Covalently bonded double-charged ionic liquid on magnetic graphene oxide as a novel, efficient, magnetically separable and reusable sorbent for extraction of heavy metals from medicine capsules

Abstract

A new, efficient and reusable double-charged ionic liquid modified magnetic graphene oxide (DIL-MGO) was used as a special sorbent in the solid-phase extraction process for simultaneous separation and preconcentration of Pb(II), Cd(II), Ni(II), Cu(II) and Cr(III). The ionic liquid was chemically attached to magnetic graphene oxide nano sheets and then characterized by field emission scanning electron microscopy (FE-SEM), thermo gravimetric analysis (TGA), X-ray diffraction (XRD), vibrating sample magnetometry (VSM) and Fourier transform infra-red spectroscopy (FT-IR). This sorbent employs a hybrid combination of an ionic liquid and graphene oxide extraction properties, which contributed to the increased extraction efficiency via improving mass transfer yield. An ultrasound-assisted double charged ionic liquid-linked magnetic micro solid phase extraction method (US-assisted DIL-MμSPE) coupled with micro sampling atomic absorption spectrometry was utilized for the determination of heavy metals. Effecting parameters, including sample pH, amount of DIL-MGO, extraction time, type and concentration of desorption solvent, were investigated and optimized. Good recoveries and satisfactory relative standard deviations (RSDs) under the optimal conditions were achieved. The limits of detection were from 0.2 to 1.8 μg L⁻¹. Method performance was investigated by the determination of the mentioned heavy metals in complex medicine capsule matrixes with good recoveries. Finally, the sorbent was reused for eight runs without significant loss of its recovery.

---

1. Introduction

The control of drug impurities in the pharmaceutical industry is an important issue. The importance of this matter is to ensure that the quality of drugs and the correct amount of medication is being prescribed to the patient; therefore, drug impurities must be carefully monitored. There are three types of impurities, organic, inorganic and residual solvents, in drugs and medicines. Metals are one of inorganic impurities that must be controlled in medicines and their ingredients.¹ The main sources of heavy metals are catalysts for chemical reactions, water used in the processes, and reactors (if stainless steel reactors are used). An important, but often overlooked, component of the pharmaceutical industry is associated with medicine capsules. The major metals with potential health concern found in medicines and their ingredients are lead, cadmium, nickel, copper and chromium.² These heavy metals are of great concern even at trace levels because they are not metabolized and may cause serious hurt and even cancer. Therefore, selective analytical methods need to be developed to monitor heavy metals. Despite major developments in analysis of heavy metals, their determination at trace amounts is often difficult with additional problems due to trace amount of analytes and matrix interferences. New separation and preconcentration techniques as prerequisite steps for the elimination of any matrix components, consequently improving the detection limits, are of significant importance.³ Therefore, for separation and preconcentration of heavy metals, various techniques have been proposed such as liquid–liquid extraction (LLE),⁴ single-drop microextraction (SDME),⁵ solid phase extraction (SPE),^6–9 cloud point extraction (CPE)¹⁰ and dispersive liquid–liquid microextraction (DLLME).¹¹

Nowadays, ionic liquids have found numerous applications in the extraction process; however, these IL-based techniques have some disadvantages such as lower rates of mass transfer and loss of ILs to the aqueous phase due to the relatively high viscosity of the ionic liquids, slow diffusion due to the bulk ionic liquid, as well as longer equilibrium times, difficulty in separate-phase, need of some special instruments (syringes, tubes, centrifuge and others), difficulty in collecting ionic liquids dispersed in a solution and inefficiency against complex matrixes. Therefore, from both economic and technical points of view, the productivity of reactions based on bulk ILs is hampered since a large part of the ILs are not contributing to the overall process.^12–14 Most of these inherent limitations may be improved by immobilizing the ionic liquids on a suitable solid base, such as silica, magnetic nanoparticles or nano carbons, which support and fixes the ILs on its surface.¹⁵ If the platform has magnetic properties, this eliminates the problem of collecting ionic liquids in solutions, and magnetic ionic liquid materials with captured analytes can be readily separated from the sample matrix by an external magnetic field.¹⁶ Immobilized ionic liquids offer several other advantages over homogeneous IL systems, such as enhanced mass transfer, easy handling, reduced IL amounts utilized in a typical process relative to homogeneous systems and recycle ability. Moreover, by applying such sorbents, unique and tunable properties of ionic liquids together with supporting characteristics of heterogeneous substrates would be employed at the same time. Thus, not only can most of the problems for the ILs-based microextraction techniques be ameliorated with this hybrid of ILs and solid sorbents, but also various advantages are found for catalysis, removal, separation, extraction processes¹⁷ and as a stationary phase in chromatographic applications.¹⁸

Recently, carbon-based materials, such as carbon nanotubes, graphene, and fullerene, are widely used as a sorbent in solid phase extractions and new modes of SPE. Graphene sheets, single or multi layers of sp² hybridized carbon atoms arranged in a honeycomb lattice, have unique structures, ultrahigh specific surface areas, good thermal conductivities, and fast mobility of charge carriers. Graphene performance as a sorbent in SPE is better than that of carbon nanotubes. This is due to the special morphology of graphene in which both sides of its planar sheets are available for molecule adsorption and leads to a faster adsorption equilibrium and analyte elution.^19–21

To the best of our knowledge, this is the first report of using chemically bonded double charged ionic liquids on magnetic graphene oxide for simultaneous solid phase extraction and preconcentration of heavy metal ions. Herein, it is believed that integrating MGO and IL provides new hybrid materials and can open a novel path towards the expansion of solid phase microextraction techniques.

2. Experimental

2.1. Apparatus

Pb(II), Cd(II), Ni(II), Cu(II) and Cr(III) were measured using an Agilent 200 Series AA (model 240 AA) flame atomic absorption spectrometer (USA) including an air–acetylene flame. The conditions for flame atomic absorption spectrometry for the mentioned heavy metal ions are summarized in Table 1. The morphology of (DIL-MGO) was characterized using a field emission scanning electron microscope (FESEM). FESEM was performed by a MIRA3 TESCAN field emission scanning electron microscope. Fourier transform infra-red (FTIR) spectra were obtained using an S, 8400 Shimadzu at a resolution of 1 cm⁻¹. Pressed potassium bromide (KBr) pellets at a sample/KBr weight ratio of 1 : 100 were scanned and recorded between 3400 and 400 cm⁻¹. XRD patterns of samples were obtained with an INEL EQuinox 3000 X-ray diffractometer (INEL. France) using Cu-Kα radiation (λ = 1.541874 Å), (40 kV and 45 mA conditions). Thermo-gravimetric analyses (TGA) were conducted with a LINSEIS model STS PT 16000 thermal analyzer at a heating rate of 5 °C min⁻¹.

Table 1 The FAAS conditions for determination of Pb(II), Cd(II), Ni(II), Cu(II) and Cr(III)

Parameters	FAAS conditions
	Pb(II)	Cd(II)	Ni(II)	Cr(III)	Cu(II)
Wavelength (nm)	217.0	228.8	232.0	357.9	324.7
Slit (nm)	1.0	0.5	0.2	0.2	0.5
Lamp current (mA)	10.0	10.0	10.0	10.0	10.0
Working range (ppm)	0.1–30	0.02–3	0.1–20	0.06–15	0.03–10
Volume injection (μL)	50	50	50	50	50
Mode	Peak area	Peak area	Peak area	Peak area	Peak area
Gas	Air–acetylene	Air–acetylene	Air–acetylene	Air–acetylene	Air–acetylene

---

2.2. Chemical reagents and materials

The chemical compounds and reagents used in this study were of analytical grade and purchased from Merck (Darmstadt, Germany) unless otherwise stated. The stock standard solutions of Pb(II), Cd(II), Ni(II), Co(II) and Cr(III) (1000 mg L⁻¹) were prepared from proper amounts of their nitrate salts in ultra-pure water and working standard solutions were prepared by right stepwise dilutions of the stock standard solutions. High purity sodium hydroxide (BioUltra grade, ≥98%, Sigma-Aldrich), ethanol (absolute, Merck Millipore, Darmstadt, Germany), nitric acid (Trace SELECTR Ultra, 65–71%, Sigma-Aldrich), and hydrochloric acid (Trace SELECTR Ultra, 30–35%, Sigma-Aldrich) were used. For the synthesis of the sorbent, 1,4-diazabicyclo[2.2.2]octane (Dabco) (furnished by Aldrich, Germany, purity grade 98%); 3-chloropropyltrimethoxysilane (CPTMS) (Aldrich, Germany, purity grade 97%); iron(II) chloride tetra hydrate (99%); and iron(III) chloride hexa hydrate (98%) were used. To decrease the risk of contamination, the laboratory glassware and plastics (polypropylene) used in this study were cleaned by soaking them in 10% (v/v) nitric acid for at least 24 hours and then rinsed with deionized water and dried in a clean oven prior to use.

2.3. Synthesis of double-charged ionic liquid magnetic graphene oxide

For preparation of DIL-MGO, graphene oxide (GO) was synthesized from natural graphite powders by a modified Hummers' method.²² First, 120 mL of H₂SO₄ was added into a 500 mL flask and then cooled followed by stirring. 5.0 g of graphite powder and 2.5 g of NaNO₃ were subsequently added under vigorous stirring. Then, 15 g of KMnO₄ was added gradually under stirring and the temperature of the mixture was maintained below 10 °C. The mixture was stirred at room temperature. The diluted suspension was stirred at 98 °C, followed by the addition of 50 mL of 30% H₂O₂. Finally, the mixture was filtered and washed with a 5% HCl solution, followed by water washing until the pH = 7.0. GO was obtained as a grey powder after filtration and dried at 65 °C under vacuum.

Magnetic graphene oxide (MGO) was prepared by the chemical coprecipitation of FeCl₂ and FeCl₃ in an alkaline solution in the presence of GO. Briefly, it was synthesized by suspending 50.0 mg GO in 50.0 mL of a solution containing 50.0 mg FeCl₃ and 35.0 mg FeCl₂ under an N₂ atmosphere. After the solution was sonicated for 30 min, a NH₄OH aqueous solution was added dropwise to precipitate the iron oxides while the mixture solution was under constant mechanical stirring at 90 °C. The pH of the final mixture should be 11.0 and the reaction was continued for about 60 min. The obtained MGO precipitate was isolated in the magnetic field, and the supernatant was separated from the precipitate by decantation. The obtained MGO composites were then washed with 150 mL of absolute alcohol and dried at 65 °C under vacuum.

Si-MGO was prepared by dispersing 100.0 mg of MGO into an anhydrous toluene solution containing 3 mL of CPTMS. The solution was refluxed for 48 h at 90 °C under the protection of N₂, and the resulting solid was washed with toluene several times. Si-MG was collected with an external magnetic field, and then washed by redistilled water and dried under vacuum.

For preparation of DABCO-Si-MGO, MGO-Si (100 mg) was added to a flask containing 20 mL of anhydrous toluene and an excess of DABCO (100 mg). After degassing and nitrogen purging for about 5 min, the solution was placed at 90 °C and reacted for 24 h. Then, the reaction mixture was cooled to room temperature, transferred to a vacuum glass filter, and washed with toluene and absolute alcohol. MGO chemically bonded with DABCO (MGO–DABCO) was dried at 50 °C for 8 h.²³ Preparation of the ionic liquid functionalized magnetic graphene oxide is shown in Fig. 1.

Fig. 1 Schematic of synthesis of DIL-MGO.

2.4. Ultrasound-assisted double charged ionic liquid-linked magnetic micro solid phase extraction procedure (US-assisted DIL-MμSPE)

50 mL of ultra-pure water containing 50 μg L⁻¹ of Pb(II), Ni(II), Cr(III), Cu(II) and 10 μg L⁻¹ of Cd(II) with 10 μL of a 10% (m/v) APDC solution (as a chelating agent) was placed in a beaker. The pH of the sample solution was adjusted to 6 by adding an appropriate amount of diluted NaOH and HNO₃. Subsequently, 40 mg of sorbent was added into the sample solution. The mixture was sonicated at room temperature for 4 min to form a homogeneous dispersion solution. During this stage, the heavy metal hydrophobic complexes were extracted into the hydrophobic IL layer on the surface of the magnetic graphene oxide. Then, the sorbent was held in place by placing a Nd-Fe-B magnet around the sample beaker and the upper phase was decantated completely. Then, preconcentrated target analytes were desorbed from the sorbent via 250 μL of 1 mol L⁻¹ nitric acid with the aid of ultrasound for 1 min. Finally, 50 μL of eluent was injected directly into the micro sampler atomic absorption spectrometer for each heavy metal ion analysis (Fig. 2).

Fig. 2 Schematic of the US-assisted DIL-MμSPE process.

2.5. Sample preparation

For sample preparation, laboratory glassware and vessels were cleaned by soaking in 10% (v/v) nitric acid for at least 24 h and then rinsed with deionized water prior to use. As heavy metal ions concentrations in medicine capsule must be very low, even minor contamination at any stage of sampling, sample storage and handling, or analysis has the potential to affect the accuracy of the results. Two types of empty medicine capsules were considered as real samples and 0.5 g from each type of them were treated with a mixture of concentrated HNO₃ and H₂O₂ (1 : 3) and the mixture was sonicated for 5 min. After completing the dissolution process, the sample solution was filtered through a cellulose filter paper. The filter paper was washed with 2 mL of 0.1 mol L⁻¹ HNO₃. Finally, the solution was diluted to 50.0 mL with distilled–deionized water by adjusting the pH to 6. These sample solutions were accurately analysed using a US-assisted DIL-MμSPE method, and the heavy metals in the final solution were determined by AAS or ICP OES.²⁴

3. Result and discussion

3.1. Characterization of adsorbent

The morphological characterization of DIL-MGO was performed by field emission scanning electron microscopy (FE-SEM). As shown in Fig. 3, the graphene oxide sheets had wrinkled surfaces and the Fe₃O₄ nanoparticles were well dispersed on the surface of GO to form magnetic graphene oxide. The porous structure of DIL-MGO with holes in the surface of the sheets can be clearly observed. DIL-MGO is composed of large amounts of sheets, and the SEM image of the synthesized DIL-MGO clearly confirms the porosity of sorbent. The existence of cavities on the surface of the sorbent increases the surface area of it and therefore increases its ability as a sorbent.^25,26

Fig. 3 Porous structure of DIL-MGO in scanning electron microscopy image at different magnifications (a) 10 μm, (b) 1 μm and (c) 200 nm.

Magnetic properties of DIL-MGO were studied using a vibrating sample magnetometer (VSM). As shown in Fig. 4a, the room-temperature magnetization curves showed magnetic hysteresis curves for DIL-MGO. Obviously, DIL-MGO exhibits negligible coercivity (H_c) and remanence, reaching a saturation magnetization value of 21.73 emu g⁻¹ at room temperature. This feature would be favorable for their use as adsorbents and it is important for the convenient recycling of DIL-MGO.²⁷

Fig. 4 (a) VSM curve of DIL-MGO. (b) TGA curve of DIL-MGO. (c) X-ray diffraction pattern of the GO, Fe₃O₄ and DIL-MGO. (d) FTIR spectra of the GO, Fe₃O₄ and DIL-MGO.

Thermal stability is an important factor that affects the application of an adsorbent. Thermo gravimetric analyses on the GO and DIL-MGO in the temperature range from 30 to 600 °C are shown in Fig. 4b. GO shows a weight loss at 100 °C, primarily due to the loss of residual water in the GO sheet layers. A sharper weight loss near 200 °C is also observed, due to the pyrolysis of oxygen-containing groups from the surface of the GO sheets. Finally, the loss that occurred at 500 °C was mainly due to the combustion of the carbon skeleton. In DIL-MGO, weight loss below 300 °C is attributed to loss of water and DIL combustion. The initial loss of weight, which occurred below 100 °C, is attributed completely to the loss of physically adsorbed water molecules. The second weight loss at about 260 °C is associated with the thermal decomposition of organic moieties at the surface and the third one is associated with decomposition of the residual methoxy side groups. It is interesting to observe that the weight loss is smaller for DIL-MGO than GO at high temperatures (over 500 °C), which may be attributed to the slight reduction of GO by DIL.²⁸

The phase structure of GO, Fe₃O₄ and DIL-MGO was characterized by XRD measurements, as shown in Fig. 4c. The peak observed at a diffraction angle (2θ) of 9.98° is highly specific to the crystalline nature of GO nanosheets in the GO pattern, and the XRD pattern of a standard Fe₃O₄ crystal with the spinel structure has characteristic peaks at 2θ = 30.2°, 35.8°, 43.4°, 57.5°, and 63.1°. These peaks remain after modifications in DIL-MGO. Therefore, it can be concluded that the obtained DIL-MGO shows a spinel structure and these modifications do not cause a phase change in Fe₃O₄.²⁹ Noticeably, the GO peak disappeared in the XRD pattern of DIL-MGO. This phenomenon was also observed by some researchers.^30,31 They considered that the disappearance of the GO peak in the XRD pattern of DIL-MGO may be derived from the following reasons: (1) more monolayers of graphene caused due to the reduction of graphene sheet aggregation in the presence of magnetite, resulting in weaker peaks from carbon being observed; (2) the strong signals of the iron oxides overwhelming the weak carbon peaks.³¹

The FT-IR spectrum of GO, Fe₃O₄ and DIL-MGO in the 400–3400 cm⁻¹ is shown in Fig. 4d. The FT-IR analysis of the DIL-MGO exhibits a basic characteristic peak at approximately 590 cm⁻¹, which was attributed to the presence of Fe–O stretching vibrations. The presence of bands around 702, 831 and 1007 cm⁻¹ was most probably due to the formation of a condensed silica network and symmetric and asymmetric stretching vibrations of the framework and terminal Si–O groups. The absorbance peak of C–N with a wavelength of about 3000 cm⁻¹ indicated that the surface of MGO is successfully modified by double-charged diazoniabicyclo[2.2.2] octane chloride. A band at 1400 cm⁻¹ is characteristic of the tertiary amine group. The bands around 1635 and 3400 cm⁻¹ are also due to the bending vibration of water molecules, which were adsorbed on the surface.³²

3.2. Application of DIL-MGO in (US-assisted DIL-MμSPE) method and optimization effective parameter

3.2.1. Effect of pH. Both the major factors affecting the extraction of metal ions, namely, complex formation and chemical stability of the hydrophobic complex, are influenced by the pH of the aqueous phase as an exclusive parameter. The pH value of the aqueous solution should be appropriate for hydrophobic complex formation to be sufficiently hydrophobic to facilitate mass transfer into the thin ionic liquid layer on the surface of DIL-MGO. The effect of pH on the extraction of heavy metals was studied in the range of 1–8 by adjusting with diluted HNO₃/NaOH solutions. As the results show in Fig. 5a, extraction efficiency at a pH of 6 is maximized for analytes. This may be due to the higher stability of the metal ions–APDC complex at this pH. Therefore, the optimum pH of 6 was selected for subsequent work because it is near the neutral ambience. At higher pH values, heavy metal ions precipitate with OH⁻ and their retention is changed. At lower pH values, the ligand is protonated, so recoveries are decreased due to competition between hydronium ions and desired metal ions.

Fig. 5 Optimization US-assisted DIL-MμSPE parameters (a) effect of pH. (b) Effect of the amount of DIL-MGO. (c) Effect of the type of eluent. (d) Effect of concentration of eluent. (e) Effect of time. (f) Effect of concentration of APDC (% w/v), on the recovery of Cd(II), Cu(II), Ni(II), Pb(II) and Cr(III) (n = 3). Microextraction condition: 50 mL solution containing 50 μg L⁻¹ Cu(II), Ni(II), Pb(II) and Cr(III) and 10 μg L⁻¹ Cd(II); pH: 6.0; 40 mg DIL-MGO; 4 min ultrasonic bath; 250 μL of 1 mol L⁻¹ nitric acid.

3.2.2. Amount of DIL-MGO sorbent. An optimum amount of sorbent for maximum recovery was determined by varying its amount in the range of 5–50 mg. As shown in Fig. 5b, the extraction efficiencies of heavy metals firstly increased when the amount of sorbent was increased from 5 to 40 mg. However, the extraction efficiency was decreased when more than 40 mg of DIL-MGO was used. This is due to the fact that a small quantity of DIL-MGO could not be dispersed due to the hydrophobic surface of it or the attachment of DIL-MGO onto the magnet decreased and DIL-MGO was suspended in the solution due to the reduced magnetism of the magnet. This hindered the collection of all the DIL-MGO by an external magnet, which resulted in a slight decrease in the extraction efficiency. Based on the results, 40 mg of sorbent provided the highest extraction efficiency for all analytes.

3.2.3. Choice of the eluent: type, concentration and volume. A suitable desorption solvent is critical in desorption of heavy metals. Therefore, in order to choose the most effective eluent for the simultaneous and quantitative recovery of the mentioned metal ions, different acidic eluents were used and tested according to these facts; at low pH, the adsorption of the investigated metal ions was almost negligible, and most metal–APDC complexes are dissociable in such media. As shown in Fig. 5c, all the tested heavy metals achieved their highest recoveries when HNO₃ was used as an eluent. Thus, HNO₃ was selected for desorption. Then, an acidic solution of HNO₃ was prepared in the concentration range of 0.5–3.0 mol L⁻¹. As can be seen in Fig. 5d, for almost all heavy metals, by increasing the concentration of the acid concentration up to 1.0 mol L⁻¹, elution recoveries increased and then stayed approximately constant. Accordingly, a concentration of 1.0 mol L⁻¹ was selected as the optimum for the subsequent work. Moreover, for the estimation of the minimum required volume of this eluent, various volumes in the range from 150 μL to 1.0 mL of the selected eluent (1.0 mol L⁻¹) were applied. Results indicated that volumes larger than 250 μL are proper for the complete elution. Thus, for achieving the highest enrichment, a volume of 250 μL of this eluent was selected for the following experiments.

3.2.4. Adsorption and desorption time. The adsorption time is an important parameter and it is a major sign of the performance of a sorbent. Therefore, the relationship between contact time and metal ions sorption onto the sorbent was studied. The total extraction time relates to the time required for complex formation and then transfer of the hydrophobic complexes to the sorbent. Due to the relatively large surface area, a very fast mass transfer process occurs and generally offers a fast extraction process compared to other methods based on ionic liquids. The extraction recovery was studied at time intervals in the range of 1–7 min and the results are shown in Fig. 5e. The extraction efficiency of all the heavy metals reached a maximum value when the adsorption time was at 4 min and the extraction efficiency remained fixed after 4 min. The results indicated that an extraction equilibrium could be achieved in a very short time, which could be ascribed to the high surface area of DIL-MGO and the homogeneous distribution of DIL-MGO in the sample solution during extraction.³³ Overall, 4 min was chosen as the optimum adsorption time for extraction. Desorption time had a remarkable influence on the extraction efficiency of all the heavy metals; therefore, desorption time was studied in the range from 30 s to 4 min. When the desorption time was less than 1 min, the extracted analytes could not be fully desorbed from DIL-MGO. It was observed that desorption equilibrium was obtained for all target heavy metals after 1 min. Thus, desorption time of 1 min was applied for all subsequent experiments.

3.2.5. Effect of APDC (chelating agent) concentration. Ammonium Pyrrolidine Dithio Carbamate (APDC) was used as a ligand and can form very stable complexes with metal ions. Metal ion complexes with APDC are soluble in water and can be extracted into a sorbent. The effect of the APDC concentration on the extraction yields of heavy metals was studied in the concentration range of 1–20% w/v of APDC. The obtained results in Fig. 5f show that the extraction efficiencies of metals increase as APDC concentration increases from 1% to 10% w/v; furthermore, it remains approximately constant up to 30% w/v. Therefore, an amount of 10% w/v of APDC is sufficient and chosen for total complex formation and maximum extraction efficiency.

3.2.6. Salt effect. In order to investigate the effect of ionic strength on μSPE based on the ionic liquid modified magnetic graphene oxide sorbent, recoveries of five mentioned metal ions in the presence of NaCl (as one of the most common electrolytes present in environmental samples) were studied. To investigate the effect of the addition of salt in the present study, different concentrations of NaCl (0 to 30%, w/v) were applied. According to the results, no significant impact on the recovery was observed by the addition of NaCl. Therefore, the experiments were carried out without the addition of NaCl.

3.3. Stability and reusability of DIL-MGO sorbent

Finally, the stability, reusability and regeneration of the sorbent were evaluated by selecting solutions containing lead as a test analyte. It was used in consecutive adsorption/desorption cycles in accordance with the extraction procedure. The extraction efficiency of test objects decreased approximately 12% after 8 recycles. It indicated that even with a slight loss in the sorption capacity, DIL-MGO was still able to quantitatively extract metal ions for more than 8 cycles. After these uses, we observed a lost in the extraction efficiency. Probably, after these uses, the ionic liquids on the surface of the magnetic graphene oxide were altered and hence a decrease in the extracting capacity was observed. Moreover, under the influence of ultrasound and elution with an acidic solution, magnetic particles lose their properties and an external magnet cannot remove them completely, and therefore the analytes adsorbed on them are thrown away. Finally, the data showed that the developed sorbents were mechanically stable, possessed an excellent reusability and had a rather high durability.

3.4. Sorption capacity

In order to evaluate the maximum sorption capacity of the sorbent for each heavy metal ion, the adsorption of metal ions on MGO-IL was simulated using Langmuir and Freundlich isotherm models. The batch adsorption experiments were carried out with 60 mg of MGO-IL and 50 mL of heavy metals aqueous solutions with the desired concentration and appropriate pH. Then, the suspensions were sonicated for 10 min, and after decantation by applying an external magnetic field, the retained metal ions in the supernatant solution were determined using AAS. Herein, Langmuir and Freundlich isotherm models are applied to simulate and understand the adsorption mechanism. The Langmuir and Freundlich isotherm models are expressed below:

where C_e is the equilibrium concentration of metal ions in the solution phase (mg L⁻¹), q_e is the amount of metal ions adsorbed at equilibrium (mg g⁻¹), q_m is the maximum amount of metal ions adsorbed per unit weight of adsorbent, and k_l is the Langmuir constant (L mg⁻¹), related to the free energy of adsorption. K_f and n are Freundlich constants related to the adsorption capacity and adsorption intensity, respectively. The high values for the correlation coefficients indicate that the Langmuir equation fits the experimental data better than the Freundlich model. The maximum adsorption capacities (q_m) of Pb, Cu, Cr, Cd and Ni on MGO-IL at 298 K are 47.61, 33.55, 29.49, 24.93 and 18.11 mg g⁻¹, respectively.

3.5. Analytical performance of method

Under optimal conditions, analytical performance of this work was evaluated by investigating the linear range, correlation coefficient, limits of detection (LODs), limits of quantification (LOQs), repeatability, enrichment factor and consumptive index (CI). The results of this study are shown in Table 2. The linearity of Cd(II), Cu(II), Ni(II), Pb(II) and Cr(III) calibration plots were investigated over a concentration range of 1.1–15, 6.5–50, 1–100, 3.5–150 and 2.6–75 μg L⁻¹ with the regression equations of A = 0.04[Cd] + 0.0072, A = 0.01[Cu] + 0.0877, A = 0.0096[Ni] + 0.031, A = 0.0065[Pb] + 0.0287 and A = 0.01[Cr] + 0.0142, respectively. The correlation coefficients (R²) for each metal ion were greater than 0.997, indicating good linearity. The limits of detection (LODs) and limits of quantification (LOQs) were calculated as 3s/m and 10s/m, where s is the standard deviation of five consecutive blanks and m is slope of the calibration curve. To calculate precision, relative standard deviations (RSD%) were calculated from several individual standards. The results showed that the method is precise. The enrichment factors were measured as the ratio of sample volume in this study (50 mL) to final volume (250 μL). The consumptive index, CI, is defined as follows:
where V_s and EF are the sample volume (mL) and enrichment factor, respectively. A lower CIs mean that a higher enrichment factor from lower volumes of sample can be obtained. This is a key phrase to compare the different methods in which various volumes of sample are used.

Table 2 Analytical performance data of the US-assisted, DIL-MSPE method for Cd(II), Cu(II), Ni(II), Pb(II) and Cr(III) determination (50 mL sample volume, pH = 6, sorbent 40 mg of DIL-MGO, ultrasonication 4 min, elution volume 250 μL, n = 3)

Parameter	Pb(II)	Cd(II)	Ni(II)	Cr(III)	Cu(II)
Linear range (μg L^−1)	3.5–150	1.1–15	1–100	2.6–75	6.5–50
Correlation coefficient	0.998	0.998	0.997	0.999	0.997
Limits of detection (LODs) (μg L^−1)	0.923	0.2325	0.2	0.742	1.811
Limits of quantification (LOQs) (μg L^−1)	3.076	0.775	0.66	2.47	6.1
Relative standard deviation (RSD%, n = 5)	2.057	3.067	2.33	1.80	4.04
Enrichment factor	200	200	200	200	200
Regression equation (mg L^−1)	A = 0.006[Pb] + 0.0287	A = 0.04[Cd] + 0.0072	A = 0.009[Ni] + 0.031	A = 0.01[Cr] + 0.0142	A = 0.01[Cu] + 0.0877
CI	0.25	0.25	0.25	0.25	0.25

---

3.6. Effect of interference ions

Under optimal conditions, effect of interference ions on extraction of heavy metals was studied. The results of this study are shown in Table 3. To perform this experiments, different concentration of Na⁺, Li⁺, K⁺, F⁻, Cl⁻, NO₃⁻, CH₃COO⁻, Ag⁺, NH₄⁺, Mg²⁺, Ca²⁺, Ba²⁺, Zn²⁺, Fe³⁺, Al³⁺ and Co²⁺ was added to a heavy metal solution. The tolerance limit was defined as the highest amount of interfering ions causing a change of less than ±5% in the recovery of the heavy metals. Results showed that the presence of major coexisting ions had no significant influence on the simultaneous preconcentration of the mentioned heavy metals, and DIL-MGO is a selective sorbent for extraction of heavy metals in complicated matrices with interfering ions.

Table 3 Effect of potential interfering ions on the recovery of target ions

Ion	Concentration (mg L^−1)	Added as	Recovery (%)
			Pb(II)	Cd(II)	Ni(II)	Cr(III)	Co(II)
Na^+	1000	NaCl	97.9	98.5	95.2	100.8	96.4
Li^+	1000	LiF	98.2	96.6	98.9	98.9	98.8
K^+	1000	KCl	102.1	96.3	97.7	95.1	97.3
F^−	1000	NaF	97.4	95.9	100	99.1	95.4
Cl^−	1000	NaCl	98.3	97.3	98.5	96.4	98.3
NO3^−	1000	NaNO3	96.9	102.1	102.4	102.8	102.01
CH3COO^−	500	NaOAC	98.6	99.2	98.3	99.3	97.5
Ag^+	500	AgNO3	101.04	95.9	101.8	97.9	95.1
NH4^+	500	NH4NO3	95.9	96.2	95.2	96.7	97.8
Mg^2+	100	MgCl2	96.6	95.7	99.2	101.8	101.92
Ca^2+	100	CaCl2	97.1	96.5	98.2	98.1	96.3
Ba^2+	100	BaCl2·6H2O	98.4	96.2	101.4	97.8	98.4
Zn^2+	50	Zn(NO3)2·6H2O	96.3	98.86	97.1	96.02	97.2
Fe^3+	50	Fe(NO3)3·9H2O	101	102.47	98.5	103.32	98.7
Al^3+	50	Al(NO3)3·9H2O	98.9	98.53	102	97.65	95.8
Co^2+	10	Co(NO3)2·6H2O	95.23	97.2	95.45	95.8	96.07

---

3.7. Analysis of real samples

According to the United States Pharmacopeial Convention (USP), the presence of elemental impurities in drug substances and excipients must be controlled and where present, reported. The acceptable levels for these impurities depend on the material's ultimate use. Therefore, drug product manufacturers must determine the acceptable level of elemental impurities in the drug substances and excipients used to produce their products.³⁴

In order to validate and evaluate the accuracy of this method, it was applied in the extraction and micro-AAS determination of investigated toxin metal ions in two types of complex medicine capsule matrices. As it will demonstrate, US-assisted and DIL-MμSPE/AAS results are comparable to values obtained with ICP-OES and there is no significant differences between the US-assisted, DIL-MμSPE/AAS or ICP-OES results, at 95% level. Then, each type of sample was spiked with 10 μg L⁻¹ of the mentioned heavy metals according to its linear range to assess recoveries. The results of these experiments are shown in Table 4. The results demonstrated that there are good agreements between the recommended and obtained values of the elements and it can be concluded that the synthesized sorbent is suitable for determination of heavy metals in complicated real samples.

Table 4 Analytical results for accuracy investigation of the US-assisted, DIL-MSPE method in spiked medicine capsules samples

Sample	Heavy metals	Determined concentration by (US-assisted, DIL-MμSPE)/ICP-OES^a (μg L^−1)	Determined concentration by (US-assisted, DIL-MμSPE)/AAS (μg L^−1)	Determined concentration after spike^c by US-assisted, DIL-MSPE^d (μg mL^−1)	Recovery (%)
a Not significantly different from the ICP-OES and AAS value, at the 95% level.b Below detection limit.c Sample was spiked with 10.0 μg L^−1 of each element.d Mean value ± standard deviation based on three replicate measurements.
Medicine capsule type 1	Cadmium	1.1 ± 0.1	1.4 ± 0.1	11.22 ± 0.04	98.4
	Lead	BDL^b	BDL	10.13 ± 0.23	101.3
	Chromium	3.4 ± 0.1	3.27 ± 0.02	12.83 ± 0.13	96.7
	Nickel	14.06 ± 0.06	14.17 ± 0.05	24.75 ± 0.05	102.4
	Copper	19.03 ± 0.1	18.90 ± 0.12	28.56 ± 0.05	98.8
Medicine capsule type 2	Cadmium	1.31 ± 0.04	1.29 ± 0.04	11.41 ± 0.11	101.6
	Lead	BDL	BDL	9.88 ± 0.18	98.8
	Chromium	3.68 ± 0.12	3.74 ± 0.06	13.11 ± 0.19	95.4
	Nickel	27.28 ± 0.02	27.36 ± 0.06	37.07 ± 0.03	99.2
	Copper	24.71 ± 0.02	24.82 ± 0.08	34.94 ± 0.06	100.3

---

3.8. Comparison of analytical performance data with literatures

A comparison between presented sorbent and some of the other IL based sorbents that were used for extraction and preconcentration of some toxic heavy metals from various real samples was performed and the result are shown in Table 5. These results signify that the presented method has made some improvements compared with earlier reported methods with significant merits. For example, the CIs factored in this work compared to other works are the lowest, as well as the sorbent dosage is less. Moreover, the utilized method provides an easy operation of extraction (based on the use of an ionic liquid as the surface modifier for MGO sorbent), as well as facile AAS determination by micro sampler introduction.

Table 5 Comparison between the presented sorbent and some of the other published IL based sorbents

Metal ions	Method	Sorbent	Amount of sorbent (mg)	Instrumental technique	Matrix	CI^a	Ref.
a Consumptive index.b Magnetic solid phase extraction.c Flame atomic absorption spectrometry.d Dispersive solid phase extraction.e Electrothermal atomic absorption spectrometry.
Cd, Co, Ni, Pb and Cr	Ultrasound-assisted, dual charged ionic liquid-linked, magnetic solid phase extraction	Dual charged IL-magnetic graphene oxide	40	FAAS	Medicine capsule samples	0.25	This work
Pb and Cd	MSPE^b	IL-magnetic nanoparticles	40	FAAS^c	Milk and water samples	0.25	35
Cu, Cd, Pb, Ni, Zn and Co	SPE	Modified MWNTs	300	FAAS	Food and real water samples	5	36
Ni	Column-SPE	IL-TiO2	30	FAAS	Food samples	1	37
Cd, Pb, Ni, and Zn	D-SPE^d	Magnetic metal–organic framework	25	FAAS	Fish, sediment, soil and water samples	7.8	38
Cd	Column-SPE	IL-silica	30	FAAS	Water samples	2	39
Pb	Column-SPE	IL-silica	100	FAAS	Water samples	1	40
Cd, Co, Mn, Ni and Pb	Column-SPE	IL-silica	200	FAAS	Water and tobacco samples	0.75	41
Cd, Pb, Zn and Ni	SPE	MWNTs	200	FAAS	Food samples	3	42
Cd, Pb and Ni	SPE	1-(2-Pyridylazo)-2-naphthol impregnated activated carbon	—	FAAS	Soil and, environmental, water samples	10	43
Pb and Cd	Column SPE	MnO2–CNT	50	ETAAS^e	Water	1.5	44

---

This method has some advantages including the following:

(1) In this sorbent, the ionic liquids were chemically bonded on the surface of the MGO, and chemically bonded ILs are stronger than ILs physically adsorbed onto the surface like some reported works.

(2) By immobilizing the ionic liquids onto wide graphene nano sheets, the mass transfer increased in comparison to just ionic liquids and reduced equilibrium times were also seen.

(3) By immobilizing the ionic liquids, the inherent problem of ILs, namely, slow diffusion and loss in the aqueous phase due to the relatively high viscosity of the ionic liquids, could be solved.

(4) Collecting ionic liquids that are dispersed in a solution is difficult. By immobilizing ionic liquids on a magnetic base, the difficulty in separating and need for special instruments (syringes, tubes, centrifuge and others) goes away.

(5) By immobilizing ionic liquids, numerous advantages appears such as easy handling, reduction in the amount of ILs utilized in a typical process relative to homogeneous systems and recycle ability.

(6) A small amount of solvent is used for elution; therefore, it is an environment friendly method.

(7) By consuming a small amount of sorbent, high recoveries can be achieved.

(8) Dispersing the sorbent creates a high contact surface between the sorbent and analytes. Moreover, extraction time will be reduced.

(9) According to Table 5, this IL base microextraction work is comparable with other reported IL base microextraction works used to extract metal ions.

4. Conclusion

To the best of our knowledge, this is the first report using chemically bonded double charged ionic liquids on magnetic graphene oxide for simultaneous extraction and preconcentration of heavy metal ions and mass transfer enhancement in ionic liquid–solid phase microextraction. Herein, it is believed the integration of MGO and ILs will provide new hybrid materials that can open novel branches toward the expansion of solid phase microextraction techniques. Various analytical parameters affecting the recovery, such as pH, sorbent dosage, type and concentration of eluent solvent, amount of complexion agent and others, are evaluated. A very simple, easy to use, inexpensive and environmentally benign method based on preconcentration with a DIL-MGO sorbent and FAAS determination by micro-sample introduction mode was applied for simultaneous extraction and determination of mentioned heavy metals in medicine capsules, which showed enough sensitivity for trace metals determination and satisfactory precision and accuracy. The recoveries for the real samples were between 95% and 102%. In addition, after a total of eight cycles, this sorbent still exhibited satisfactory reusability. Taking advantage of the mentioned properties, the prepared double charged ionic liquid modified magnetic graphene oxide could be expected to be widely used in complex matrix sample preparations.

References

1. A. Jacoby, M. Sudell, C. Tudur Smith, J. Crossley, A. G. Marson and G. A. Baker, Epilepsia, 2015, 56, 460–472.
2. H. Maeda and T. Kurokawa, Ann. Oncol., 2015, 26, 211–216.
3. H. Fazelirad and M. A. Taher, Environ. Technol., 2016, 37, 300–307.
4. M. D. Joshi and J. L. Anderson, RSC Adv., 2012, 2, 5470–5484.
5. L. Shastri, H. N. Abdelhamid, M. Nawaz and H.-F. Wu, RSC Adv., 2015, 5, 41595–41603.
6. E. Yavuz, Ş. Tokalıoğlu, H. Şahan and Ş. Patat, RSC Adv., 2013, 3, 24650–24657.
7. D. Chen, Y. Sun, Y. Yang and Q. Jia, RSC Adv., 2013, 3, 22111–22119.
8. R. Saxena and P. L. Meena, RSC Adv., 2014, 4, 20216–20225.
9. J. Li, X. Hu, P. Guan, X. Zhang, L. Qian, R. Song, C. Du and C. Wang, RSC Adv., 2015, 5, 62697–62705.
10. S. A. Arain, T. G. Kazi, H. I. Afridi, M. S. Arain, A. H. Panhwar, N. Khan, J. A. Baig and F. Shah, Ecotoxicol. Environ. Saf., 2016, 126, 186–192.
11. F. Aydin, E. Yilmaz and M. Soylak, RSC Adv., 2015, 5, 40422–40428.
12. W. Zhu, J. Zhang, H. Li, Y. Chao, W. Jiang, S. Yin and H. Liu, RSC Adv., 2012, 2, 658–664.
13. P. S. Barber, S. P. Kelley and R. D. Rogers, RSC Adv., 2012, 2, 8526–8530.
14. S. Shahriari, L. C. Tomé, J. M. Araújo, L. P. N. Rebelo, J. A. Coutinho, I. M. Marrucho and M. G. Freire, RSC Adv., 2013, 3, 1835–1843.
15. O. Nakari, A. Pihlajamäki and M. Mänttäri, Sep. Purif. Technol., 2016, 163, 267–274.
16. A. Spietelun, Ł. Marcinkowski, M. de la Guardia and J. Namieśnik, Talanta, 2014, 119, 34–45.
17. W. Zhu, P. Wu, L. Yang, Y. Chang, Y. Chao, H. Li, Y. Jiang, W. Jiang and S. Xun, Chem. Eng. J., 2013, 229, 250–256.
18. L. Wang, W. Wen, H. Xiong, X. Zhang, H. Gu and S. Wang, Anal. Chim. Acta, 2013, 758, 66–71.
19. N. Sun, Y. Han, H. Yan and Y. Song, Anal. Chim. Acta, 2014, 810, 25–31.
20. O. Akhavan, E. Ghaderi, R. Rahighi and M. Abdolahad, Carbon, 2014, 79, 654–663.
21. X. Luo, F. Zhang, S. Ji, B. Yang and X. Liang, Talanta, 2014, 120, 71–75.
22. X. Qiao, L. Zhang, N. Zhang, X. Wang, X. Qin, H. Yan and H. Liu, J. Chromatogr. A, 2015, 1400, 107–116.
23. T. Daşbaşı, Ş. Saçmacı, A. Ülgen and Ş. Kartal, Food Chem., 2016, 197, 107–113.
24. H. Qiu, A. K. Mallik, M. Takafuji, X. Liu, S. Jiang and H. Ihara, Anal. Chim. Acta, 2012, 738, 95–101.
25. P. R. Silva, V. O. Almeida, G. B. Machado, E. V. Benvenutti, T. M. Costa and M. R. R. Gallas, Langmuir, 2011, 28, 1447–1452.
26. D. S. Gay, T. H. Fernandes, C. V. Amavisca, N. F. Cardoso, E. V. Benvenutti, T. M. Costa and E. C. Lima, Desalination, 2010, 258, 128–135.
27. M. Bagherzadeh, M. Amrollahi and S. Makizadeh, RSC Adv., 2015, 5, 105499–105506.
28. S. Rezgui, B. C. Gates, S. L. Burkett and M. E. Davis, Chem. Mater., 1994, 6, 2390–2397.
29. H. Li, L. Wang, M. Yang and Y. Qi, Catal. Commun., 2012, 17, 179–183.
30. X. Yang, C. Chen, J. Li, G. Zhao, X. Ren and X. Wang, RSC Adv., 2012, 2, 8821–8826.
31. J.-L. Gong, X.-Y. Wang, G.-M. Zeng, L. Chen, J.-H. Deng, X.-R. Zhang and Q.-Y. Niu, Chem. Eng. J., 2012, 185, 100–107.
32. A. M. Medeiros, F. Machado and J. C. Rubim, Eur. Polym. J., 2015, 71, 152–163.
33. J.-H. Li, X.-C. Hu and Y.-X. Xie, Tetrahedron Lett., 2006, 47, 9239–9243.
34. U. P. G. Chapter, The United States Pharmacopeial Convention, Chapter 232: Elemental Impurities – Limits, Revision Bulletin, 2013.
35. M. D. Farahani and F. Shemirani, Microchim. Acta, 2012, 179, 219–226.
36. M. Tuzen, K. O. Saygi and M. Soylak, J. Hazard. Mater., 2008, 152, 632–639.
37. M. Amjadi and A. Samadi, Colloids Surf., A, 2013, 434, 171–177.
38. M. Taghizadeh, A. A. Asgharinezhad, M. Pooladi, M. Barzin, A. Abbaszadeh and A. Tadjarodi, Microchim. Acta, 2013, 180, 1073–1084.
39. P. Liang and L. Peng, Talanta, 2010, 81, 673–677.
40. S. Ayata, S. S. Bozkurt and K. Ocakoglu, Talanta, 2011, 84, 212–215.
41. M. Gharehbaghi and F. Shemirani, Anal. Methods, 2012, 4, 2879–2886.
42. Z. A. ALOthman, M. Habila, E. Yilmaz and M. Soylak, Microchim. Acta, 2012, 177, 397–403.
43. Z. A. Alothman, E. Yilmaz, M. Habila and M. Soylak, Ecotoxicol. Environ. Saf., 2015, 112, 74–79.
44. B. Yang, Q. Gong, L. Zhao, H. Sun, N. Ren, J. Qin, J. Xu and H. Yang, Desalination, 2011, 278, 65–69.

---