Efficient extraction of patuletin and quercetagetin from their plant extract using molecularly imprinted polymer based syringe: A green technology

2.1. Reagents

The procurement details of different reagents used in this study are as under: 2-VP (97%), methacrylamide (MAAm) (98%), methacrylic acid (MAA) (99%), acrylic acid (AA) (99%), formic acid, styrene (ST), 4-vinyl 1,3 dioxolane 2 one (4-VD), 4-vinyl pyridine (4-VP), N-isopropylacrylamide (NIPAAM), MG, ethanol (EtOH), and HCl from Sigma Aldrich (USA); acrylamide (AAm) (98%) from DUKSAN (Korea); itanoic acid (IA) (99%), 2-dimethyl amino ethyl methacrylate (2-DAEM) (98%) technical grade, 2 2-azobisisobutyronitrile (AIBN), and from DAEJUNG (Korea); acetic acid from Merck (India); MeOH from Fisher Scientific (UK); acetone and acetonitrile (ACN) from Honeywell (Germany), and EGDMA (97%) from Tokyo chemical industry (Japan). PL, gallic acid (GA), cinnamic acid (CA), QTG, RA, rutin (RU), and ferulic acid (FA) were obtained from the compound bank at the International Center for Chemical and Biological Sciences (ICCBS), University of Karachi. Deionized water produced by the Millipore Milli-Q plus water purification system, having conductivity = 0.055 µS/cm was used throughout. AIBN was used after recrystallization from MeOH. All other compounds and reagents were used without additional purification.

2.2 Synthesis of PL-MIP and corresponding NIP

Different NIPs were first synthesized using 11 functional monomers, namely, MAA, AA, MAAm, IA, acrylamide, ST, 4-VD, 4-VP, 2-DAEM, NIPAAM, and 2-VP with EGDMA as a crosslinker, AIBN as an initiator, and MeOH as a porogenic solvent. After mixing in the order elaborated below (without template) and purging with N₂ gas to eliminate oxygen gas, the reaction vessel was sonicated for 10 min and finally heated to 80°C for 24 hrs.

For the PL-MIP, a composition ratio of 1:6:80 was used for the template, FM, and crosslinker, respectively. In a reaction tube, template (PL; 0.015 mmol, 50 mg) and FM (2-VP; 0.09 mmol, 10 µL) were dissolved in 0.3 mL of MeOH under gentle stirring. After 2 h, EGDMA (1.2 mmol, 300 µL) and AIBN (10 mg) were added to the mixture. The solution was purged with nitrogen gas while being sonicated for 10 min and then heated at 80°C for 24 hrs to initiate polymerization.

The resultant polymer was a solid bead, which was grounded and sieved through a 115 µm mesh size (130 mesh) to ensure uniform particle size. To remove the template before the final application, a Soxhlet extraction was performed using a stepwise washing protocol. First, the polymer was extracted with a 1:1 mixture of MeOH and DI water containing 1% acetic acid for 48 h. This was followed by washing with pure MeOH for an additional 24 h, with the eluent monitored by UV-Vis spectroscopy to confirm the complete removal of PL. Washing was continued until no trace of PL was detected. The counterpart NIP was prepared using the same procedure but without the addition of the template. Both polymers were dried in an oven (MMM Medcenter Einrichtungen GmbH oven) at 60°C for 8 h and stored at room temperature for further use. The schematic protocol of MIP synthesis has been depicted in Figure 1-top.

Close

Figure 1.

Synthesis of PL-MIP (Top), construction of PL-MISPE frit in four simple steps (right bottom), and extraction of PL with PL-MISPE frit in four steps (left bottom).

Export to PPT

2.3. Characterization of PL-MIP and NIP

Polymer characterization, in the context of the determination of involved functional groups in the process, was done by FT-IR spectroscopy (Bruker Vector 22, Germany). The instrument was operated in the mid-infrared region, scanning wavelengths from 400 to 4000 cm⁻¹.

The morphological features of the polymer particles were examined through SEM using the Apreo 2 C LoVac model (Thermo Fisher Scientific, USA). Before imaging, the polymer samples were coated with a thin layer of gold using a sputter coater (JEOL SC7620-Quorum Technologies), with the applied coating measuring 153 Å in thickness.

Specific pore size distribution and surface area were determined through nitrogen gas adsorption experiments. BET and Barrett-Joyner-Halenda (BJH) analyses were conducted using an automated gas sorption analyzer (Quantachrome Autosorb-iQ TPX, USA).

Thermogravimetric analysis (TGA) and differential scanning calorimetry (DSC) were performed simultaneously to assess the thermal stability of the PL-MIP and NIP samples. For these analyses, 4.0 mg of each sample was analyzed using an SDT Q600 (TA Instruments, USA). The thermal behavior was monitored as the temperature increased from 25°C to 1000°C at a rate of 10°C per min under a constant nitrogen atmosphere with a flow rate of 50 mL/min.

2.4. Binding experiments

The standard procedure involved preparing solutions of 0.5 mg/mL of PL, analog compound, binary mixture of analog compound and PL, a mixture of all the analogous compounds with PL, and a real sample in water: MeOH (1:1, v/v) mixture. Each experiment utilized 2.5 mg of PL-MIP/NIP polymers, suspended in 0.5 mL of the standard solution, and stirred at 500 rpm for 30 min. Following adsorption, the supernatants were filtered using syringe filters (0.45 μm), and analysis was carried out using the UV–Vis spectrophotometer (Shimadzu UV-1800, Japan) for absorption at 365 nm to determine the remaining PL. An ultra-performance liquid chromatography (UPLC) method was used to evaluate the competitive adsorption capacity of the PL-MIP when used as a mixture. The solvent of 0.3 mL supernatant was dried in a freeze dryer for 6 h, and the residue was again prepared in 0.3 mL of high performance liquid choromatography garde (HPLC)-grade MeOH and analyzed via UPLC.

The binding extent, B_EXT (mg/g) of PL-MIP and NIP was calculated using the Eq. (1);

Where Co and Ce are the PL equilibrium concentrations in mg/mL at the initial and final stages, respectively, W is the weight of MIP/NIP in grams, and V is the volume of the PL solution in mL. Different experimental parameters of the adsorption experiment, namely contact time, PL concentration, and pH, were optimized.

2.4.1. Selectivity study

The specificity of PL-MIPs towards PL was examined by testing various analogous compounds such as QTG, FA (CA), RU, MG, GA, and RA. One mM individual solution of these compounds was utilized to assess the binding specificity of PL-MIP and NIP. Subsequently, binary solutions were prepared, each containing 1.0 mM of PL alongside one of the analogous compounds, and the same adsorption procedure was followed as outlined in Section 2.4. Lastly, a complex mixture containing PL along with all the tested analogs at a total concentration of 1.0 mM was analyzed to further evaluate the binding performance. The PL selectivity of PL-MIP was determined using the selectivity factor (β) and imprinting factor (α) as defined by Eqs. (2) and (3)[46,65,66]:

In these equations, B_PL-MIP and B_NIP are binding extent (mmol/g) of PL-MIP and NIP, while B_PL and B_ana are binding amounts (mM) of PL and analogous using PL-MIP.

2.4.2. Desorption and reusability

The recovery parameters for the extracted PL by PL-MIP were fine-tuned, focusing on the elution solvent, its quantity, and contact time for complete desorption. Various solvents, including deionized water, EtOH, MeOH, ACN, acetone, as well as different combinations of EtOH and MeOH with DI water with or without acetic acid, were employed for desorbing the extracted PL on PL-MIP. The desorption process involved stirring the mixture at 500 rpm for 5 min, followed by filtration. A 0.3 mL portion of the eluate was collected and processed for freeze-drying and dissolution in 0.3 mL MeOH for UPLC analysis. The optimal solvent volume and contact time were further optimized using the most effective desorption solvent. Additionally, the reusability of the PL-MIP was evaluated over five consecutive adsorption-desorption cycles.

2.4.3. UPLC protocols

The UPLC-UV analysis was conducted using an Agilent Technologies Infinity 1260 series system, which included a column oven, auto-sampler, binary pump, and UV detector. A Hypersil C-18 ODS column (5 µm, 4.6 mm × 150 mm) from Thermo Scientific was utilized. A gradient methodology was used: solvent A (water with 0.1% formic acid) and solvent B (MeOH containing 0.1% formic acid) at a flow rate of 0.8 mL/min for a total run-time of 12 min. The gradient profile was 0–1 min 40% B, 1→8 min 40→100% B, 8–10 min 100% B, 10→11 min 100→40% B, and 11-12 min 40% B. The injection volume was 5 μL, and the wavelength of the UV detector was set at 254 nm for PT, QTG, and RU, 330 nm for RA and FA, 271 nm for MG and CA, and 265 nm for GA. Method validation of the developed UPLC method in the context of intra-day and inter-day precision, the limit of detection (LoD), the limit of quantification (LoQ), correlation coefficient (R²), linearity, and accuracy have been elaborated in Section S1.

2.4.4. Preparation of MI-SPE frits

Following the optimization of adsorption and recovery protocols, MISPE frits were developed utilizing a syringe, membrane, and PL-MIP. In this process, a 4 mm square section of 0.45 μm membrane was fixed between the barrel and needle hub of a 3.0 mL plastic syringe. Subsequently, the barrel was loaded with 2.5 mg of PL-MIP by removing the plunger (see Figure 1-right bottom) [44].

The setup enables the smart MI-SPE to extract PL from samples effectively. The conditioning of the PL-MIP in the syringe involves a 30 s treatment with 50% MeOH in DI water (pH adjusted to 4). The sample was then drawn through the needle into the syringe using the same solvent, shaken vigorously for 120 s, and disposed of into the sample container. Following this, 1.0 mL DI water (washing solvent) was passed through the needle into the syringe, shaken for 30 s, and disposed of into a waste container. Finally, 1.0 mL EtOH (elution solvent) was introduced, shaken for 60 s, and injected into the PL collecting container (see Figure 1-left bottom). The MI-SPE was again ready for multiple extractions following the same procedure.

2.4.5. Extraction of PL and QTG from tagetes patula and tagetes erecta methanolic extract using smart MI-SPE frits

Freshly obtained 5.2 kg uncrushed mix color T. patula flowers were dried in the shade and extracted thrice with petroleum ether (PE) at room temperature. The PE extract was vacuum evaporated, resulting in a gummy residue (JFP). The remainder was extracted with MeOH and evaporated under reduced pressure to get dried residue JFM (final extract) [18]. For QTG, a total of 3.63 kg of fresh, uncrushed yellow T. erecta flowers were shade-dried and subjected to three rounds of MeOH extraction at room temperature. The resulting extract was filtered and concentrated under reduced pressure, yielding 218 g of dry residue. Thereafter, standard solutions of both dried extracts were prepared in a mixture (1:1 MeOH and water maintained at pH 4) and treated as T. patula and T. erecta real samples. These real samples were spiked with the respective analyte, either PL or QTG, for evaluation of recoveries, Table 1. PL-MISPE frit containing 2.5 mg of PL-MIP was applied to extract PL and QTG from the prepared real samples following the procedure elaborated in Section 2.4.4. The extraction process spanned 4 min, including sequential steps of conditioning, extraction, washing, and elution. After elution, 0.3 mL of the solvent was freeze-dried and re-dissolved in 0.3 mL of MeOH for subsequent UPLC analysis.

3.1. Synthesis and characterization PL-MIP

A judicious selection of functional monomer/s, crosslinkers, and porogens was extremely important for the optimal performance of any MIP for the target molecule. In this context, we employed a preselection procedure that entails the synthesis of numerous NIPs using different functional monomers, namely, MAAm, MAA, AA, itaconic acid (IA), acrylamide (AAm), ST, 4-VD, 4-VP, 2 dimethyl amino ethyl methacrylate (2-DAEM), NIPAAM, and 2-VP. EDGMA was used as a crosslinker, and MeOH as a porogen. The binding extent of different NIPs for PL was evaluated for the appropriate selection of the functional monomer, Supplementary Table. NIPs synthesized with 4-VD, 4-VP, 2-DAEM, NIPAAM, and 2-VP (serial #7-11) had a greater binding extent for PL compared to the NIPs synthesized by using other functional monomers. These functional monomers were subsequently chosen for the synthesis of MIPs and the corresponding NIPs for PL. Among the selected monomers (serial# 12-16), the maximum binding extent and imprinting factor were obtained for MIP synthesized using 2-VP as a functional monomer. As a final step, the amount of crosslinker (EDGMA) was varied to optimize the imprinting factor further (Serial# 17-18). A higher concentration of EGDMA resulted in a low binding extent for both MIP and NP (MIP: 6.7 mg/g and NIP: 1.9 mg/g). On the contrary, a low concentration of EDGMA resulted in a higher binding extent of NIP, which consequently rendered a low imprinting factor. Hence, the composition PL: 2-VP: EGDMA = 1:6:80 was finalized. This composition rendered an imprinting factor of 3.7 (serial #16).

FTIR spectra of PL (template), the 2-VP (functional monomer), the EGDMA (crosslinker), NIP, and the PL-MIP have been displayed in Figure 2. Several similar bands across these spectra reflect the presence of common functional groups among PL, 2-VP, and EGDMA. For instance, the peaks at 2960 cm⁻¹ and 1451 cm⁻¹ were attributed to the stretching and bending of sp³ C-H bonds in methylene groups. The peaks at 1716 cm⁻¹ and 1245 cm⁻¹ were associated with stretching vibrations of ester carbonyl (-C=O) and ether groups (-C-O-), respectively. In the PL spectrum, additional absorption bands at 1619 cm⁻¹ and 1245 cm⁻¹ correspond to -C=C- (aromatic ring) and -C-O- (polyols), respectively, with the characteristic broadband in a region of 3200–3500 cm⁻¹ for O–H stretching [67]. The 2-VP spectrum had peaks at 3000, 1435, and 1470 cm⁻¹, attributed to aromatic -C-H stretching and -C=C vibrations, respectively. It also contained distinctive peaks at 1585 cm⁻¹ and 1595 cm⁻¹, which indicate the -C=N stretching in the pyridine ring. Moreover, a marker peak of 2-VP at 745 cm^-1 corresponds to C–H out-of-plane stretching [68-70]. The EGDMA spectrum reveals a characteristic band at 1150 cm⁻¹, attributed to the uncommon O-CH₂-O group [46]. In the unwashed PL-MIP, the presence of characteristic peaks of PL, 2-VP, and EGDMA confirmed their incorporation into the polymer matrix. The significant reduction in the intensity of the O-H band in the unwashed PL-MIP was likely due to secondary interactions with the pyridine group of 2-VP. Additionally, the disappearance of the O-H band and suppression of aromatic stretching vibrations in the washed PL-MIP indicated the successful removal of the PL molecule. Moreover, the aromatic CH stretching at 3000 cm^-1 of PL and 2-VP also appeared in PL-MIP washed, PL-MIP unwashed, and NIP. Similar chemical composition of the washed PL-MIP and NIP was confirmed by their identical spectrum.

Close

Figure 2.

FTIR spectra of PL, 2-VP, EGDMA, PL-MIP-unwashed, PL-MIP-washed, and corresponding NIP.

Export to PPT

SEM imaging at different magnifications reveals distinct differences in the surface morphology, roughness, and porosity between PL-MIP and NIP despite their similar chemical nature, Figure 3. PL-MIPs exhibited a uniform texture and notably higher porosity compared to NIP, which displayed a smooth surface even at higher magnification. This stark contrast in morphology indicated that PL-MIPs possessed a greater surface area and porous structure that facilitated more defined binding sites. These evident morphological disparities between the two polymers contribute to their effectiveness in achieving binding and selectivity targets for the template.

Close

Figure 3.

SEM images of PL-MIP and the corresponding NIP at different magnifications.

Export to PPT

Additionally, investigation of the pore volume, surface area, and pore size of both PL-MIP and the corresponding NIP was also carried out by performing N₂ adsorption-desorption experiments, Figure 4. The BET method was employed to analyze surface area, while pore volume and pore diameter were evaluated using the BJHs (Figure 4c and 4d) method. The analysis revealed substantial differences in porosity, surface area, and pore volume between PL-MIP and NIP. The adsorption hysteresis loop of PL-MIP indicates its mesoporous nature, with a high surface area of 124 m²/g and a pore volume of 0.29 cm³/g, highlighting its multilayer adsorption behavior. On the other hand, NIP exhibited a much lower surface area (14.8 m²/g) and pore volume (0.04 cm³/g), consistent with a nonporous structure. The enhanced surface area and porosity of PL-MIP are attributed to the imprinting effect. The BJH plot further confirmed that PL-MIP possesses both microporous and mesoporous structures, underscoring the role of the pre-polymerization complex between the template and functional monomer in determining pore size. The highly porous structure and increased surface area contribute to the formation of highly specific binding sites, improving the selectivity and binding capacity of PL-MIP.

Close

Figure 4.

BET isotherms of (a) PL-MIP, and (b) NIP; BJH isotherms of (c) PL-MIP, and (d) NIP.

Export to PPT

Thermal stability, a critical property for assessing the structural integrity of the synthesized polymers (PL-MIP and NIP), endorses their similar nature [46,71,72]. TGA analysis of washed PL-MIP and NIP has been illustrated in Figure S1, depicting derivative mass loss as a function of temperature and weight loss percentage. Both PL-MIP and NIP exhibited initial mass loss around 30°C, likely due to the evaporation of trapped MeOH within the polymer matrices. Subsequently, a second significant weight loss occurred at approximately 283°C, with NIP showing 3.1% mass loss and PL-MIP 2.9%, indicating the degradation of unreacted components [73]. The thermal decomposition patterns of both polymers were similar, as evidenced by derivative weight loss curves at 350°C and 440°C, confirming their comparable chemical composition [46]. Slightly faster degradation of PL-MIP, possibly due to its porous structure, can be attributed to the impact of heat conduction. The smoother and denser structure of the NIP facilitates a more uniform temperature transition across the polymer, leading to its earlier degradation [74,75].

3.2. Extraction performance of PL-MIP

The key advantage of MIP technology lies in its ability to selectively extract target analytes beyond only polarity differences [42,76]. The efficiency of extraction can be greatly affected by various experimental parameters, which must be carefully optimized to unlock the material’s full potential.

3.2.1. Hydrogen ion concentration

The occurrence of various functional groups on both the analyte and the polymer matrix significantly affects the binding capacity, particularly in response to changes in the pH of the medium. Figure 5(a) demonstrates the effect of pH on the binding performance of PL-MIP, while the binding ability of the corresponding NIP remains relatively stable across different pH levels. While the binding extent of NIP is not affected to a large extent as a response to the medium’s pH, the maximum binding extent for PL-MIP is achieved at pH 4. A lower binding extent at pH 2 and 3 can be explained by the fact that the electron pair of pyridine is in a sp²-hybridized orbital, which is more strongly bound to the other atoms and not easily available for hydrogen bonding with PL at high H⁺ ion concentrations. However, at slightly low H⁺ concentration, the tertiary amine of 2-VP can easily form secondary bonds with PL. At pH above 4, the binding extent again decreases due to an inappropriate environment for PL being an acidic compound, the aromatic OH loses its H, and the negative charge gets delocalized over aromatic rings, resulting in a decrease in the surface activity of PL.

Close

Figure 5.

Optimization of key experimental parameters affecting the binding performance of PL-MIP and NIP: (a) pH using 0.5 mL of 0.5 mg/mL PL solution in a 1:1 MeOH/DI water mixture, with a contact time of 30 min; (b) PL concentration using 0.5 mL solutions in MeOH/DI water (1:1) at pH 4, with a 30-min contact time; (c) Contact time using 0.5.mL of 0.5 mg/mL PL solution (in 1:1 MeOH/DI water) at pH 4. In all cases, 2.5 mg of PL-MIP/NIP was used, with stirring speed maintained at 500 rpm.

Export to PPT

3.2.2. Initial concentration of PL

The binding extent is closely linked to the relative concentrations of the adsorbent and the analyte. To assess the maximum binding extent of PL-MIP, the concentration of the PL solution was varied while keeping the amount of PL-MIP/NIP constant. Initially, the adsorption of PL increased for both PL-MIP and NIP, indicating that functional groups on the polymers were actively engaging with the analyte. However, as PL concentrations exceeded 0.5 mg/mL, the adsorption rate decreased, making a plateau, Figure 5(b). Consequently, the binding extents were determined to be 38.0 mg/g for PL-MIP and 7.9 mg/g for NIP. The resulting imprinting factor (IF), defined as the ratio of PL-MIP to NIP binding, was 4.8. This high IF highlights the superior selectivity of PL-MIP, attributed to the formation of specific recognition cavities that align with the molecular structure of the PL. The high-affinity binding sites were absent in NIP, resulting in its lower adsorption performance [77,78].

3.2.3. Contact time

Optimizing contact time is also crucial for achieving optimal binding performance. Both PL-MIP and NIP showed a rapid increase in binding in the first 15 min. However, the rate of binding gradually decreased, and equilibrium was reached at 30 min, Figure 5(c). This behavior reflects the initial ease with which PL interacts with accessible binding sites on the polymer surface. Later, the sites inside the cross-linked polymeric matrix become slowly accessible [46,61]. NIP, lacking high-affinity recognition sites, demonstrated a lower binding extent compared to PL-MIP. A slight decline in binding after 30 min may be attributed to the detachment of PL from weakly bound low-affinity sites.

3.3. Binding isotherms

The distribution of PL between the liquid phase and the polymer matrix, along with the attainment of maximum uptake of PL, can be elucidated using the Langmuir and Freundlich isotherm models, as represented in Eqs. (4 and 5) [46,65,66],

In these equations, C_e (mg/L) represents PL concentration at equilibrium, B_e (mg/g) is the binding extent at specific applied concentration, the maximum binding capacity is B_max (mg/g), and K_L (L.mg^-1) is the Langmuir constant, reflecting the theoretical maximum monolayer capacity. The constants n and K_f (mg^-(n+1)/n.L^1/n.g^-1) pertain to the Freundlich model.

The Langmuir plot allows for the determination of K_L and B_max, from its slope and intercept, respectively. Meanwhile, the Freundlich constants, n and K_f, can be derived from the slope and intercept of the Freundlich plot, which indicates the adsorption favorability and capacity. Figure 6(a) and (b) illustrate the linearized forms of both isotherm models, with the calculated parameters detailed in Table S1 and S2. Langmuir and Freundlich models are triggered by their behavior using the LF model, a flexible expression suitable for various adsorption scenarios. It is important to note that the Freundlich isotherm data may be less reliable for heterogeneous systems like PL-MIPs, particularly at elevated adsorbate concentrations. MIPs represent unique heterogeneous adsorption systems that can effectively capture PL across a broad concentration range, making the LF isotherm a better fit for characterizing their adsorption behavior [79,80].

Close

Figure 6.

(a) Langmuir isotherm model and (b) Freundlich isotherm model for MIP and NIP; (c) Scatchard plot of PL-MIP and (d) NIP.

Export to PPT

Assessment of the surface heterogeneity by LF isotherm (Eq. 6) of the synthesized PL-MIP and NIP was conducted, which involves three fitting coefficients: m, a, and Nt [46,60].

In this case, N_t represents the number of binding sites, while ‘a’ signifies the median binding affinity with K_o = a^1/m. The heterogeneity index ‘m’ ranges from 0 to 1, with higher values indicating greater homogeneity and lower values suggesting heterogeneity due to the presence of distinctive imprinted cavities.

In order to derive the fitting parameters from experimental data for both PL-MIP and NIP, the DDsolver program in Microsoft Excel was employed. PL-MIP exhibited a greater concentration of binding locations per gram (N_t = 44.7 mg/g) than NIP (N_t = 9.50 mg/g), as detailed in Table S3. Additionally, PL-MIP had greater heterogeneity (m = 0.7) compared to NIP (m = 0.8), attributed to the abundance of specifically tailored shapes inside the PL-MIP matrices. The K_o value corroborated the accuracy of the estimated values.

Scatchard analysis renders differentiation between specific and non-specific binding interactions. By plotting the binding extent per free PL against the free PL concentration, valuable insights into the adsorption events between the target and receptor can be gained. Figures 6(c) and (d) show plots of binding extent /available PL in L/g as the X-axis and binding extent in mg/g as the Y-axis for the polymers, respectively, using the Eq. (7) below [46].

Here, B denotes the binding extent (mg/g) at a specific PL concentration C (mg/L), and B_max represents the maximum binding extent (mg/g) while K_d corresponds to the equilibrium dissociation constant (mg/L). The slope K_d reflects the nature and extent of binding, whereas the intercept (B_max) indicates the number of available binding sites.

The Scatchard plot for PL-MIP reveals two distinct linear regions, signifying the existence of both high-affinity and low-affinity binding sites. However, NIP displays only low-affinity binding sites [81]. The values for K_d and B_max associated with the high-affinity sites are 11.34 and 11 mg/g, respectively, while the low-affinity sites have K_d and B_max values of 208 and 38 mg/g, respectively. Adsorption at low-affinity binding sites may correspond to specific binding sites for PL. The big structure of PL requires slow adjustment to the cavities of PL-MIP. Furthermore, the entry in deeper cavities may result in slow kinetics. On the other hand, high-affinity sites with fast adsorption may correlate with non-specific binding events with at least one OH group that allows early adsorption. The presence of only non-specific binding sites in NIP is confirmed by its K_d (417) and B_max (8) values [62,82]. Thus, PL-MIP has a high binding efficiency for PL, which makes it a promising choice as an SPE adsorbent for PL.

3.4. Binding kinetics

The binding performance of PL-MIP and NIP over varying contact times is presented in Figure 5(c). The graph shows an initial rapid increase in binding followed by a gradual decline, which can be attributed to the quick occupation of the surface binding sites [46,61]. After 30 min, no additional binding of PL is observed in either case, which indicates their saturation. To achieve the best-fitted adsorption behavior towards PL-MIP and NIP, pseudo-first-order (Eq. 8) and pseudo-second-order (Eq. 9) kinetic models were employed to gain a comprehensive understanding of the regulating mechanism. The pseudo-second-order model suggests that chemical interactions occur between the adsorbate and the polymer, while the pseudo-first-order model implies that the adsorption depends on the availability of binding sites and adsorbate concentration [83].

In these equations, B_e and B_t (mg/g) represent the binding capacity at equilibrium and at time t, respectively. The rate constants for the pseudo-first-order and pseudo-second-order models are denoted as k₁ (1/min) and k₂ (g.mg^-1.min^-1), respectively. As shown in Figure 7, the slope and intercept of the linear plots establish the values of k₁ and k₂ through Log(Be-Bt) vs t (Eq. 8) and t/Bt vs t (Eq. 9), respectively. The higher R² value for the pseudo-second-order model for PL-MIP suggests the chemisorption by the cavities through π-π interactions and hydrogen bonding. The slope of the linear equation yields binding extent, 38.9 mg/g for PL-MIP and 8.83 mg/g for NIP, which are consistent with experimental observations [84]. The rate constant (k₂) for PL-MIP, derived from the intercept of the linear regression line, is 9.54 х 10^-3 g.mg^-1.min^-1, which is lower in comparison to that of NIP, 29.7 х 10^-3 g.mg^-1.min^-1, Table S4. These findings suggest a slower diffusion of PL into the specifically tailored cavities of PL-MIP than the non-specific binding sites on both PL-MIP and NIP.

Close

Figure 7.

Pseudo-first-order (a) and pseudo-second-order, (b) kinetic model for PL-MIP and NIP.

Export to PPT

3.5. Binding selectivity

The primary advantage of MIP architecture over other separation approaches is its high selectivity for template shapes relative to NIPs. To assess the selectivity of PL-MIP for PL, its performance against structurally similar compounds such as GA, MG, CA, FA, QTG, RU, and RA, was evaluated. The structures of PL and these analogous compounds have been depicted in Figure S2.

Initially, the selectivity of PL-MIP relative to NIP for each of the above-mentioned compounds is individually evaluated. The selectivity was assessed in terms of selectivity factor, imprinting factor, and binding extent, Figure 8. PL-MIP adsorbed a greater amount of QTG and PL compared to other analogous compounds. The higher adsorption of QTG can be attributed to its very similar structure to that of PL, with only a difference of an extra hydroxyl group in place of the methoxy group (-OCH₃). The additional OH group makes QTG compact, charged, and more accessible for the binding site as compared to PL. Thus, QTG had a 1.25-fold greater binding extent than PL on PL-MIP. Hence, the PL-MIP can also be employed for elective extraction of QTG following the dummy template protocol. A key aspect of this approach lies in evaluating the selectivity of PL-MIP compared to NIP. A significant difference in the binding capacity between PL-MIP and NIP for PL demonstrates the high selectivity of the imprinted polymer for PL. However, for other tested compounds, PL-MIP did not exhibit enhanced selectivity over NIP except for QTG. In some instances, NIP even adsorbs more than PL-MIP, likely due to the availability of unbound functional groups in its polymer matrix. However, it is pertinent to mention that despite the high binding extent for QTG, the imprinting factor (3.2) is lower than PL (4.8). The imprinting factor of PL-MIP is high for PL and QTG, while it is low or even negative for other structurally similar compounds (Figure 8a). The UV-Vis spectra of varying concentrations of PL have been provided in Figure S3(a), and the calibration curve correlating PL concentration with absorbance at 265 nm has been demonstrated in Figure S3(b). This calibration curve was used to quantify the adsorption of PL on PL-MIP and NIP.

Close

Figure 8.

(a) PL-MIP/NIP binding extent profile using pure analogous compound with imprinting factor values, (b) PL-MIP binding extent profile using binary solutions of PL with one analogous compound along selectivity factor values; experimental conditions: 0.5 mL of 1.0 mmol/L solution (MeOH/DI water, 1:1), contact time: 30 min, pH: 4, stirring speed: 500 rpm, and 2.5 mg of PL-MIP/NIP.

Export to PPT

The selectivity of PL-MIP for PL was also tested in the presence of competing compounds in binary mixtures. The competitive compounds had little effect on the binding capacity of PL-MIP for PL, except QTG and RU, Figure 8(b). QTG is the most adsorbed compound, even in the presence of PL, being structurally more adjustable, polar, and compact than PL. On the other hand, RU has a comparatively bigger structure that may block the binding sites and may also weaken the affinity of PL for binding sites by forming the PL-RU complex. Additionally, a mixture containing all competitive compounds and PL was processed to assess the selectivity of PL-MIP and NIP under complex sample conditions. In this case, a UPLC approach was devised to achieve baseline separation of all mixed compounds, which are GA, MG, CA, FA, (QTG), RU, RA, and PL presents in equimolar concentration, Figure 9-top. The same composition is processed with the adsorption protocol on both PL-MIP and NIP, followed by the collection of supernatants and their analysis with UPLC under the same conditions. The elugrams of supernatants of PL-MIP and NIP mirror the selectivity study using individual compounds and their binary mixtures. A noticeable decrease in the peaks of PL and QTG in the PL-MIP supernatant compared to other compounds reflects the PL-MIP’s selective adsorption. In contrast, the NIP supernatant displayed similar peak intensities for all components, indicating a lack of selectivity. Further analysis of the desorption solvents confirmed that PL-MIP predominantly retained PL and QTG, while NIP adsorbed all the components of the mixture without any selectivity. Concentration-dependent chromatograms for PL have been provided in Figure S4(a), while the calibration curve plotting peak area versus PL concentration has been depicted in Figure S4(b).

Close

Figure 9.

UPLC chromatograms of a mixture of PL with galic acid (GA), MG, FA, quercetagetin (QTG), RU, and RA, and CA prepared in MeOH: DI water (1:1), PL-MIP and NIP supernatants, and PL-MIP and NIP extract, performed at optimized conditions (pH =4, t = 30 min, stirring speed = 500 rpm, solution volume = 0.5 mL, and PL-MIP amount = 2.5mg); UPLC protocols: column: Hypersil C-18 ODS column (5 μm, 4.6 mm × 150 mm), solvent A (aqueous formic acid 0.1%) and solvent B (formic acid, 0.1% in MeOH) at a flow rate of 0.8 mL/min for a total run-time of 12 min. The gradient profile was 0–1 min 40% B, 1→8 min 40→100% B, 8–10 min 100% B, 10→11 min 100→40% B, and 11-12 min 40% B. The injection volume was 5 μL and the wavelength of UV detector was set at 254 nm for PT, QTG and RU, 330 nm for RA and FA, 271 nm for MG and CA, and 265 nm for GA.

Export to PPT

According to the calibration curve, the initial concentration of PL in the mixture was 0.14 mg/mL. After adsorption, the supernatant of PL-MIP contained 0.035 mg/mL, whereas the NIP supernatant held 0.112 mg/mL. Analysis of the extraction solvent confirmed that PL-MIP extracted 0.1 mg/mL of PL, compared to just 0.03 mg/mL in NIP, resulting in a recovery of 72% for PL. On the other hand, the QTG amount in the mixture is found to be 0.14 mg/mL, calculated through the calibration curve. The supernatant of PL-MIP carried 0.045 mg/mL of QTG, while the NIP supernatant contained 0.1 mg/mL. Analysis of the extraction solvent indicated higher QTG presence in PL-MIP (0.09 mg/mL) compared to NIP (0.045 mg/mL), QTG recovery for PL-MIP was 64%.

3.6. Desorption and reusability

An effective desorption strategy is essential for the success of SPE methods in order to ensure the recovery of the target compound in a sufficiently pure form. Optimizing desorption involves fine-tuning of the washing and elution solvents, solvent volume, and contact time [65]. In this regard, various solvents and combinations were tested to maximize PL and QTG recovery, with EtOH proving to be the most efficient, achieving over 90% recovery compared to all other solvent combinations, Figure 10(a). Moreover, selecting a washing solvent that removes loosely bound impurities without affecting the target compound is equally important—DI water was identified as the most suitable option. Consequently, DI water is selected as the washing solvent, while EtOH is chosen as the desorption solvent. The next important question was the optimal amount of elution solvent required for maximum recovery of the extracted PL. The amount of retrieved PL from PL-MIP initially increased with the amount of the elution solvent, reaching an optimum while using 2 mL, which extracted more than 90% of the trapped PL (Figure 10b). On a similar note, the contact time of the elution solvent with the PL-MIP containing extracted PL is a key parameter. EtOH is an efficient solvent, extracting the maximum of the trapped PL in the first 5 min. Further increases in contact time did not result in a significant gain in extraction efficiency; instead, extraction efficiency was slightly reduced (Figure 10c). As a result, the optimal desorption protocol is 2 mL of EtOH with a 5-min contact time. The same desorption protocol was found to be efficient for QTG. MIPs being cross-linked polymers with stable and well-defined morphology retain their shape that demonstrates their high potential for multiple-time use. The reusability potential of PL-MIP over five adsorption-desorption cycles using the optimized protocols is confirmed by only a slight reduction in the performance, Figure 11.

Close

Figure 10.

Desorption protocols for recovery of adsorbed PL from PL-MIP (a) the solvent, (b) volume of the EtOH, (c) contact time of EtOH.

Export to PPT

Close

Figure 11.

PL recovery in repeated five cycles of adsorption-washing-desorption using optimized protocols.

Export to PPT

3.7. Application to the real sample of T. patula and T. erecta (methanolic extract) using PL-MIP containing syringe

Finally, a syringe is prepared using PL-MIP following the procedure elaborated in the experimental section and is applied for the extraction of PL and QTG from real plant extracts. Five milligrams of dried methanolic extract of Tagetes patula was diluted in 10 mL of 1:1 MeOH and pre-adjusted pH 4 deionized water. Initially, 6.5 µg of PL is found in 0.5 mL of solution. Moreover, specific amounts of PL were added to the extract to evaluate the percent recovery of PL. The same procedure was adopted for Tagetes erecta with the addition of QTG, Table 1. The optimized extraction procedure was employed using the smart syringe, as elaborated in Section 2.4.4. The enrichment procedure consists of four steps, namely conditioning, extraction, washing, and elution (Figure 1). In all cases, recovery of PL after repeated purifying cycles was found to be more than 96%. Each purification cycle was completed in 4 min, and 4-6 purification cycles are required for about 100% PL recovery. The recovery from non-spiked samples approached 98±l%. A slight decrease in recovery to 96% for spiked samples is attributed to the high concentration of PL. The same results were found in the case of QTG, where recoveries were found at 97.4% and 98% with the spiked and non-spiked samples, respectively. The UPLC profiles of the methanolic extract of Tagetes patula and Tagetes erecta have been depicted in Figure 12, which contains two major and three minor components. A clear reduction in the intensity of the PL and QTG peaks was noticed in the supernatant of the PL-MIP sample compared to other compounds. The supernatant of the same sample treated with NIP did not show any significant reduction in the PL and QTG peaks compared to other peaks. The analysis of supernatant demonstrates the selective extraction of PL and QTG by PL-MIP compared to NIP. The result was further confirmed by the excessive presence of PL and QTG in the desorbed sample from PL-MIP compared to a mixture from NIP. The strategy can be used to selectively enrich any compound of interest using its MIP as an adsorbent. The selective extraction method is fairly green with a minimal amount of solvents and a swift protocol without requiring any high-tech machines, rendering significantly high selectivity for the target compound in comparison to conventional extraction methods.

Close

Figure 12.

UPLC chromatograms of real samples of (a) T. patula and (b) T. erecta, prepared in MeOH: DI water (1:1), PL-MIP and NIP supernatants, and PL-MIP and NIP extract, performed at optimized conditions (pH = 4, t = 30 min, stirring speed = 500 rpm, solution volume = 0.5 mL, and PL-MIP amount = 2.5 mg).

Export to PPT