples and relative recoveries of the compounds studied were in the range of 90.0–98.8%. © 2013 Elsevier B.V. All rights reserved.
1. Introduction
∗ Corresponding author. Tel.: +98 21 82883417; fax: +98 21 88006544. E-mail addresses: [email protected], [email protected] (Y. Yamini). 0003-2670/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.aca.2013.10.026
Benzodiazepines (BZDs) are a large group of drugs with a broad range of therapeutic effects, used as hypnotics, anxiolytics, muscle relaxants, and anticonvulsants [1]. The sedative and amnestic
136
F. Rezaei et al. / Analytica Chimica Acta 804 (2013) 135–142
properties of some BZDs are also considered useful in anesthesia. In addition, the BZDs have a rapid onset of action combined with low acute toxicity [2]. Several major metabolites of BZDs possess pharmacological profiles similar to the parent drugs. BZDs are now among the most commonly prescribed drugs, which increases their potential for addiction and abuse in cases of crime, driving under the influence of drugs, suicide, and drug facilitated sexual assault. For these reasons, simultaneous analysis of BZDs and their metabolites in complicated matrices is of great interest to clinicians and forensic toxicologists [1]. A variety of methods exist in the literature for detection of benzodiazepines in complicated matrices. Gas chromatography–mass spectrometry (GC–MS) determinations of BZDs have been reviewed by Maurer [3] and further studies have also been carried out [4–7]. High-performance liquid chromatography (HPLC) methods have been studied [8–13] and a critical evaluation of the application of capillary electrophoresis (CE) to detection of BZDs in formulations and body materials have been made [14]. Chromatographic techniques usually require complex isolation procedures (sample preparation) to separate BZDs from complicated matrices. Several sample preparation procedures such as liquid-liquid extraction (LLE) [15], solid phase extraction [16,17], and liquid phase microextraction [18,19] have been developed for separation and preconcentration of drugs from different matrices. However, the LLE and SPE methods often require large amounts of organic solvents, some of which are harmful and contaminate the environment due to their high vapor pressure. The pressure to decrease organic solvent usage in laboratories is increasing. Thus, miniaturization and improvement of sample handling using alternatives is a challenge that has been discussed by several researchers. From this perspective, supramolecular solvents (SUPRASs)-based extractions were an appropriate choice [20]. In 2006, Pérez-Bendito’s research group for the first time investigated the potential of the tetrabutylammonium (TBA)induced liquid-phase separation in alkyl carboxylic acid vesicular solutions for extraction of organic compounds prior to HPLC [21]. SUPRASs are water-immiscible liquids made up of supramolecular assemblies dispersed in a continuous phase. Two characteristics give the alkyl carboxylic acid-based coacervates a high potential for analytical extraction processes. First, the polar region of molecular aggregates consists of protonated and deprotonated carboxylic groups and ammonium groups; so, a number of interactions (e.g., electrostatics, -cation, hydrogen bonds, formation of mixed aggregates, etc.) can be established with analytes, in addition to hydrophobic interactions in the hydrocarbon region. Second, vesicles have a number of available solubilization sites; so, high concentrations of polar and apolar molecules can be solubilized in each aggregate. Formation of vesicles in the aqueous solution before adding Bu4 N+ ions was not essential to achieve liquid-liquid-phase separation. Several extraction methods, such as single-drop microextraction (SDME) [22], floating drop microextraction (FDME) [23], dispersive liquid–liquid microextraction (DLLME) [24] have been reported based on SUPRASs. In our previous work, SUPRAS comprised of tetrabutylammonium-induced vesicles of decanoic acid are proposed as valuable solvents for hollow-fiber supported LPME (HF-LPME) of organic pollutants from water samples [25]. HF-LPME uses a porous-walled polypropylene hollow fiber to stabilize and protect the organic phase to enhance the mechanical robustness and the extraction efficiency [26,27]. This technique can provide a high analyte preconcentration and excellent sample cleanup, with the advantage that the fiber is disposable after use because of its low price. However, there is still a loss of the organic solvent with a lower boiling point while samples are stirred vigorously. The use of SUPRAS as a liquid membrane phase could overcome these inconveniences due to their negligible vapor pressure and high viscosity.
The aim of the present study is to develop an HF-LPME method using SUPRAS for microextraction of BZDs to eliminate the use of organic solvents. Several factors that affect the extraction efficiency of hollow fiber vesicular-mediated microextraction (HF-VMME) such as HF length, sample pH, stirring rate, extraction time, and salt content were optimized. The proposed method was successfully applied for the determination of trace levels of BZDs in different matrices. 2. Experimental 2.1. Chemicals and reagents All the reagents used were of analytical grade. Five selected BZDs (alprazolam, nitrazepam, clonazepam, midazolam and diazepam) were kindly donated from the Department of Toxicology, Tarbiat Modares University (Tehran, Iran). Decanoic acid was purchased from Fluka (Buchs, Switzerland). Tetrabutylammonium hydroxide (Bu4 NOH, 40%, w/v in water) was obtained from Sigma–Aldrich (Milwaukee, WI, USA). The ultra-pure water was prepared by a model Aqua Max-Ultra Youngling ultra-pure water purification system (Dongan-gu, South Korea). HPLC grade methanol and acetonitrile were purchased from Caledon (Ontario, Canada). Microliter syringes (25–500 L) were purchased from Hamilton (Bonaduz, Switzerland). The Q3/2 Accurel polypropylene microporous hollow-fiber membrane (200 m wall thickness, 600 m inner diameter, 0.2 m pore size, 75% porosity) was obtained from Membrana (Wuppertal, Germany). Stock standard solutions of 1000 g mL−1 BZDs were prepared by dissolving appropriate amount of the compounds in methanol and stored at 4 ◦ C. Working standard solutions were prepared daily by diluting the stock standard solution with ultra-pure water to the required concentrations. 2.2. Apparatus The HPLC instrument consisted of an Agilent (Wilmington, DE, USA) 1200 Series binary pump, micro vacuum degasser, an injector equipped with a 20-L sample loop, a 1200 Series diode array detector, and B.04.03 ChemStation software. An ODS-3 column (250 mm × 4.6 mm, with 5-m particle size) from MZAnalysentechnik (Mainz, Germany) was applied to separate the BZDs. The mobile phase was acetonitrile and 15 mM sodium dihydrogen phosphate with pH 4.6 (45:55) for 15 min and 100% acetonitrile for 5 min at a flow rate of 1 mL min−1 . The diodearray detector allowed the wavelength range from 200 to 400 nm to be scanned to obtain three-dimensional (absorbance × wave length × time) chromatograms. In the present work, absorption measurements were performed at 220 nm, a wavelength very close to the absorption maxima of the five BZDs. Stirring of the solution was carried out on an IKA multi-position magnetic stirrer (Staufen, Germany) using a 12 mm × 4 mm stirring bar. 2.3. Sample preparation (a) Human urine sample was collected from a healthy volunteer. The sample was filtered through a 0.45-m pore size cellulose acetate filter from Millipore (Madrid, Spain). The filtrate was collected in a glass container, which was carefully cleaned with hydrochloric acid and washed with deionized water, and stored at 4 ◦ C to prevent bacterial growth and proteolysis. Twelve milliliter of the urine sample was spiked with the mixed standard solution to obtain desired concentration and diluted to 24 mL with deionized water. Then, pH of the solution was
F. Rezaei et al. / Analytica Chimica Acta 804 (2013) 135–142
137
adjusted at 9.0 by dropwise addition of 0.1 mol L−1 NaOH. These samples were subsequently submitted to HF-VMME procedure. (b) Drug-free human plasma (blood group A+ and O+ ) was obtained from the Iranian Blood Transfusion Organization (Tehran, Iran). The samples were stored at 4 ◦ C, thawed, and shaken before extraction. Five milliliter of the plasma sample was spiked with the mixed standard solution to obtain desired concentration, prepared by the same procedure as urine sample and extracted using HF-VMME method. (c) Fruit juice samples were filtered to remove any suspended material. Then, 8 mL of the filtrate was diluted at 1:2 ratio with deionized water. For preconcentration, pH values of the diluted samples were adjusted at 9.0 using 0.1 mol L−1 NaOH solution before extracting them by the described procedure. (d) Tap water was collected from our lab (Tehran, Iran) and the HF-VMME method was applied to extract the BZDs. The pH of the water samples were adjusted at 9.0 using 0.1 mol L−1 NaOH before analysis. It was found that HF-VMME could not be implemented in some untreated real samples because the solvent became turbid owing to the precipitation of calcium decanoate. The interference was removed by adding EDTA (2 g L−1 ) to the samples before immersing the hollow fiber into the sample solution.
Before extraction, the syringe was rinsed with acetone followed by SUPRAS to avoid carryover and air bubble formation. About 35 L of the supramolecular solvent was withdrawn into the syringe. A piece of the hollow fiber (10 cm length) was fixed onto the tip of the syringe needle and the assembly was immersed in the coacervate phase for 2 min to impregnate the pores of the fiber wall. The fiber was inserted into ultra-pure water for 10 s to wash extra SUPRAS from the hollow fiber surface. Then, the syringe plunger was depressed to fill the hollow fiber with the vesicular coacervate phase. Finally, the end of the hollow fiber was sealed by a piece of aluminum foil. The assembly was then directly immersed into the sample. For each extraction, a new hollow fiber was used to avoid any possible memory effects. The extraction was carried out during the recommended time. After extraction, the acceptor phase solution was retracted into the syringe and injected into the 20-L loop of the HPLC. It is worthy to note that the viscosity of SUPRAS is located between organic solvent and ionic liquid [25]. Regarding application of ionic liquid with higher viscosity in hollow fiber as liquid membrane [28], SUPRAS with less viscosity can be used in this technique more easily. In this work, 30 L of SUPRAS was introduced into the lumen of the hollow fiber and about 28 L was retracted into the syringe at the end of the extraction.
It is worthy to note that if the samples are stored at −20◦ C, benzodiazepine and their metabolites will be stable for at least 12 months [9,10].
3. Results and discussion
2.4. Vesicular coacervate phase formation Vesicular coacervate was obtained by mixing 10.3 g of decanoic acid and 7.8 g of tetrabutylammonium hydroxide in 400 mL distilled water at pH 7.0 (±0.1). To dissolve the decanoic acid, the mixture was stirred at 1200 rpm for 10 min. Finally, phase separation was obtained by centrifugation of the mixture at 4000 rpm for 5 min and the vesicular coacervate solvent obtained was used for further experiments. 2.5. HF-VMME procedure The extractions were performed according to the following procedure: A 24 mL aqueous sample solution (pH ≈ 9) containing 50 g L−1 of each BZD was placed in a 25-mL vial, and a magnetic bar was placed into the solution to ensure efficient stirring during the extraction. The hollow fiber was sonicated for 5 min in acetone to remove any contaminants in the fiber. It was removed from acetone, and the solvent was allowed to evaporate completely.
3.1. Optimization of HF-VMME The SUPRAS used for BZDs microextraction consisted of unilamellar vesicles of decanoic acid (DeA)/tetrabutylammonium decanoate (Bu4 NDe) in the nano- and micro-scale regimes dispersed in a continuous phase. The solvent was produced through a sequential self-assembly process involving two steps. Firstly, DeA and De− formed vesicles in an aqueous solution. The amount and stability of these aggregates were maximal at DeA/De− molar ratios of ca. 1; since in addition to hydrophobic forces, hydrogen bonds between the polar groups of carboxylic and carboxylate molecules were the major force driving their self-assembly. Secondly, aggregate growth was promoted by reduction of ionic head group repulsion with the counterion Bu4 N+ . Thus Bu4 N+ acts as a coacervating agent. In its presence, large non-water-soluble vesicles are formed as a result of neutralization of De− molecules, which avoids repulsive electrostatic interactions between them in the aggregate. The DeA–Bu4 NDe vesicles formed were separated from the aqueous solution as an immiscible liquid (the supramolecular solvent) with density lower than water [29–31]. The aggregates in the SUPRAS are expected to efficiently extract BZDs according to
Fig. 1. A schematic representation of HF-VMME.
138
F. Rezaei et al. / Analytica Chimica Acta 804 (2013) 135–142
Fig. 2. Effect of sample pH on the peak area of the drugs. Extraction conditions: volume of the aqueous solution, 24.0 mL; hollow fiber length, 8 cm; stirring rate, 800 rpm; extraction time, 40 min.
their structure and the parameters related to their extraction, as schematically shown in Fig. 1. The pH value of the sample solution is one of the most important factors affecting the stability of the coacervate phase and extraction efficiency. Carboxylic acid molecules were strongly hydrogen-bonded to each other and it is proposed that hydrogen bond intermolecular forces strongly increase the cohesion among molecules. Generally, sample solution pH determines the state of DeA in the aqueous solution, which plays an important role in the stability of vesicles in aqueous samples. Experimental results show that the maximum amount of vesicles are formed at pH = pKa,DeA (≈7) [21]. However, since the target compounds are weak bases, the extraction efficiency was studied in a wide range of acidic to basic medium (pH values from 3.0 to 11.0). As can be seen in Fig. 2, at pH values lower than 5.0; especially under acidic conditions (pH 3.0), the signals for all BZDs significantly decreased. Extraction of the neutral analyte into the extraction phase is eligible because of conventional interactions. Furthermore, extraction of positively charged (protonated) species seems as an interesting phenomenon, which occurred in the acidic medium for BZDs. This can be a result of ion pair formation between decanoate and protonated analytes. The peak area increases when the pH value of sample solution is increased from 5.0 to 9.0, but then decreases by further increase of pH. It was thus deduced that pH 9.0 is the most suitable value for the extraction. Addition of salt to the sample may have different effects. It can reduce the amount of water available to dissolve analyte molecules due to formation of hydration spheres around the ionic salt molecules and can improve the extraction efficiency for the target analytes into the extraction phase. On the other hand, addition of salt increases the ionic strength of the sample and changes the physical properties of the Nernst diffusion film; it can reduce the rate of diffusion of the analyte into the extraction phase and reduces the extraction efficiency. In this work, effect of salt on the extraction efficiency of the target compounds was investigated by adding sodium chloride from 0 to 15% (w/v) to the sample. According to Fig. 3, addition of salt in this range has a negative effect on the extraction efficiency. This may be due to the increase in viscosity that in turn decreases the mass transfer of the analyte to the extraction solvent. Hence, further extractions were performed without adding NaCl to the solutions. The acceptor phase volume is another important factor affecting the preconcentration factor of the analytes. By increasing the acceptor phase volume, because of dilution effect, the preconcentration factor is reduced. On the other hand, the acceptor phase volume should be large enough to promote analyte transport to the acceptor phase. Generally, an increase in the acceptor volume enhances
Fig. 3. Effect of salt addition on the peak area of BZDs. Extraction conditions: as in the Fig. 2; except that pH of the aqueous solution is 9.0.
the preconcentration factor as much as it does not lead to dilution of the extractant. The acceptor phase volume is proportional to the length of the hollow fiber. Therefore, the influence of the acceptor phase volume was studied by varying of hollow fiber length in the range of 6–11 cm. As demonstrated in Fig. 4, the relative peak area increases up to 10 cm, because the rate of the analyte transport into a solvent is directly related to the interfacial area between the two liquid phases. The results demonstrated that an increase in the HF length from 10 to 11 cm will result in a decrease in the peak areas of the analytes because of the dilution effect. Therefore, HF length of 10 cm was selected in the subsequent experiments. As is known, stirring speed plays an essential role in increasing the kinetics and efficiency of extraction by increasing the mass transfer and reducing the thickness of double layer around SLM. In HF-LPME, the extraction solvent is protected by the hydrophobic hollow fiber; so, high stirring speed can be applied compared with single-drop LPME. On the other hand, this parameter has to be carefully considered because a high stirring speed may cause the fall of the fiber in the sample and a whirlpool appears in the vial, which reduces the effective extraction surface between the hollow fiber and the aqueous sample. In addition, very high stirring rate may lose solvent that could affect the precision. To attain optimal stirring speeds, different stirring speeds ranging from 400 to 1000 rpm were examined. The results showed that the stirring speed of 900 rpm yielded the highest peak area and it was selected for subsequent experiments. In HF-LPME, mass transfer is a time-dependent process; thus, one of the most important parameters affecting the extraction efficiency is the extraction time. As SPME, HF-LPME is not an exhaustive extraction. It can take from several minutes to hours to reach an equilibrium that ensures an optimal extraction recovery
Fig. 4. Effect of hollow fiber length on the peak area of the drugs. Extraction conditions: volume of the aqueous solution, 24.0 mL; pH of the aqueous solution, 9.0; stirring rate, 800 rpm; extraction time, 40 min; concentration of the drugs, 50 g L−1 .
F. Rezaei et al. / Analytica Chimica Acta 804 (2013) 135–142
139
Table 1 Figures of merit of HF-VMMEa method for analyzing the selected BZDs in water, urine, and plasma samples. Sample
Analyte
R2
LOD (g L−1 )b
Deionized water
Alprazolam Nitrazepam Clonazepam Midazolam Diazepam
0.9956 0.9985 0.9993 0.9954 0.9966
0.7 0.7 0.7 0.7 0.5
Urine
Alprazolam Nitrazepam Clonazepam Midazolam Diazepam
0.9976 0.9907 0.9916 0.9945 0.9989
Plasma
Alprazolam Nitrazepam Clonazepam Midazolam Diazepam
0.9944 0.9937 0.9967 0.9984 0.9946
a b c d e
LDR (g L−1 )c
RSD (%)d
PFe
2.0–200.0 2.0–200.0 2.0–200.0 2.0–200.0 2.0–200.0
4.3 5.8 5.2 6.1 5.7
133 112 141 169 198
2.0 2.0 2.0 2.0 2.0
5.0–200.0 5.0–200.0 5.0–200.0 5.0–200.0 5.0–200.0
7.6 5.2 7.3 8.9 6.7
61 45 69 75 90
3.0 3.0 3.0 3.0 3.0
7.0–200.0 7.0–200.0 7.0–200.0 7.0–200.0 7.0–200.0
6.2 4.8 5.9 7.0 5.3
53 38 57 66 81
Hollow fiber vesicular-mediated microextraction. Limit of detection. Linear dynamic range. Relative standard deviation. Preconcentration factor.
and preconcentration factor. If time is chosen too short, the efficiency of extraction equilibrium may not reach. On the contrary, if the extraction time is chosen too long, solvent loss and formation of air bubbles may occur, which would decrease the extraction efficiency. Since SUPRAS have higher viscosity than conventional organic solvents, the diffusion of target species through the membrane and so the extraction rate will slightly be slower. The effect of extraction times in the range of 20–60 min on the extraction efficiency was investigated. The amounts of extracted analytes were increased dramatically by increasing the exposure time from 20 to 50 min and thereafter the curves became flat. On the basis of the results, 50 min was selected as the extraction time. 3.2. Quantitative analysis To evaluate practical applicability of the proposed HF-VMME technique; linearity, relative standard deviations (RSDs), limits of detection (LODs), and preconcentration factors (PFs) were investigated by extraction of BZDs from water samples under the optimal conditions (Table 1). In the first experiment, three-replicate extractions and determinations of BZDs (10 g L−1 of each analyte in the
water samples) were performed by HF-VMME/HPLC-UV and the RSDs% in the range of 4.3–6.1% were obtained for extraction of the analytes from water samples. In a subsequent experiment, linear ranges of 1.0–200.0 g L−1 for diazepam and 2.0–200.0 g L−1 for other analytes with coefficient of determinations (R2 ) of 0.9956–0.9993 for BZDs were obtained in the water samples. Based on the signal to noise ratio of 3 (S/N = 3), LODs of BZDs varied in the range as 0.5–0.7 g L−1 . For comparison, some figures of merit of previously reported methods such as LODs, LDRs, and RSD% for extraction and determination of BZDs are summarized in Table 2. As can be seen, the proposed HF-VMME method has a good sensitivity and proper precision with a suitable dynamic linear range. Also, the LODs obtained for the drugs by the present method are better or comparable with those obtained by other methods with similar detection instruments. 3.3. Analysis of real samples In order to evaluate applicability of the developed extraction method for analysis of BZDs in real samples with complex matrices,
Fig. 5. HPLC-DAD chromatograms of fruit juice (a) non-spiked and (b) spiked by 10 g L−1 of the target analytes, after HF-VMME; (1) alprazolam, (2) nitrazepam, (3) clonazepam, (4) midazolam, and (5) diazepam.
140
F. Rezaei et al. / Analytica Chimica Acta 804 (2013) 135–142
Table 2 Comparison of proposed method with other methods for analyzing the selected BZDs. Method
Analyte
LOD (g L−1 )
LDR (g L−1 )
RSD (%)
Ref.
LLE-HPLC
Diazepam Clonazepam
30 50
50–1200 75–1200
12.3 9.7
[15]
SPE-LC-MS
Alprazolam Nitrazepam Clonazepam Midazolam Diazepam
0.4 0.2 0.3 1.0 4.0
1–500 1–1000 1–1000 5–2000 13–1000
6 6 6 8 4
a
Alprazolam Clonazepam Diazepam
3.5 8.1 2.8
8.1–5000 18.8–5000 6.4–5000
5.6 12.7 3.3
[12]
SPE-HPLC-UV
Alprazolam Clonazepam Diazepam
1.0 2.0 2.5
5.0–200.0 5.0–200.0 7.5–500.0
7.1 7.1 7.1
[32]
SPME-LC-ESI-MS
Clonazepam Diazepam
2 1
2–500 1–500
4.0 4.0
[33]
SPME-LC
Nitrazepam Diazepam
4 1
5–1000 5–400
9.5 13.0
[34]
SPE-LC-ESI-MS/MS
Alprazolam Clonazepam Midazolam
0.02 0.02 0.02
0.05–50 0.05–50 0.05–50
15 15 15
[35]
b
Alprazolam Diazepam
7.4 6.2
– –
2.4 2.1
[36]
SPE-HPLC-DAD
Alprazolam Diazepam
8.8 10
– –
3.5 5.5
[36]
HF-VMME-HPLC-DAD
Alprazolam Nitrazepam Clonazepam Midazolam Diazepam
2.0–200.0 2.0–200.0 2.0–200.0 2.0–200.0 1.0–200.0
4.3 5.8 5.2 6.1 5.7
Proposed method
a b
UA-DLLME-HPLC
MAE-HPLC-DAD
0.7 0.7 0.7 0.7 0.5
[8]
Ultrasound-assisted dispersive liquid–liquid microextraction. Microwave assisted extraction.
tap water, fruit juice, urine and plasma samples were selected and the drugs were extracted using the proposed method under the optimal conditions. Sample preparation for real samples was performed according to Section 2.3. One of the criminal applications of BZDs is the addition of these drugs into fruit juice to make the consumer sleepy for robbery. Extraction of BZDs in fruit juice which was performed in this paper recommended by Research Center of Antinarcotics Police of Iran (Tehran) to develop a simple and easy method for analysis of these drugs in fruit juice. Different samples were analyzed by HF-VMME followed with HPLC-DAD. The results showed that the samples were free from contamination of the drug. Thus, all of the real aqueous samples were spiked with the BZDs standard and then they were prepared as it was discussed in Section 2.3 to assess the matrix effects. The results of analysis are shown in Table 3. Accuracy was calculated as the relative recoveries for analysis of known amounts of target analytes added to actual water samples using the proposed method. Relative recovery (RR%) was acquired from the following equation: RR% =
Cfound − Creal × 100 Cadded
10 g L−1 -spiked fruit juice sample. Fig. 6 shows chromatograms of the non-spiked and 15 g L−1 -spiked urine sample and Fig. 7 shows chromatograms of non-spiked and spiked plasma sample at the concentration level of 15 g L−1 BZDs standards. As can be seen in Figs. 6 and 7, this method is a good clean-up procedure that can be used for trace analysis of BZDs in complicated matrices such as urine and plasma samples.
(1)
where Cfound , Creal , and Cadded are concentration of analyte after addition of a known amount of standard into the real sample, concentration of analyte in real sample, and concentration of a known amount of standard, which was spiked into the real sample, respectively. The RR% of the BZDs were in the range of 90.0–98.8%. Intraand inter-assay precisions ranged between 4.4–8.2% and 7.6–12.1%, respectively. Fig. 5 shows chromatograms of the non-spiked and
Fig. 6. HPLC-DAD chromatograms of urine (a) non-spiked and (b) spiked by 15 g L−1 of the target analytes, after HF-VMME; (1) alprazolam, (2) nitrazepam, (3) clonazepam, (4) midazolam, and (5) diazepam.
F. Rezaei et al. / Analytica Chimica Acta 804 (2013) 135–142
141
Table 3 Results obtained for analysis of BZDs in real samples. Samplea
Alprazolam
Nitrazepam
Clonazepam
Midazolam
Diazepam
Tap water
Foundb (addedc ) RR% RSD (n = 3) Found (added) RR% RSD (n = 3)
9.64 (10.00) 96.4 6.3 24.2 (25.0) 96.8 5.4
9.88 (10.00) 98.8 5.2 23.5 (25.0) 94.0 4.8
9.54 (10.00) 95.4 5.4 24.5 (25.0) 98.0 5.9
9.16 (10.00) 91.6 6.6 23.8 (25.0) 95.2 6.1
9.40 (10.00) 94.0 5.8 23.2 (25.0) 92.8 4.4
Orange juice
Found (added) RR% RSD (n = 3) Found (added) RR% RSD (n = 3)
9.16 (10.00) 91.6 8.2 22.8 (25.0) 91.2 7.1
9.36 (10.00) 93.6 5.6 23.1 (25.0) 92.4 6.4
9.27 (10.00) 92.7 7.4 23.3 (25.0) 93.2 6.8
9.44 (10.00) 94.4 5.6 24.1 (25.0) 96.4 7.5
9.37 (10.00) 93.7 6.3 22.7 (25.0) 90.8 5.9
Sour cherry juice
Found (added) RR% RSD (n = 3) Found (added) RR% RSD (n = 3)
9.24 (10.00) 92.4 5.7 23.6 (25.0) 94.4 6.1
9.56 (10.00) 95.6 6.2 22.7 (25.0) 90.8 5.6
9.19 (10.00) 91.9 6.4 23.9 (25.0) 95.6 6.1
9.08 (10.00) 90.8 7.3 23.5 (25.0) 94.0 6.0
9.52 (10.00) 95.2 5.5 24.6 (25.0) 98.4 5.3
Delster
Found (added) RR% RSD (n = 3) Found (added) RR% RSD (n = 3)
9.35 (10.00) 93.5 5.9 23.9 (25.0) 95.6 6.0
9.76 (10.00) 97.6 6.1 24.1 (25.0) 96.4 6.3
9.49 (10.00) 94.9 7.2 23.6 (25.0) 94.4 6.5
9.43 (10.00) 94.3 6.5 24.1 (25.0) 96.4 5.9
9.58 (10.00) 95.8 6.8 23.5 (25.0) 94.0 6.4
Urine
Found (added) RR% RSD (n = 3) Found (added) RR% RSD (n = 3)
9.02 (10.00) 90.2 5.3 23.2 (25.0) 92.8 6.4
9.06 (10.00) 90.6 5.6 23.6 (25.0) 94.4 7.2
9.17 (10.00) 91.7 7.1 22.9 (25.0) 91.6 6.9
9.34 (10.00) 93.4 5.6 23.1 (25.0) 92.4 5.7
9.22 (10.00) 92.2 6.5 23.4 (25.0) 93.6 6.1
Plasma
Found (added) RR% RSD (n = 3) Found (added) RR% RSD (n = 3)
14.5 (15.0) 96.7 6.7 27.2 (30.0) 90.7 6.5
13.8 (15.0) 92.0 7.9 28.9 (30.0) 96.3 4.8
13.6 (15.0) 90.7 7.4 27.6 (30.0) 92.0 7.4
14.4 (15.0) 96.0 5.3 29.4 (30.0) 98.0 6.9
13.5 (15.0) 90.0 6.1 28.3(30.0) 94.3 5.8
Fig. 7. HPLC-DAD chromatograms of plasma (a) non-spiked and (b) spiked by 15 g L−1 of the target analytes, after HF-VMME; (1) alprazolam, (2) nitrazepam, (3) clonazepam, (4) midazolam, and (5) diazepam.
4. Conclusions The supramolecular solvent based on DeA–Bu4 NDe vesicles constitutes a suitable alternative to organic molecular ones to extract BZDs from water samples. This solvent is produced from environmental friendly and available reagents in a simple way and
the extraction process is simple, rapid, and cheap. SUPRASs consist of amphiphilic nanostructures that provide multiple binding sites and regions of different polarity. These outstanding properties make them suitable for extraction of a variety of analytes with high efficiency and render them ideal for microextractions. The high concentration of DeA in the extractant phase (∼1 mg L−1 )
142
F. Rezaei et al. / Analytica Chimica Acta 804 (2013) 135–142
and capability of analyte solubilization of the vesicles permit favorable partition of analytes using a quite low volume of SUPRAS (30 L). Extractions were based on hydrophobic, hydrogen bond, and -cation interactions between the analytes and the vesicular aggregates. So, coacervates can be considered as multifunctional solvents that have the capability to efficiently extract a wide variety of compounds. The vesicular coacervate phase possesses relatively high viscosities, low vapor pressure, and high cohesion between the aggregates. These properties make its use very attractive in obtaining stable SLM. In this work, BZDs were extracted from water, urine, and juice samples into the supramolecular solvent impregnated in the pores and also filled inside the porous polypropylene hollow fiber membrane. Moreover, by using SUPRAS as the extraction phase, the need for using toxic organic solvents was eliminated. References [1] M. Nakamura, Biomed. Chromatogr. 25 (2011) 1283–1307. [2] E.N. Sauve, M. Langødegård, D. Ekeberg, Å.M.L. Øiestad, J. Chromatogr. B 883–884 (2012) 177–188. [3] H.H. Maurer, J. Chromatogr. 580 (1992) 3–41. [4] S. Bourcier, Y. Hoppilliard, Eur. J. Mass Spectrom. 7 (2001) 359–371. [5] L. Wang, H. Zhao, Y. Qiu, Z. Zhou, J. Chromatogr. A 1136 (2006) 99–105. [6] R. Cordero, S. Paterson, J. Chromatogr. B 850 (2007) 423–431. [7] K. Arnhard, R. Schmid, U. Kobold, R. Thiele, Anal. Bioanal. Chem. 403 (2012) 755–768. [8] B.E. Smink, J.E. Brandsma, A. Dijkhuizen, K.J. Lusthof, J.J. de Gier, A.C.G. Egberts, D.R.A. Uges, J. Chromatogr. B 811 (2004) 13–20. [9] A. El Mahjoub, C. Staub, J. Pharm. Biomed. Anal. 23 (2000) 1057–1063. [10] A. El Mahjoub, C. Staub, J. Pharm. Biomed. Anal. 23 (2000) 447–458. [11] K.E. Rasmussen, S. Pedersen-Bjergaard, M. Krogh, H.G. Ugland, T. Grønhaug, J. Chromatogr. A 873 (2000) 3–11.
[12] P. Fernández, C. Gonzálezb, M.T. Pena, A.M. Carro, R.A. Lorenzo, Anal. Chim. Acta 767 (2013) 88–96. [13] H.L. Lord, M. Rajabi, S. Safari, J. Pawliszyn, J. Pharm. Biomed. Anal. 44 (2007) 506–519. [14] W.F. Smyth, S. McClean, Electrophoresis 19 (1998) 2870–2882. [15] K.B. Borges, E.F. Freire, I. Martins, M.E.P.B. Siqueira, Talanta 78 (2009) 233–241. [16] R.A. Anderson, M.M. Ariffin, P.A.G. Cormack, E.I. Miller, Forensic Sci. Int. 174 (2008) 40–46. [17] V.F. Samanidou, A.P. Pechlivanidou, I.N. Papadoyannis, J. Sep. Sci. 30 (2007) 679–687. [18] H.G. Ugland, M. Krogh, K.E. Rasmussen, J. Chromatogr. B 749 (2000) 85–92. [19] H.G. Ugland, M. Krogh, L. Reubsaet, J. Chromatogr. B 798 (2003) 127–135. [20] F. Rezaei, Y. Yamini, M. Moradi, B. Ebrahimpour, Talanta 105 (2013) 173–178. [21] F.J. Ruiz, S. Rubio, D. Pérez-Bendito, Anal. Chem. 78 (2006) 7229–7239. [22] F.J. López-Jiménez, S. Rubio, D. Pérez-Bendito, J. Chromatogr. A 1195 (2008) 25–33. [23] M. Moradi, Y. Yamini, J. Chromatogr. A 1229 (2012) 30–37. [24] S. Jafarvand, F. Shemirani, J. Sep. Sci. 34 (2011) 455–461. [25] M. Moradi, Y. Yamini, F. Rezaei, E. Tahmasebi, A. Esrafili, Analyst 137 (2012) 3549–3557. [26] S. Pedersen-Bjergaard, K.E. Rasmussen, Anal. Chem. 71 (1999) 2650–2656. [27] S. Pedersen-Bjergaard, K.E. Rasmussen, J. Chromatogr. A 1184 (2008) 132– 142. [28] R. Fortunato, C.A.M. Afonso, M.A.M. Reis, J.G. Crespo, J. Membr. Sci. 242 (2004) 197–209. [29] F.J. Ruiz, S. Rubio, D. Pérez-Bendito, J. Chromatogr. A 1163 (2007) 269–276. [30] A. Moral, M.D. Sicilia, S. Rubio, Anal. Chim. Acta 650 (2009) 207–213. [31] A. Moral, M.D. Sicilia, S. Rubio, J. Chromatogr. A 1216 (2009) 3740–3745. [32] L. Mercolini, R. Mandrioli, M. Amore, M.A. Raggi, J. Sep. Sci. 31 (2008) 2619–2626. [33] H. Yuan, Z. Mester, H. Lord, J. Pawliszyn, J. Anal. Toxicol. 24 (2000) 718–725. [34] K. Jinno, M. Taniguchi, M. Hayashida, J. Pharm. Biomed. Anal. 17 (1998) 1081–1091. [35] O. Quintela, F.L. Sauvage, F. Charvier, J.M. Gaulier, G. Lachâtre, P. Marquet, Clin. Chem. 52 (2006) 1346–1355. [36] P. Fernández, C. Vázquez, R.A. Lorenzo, A.M. Carro, I. Álvarez, P. Cabarcos, Anal. Bioanal. Chem. 397 (2010) 677–685.
