Application of Cu-based metal-organic framework (Cu-BDC) as a sorbent for dispersive solid-phase extraction of gallic acid from orange juice samples using HPLC-UV method

2.1 Materials and instruments

GA, Cu (II) nitrate threehydrate (Cu(NO₃)₂·3H₂O), terephthalic acid (H₂BDC), dimethylformamide (DMF), NaOH, HCl, ethanol, methanol, hexane, acetone, acetic acid, and acetonitrile were purchased from Merck (Darmstadt, Germany). Ultrapure water was prepared from a MilliQ gradient water purification system. All instruments including FT-IR, XRD, EDS and SEM for analysis were applied according to our previous works (Arabi et al., 2017b; Bagheri and Ghaedi, 2019). Also, HPLC analysis was done according to our previous work using the mobile phase consisted of water-methanol (80/20, v/v) in which the pH was adjusted at 2.5 by phosphoric acid and the flow rate was set to be 1.0 mL min⁻¹ (Arabi et al., 2017b).

2.2 Synthesis of Cu-BDC

For the synthesis of Cu-BDC, 1.0 g H₂BDC and 4 g Cu(NO₃)₂·3H₂O were fully dispersed in 75.0 mL (DMF: ethanol, 50:25 mL) and then the obtained solution was transferred to a 100 mL polytetrafluoroethylene lined stainless steel autoclave and placed in an oven at 120 °C for 12 h. After that, it was taken out and cooling down at room temperature. Then, the product was separated and washed several times with DMF and water. Finally, MOF was obtained by drying the product at 80 °C oven for 12 h.

2.3 Adsorption test

To investigate the adsorption capacity of Cu-BDC, 30.0 mg of Cu-BDCwas suspended into 15 mL solution of GA with varying concentrations (5.0–50.0 mg L^{− 1}) by mechanical stirring for 30 min at ambient temperature. Then, the Cu-BDCwas separated from the sample solution by centrifuge and residual (non-adsorbed) concentrations of GA in the supernatant were analyzed by HPLC. The equilibrium adsorption capacity (Q, mg g⁻¹) of Cu-BDCwas considered based on the following equation:

2.4 Preparation of calibration standard and Cu-BDC-DSPE procedure for the determination of GA from orange juice sample

Orange juice samples were purchased from the local supermarket in Shiraz and stored at 4 ^◦C before use. All samples were centrifuged and filtered through a 0.45 µm filter before DSPE. A standard stock solution of GA (100.0 mg L⁻¹) was prepared by dissolving 10 mg of GA in 100 mL of ultrapure water. The working standard solution of GA was prepared by diluting the stock solution with water. Cu-BDC-DSPE-HPLC/UV procedure for the extraction of GA from orange juice sample is presented in Fig. 1a. In this regard, 15 mL of orange juice solution with the GA spiked concentration of 200.0 µg L⁻¹ was adjusted at pH 8.0 and the solution was mixed with 30.0 mg of Cu-BDC in the beaker under the ultrasound irradiation for 5.0 min. Then, the Cu-BDC was separated using centrifuge and washed by 3.0 mL of hexane as washing solvent and subsequently with 2 mL of methanol: acetic acid (90:10, v/v) as eluent. The eluent was evaporated to dryness under nitrogen stream in water bath (35 °C) and the residue was subsequently reconstituted in 100 µL mobile phase (water-methanol (80/20, v/v)) and 20 µL of it ready for HPLC-UV analysis.

Close

Fig. 1

The Cu-BDC based DSPE procedure for gallic acid extraction (a) and the basic preparation procedure of Cu-BDC (b).

Export to PPT

3.1 Synthesis and characterization of Cu-BDC

The Cu-BDC was prepared using a simple and one-pot solvothermal method (Fig. 1b). In this regard, Cu(NO₃)₂·3H₂O was used as a source of Cu²⁺ and H₂BDC was used as ligand. The COO- groups of H₂BDC can interact with Cu²⁺ ions and form a Cu-BDC material by simple solvothermal method. Happily, the presented method represents a significant advantage of eco-friendliness, cost-savings, simplicity, rapidity, and efficiency. After synthesis of Cu-BDC, it was fully characterized using FT-IR, XRD, SEM, EDS and BET analysis.

Fig. 2a, represents the FT-IR spectra of prepared Cu-BDC. The observed peaks at 488 cm⁻¹ and 721 cm⁻¹ could be attributed to the bending and stretching modes of Cu-O, respectively (Abbasi et al., 2017). The peaks at 663–766 cm⁻¹ are related to the ring in H₂BDC. In addition, aromatic carbon C—C vibrational mode resulted in the bands at about 1390 cm⁻¹. Also, the peaks at 970 cm^−1, 1500 cm⁻¹, and 1640 cm⁻¹ are related to C—O, to —C⚌C— and C⚌O groups of H₂BDC, respectively that proved the successful synthesis of proposed Cu-BDC. Fig. 2b, shows the XRD pattern of Cu-BDC in which the peaks at 2θ of 7.0°, 13.8°, 15.7°, 19.6°, 30.0°, 37.0° 42.5°, 62.5°, and 74.0 ° are related to (2 2 0), (2 2 2), (3 3 3), (4 2 0), (4 2 2), (7 7 3), (8 8 2), (4 4 0) and (5 3 3) crystal planes respectively and are consistent with previous reported literature (Cambridge crystallographic information data with deposit number of 112954) (Kaur et al., 2019; Lin et al., 2014; Liu et al., 2012; Shete et al., 2018; Tari et al., 2016). Fig. 3a and b show the SEM images of prepared Cu-BDCs with different magnitude. SEM images of prepared Cu-BDCs represent the regular and ordered structure of Cu-BDCs which provides direct evidence for the successful synthesis of the Cu-BDCs. As can be seen from SEM images, the prepared Cu-BDCs have a rod and long bar-like structure. In addition, some of the prepared Cu-BDCs have an octahedral structure. Also, according to SEM images, the prepared Cu-BDCs have a 3D structure. Fig. 4a and b show the EDS and BET analysis of prepared Cu-BDC in which the presence of C, O and Cu elements confirmed the successful synthesis of Cu-BDC. Also, the prepared Cu-BDC has high specific surface area (400 m² g⁻¹) which make it as good sorbent for using in extraction field.

Close

Fig. 2

The FT-IR spectra (a) and XRD pattern (b) of prepared Cu-BDC.

Export to PPT

Close

Fig. 3

The SEM images of prepared Cu-BDC.

Export to PPT

Close

Fig. 4

The EDS and BET analysis of prepared Cu-BDC.

Export to PPT

3.2 Optimization of parameters for DSPE based Cu-BDC of GA from orange juice sample

To achieve maximum extraction recovery of GA in orange juice samples, the basic influences of qualitative factors such as pH and kind of washing and eluent solvents were assessed, at the spiked concentration level of 0.5 mg L⁻¹.

One of the most important factors that can affect the structure of both analyte and sorbent is pH. To achieve the maximum extraction recovery of GA, the effect of sample pH was evaluated in the range of 3.0–10.0. The results confirmed that the maximum extraction recovery was achieved at pH 8.0. This phenomenon could be related to structure of GA and MOF. Indeed, GA acid has carboxylic acid and hydroxyl groups which have different structure in different pH. The pKa values of carboxylic acid and hydroxyl groups are 4.1 and 8.3, respectively (Arabi et al., 2017b). In pH values higher than 4.1, the carboxylic acid group of GA get negative charge and change to carboxylate group, while the hydroxyl groups of GA are neutral at pH values lower than 8.3. Based on the structures of GA and Cu-BDC, the adsorption of GA from orange juice sample by the Cu-BDC, might be attributed to the combined action of hydrogen bonding, π-π interaction, and also coordination bonds. On the other hand, we envisioned that the dominant adsorption mechanism is hydrogen bonding. In pH values higher than 4.1, the carboxylate group of GA can interact with Cu²⁺ ions via coordination bonds. In addition, the aromatic structure of GA allows it to interact with the aromatic ring of H₂BDC via π-π interaction. Also, in the pH values between 4.1 and 8.3, the hydroxyl groups of GA can interact with COO- groups of H₂BDC through hydrogen bonding. Overall, at a sample pH of 8.0, the maximum extraction recovery was achieved that is logical. Therefore, the sample pH of 8.0 was set for the next experiments.

The orange juice sample contains complicated matrices and impurities which not only can retain on the sorbent and HPLC column but also can overlap with the GA peak and subsequently decrement the performance of the proposed method. So, the washing step is required to remove these impurities. Therefore, the different washing solvent with different polarities including deionized water, methanol, ethanol, acetonitrile, acetone, and hexane were examined to attain the cleanest extracts. The results displayed that the cleanest extraction was achieved when hexane was applied as a washing solvent. This phenomenon could be related to the rinsing strength of hexane as a nonpolar solvent for washing nonpolar interferents while adsorbed GA retain on Cu-BDC. Another important factor in extraction process is choosing of appropriate eluent solvent which should be able to desorb GA from Cu-BDC. In this regard, different solvents and their mixture including methanol, ethanol, acetonitrile, methanol: ethanol (50:50, v/v), methanol: acetic acid 90:10 and ethanol: acetic acid 90:10) were assessed. The results proved that maximum extraction efficiency was achieved when methanol: acetic acid 90:10 was applied as eluent that this result might be due to the similarity between the polarity of GA and methanol and ethanol and high rinsing strength of them. The addition of acetic acid could cause hindering the formation of hydrogen bonding between GA and sorbent and thus leading to easier removal of GA.

3.3 Optimization of effective parameters via CCD based RSM

Different optimization approaches have been used in the analytical methods to investigate the effect of factors and estimate the optimum levels. One of the common optimization approaches is one-variable-at-a-time in which the effect of one factor is monitored in the fix values of other factors. In spite of simplicity, it suffers from some disadvantages including a large number of experimental runs, labor effort, high consumption of chemicals and solvents, time consuming and high cost. Most importantly, it cannot consider the effect of factors concurrently. To overcome these drawbacks, different optimization methods have been used that amongst CCD based RSM is a versatile and powerful optimization technique which not only can resolve the limitations of the traditional optimization approaches but also can consider the effect of effective factors simultaneously. Most interestingly, it can estimate the true optimum levels (Arabi et al., 2017a, 2017c; Arabi e tal., 2016; Ostovan et al., 2018). So, in this regard, in this study, CCD based RSM was applied to investigate the effect of four effective factors on extraction recovery of GA including sorbent dosage, ultrasonic time, washing solvent volume and also eluent solvent volume in five levels (−2, −1, 0, 1, 2) according to 21 experimental runs, as listed in Table S1. To appraise the consequence of parameters, their interactions, dependability, and accuracy of the model, the experimental results were fitted to the quadratic model and analysis of variance (ANOVA) parameters (Table 1) including p and F values, lack-of-fit test, coefficients of determination (R²), Adjusted R², Adequate Precision, standard deviation and coefficient of variation at a certain confidence level (α = 0.05) were considered. Based on Table 1 and Pareto chart (Fig. S1), the factors with higher values of F, and p values < 0.05, had a significant contribution to extraction recovery. Also, the lack-of-fit value more than 0.05 (0.1105) and the correlation coefficient (R²) more than 0.98 (0.9822) verified the suitability, dependability, and qualification of the suggested model. R² values were calculated according to (2) equation and consider as a measure of the variation around the mean calculated by the model:

The Eq. (7) shows the suitable relation between ER (%), factors, their interactions and also a quadratic model in which the coefficients show the intensity of each factor on ER (%). Also, the positive signs display the increasing of ER (%) while negative signs have the opposite effects and express declining of ER (%). In addition, to assess the concurrent influence of parameters on ER (%), RSM 3D plots were used (Fig. 5). As can be seen from Fig. 5a, the concurrent increase of sorbent dosage and ultrasonic time, enhance ER (%). Indeed, increasing of sorbent dosage, enhances active sites and specific surface area that subsequently enhances adsorption capacity and ER (%). On the other hand, ultrasonic irradiation increases the contact area between sorbent and GA and also enhances the mass transfer of analyte from sample solution to Cu-BDC which enhances ER (%). As orange juice is complicated, it is necessary to remove impurities and complicated matrices. So. The effect of washing solvent volume in the range of 2.0 to 6.0 mL was evaluated and the results showed that the cleanest reaction was achieved when 3.0 mL of hexane was applied as washing. Indeed, 3.0 mL of hexane was enough to remove impurities. On the other hand, the desorption of adsorbed GA is the key step. In this regard, the effect of methanol: acetic acid (90:10 v/v) as eluent solvent was investigated in the range of 2.0–6.0 and results confirmed that 2.0 mL of methanol: acetic acid (90:10 v/v) was enough to desorb GA molecules from sorbent. After RSM evaluation, for simultaneous optimization of the analyzed factors, desirability function (DF) was applied based on the (8) equation (Bandara et al., 2019):

Close

Fig. 5

Response surface showing gallic acid recovery efficiency as a function of two independent variables: (a) Ultrasonic time (min) – Sorbent dosage (mg), (b) Washing solvent volume (mL)- Sorbent dosage (mg), (c) Eluent solvent volume (mL)- Sorbent dosage (mg), (d) Washing solvent volume (mL) - Ultrasonic time (min), (e) Eluent solvent volume (mL) -Ultrasonic time (min) and (f) Eluent solvent volume (mL) - Washing solvent volume (mL) at 25 °C.

Export to PPT

3.4 Assessment of adsorption isotherms for determination of GA by Cu-BDC

One of the main significant features in adsorption study is adsorption isotherm that displays the dependency, propensity and also adsorption mechanism of analyte toward sorbent (Sarma et al., 2019). To evaluate the adsorption isotherms of GA onto Cu-BDC, the quantitative analysis was completed in the concentration range of 5.0–50.0 mg L⁻¹ of GA. In the following, different isotherm models including Langmuir, Freundlich, Temkin, and Dubinin–Radushkevich were applied to investigate the adsorption properties of Cu-BDC. These models and their parameters are presented in Table S2. As can be seen from Table S2, q_e illustrations the adsorbed GA at equilibrium, Q_m and Q_s are the maximum adsorption capacity, C_e is the GA concentration in sample solution at equilibrium and K_L, R_L, α, m, B, K_T and β are the Langmuir, Freundlich, Temkin and Dubinin–Radushkevich constants, respectively. According to Table S2, Q_m for the Langmuir model was attained to be 384.6 mg g⁻¹ which is close to the value gained from experimental results. Based on the results, the Langmuir isotherm model was more suitable than other models due to its higher R² (0.9965) that proved the adsorption of GA onto Cu-BDC can be well described by Langmuir isotherm. Langmuir isotherm illustrates that the adsorption of the analyte onto sorbent is monolayer and the surface of the sorbent is homogenous while the Freundlich model assumes that the adsorption of the analyte is multilayer and on the heterogeneous surface of the sorbent (Araújo et al., 2018). R_L explains the separation factor which takes values between 0 and 1 that represent desire (irreversible) and undesired (reversible) adsorption, respectively. The obtained R_L for GA was between 0.029 and 0.230 which are near to 0 and confirmed a desire (irreversible adsorption) process. The physical and/or chemical nature of adsorption is signified by Dubinin–Radushkevich isotherm model. β can take different values in which the β values between 8 and 16 KJ mol⁻¹ show that the adsorption mechanism is chemical while for β values lower than 8 KJ mol⁻¹, the adsorption process occurs through a physical mechanism (D. Liu et al., 2019; Razmi et al., 2019).

3.5 Examination of adsorption kinetic for adsorption of GA by Cu-BDC

Dynamic of the adsorption process was evaluated in terms of the order of rate constant using different adsorption kinetics models including First-order-kinetic, Pseudo-second-order-kinetic, Intraparticle diffusion, and Elovich. According to Table S3, q_e and q_t are the adsorbed GA at equilibrium and time t, K₁, K₂, K_diff, C, β and α are the constants of models, respectively (Noormohamadi et al., 2018). According to Table S3, the Pseudo-second-order-kinetic model was more appropriate than other models due to its higher R² (0.9921) which confirmed that adsorption kinetic followed by Pseudo-second-order. Moreover, the theoretical qe(cal) (41.15) value was close to the experimental qe(exp) (38.23) value for the Pseudo-second-order-kinetic model which implied that the second-order model is in good agreement with experimental data and can be used to favorably explain the GA adsorption on Cu-BDC.

3.6 Method validation and real sample analysis

Under the optimized conditions, the proposed method was validated. In this regard, the linear range, the limit of detection (LOD), the limit of quantification (LOQ), precision, and accuracy of the presented method was evaluated. The linear range was achieved in the range of 0.5–15.0 mg L⁻¹ for GA. In addition, the LOD and LOQ values were attained to be 0.145 mg L⁻¹ and 0.48 mg L⁻¹ using 3 σ/slope and 10 σ/slope ratios, respectively; in which σ is the standard deviation in the mean value for chromatograms obtained from the blank orange juice sample according to IUPAC recommendation (Committee, 1987). The precision of the method was examined by performing replicate analyses of blank orange juice samples at five different concentration levels (0.5, 1.0, 5.0, 10.0 and 15.0 mg L⁻¹) on the same day and in the five successive days expressed as relative standard deviation (RSD). Based on Table 2, The RSD values were lower than 6.5% for all intra and inter-day precision while highest extraction recoveries (higher than 86.0%) was achieved, confirmed qualified reproducibility and high accuracy of the proposed method. Different orange juice samples were applied to approve the analytical applicability of the proposed method using the standard addition method. In order to investigate the effect of sample preparation and enrichment of ER, the orange juice sample was injected to HPLC directly, as can be seen, the peak of GA wasn’t detected that is due to the complexity of the orange juice samples and low concentration of GA (Fig. 6a). After sample preparation using Cu-BDC, owing to the high potential of Cu-BDC-DSPE for enriching GA, trace quantities of GA has a remarkable signal (Fig. 6b). In the following, the accuracy of the model was investigated using a recovery test (Table 3). The ER% were found to be in the range of 98.0–101.2% while RSD values were lower than 6.0% for five replicate continuous injections. These results indicate that the proposed Cu-BDC-DMSPE -HPLC/UV method was mighty for the clean extraction of GA at trace level from intricate orange juice samples with supreme precision and accuracy.

Close

Fig. 6

Typical chromatograms of (a) direct injection of orange juice sample and (b) spiked orange juice sample after Cu-BDC -DSPE procedure.

Export to PPT

3.7 Reusability of the prepared Cu-BDC

The practical aspect of prepared Cu-BDCwas evaluated using a reusability test in which, several adsorption-desorption recycles of Cu-BDCwere completed and results represented that the prepared Cu-BDCis applicable for at least six times (Fig. S4.). Consequently, the proposed Cu-BDCis not only easy of synthesis, low-cost, and simple but also it is a good candidate for practical application for extraction and determination of GA from complicated orange juice samples.

3.8 Method comparison

Our presented method was compared with some other reports in terms of adsorption capacity, adsorption equilibrium time, reusability and figure of merits. The prepared Cu-BDC has high adsorption capacity (300.0 mg g⁻¹) and the short adsorption equilibrium time (5 min) and outstanding reusability (6 times). From the view point of the analytical aspect, the linear range and LOD of the presented method are approximately the same or better than some other methods. Molecularly imprinted polymers (MIPs) have these ability to distinguish target analyte from other compounds. So, they have excellent figures of merit. In spite of their selectivity ability, MIPs don’t have high adsorption capacity and have high adsorption equilibrium time while our prepared Cu-BDC has high adsorption capacity and low adsorption equilibrium time. These facts signify the promising future of our prepared Cu-BDC for clean-up, pre-concentration, and determination of GA from complicated orange juice samples. (Arabi et al., 2017b; Denderz and Lehotay, 2014; Song et al., 2012) (Table 4).