Magnetic Solid-phase Extraction with Fe3O4/Molecularly Imprinted Polymers Modified by Deep Eutectic Solvents and Ionic Liquids for the Rapid Purification of Alkaloid Isomers (Theobromine and Theophylline) from Green Tea

Different kinds of deep eutectic solvents (DES) based on choline chloride (ChCl) and ionic liquids (ILs) based on 1-methylimidazole were used to modify Fe3O4/molecularly imprinted polymers (Fe3O4/MIPs), and the resulting materials were applied for the rapid purification of alkaloid isomers (theobromine and theophylline) from green tea with magnetic solid-phase extraction (M-SPE). The M-SPE procedure was optimized using the response surface methodology (RSM) to analyze the maximum conditions. The materials were characterized by Fourier transform infrared spectroscopy (FI-IR) and field emission scanning electron microscopy (FE-SEM). Compared to the ILs-Fe3O4/MIPs, the DESs-Fe3O4/MIPs were developed for the stronger recognition and higher recoveries of the isomers (theophylline and theobromine) from green tea, particularly DES-7-Fe3O4/MIPs. With RSM, the optimal recovery condition for theobromine and theophylline in the M-SPE were observed with ratio of methanol (80%) as the washing solution, methanol/acetic acid (HAc) (8:2) as the eluent at pH 3, and an eluent volume of 4 mL. The practical recoveries of theobromine and theophylline in green tea were 92.27% and 87.51%, respectively, with a corresponding actual extraction amount of 4.87 mg·g−1 and 5.07 mg·g−1. Overall, the proposed approach with the high affinity of Fe3O4/MIPs might offer a novel method for the purification of complex isomer samples.


Introduction
Green tea has many beneficial effects, such as reducing the risk of cancer and cardiovascular diseases, antioxidant effects, anti-inflammation, and anti-obesity. These effects are due mainly to its purine alkaloids, such as methylxanthines [1,2]. Theobromine (3,7-dimethylxanthine) and theophylline (1,3-dimethylxanthine) are common alkaloid isomers that belong to the methylxanthine family, which is most abundant in food [3,4]. The traditional approaches to determining the two methylxanthines have some disadvantages, such as being time-consuming and tedious, low sensitivity, and consuming large amounts of toxic organic solvents [5,6]. Owing to the complexity of the sample matrices, sample pretreatment is a crucial step in the analytical method. Thus far, the most widely used sample pretreatment method is solid-phase extraction (SPE) [7]. On the other hand, SPE is limited when the concentrations of the target compounds are extremely low or there is some interference of the complex in biological samples. Compared to traditional SPE, magnetic SPE (M-SPE) can be used effectively for the purification of trace amounts of the targets [8,9].

Purification of the Extracts with Solid-Phase Extraction
The recoveries of the extracted samples with DESs-Fe 3 O 4 /MIPs and ILs-Fe 3 O 4 /MIPs indicated that they had high selectivity and affinity for the analytes in an aqueous environment. Compared to the ILs-Fe 3 O 4 /MIPs, DESs-Fe 3 O 4 /MIPs had a better effect. The method was reliable and could be used for the trace analysis of theobromine and theophylline in green tea. In particular, DES-7-Fe 3 O 4 /MIPs were observed with the highest recoveries and recognition towards theobromine and theophylline in green tea. Figure 1 presents the recoveries of theobromine and theophylline by the different types of Fe 3 O 4 /MIPs.  Table 2. Intra-day and Inter-day precisions and accuracies of theobromine and theophylline.

Purification of the Extracts with Solid-Phase Extraction
The recoveries of the extracted samples with DESs-Fe3O4/MIPs and ILs-Fe3O4/MIPs indicated that they had high selectivity and affinity for the analytes in an aqueous environment. Compared to the ILs-Fe3O4/MIPs, DESs-Fe3O4/MIPs had a better effect. The method was reliable and could be used for the trace analysis of theobromine and theophylline in green tea. In particular, DES-7-Fe3O4/MIPs were observed with the highest recoveries and recognition towards theobromine and theophylline in green tea. Figure 1 presents the recoveries of theobromine and theophylline by the different types of Fe3O4/MIPs. DES-7, the type of functional synthetic molten salt composed of an organic cation (ChCl) and anion, was introduced in the procedure of Fe3O4/MIPs synthesis to improve the affinity and selectivity of the resulting nanoparticles. The electrostatic and ion-exchange interactions of the DES-7 acylamino group was firmly embedded in the Fe3O4/MIPs, while -NH of choline chloride and urea could be the functional group on the surface of the nanoparticles to form a hydrogen bond with the template, which was beneficial to the resolution of complex samples comprising the components with a wide range of polarities.
In addition, excessive DES-7 could be linked to the surface of the materials because an auxiliary solvent can also promote the functional solvents to form the specific binding sites and make Fe3O4/MIPs more rigid without shrinking or swelling, which can also improve the affinity and selectivity of the Fe3O4/MIPs. which was beneficial to the resolution of complex samples comprising the components with a wide range of polarities.
In addition, excessive DES-7 could be linked to the surface of the materials because an auxiliary solvent can also promote the functional solvents to form the specific binding sites and make Fe 3 O 4 /MIPs more rigid without shrinking or swelling, which can also improve the affinity and selectivity of the Fe 3 O 4 /MIPs.  Figure 2 shows many macropores and flow-through channels inlaid in the network skeletons of the Fe3O4/MIPs and DESs-7-Fe3O4/MIPs. In addition, compared to the Fe3O4/MIPs, the surface of the DESs-7-Fe3O4/MIPs was irregular and agglomeration was observed. The rough surface of the DESs-7-Fe3O4/MIPs was porous, which corresponded to the decomposition of the magnetic microsphere surface. The looser skeleton not only decreased the mass-transfer resistance but also increased the surface area of the nanoparticles in favor of embedding theobromine and theophylline into the cavity of the DESs-7-Fe3O4/MIPs, which resulted in easier adsorption and desorption and higher stability and reproducibility. The FT-IR spectra of the Fe3O4/MIPs and DESs-Fe3O4/MIPs ( Figure 3) showed a great difference in the fingerprint region. Fe3O4/MIPs displayed the characteristic peaks of Fe-O at 568 cm −1 , while the DESs-7-Fe3O4/MIPs exhibited the relatively strong band of the -OH group at 1560 cm −1 and 3320 cm −1 . In addition, the band at approximately 2365 cm −1 was assigned to the stretching vibration of -CH2 or -CH3 from choline chloride; the band at 1340 cm −1 was assigned to absorption by the C-O group; the band at 2365 cm −1 was ascribed to the stretching vibrations of the tertiary amine group (C-N) from choline chloride. Compared to Fe3O4/MIPs, FT-IR spectra of DESs-7-Fe3O4/MIPs showed that the hydrogen bond was strengthened and enhanced. The FT-IR spectra of the Fe 3 O 4 /MIPs and DESs-Fe 3 O 4 /MIPs ( Figure 3) showed a great difference in the fingerprint region. Fe 3 O 4 /MIPs displayed the characteristic peaks of Fe-O at 568 cm −1 , while the DESs-7-Fe 3 O 4 /MIPs exhibited the relatively strong band of the -OH group at 1560 cm −1 and 3320 cm −1 . In addition, the band at approximately 2365 cm −1 was assigned to the stretching vibration of -CH 2 or -CH 3 from choline chloride; the band at 1340 cm −1 was assigned to absorption by the C-O group; the band at 2365 cm −1 was ascribed to the stretching vibrations of the tertiary amine group

Adsorption Properties of Fe3O4/MIPs and DESs-7-Fe3O4/MIPs
The static and dynamic adsorption curves were used to evaluate the binding property of the Fe3O4/MIPs and DES-7-Fe3O4/MIPs at room temperature. The static and dynamic adsorption experiments (Figures 4 and 5, respectively) showed that the amounts of theobromine and theophylline sorbed by the nanoparticles increased with increasing concentration (5.00-300.00 μg·mL −1 ), and DES-7-Fe3O4/MIPs displayed higher affinity than Fe3O4/MIPs. The difference in the adsorption capacity between the Fe3O4/MIPs and DES-7-Fe3O4/MIPs increased with increasing theobromine and theophylline concentration until both reached equilibrium. The static adsorption capacity of the analytes on the DES-7-Fe3O4/MIPs (0.0441 mg/g for theobromine, and 0.0418 mg/g for theophylline) revealed better affinity than the Fe3O4/nanoparticles.

Adsorption Properties of Fe 3 O 4 /MIPs and DESs-7-Fe 3 O 4 /MIPs
The static and dynamic adsorption curves were used to evaluate the binding property of the

Adsorption Properties of Fe3O4/MIPs and DESs-7-Fe3O4/MIPs
The static and dynamic adsorption curves were used to evaluate the binding property of the Fe3O4/MIPs and DES-7-Fe3O4/MIPs at room temperature. The static and dynamic adsorption experiments (Figures 4 and 5, respectively) showed that the amounts of theobromine and theophylline sorbed by the nanoparticles increased with increasing concentration (5.00-300.00 μg·mL −1 ), and DES-7-Fe3O4/MIPs displayed higher affinity than Fe3O4/MIPs. The difference in the adsorption capacity between the Fe3O4/MIPs and DES-7-Fe3O4/MIPs increased with increasing theobromine and theophylline concentration until both reached equilibrium. The static adsorption capacity of the analytes on the DES-7-Fe3O4/MIPs (0.0441 mg/g for theobromine, and 0.0418 mg/g for theophylline) revealed better affinity than the Fe3O4/nanoparticles.  they were studied to reach adsorption equilibrium at different times (270 min for theobromine and 300 min for theophylline). Having specific recognition sites for theobromine and theophylline, the DES-7-Fe 3 O 4 /MIPs showed good sensitivity for the two targets. they were studied to reach adsorption equilibrium at different times (270 min for theobromine and 300 min for theophylline). Having specific recognition sites for theobromine and theophylline, the DES-7-Fe3O4/MIPs showed good sensitivity for the two targets.

Optimization of M-SPE Procedure
Among the materials, the DES-7-Fe3O4/MIPs showed the best recognition for the two targets. Figure 6 shows the effects of MeCN, EtOH, H2O, MeOH, EtOAc, and acetone, respectively, as the washing solution in the M-SPE procedure with DES-7-Fe3O4/MIPs. Methanol was observed as the best washing solution among the five types of washing solution.

Optimization of M-SPE Procedure
Among the materials, the DES-7-Fe 3 O 4 /MIPs showed the best recognition for the two targets. Figure 6 shows the effects of MeCN, EtOH, H 2 O, MeOH, EtOAc, and acetone, respectively, as the washing solution in the M-SPE procedure with DES-7-Fe 3 O 4 /MIPs. Methanol was observed as the best washing solution among the five types of washing solution. they were studied to reach adsorption equilibrium at different times (270 min for theobromine and 300 min for theophylline). Having specific recognition sites for theobromine and theophylline, the DES-7-Fe3O4/MIPs showed good sensitivity for the two targets.

Optimization of M-SPE Procedure
Among the materials, the DES-7-Fe3O4/MIPs showed the best recognition for the two targets. Figure 6 shows the effects of MeCN, EtOH, H2O, MeOH, EtOAc, and acetone, respectively, as the washing solution in the M-SPE procedure with DES-7-Fe3O4/MIPs. Methanol was observed as the best washing solution among the five types of washing solution.    Figure 7 shows the 3D response surface plots. In this figure, the recoveries of theobromine increased when ratio of methanol as the washing solution was in the designed range from 0:1 to 100:0, and the response increased initially then decreased when the pH of the eluent was increased from 2 to 8, and there was a small increase followed by a decrease as the eluent volume ranged from 1 mL to 6 mL. Moreover, the recoveries of theophylline ( Figure 8) showed similar changes. increased when ratio of methanol as the washing solution was in the designed range from 0:1 to 100:0, and the response increased initially then decreased when the pH of the eluent was increased from 2 to 8, and there was a small increase followed by a decrease as the eluent volume ranged from 1 mL to 6 mL. Moreover, the recoveries of theophylline ( Figure 8) showed similar changes.  The recoveries of theobromine and theophylline decreased with decreasing ratio of ethanol. As the ratio of ethanol decreased, the polarity of the washing solvents decreased and in this way, the interference of the impurities in the matrix of samples and materials were reduced. The different pH values were attributed to the destruction of the hydrogen-bonding interaction between the template increased when ratio of methanol as the washing solution was in the designed range from 0:1 to 100:0, and the response increased initially then decreased when the pH of the eluent was increased from 2 to 8, and there was a small increase followed by a decrease as the eluent volume ranged from 1 mL to 6 mL. Moreover, the recoveries of theophylline ( Figure 8) showed similar changes.  The recoveries of theobromine and theophylline decreased with decreasing ratio of ethanol. As the ratio of ethanol decreased, the polarity of the washing solvents decreased and in this way, the interference of the impurities in the matrix of samples and materials were reduced. The different pH values were attributed to the destruction of the hydrogen-bonding interaction between the template The recoveries of theobromine and theophylline decreased with decreasing ratio of ethanol. As the ratio of ethanol decreased, the polarity of the washing solvents decreased and in this way, the interference of the impurities in the matrix of samples and materials were reduced. The different pH values were attributed to the destruction of the hydrogen-bonding interaction between the template molecules and imprinted cavities. The eluent volume is an important parameter that influences the recoveries of theophylline and theobromine in the M-SPE procedure.
The optimal recoveries condition for theobromine and theophylline were observed under the same condition (ratio of methanol in water (80%), pH of the eluent (PH = 3), and the eluent volume (4 mL) and was estimated using the model equation by solving the regression equation and analyzing the response surface contour plots. The theoretical recoveries of theobromine and theophylline under the above conditions were 92.27% and 87.51%, respectively.

Analysis of Green Tea Samples
The proposed M-SPE method was used to determine the alkaloid isomers (theobromine and theophylline) in green tea under the optimal conditions. The optimal conditions for the recovery of theobromine and theophylline were ratio of methanol (80%) as the washing solution, methanol/HAc (8:2) eluent at pH 3, and an eluent volume of 4 mL. The practical recoveries of theobromine and theophylline in green tea were 92.27% and 87.51%, with corresponding extraction amounts of 4.87 mg·g −1 and 5.07 mg·g −1 . Figure 9 presents chromatograms of the sample extracts using molecules and imprinted cavities. The eluent volume is an important parameter that influences the recoveries of theophylline and theobromine in the M-SPE procedure. The optimal recoveries condition for theobromine and theophylline were observed under the same condition (ratio of methanol in water (80%), pH of the eluent (PH = 3), and the eluent volume (4 mL) and was estimated using the model equation by solving the regression equation and analyzing the response surface contour plots. The theoretical recoveries of theobromine and theophylline under the above conditions were 92.27% and 87.51%, respectively.

Analysis of Green Tea Samples
The proposed M-SPE method was used to determine the alkaloid isomers (theobromine and theophylline) in green tea under the optimal conditions. The optimal conditions for the recovery of theobromine and theophylline were ratio of methanol (80%) as the washing solution, methanol/HAc (8:2) eluent at pH 3, and an eluent volume of 4 mL. The practical recoveries of theobromine and theophylline in green tea were 92.27% and 87.51%, with corresponding extraction amounts of 4.87 mg·g −1 and 5.07 mg·g −1 . Figure 9 presents chromatograms of the sample extracts using Fe3O4/MIPs and DES-7-Fe3O4/MIPs. Compared to the two chromatograms, the chromatogram of the DES-7-Fe3O4/MIPs with fewer interfering peaks and a better chromatogram shape, and the peaks of the two isomers were easier to distinguish.

Reagents and Materials
Green tea was purchased from a local market (Incheon, Korea). Theophylline and theobromine were obtained from Sigma-Aldrich.

Reagents and Materials
Green tea was purchased from a local market (Incheon, Korea). Theophylline and theobromine were obtained from Sigma-Aldrich. Co., Ltd.

Chromatography and Sample Pretreatment
The chromatography system consisted of a Waters 600 s Multi solvent Delivery System, Waters 1515 liquid chromatography pump (Waters, MA, USA), a Rheodyne injector (20 µL sample loop) and a variable wavelength 2489 UV dual channel detector. EmpowerTM 3 software (Waters) was used as the data acquisition system. The analysis was performed on an OptimaPak C 18 column (5 µm, 250 × 4.6 mm, i.d., RStech Corporation, Daejeon, Korea). The mobile phase was methanol-water-acetic acid (20/80/2, v/v/v). The flow rate was 0.8 mL·min −1 , and the detection wavelength was 280 nm.
The standard theobromine and theophylline were dissolved in methanol to a concentration of 1000.00 µg·mL −1 . For method development, a series of standard solutions containing theobromine and theophylline were prepared at five concentrations ranging from 5.00-100.00 µg·mL −1 . The standard curve of theobromine and theophylline were linear by assaying five data points and measured twice.

Preparation of DESs and ILs
The DESs were formed by ChCl and ethylene glycol, glycerol, and 1,4-butanediol, urea, formic acid, acetic acid, and propionic acid (1/2, n/n) in a conical flask, and heated to 80 • C with constant stirring for 2 h until a homogeneous liquid formed. Table 3 lists the DESs. ethanol were obtained from Fisher Scientific Co., Ltd. (Seoul, Korea). All other solvents used in the experiment were of HPLC or analytical grade.

Chromatography and Sample Pretreatment
The chromatography system consisted of a Waters 600 s Multi solvent Delivery System, Waters 1515 liquid chromatography pump (Waters, MA, USA), a Rheodyne injector (20 μL sample loop) and a variable wavelength 2489 UV dual channel detector. EmpowerTM 3 software (Waters) was used as the data acquisition system. The analysis was performed on an OptimaPak C18 column (5 μm, 250 × 4.6 mm, i.d., RStech Corporation, Daejeon, Korea). The mobile phase was methanol-water-acetic acid (20/80/2, v/v/v). The flow rate was 0.8 mL·min −1 , and the detection wavelength was 280 nm.
The standard theobromine and theophylline were dissolved in methanol to a concentration of 1000.00 μg·mL −1 . For method development, a series of standard solutions containing theobromine and theophylline were prepared at five concentrations ranging from 5.00-100.00 μg·mL −1 . The standard curve of theobromine and theophylline were linear by assaying five data points and measured twice.

Preparation of DESs and ILs
The DESs were formed by ChCl and ethylene glycol, glycerol, and 1,4-butanediol, urea, formic acid, acetic acid, and propionic acid (1/2, n/n) in a conical flask, and heated to 80 °C with constant stirring for 2 h until a homogeneous liquid formed. Table 3 lists the DESs.

Chromatography and Sample Pretreatment
The chromatography system consisted of a Waters 600 s Multi solvent Delivery System, Waters 1515 liquid chromatography pump (Waters, MA, USA), a Rheodyne injector (20 μL sample loop) and a variable wavelength 2489 UV dual channel detector. EmpowerTM 3 software (Waters) was used as the data acquisition system. The analysis was performed on an OptimaPak C18 column (5 μm, 250 × 4.6 mm, i.d., RStech Corporation, Daejeon, Korea). The mobile phase was methanol-water-acetic acid (20/80/2, v/v/v). The flow rate was 0.8 mL·min −1 , and the detection wavelength was 280 nm.
The standard theobromine and theophylline were dissolved in methanol to a concentration of 1000.00 μg·mL −1 . For method development, a series of standard solutions containing theobromine and theophylline were prepared at five concentrations ranging from 5.00-100.00 μg·mL −1 . The standard curve of theobromine and theophylline were linear by assaying five data points and measured twice.

Preparation of DESs and ILs
The DESs were formed by ChCl and ethylene glycol, glycerol, and 1,4-butanediol, urea, formic acid, acetic acid, and propionic acid (1/2, n/n) in a conical flask, and heated to 80 °C with constant stirring for 2 h until a homogeneous liquid formed. Table 3 lists the DESs.

Chromatography and Sample Pretreatment
The chromatography system consisted of a Waters 600 s Multi solvent Delivery System, Waters 1515 liquid chromatography pump (Waters, MA, USA), a Rheodyne injector (20 μL sample loop) and a variable wavelength 2489 UV dual channel detector. EmpowerTM 3 software (Waters) was used as the data acquisition system. The analysis was performed on an OptimaPak C18 column (5 μm, 250 × 4.6 mm, i.d., RStech Corporation, Daejeon, Korea). The mobile phase was methanol-water-acetic acid (20/80/2, v/v/v). The flow rate was 0.8 mL·min −1 , and the detection wavelength was 280 nm.
The standard theobromine and theophylline were dissolved in methanol to a concentration of 1000.00 μg·mL −1 . For method development, a series of standard solutions containing theobromine and theophylline were prepared at five concentrations ranging from 5.00-100.00 μg·mL −1 . The standard curve of theobromine and theophylline were linear by assaying five data points and measured twice.

Preparation of DESs and ILs
The DESs were formed by ChCl and ethylene glycol, glycerol, and 1,4-butanediol, urea, formic acid, acetic acid, and propionic acid (1/2, n/n) in a conical flask, and heated to 80 °C with constant stirring for 2 h until a homogeneous liquid formed. Table 3 lists the DESs.

Chromatography and Sample Pretreatment
The chromatography system consisted of a Waters 600 s Multi solvent Delivery System, Waters 1515 liquid chromatography pump (Waters, MA, USA), a Rheodyne injector (20 μL sample loop) and a variable wavelength 2489 UV dual channel detector. EmpowerTM 3 software (Waters) was used as the data acquisition system. The analysis was performed on an OptimaPak C18 column (5 μm, 250 × 4.6 mm, i.d., RStech Corporation, Daejeon, Korea). The mobile phase was methanol-water-acetic acid (20/80/2, v/v/v). The flow rate was 0.8 mL·min −1 , and the detection wavelength was 280 nm.
The standard theobromine and theophylline were dissolved in methanol to a concentration of 1000.00 μg·mL −1 . For method development, a series of standard solutions containing theobromine and theophylline were prepared at five concentrations ranging from 5.00-100.00 μg·mL −1 . The standard curve of theobromine and theophylline were linear by assaying five data points and measured twice.

Preparation of DESs and ILs
The DESs were formed by ChCl and ethylene glycol, glycerol, and 1,4-butanediol, urea, formic acid, acetic acid, and propionic acid (1/2, n/n) in a conical flask, and heated to 80 °C with constant stirring for 2 h until a homogeneous liquid formed. Table 3 lists the DESs.

Chromatography and Sample Pretreatment
The chromatography system consisted of a Waters 600 s Multi solvent Delivery System, Waters 1515 liquid chromatography pump (Waters, MA, USA), a Rheodyne injector (20 μL sample loop) and a variable wavelength 2489 UV dual channel detector. EmpowerTM 3 software (Waters) was used as the data acquisition system. The analysis was performed on an OptimaPak C18 column (5 μm, 250 × 4.6 mm, i.d., RStech Corporation, Daejeon, Korea). The mobile phase was methanol-water-acetic acid (20/80/2, v/v/v). The flow rate was 0.8 mL·min −1 , and the detection wavelength was 280 nm.
The standard theobromine and theophylline were dissolved in methanol to a concentration of 1000.00 μg·mL −1 . For method development, a series of standard solutions containing theobromine and theophylline were prepared at five concentrations ranging from 5.00-100.00 μg·mL −1 . The standard curve of theobromine and theophylline were linear by assaying five data points and measured twice.

Preparation of DESs and ILs
The DESs were formed by ChCl and ethylene glycol, glycerol, and 1,4-butanediol, urea, formic acid, acetic acid, and propionic acid (1/2, n/n) in a conical flask, and heated to 80 °C with constant stirring for 2 h until a homogeneous liquid formed. Table 3 lists the DESs.

Chromatography and Sample Pretreatment
The chromatography system consisted of a Waters 600 s Multi solvent Delivery System, Waters 1515 liquid chromatography pump (Waters, MA, USA), a Rheodyne injector (20 μL sample loop) and a variable wavelength 2489 UV dual channel detector. EmpowerTM 3 software (Waters) was used as the data acquisition system. The analysis was performed on an OptimaPak C18 column (5 μm, 250 × 4.6 mm, i.d., RStech Corporation, Daejeon, Korea). The mobile phase was methanol-water-acetic acid (20/80/2, v/v/v). The flow rate was 0.8 mL·min −1 , and the detection wavelength was 280 nm.
The standard theobromine and theophylline were dissolved in methanol to a concentration of 1000.00 μg·mL −1 . For method development, a series of standard solutions containing theobromine and theophylline were prepared at five concentrations ranging from 5.00-100.00 μg·mL −1 . The standard curve of theobromine and theophylline were linear by assaying five data points and measured twice.

Preparation of DESs and ILs
The DESs were formed by ChCl and ethylene glycol, glycerol, and 1,4-butanediol, urea, formic acid, acetic acid, and propionic acid (1/2, n/n) in a conical flask, and heated to 80 °C with constant stirring for 2 h until a homogeneous liquid formed. Table 3 lists the DESs.

Chromatography and Sample Pretreatment
The chromatography system consisted of a Waters 600 s Multi solvent Delivery System, Waters 1515 liquid chromatography pump (Waters, MA, USA), a Rheodyne injector (20 μL sample loop) and a variable wavelength 2489 UV dual channel detector. EmpowerTM 3 software (Waters) was used as the data acquisition system. The analysis was performed on an OptimaPak C18 column (5 μm, 250 × 4.6 mm, i.d., RStech Corporation, Daejeon, Korea). The mobile phase was methanol-water-acetic acid (20/80/2, v/v/v). The flow rate was 0.8 mL·min −1 , and the detection wavelength was 280 nm.
The standard theobromine and theophylline were dissolved in methanol to a concentration of 1000.00 μg·mL −1 . For method development, a series of standard solutions containing theobromine and theophylline were prepared at five concentrations ranging from 5.00-100.00 μg·mL −1 . The standard curve of theobromine and theophylline were linear by assaying five data points and measured twice.

Preparation of DESs and ILs
The DESs were formed by ChCl and ethylene glycol, glycerol, and 1,4-butanediol, urea, formic acid, acetic acid, and propionic acid (1/2, n/n) in a conical flask, and heated to 80 °C with constant stirring for 2 h until a homogeneous liquid formed. Table 3 lists the DESs.

Chromatography and Sample Pretreatment
The chromatography system consisted of a Waters 600 s Multi solvent Delivery System, Waters 1515 liquid chromatography pump (Waters, MA, USA), a Rheodyne injector (20 μL sample loop) and a variable wavelength 2489 UV dual channel detector. EmpowerTM 3 software (Waters) was used as the data acquisition system. The analysis was performed on an OptimaPak C18 column (5 μm, 250 × 4.6 mm, i.d., RStech Corporation, Daejeon, Korea). The mobile phase was methanol-water-acetic acid (20/80/2, v/v/v). The flow rate was 0.8 mL·min −1 , and the detection wavelength was 280 nm.
The standard theobromine and theophylline were dissolved in methanol to a concentration of 1000.00 μg·mL −1 . For method development, a series of standard solutions containing theobromine and theophylline were prepared at five concentrations ranging from 5.00-100.00 μg·mL −1 . The standard curve of theobromine and theophylline were linear by assaying five data points and measured twice.

Preparation of DESs and ILs
The DESs were formed by ChCl and ethylene glycol, glycerol, and 1,4-butanediol, urea, formic acid, acetic acid, and propionic acid (1/2, n/n) in a conical flask, and heated to 80 °C with constant stirring for 2 h until a homogeneous liquid formed. Table 3 lists the DESs.

Preparation of Fe3O4/MIPs
The Fe3O4/MIPs were prepared by a chemical coprecipitation method [37,38]. FeCl2·4H2O (6.0 g), FeCl3·6H2O (15.6 g), and hydrochloric acid (12 M, 2.55 mL) were dissolved in pure water (50 mL). The mixture was added dropwise to a NaOH solution (250 mL, 1.5 M) with vigorous stirring with nitrogen gas passing continuously through the solution during the reaction. Subsequently, the magnetic precipitates were isolated from the solution using a magnet, and washed sequentially with water and ethanol before being dried at 50 °C in a vacuum for 24 h.

Preparation of DESs (ILs)-Fe3O4/MIPs
A 200 mg sample of Fe3O4/MIPs was added to a 250 mL round-bottom flask and dispersed ultrasonically for 20 min. Subsequently, isopropanol (8 mL) with the DESs (ILs, 2 mL) was added to the flask and reacted at 80 °C by stirring these two components for 24 h. After the reaction, the product was separated from the reaction medium under an applied magnetic field, rinsed three times with pure water, twice with isopropanol, and then separated using an external magnetic field. The product was then dried in the vacuum.
MAA was each added to Fe3O4 particles in a clean, dry round bottomed flask containing a magnetic stirring bar. Template molecules of theobromine and theophylline, were then allowed to form hydrogen bonds between them. The emulsions were sonicated for 20 min and stored at 4 °C in the dark. EDMA (20 mmol) and AIBN (1.0 mmol) were added to the mixed solutions together at 60 °C for 12 h. After polymerization, the bulk polymers were ground and sieved through a 105 μm stainless-steel mesh. The resulting polymers were washed with MeOH-HOAc (9:1, v/v) in a Soxhlet apparatus to remove the templates, and dried in a vacuum for 12 h. In the same case as the other synthetic processes, deep eutectic solvents-polymers without templates (DESs-Fe3O4/NIPs) were prepared in the absence of DESs. The magnetic polymer without templates (Fe3O4/NIPs) was also synthesized using the same procedure but in the absence of the templates and without DESs.

Characterization of the Fe3O4/MIPs and DESs (ILs)-Fe3O4/MIPs
The Fe3O4/MIPs and DESs (ILs)-Fe3O4/MIPs were dried at 50 °C for 24 h. The amount of the obtained materials was ground together with KBr for tablets. The spectra were then analyzed using a Fourier transform infrared (FTIR) spectrometer (VERTEX 80V) in the wave number range of 400-4000 cm −1 .
The morphological microstructures of these materials were observed by field emission scanning electron microscopy (FE-SEM, SE-4200, MERLIN Compact, ZEISS, Jena, Germany). The ILs were formed with 1-methylimidazole (1 mL) and an excess of bromoethane, bromobutane, bromohexane, and bromohexane in a round-bottom flask at 80 °C for 6 h and washed with ethyl acetate after cooling to room temperature.

Preparation of Fe3O4/MIPs
The Fe3O4/MIPs were prepared by a chemical coprecipitation method [37,38]. FeCl2·4H2O (6.0 g), FeCl3·6H2O (15.6 g), and hydrochloric acid (12 M, 2.55 mL) were dissolved in pure water (50 mL). The mixture was added dropwise to a NaOH solution (250 mL, 1.5 M) with vigorous stirring with nitrogen gas passing continuously through the solution during the reaction. Subsequently, the magnetic precipitates were isolated from the solution using a magnet, and washed sequentially with water and ethanol before being dried at 50 °C in a vacuum for 24 h.

Preparation of DESs (ILs)-Fe3O4/MIPs
A 200 mg sample of Fe3O4/MIPs was added to a 250 mL round-bottom flask and dispersed ultrasonically for 20 min. Subsequently, isopropanol (8 mL) with the DESs (ILs, 2 mL) was added to the flask and reacted at 80 °C by stirring these two components for 24 h. After the reaction, the product was separated from the reaction medium under an applied magnetic field, rinsed three times with pure water, twice with isopropanol, and then separated using an external magnetic field. The product was then dried in the vacuum.
MAA was each added to Fe3O4 particles in a clean, dry round bottomed flask containing a magnetic stirring bar. Template molecules of theobromine and theophylline, were then allowed to form hydrogen bonds between them. The emulsions were sonicated for 20 min and stored at 4 °C in the dark. EDMA (20 mmol) and AIBN (1.0 mmol) were added to the mixed solutions together at 60 °C for 12 h. After polymerization, the bulk polymers were ground and sieved through a 105 μm stainless-steel mesh. The resulting polymers were washed with MeOH-HOAc (9:1, v/v) in a Soxhlet apparatus to remove the templates, and dried in a vacuum for 12 h. In the same case as the other synthetic processes, deep eutectic solvents-polymers without templates (DESs-Fe3O4/NIPs) were prepared in the absence of DESs. The magnetic polymer without templates (Fe3O4/NIPs) was also synthesized using the same procedure but in the absence of the templates and without DESs.

Characterization of the Fe3O4/MIPs and DESs (ILs)-Fe3O4/MIPs
The Fe3O4/MIPs and DESs (ILs)-Fe3O4/MIPs were dried at 50 °C for 24 h. The amount of the obtained materials was ground together with KBr for tablets. The spectra were then analyzed using a Fourier transform infrared (FTIR) spectrometer (VERTEX 80V) in the wave number range of 400-4000 cm −1 .
The morphological microstructures of these materials were observed by field emission scanning electron microscopy (FE-SEM, SE-4200, MERLIN Compact, ZEISS, Jena, Germany). The ILs were formed with 1-methylimidazole (1 mL) and an excess of bromoethane, bromobutane, bromohexane, and bromohexane in a round-bottom flask at 80 °C for 6 h and washed with ethyl acetate after cooling to room temperature.

Preparation of Fe3O4/MIPs
The Fe3O4/MIPs were prepared by a chemical coprecipitation method [37,38]. FeCl2·4H2O (6.0 g), FeCl3·6H2O (15.6 g), and hydrochloric acid (12 M, 2.55 mL) were dissolved in pure water (50 mL). The mixture was added dropwise to a NaOH solution (250 mL, 1.5 M) with vigorous stirring with nitrogen gas passing continuously through the solution during the reaction. Subsequently, the magnetic precipitates were isolated from the solution using a magnet, and washed sequentially with water and ethanol before being dried at 50 °C in a vacuum for 24 h.

Preparation of DESs (ILs)-Fe3O4/MIPs
A 200 mg sample of Fe3O4/MIPs was added to a 250 mL round-bottom flask and dispersed ultrasonically for 20 min. Subsequently, isopropanol (8 mL) with the DESs (ILs, 2 mL) was added to the flask and reacted at 80 °C by stirring these two components for 24 h. After the reaction, the product was separated from the reaction medium under an applied magnetic field, rinsed three times with pure water, twice with isopropanol, and then separated using an external magnetic field. The product was then dried in the vacuum.
MAA was each added to Fe3O4 particles in a clean, dry round bottomed flask containing a magnetic stirring bar. Template molecules of theobromine and theophylline, were then allowed to form hydrogen bonds between them. The emulsions were sonicated for 20 min and stored at 4 °C in the dark. EDMA (20 mmol) and AIBN (1.0 mmol) were added to the mixed solutions together at 60 °C for 12 h. After polymerization, the bulk polymers were ground and sieved through a 105 μm stainless-steel mesh. The resulting polymers were washed with MeOH-HOAc (9:1, v/v) in a Soxhlet apparatus to remove the templates, and dried in a vacuum for 12 h. In the same case as the other synthetic processes, deep eutectic solvents-polymers without templates (DESs-Fe3O4/NIPs) were prepared in the absence of DESs. The magnetic polymer without templates (Fe3O4/NIPs) was also synthesized using the same procedure but in the absence of the templates and without DESs.

Characterization of the Fe3O4/MIPs and DESs (ILs)-Fe3O4/MIPs
The Fe3O4/MIPs and DESs (ILs)-Fe3O4/MIPs were dried at 50 °C for 24 h. The amount of the obtained materials was ground together with KBr for tablets. The spectra were then analyzed using a Fourier transform infrared (FTIR) spectrometer (VERTEX 80V) in the wave number range of 400-4000 cm −1 .
The morphological microstructures of these materials were observed by field emission scanning electron microscopy (FE-SEM, SE-4200, MERLIN Compact, ZEISS, Jena, Germany).

Preparation of Fe3O4/MIPs
The Fe3O4/MIPs were prepared by a chemical coprecipitation method [37,38]. FeCl2·4H2O (6.0 g), FeCl3·6H2O (15.6 g), and hydrochloric acid (12 M, 2.55 mL) were dissolved in pure water (50 mL). The mixture was added dropwise to a NaOH solution (250 mL, 1.5 M) with vigorous stirring with nitrogen gas passing continuously through the solution during the reaction. Subsequently, the magnetic precipitates were isolated from the solution using a magnet, and washed sequentially with water and ethanol before being dried at 50 °C in a vacuum for 24 h.

Preparation of DESs (ILs)-Fe3O4/MIPs
A 200 mg sample of Fe3O4/MIPs was added to a 250 mL round-bottom flask and dispersed ultrasonically for 20 min. Subsequently, isopropanol (8 mL) with the DESs (ILs, 2 mL) was added to the flask and reacted at 80 °C by stirring these two components for 24 h. After the reaction, the product was separated from the reaction medium under an applied magnetic field, rinsed three times with pure water, twice with isopropanol, and then separated using an external magnetic field. The product was then dried in the vacuum.
MAA was each added to Fe3O4 particles in a clean, dry round bottomed flask containing a magnetic stirring bar. Template molecules of theobromine and theophylline, were then allowed to form hydrogen bonds between them. The emulsions were sonicated for 20 min and stored at 4 °C in the dark. EDMA (20 mmol) and AIBN (1.0 mmol) were added to the mixed solutions together at 60 °C for 12 h. After polymerization, the bulk polymers were ground and sieved through a 105 μm stainless-steel mesh. The resulting polymers were washed with MeOH-HOAc (9:1, v/v) in a Soxhlet apparatus to remove the templates, and dried in a vacuum for 12 h. In the same case as the other synthetic processes, deep eutectic solvents-polymers without templates (DESs-Fe3O4/NIPs) were prepared in the absence of DESs. The magnetic polymer without templates (Fe3O4/NIPs) was also synthesized using the same procedure but in the absence of the templates and without DESs.

Characterization of the Fe3O4/MIPs and DESs (ILs)-Fe3O4/MIPs
The Fe3O4/MIPs and DESs (ILs)-Fe3O4/MIPs were dried at 50 °C for 24 h. The amount of the obtained materials was ground together with KBr for tablets. The spectra were then analyzed using a Fourier transform infrared (FTIR) spectrometer (VERTEX 80V) in the wave number range of 400-4000 cm −1 .
The morphological microstructures of these materials were observed by field emission scanning electron microscopy (FE-SEM, SE-4200, MERLIN Compact, ZEISS, Jena, Germany).

Preparation of Fe3O4/MIPs
The Fe3O4/MIPs were prepared by a chemical coprecipitation method [37,38]. FeCl2·4H2O (6.0 g), FeCl3·6H2O (15.6 g), and hydrochloric acid (12 M, 2.55 mL) were dissolved in pure water (50 mL). The mixture was added dropwise to a NaOH solution (250 mL, 1.5 M) with vigorous stirring with nitrogen gas passing continuously through the solution during the reaction. Subsequently, the magnetic precipitates were isolated from the solution using a magnet, and washed sequentially with water and ethanol before being dried at 50 °C in a vacuum for 24 h.

Preparation of DESs (ILs)-Fe3O4/MIPs
A 200 mg sample of Fe3O4/MIPs was added to a 250 mL round-bottom flask and dispersed ultrasonically for 20 min. Subsequently, isopropanol (8 mL) with the DESs (ILs, 2 mL) was added to the flask and reacted at 80 °C by stirring these two components for 24 h. After the reaction, the product was separated from the reaction medium under an applied magnetic field, rinsed three times with pure water, twice with isopropanol, and then separated using an external magnetic field. The product was then dried in the vacuum.
MAA was each added to Fe3O4 particles in a clean, dry round bottomed flask containing a magnetic stirring bar. Template molecules of theobromine and theophylline, were then allowed to form hydrogen bonds between them. The emulsions were sonicated for 20 min and stored at 4 °C in the dark. EDMA (20 mmol) and AIBN (1.0 mmol) were added to the mixed solutions together at 60 °C for 12 h. After polymerization, the bulk polymers were ground and sieved through a 105 μm stainless-steel mesh. The resulting polymers were washed with MeOH-HOAc (9:1, v/v) in a Soxhlet apparatus to remove the templates, and dried in a vacuum for 12 h. In the same case as the other synthetic processes, deep eutectic solvents-polymers without templates (DESs-Fe3O4/NIPs) were prepared in the absence of DESs. The magnetic polymer without templates (Fe3O4/NIPs) was also synthesized using the same procedure but in the absence of the templates and without DESs.

Characterization of the Fe3O4/MIPs and DESs (ILs)-Fe3O4/MIPs
The Fe3O4/MIPs and DESs (ILs)-Fe3O4/MIPs were dried at 50 °C for 24 h. The amount of the obtained materials was ground together with KBr for tablets. The spectra were then analyzed using a Fourier transform infrared (FTIR) spectrometer (VERTEX 80V) in the wave number range of 400-4000 cm −1 .
The morphological microstructures of these materials were observed by field emission scanning electron microscopy (FE-SEM, SE-4200, MERLIN Compact, ZEISS, Jena, Germany).

Preparation of Fe 3 O 4 /MIPs
The Fe 3 O 4 /MIPs were prepared by a chemical coprecipitation method [37,38]. FeCl 2 ·4H 2 O (6.0 g), FeCl 3 ·6H 2 O (15.6 g), and hydrochloric acid (12 M, 2.55 mL) were dissolved in pure water (50 mL). The mixture was added dropwise to a NaOH solution (250 mL, 1.5 M) with vigorous stirring with nitrogen gas passing continuously through the solution during the reaction. Subsequently, the magnetic precipitates were isolated from the solution using a magnet, and washed sequentially with water and ethanol before being dried at 50 • C in a vacuum for 24 h.

Preparation of DESs (ILs)-Fe 3 O 4 /MIPs
A 200 mg sample of Fe 3 O 4 /MIPs was added to a 250 mL round-bottom flask and dispersed ultrasonically for 20 min. Subsequently, isopropanol (8 mL) with the DESs (ILs, 2 mL) was added to the flask and reacted at 80 • C by stirring these two components for 24 h. After the reaction, the product was separated from the reaction medium under an applied magnetic field, rinsed three times with pure water, twice with isopropanol, and then separated using an external magnetic field. The product was then dried in the vacuum.
MAA was each added to Fe 3 O 4 particles in a clean, dry round bottomed flask containing a magnetic stirring bar. Template molecules of theobromine and theophylline, were then allowed to form hydrogen bonds between them. The emulsions were sonicated for 20 min and stored at 4 • C in the dark. EDMA (20 mmol) and AIBN (1.0 mmol) were added to the mixed solutions together at 60 • C for 12 h. After polymerization, the bulk polymers were ground and sieved through a 105 µm stainless-steel mesh. The resulting polymers were washed with MeOH-HOAc (9:1, v/v) in a Soxhlet apparatus to remove the templates, and dried in a vacuum for 12 h. In the same case as the other synthetic processes, deep eutectic solvents-polymers without templates (DESs-Fe 3 O 4 /NIPs) were prepared in the absence of DESs. The magnetic polymer without templates (Fe 3 O 4 /NIPs) was also synthesized using the same procedure but in the absence of the templates and without DESs.

Characterization of the Fe 3 O 4 /MIPs and DESs (ILs)-Fe 3 O 4 /MIPs
The Fe 3 O 4 /MIPs and DESs (ILs)-Fe 3 O 4 /MIPs were dried at 50 • C for 24 h. The amount of the obtained materials was ground together with KBr for tablets. The spectra were then analyzed using a Fourier transform infrared (FTIR) spectrometer (VERTEX 80V) in the wave number range of 400-4000 cm −1 .
The morphological microstructures of these materials were observed by field emission scanning electron microscopy (FE-SEM, SE-4200, MERLIN Compact, ZEISS, Jena, Germany).

Absorption Capacity of Fe 3 O 4 /MIPs and DESs-Fe 3 O 4 /MIPs
At room temperature, for the static absorption experiment, 20.0 mg each of the proposed materials was mixed with 2.0 mL of the theobromine and theophylline standard solutions (5.0-300.0 µg·mL −1 ) in centrifuge tubes. After shaking for 8 h, the mixture was centrifuged, and the theobromine and theophylline concentrations in the upper solution were measured to calculate the adsorption capacities.
For the dynamic adsorption experiment, 1.0 mL of the theobromine and theophylline standard solution (150.0 µg·mL −1 ) was mixed with 10.0 mg each of the proposed nanoparticles and shaken for 30-360 min. After centrifuging, the theobromine and theophylline levels in the upper solution for various times were determined to calculate the adsorption capacities.
The adsorption quantity (Q) was calculated based on the change in the free concentration (C free ) and the initial concentration (C 0 ) of the template by Equation (1), where V is the volume of the solution and W is the mass of the material powder:

Purification of Theobromine and Theophylline from Green Tea by M-SPE
Green tea was dried in an oven at 50 • C and ground to a powder. The designated conditions were an ultrasonic time of 1 h, an absolute ethyl alcohol as the extracting solution, and the ratio of material to liquid ratio of 1:20 (g·mL −1 ). The suspension was then filtered to obtain the extraction samples. After each SPE cartridge was preconditioned sequentially by methanol (1.5 mL) and deionized water (1.5 mL) to clean the cartridge, 1.0 mL of the extract solution was loaded on the cartridge followed by 1.5 mL of MeCN, EtOH, H 2 O, MeOH, EtOAc, and acetone, respectively, as the washing solution and methanol/HOAc (8:2) (1.5 mL) mixture solution, as the commonly used elution in SPE was chosen as the elution solution. The effluents at every step were collected using a 1.0 mL syringe, which was connected to the bottom of the SPE cartridge to ensure a suitable and constant flow rate.

Optimization of M-SPE Procedure
To achieve the excellent extraction efficiency, several parameters involving washing solvent, the pH of the eluent solvent and elution volume were optimized, and the 17-run BBD was applied to optimize the procedure statistically. In Table 5, these three factors were designated as X 1 , X 2 , and X 3 prescribed into three levels, coded +1, 0, and −1 for high, intermediate, and low values, respectively. The three test variables were coded according to the following equation: Ratio of methanol (X 1 ) (%) 0 50 100 PH of eluent (X 2 ) 2 5 8 Volume of eluent (X 3 ) (mL) 1 3. 5 6 In this equation, x i is the coded value of the independent variable, Xi is the actual value of the independent variable, X 0 is the actual value of the independent variable at the center point, and X is the step change value of the independent variable. A second-order polynomial model was fitted to correlate the relationship between the independent variables and the response (target recovery) to predict the optimized conditions: In this equation, Y is the dependent variable, A 0 is a constant, and A i , A ii , and A ij are coefficients estimated by the model. X i and X j are the levels of the independent variables that represent the linear, quadratic, and cross-product effects of the X 1 , X 2 , and X 3 factors on the response, respectively. The model evaluated the effects of each independent variable on the response.

Conclusions
Different types of DESs based on choline chloride and ILs based on 1-methylimidazole were used to modify Fe 3 O 4 /MIPs. The resulting nanoparticles were applied to the rapid purification of alkaloid isomers (theobromine and theophylline) from green tea by M-SPE. The M-SPE procedure was optimized by RSM to determine the best conditions. The materials were characterized by FT-IR spectroscopy and FE-SEM. Compared to the ILs-Fe 3 O 4 /MIPs, the DESs-Fe 3 O 4 /MIPs were developed for stronger recognition and higher recoveries of theophylline and theobromine from green tea, particularly DES-7-Fe 3 O 4 /MIPs. With RSM, the optimal recovery conditions for theobromine and theophylline were ratio of methanol (80%) as the washing solution, methanol/HAc (8:2) eluent at pH 3, and an eluent volume of 4 mL. The practical recoveries of theobromine and theophylline in green tea were 92.27% and 87.51%, respectively, with corresponding extraction amounts of 4.87 mg·g −1 , and 5.07 mg·g −1 . Overall, the proposed approach with the high affinity of Fe 3 O 4 /MIPs might offer a novel method for the purification of complex isomers samples.