Introduction

The pollution of rivers and streams with chemical contaminants has become one of the most critical environmental problems of the century. Heavy metals as persistent environmental contaminants are of great importance amongst chemical pollutants. At trace levels several heavy metals such as Mn, Zn, Cu, Co and Ni are essential micronutrients for plants, living organisms and the human body whilst most of them are toxic or even carcinogenic at high concentrations. Typically they enter the body via the food chain, ambient air or drinking water. As a consequence, contamination levels in urban and industrial waste water need to be controlled and strict regulations have been drawn up and proposed in this regard. Compliance with the regulations requires evaluation of robust, reliable, sensitive and environmental friendly analytical methods and instruments. However, in many cases where the sample matrices are complex or the concentration of elements fall below the detection limit values of common techniques of trace metal determination such as flame or electrothermal atomic absorption spectroscopy (FAAS and ETAAS) and inductively coupled plasma-optical emission spectrometry (ICP-OES), a separation or preconcentration method such as solid phase extraction or microextraction [1–6], ion exchange [7, 8], cloud point [9, 10] and liquid-liquid extraction or microextraction [11] must be applied.

Dispersive liquid-liquid microextraction (DLLME) utilizes the use of a few milliliters of water-soluble disperser solvents such as methanol, acetonitrile or acetone to disperse microliter volume of water-immiscible extraction solvent into the aqueous medium [12]. In this method, a cloudy solution is formed right after injecting the mixture of extraction and disperser solvents into the aqueous sample where the distribution of the target analytes from sample into the organic solvent takes place. Since its introduction, DLLME has been successfully applied for extraction of organic [13–16] and inorganic [17–20] species in different samples. Despite its several powerful aspects, DLLME’s main drawback lays in the choice of the extraction solvent. In this microextraction technique, solvents of higher densities than water are preferred which are not often compatible with instruments such as ICP-OES and reversed-phase high-performance liquid chromatography (RP-HPLC).

Ionic liquids (ILs) applied at room or non-ambient (controlled) temperatures are considered as relatively recent chemicals of unique properties. Since the first use of an ionic liquid as an alternative to traditional volatile organic solvents for two-phase liquid-liquid separations in 1998 [21], in many cases ILs have demonstrated advantages compared to common solvents used in separation and extraction processes. Liu et al. reported that ILs can be used as extraction solvents in liquid-phase microextraction (LPME), the authors found that the ILs [C6mim] [PF6] and [C8mim] [PF6] both outperformed 1-octanol in direct-immersion and headspace LPME of United States Environmental Protection Agency (US-EPA) high priority polycyclic aromatic hydrocarbons (PAHs) [22]. Several studies have been carried out investigating extraction capability of various types of ILs in different liquid phase microextraction approaches isolating organic [23–32] and inorganic [33–37] species. ILs unique properties has led to evolution of new concepts of phase separation in microextraction techniques based on direct contact of extraction and sample phases such as temperature-controlled ionic liquid based dispersive liquid-liquid microextraction [38] and dispersive liquid–liquid microextraction using an in situ metathesis reaction to form an ionic liquid extraction phase [39].

The major aim of this study is to find a substitute for common DLLME extraction solvents which can be introduced to ICP-OES without further need for either time-consuming evaporation, solidification or any manipulation. This paper represents the practical procedure, characteristics and application of conjoining a green ionic liquid based dispersive liquid-liquid microextraction (IL-DLLME) with FI-ICP-OES for determination of cobalt, copper, manganese, nickel and zinc in various environmental water samples.

Experimental

Apparatus

A radial view Varian Vista-Pro simultaneous inductively coupled plasma optical emission spectrometry (Springvale, Australia, www.varianinc.com) equipped with a V-groove nebulizer and Scott spray chamber utilizing a charge coupled device (CCD) as detector was applied for determination of target elements. The flow-injection (FI) manifold consisted of an adjustable injection loop, made of silicon tube, connected to a six-port two-position injection valve (Tehran University, Iran), in which the 1-propanol/water mixture (80:20% v/v) was used as the eluting solvent. The loop volume was adjusted in order to enable introduction of the entire extraction IL volume in each steps of optimization and real sample analysis. FI-ICP-OES optimal operating conditions and employed spectral lines of the elements are shown in Table 1.

Table 1 FI-ICP-OES operating conditions and spectral lines of the elements
Full size table

The pH measurement and adjustment was performed using an Inolab WTW pH meter (Weilheim, Germany, www.wtw.com) equipped with a combined electrode.

Chemicals and reagents

Stock standard solutions (1000 mg L−1) of Co(II), Cu(II), Mn(II) and Y(III) were prepared by dissolving proper amounts of analytical grade CoCl2.6H2O, Cu(NO3)2.3H2O, MnCl2.4H2O and Y(NO3)3.H2O salts from Merck (Darmstadt, Germany, www.merck-chemicals.com) in ultra-pure water. Ni(II) and Zn(II) standard stock solutions (1000 mg L−1) were purchased from Sigma–Aldrich (Milwaukee, WI, USA, www.sigmaaldrich.com). Mixed working standard solutions were prepared by diluting the stock solutions with ultra-pure water and in case of the internal standard solution of 5 mg L−1 Y3+ with 1-propanol.

Solutions of chelating agents 1-(2-thenoyl)-3,3,3-trifluoracetone (TTA) and 1-(2-pyridilazo)-2-naphthol (PAN) (reagent grade, Merck) were prepared in methanol in 0.5 M concentration.

Ionic liquids 1-hexyl-3-methyl imidazolium bis(trifluoromethylsulfonyl)imide [C6mim][Tf2N] and 1-hexyl-3-methyl imidazolium hexafluorophosphate [C6mim][PF6] were purchased from KimiaExir (Tehran, Iran, www.kimiaexir.com). Ethanol, methanol, acetone, acetonitrile, ammonium acetate, nitric acid, sodium hydroxide and sodium chloride, were obtained from Merck. 1-Propanol was purchased from Panreac (Barcelona, Spain, www.panreac.es). The ultra-pure water was purified using an aquaMAX–Ultra Younglin Ultra-pure water purification system (Seoul, South Korea, www.younglin.com).

Peak area calculation and experimental design softwares

Integration was performed using the freeware CHROMuLAN version 0.79 from PiKRON Ltd (Prague, Czech Republic, www.pikron.com) to which the spectral data were loaded in txt format.

Experimental design data analysis was performed by using the software Design-Expert trial version 7.0.0 from Stat-Ease Inc (Minneapolis, MN, USA, www.statease.com).

Procedure of ionic liquid based dispersive liquid-liquid microextraction (IL-DLLME)

A 30.0 mL portion of sample containing 50 mM of ammonium acetate solution was placed in a 45 mL screw cap glass tube and spiked with 1-(2-thenoyl)-3,3,3-trifluoracetone (TTA), as complexing agent, after appropriate pH (8.0) adjustment, 4 mL of ethanol containing 210 μL of [C6mim][Tf2N] ionic liquid was injected into the sample solution. A cloudy solution was formed and extraction was completed rapidly. After centrifugation for 10 min at 4200 rpm, the upper solution was removed and the 100 μL sedimented IL containing target analytes was collected, diluted with 100 μL 1-propanol and introduced to ICP-OES by the flow injection manifold.

Sample preparation

Sample collection, preservation and storage were all performed according to the US-EPA method 200.7 [40]. Samples were collected in polyethylene bottles, filtered through a 0.45 μm pore diameter membrane filter and acidified with nitric acid to pH below 2.0 following filtration. All samples were preserved at 4 °C before analysis.

Results and discussion

Ionic liquid as extraction solvent and its introduction to ICP-OES

Selection of the extraction solvent is considered a key step of all liquid phase microextraction techniques. In traditional DLLME organic solvents are selected on the basis of their higher density than water, extraction capability of interested compounds along with good instrumental behavior. As a result, halogenated hydrocarbons such as chlorobenzene, chloroform, carbon tetrachloride and tetrachloroethylene are the most widely used solvents. However, there have been reports on application of low density solvents, either collected by using especially designed vials [41] or solidification of the floating organic drop [42], still the traditional sedimentation is the most convenient and widely used way of collecting the organic solvent due to unavailability of the designed vials or the time-consuming solidification step.

Introduction of organic solvents to the ICP-OES usually results in plasma cooling, instability or even shut down due to high vapor pressure and intensive background emission of such solvents, whereas such phenomena are not observed for 1-propanol. Lower vapor pressure of 1-propanol (b.p. 98 °C) resembling water more than commonly used organic solvents such as methanol, ethanol and acetonitrile and its miscibility with both aqueous and organic media, either pure or mixed with water, has led to its application as the eluting solvent for easy introduction of different types of solvents to the ICP-OES without any plasma instability problem or need for further manipulation or evaporation of the solvent [43]. Nevertheless, with the use of 1-propanol as the eluting solvent not every type of organic solvents can be introduced to the ICP-OES. High vapor pressure and emissive behavior of the halogenated hydrocarbons, considered as the traditional DLLME extraction solvents, would lead to plasma shut down when applied directly and would result in plasma instability and baseline intensive drift when applied as a 1:1 mixture with 1-propanol. In this regard, non-volatile dense ionic liquids can be applied as the suitable substitutes of common halogenated extraction solvents. The ionic liquid used in the current study ([C6mim][Tf2N]) was selected due to its less solubility in water, less viscosity which enables volumetric measurements with high surface area syringes and better emissive behavior in ICP-OES compared to [C6mim][PF6]. Figure 1 compares the emissive behaviour of common dense solvents and the suggested IL introduced to the FI-ICP-OES. It represents injection of 1:1 mixture of CHCl3, CH2Cl2 and chlorobenzene in 1-propanol; each 2 injections in a row as well as the effect of [C6mim][Tf2N]/1-propanol 1:1 mixture injection on the FI-ICP-OES baseline. The smooth and stable baseline for IL injection proves its compatibility with ICP-OES.

Fig. 1
figure 1

Baseline obtained for introduction of (a) common DLLME solvents (b) [C6mim][Tf2N] to the ICP-OES at 238.892 nm; 1: CHCl3, 2: CH2Cl2 and 3: C6H5Cl

Full size image

Complexing agent effect

The ionic liquid was insufficient for extraction of the charged metal ions, it was necessary to improve their affinity for the IL phase by complex formation with a suitable reagent. Imidazolium based ILs present a high chemical affinity towards compounds containing one or more aromatic rings in their structures [44]. TTA and PAN, two complexing reagents containing aromatic rings, were tested and TTA was found to be more efficient one for extraction of target metal ions as shown in Fig. S1 (Electronic Supplementary Material, ESM). The extracted species were anionic M(TTA) -3 for M = Mn, Co, Ni and Zn and neutral M(TTA)2 in case of Cu [45].

Selection of disperser solvent

The choice of the disperser solvent was done considering its miscibility with the IL phase and aqueous sample. Two milliliters of acetonitrile, acetone, methanol and ethanol containing proper amount of the IL to obtain the same settled phase volume were evaluated and as demonstrated in Fig. S2 (ESM) ethanol yielded the highest peak area.

Experimental design and data analysis

A half fractional factorial design was employed in 20 runs containing 4 center points to investigate the most influential of the five variables including A: salt% (0-10%), B: pH (3.0–9.0), C: ligand to metal molar ratio (5.0–30.0), D: volume of the extraction solvent (100–250 μL) and E: volume of the disperser solvent (1.0–6.0 mL) affecting the extraction of the target elements. The total normalized peak area was used as the response. The normalization was done by using Eq. 1, where M j represents M = Co, Cu, Mn, Ni and Zn with j = 1–5 values, i refers to the run number (i = 1–20) and k refers to the run having the smallest peak area values amongst the entire runs. As the Pareto chart in Fig. 2 presents, all of the factors apart from salt% were evaluated influential at the 0.05% level so salt addition was excluded from the optimization step.

Fig. 2
figure 2

Pareto chart of the main effects of variables obtained from half-fractional factorial design

Full size image
$$ Total\,Normalized\,Peak\,Area = \sum\limits_{{i = 1,j = 1}}^{{i,j = 5}} {\left( {\frac{{{M_j}\,Peak\,area\,of\,run(i)}}{{{M_j}\,Peak\,area\,of\,run(k)}}} \right)} $$
(1)

A face-centered central composite design (FC-CCD) consisting of 30 runs including 6 center points was applied to optimize the four factors of effectiveness chosen in the screening design. Some of the obtained response surfaces are shown in Fig. 3. As Table 2 represents the predicted optimal response and the observed experimental responses concurred.

Fig. 3
figure 3

The three-dimensional response surfaces obtained from face-centered central composite design. Each surface represents the variating pattern of the response with the two indicated factors changing values while the other ones are kept constant at their central values

Full size image
Table 2 Predicted optimal values and response along with the corresponding observed experimental response
Full size table

Effect of potentially interfering elements

The impact of potentially interfering elements on the extraction and determination of heavy metals was investigated adding different amounts of Na+, K+, Li+, Mg2+, Ba2+, Ca2+, Al3+, Hg2+, Pb2+ and Fe3+ into the tested sample solution containing 100 ng mL−1 of the target elements. The experimental results are shown in Table 3. The tolerance limit is defined as the highest amount of potentially interfering element which results in the recovery of the studied species in the 90–110% interval. Amongst the aforesaid elements, two had to be masked with suitable masking agents; Al3+ was masked with F- in order to form soluble aluminum complexes preventing its precipitation as Al(OH)3 at the extraction pH of 8.0. Fe3+ was also masked with SCN- due to its high affinity for complex formation with TTA compared to those of target elements as a result of which extraction in its presence resulted in very low extraction recoveries of the desired elements.

Table 3 Effect of presence of the potentially interfering elements on the extraction and determination of heavy metals
Full size table

Analytical performance

Under the optimized extraction conditions, the analytical performance of the developed method in terms of enhancement factor, detection limit, linearity and repeatability was evaluated in ultra-pure water, the results are shown in Table 4. The calibration curves were obtained after the standard series were subjected to the IL-DLLME followed by introduction to FI-ICP-OES as described previously. It is noteworthy to mention that the diluting 1-propanol contained Y3+ as internal standard, in this regard peak area of each point of the curves was divided to its analogous Y3+ peak area at 377.433 nm.

Table 4 Analytical performance of the IL-DLLME-FI-ICP-OES method
Full size table

Table 5 offers a comparative vision of the characteristics of the current method with previously developed methods. As it shows, the evaluated method offers comparable analytical performance along with rapid and sensitive trace multi-element analysis.

Table 5 Comparison of the characteristics of the current method with previously developed methods
Full size table

Real sample analysis

To evaluate the practical applicability of the developed method, determination of target elements concentration was done in three different environmental water samples consisting of fish farming pond water, factory waste water and tap water. Farming pond water sample was collected from water outlets of a salmon farming pond constructed in one of the Haraz river branches in north of Iran. Factory waste water was the final outfall of waste water treatment plant of one of the Iran’s largest automobile industrial complexes. Tap water, the main source of the city inhabitants drinking water, was collected from Tehran city tap water. The results are given in Table S1 (ESM). As the recovery values of the spiked samples were not in the range of 90-110% for the last two samples, the standard addition approach was used for quantification of the target elements.

Conclusion

High sample throughput, low detection limits, sensitivity and reproducibility are some of the advantages offered by the IL-DLLME-FI-ICP-OES method, combining DLLME as an efficient, rapid preconcentration method and ICP-OES as a powerful sensitive multi-element determination technique has created a powerful method of trace metal analysis in different matrices. Application of the current method for determination of the target elements in different water samples showed satisfactory results, implying robustness of this combination. To the best of our knowledge, this is the first work reporting the introduction of the green unique ionic liquids to the ICP-OES. This can lead to the use of task-specific ILs for simultaneous selective extraction and speciation of different metallic species by the ICP-OES. Furthermore the current study introduces a new concept for introduction of ILs to the coupled separation-spectroscopic techniques such as HPLC-ICP-OES/MS widely used in drug-metabolite profiling [46]. Problems of introducing large amounts of organic solvents to the plasma in such systems may be overcome using 1-propanol, the plasma-compatible organic solvent, as the organic portion of the mobile phase [47].