A Novel Photo-Responsive Molecularly Imprinted Silica as a Sustainable Solid-Phase Extraction Filler for Highly Selective Adsorption of Chlorogenic Acid

Abstract

Chlorogenic acid (CA) is an important natural antioxidant component and holds strong potential for health food and cosmetic applications. In this study, a silica-based photo-responsive molecular imprinting material (PMI-PDA@NH₂-SiO₂) was designed as a solid-phase extraction (SPE) adsorption filler and applied for the efficient separation of CA in Ficus carica L. Using polydopamine-modified NH₂-SiO₂ silica as the base, combined with the photo-responsive monomer 4-methacryloyloxyazobenzene (AZO-MAA), a molecular imprinting layer with photo-responsive regulation function was constructed. Under 365 nm ultraviolet light irradiation, the azobenzene group was isomerized to the cis structure, causing the imprint cavity to shrink, thereby enabling controlled release of CA, with complete desorption within 40 min, and a desorption rate of 94.33%. Importantly, the material retained 85.56% of its initial adsorption capacity and 91.38% of its original desorption efficiency after 6 consecutive adsorption/desorption cycles, confirming robust operational stability and reproducibility. PMI-PDA@NH₂-SiO₂ was applied to the extract of Ficus carica L., achieving an adsorption rate of 92.3% for CA and a desorption rate of 87.27%. Density functional theory (DFT) calculations and NOESY spectroscopic analyses revealed that CA interacted with functional monomers via hydrogen bonding and van der Waals forces. This study advances a green separation strategy for bioactive phytochemicals in complex natural matrices.

1. Introduction

Ficus carica L. belongs to the Ficus genus of the Moraceae family and has been highly valued as a resource with both medicinal and edible properties since ancient times [1]. Chlorogenic acid (CA) is an important phenolic compound in F. carica fruits, which possesses various biological activities such as antioxidant, anti-inflammatory, and antiviral properties [2]. To enable the broader application of CA in the fields of cosmetics, healthcare, and chemical engineering, it is imperative to establish an efficient strategy for its separation [3]. Current separation methods for CA primarily include membrane separation [4], polyamide column chromatography [5], macroporous adsorption resin chromatography [6], and high-performance liquid chromatography (HPLC) [7]. However, these methods suffer from poor selectivity and are associated with several practical limitations, including high operational costs, substantial solvent consumption, prolonged pretreatment times, and low separation efficiency [8].

Molecular imprinting technology (MIT) provides an efficient solution to the challenge of low selectivity in natural compound separation [9]. This technique involves initiating the copolymerization of functional monomers and cross-linkers on the surface of a support matrix in the presence of a template molecule, yielding molecularly imprinted polymers (MIPs) [10]. Endowed with pre-engineered recognition sites, MIPs have emerged as pivotal materials in separation technology, owing to their exceptional selectivity and robust binding affinity toward target analytes [9,11].For the separation of CA, a range of high-performance MIP-based materials have been developed [12,13,14]. Among these materials, surface molecularly imprinted polymers (SMIPs) are constructed by forming an imprinted layer on the surface of nano- or microscale carriers. These materials, which possess a high specific surface area, well-defined morphology, enhanced binding capacity, accelerated mass transfer kinetics, and a high density of recognition sites, all of which contribute to significantly improved separation performance [15].

Silica nanoparticles (SiO₂) are ideal MIP carriers, as their large pore volumes, high specific surface areas, and tunable mesopore sizes can enhance the adsorption properties of imprinted materials. Further functionalization with amino groups produces amino-functionalized mesoporous silica nanoparticles (NH₂-SiO₂), which not only retain high specific surface areas and mesopore capacities but also feature abundant surface amino groups that can act as direct adsorption sites or undergo further modification to boost adsorption efficiency [16]. The silica framework is inherently polar due to its dense silanol groups (-SiOH), and the introduction of strongly polar amino groups (-NH₂) further amplifies the surface polarity and chemical reactivity, facilitating more intimate interactions with polar target molecules [17,18]. In recent years, NH₂-SiO₂ has demonstrated substantial application potential across fields including drug delivery, catalysis, and functional material engineering.

Stimuli-responsive polymers (SRPs), which exhibit dynamic changes in morphology, chemical structure, and physicochemical properties in response to external cues (e.g., light, temperature, pH, and solvent polarity), have emerged as a prominent research focus in the field of molecular imprinting technology [19,20]. Light has garnered increasing attention from researchers relative to traditional stimuli, owing to its distinct advantages of being clean, non-polluting, highly controllable, and non-contact.

Photo-responsive molecularly imprinted polymers (P-MIPs) incorporate photosensitive moieties (e.g., azobenzene) that undergo cis-trans isomerization upon irradiation with specific wavelengths of light. This structural transition enables dynamic regulation of the spatial configuration and chemical microenvironment of imprinted cavities, facilitating reversible adsorption and release of target compounds and thereby significantly enhancing the recyclability of the imprinted material [21,22]. A photo-responsive functional monomer, 4-methacryloyloxyazobenzene (Azo-MAA), was successfully synthesized by conjugating azobenzene with methacryloyl chloride, capitalizing on the intrinsic photo-responsive properties of azobenzene derivatives [23].

Polydopamine (PDA) is a superior functional monomer that is formed through the self-polymerization of dopamine (DA), a molecule rich in catechol and amine functional groups, under optimized conditions. Beyond its exceptional adhesive properties, PDA can be further chemically modified to introduce additional functional moieties, thereby enhancing the density of imprinting sites and endowing the material with expanded functionality [24,25]. Thus, employing PDA and Azo-MAA as dual functional monomers for molecularly imprinted materials is anticipated to integrate the advantages of high imprinting site density and photo-responsive behavior [26,27,28], offering a novel material platform for the efficient and intelligent separation of CA.

Photo-responsive molecularly imprinted composites represent an advanced class of materials that integrate selective recognition with stimuli-regulated binding behavior. The incorporation of photo-responsive functional monomers serves as the key responsive component of the photo-responsive molecularly imprinted (PMI) layer. These materials not only achieve high-affinity capture of target analytes but also enable spatiotemporally controlled release through light irradiation, offering a sustainable and precise approach for separation processes. In this work, the developed PMI-PDA@NH₂-SiO₂ composite exhibits outstanding affinity and selectivity toward CA, allowing reversible and efficient adsorption/desorption cycling under mild photo-stimulation with minimal loss of capacity. Compared to conventional isolation methods for CA [12,13,14], this material demonstrates enhanced recognition sensitivity, accelerated binding kinetics, and remarkable cycling stability, positioning it as a promising and green alternative for the selective extraction and controllable recovery of CA from complex samples.

The elucidation of the adsorption mechanism provides a critical theoretical foundation for optimizing material design and enhancing adsorption efficiency. Density functional theory (DFT) is a quantum mechanical computational approach that uses electron density as the fundamental variable to investigate the structure of multi-electron systems, and it stands as one of the most powerful tools for studying intermolecular interactions [29,30]. In addition, two-dimensional nuclear magnetic resonance-nuclear Overhauser effect spectroscopy (2D NMR-NOESY) enables structural characterization of complex organic molecules, with its 2D correlation signals clearly resolving proton connectivity. As a robust analytical technique for characterizing intermolecular interactions, it offers experimental evidence to support the recognition mechanism.

In this study, a photo-responsive amino-functionalized silica gel molecularly imprinted material (PMI-PDA@NH₂-SiO₂) was rationally designed using NH₂-SiO₂ as the support, Azo-MAA and PDA as dual functional monomers, ethylene glycol dimethacrylate (EGDMA) as the crosslinker, azobisisobutyronitrile (AIBN) as the initiator, and dimethyl sulfoxide (DMSO) as the porogen. This material enables highly selective and reproducible separation of CA from F. carica extracts. Furthermore, PDA modification introduced additional imprinting sites to the support, and the photoisomerization property of Azo-MAA conferred upon the imprinted material the ability to rapidly and reversibly regulate adsorption and desorption processes of CA. Finally, the mechanisms underlying the adsorption and desorption of CA by PMI-PDA@NH₂-SiO₂ were elucidated through density functional theory (DFT) calculations and 2D ¹H-¹H NOESY spectroscopy.

2. Materials and Methods

2.1. Reagents and Materials

Ficus carica L. fruits were harvested from Chengshan Town, Rongcheng City, Shandong Province, China. Fresh fruits were dried at 40 °C to a constant weight, ground into powder, and passed through a 60-mesh sieve. The fine powder was stored under dry, light-proof conditions at room temperature until analysis. NH₂-SiO₂ was purchased from Zhengzhou Hecheng New Materials Technology Co., Ltd. (Zhengzhou, China). Triethylamine (99.0%) was sourced from Shanghai Saen Chemical Technology Co., Ltd. (Shanghai, China). Dopamine hydrochloride (DA, 98.0%) was provided by Shanghai Debai Biotechnology Co., Ltd. (Shanghai, China). Rutin (RU, 98.0%), epicatechin (EC, 98.0%), and chlorogenic acid (CA, 98.0%) were procured from Shanghai Yuanye Bio-Technology Co., Ltd. (Shanghai, China). Acetonitrile (99.9%) and phosphoric acid (99.0%) were acquired from Shanghai Macklin Biochemical Technology Co., Ltd. (Shanghai, China). All other analytical grade reagents were provided by Shanghai Macklin Biochemical Technology Co., Ltd. (Shanghai, China). Ultrapure water was prepared using an ultrapure water system from Research Water Purification Technology Co., Ltd. (Xiamen, China).

2.2. Synthesis of PMI-PDA@NH₂-SiO₂ MIPs

2.2.1. Preparation of AZO-MAA

First, 4.96 g of 4-hydroxyazobenzene, 4.30 g of triethylamine, and 0.15 g of N,N-dimethyl aminopyridine were dissolved in 150 mL of anhydrous acetonitrile. The solution was cooled to 0–5 °C in an ice bath. Under continuous stirring, 4.5 g of methacryloyl chloride was added dropwise. After the addition was completed, the reaction was stirred continuously at 40 °C for 24 h, then cooled to room temperature, followed by the addition of 50 mL of saturated NaOH solution. The precipitated solid was collected by suction filtration, washed successively with 2.0 mol/L HCl solution, and finally purified by recrystallization from an ethanol–water mixed solvent to afford the photo-responsive monomer AZO-MAA.

2.2.2. Preparation of PDA@NH₂-SiO₂

First, 1.0 g of dried NH₂-SiO₂ was immersed in 150 mL of 0.1 mol/L Tris-HCl buffer solution. Then, 0.25 g of DA was added to the suspension and magnetically stirred at 500 rpm until completely dissolved. The polymerization reaction was carried out at 25 °C and 150 rpm under light-shielded conditions (wrapping the reaction vessel with aluminum foil) for 12 h, allowing a uniform polydopamine film to form on the surface of NH₂-SiO₂ via dopamine self-polymerization. After the reaction, the product was collected by vacuum filtration and repeatedly washed with deionized water until the filtrate became colorless. The final product was dried at 40 °C for 24 h to yield polydopamine-modified particles, denoted as PDA@NH₂-SiO₂.

2.2.3. Preparation of PMI-PDA@NH₂-SiO₂

First, 0.1 mmol of CA and 0.4 mmol Azo-MAA were dissolved in 100 mL of DMSO and pre-assembled at 25 °C for 6 h. Next, 1.0 g of PDA@NH₂-SiO₂ and 2.0 mmol of EGDMA were added to the pre-assembled solution in sequence and sonicated until dissolved. Then, the solution was degassed under nitrogen protection for 10 min, and 0.05 mmol of AIBN was added. The entire system was continuously reacted for 24 h under a 60 °C water bath and nitrogen protection. After the reaction, the resulting product was collected and washed exhaustively with distilled water and a methanol/acetic acid mixture (9/1, v/v) to remove the template molecule and finally washed with a large amount of methanol to remove the residual acetic acid. The product was dried to obtain the target molecularly imprinted material PMI-PDA@NH₂-SiO₂. Ultimately, a non-imprinted polymer (PNI-PDA@NH₂-SiO₂) was prepared as the control material following the identical synthetic protocol except that CA was not added. The preparation process of PMI-PDA@NH₂-SiO₂ is shown in Figure 1.

2.3. Sample Characterization

The successful synthesis of the photo-responsive monomer AZO-MAA was verified using ultraviolet-visible (UV-vis) spectroscopy (Hitachi U-3900 spectrophotometer, High-Technologies Corporation, Tokyo, Japan) and electrospray ionization mass spectrometry (ESI-MS, Thermo Fisher Scientific TSQ Fortis Plus system, Thermo Fisher Scientific Inc., Waltham, MA, USA). The surface topographies of NH₂-SiO₂, PDA@NH₂-SiO₂, and PMI-PDA@NH₂-SiO₂ were visualized via scanning electron microscopy (SEM, Thermo Fisher Scientific Apreo C, Thermo Fisher Scientific Inc., Waltham, MA, USA) at an accelerating voltage of 5 kV. Fourier-transform infrared spectroscopy (FT-IR, Thermo Fisher Scientific Nicolet iS50, Thermo Fisher Scientific Inc., Waltham, MA, USA) was conducted to characterize the chemical structure of AZO-MAA and identify changes in surface functional groups of NH₂-SiO₂, PDA@NH₂-SiO₂, and PMI-PDA@NH₂-SiO₂.

2.4. HPLC Analysis

Quantification was performed using the external standard method. The quantification of sample and standard was conducted using high-performance liquid chromatography (HPLC, Agilent 1260, Agilent Technologies, Inc., Santa Clara, CA, USA) equipped with a HiQ Sil C18 column (4.6 mm× 250 mm, 5 μm). The sample solution consisted of a methanol extract of Ficus carica L., and the HPLC mode employed was isocratic elution. The mobile phase for HPLC consisted of a 30% A phase (0.2% aqueous phosphoric acid) and a 70% B phase (acetonitrile). The flow rate was maintained at 1.0 mL/min, the column temperature was set to 30 °C, the detection wavelength was fixed at 275 nm, and the injection volume was 20 μL.

2.5. Photo-Controlled Adsorption/Desorption Experiments

This experiment investigates the photo-controlled adsorption/desorption performance of PMI-PDA@NH₂-SiO₂. First, 20 mg of PMI-PDA@NH₂-SiO₂ was added to 20 mL of a 0.1 mg/mL CA standard solution. The mixture was then placed in a constant temperature shaker under dark conditions at 25 °C and 100 rpm until adsorption equilibrium was reached. The initial concentration of CA (C₀) in the supernatant was determined by HPLC. Subsequently, the sample was alternately exposed to 365 nm ultraviolet light and natural light for 1 h each time. After each light exposure, the supernatant was centrifuged (5000 rpm) and filtered (0.22 μm filter membrane), and the CA concentration in the supernatant was measured by HPLC. The binding rate (D%) of PMI-PDA@NH₂-SiO₂ to CA was calculated according to Equation (1). This process was repeated for 6 cycles of alternating ultraviolet/natural light exposure. The CA concentration in each cycle was determined. All experiments were conducted in triplicate.

$$D\%=(1-{\displaystyle \frac{C_e}{C_0}}) \times 100\%$$

(1)

where C₀ (mg/L) represents the initial concentration of CA during the adsorption process, and C_e (mg/L) represents the equilibrium concentration of CA.

2.6. Rebinding Experiment

2.6.1. Adsorption Kinetics

First, 0.1 g of PMI-PDA@NH₂-SiO₂ was dispersed in 50 mL of 0.1 mg mL CA standard solution and incubated in a thermostatic shaker (150 rpm, 25 °C) to initiate the adsorption process. Aliquots (2 mL) were withdrawn at predefined time intervals (1, 2, 3, 4, 5, 10, 15, 30, 60, 90, and 120 min) within 0–120 min. Each sample was immediately centrifuged at 8000 rpm for 5 min and filtered through a 0.22 μm membrane to remove solid particles. The residual CA concentration in the supernatant was quantified via HPLC. The adsorption capacity at time t (Q_t, mg/g) was calculated using Equation (2):

$$Q_e = {\displaystyle \frac{(C_0-C_t)\mathrm{V}}{m}}$$

(2)

where C₀ (mg/L) and C_t (mg/L) represent the initial and real-time concentration of CA, V (L) represents the volumes of CA solutions, and m (g) represents the mass of the adsorbent.

To elucidate the adsorption mechanism, the kinetic data were fitted to the pseudo-first-order kinetic model Equation (3) and pseudo-second-order kinetic model Equation (4). The equations of models are as follows:

$$\ln (Q_e-Q_t)=\ln Q_e-k_{1}t$$

(3)

$${\displaystyle \frac{t}{Q_t}}={\displaystyle \frac{1}{k_2{Q}_e^2}}+{\displaystyle \frac{t}{Q_e}}$$

(4)

where Q_e (mg/g) and Q_t (mg/g) represent the adsorption capacity at adsorption equilibrium and at time t (min), respectively. k₁ (min⁻¹) and k₂ (g/mg/min) represent the constants in Equations (3) and (4).

2.6.2. Adsorption Isotherms

First, 20 mg of PMI-PDA@NH₂-SiO₂ was added to 5 mL of CA standard solutions with different concentrations (0.02, 0.04, 0.06, 0.08, 0.10, 0.12, 0.14, 0.16, 0.18, and 0.2 mg/mL), followed by shaking at 200 rpm for 60 min. Subsequently, the concentration of CA was determined by HPLC. The performance and adsorption behavior of PMI-PDA@NH₂-SiO₂ were evaluated and predicted by Langmuir and Freundlich isothermal adsorption models. The Langmuir model and Freundlich model are represented by Equations (5) and (6):

$$Q_e = {\displaystyle \frac{Q_m{K}_L{C}_e}{1 + K_L{C}_e}}$$

(5)

$$\ln Q_e=\ln K_F+{\displaystyle \frac{1}{\mathrm{n}}}\ln C_e$$

(6)

where Q_e (mg/g) and C_e (mg/mL) represent the adsorption capacity of the material and the equilibrium concentration of CA, respectively; Q_m (mg/g) represents the maximum theoretical adsorption capacity; n stands for the empirical constant; and K_F and K_L denote the Freundlich constant and Langmuir constant respectively [31].

2.6.3. Adsorption Selectivity

To evaluate the adsorption selectivity of PMI-PDA@NH₂-SiO₂ for CA, rutin (RU) and epicatechin (EC) were used as competitive adsorbates. The chemical structures of RU and EC are provided in Figure S2. A mixed solution, with an initial concentration of 0.1 mg/mL for CA, RU, and EC, was prepared. First, 20 mg of PMI-PDA@NH₂-SiO₂ was added to the mixed solution, and the adsorption experiment was carried out in a constant-temperature shaker for 60 min. The concentrations of CA, RU, and EC in the solution were determined by HPLC. The imprinting factor (IF), distribution coefficient (K_d), selectivity coefficient (k^sel), and relative selectivity coefficient (k^rel) of PMI-PDA@NH₂-SiO₂ were expressed by Equations (7)–(10).

$$IF ={\displaystyle \frac{Q_{PMI\text{-}PDA@{NH}_2\text{-}Si{O}_2}}{Q_{PNI\text{-}PDA@{NH}_2\text{-}Si{O}_2}}}$$

(7)

where $Q_{PMI\text{-}PDA@{NH}_2\text{-}Si{O}_2}$ (mg/g) and $Q_{PNI\text{-}PDA@{NH}_2\text{-}Si{O}_2}$ (mg/g) are the equilibrium adsorption capacities of PMI-PDA@NH₂-SiO₂ (molecularly imprinted polymer, MIP) and PNI-PDA@NH₂-SiO₂ (non-imprinted polymer, NIP) for CA, respectively.

$$K_{ d}={\displaystyle \frac{Q_e}{C_e}}$$

(8)

$$k^{sel}={\displaystyle \frac{K_{d, template molecule}}{K_{d, referent molecule}}}$$

(9)

$$k^{rel}={\displaystyle \frac{k_{MIPs}^{sel}}{k_{NIPs}^{sel}}}$$

(10)

where $K_d$ (mL/g) is the distribution coefficient, Q_e (mg/g) is the equilibrium adsorption capacity of PMI-PDA@NH₂-SiO₂ for the target compound or analog, and C_e (mg/mL) is the equilibrium concentration of the solute in the solution, K_{d, template molecule} is the partition coefficient of the template molecule; K_{d, referent molecule} is the partition coefficient of the competitor; k^sel_MIPs is the selectivity coefficient of MIPs; and k^sel_NIPs is the selectivity coefficient of NIPs [32,33].

2.7. Reusable Performance Test

To evaluate the reusability of the prepared PMI-PDA@NH₂-SiO₂ for CA, a cyclic adsorption/desorption experiment was conducted. The specific steps were as follows: 20 mg of PMI-PDA@NH₂-SiO₂ was added to a 0.1 mg/mL CA standard solution, and the adsorption experiment was carried out in a constant-temperature shaker at 25 °C and 100 rpm until adsorption equilibrium was reached. Subsequently, the adsorption-saturated PMI-PDA@NH₂-SiO₂ was separated from the solution and desorption was performed under 365 nm ultraviolet light until desorption equilibrium was achieved. After desorption, the resin was washed with deionized water, dried, and reused. The content of CA in each cycle was determined, and the adsorption capacity and desorption rate of each adsorption/desorption process were calculated to evaluate the recyclability of PMI-PDA@NH₂-SiO₂.

2.8. Recognition Mechanism

2.8.1. DFT Simulation

The interaction mode, charge information, and interaction force of PDA, Azo-MAA, and CA systems were studied by DFT simulation, and the recognition mechanism of PMI-PDA@NH₂-SiO₂ was clarified. All calculations were performed using Gaussian 16 software. Firstly, the geometric structures of PDA, Azo-MAA, and CA and their complexes were optimized to reach the stable configuration with the lowest energy, and the 3D visualization structures were verified by Gauss View 6.0. The intermolecular charge transfer characteristics were investigated based on Mulliken atomic charge distribution, and the donor–acceptor regions were characterized by electrostatic potential mapping. The energy gap values of the system were calculated using the frontier molecular orbital theory (HOMO-LUMO) to evaluate the reactivity and stability of the system. Moreover, non-covalent interactions (NCIs) in the system were characterized using electron density analysis, and RDG isosurfaces generated with Multiwfn 3.7 were visualized in real space to reveal weak intermolecular interactions.

2.8.2. NMR Measurements

2D nuclear magnetic resonance (2D-NMR) is a key technique in modern structural analysis, capable of precisely assigning spectral peaks for complex organic compounds. In the 2D ¹H-¹H NOESY spectrum, when the spatial distance between two protons is less than 5 Å, corresponding cross-peaks will appear in the spectrum, and this signal can be used to infer the stereostructure and spatial conformation of the molecule. In this study, the interaction between the functional monomer DA and the template CA was analyzed using the 2D ¹H-¹H NOESY spectrum. DA and the template CA were mixed at a molar ratio of 1:1. Then, the mixture was dissolved in DMSO-d6 for NMR characterization. The NMR measurements were carried out using a nuclear magnetic resonance spectrometer (NMR, Qone AS400, Wuhan Zhongke Niujin Wave Spectrum Technology Co., Ltd., Wuhan, China).

2.9. Purification of CA from Ficus carica L. Using PMI-PDA@NH₂-SiO₂ SPE Column

Using PMI-PDA@NH₂-SiO₂ as the solid-phase extraction (SPE) column packing material, the filling volume was controlled at approximately 2/3 of the column volume. After completing the column packing, 30 mL of the recovered F. carica L. extract was loaded onto the solid-phase extraction column at a flow rate of 2 BV/h. The column was allowed to stand overnight, and the eluate I was collected. Then, a mixed elution solution composed of methanol and acetic acid in a ratio of 9:1 (v/v) was prepared. This elution solution was used to desorb CA from the stationary phase, with a volume controlled with a volume of 4 BV/h. Eluate II was collected and its volume was recorded. Finally, eluate II was subjected to rotary evaporation in an oil bath to concentrate the CA content by evaporating the solvent. Following rotary evaporation, the crude product (m_eluate) was accurately weighed. The adsorption recovery rate (R_ads) of CA by the PMI-PDA@NH₂-SiO₂ extraction column, the desorption recovery rate (R_des) of CA in the final elution solution, and the purity of the obtained CA(P_CA) are calculated using the following Equations (11)–(13):

$$R_{ads} = {\displaystyle \frac{\left(C_{0 }-{ C}_I\right) \times V_0}{C_0 \times V_0}}\times 100\%$$

(11)

where R_ads (%) represents the adsorption efficiency of PMI-PDA@NH₂-SiO₂ extraction column for CA, C₀ (mg/mL) is the concentration of CA in the extract solution, C_I (mg/mL) is the concentration of CA in effluent I, and $V_0$ (mL) is the volume of the extract solution.

$$R_{des }={\displaystyle \frac{C_{II} \times V_{II}}{\left(C_0\text{-} C_I\right)\times V_0}}\times 100\%$$

(12)

where R_des (%) represents the desorption efficiency of PMI-PDA@NH₂-SiO₂ extraction column for CA, C_II (mg/mL) represents the concentration of CA in the elution solution II, and V_II (mL) denotes the volume of the eluent II.

$$P_{CA}={\displaystyle \frac{C_{II}\times V_{II}}{m_{eluate}}}\times 100\%$$

(13)

where P_CA (%) represents the purity of the obtained CA, and m_eluate (mg) refers to the total mass after the rotary evaporation of eluate II.

3. Results and Discussion

3.1. Characterization of AZO-MAA

3.1.1. UV-Vis Spectroscopic Analysis of Azo-MAA

The photo-responsive behavior of Azo-MAA was characterized by UV-Vis absorption spectroscopy, with the results presented in Figure 2. In the absence of UV irradiation, Azo-MAA displayed a prominent absorption peak at 340 nm, which is ascribed to the π→π* transition of the trans-Azo group [34]. Upon exposure to UV light, the trans-Azo underwent photoisomerization to the cis configuration via light energy absorption. As the UV irradiation duration increased, the relative abundance of the trans-Azo moiety decreased, leading to a gradual attenuation of the absorption peak intensity. When the irradiation time reached 20 min, the system reached a photostationary state. Further extension of the irradiation time did not induce significant changes in the peak shape or position of the absorption curve.

Notably, the spectral data revealed that UV irradiation only altered the absorption peak intensity without generating new peaks or causing peak disappearance. This observation confirms that UV light exclusively triggered the photoisomerization of the Azo group rather than inducing side reactions such as photocrosslinking or photodegradation.

3.1.2. FT-IR Spectroscopy of Azo-MAA

To confirm the chemical structure of Azo-MAA and the successful incorporation of target functional groups, the as-synthesized Azo-MAA was characterized by FT-IR spectroscopy, with the results presented in Figure 3. The C-O bond of the carboxyl group on MAA disappeared at 1305 cm⁻¹. A distinct peak appeared at 1008 and 1058 cm⁻¹, corresponding to the asymmetric stretching vibration of the ether bond (C-O-C) in the ester group. This confirmed that an esterification reaction occurred between MAA and Azo. Meanwhile, the stretching vibration peak of the hydroxyl group (-OH) at 3500 cm⁻¹ was significantly attenuated. This phenomenon is primarily ascribed to the consumption of hydroxyl groups during covalent bond formation. The disappearance of the C-O peak, the appearance of the C-O-C peak, and the attenuation of the -OH peak indicated the successful synthesis of Azo-MAA [31]. Additionally, the C–H stretching vibration peaks of methylene (–CH₂–) and methyl (–CH₃) were observed at 2930 and 2948 cm⁻¹, respectively, together with the aromatic C=C stretching at 1572 cm⁻¹ and the –N=N– stretching at 1450 cm⁻¹, further confirming the successful synthesis of Azo-MAA [35].

3.1.3. LC-MS Analysis of Azo-MAA

Liquid chromatography mass spectrometry (LC-MS) was employed for the qualitative characterization of Azo-MAA, with the results presented in Figure 4. In the negative ion mode ([M-H]⁻), a distinct quasi-molecular ion peak was detected at a mass-to-charge ratio (m/z) of 265.26. This observed m/z value is in excellent agreement with the theoretical molecular weight of the deprotonated Azo-MAA ([M-H]⁻), which confirms the successful synthesis of Azo-MAA. Further structural elucidation of fragmentation patterns was achieved through MS/MS analysis, as detailed in Figure S1.

3.2. Characterization of PMI-PDA@NH₂-SiO₂

The surface morphology of the materials was characterized via scanning electron microscopy (SEM), with the results presented in Figure 5. The unmodified NH₂-SiO₂ appeared as irregular white powder particles with a relatively smooth surface, dense structure, and high specific surface area (Figure 5A) [32]. After PDA modification, PDA@NH₂-SiO₂ transformed into irregular black powder particles with a significantly increased surface roughness (Figure 5B). This morphological transformation confirms the successful deposition of PDA onto the NH₂-SiO₂ surface.

The final PMI-PDA@NH₂-SiO₂ exhibited a dark brown appearance (Figure 5C). While retaining the intrinsic framework of NH₂-SiO₂, its surface roughness was further enhanced, with granular substances observed locally adhering between the layered structures. Collectively, the SEM results indicate that the surface morphology of the NH₂-SiO₂ support evolved from smooth and dense to a rougher texture after stepwise synthesis and modification. The observed surface irregularities are consistent with the presence of template-specific cavities typically formed during molecular imprinting, which are considered essential for selective molecular recognition.

The stepwise preparation process of PMI-PDA@NH₂-SiO₂ was characterized by tracking the infrared characteristic absorption peaks of key functional groups during the preparation process. The corresponding FT-IR spectra are presented in Figure 6. NH₂-SiO₂ showed a peak at 2953 cm⁻¹, attributed to the stretching vibration of aliphatic C-H; a broad peak in the range of 3400–3100 cm⁻¹ was attributed to the stretching vibrations of N-H and O-H; and an asymmetric stretching vibration peak of Si-O-Si at 1050 cm⁻¹ was consistent with the reported framework [36]. After modification with PDA, PDA@NH₂-SiO₂ exhibited a broader peak in the range of 3400–3100 cm⁻¹, attributed to the superposition of the stretching vibrations of O-H and N-H of PDA. Simultaneously, a bending vibration peak of the benzene ring C=C at 1503 cm⁻¹ confirmed the successful anchoring of PDA [37]. For PMI-PDA@NH₂-SiO₂, the introduction of Azo-MAA was confirmed by the C=O stretching vibration peak at 1730 cm⁻¹ and the N=N stretching vibration peak at 1590 cm⁻¹. Additionally, the broadening and intensification of the absorption peak at 1064 cm⁻¹ might be the result of the superposition of the stretching vibration peak of Si-O-Si and the stretching vibration peak of the ether bond (C-O-C) in Azo-MAA, further confirming the successful introduction of the Azo-MAA [31]. The broad peak in the range of 3400–3100 cm⁻¹ significantly weakened after the formation of the molecularly imprinted cavity, indicating that the phenolic hydroxyl and amino groups on the PDA surface as active sites might have participated in the imprinting polymerization process, thereby providing crucial spectroscopic evidence for the successful in situ polymerization of the molecularly imprinted layer on the PDA surface.

3.3. Results of the Photo-Controlled Adsorption/Desorption Experiments

To evaluate the photo-regulated adsorption/desorption performance of PMI-PDA@NH₂-SiO₂, photo-controlled adsorption/desorption curves (Figure 7) were constructed based on the variation of binding rates of PMI-PDA@NH₂-SiO₂ and NH₂-SiO₂ toward CA with irradiation time under different wavelengths. At adsorption equilibrium, the adsorption rate of PMI-PDA@NH₂-SiO₂ for CA reached up to 93.5%. The total adsorption capacity of NH₂-SiO₂ for CA was 60.8%, indicating that 32.7% of CA was adsorbed via the specific recognition sites of the molecularly imprinted material. Under 365 nm UV light irradiation, 84.2% of CA was released from the imprinted material into the solution. Upon switching to natural light, 80.0% of CA was re-adsorbed. In contrast, the binding capacity of NH₂-SiO₂ for CA decreased from 60.8% to 14.14% under alternating light irradiation.

The observed light-controlled adsorption/desorption behavior is primarily attributed to the cis-trans isomerization of the Azo chromophore. Isomerization of Azo induces reversible conformational changes in the specific molecularly imprinted cavities. Owing to the sufficient free space within the material and the higher energy required for Azo isomerization compared to the interaction energy between the template molecule and receptor sites, CA is driven to release from the imprinted sites [38]. These results confirm that PMI-PDA@NH₂-SiO₂ exhibits significantly superior light-regulated adsorption/desorption performance relative to NH₂-SiO₂.

After five cycles of cis-trans isomerization, PMI-PDA@NH₂-SiO₂ retained a high adsorption capacity for CA, demonstrating excellent reversibility and cycling stability of the material’s specific receptor sites and imprinting affinity. No significant attenuation of specific recognition ability was observed during multiple light-controlled cycles.

3.4. Results of the Adsorption Experiments

3.4.1. Adsorption Kinetics

To investigate the adsorption behavior of PMI-PDA@NH₂-SiO₂ and PNI-PDA@NH₂-SiO₂ toward CA, adsorption kinetic experiments were performed. The adsorption kinetic results are presented in Figure 8. The adsorption capacity of PNI-PDA@NH₂-SiO₂ for CA increased rapidly within the first 10 min, followed by a slower rate of increase, and stabilized at 20 min. In contrast, the adsorption capacity of PMI-PDA@NH₂-SiO₂ for CA increased more rapidly within the first 20 min and reached equilibrium at 30 min. Although both silica-based materials exhibited similar trends in adsorption capacity, their saturated adsorption capacities differed significantly: the saturated adsorption capacity of PMI-PDA@NH₂-SiO₂ was approximately twice that of PNI-PDA@NH₂-SiO₂. This discrepancy is attributed to the abundant molecularly imprinted cavities on the surface of PMI-PDA@NH₂-SiO₂, which are structurally matched to CA, thereby conferring superior adsorption performance for CA.

The kinetic fitting results are summarized in

Table 1. Pseudo-first-order and pseudo-second-order adsorption kinetics model fitting parameters of PMI-PDA@NH₂-SiO₂ and PNI-PDA@NH₂-SiO₂.

Sample	Pseudo-First-Order			Pseudo-Second-Order
	Qe1 (mg/g)	k1 (min−1)	R2	Qe2 (mg/g)	k2 (g/mg/min)	R2
PMI-PDA@NH2-SiO2	28.85 ± 0.75	0.195 ± 0.030	0.949	31.14 ± 1.58	0.008 ± 0.003	0.877
PNI-PDA@NH2-SiO2	13.72 ± 0.59	0.236 ± 0.032	0.952	15.13 ± 0.96	0.019 ± 0.005	0.918

Note: The reported values represent the mean ± standard deviation from three independent determinations (n = 3).

. The pseudo-first-order model (R² = 0.949) demonstrated a higher correlation coefficient than the pseudo-second-order model (R² = 0.877), and the theoretical equilibrium adsorption capacity (Q_e1 = 28.85 mg/g) was close to the experimental value. These results suggest that adsorption occurs mainly on the surface, aligning with the surface imprinting feature of PMI-PDA@NH₂-SiO₂. This is because the surface-located recognition sites promote rapid diffusion of the target to the active sites. This further confirms the affinity and selectivity of PMI-PDA@NH₂-SiO₂ for CA [39].

3.4.2. Adsorption Isotherms

The adsorption capacities of PMI-PDA@NH₂-SiO₂ and PNI-PDA@NH₂-SiO₂ in CA solutions of varying concentrations were evaluated through adsorption isotherm experiments, and the adsorption data were fitted and analyzed using the Langmuir and Freundlich models. As shown in Figure 9 and

Table 2. The fitting parameters of the Langmuir and Freundlich models of PMI-PDA@NH₂-SiO₂ and PNI-PDA@NH₂-SiO₂.

Types	Langmuir Model			Freundlich Model
	Qm (mg/g)	KL (mL/mg)	R2	KF (mg1 − n · mLn · g−1)	1/n	R2
PMI-PDA@NH2-SiO2	35.90 ± 3.59	52.770 ± 33.46	0.992	127.628 ± 14.63	0.797 ± 0.035	0.961
PNI-PDA@NH2-SiO2	16.43 ± 1.27	18.561 ± 4.16	0.908	25.223 ± 3.70	0.392 ± 0.066	0.833

Note: The reported values represent the mean ± standard deviation from three independent determinations (n = 3).

, the results indicate that the adsorption capacities of both PMI-PDA@NH₂-SiO₂ and PNI-PDA@NH₂-SiO₂ increased with rising initial CA concentration, reaching saturation at approximately 0.10 mg/mL. Notably, PMI-PDA@NH₂-SiO₂ exhibited significantly superior adsorption performance compared to PNI-PDA@NH₂-SiO₂, with a maximum adsorption capacity of 35.90 mg/g, approximately twice that of PNI-PDA@NH₂-SiO₂. This enhanced performance suggests that PMI-PDA@NH₂-SiO₂ possesses specific recognition sites capable of forming interactions with CA molecules, resulting in higher adsorption capacity and selectivity [40]. The Langmuir model provided a better fit for PMI-PDA@NH₂-SiO₂ (R² = 0.992), indicating that CA adsorption on PMIP-PDA@NH₂-SiO₂ primarily occurs as monolayer adsorption, with a uniform surface and consistent adsorption site sizes.

3.4.3. Adsorption Selectivity

To further assess the selective adsorption capability of PMI-PDA@NH₂-SiO₂ toward chlorogenic acid (CA) and its structural analogues (epicatechin, EC; rutin, RU), selective adsorption experiments were performed and the IF values were calculated. The results are presented in Figure 10 and

Table 3. K_d, k^sel, and k^rel of PMI-PDA@NH₂-SiO₂ and PNI-PDA@NH₂-SiO₂ for CA, EC, and RU.

Analytes	${\mathit{K}}_{\mathit{d}, \mathit{P}\mathit{M}\mathit{I}\mathit{s}}$ (g/L)	${\mathit{k}}_{\mathit{PMIs}}^{\mathit{sel}}$	${\mathit{K}}_{\mathit{d}, \mathit{P}\mathit{N}\mathit{I}\mathit{s}}$ (g/L)	${\mathit{k}}_{\mathit{PNIs}}^{\mathit{sel}}$	krel
CA	0.453 ± 0.057	-	0.093 ± 0.025	-	-
EC	0.089 ± 0.011	5.172 ± 1.125	0.040 ± 0.009	2.424 ± 0.667	2.162 ± 0.161
RU	0.072 ± 0.014	6.279 ± 0.395	0.066 ± 0.012	1.420 ± 0.144	4.470 ± 0.456

Note: The reported values represent the mean ± standard deviation from three independent determinations (n = 3).

. The adsorption capacities of PMI-PDA@NH₂-SiO₂ for CA, EC, and RU were 21.18 mg/g, 7.27 mg/g, and 6.08 mg/g, respectively, with corresponding IF values of 2.84, 2.05, and 1.09. The selectivity factors for CA relative to EC and RU were 5.08 and 6.29, respectively, indicating that the adsorption capacity of PMI-PDA@NH₂-SiO₂ for CA was significantly higher than that for its structural analogs. This enhanced selectivity is attributed to the presence of specific imprinted sites on the surface of PMI-PDA@NH₂-SiO₂, which are precisely matched to the molecular structure of CA. In contrast, PNI-PDA@NH₂-SiO₂, which lacks specific recognition cavities, exhibited substantially lower adsorption capacities compared to PMI-PDA@NH₂-SiO₂.

The adsorption mechanism of PMI-PDA@NH₂-SiO₂ for CA is illustrated in Figure 11. During the molecular imprinting process, CA binds to PDA to form cavity structures with memory functionality. After CA removal, these cavities retain the spatial configuration features of CA, analogous to the “key-and-lock” specific recognition principle [41]. Thus, when CA re-interacts with PMI-PDA@NH₂-SiO₂, it can precisely fit into these cavities and exhibit strong adsorption.

3.4.4. Recyclability Experiments

Six consecutive adsorption/desorption cycles of CA were performed using PMI-PDA@NH₂-SiO₂, and the results are presented in Figure 12. The experimental data reveal that both the adsorption capacity and desorption efficiency of PMI-PDA@NH₂-SiO₂ toward CA exhibit a gradual downward trend with increasing cycle numbers. After six cycles, the adsorption capacity of PMI-PDA@NH₂-SiO₂ still retains 85.56% of its initial maximum capacity, while the desorption efficiency remains at 91.38% of its initial level. These findings confirm that PMI-PDA@NH₂-SiO₂ can be effectively recycled for at least six cycles to achieve efficient separation of CA, indicating that the material not only possesses excellent adsorption performance but also maintains stable separation capability over multiple cycles.

3.5. Analysis of the Recognition Mechanism

3.5.1. Mulliken Atomic Charge Distribution

The optimized structure of the interaction system between the template molecule (CA) and the functional monomer (DA) is depicted in Figure 13A. The mechanism underlying the intermolecular interactions between CA and DA was elucidated via Mulliken charge distribution analysis. As illustrated in Figure 13A, the 50th oxygen (O) atom in CA bears a pronounced negative charge, acting as an electron acceptor, whereas the 22nd hydrogen (H) atom in DA carries a distinct positive charge, functioning as an electron donor. Concurrently, the optimized structure of the interaction system between CA and Azo-MAA is presented in Figure 13B. The 69th O atom in CA exhibits a significant negative charge (–0.527), serving as an electron acceptor, while the 30th H atom in Azo-MAA possesses a notable positive charge (0.440), acting as an electron donor. This complementary charge distribution suggests that intermolecular hydrogen bonds (O50···H22, O69···H30) can be formed between the two molecules via electrostatic interactions. Electrostatic potential analysis further validates the prominent charge polarization effect within the system. The electron-rich regions surrounding oxygen atoms in CA and the electron-deficient regions around hydrogen atoms in the two functional monomers exhibit a clear spatial correspondence.

3.5.2. IGMH Analysis

The independent gradient model based on Hirshfeld partitioning (IGMH) was employed to probe the non-covalent interaction mechanisms within the CA-PDA and CA-Azo-MAA systems [42]. By computing the electron density distribution and its gradient variations, IGMH delineates the spatial distribution and intensity of intermolecular interactions, including hydrogen bonds, van der Waals forces, and repulsive forces. As illustrated in Figure 14, in the IGMH isosurface plots of the CA-PDA and CA-Azo-MAA systems, blue regions denote attractive interactions, green regions correspond to van der Waals interactions, and red regions signify repulsive interactions. Notably, blue and green isosurface features were observed between the C=O group of CA and the O-H groups of DA and Azo-MAA, indicating the existence of strong hydrogen bonds and van der Waals forces. This thus validates the selective binding mechanism between CA and DA/Azo-MAA mediated by hydrogen bonds and van der Waals forces [43].

3.5.3. Molecular Orbital Theory

Based on molecular orbital theory, the reactivity and stability of the CA-DA system were analyzed. The frontier molecular orbitals (FMOs) of the CA-DA system are depicted in Figure 15A. The highest occupied molecular orbital (HOMO) and the lowest unoccupied molecular orbital (LUMO) energies of the CA-DA system are –5.2197 eV and –1.8806 eV, respectively, corresponding to a HOMO–LUMO energy gap of 3.3391 eV. The frontier molecular orbitals (FMOs) of the CA-Azo-MAA system are displayed in Figure 15B. For CA-Azo-MAA, the HOMO and LUMO energies are –5.7216 eV and –2.4481 eV, respectively, giving a HOMO–LUMO gap (ΔE) of 3.2734 eV, which reflects a relatively small separation between the occupied and unoccupied orbitals. A narrower energy gap is generally associated with a higher probability of electronic transitions and enhanced electron transport properties [44,45], further corroborating the excellent stability of CA-PDA and CA-Azo-MAA systems.

3.5.4. NMR Analysis

Nuclear magnetic resonance (NMR) spectroscopy, particularly ¹H NMR and 2D NOESY, serves as a robust tool for investigating intermolecular interactions [46]. In this study, a 1:1 molar ratio of CA to DA was employed to probe their molecular interactions. The ¹H NMR spectrum of the CA-DA complex is presented in Figure 16A, and corresponding proton assignments are presented in the Supporting Information (Tables S1 and S2). Comparative analysis against the spectra of the pure components revealed significant chemical shift perturbations for the 3-OH and 4-OH protons of DA, as well as the 1-OH, 2-OH, NH₂, 2′-OH, and 3′-OH protons of CA.

2D NOESY spectral analysis of the complex further elucidated its binding mode (Figure 16B). Distinct cross-peaks were observed between δ 9.21 and 8.85 and between δ 9.64 and 8.85, indicating spatial proximity between the aromatic ring of DA and the phenolic hydroxyl groups of CA, attributed to π-π stacking interactions. Additionally, the cross-peak at δ 7.92 and δ 12.41 confirmed hydrogen bond formation between the amino group of DA and the carboxyl proton of CA.

Other hydrogen bond interactions were validated by cross-peaks between δ 9.21/δ 9.64/δ 4.80/δ 5.55 and δ 7.92, corresponding to interactions between the hydroxyl groups of CA and the amino group of DA. Furthermore, the cross-peak between δ 5.55/δ 4.80 and δ 8.90 suggested the formation of a hydrogen bond network between the hydroxyl groups of CA and the aromatic hydroxyl groups of DA. Collectively, these findings indicate a potential synergistic interplay between hydrogen bonding and van der Waals forces in the CA-DA interaction.

3.6. Comparison of PMI-PDA@NH₂-SiO₂ with Other Adsorbents

The prepared adsorbent was systematically compared with other molecularly imprinted materials reported in the literature for CA separation, and the results are shown in

Table 4. Comparison with previously reported molecularly imprinted materials for CA.

Adsorbents	Adsorption Capacity (mg/g)	Imprinting Factor (IF)	Synthetic Agents	Recycle Time	References
CGA-MIPC	14.72	9.68	4-VP, EDMA, AIBN, [BMIM]BF4–DMSO	-	[13]
3D-MMIPs	10.88	3.28	M-GO-MWNTs, MAA, EGDMA, AIBN	5	[47]
Fe3O4-Cu@Lyz-MIPs	10.82	2.85	Fe3O4-Cu NPs, PT Lyz	10	[12]
PMIP-PDA@NH2-SiO2	28.85	2.84	AZO-MAA, EGDMA, AIBN, NH2-SiO2	6	This work

. Specifically, 3D-MMIPs use M-GO-MWNTs as a carrier, significantly increasing the specific surface area and improving selectivity. Unlike 3D-MMIPs that require repeated strong-acid elution, our PMI-PDA@NH₂-SiO₂ material eliminates this step and is therefore more sustainable. Fe₃O₄-Cu@Lyz-MIPs employ phase-transformed lysozyme as the MIP imprinting shell, exhibiting good acid resistance and structural stability. CGA-MIP uses [BMIM]BF₄–DMSO as a porogen, enabling a distinct imprinting effect even at a low template-to-functional monomer molar ratio of 1:240. Nevertheless, none of these materials possess photo-responsive properties. In contrast, the material developed in this work achieves adsorption and desorption through a photo-responsive mechanism, avoiding the use of large amounts of organic solvents required in conventional elution methods. Moreover, the light-controlled process does not require direct contact with the material, operates under mild conditions, and exhibits excellent recyclability, offering clear advantages in green separation and sustainability.

3.7. Purification of CA from Ficus carica L. Using PMI-PDA@NH₂-SiO₂ SPE Column

The as-prepared PMI-PDA@NH₂-SiO₂ was packed into a solid-phase extraction (SPE) column for the selective separation and purification of CA from F. carica L. extracts, with the setup illustrated in Figure 17A. The SPE procedure was carried out as follows. First, PMI-PDA@NH₂-SiO₂ was loaded into the SPE column, followed by the introduction of the F. carica L. extract sample for dynamic adsorption. Once PMI-PDA@NH₂-SiO₂ reached adsorption saturation, the adsorbed CA was eluted using an appropriate eluent. The elution mechanism involves the competition between free H⁺ ions in acetic acid and the -OH groups of CA. This competitive interaction disrupts the hydrogen bonding between CA and PMI-PDA@NH₂-SiO₂, thereby facilitating the release of CA.

HPLC chromatograms of the F. carica L. extract before adsorption and the eluent after adsorption are presented in Figure 17B and Figure 17C, respectively, with quantitative analysis performed via HPLC. In the HPLC profile of the F. carica L. extract, a characteristic peak corresponding to CA was observed at a retention time of 3.58 min. The PMI-PDA@NH₂-SiO₂-packed SPE column exhibited an adsorption rate (R_ads) of 92.3%, a desorption rate (R_des) of 87.27%, a recovery rate of 80.55%, and a CA purity (P_CA) of 82.62% in the eluent. This confirms the outstanding adsorption performance of PMI-PDA@NH₂-SiO₂ toward CA. These results demonstrate that PMI-PDA@NH₂-SiO₂ can serve as an efficient SPE packing filler for the separation of CA from F. carica L. fruits.

3.8. Greenness Assessment and AGREEMIP Analysis

AGREEMIP is a measurement tool and software for evaluating and comparing the environmental sustainability of the MIP synthesis process [48].This study assesses the greenness of the PMI-PDA@NH₂-SiO₂ molecular imprinting process using the AGREEMIP software (Version 1.0) metric. Each of its 12 principles is scored based on experimental data, normalized, and weighted to obtain an overall greenness value. As seen in Figure S3, PMI-PDA@NH₂-SiO₂ can be categorized as good greenness based on the final score of 0.76. As shown in Table S3, the favorable greenness performance is mainly attributed to several high-scoring criteria. In particular, the removal of polymerization inhibitors, template usage, porogen/solvent selection, other reagents/adjuvants/carriers, and core/particle preparation or surface modification all received scores of 1.0, indicating low environmental burden in these aspects. The template elution solvent also showed a high score of 0.92, suggesting relatively limited solvent-related hazard during template removal. In addition, the functional monomer and cross-linking agent obtained moderate-to-high scores of 0.65 and 0.62, respectively, further supporting the overall acceptable sustainability of the synthesis route. Although some steps, such as polymerization initiation, particle size control, template elution technique, and final product reusability, showed comparatively lower scores, their influence did not prevent the process from achieving an overall good greenness rating.

Therefore, the AGREEMIP evaluation demonstrates that the PMI-PDA@NH₂-SiO₂ molecular imprinting process has favorable environmental sustainability, especially in terms of reagent consumption, solvent use, and template elution solvent safety.

4. Conclusions

A photo-responsive molecularly imprinted polymer (PMI-PDA@NH₂-SiO₂) was successfully prepared. In this system, DA served a dual function by self-polymerizing on the NH₂-SiO₂ surface to form a PDA layer and acting as a green functional monomer to facilitate the formation of the imprinted layer. This dual functionality thereby increased adsorption sites and conferred excellent selectivity and affinity toward the template molecule CA. Additionally, Azo-MAA was incorporated as a photo-responsive monomer, endowing the polymer with photo-responsive properties and enhancing the efficiency of its adsorption/desorption processes.

Subsequently, PMI-PDA@NH₂-SiO₂ was employed as an efficient adsorbent filler for solid-phase extraction (SPE) columns, enabling the effective separation of CA from F. carica L. extracts. The maximum adsorption capacity of PMI-PDA@NH₂-SiO₂ for CA reached 28.6 mg/g at a CA concentration of 0.1 mg/mL. The SPE column filled with PMI-PDA@NH₂-SiO₂ exhibited an adsorption rate of 92.33% and a desorption rate of 87.27% for CA in F. carica L. extracts, confirming its superior adsorption performance for CA. Moreover, PMI-PDA@NH₂-SiO₂ demonstrated high selectivity and affinity for CA, with an IF of 2.84.

After six consecutive adsorption/desorption cycles, PMI-PDA@NH₂-SiO₂ retained robust adsorption performance and reusability. Its adsorption capacity remained at 86.17%, with only a 4.13 mg/g decrease in adsorption amount. Furthermore, the recognition mechanism of PMI-PDA@NH₂-SiO₂ was elucidated to be primarily the synergistic effect of hydrogen bonding and van der Waals forces between the template molecule CA and the functional monomer DA. This work proposes a novel photo-responsive, non-contact regulation strategy for the green and efficient separation of CA. The photo-responsive reversibility of this material significantly improves its reusability, promoting the greening and sustainable development of chemical separation processes and demonstrating considerable potential for industrial application.