Ultrasound-assisted dispersive solid-phase extraction for the preconcentration of dyes in water samples

2.1 Materials and instrumentation

The chemicals used in this research, such as sodium hydroxide (97%), zinc nitrate (99%), methanol (99%), hydrochloric acid (37%), acetonitrile (99%), tetrahydrofuran (99%), iron(III) nitrate (99%), magnesium nitrate (99%), hydrogen peroxide (30%), and dimethylformamide (99%), were purchased from Merck company. Safranin O (99%) and methylene blue (99%) were supplied by Fluka. All chemicals were of analytical grade purity and were used without any additional purification. To prepare a standard solution of 100 mg/L, a quantity of 0.01 g of the dye was dissolved in a small amount of distilled water and made up to a volume of 100 mL in a volumetric flask using distilled water. Daily solutions were prepared by diluting the aforementioned solution with distilled water. pH values were measured using a Metrohm 780 pH meter (Metrohm, Switzerland), and weighing was done using a Sartorius AG Germany balance. Phase separation was performed using a centrifuge device (Hermle, Z 300 K, Germany). Dispersion of the solid and liquid phases was carried out in an ultrasonic bath (UC-3360, China). UV–vis spectra were measured using a UV/Vis spectrophotometer from Shimadzu Corporation, Japan. XRD patterns were obtained using a Philips PW3040/60 powder diffraction instrument with Cu K_α radiation (40 kV and 40 mA) in the 2θ range of 20–80°. TEM images were captured using an H-600 transmission electron microscope from Hitachi, Japan. The magnetic properties of the magnetic particles were evaluated using a vibrating sample magnetometer (VSM) model Lake Shore 7307.

2.2 Synthesis of magnesium ferrite/zinc oxide (MgFe₂O₄/ZnO) nanocomposite

2.3 Experimental procedures

The extraction method using ultrasound was performed as follows: 10 mL of the desired concentration of dye (500 µg/L) was poured into a 15 mL glass tube and adjusted to pH = 7. Then, 1.53 mg of MgFe₂O₄/ZnO nanocomposite was added to the test tube. Subsequently, the solution was placed in an ultrasonic bath for 8 min to facilitate mass transfer and dye adsorption onto the MgFe₂O₄/ZnO nanocomposite. After the desired time elapsed, the adsorbent was separated from the solution using a centrifugation method (5 min at 3500 rpm) and an external magnetic field. To complete the extraction process, 140 μL of methanol was added to the settled adsorbent. Then the test tube was subjected to ultrasonic treatment for 5 min to enhance the desorption of the analyte from the surface of the adsorbent. The test tube was once again placed inside a centrifuge device for 5 min at 3500 rpm to settle the adsorbent. Then, 20 µL of the liquid extraction phase containing the dye was drawn using a micropipette and analyzed using a UV–Vis spectrophotometer. The response (extraction recovery, ER%) was calculated based on Eq. (1).

In this equation, C_sed represents the concentration of dye in the precipitated phase, and C₀ denotes the concentration of dye in the sample before extraction. Additionally, V_sed represents the volume of the precipitated phase, and V_aq corresponds to the total volume of the sample.

2.4 Experimental design

RSM is a collection of statistical techniques used in process optimization, where the desired response is influenced by several variables (Rind et al., 2023). With the aid of this statistical design, the number of experiments is reduced, and all second-degree regression model coefficients and interaction effects of the factors can be estimated (Sirichan et al., 2022, Shojaei et al., 2021). Additionally, RSM allows for extracting more information compared to the OFAT method. In the present study, Box-Benken design (BBD) based on RSM method was used to evaluate the coefficients of the selected model. Therefore, the four factors (adsorbent amount, ultrasonic time, pH, and eluent volume) were utilized at three different levels, resulting in a total of 29 experiments conducted based on the BBD, as outlined in Table 1. The experimental conditions included adsorbent amount (X₁) of 0.5, 1.5, and 2.5 mg, pH (X₂) values of 4, 7, and 10, eluent volume (X₃) of 50, 150, and 250 µL, and ultrasonic time (X₄) of 4, 8 and 12 min. The optimal conditions were determined using a quadratic equation. To examine the effects and interactions of these parameters on the recovery, significance tests and analysis of variance (ANOVA) were performed, and the results were visually represented through surface plots generated using Design Expert 10 software.

3.1 Characterization of the synthesized nanocomposite

Fig. 1a illustrates the XRD pattern of ZnO nanoparticles, MgFe₂O₄ nanoparticles, and MgFe₂O₄/ZnO nanocomposite. According to the Fig. 1a, it was observed that only in the pure ZnO sample, the peaks were broader compared to the other samples. A comparison of the XRD patterns showed no shifts or impurity peaks in the samples. Additionally, the formed peaks of the MgFe₂O₄/ZnO nanocomposite indicated a decrease in the intensity of the MgFe₂O₄ peaks upon the addition of ZnO. This could be attributed to the surface coverage of MgFe₂O₄ nanoparticles by ZnO nanoparticles. Based on Fig. 1a, it can be observed that there were no impurity peaks in the XRD pattern of the prepared ZnO nanoparticles, and the formed peaks perfectly matched the standard reference (JCPDS 36–1451). Similarly, there were no impurity peaks in the XRD pattern of the MgFe₂O₄ nanoparticles, and the formed peaks were in complete agreement with the standard reference (JCPDS 73–1720) for MgFe₂O₄ nanoparticles (Su et al., 2014). The crystallite size of the MgFe₂O₄/ZnO nanocomposite was approximately 10 nm, as determined using the Debye-Scherrer equation. Fig. 1b shows the FE-SEM image of the MgFe2O4/ZnO nanocomposite. As depicted in Fig. 1b, the MgFe₂O₄/ZnO nanocomposite appeared spherical and exhibited a uniform distribution. No significant aggregation was observed in the image. The average nanoparticle size was calculated to be 20 nm. Fig. 1c displays the TEM image of the MgFe₂O₄/ZnO nanocomposite. Based on Fig. 1c, the formation of fine crystallites with an average size of approximately 10 nm can be observed. The magnetization curves of MgFe₂O₄nanoparticles and MgFe₂O₄/ZnO nanocomposite as a function of the magnetic field are shown in Fig. 1d. In Fig. 1d, it can be observed that the samples exhibit a soft magnetic nature with low coercivity. The values obtained from these curves, including saturation magnetization (M_s), residual magnetization (M_r), and coercivity (H_c) of the samples, are presented in Table 2. Furthermore, based on Fig. 1c, it was observed that the introduction of non-magnetic zinc oxide nanoparticles resulted in a reduction of the saturation magnetization of pure magnesium ferrite nanoparticles.

Close

Fig. 1

A) xrd patterns, b) fesem image, c) tem image, and d) vsm spectra of mgfe₂O₄/ZnO nanocomposite.

Export to PPT

3.2 Effects of type of eluent

One of the important parameters in the USP-DSPE method is the selection of a suitable eluent. The eluent used should have the ability to desorb the maximum amount of the desired compound. To achieve this, a series of preliminary experiments were conducted before the actual experimental design to select the appropriate eluent with efficient and reproducible analyte desorption from the surface of the nanoparticles. In this study, under identical conditions, methanol, tetrahydrofuran, acetonitrile, and dimethylformamide were examined for selecting the appropriate eluent. Based on Fig. 2, among the investigated solvents, methanol was chosen as the eluent due to its superior extraction results and efficiency. Consequently, methanol was used as the eluent for further studies.

Close

Fig. 2

Effect of eluent type on the recovery of MB and SO.

Export to PPT

3.3 Effect of the type of mixing

Mixing is another important parameter in the solid-phase extraction method, as it ensures the uniform distribution of the adsorbent in the solution and enhances its interaction with the analyte. For this reason, different mixing techniques were employed to assess their effects on the extraction performance. For this purpose, three types of mixing, including manual, mechanical, and ultrasonication, were applied to the samples for a 5 min. Additionally, a control extraction without mixing was conducted, and its results were compared to those obtained from the mixing methods. As shown in Fig. 3, all three mixing techniques exhibited a more significant impact on the extraction compared to the non-mixed experiment. Based on the results, ultrasonication was selected to achieve the highest extraction efficiency.

Close

Fig. 3

Effect of mixing type on the recovery of MB and SO.

Export to PPT

3.4 Fitting of the process model and statistical analysis

In order to obtain the multivariate regression model and predict the response, various linear, quadratic, and polynomial relations were fitted. Then, these models were statistically analyzed so that the appropriate model could be selected. The results of ANOVA of the regression model are shown in Tables S1 and S2. For the quadratic model, it was found that the value of R² is equal to 1, implying the significance of the fitted model in describing response changes as a function of independent variables (Mohammed et al., 2023, Ameer et al., 2017). The R² values for MB and SO were 0.9965 and 0.9987, respectively, which means that more than 99.6% of the data changes can be explained. The proximity of R² and Adjusted R² values indicates no significant error in the model (Igwilo et al., 2022). In addition, actual values are measured through measurable experiments, and predicted values are estimated through the model. The Adjusted R² values for MB and SO were 99.29% and 99.74%, respectively, which can be used as a regression model to predict the values of the function variables at different levels of F-value. The coefficients of the dependent variables showed that most of the coefficients were significant (P-value < 0.0001). Adeq-precision measures the signal-to-noise ratio (Kushwah et al., 2023). The most suitable case is a value greater than 4, with 51.60 for MB and 82.61 for SO, indicating a reasonable signal. Also, the quadratic model was significant for efficiency percentage (P-value < 0.0001). Based on the results, the fitted models for each of the extraction response are in terms of Eqs. (2) and (3):

In addition to the statistical criteria to evaluate the model's accuracy, the difference between the predicted and experimental responses (residuals) was used to check the model's accuracy graphically. Also, residuals are considered as unadjusted changes by the model (Sizykh et al., 2023, Saleh et al., 2017). In order to analyze the normality of the test data, the normal probability curve is shown in Fig. S1a and S1b. The proximity of the points to the straight line indicates the normal distribution of the error with a mean of zero and a constant value. The residual diagram is given in Fig. S1c and S1d to evaluate the normal distribution of residues. Fig. S1c and S1d show the consistency between the experimental and predicted values, indicating a good match between R² and Adjusted R².

3.5 Study of interactive effects of different parameters on the extraction efficiency

A perturbation plot was utilized to analyze the effects of various factors on a specific point in the design space (Fig. 4a–c). The plot displayed the extraction efficiency (ER%) while changing one factor at a time within its range, while maintaining the other factors constant. This allowed for examining the effects of all factors at their central point, which included a pH value of 7, 1.53 mg of MgFe₂O₄/ZnO, 140 µL of methanol, and 8 min of sonication.

Close

Fig. 4

Response surface and contour plots for combined effect of pH and adsorbent amount (a), ultrasonic time and adsorbent dosage (b), eluent volume and pH (c) on the ER% of MB and SO by USA-DSPE method.

Export to PPT

Fig. 4a and 4c demonstrate that the maximum recovery of MB and SO (95%) was achieved at pH = 7, while the minimum recovery (50%) was observed at pH values below 4. This behavior can be explained by considering the protonation and deprotonation of surface functional groups of dye and the sorbent (Zahara et al., 2023). In alkaline conditions, the positively charged dye was strongly attracted to the negatively charged MgFe₂O₄/ZnO nanocomposite, resulting in a higher sorption recovery of dye. Conversely, the lowest recovery in acidic conditions is attributed to the increased interionic repulsion between the cationic analyte species and the MgFe₂O₄/ZnO nanocomposite adsorbent. Alhmaunde et al., Ozalp et al., and Zhang et al. have also reported similar finding regarding the influence of pH on the recovery of dyes (Alhmaunde et al., 2022; Ozalp et al., 2023; Zhang et al., 2020).

The experimental results presented in Fig. 4b confirm that the maximum extraction efficiency (ER%) of SO can be achieved with sonication times exceeding 8 min, indicating the method's rapidity and high efficiency. The initial rapid sorption is attributed to the increased available surface area and vacant sites, which result from the effective dispersion of the adsorbent in the solution through the application of ultrasonic power (Li et al., 2019). Dil et al., in their study, obtained similar results regarding the increase in extraction recovery with the increase in ultrasonic time in the extraction of dye using the ultrasound-assisted dispersive solid-phase microextraction method (Dil et al., 2016).

The combined effects of eluent volume and pH on the extraction of SO (Fig. 4c) indicate that increasing the pH and eluent volume results in higher recovery rates. Furthermore, elution with a small volume of eluent allows for maximum preconcentration and enrichment factor (Ragheb et al., 2021). However, using an excessive amount of adsorbent in the USA-DSPE method reduces the dispersibility of the eluent volume and increases the recovery of SO (Nguyen et al., 2023).

Increasing the amount of adsorbent leads to an increase in the contact surface area and active binding sites. Therefore, a higher presence of binding sites provides more effective functional groups that play a crucial role in the adsorption process of SO. As shown in Fig. 4a and 4b, increasing the amount of adsorbent enables the establishment of more bonds between the cations of the dye and the opposite charges on the functional groups of the adsorbent. The increase in the number of these interactions between dye molecules and the surface of the adsorbent results in higher dye adsorption from the aqueous solution and improves the extraction efficiency (An et al., 2022). In a study, Shojaei et al. achieved similar results in evaluating the effect of adsorbent amount on the extraction recovery of MG and MB dyes. Their results showed that the extraction efficiency of MG and MB dyes increased with the increase of adsorbent amount (Shojaei et al., 2023).

3.6 Optimization of extraction conditions

Aside from achieving a higher recovery percentage, less solvent and adsorbent consumption is considered to reduce the cost of operation. In order to verify the extraction of dyes, the effect of four factors (eluent volume, ultrasonic time, adsorbent amount, and pH) was investigated. In order to investigate the effect of these factors on dyes recovery, the optimization procedure was used. Table 3 shows the levels associated with each of the factors. All experiments were performed based on the proposed method. In all these experiments, extraction recovery was considered as the answer. Based on the results, the highest recovery percentage of dyes extraction in optimal parameter conditions by the USA-DSPE method was 98.60% and 97.23% for MB and SO, respectively.

3.7 Study of sample volume and preconcentration factor

In order to measure very low concentrations and achieve high preconcentration factors in extraction methods, the maximum possible sample volume containing the analyte (which leads to the highest extraction efficiency under optimal conditions) is utilized. Based on this, sample volumes ranging from 5 to 20 mL containing 500 µg/L of the investigated dye were used. As observed in Fig. 5, the extraction efficiency remains approximately constant up to a volume of 10 mL, and it decreases at higher volumes. Therefore, by selecting 10 mL as the sample volume, a preconcentration factor of 71.5 was obtained.

Close

Fig. 5

Effect of sample volume on the recovery MB and SO dyes.

Export to PPT

3.8 Analytical performance of the USA-DSPE method

The USA-DSPE method was investigated for preconcentration and determination of MB and SO value in optimal conditions. Analytical properties include linear range (LR), the limit of detection (LOD), the limit of quantification (LOQ), relative standard deviation (%RSD), coefficient of determination (R²), enrichment factor (EF), preconcentration factor (PF), and extraction recovery (ER) was calculated by optimized variables to evaluate the performance of the extraction method and resulted presented in Table 4. The detection limit was calculated by 3Sb/m. Sb is the standard deviation of the control, and m is the slope of the calibration curve. LOD values for MB and SO were 1.5 µg/L and 3 µg/L, respectively. The relative standard deviation for five measurements of dyes with a concentration of 500 µg/L was<3% for both dyes. The dyes measurement calibration diagram was linear in the range of 5–1000 µg/L for MB and 10–1100 µg/L for SO. The results showed that the extraction efficiency of method is in the range of 96.16%–98.02%.

3.9 Study of interferences

One of the key parameters that contribute to understanding the capability of an analytical method is evaluating its performance in the presence of interfering species. By examining the interferences, the selectivity of a method can be determined. Essentially, the method's capability should be sufficient to accurately measure the target analyte among various other species with different concentrations. To achieve this, the interferences of different species on dye extraction were investigated under optimal conditions. In this investigation, if the target species resulted in more than a 5% error in dye measurement, it was considered as an interference. The results of this study are presented in Table 5, indicating that the proposed method exhibits relatively high selectivity.

3.10 Reusability of the MgFe₂O₄/ZnO nanocomposite

MgFe₂O₄/ZnO nanocomposite can be easily separated from the solution, washed, and reused without changing its structure. To separate MgFe₂O₄/ZnO nanocomposite, after completing the reaction, an external magnetic field (magnet) was applied to the container, and magnetic nanoparticles intensely moved toward the magnet and accumulated. In this case, the solution overflowed into another container, and the remaining MgFe₂O₄/ZnO nanocomposite in the first container was washed with methanol. Afterward, the remaining MgFe₂O₄/ZnO nanocomposite was dried at ambient temperature to be prepared for the subsequent use. The process of separating and using MgFe₂O₄/ZnO nanocomposite was replicated in 6 consecutive times. The recovery rate of MgFe₂O₄/ZnO nanocomposite and the activity of MgFe₂O₄/ZnO nanocomposite in the sample reaction were evaluated. On average, more than 95% of MgFe₂O₄/ZnO nanocomposite was recovered from the reaction medium each time, and maintaining the properties of MgFe₂O₄/ZnO nanocomposite yielded more than 90% efficacy under optimal conditions. The incurred results approved the recovery property for MgFe₂O₄/ZnO nanocomposite (Fig. 6a and 6b).

Close

Fig. 6

The reusability of adsorbent for extraction of a) MB and b) SO.

Export to PPT

3.11 Dyes measurement in environmental water samples

After determining the optimal conditions, several environmental water samples were analyzed to assess the applicability of the USA-DSPE method. The water samples included tap water, river water, mineral water, and wastewater. Prior to analysis, the samples were filtered using 450 μm filter paper to remove impurities. Subsequently, a specific amount of dyes was added to the water samples, and the analysis was conducted under optimal conditions. The laboratory data were documented in Table S3. According to the results presented in Table S3, the extraction efficiency of MB and SO dyes from natural water samples ranged from 93.74% to 99.11%, with an RSD (n = 3) ranging from 1% to 4%. Based on these findings, the USA-DSPE method exhibits the potential for extracting MB and SO dyes in various water samples. The laboratory data represent the average of three replicates with a 95% confidence level.

3.12 Comparison of the presented procedure with other methods

To compare the performance of the USA-DSPE method used in this study, these results were compared with various studies used for dyes extraction (Table 6). The amount of LDs and LODs reported in the present research can be compared with other methods. Also, the relative standard deviation (RSD) values of the methods in this study are lower than other methods (RSD < 3), while in most similar studies, the declared RSD is higher than 3%, which has high accuracy. Most of the aforementioned methods have utilized high-sensitivity and selective analysis devices such as HPLC or MS. However, the extensive application of these devices for dye analysis is impractical in developing countries due to their high cost. Consequently, the proposed approach, in conjunction with spectrophotometry, offers a suitable alternative for the preconcentration and extraction of trace amounts of dyes from aqueous samples. This method exhibits notable advantages, including high sensitivity, simplicity, rapidity, and reproducibility. Also, the results of this study showed that using the RSM technique reduces the number of tests to determine the optimal extraction conditions, and it can also reduce investigation time and costs.