Generic strategy for high-performance bio-based imprinted membranes free of non-specific adsorption: machine learning-assisted sensing

Generic strategy for high-performance bio-based imprinted membranes free of non-specific adsorption: machine learning-assisted sensing | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Generic strategy for high-performance bio-based imprinted membranes free of non-specific adsorption: machine learning-assisted sensing Ouarda El Hani, Khalid Digua, Aziz Amine This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7609402/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 21 Mar, 2026 Read the published version in Microchimica Acta → Version 1 posted 8 You are reading this latest preprint version Abstract Molecularly imprinted polymers offer exceptional sensing capabilities; however, their use is limited by the complex synthesis process, which relies on toxic and costly reagents. Additionally, issues with non-specific adsorption (NSA) further diminish their selectivity, limiting their overall effectiveness. Thus, a rapid, green, and cost-effective strategy was proposed for the preparation of bio-based molecularly imprinted membranes (MIMs) by studying diverse functional biopolymers, including chitosan, sodium alginate, carboxymethyl cellulose, sulfonated cellulose nanocrystalline, cellulose acetate, and gelatin. These MIMs were tested for tetracycline and caffeic acid (Caf) templates. The MIMs were prepared in less than 2h, without the need for complex synthesis, toxic, or expensive reagents. Furthermore, various approaches were introduced to eliminate the NSA for the first time, using covalent and non-covalent crosslinking. After appropriate selection of the biopolymer and crosslinker, the affinity of the non-imprinted membranes, free of cavities, toward the templates became negligible, thereby confirming the improvement in imprinting performance and the suppression of NSA. Coupling these NSA-free MIMs with smartphone colorimetric detection offers rapid, cost-effective, and on-site sensing. Sensor arrays were developed for the detection of Caf in pears, plums, and apples. The RGB spectral data was processed using machine learning. The artificial neural network model showed excellent regression performance, with high R 2 (0.989–0.970) and low RMSE (0.01–0.05), confirming the strategy's precision and the MIMs selectivity efficiency in complex matrices. This work provides a pathway to transition from conventional synthetic to sustainable artificial antibodies for an ultra-selective, green, cost-effective, and efficient future for imprinting technologies. Molecularly imprinted membrane biopolymer non-specific adsorption crosslinking sensor array machine learning Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Introduction Molecular imprinting technology has emerged as an important approach to creating synthetic recognition materials via a lock-and-key mechanism. This process involves the template-induced formation of specific recognition sites within a polymer matrix. The resulting molecularly imprinted polymers (MIPs) have exceptional properties such as selectivity, cost-effectiveness, and stability [ 1 , 2 ]. These attributes make MIPs good alternatives to natural antibodies [ 3 ] with diverse applications in sensing [ 4 ] and separation processes [ 5 ]. Recently, the scope of molecular imprinting technology has extended to the development of molecularly imprinted membrane (MIM), which integrates the molecular recognition capabilities of MIPs with the membrane characteristics. Beyond the membranes' robustness and ease of handling, they offer unique advantages in selective adsorption and separation, owing to the presence of tailor-made recognition sites embedded within a permeable membrane matrix. Their controllable permeability and surface and internal imprinting capabilities enable precise interaction with target analytes while minimizing interference from structurally similar compounds [ 6 ]. Unlike conventional MIP synthesis methods, which require long-time synthesis, high energy inputs, and expensive or hazardous monomers and initiators [ 7 ]. An alternative synthesis approach of alginate and cellulose acetate-based MIM was proposed recently in our laboratory; these biopolymers simultaneously act as the imprinted polymer due to their functional groups, and as the membrane matrix [ 8 , 9 ]. However, these studies have been limited to only two biopolymers. For this reason, in the present work we propose a generic strategy for successfully preparing bio-based MIMs by testing various well-known biopolymers in membrane science with diverse functional groups, including cellulose derivatives (e.g., cellulose nanofibers, carboxymethyl cellulose (CMC), sulfonated cellulose nanocrystalline (SCNC), and cellulose acetate (CA)), sodium alginate (SA), chitosan (Cs), and gelatin (Gel) [ 10 ]. The MIMs demonstrate not only their capability as selective adsorbents for targeted separation but also their versatility as sensing platforms across various applications, including environmental monitoring, food safety, and pharmaceutical analysis, enhancing both selectivity and sensitivity[ 11 ]. Combining them with smartphone-based detection makes sensor arrays more efficient and suitable for on-site monitoring [ 12 ]. These advanced systems enable real-time and accurate monitoring of various analytes. The integration of machine learning algorithms [ 13 ] further enhances the performance of sensor arrays by analyzing complex datasets to improve the sensitivity, rapidity, and precision of detection. Among the widely used machine learning models in sensing were artificial neural networks (ANN) [ 14 ] and partial least squares (PLS) [ 12 ]. Despite the advantages of MIPs and MIMs, these materials still face the problem of non-specific adsorption (NSA), which compromises their selectivity and, consequently, their applicability in real-world scenarios. NSA refers to the unintended binding of target molecules to non-specific sites within the material, which significantly diminishes selectivity [ 15 – 17 ]. In the literature, it is often observed that non-imprinted materials, prepared using the same protocol as MIPs/MIMs but without the template, still exhibit an affinity for target molecules, indicating the presence of numerous non-specific sites within the materials. Despite the severity of the NSA issue, it has been addressed in only a limited number of studies [ 15 – 18 ]. In our laboratory, we have proposed innovative non-covalent approaches to address NSA in both MIPs and MIMs. For MIPs, we developed a strategy involving the modification of MIPs with surfactants [ 19 ]. Similarly, for MIMs, we introduced a method to eliminate NSA in alginate-based MIMs, involving crosslinking with divalent cations, followed by chelation before template removal [18 ]. This approach significantly enhanced the imprinting performance. On the same concept, Noof et al. modified a methacrylic acid-based MIP using diazomethane to remove NSA, thereby enhancing its selectivity for bisphenol A. These studies highlight the crucial role of NSA elimination in MIP synthesis [ 20 ]. However, these studies are still in their infancy. To enlarge the control of this problem for large types of MIMs, this work proposes a generic innovative approach to bio-based MIMs preparation using diverse biopolymers. Different crosslinking approaches were evaluated to eliminate NSA: covalent and non-covalent. Crosslinkers were selected based on the functional groups present in the biopolymers. Overall, this study aims to present a generic strategy for preparing bio-based imprinted membranes and studying various biopolymers, such as CMC, SA, CA, Cs, Gel, and SCNC. The research emphasizes a significant step toward developing green and sustainable imprinting technologies. This strategy was tested on two different templates: antibiotic (tetracycline) and phenol (caffeic acid). Moreover, both covalent and non-covalent modification approaches were implemented for NSA elimination. Isothermal evaluations were performed to evaluate the imprinting performance of the bio-based MIMs using various crosslinkers. Additionally, a sensor array of MIMs coupled with smartphone-based colorimetric detection and machine learning was proposed for the precise, selective detection and prediction of caffeic acid in fruits. This work provides a generic framework for preparing bio-based MIM, offering efficient solutions for NSA elimination. 1. Materials and methods 1.1. Reagents and instrumentation CMC (high viscosity extra pure, with carboxymethyl substitution degree of 0.5–0.8), SA (C 6 H 7 O 6 Na) n (M w =10000–600000 g/mL, 91%-106%) (SA, BioChemica), Cs, Gel, CA (39.8 wt. % acetyl content. Mn 50000), SCNC, calcium, phosphate buffer solution (PB, 0.5 M, pH = 7.0, prepared from anhydrous sodium monobasic phosphate and sodium dibasic phosphate dodecahydrate), sodium sulfate (Na 2 SO 4 , 98%), sodium tripolyphosphate (STPP), cetyltrimethylammonium bromide (CTAB, ≥ 98%, LaboChemie), Sodium dodecyl sulfate (SDS, LaboChemie), Borax, glutaraldehyde solution (25% in H 2 O) (GA), epichlorohydrin (EPC), 1-Ethyl-3-(3 dimethylaminopropyl)carbodiimide/N-Hydroxysuccinimide (EDC/NHS, ≥ 99%, Sigma), copper, ethylenediamine (EDA, ≥ 98%, Labo Chemie), phenylenediamine (PDA, 98%, Sigma), Folin-Ciocalteu (FC, 2N, Sigma), sodium carbonate (Na 2 CO 3 ), sodium hydroxide (NaOH), caffeic acid (Caf, ≥ 98%, Sigma), and tetracycline (TC, 98–102%, Sigma). The spectrofluorometer POLARStar Omega, from BMG LABTECH (Germany), was used for the absorbance measurement. A Samsung S10 smartphone served as the readout device for smartphone-based colorimetric detection, utilizing the high-efficiency image format. Lab Filter Holders (PP 25mm) integrated to a syringe (10 mL) were used to perform the extraction. FTIR spectroscopy (ThermoFisher Scientific, USA) in ATR mode, covering the spectral range of 650–4000 cm¹, was used to analyze the chemical structure of the synthesized MIMs. 1.2. Preparation of the bio-based MIMs The preparation of MIMs using each time SA, CMC, SCNC, Cs, CA, or Gel as imprinting polymer was conducted as follows. A 1.5% (w/v) solution of SA, CMS, or SCNC was prepared in water, while 1.5% Cs was dissolved in 1% acetic acid and 5% CA in DMSO. For Gel-based MIMs, a 5% (w/v) solution was solubilized in water. All solutions were stirred for 15 minutes. Subsequently, 0.2% of the template (TC or Caf) was added to the polymer solution. The mixture was stirred for 15 minutes to ensure homogenization and to promote hydrogen bonding between the template and the functional groups of the biopolymer. Then, polymer chains were linked with 1% of borax (due to its affinity to hydroxyl groups [ 21 ]) to create the initial three-dimensional network around the template. 10 mL of the resulting gel was cast onto an 8.5 cm petri dish and dried for 1h at 50°C to allow solvent evaporation. Bio-based NIMs were prepared using the same protocols as MIMs, without the addition of the template. TC and Caf templates interacted via hydrogen bonding with the carboxylic groups of SA, CMC, and Gel, the amine groups of Cs and Gel, the sulfone groups of SCNC, and the acetyl functional groups in CA. However, some functional groups are not involved in the cavity creation during binding. Thus, the prepared MIMs and corresponding NIMs were subjected to crosslinking before template removal to block the non-specific sites (as shown in Scheme 1 ). The crosslinkers are selected based on the functional groups in each bio-based MIM (as shown in Table 1 ). All further experiments were performed in triplicate. 1.2.1. Elimination of NSA from SA, CMC, or SCNC-based MIMs For SA, CMC, and SCNC-MIMs/NIMs, various non-covalent crosslinkers and modification agents were tested, including 5% Ca²⁺/1 mM PB, 1% CTAB, or 1% Cu²⁺ (known for their affinity to anionic carboxylic groups [ 18 , 22 , 23 ]). Also, covalent crosslinking with 0.2 M EDC/NHS activating carboxylic groups, followed by reaction with diamines (1% EDA or 1% PDA) [ 24 – 26 ] was performed for SA and CMC-based MIM/NIMs or by adding 1% EPC (to interact with sulfates SO 3 ⁻ [ 27 ]) for SCNC-based MIM/NIM. 1.2.2. Elimination of NSA from Cs-based MIM To remove NSA from Cs-MIM, electrostatic (non-covalent) modification was performed, testing various agents, including 1% Na 2 SO 4 , 1% STPP, 1% SDS, 1% Cu²⁺ (known for their affinity to NH₃⁺ [ 28 – 31 ]). Covalent crosslinking using 1% GA (to interact with NH₃⁺ [ 32 ]). 1.2.3. Elimination of NSA from Gel and CA-based MIMs For Gel-MIM, non-covalent modification was tested using 1% Na 2 SO 4 , 1% STPP, or 1% SDS, for interactions with NH 3 ⁺ groups, and 1% SDS to interact with COO - groups. For covalent crosslinking, 1% GA was employed to interact with amine groups in Gel by using 0.2 M EDC/NHS by activating the carboxylic groups of Gel, followed by reaction with amine groups (present in Gel) via self-crosslinking. For CA-MIM, 1% GA and 1% EPC were used as crosslinkers for OH groups. Table 1 Tested covalent and non-covalent modification mechanisms for NSA elimination in prepared bio-based MIMs. Bio-based MIM Crosslinking and modification agent Mechanism Function SA and CMC 0.5% Ca²⁺ + 1 mM PB 1% CTAB 1% Cu²⁺ Electrostatic (non-covalent) COO - 0.2 M EDC/NHS/ 1% EDA (or 1% PDA) Covalent SCNC 0.5% Ca²⁺ + 1 mM PB 1% CTAB 1% Cu²⁺ Electrostatic (non-covalent) SO 3 - 1% EPC Covalent Cs 1% Na 2 SO 4 1% STPP 1% SDS 1% Cu²⁺ Non-covalent NH₃⁺ 1% GA Covalent Gel 1% Na 2 SO 4 1% STPP 1% SDS 1% Cu²⁺ Non-covalent NH₃⁺ 1% CTAB Non-covalent COO - 1% GA Covalent NH 3 0.2 M EDC/NHS Covalent COOH and NH 3 CA 1% GA 1% EPC Covalent OH 1.2.4. Template removal After the preparation of bio-based MIMs/NIMs, they were cut into 0.5 cm diameter discs using a perforator and then mounted in a membrane holder integrated into a 10 mL syringe. The membranes were initially rinsed with 1 mL of methanol to remove unreacted reagents. Template removal was then carried out using 10 mL of 0.5 N Folin-Ciocalteu (FC)/0.1% NaOH. In a continued way to ensure complete template removal (Scheme 2 A). The developed approach of extraction was compared to the conventional approach using methanol and methanol: acetic acid (9:1) mixture as extraction solvents. In the conventional approach, the extraction solvents were added, and the process continued until complete removal of the target analytes was achieved. Afterward, the membranes were thoroughly washed with distilled water to eliminate residual molecules such as FC. Template removal was monitored visually and also assessed by UV-Vis spectroscopy. 1.3. Binding study using the prepared bio-based MIMs The binding performance of the bio-based MIMs and NIMs prepared from each biopolymer was assessed by evaluating their binding capacity for templates (TC and Caf) using MIM/NIM disc (0.5 cm, 4 mg) as sorbents in 10 mL of ethanol: water (1:1) template solution at different concentrations (0.5-8 µg/mL for Caf and 0.5-5 µg/mL for TC ) within 15 minutes. To further investigate the adsorption behavior of the membranes, the adsorption capacity (Q(mg/g)) was determined through isotherm studies by analyzing the uptake capacities of MIMs and NIMs across varying template concentrations. The adsorption capacity was calculated using the following equation ( Eq. 1 ): Eq. 1 \(\:\:\:\:Q\:\left(mg/g\right)=\frac{({C}_{initial}-{C}_{final})}{m}\times\:V\) Where: C initial ​ (µg/mL) is the initial template concentration, C final (µg/mL) is the template concentration remaining in the solution after adsorption, m(mg) is the weight of the membrane (MIM or NIM), and V (mL) is the loading volume. Imprinting factor (IF) represents the efficiency of each bio-based MIM compared to its corresponding NIM. An IF > 1 indicates that the selective sites of the template were successfully created in MIM (the higher the IF, the more performant and selective the MIM is). IF is calculated according to the following equation ( Eq. 2 ): Eq. 2 \(\:\:\:\:\:\:IF=Q\left(MIM\right)/Q\left(NIM\right)\:\:\) Where: Q (MIM) and Q (NIM) were the adsorption capacities of MIM and NIM, respectively. 1.4. Stability and reusability study Stability tests were conducted by storing the membranes (Cs-MIM, Gel-MIM, SA-MIM) at room temperature for 90 days, followed by evaluating their adsorption efficiency for 5 mg/L Caf, tested on Cs-MIM (GA) and Cs-MIM (Cu 2+ ) or TC, tested on SA-MIM (EDC/NHS/PDA) and Gel-MIM (EDC/NHS). For reusability, these bio-based MIMs underwent 10 consecutive adsorption/desorption cycles: adsorption in water (5 mg/L Caf or TC) and desorption using FC extracting agent. Recoveries are quantified via the developed smartphone-based colorimetric detection. 1.5. Smartphone-based colorimetric detection of phenolic-based compounds using MIMs sensor and FC reagent For both Caf and TC detection, bio-based MIMs (Cs-MIM (GA) for Caf and SA-MIM (EDC/NHS/PDA) for TC), each measuring 0.5 cm in diameter, were placed in the wells of a microplate. Then, 100 µL of the analytes (added in four equal drops of 25 µL, with a 5-minute interval between additions) were introduced at concentrations ranging from 0.1 to 5 µg/mL for TC and 0.05 to 8 µg/mL for Caf. Subsequently, 10 µL of the FC reagent and 100 µL of 0.6 M Na 2 CO 3 were added. The resulting blue coloration was measured using an ELISA reader at 750 nm and analyzed via smartphone imaging, processed with ImageJ software (Scheme 2 A). 1.6. MIMs-based sensor arrays for the smartphone detection of Caf in fruits using machine learning 1.6.1. Sensor arrays preparation For the detection of Caf in fruit samples, four bio-based MIMs were selected based on their superior affinity and selectivity. These MIMs were Cs-MIM (Cu²⁺), Cs-MIM (GA), CMC-MIM (EDC/NHS/PDA), and CMC-MIM (Ca²⁺/PB). The selected MIMs were tested at three concentrations of Caf: 0.1 µg/mL, 0.5 µg/mL, and 1.5 µg/mL. The fruit samples were prepared as follows. Fresh fruits were ground using a garlic press to extract their juice, which was then filtered using a nylon filter with a pore size of 0.45 µm. The resulting filtrate was diluted 10-fold with distilled water to ensure compatibility with the detection system. The experimental sensing process followed the same detection protocol as previously established to maintain consistency across all tests. 1.6.2. Calibration and validation using PLS and ANN approaches For calibration and validation data set collection, the [Caf] response for each further sample was measured using a smartphone integrated with image processing software (ImageJ), which analyzed the RGB color space. Each RGB channel was presented as a spectrum ranging from 0 to 255 pixels, resulting in 265 × 3 channels = 795 spectral variables per sample. Consequently, the size of the calibration dataset was calculated as 4 MIMs × 3 added [Caf] × 4 replicates = 48 samples; and for each sample, 795 variables were collected. The RGB data collected from ImageJ were processed and analyzed using Unscrambler software and JMP-Pro. Two multivariate approaches, PLS and ANN, were employed for the regression. PLS was used for modeling relationships between independent variables and dependent variables, reducing dimensionality while maximizing variance [ 33 ]. The validation involved a 6-fold cross-validation approach. ANN was also employed to model the relationship between input variables and a continuous output, effectively capturing complex patterns and dependencies within the data [ 34 ]. The ANN model implemented in this study consisted of three hidden layers, allowing for the extraction of complex patterns from RGB data to enhance predictive accuracy. The model was trained using boosting with 100 iterations and a 0.1 learning rate, an iterative technique that improves predictions by continuously adjusting weights based on previous errors. 1.6.3. Prediction of [Caf] in fruit samples The obtained PLS and ANN models acquired from the calibration were applied for the prediction of [Caf] in fruit samples, including plum, pear, and apple. The size of the prediction dataset was 3 fruits × 4 MIMs = 12 samples (without adding Caf), with 795 variables per sample collected similarly from ImageJ software. 2. Results and discussions 2.1. Bio-based MIMs as an alternative approach in imprinting technology In the present study, a generic strategy of bio-based MIMs preparation was proposed to enhance both the synthesis performance and the selectivity. The conventional synthesis of MIPs faces several challenges, including long synthesis times (> 24 hours), high energy inputs (thermal, ultrasonic, microwave irradiation, or UV exposure [ 35 ]), and the use of toxic and expensive reagents [ 7 ]. To overcome these limitations, our approach utilized readily available biopolymers as functional ²polymers for MIMs preparation. These biopolymers are abundant, cost-effective, and green. They serve as ideal candidates for the preparation of MIMs. Additionally, biopolymers with diverse functional groups are tested, offering a versatile strategy for imprinting a wide range of templates, thus enhancing the scope of MIM applications. The bio-based MIMs were synthesized in less than 2h without the need for toxic or expensive reagents. The results (summarized in Fig. 1 ) showed that before modification, Qs(MIM) was higher than that of Qs(NIM), confirming the creation of TC cavities in all bio-based MIMs registering IF values between 1.2 and 1.7, confirming the successful creation of template cavities in all bio-based MIMs. The imprinting performance of these MIMs was further enhanced using both covalent and electrostatic crosslinking approaches. 2.2. NSA elimination from the prepared bio-based MIMs The crosslinking in bio-based MIMs, conducted in the presence of the template, not only facilitates the formation of a stable 3D network around the template, thereby enhancing the recognition cavities by making them complementary in both functional groups and spatial configuration, but also plays a crucial role in eliminating NSA. Specifically, this strategy enables the selective crosslinking of polymer chains in the functional groups not involved in reaction with the template, effectively occupying and blocking free non-specific sites that would otherwise contribute to undesired analyte retention during the rebinding step. By blocking these groups before the template is removed, the resulting material presents minimal accessible non-imprinted sites, ensuring that specific interactions within the imprinted cavities primarily drive recognition. Thus, negligible adsorption in NIM signifies the successful NSA elimination. Based on Fig. 1 , before crosslinking, all bio-based MIMs exhibited IF ranging from 1.3 to 1.5 (> 1), indicating moderate imprinting performance and confirming the successful formation of template sites via the proposed approach. The use of biopolymers as both functional polymers and membrane matrices, combined with borax-mediated crosslinking, facilitated the initial creation of a 3D network around the template by occupying the diol functions from the two chains of biopolymers. Despite this, the IF values remained moderate, and the NIMs still displayed high affinity for the template due to the NSA, highlighting the need for further optimization of the system. Furthermore, the mild conditions applied in this study (room temperature, minimal mechanical stress) ensured that the possible dynamic boronate ester bonds remained intact, preserving the integrity of the imprinted cavities. These findings align with our further experiments, which demonstrated the effectiveness and robustness of the MIMs, particularly after additional covalent or non-covalent crosslinking steps to reinforce the imprinting architecture and eliminate NSA. For SA- and CMC-based MIMs (Fig. 1 A, B), after modification with CTAB or Cu²⁺, a noticeable reduction in Q(NIM) was observed, suggesting partial elimination of NSA. Crosslinking with Ca²⁺/PB or EDC/NHS/(EDA or PDA) resulted in the lowest Q(NIM) with IFs of 20–30 and 67–80, respectively, indicating effective imprinting enhancement. For SCNC-based MIM (Fig. 1 C), CTAB or Cu²⁺ reduced significantly Q(NIM). EPC was the most effective in enhancing the imprinting performance (IF = 45). For Cs-based MIM (Fig. 1 D), Na 2 SO 4 , STPP, and SDS reduced the Q(NIM) significantly but left moderate NSA. Cu²⁺ and GA, with IF values of 90 and 80, respectively, provided optimal NSA elimination. For Gel-based MIM (Fig. 1 E), Na 2 SO 4 , STPP, SDS, CTAB, GA, and Cu 2+ provided lower Q(NIM). Self-crosslinking with EDC/NHS further enhanced the imprinting performance (IF = 10). Despite the improvement of CA-based MIM (Fig. 1 F), NIM adsorption capacity could not be completely eliminated due to the inherent difficulty in crosslinking acetyl functions in CA without decomposing them, which limits the complete suppression of NSA. The previously discussed modifications for NSA elimination were also tested on different days to evaluate the variability of the generic strategy. Consistent results were obtained, demonstrating the robustness of the method with a coefficient of variation below 5%. 2.3. Template removal The removal of templates from MIMs is a critical step that significantly influences their selectivity and performance. Conventional extraction methods often require prolonged processing times due to the high affinity of the templates for the imprinted matrix, and the use of toxic, harsh solvents can damage the cavity structures, ultimately reducing the material’s selectivity. Therefore, developing a green, rapid, and simple extraction approach is of high importance. In this work, template removal was achieved via complexation with the FC reagent. The FC is well-known for its high affinity to phenolic compounds, forming a stable blue complex that is structurally incompatible with the original cavity architecture. This prevents re-adsorption and facilitates the template release while preserving the integrity of the imprinted sites. In addition, the developed MIMs were designed such that template binding occurred predominantly at or near the surface of the membrane, as is typical in membrane-imprinted systems. This superficial binding significantly reduces the diffusion path length, making template accessibility feasible even in the presence of covalent crosslinking (e.g., via GA in Cs-MIM or EDC/NHS in SA-MIM ). A comparative investigation was conducted to examine the removal of TC and Caf using two approaches: the complexation reaction with FC versus conventional elution with methanol: Acetic acid (9:1) and pure methanol. As depicted in Table S1 , the conventional methods required a time from 6 h to 10 h for extraction, whereas the FC-based complexation achieved complete template removal within only 20 min. This high extraction efficiency is attributed to the superior complexation capability of the FC reagent. Moreover, the continuous extraction strategy used here offers a distinct advantage over traditional methods that involve soaking the MIM in an elution solution, which often face the risk of template re-adsorption and release equilibrium. The use of a syringe-coupled MIM scaffold further enhances the process by enabling extraction under controlled flux, thereby accelerating continuous extraction and ensuring effective template removal. This highlights the effectiveness of the proposed removal method as a rapid and environmentally friendly method for phenolic templates removal in MIMs. 2.4. Binding study of the bio-based MIMs for TC and Caf templates After selecting the optimum modifications for each bio-based MIM, the proposed generic strategy was applied to two different templates: Caf and TC. The results presented in Fig. 2 highlight the Q(mg/g) of Caf and TC using bio-based MIMs and their corresponding NIMs. The significant differences in Q(MIM) and Q(NIM) confirm the successful creation of template cavities in the MIMs. However, the affinity of each template for each bio-based MIM varies due to differences in the interactions between the functional groups in the bio-based MIMs and the templates, as well as the effectiveness of the crosslinking strategies. Despite the difference in the affinity of the bio-based MIMs for each template, NSA elimination using the optimal crosslinking agents showed similar improvements in imprinting performance. Therefore, this generic strategy can be applied regardless of the template. For Caf (Fig. 2 A), the highest Q(mg/g) was observed in Cs-based MIMs. This is attributed to the abundant amino groups in Cs, which enable strong hydrogen interactions with phenolic and carboxylic groups of Caf. Then, CMC and SA-based MIMs demonstrated lower, but still significant, affinity. In these cases, the carboxyl groups in CMC and SA facilitated moderate interactions with Caf, though their affinity was slightly lower than that of Cs-based MIMs. This difference in affinity may be attributed to the absence of amino groups, which are present in the Cs-based MIMs and likely contribute to stronger interactions. Gel-based MIMs, despite having abundant amine functions, showed reduced affinity due to limited spatial accessibility for interactions compared to Cs. CA and SCNC-based MIMs exhibited the lowest adsorption capacities, as acetyl groups and sulfate have limited interaction potential with phenols and acids. In the case of TC (Fig. 2 B), Gel-, SA-, and CMC-based MIMs showed the highest affinity. This is due to the multiple carboxyl groups in SA and CMC, which interact favorably with the amino and amide groups in TC, forming strong hydrogen bonds, as well as the amide and carboxyl groups in Gel. Cs-based MIMs demonstrated moderate adsorption, while CA- and SCNC-based MIMs exhibited the weakest binding. 2.5. Isothermal study of Caf and TC using the MIMs After selecting the optimum bio-based MIMs for each template, the isothermal study for Caf and TC, conducted over specific concentration ranges (0.5-8 mg/L for Caf and TC), highlights the adsorption behavior of the selected MIM and NIM. For Caf (Fig. 2 C), the Q increases progressively with concentration of Caf for both Cs-MIM (Cu 2+ ) and Cs-MIM (GA). The Q of both MIMs far exceeds that of their respective NIMs, confirming the successful creation of the cavity and elimination of NSA. For TC (Fig. 2 D ) , Gel-MIM (EDC/NHS), SA-MIM (Ca 2+ /PB), and SA-MIM (EDC/NHS/PDA) demonstrate a significant increase in Q(MIMs). The corresponding NIMs display non-adsorption across all concentrations, confirming the elimination of NSA. For Gel-NIM (EDC/NHS), NSA was not observed at lower concentrations except at 5 mg/L, though it remained insignificant compared to its corresponding Gel-MIM. 2.6. Characterizations 2.6.1. FT-IR results In Fig. 3 A, the FT-IR spectrum for Cs-MIM showed the prominent peaks of Cs, including a broad band at 3200 cm⁻¹ attributed to O–H and NH₂ groups, a peak at 2880 cm⁻¹ for CH vibrations, a peak at 1560 cm⁻¹ for NH₂ stretching, and a peak at 1030 cm⁻¹ corresponding to C–O vibrations from the pyranose ring. After modification with Cu²⁺, the 2880 cm⁻¹ peak disappears, and the NH₂ stretching peak shifts to 1650 cm⁻¹ due to Cu 2+ -induced crosslinking that alters the chemical environment of the amino groups; concurrently, the pyranose ring peak becomes sharper and more intense, suggesting a more ordered structure. After modification with GA, significant spectral changes are observed: the appearance of a new peak at 1660 cm⁻¹, indicative of C = N stretching and confirming Schiff base formation, an increase in C–H stretching intensity from the additional alkyl groups in GA, notable shifts and enhanced intensity of the C–O related peak at 1050 cm⁻¹ reflecting alterations in the polysaccharide backbone, also a new peak at 950 cm⁻¹ attributed to changes in glycosidic linkages after crosslinking. For Gel-MIM (Fig. 3 B), before modification exhibited characteristic peaks including a broad band around 3300 cm⁻¹ for O–H and N–H stretching, and distinct peaks around 1650 cm⁻¹ and 1550 cm⁻¹ for amide groups, with additional bands between 1100 and 1500 cm⁻¹ corresponding to C-O and C-N free functional groups. After self-crosslinking with EDC/NHS, there is a noticeable reduction in the intensity of these primary peaks and the disappearance of the bands in the 1100–1500 cm - ¹ range. This indicates that the free amine and carboxyl groups have been linked, leading to a crosslinked Gel structure. In Fig. 3 C, the FTIR spectrum of SA-MIM displays the characteristic peaks of SA: a broad band at 3270 cm⁻¹ due to OH stretching, peaks at 1580 cm⁻¹ and 1400 cm⁻¹ corresponding to the asymmetric and symmetric stretching vibrations of –COO⁻ groups, and a peak at 1020 cm⁻¹ from C–O stretching in the polysaccharide backbone. Upon crosslinking with Ca²⁺/PB, the peak at 1580 cm⁻¹ shifts to 1600 cm⁻¹ and that at 1400 cm⁻¹ shifts to 1430 cm⁻¹, suggesting the formation of ionic bonds between Ca²⁺ and the –COO⁻ groups, with the appearance of a distinct peak around 950 cm⁻¹ further indicating the establishment of a crosslinked network. Additionally, modification with EDC/NHS/PDA causes a shift from 1580 cm⁻¹ to 1607 cm⁻¹, which is attributed to the conversion of the –COO⁻ into amide bonds after activation by EDC/NHS and reaction with PDA, thereby confirming the successful modification of SA-MIM. 2.6.2. Water contact angle results As shown in Fig. 3 (D, E , and F) , Cs-MIM, Gel-MIM, and SA-MIM exhibited remarkable hydrophilicity before crosslinking with contact angles of 43°, 39°, and 27°, respectively. After non-covalent crosslinking, with Cu²⁺ for Cs-MIM and Ca²⁺/PB for SA-MIM, the contact angles increased to 75° and 74°, respectively. Covalent crosslinking further decreased hydrophilicity, with contact angles rising to 94° for Cs-MIM (GA), 78° for Gel-MIM (EDC/NHS), and 74° for SA-MIM (EDC/NHS/PDA); this is due to the formation of stable covalent bonds that consume free hydrophilic groups and introduce more hydrophobic moieties. 2.7. Stability and reusability performance As shown in Fig. 4 A, the stability over 90 days demonstrates that all membranes exhibit excellent long-term performance, with recovery rates consistently above 80%. While Cs-MIM (Cu²⁺) shows slightly lower values, its overall performance remains remarkable. Similarly, Fig. 4 B illustrates the reusability over 10 cycles. Cs-MIM (GA) maintains reliable performance for Caf adsorption/desorption, retaining high recovery rates even after repeated cycles (> 80%). Likewise, SA-MIM (EDC/NHS/PDA) and Gel-MIM (EDC/NHS) exhibit stable reusability for TC adsorption/desorption, underscoring the practical suitability of bio-based MIMs for repeated applications due to efficient covalent crosslinking, which enhances their mechanical properties. Cs-MIM (Cu 2+ ) and SA-MIM (Ca 2+ /PB) showed as well as remarkable recoveries up to cycle 8. After that, its regeneration efficiency reduces due to the limited stability of non-covalent crosslinking compared to covalent bonding. Overall, the MIMs developed via electrostatic crosslinking effectively reduce NSA but exhibit limited stability. To overcome this limitation, covalent crosslinking was introduced to enhance imprinting performance and expand the applicability of these MIMs, making these MIMs more suitable for complex matrices, even under challenging conditions. 2.8. The smartphone-based colorimetric detection using MIMs and FC reagent FC reagent is widely recognized for its ability to interact with phenols, such as Caf, It reacts with phenolic hydroxyl groups under alkaline conditions, producing a blue chromophore that absorbs at 750 nm [ 36 ]. This reaction occurs due to the interaction of the FC reagent with the phenol group, making the method suitable for the colorimetric detection of any molecule containing phenolic groups. In this study, the FC-based colorimetric method applied for the first time for TC detection, to the best of our knowledge. After the development of blue coloration, images of the resulting color intensities were captured using a smartphone and analyzed with ImageJ software. The RGB (red-green-blue) intensities were measured and correlated with analyte concentrations. After subtracting the blank values, significant increases in all three channels (R, G, and B) were observed with rising analyte concentrations, as illustrated in Fig. 5 . Subsequently, specific channels were selected to construct calibration curves. For Caf (Fig. 5 A), a remarkable R² value of 0.985, a limit of detection (LOD) of 0.003 mg/L, and a limit of quantification (LOQ) of 0.015 mg/L were obtained. These limits were calculated based on calibration curve approach as follows ( Eq. 3 ): Eq. 3 \(\:LOQ=10\times\:Blank\:standard\:deviation/Slope\:\) Similarly, for TC (Fig. 5 B), R² value of 0.983, LOD of 0.03 mg/L, and LOQ of 0.1 mg/L were obtained. These results confirm the effectiveness of smartphone-based colorimetric detection and bio-based MIM binding performance in quantifying the concentrations of both analytes with high sensitivity and reliability. 2.9. Application of MIMs in sensor arrays for the detection of Caf in fruits 2.9.1. PLS regression model The PLS regression results for Caf detection are detailed in the Supplementary material . 2.9.2. ANN regression model To assess the accuracy of the ANN model, the correlation between measured and predicted values was analyzed. A strong correlation was observed in the calibration, with R 2 = 0.989, R 2 adj = 0.974, and a low Root Mean Square Error of Calibration (RMSEC) of 0.01, indicating excellent model fit and accuracy. Similarly, in validation, the ANN model exhibited good performance, achieving R 2 = 0.979, R 2 adj = 0.958, and a RMSEV = 0.02 (Fig. 5 C). Furthermore, the prediction results (Fig. 5 D) showed R 2 = 0.970, R 2 adj = 0.967, and a RMSEP = 0.05. Additionally, the notably lower RMSE values and higher R 2 compared to PLS further highlight the superior performance of ANN. Among the tested samples based on Cs-MIM (Cu 2+ ), plum exhibited the highest Caf concentration (1.9 mg/L), followed by pear (1.3 mg/L), while apple had the lowest concentration (0.8 mg/L) using the ANN model. In addition to the machine learning efficiency in sensing, the remarkable results are attributed to the high imprinting performance of the developed bio-based MIMs, which enable selective detection of Caf in complex matrices, as well as the reliability of the smartphone-based detection method. Together, contributed to the overall effectiveness and robustness of the proposed sensing strategy. Conclusions This study demonstrates the successful development of bio-based MIMs as a promising alternative to conventional synthetic antibodies. By utilizing various biopolymers with distinct functional groups, including Cs, SA, CMC, SCNC, CA, and Gel, we have shown the potential of the proposed strategy to address common limitations of synthetic antibodies. These MIMs were prepared using a rapid, green, and effective strategy. Also, the selectivity was significantly enhanced. The IF was improved considerably through covalent and non-covalent crosslinking with specific agents, increasing from initial values around 1.5 (for 5 mg/L) before NSA elimination to much higher values. For SA and CMC, crosslinking with Ca²⁺/phosphate achieved an IF of 30, and EDC/NHS/diamine crosslinking improved the IF to 80. SCNC showed an IF of 45 with EPC, while Cs achieved an IF of 90 with Cu²⁺ and 80 with GA. For Gel, crosslinking with EDC/NHS resulted in an IF of 10. These enhancements were observed for both TC and Caf templates. In addition, a sensor array using MIMs for smartphone-based detection of Caf coupled with PLS and ANN models was developed. The ANN model demonstrated strong predictive performance, with high values (0.989–0.970) and low RMSE (0.01–0.05) across calibration, validation, and prediction. These results confirm its accuracy, reliability, and potential for rapid Caf detection in complex fruit matrices. Overall, this work not only enhances the selectivity and performance of bio-based MIMs but also paves the way for sustainable, cost-effective, and efficient alternatives to synthetic antibodies. Declarations CRediT authorship contribution statement Ouarda El Hani : Investigation; Conceptualization; Data curation; Formal analysis; Methodology; Software; Writing-original draft . Khalid Digua: Supervision; Writing – review & editing; Validation . Aziz Amine: Supervision; Conceptualization; Resources; Writing – review & editing; Validation. Acknowledgements We thank Prof. Nabil El Moçayd from Mohammed VI Polytechnic University (Morocco) for his valuable remarks and recommendations regarding machine learning models. We also acknowledge the support of the PRIMA project, “Agro Food Waste Recovery: New Processing Technologies for Food Safety and Packaging (FoWRSaP)”. Ethics approval : Not applicable. Source of biological material: Not applicable. Statement on animal welfare : Not applicable. Conflict of interest: The authors have no financial or non-financial conflict of interest to disclose. Competing of interest: The authors have no financial or non-financial competing interests to disclose. Funding : No funding was received to assist with the preparation of this manuscript. References BelBruno JJ (2019) Molecularly Imprinted Polymers. Chem Rev 119:94–119. https://doi.org/10.1021/acs.chemrev.8b00171 Teixeira SPB, Reis RL, Peppas NA, et al (2021) Epitope-imprinted polymers: Design principles of synthetic binding partners for natural biomacromolecules. Sci Adv 7:eabi9884. https://doi.org/10.1126/sciadv.abi9884 Smolinska-Kempisty K, Guerreiro A, Canfarotta F, et al (2016) A comparison of the performance of molecularly imprinted polymer nanoparticles for small molecule targets and antibodies in the ELISA format. Sci Rep 6:37638. https://doi.org/10.1038/srep37638 Saadatidizaji Z, Sohrabi N, Mohammadi R (2024) Development of a simple polymer-based sensor for detection of the Pirimicarb pesticide. Sci Rep 14:10293. https://doi.org/10.1038/s41598-024-60748-6 Rutkowska M, Płotka-Wasylka J, Morrison C, et al (2018) Application of molecularly imprinted polymers in analytical chiral separations and analysis. TrAC Trends in Analytical Chemistry 102:91–102. https://doi.org/10.1016/j.trac.2018.01.011 Ratnaningsih E, Kadja GTM, Putri RM, et al (2022) Molecularly Imprinted Affinity Membrane: A Review. ACS Omega 7:23009–23026. https://doi.org/10.1021/acsomega.2c02158 Martins RO, Bernardo RA, Machado LS, et al (2023) Greener molecularly imprinted polymers: Strategies and applications in separation and mass spectrometry methods. TrAC Trends in Analytical Chemistry 168:117285. https://doi.org/10.1016/j.trac.2023.117285 El Hani O, García-Guzmán JJ, Palacios-Santander JM, et al (2024) Development of a molecularly imprinted membrane for selective, high-sensitive, and on-site detection of antibiotics in waters and drugs: Application for sulfamethoxazole. Chemosphere 350:141039. https://doi.org/10.1016/j.chemosphere.2023.141039 El Hani O, García-Guzmán JJ, Palacios-Santander JM, et al (2024) Molecularly imprinted alginate-PVP membrane for the selective fluorescence detection of flavonoids in foods: A practical case for rutin. Sensors and Actuators B: Chemical 403:135232. https://doi.org/10.1016/j.snb.2023.135232 Galiano F, Briceño K, Marino T, et al (2018) Advances in biopolymer-based membrane preparation and applications. Journal of Membrane Science 564:562–586. https://doi.org/10.1016/j.memsci.2018.07.059 Wang M, Cetó X, Del Valle M (2022) A novel electronic tongue using electropolymerized molecularly imprinted polymers for the simultaneous determination of active pharmaceutical ingredients. Biosensors and Bioelectronics 198:113807. https://doi.org/10.1016/j.bios.2021.113807 Yang M, Zhang M, Jia M (2023) Optical sensor arrays for the detection and discrimination of natural products. Nat Prod Rep 40:628–645. https://doi.org/10.1039/D2NP00065B Yang J, Lu S, Chen B, et al (2023) Machine learning-assisted optical nano-sensor arrays in microorganism analysis. TrAC Trends in Analytical Chemistry 159:116945. https://doi.org/10.1016/j.trac.2023.116945 Saeed R, Glamuzina B, Tuyet Nga MT, et al (2025) Supervised learning-based artificial senses for non-destructive fish quality classification. Biosensors and Bioelectronics 267:116770. https://doi.org/10.1016/j.bios.2024.116770 Lamaoui A, Palacios-Santander JM, Amine A, Cubillana-Aguilera L (2021) Molecularly imprinted polymers based on polydopamine: Assessment of non-specific adsorption. Microchemical Journal 164:106043. https://doi.org/10.1016/j.microc.2021.106043 Karrat A, Amine A (2024) Innovative approaches to suppress non-specific adsorption in molecularly imprinted polymers for sensing applications. Biosensors and Bioelectronics 250:116053. https://doi.org/10.1016/j.bios.2024.116053 Li X, Zhang B, Tian L, et al (2015) Improvement of recognition specificity of surface protein-imprinted magnetic microspheres by reducing nonspecific adsorption of competitors using 2-methacryloyloxyethyl phosphorylcholine. Sensors and Actuators B: Chemical 208:559–568. https://doi.org/10.1016/j.snb.2014.11.045 Hani OE, Digua K, Amine A (2025) Elimination of non-specific adsorption in the molecularly imprinted membrane: application for tetracycline detection. Anal Bioanal Chem. https://doi.org/10.1007/s00216-025-05804-4 Karrat A, Amine A (2024) Innovative approaches to suppress non-specific adsorption in molecularly imprinted polymers for sensing applications. Biosensors and Bioelectronics 250:116053. https://doi.org/10.1016/j.bios.2024.116053 Alenazi NA, Lai EPC, Manthorpe JM (2014) Enhanced selectivity of a molecularly imprinted polymer toward the target molecule via esterification of non‐specific binding sites with diazomethane. J of Molecular Recognition 27:755–762. https://doi.org/10.1002/jmr.2402 Tanpichai S, Phoothong F, Boonmahitthisud A (2022) Superabsorbent cellulose-based hydrogels cross-liked with borax. Sci Rep 12:8920. https://doi.org/10.1038/s41598-022-12688-2 Mirtič J, Ilaš J, Kristl J (2018) Influence of different classes of crosslinkers on alginate polyelectrolyte nanoparticle formation, thermodynamics and characteristics. Carbohydrate Polymers 181:93–102. https://doi.org/10.1016/j.carbpol.2017.10.040 Zhou Q, Kang H, Bielec M, et al (2018) Influence of different divalent ions cross-linking sodium alginate-polyacrylamide hydrogels on antibacterial properties and wound healing. Carbohydrate Polymers 197:292–304. https://doi.org/10.1016/j.carbpol.2018.05.078 Hung W-S, Tsou C-H, De Guzman M, et al (2014) Cross-Linking with Diamine Monomers To Prepare Composite Graphene Oxide-Framework Membranes with Varying d -Spacing. Chem Mater 26:2983–2990. https://doi.org/10.1021/cm5007873 Reddy N, Reddy R, Jiang Q (2015) Crosslinking biopolymers for biomedical applications. Trends in Biotechnology 33:362–369. https://doi.org/10.1016/j.tibtech.2015.03.008 Yan M, An X, Jiang Z, et al (2022) Effects of cross‐linking with EDC/NHS and genipin on characterizations of self‐assembled fibrillar gel prepared from tilapia collagen and alginate. Polymer Degradation and Stability 200:109929. https://doi.org/10.1016/j.polymdegradstab.2022.109929 Hashem A, Aniagor CO, Mohamed LA, et al (2025) Hydroxypropyl sulfonated starch and Asperlligus oryzae biomass for cationic dye adsorption: characterization, mechanism, sorption modelling. Biomass Conv Bioref 15:1211–1227. https://doi.org/10.1007/s13399-023-05248-9 Bhatia S, Shah YA, Al-Harrasi A, et al (2024) Tuning the structure and physiochemical properties of sodium alginate and chitosan composite films through sodium tripolyphosphate (STPP) crosslinking. International Journal of Biological Macromolecules 264:130463. https://doi.org/10.1016/j.ijbiomac.2024.130463 Jătariu AN, Danu M, Peptu CA, et al (2013) Ionically and Covalently Cross-Linked Hydrogels Based on Gelatin and Chitosan. Soft Materials 11:45–54. https://doi.org/10.1080/1539445X.2011.580409 Jiang S, Qiao C, Wang X, et al (2022) Structure and properties of chitosan/sodium dodecyl sulfate composite films. RSC Adv 12:3969–3978. https://doi.org/10.1039/D1RA08218C Wu M, Zhang X, Zhao Y, et al (2022) A high-performance hydroxide exchange membrane enabled by Cu2+-crosslinked chitosan. Nat Nanotechnol 17:629–636. https://doi.org/10.1038/s41565-022-01112-5 Buhus G, Popa M, Desbrieres J (2009) Hydrogels Based on Carboxymethylcellulose and Gelatin for Inclusion and Release of Chloramphenicol. Journal of Bioactive and Compatible Polymers 24:525–545. https://doi.org/10.1177/0883911509349687 Ni W, Brown SD, Man R (2009) Stacked partial least squares regression analysis for spectral calibration and prediction. Journal of Chemometrics 23:505–517. https://doi.org/10.1002/cem.1246 Wu Y, Feng J (2018) Development and Application of Artificial Neural Network. Wireless Pers Commun 102:1645–1656. https://doi.org/10.1007/s11277-017-5224-x Lamaoui A, García-Guzmán JJ, Amine A, et al (2021) Synthesis techniques of molecularly imprinted polymer composites. In: Molecularly Imprinted Polymer Composites. Elsevier, pp 49–91 Lamuela-Raventós RM (2017) Folin-Ciocalteu method for the measurement of total phenolic content and antioxidant capacity. In: Apak R, Capanoglu E, Shahidi F (eds) Measurement of Antioxidant Activity & Capacity. John Wiley & Sons, Ltd, Chichester, UK, pp 107–115 Additional Declarations No competing interests reported. Supplementary Files Supplementarymaterial.docx scheme1.jpg Scheme 1. Preparation of bio-based MIMs from various biopolymers with the elimination of NSA via (non)-covalent modifications. scheme2.jpg Scheme 2. Template removal by complexation with FC(A); smartphone-based colorimetric detection using MIMs and FC (B). Cite Share Download PDF Status: Published Journal Publication published 21 Mar, 2026 Read the published version in Microchimica Acta → Version 1 posted Editorial decision: Revision requested 25 Jan, 2026 Reviews received at journal 22 Jan, 2026 Reviewers agreed at journal 19 Jan, 2026 Reviewers agreed at journal 28 Sep, 2025 Reviewers invited by journal 27 Sep, 2025 Editor assigned by journal 16 Sep, 2025 Submission checks completed at journal 16 Sep, 2025 First submitted to journal 13 Sep, 2025 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-7609402","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":525747783,"identity":"322ed9a1-6c02-4740-88da-6aab92a0f94d","order_by":0,"name":"Ouarda El Hani","email":"","orcid":"","institution":"Hassan II University of Casablanca","correspondingAuthor":false,"prefix":"","firstName":"Ouarda","middleName":"El","lastName":"Hani","suffix":""},{"id":525747784,"identity":"045c16cd-8522-44ee-a71c-8dbaf3824014","order_by":1,"name":"Khalid Digua","email":"","orcid":"","institution":"Hassan II University of Casablanca","correspondingAuthor":false,"prefix":"","firstName":"Khalid","middleName":"","lastName":"Digua","suffix":""},{"id":525747785,"identity":"988e0ce0-3648-4da0-9959-5f8b8bf00881","order_by":2,"name":"Aziz Amine","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA6ElEQVRIie3RsQrCMBCA4QtiXQ5cHaS+QjqJIPgq6dLNN3BIKehSdPUxdNG1pdAuFdeWDCKuDnWRCiImg+AUFBeH/EPI8nEXAmAy/WWEqxMBGlH1LbHcJbDvxqHzGWkvkqCCyabbb4UX/1ZvoD2LmqtaQzpRzCmkAgfhbh0gE9DJmVWGujGxzxlYAmkx3p5AEijAKlAjegnhETwkOZyPfi1JT5LyriE0JT4nUzUFCVeL0UIN1RAnJwG4c0lyzwnQE+jk7lR0NcTeZ6equooRzRK52FDYdpak5Vn3fBV7u+Prc00mk8n0Q0/92VRpyK2K7AAAAABJRU5ErkJggg==","orcid":"","institution":"Hassan II University of Casablanca","correspondingAuthor":true,"prefix":"","firstName":"Aziz","middleName":"","lastName":"Amine","suffix":""}],"badges":[],"createdAt":"2025-09-13 20:38:19","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7609402/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7609402/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1007/s00604-026-07986-9","type":"published","date":"2026-03-21T15:58:21+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":93137452,"identity":"b4bebe10-7e7d-4de8-9c7c-b4b8d4113db4","added_by":"auto","created_at":"2025-10-09 12:26:12","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":2991651,"visible":true,"origin":"","legend":"","description":"","filename":"Figures.docx","url":"https://assets-eu.researchsquare.com/files/rs-7609402/v1/a58fcc4819bf645d4eee6505.docx"},{"id":93135781,"identity":"27d08bbe-41e3-46af-9bb5-3e487fdffcda","added_by":"auto","created_at":"2025-10-09 12:10:12","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":2684752,"visible":true,"origin":"","legend":"","description":"","filename":"Manuscript.docx","url":"https://assets-eu.researchsquare.com/files/rs-7609402/v1/43d46b1101024d498f906d50.docx"},{"id":93135790,"identity":"272d94ec-8ad8-4358-8e7c-8ea359c341aa","added_by":"auto","created_at":"2025-10-09 12:10:12","extension":"docx","order_by":2,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":20242,"visible":true,"origin":"","legend":"","description":"","filename":"Tables.docx","url":"https://assets-eu.researchsquare.com/files/rs-7609402/v1/edbc8ead18f63a825b5caae3.docx"},{"id":93135791,"identity":"8bb2e7a1-f7de-43dd-87e3-80068ab25269","added_by":"auto","created_at":"2025-10-09 12:10:12","extension":"json","order_by":3,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":5722,"visible":true,"origin":"","legend":"","description":"","filename":"c4956caa9a7c4ce3acbfc2fb93e13aee.json","url":"https://assets-eu.researchsquare.com/files/rs-7609402/v1/1012574b7d5b868a5c201df9.json"},{"id":93135794,"identity":"83502b71-e4bc-4c0d-a0ff-e5f3abe2b348","added_by":"auto","created_at":"2025-10-09 12:10:12","extension":"docx","order_by":4,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":391013,"visible":true,"origin":"","legend":"","description":"","filename":"Supplementarymaterial.docx","url":"https://assets-eu.researchsquare.com/files/rs-7609402/v1/7188f0f718a45f61f4ec58e8.docx"},{"id":93136278,"identity":"68ac070e-6f35-4009-8c18-dbd1b5c66718","added_by":"auto","created_at":"2025-10-09 12:18:12","extension":"xml","order_by":5,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":124647,"visible":true,"origin":"","legend":"","description":"","filename":"c4956caa9a7c4ce3acbfc2fb93e13aee1enriched.xml","url":"https://assets-eu.researchsquare.com/files/rs-7609402/v1/f1c84f74bf792411607bbe47.xml"},{"id":93135792,"identity":"675d03c8-c777-4e6e-b11c-914d107a99c6","added_by":"auto","created_at":"2025-10-09 12:10:12","extension":"png","order_by":6,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":452642,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-7609402/v1/e09b96f18cf3f33d83e81524.png"},{"id":93137454,"identity":"48ac7fe5-1a5a-4c00-bb83-c17341624b53","added_by":"auto","created_at":"2025-10-09 12:26:12","extension":"png","order_by":7,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":425781,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage10.png","url":"https://assets-eu.researchsquare.com/files/rs-7609402/v1/b8679a3d4982d57ba7b44b9d.png"},{"id":93225144,"identity":"bb8820c9-3ff9-418c-b533-084e637100ae","added_by":"auto","created_at":"2025-10-10 12:00:46","extension":"png","order_by":8,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":398209,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage11.png","url":"https://assets-eu.researchsquare.com/files/rs-7609402/v1/b2e9466e5e6e7592dc5ffde7.png"},{"id":93137461,"identity":"0fd3a36b-7b6d-4fac-8b3c-e880a13803e3","added_by":"auto","created_at":"2025-10-09 12:26:12","extension":"jpeg","order_by":9,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":593484,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage12.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7609402/v1/fe944a7e08de1009e9260987.jpeg"},{"id":93136279,"identity":"30b8413e-469b-4216-b483-5125153ca867","added_by":"auto","created_at":"2025-10-09 12:18:12","extension":"png","order_by":10,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":434897,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage13.png","url":"https://assets-eu.researchsquare.com/files/rs-7609402/v1/26f518c50cd9bea70c8ad3b4.png"},{"id":93137956,"identity":"327722de-9879-4539-9cf2-95aa917e6a2d","added_by":"auto","created_at":"2025-10-09 12:34:12","extension":"jpeg","order_by":11,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":358998,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage14.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7609402/v1/092ef29c1d9311504ee6151b.jpeg"},{"id":93137458,"identity":"219d6a6c-c421-4284-b3b8-ea401538673f","added_by":"auto","created_at":"2025-10-09 12:26:12","extension":"png","order_by":12,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":412388,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage15.png","url":"https://assets-eu.researchsquare.com/files/rs-7609402/v1/aea29bb9c4441e4dcf06274e.png"},{"id":93138838,"identity":"fc6bdef8-045d-40e7-b332-31d0c568d9be","added_by":"auto","created_at":"2025-10-09 12:42:12","extension":"png","order_by":13,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":425781,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage10.png","url":"https://assets-eu.researchsquare.com/files/rs-7609402/v1/9e81d920d662dac6c9baf8c8.png"},{"id":93136292,"identity":"78f7d4fa-fb65-4d05-bd54-007465fa6228","added_by":"auto","created_at":"2025-10-09 12:18:12","extension":"png","order_by":14,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":394053,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-7609402/v1/08ca5a65d3171282c44c068b.png"},{"id":93136287,"identity":"e1c493a6-8387-4db6-9a69-cc41d4365f6c","added_by":"auto","created_at":"2025-10-09 12:18:12","extension":"png","order_by":15,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":268208,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-7609402/v1/7ce36b285374b1ad3e4a0434.png"},{"id":93135799,"identity":"f77dc464-c3a0-417f-9806-ce355e5da613","added_by":"auto","created_at":"2025-10-09 12:10:12","extension":"png","order_by":16,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":434897,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage13.png","url":"https://assets-eu.researchsquare.com/files/rs-7609402/v1/58e6424fe846c78854bdf45a.png"},{"id":93135805,"identity":"09bef880-cfdd-4698-b1a5-ddf8381329fd","added_by":"auto","created_at":"2025-10-09 12:10:12","extension":"png","order_by":17,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":151698,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-7609402/v1/4ae8e737b14bd6577b79f3ac.png"},{"id":93136283,"identity":"537889de-0313-49d9-8e75-b48cc1c05ca6","added_by":"auto","created_at":"2025-10-09 12:18:12","extension":"png","order_by":18,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":463304,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage7.png","url":"https://assets-eu.researchsquare.com/files/rs-7609402/v1/0eb6475fefcee8272cec0d5a.png"},{"id":93137462,"identity":"2e2ba3b6-eae0-4aa2-b86b-fae09766ca5e","added_by":"auto","created_at":"2025-10-09 12:26:12","extension":"png","order_by":19,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":424011,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage8.png","url":"https://assets-eu.researchsquare.com/files/rs-7609402/v1/030206cf68243246d9badf3b.png"},{"id":93137960,"identity":"95165b95-9749-4d7f-9a65-28948157c8ea","added_by":"auto","created_at":"2025-10-09 12:34:12","extension":"png","order_by":20,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":452642,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-7609402/v1/b9a92854a3ec75e5fd598005.png"},{"id":93138837,"identity":"9b73ba5f-bdf3-4b4c-a885-6eb0eae49dc6","added_by":"auto","created_at":"2025-10-09 12:42:12","extension":"png","order_by":21,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":94523,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-7609402/v1/1cc6ff0e3801685390919cf5.png"},{"id":93137463,"identity":"cd5ab611-30ea-4bf0-b56f-829b7fafdafa","added_by":"auto","created_at":"2025-10-09 12:26:12","extension":"png","order_by":22,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":45781,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage10.png","url":"https://assets-eu.researchsquare.com/files/rs-7609402/v1/7895e87079b797893c2d9c67.png"},{"id":93135815,"identity":"d52f5c8e-04d2-4452-a68e-0c0d95b46187","added_by":"auto","created_at":"2025-10-09 12:10:12","extension":"png","order_by":23,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":94378,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage11.png","url":"https://assets-eu.researchsquare.com/files/rs-7609402/v1/64b5e7c97b2b7c4a2fabe1a6.png"},{"id":93135818,"identity":"1c32db6b-7149-4685-93a4-89e8c158e63e","added_by":"auto","created_at":"2025-10-09 12:10:13","extension":"png","order_by":24,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":101988,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage12.png","url":"https://assets-eu.researchsquare.com/files/rs-7609402/v1/85ce4f11d247b073e9b3a1cd.png"},{"id":93136290,"identity":"4ae3d395-ac90-498a-b6e4-f61ebf5038ce","added_by":"auto","created_at":"2025-10-09 12:18:12","extension":"png","order_by":25,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":54695,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage13.png","url":"https://assets-eu.researchsquare.com/files/rs-7609402/v1/7b2907977d0f31068c00b522.png"},{"id":93135810,"identity":"278be8aa-9f6c-46fd-9ed6-9588c964c3dc","added_by":"auto","created_at":"2025-10-09 12:10:12","extension":"png","order_by":26,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":62190,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage14.png","url":"https://assets-eu.researchsquare.com/files/rs-7609402/v1/87b2f615cd912a075ee79575.png"},{"id":93135821,"identity":"25564861-9acd-45f4-a401-1bffb1ca5d23","added_by":"auto","created_at":"2025-10-09 12:10:13","extension":"png","order_by":27,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":78152,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage15.png","url":"https://assets-eu.researchsquare.com/files/rs-7609402/v1/15a108620a2a9f4626451526.png"},{"id":93137961,"identity":"a28d4cfe-b828-40b5-9e17-98879d904cab","added_by":"auto","created_at":"2025-10-09 12:34:12","extension":"png","order_by":28,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":45781,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage10.png","url":"https://assets-eu.researchsquare.com/files/rs-7609402/v1/93e8ed313ec7df2f27b5f86d.png"},{"id":93136295,"identity":"9a936e4f-4b17-4e88-8d1b-d78a714c0e41","added_by":"auto","created_at":"2025-10-09 12:18:12","extension":"png","order_by":29,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":106408,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-7609402/v1/290987d4fd330f7b582f1fdf.png"},{"id":93135814,"identity":"f83bc9ed-15cc-490f-86e8-701b3aa25914","added_by":"auto","created_at":"2025-10-09 12:10:12","extension":"png","order_by":30,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":55892,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-7609402/v1/53fc33646f7c2cfbf6a28cec.png"},{"id":93135816,"identity":"a6e24019-7431-47af-be70-d80d139e2961","added_by":"auto","created_at":"2025-10-09 12:10:12","extension":"png","order_by":31,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":54695,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-7609402/v1/c42732e8e9bcb6aaac16e0c5.png"},{"id":93135824,"identity":"3fb5c548-bd5f-4023-9172-58f7a0ea113b","added_by":"auto","created_at":"2025-10-09 12:10:13","extension":"png","order_by":32,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":33322,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-7609402/v1/979234d789955b4209cf59fc.png"},{"id":93136296,"identity":"23ee310d-374c-4648-9825-41fcfca64353","added_by":"auto","created_at":"2025-10-09 12:18:13","extension":"png","order_by":33,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":85524,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage7.png","url":"https://assets-eu.researchsquare.com/files/rs-7609402/v1/ded8e5beee7681119052cdd4.png"},{"id":93135822,"identity":"b8257284-afe7-4a6a-9672-5c51dc9a6a79","added_by":"auto","created_at":"2025-10-09 12:10:13","extension":"png","order_by":34,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":55408,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage8.png","url":"https://assets-eu.researchsquare.com/files/rs-7609402/v1/4e3209f54f9678662639dca3.png"},{"id":93136298,"identity":"76bba308-6689-4802-b513-209fa484752c","added_by":"auto","created_at":"2025-10-09 12:18:13","extension":"png","order_by":35,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":94523,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage9.png","url":"https://assets-eu.researchsquare.com/files/rs-7609402/v1/57c12333b0175e333695ec8e.png"},{"id":93135819,"identity":"03f88820-27a1-41de-9ad4-fd8e8991885c","added_by":"auto","created_at":"2025-10-09 12:10:13","extension":"xml","order_by":36,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":126167,"visible":true,"origin":"","legend":"","description":"","filename":"c4956caa9a7c4ce3acbfc2fb93e13aee1structuring.xml","url":"https://assets-eu.researchsquare.com/files/rs-7609402/v1/8b04e3cd4cf772e95d03286a.xml"},{"id":93136297,"identity":"8752431d-3e9e-4247-b14f-22a4ce405ad4","added_by":"auto","created_at":"2025-10-09 12:18:13","extension":"html","order_by":37,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":135436,"visible":true,"origin":"","legend":"","description":"","filename":"earlyproof.html","url":"https://assets-eu.researchsquare.com/files/rs-7609402/v1/e8a23be642007d92fc2f846a.html"},{"id":93135776,"identity":"9e9514f2-bf66-4b8d-a03a-3874c666fbc3","added_by":"auto","created_at":"2025-10-09 12:10:12","extension":"jpg","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":131780,"visible":true,"origin":"","legend":"\u003cp\u003eVariation of Qs(MIM) and Qs(NIM) for TC using SA \u003cstrong\u003e(A)\u003c/strong\u003e, CMC \u003cstrong\u003e(B)\u003c/strong\u003e, SCNC \u003cstrong\u003e(C)\u003c/strong\u003e, Cs \u003cstrong\u003e(D)\u003c/strong\u003e, Gel \u003cstrong\u003e(E)\u003c/strong\u003e, and CA \u003cstrong\u003e(F)\u003c/strong\u003e-based MIMs modified with different agents\u003cstrong\u003e \u003c/strong\u003e(adsorption test for 5 mg/L of TC).\u003c/p\u003e","description":"","filename":"1.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7609402/v1/eb9d4ec2e72f58ed48da464c.jpg"},{"id":93135780,"identity":"16ab3607-f393-41e2-95b1-3edc4d71cb36","added_by":"auto","created_at":"2025-10-09 12:10:12","extension":"jpg","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":141520,"visible":true,"origin":"","legend":"\u003cp\u003eAdsorption capacities of\u003cstrong\u003e \u003c/strong\u003eCaf \u003cstrong\u003e(A)\u003c/strong\u003e and TC \u003cstrong\u003e(B)\u003c/strong\u003e to different bio-based MIMs (adsorption test at 5 mg/L).\u003cstrong\u003e \u003c/strong\u003eIsothermal study of Caf using Cs-MIM (Cu²⁺ and GA) \u003cstrong\u003e(C) \u003c/strong\u003eand TC using Gel-MIM (EDC/NHS) and SA-MIM (Cu\u003csup\u003e2+\u003c/sup\u003e/PB and EDC/NHS/PDA) \u003cstrong\u003e(D)\u003c/strong\u003e.\u003c/p\u003e","description":"","filename":"2.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7609402/v1/02130a62ee01340617dd42d0.jpg"},{"id":93137955,"identity":"34733b0e-7729-4088-a880-2d77da045543","added_by":"auto","created_at":"2025-10-09 12:34:12","extension":"jpg","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":89524,"visible":true,"origin":"","legend":"\u003cp\u003eFTIR spectra of Cs-MIMs \u003cstrong\u003e(A)\u003c/strong\u003e, Gel-MIMs \u003cstrong\u003e(B)\u003c/strong\u003e, and SA-MIMs \u003cstrong\u003e(C)\u003c/strong\u003e; water contact angle results of Cs-MIMs \u003cstrong\u003e(D)\u003c/strong\u003e, Gel-MIMs \u003cstrong\u003e(E)\u003c/strong\u003e, and SA-MIMs \u003cstrong\u003e(F)\u003c/strong\u003e.\u003c/p\u003e","description":"","filename":"3.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7609402/v1/eb58d5c64be981fa37050480.jpg"},{"id":93135779,"identity":"b8bd19f3-3d1f-473d-b09f-8b7328c88beb","added_by":"auto","created_at":"2025-10-09 12:10:12","extension":"jpg","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":100165,"visible":true,"origin":"","legend":"\u003cp\u003eRecovery performance of bio-based MIMs (Cs-MIM, Gel-MIM, SA-MIM) over 90 days (stability) \u003cstrong\u003e(A)\u003c/strong\u003eand 10 cycles (reusability) \u003cstrong\u003e(B) (\u003c/strong\u003etests for 5 mg/L of Caf and TC).\u003c/p\u003e","description":"","filename":"4.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7609402/v1/13db4b0c4e184edb6aa6522e.jpg"},{"id":93135788,"identity":"c8df750d-b22c-4339-9672-a58246b8ced1","added_by":"auto","created_at":"2025-10-09 12:10:12","extension":"jpg","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":137625,"visible":true,"origin":"","legend":"\u003cp\u003eCalibration curves for smartphone-based colorimetric detection of Caf using Cs-MIM (GA) \u003cstrong\u003e(A) \u003c/strong\u003eand TC using SA-MIM (EDC/NHS/PDA \u003cstrong\u003e(B)\u003c/strong\u003e.\u003cstrong\u003e \u003c/strong\u003eY predicted by the ANN model vs. Y measured for calibration and validation\u003cstrong\u003e (C)\u003c/strong\u003e and prediction in pear, plum, and apple \u003cstrong\u003e(D).\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"5.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7609402/v1/a36be284b21bb9907156b4a2.jpg"},{"id":105223343,"identity":"781cd9e0-46af-4b71-9c42-827f7009f595","added_by":"auto","created_at":"2026-03-23 16:04:30","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1847463,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7609402/v1/06c0cb25-3a7e-4f25-b3da-13a8606fadec.pdf"},{"id":93136274,"identity":"528945d5-ce06-4143-8d9d-9fbbc251d884","added_by":"auto","created_at":"2025-10-09 12:18:12","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":391013,"visible":true,"origin":"","legend":"","description":"","filename":"Supplementarymaterial.docx","url":"https://assets-eu.researchsquare.com/files/rs-7609402/v1/55885afb99e2f7e3911119a2.docx"},{"id":93136277,"identity":"b5a82753-cfd2-4c81-955b-a11f2eebc842","added_by":"auto","created_at":"2025-10-09 12:18:12","extension":"jpg","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":156904,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eScheme 1. \u003c/strong\u003ePreparation of bio-based MIMs from various biopolymers with the elimination of NSA via (non)-covalent modifications.\u003c/p\u003e","description":"","filename":"scheme1.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7609402/v1/127af2842d3f71a10ef6ba40.jpg"},{"id":93135787,"identity":"b8730363-4cd2-4c69-9654-2f3d46491657","added_by":"auto","created_at":"2025-10-09 12:10:12","extension":"jpg","order_by":3,"title":"","display":"","copyAsset":false,"role":"supplement","size":121315,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eScheme 2. \u003c/strong\u003eTemplate removal\u003cstrong\u003e \u003c/strong\u003eby complexation with FC\u003cstrong\u003e(A)\u003c/strong\u003e; smartphone-based colorimetric detection using MIMs and FC \u003cstrong\u003e(B).\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"scheme2.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7609402/v1/1d69cfb82cbeb22ca277bd07.jpg"}],"financialInterests":"No competing interests reported.","formattedTitle":"Generic strategy for high-performance bio-based imprinted membranes free of non-specific adsorption: machine learning-assisted sensing","fulltext":[{"header":"Introduction","content":"\u003cp\u003eMolecular imprinting technology has emerged as an important approach to creating synthetic recognition materials via a lock-and-key mechanism. This process involves the template-induced formation of specific recognition sites within a polymer matrix. The resulting molecularly imprinted polymers (MIPs) have exceptional properties such as selectivity, cost-effectiveness, and stability [\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e, \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e]. These attributes make MIPs good alternatives to natural antibodies [\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e] with diverse applications in sensing [\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e] and separation processes [\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e].\u003c/p\u003e\u003cp\u003eRecently, the scope of molecular imprinting technology has extended to the development of molecularly imprinted membrane (MIM), which integrates the molecular recognition capabilities of MIPs with the membrane characteristics. Beyond the membranes' robustness and ease of handling, they offer unique advantages in selective adsorption and separation, owing to the presence of tailor-made recognition sites embedded within a permeable membrane matrix. Their controllable permeability and surface and internal imprinting capabilities enable precise interaction with target analytes while minimizing interference from structurally similar compounds [\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e]. Unlike conventional MIP synthesis methods, which require long-time synthesis, high energy inputs, and expensive or hazardous monomers and initiators [\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e]. An alternative synthesis approach of alginate and cellulose acetate-based MIM was proposed recently in our laboratory; these biopolymers simultaneously act as the imprinted polymer due to their functional groups, and as the membrane matrix [\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e, \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e]. However, these studies have been limited to only two biopolymers. For this reason, in the present work we propose a generic strategy for successfully preparing bio-based MIMs by testing various well-known biopolymers in membrane science with diverse functional groups, including cellulose derivatives (e.g., cellulose nanofibers, carboxymethyl cellulose (CMC), sulfonated cellulose nanocrystalline (SCNC), and cellulose acetate (CA)), sodium alginate (SA), chitosan (Cs), and gelatin (Gel) [\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e].\u003c/p\u003e\u003cp\u003eThe MIMs demonstrate not only their capability as selective adsorbents for targeted separation but also their versatility as sensing platforms across various applications, including environmental monitoring, food safety, and pharmaceutical analysis, enhancing both selectivity and sensitivity[\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e]. Combining them with smartphone-based detection makes sensor arrays more efficient and suitable for on-site monitoring [\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e]. These advanced systems enable real-time and accurate monitoring of various analytes. The integration of machine learning algorithms [\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e] further enhances the performance of sensor arrays by analyzing complex datasets to improve the sensitivity, rapidity, and precision of detection. Among the widely used machine learning models in sensing were artificial neural networks (ANN) [\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e] and partial least squares (PLS) [\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e].\u003c/p\u003e\u003cp\u003eDespite the advantages of MIPs and MIMs, these materials still face the problem of non-specific adsorption (NSA), which compromises their selectivity and, consequently, their applicability in real-world scenarios. NSA refers to the unintended binding of target molecules to non-specific sites within the material, which significantly diminishes selectivity [\u003cspan additionalcitationids=\"CR16\" citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e]. In the literature, it is often observed that non-imprinted materials, prepared using the same protocol as MIPs/MIMs but without the template, still exhibit an affinity for target molecules, indicating the presence of numerous non-specific sites within the materials. Despite the severity of the NSA issue, it has been addressed in only a limited number of studies [\u003cspan additionalcitationids=\"CR16 CR17\" citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e]. In our laboratory, we have proposed innovative non-covalent approaches to address NSA in both MIPs and MIMs. For MIPs, we developed a strategy involving the modification of MIPs with surfactants [\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e]. Similarly, for MIMs, we introduced a method to eliminate NSA in alginate-based MIMs, involving crosslinking with divalent cations, followed by chelation before template removal [18 ]. This approach significantly enhanced the imprinting performance. On the same concept, Noof et al. modified a methacrylic acid-based MIP using diazomethane to remove NSA, thereby enhancing its selectivity for bisphenol A. These studies highlight the crucial role of NSA elimination in MIP synthesis [\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e]. However, these studies are still in their infancy.\u003c/p\u003e\u003cp\u003eTo enlarge the control of this problem for large types of MIMs, this work proposes a generic innovative approach to bio-based MIMs preparation using diverse biopolymers. Different crosslinking approaches were evaluated to eliminate NSA: covalent and non-covalent. Crosslinkers were selected based on the functional groups present in the biopolymers.\u003c/p\u003e\u003cp\u003eOverall, this study aims to present a generic strategy for preparing bio-based imprinted membranes and studying various biopolymers, such as CMC, SA, CA, Cs, Gel, and SCNC. The research emphasizes a significant step toward developing green and sustainable imprinting technologies. This strategy was tested on two different templates: antibiotic (tetracycline) and phenol (caffeic acid). Moreover, both covalent and non-covalent modification approaches were implemented for NSA elimination. Isothermal evaluations were performed to evaluate the imprinting performance of the bio-based MIMs using various crosslinkers. Additionally, a sensor array of MIMs coupled with smartphone-based colorimetric detection and machine learning was proposed for the precise, selective detection and prediction of caffeic acid in fruits. This work provides a generic framework for preparing bio-based MIM, offering efficient solutions for NSA elimination.\u003c/p\u003e"},{"header":"1. Materials and methods","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e\u003ch2\u003e1.1. Reagents and instrumentation\u003c/h2\u003e\u003cp\u003eCMC (high viscosity extra pure, with carboxymethyl substitution degree of 0.5\u0026ndash;0.8), SA (C\u003csub\u003e6\u003c/sub\u003eH\u003csub\u003e7\u003c/sub\u003eO\u003csub\u003e6\u003c/sub\u003eNa)\u003csub\u003en\u003c/sub\u003e (M\u003csub\u003ew\u003c/sub\u003e=10000\u0026ndash;600000 g/mL, 91%-106%) (SA, BioChemica), Cs, Gel, CA (39.8 wt. % acetyl content. Mn 50000), SCNC, calcium, phosphate buffer solution (PB, 0.5 M, pH\u0026thinsp;=\u0026thinsp;7.0, prepared from anhydrous sodium monobasic phosphate and sodium dibasic phosphate dodecahydrate), sodium sulfate (Na\u003csub\u003e2\u003c/sub\u003eSO\u003csub\u003e4\u003c/sub\u003e, 98%), sodium tripolyphosphate (STPP), cetyltrimethylammonium bromide (CTAB, \u0026ge;\u0026thinsp;98%, LaboChemie), Sodium dodecyl sulfate (SDS, LaboChemie), Borax, glutaraldehyde solution (25% in H\u003csub\u003e2\u003c/sub\u003eO) (GA), epichlorohydrin (EPC), 1-Ethyl-3-(3 dimethylaminopropyl)carbodiimide/N-Hydroxysuccinimide (EDC/NHS, \u0026ge;\u0026thinsp;99%, Sigma), copper, ethylenediamine (EDA, \u0026ge;\u0026thinsp;98%, Labo Chemie), phenylenediamine (PDA, 98%, Sigma), Folin-Ciocalteu (FC, 2N, Sigma), sodium carbonate (Na\u003csub\u003e2\u003c/sub\u003eCO\u003csub\u003e3\u003c/sub\u003e), sodium hydroxide (NaOH), caffeic acid (Caf, \u0026ge;\u0026thinsp;98%, Sigma), and tetracycline (TC, 98\u0026ndash;102%, Sigma).\u003c/p\u003e\u003cp\u003eThe spectrofluorometer POLARStar Omega, from BMG LABTECH (Germany), was used for the absorbance measurement. A Samsung S10 smartphone served as the readout device for smartphone-based colorimetric detection, utilizing the high-efficiency image format. Lab Filter Holders (PP 25mm) integrated to a syringe (10 mL) were used to perform the extraction. FTIR spectroscopy (ThermoFisher Scientific, USA) in ATR mode, covering the spectral range of 650\u0026ndash;4000 cm\u0026sup1;, was used to analyze the chemical structure of the synthesized MIMs.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec4\" class=\"Section2\"\u003e\u003ch2\u003e1.2. Preparation of the bio-based MIMs\u003c/h2\u003e\u003cp\u003eThe preparation of MIMs using each time SA, CMC, SCNC, Cs, CA, or Gel as imprinting polymer was conducted as follows. A 1.5% (w/v) solution of SA, CMS, or SCNC was prepared in water, while 1.5% Cs was dissolved in 1% acetic acid and 5% CA in DMSO. For Gel-based MIMs, a 5% (w/v) solution was solubilized in water. All solutions were stirred for 15 minutes. Subsequently, 0.2% of the template (TC or Caf) was added to the polymer solution. The mixture was stirred for 15 minutes to ensure homogenization and to promote hydrogen bonding between the template and the functional groups of the biopolymer. Then, polymer chains were linked with 1% of borax (due to its affinity to hydroxyl groups [\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e]) to create the initial three-dimensional network around the template. 10 mL of the resulting gel was cast onto an 8.5 cm petri dish and dried for 1h at 50\u0026deg;C to allow solvent evaporation. Bio-based NIMs were prepared using the same protocols as MIMs, without the addition of the template.\u003c/p\u003e\u003cp\u003eTC and Caf templates interacted via hydrogen bonding with the carboxylic groups of SA, CMC, and Gel, the amine groups of Cs and Gel, the sulfone groups of SCNC, and the acetyl functional groups in CA. However, some functional groups are not involved in the cavity creation during binding. Thus, the prepared MIMs and corresponding NIMs were subjected to crosslinking before template removal to block the non-specific sites (as shown in Scheme \u003cspan refid=\"Sch1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). The crosslinkers are selected based on the functional groups in each bio-based MIM (as shown in Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). All further experiments were performed in triplicate.\u003c/p\u003e\u003cdiv id=\"Sec5\" class=\"Section3\"\u003e\u003ch2\u003e1.2.1. Elimination of NSA from SA, CMC, or SCNC-based MIMs\u003c/h2\u003e\u003cp\u003eFor SA, CMC, and SCNC-MIMs/NIMs, various non-covalent crosslinkers and modification agents were tested, including 5% Ca\u0026sup2;⁺/1 mM PB, 1% CTAB, or 1% Cu\u0026sup2;⁺ (known for their affinity to anionic carboxylic groups [\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e, \u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e, \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e]).\u003c/p\u003e\u003cp\u003eAlso, covalent crosslinking with 0.2 M EDC/NHS activating carboxylic groups, followed by reaction with diamines (1% EDA or 1% PDA) [\u003cspan additionalcitationids=\"CR25\" citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e] was performed for SA and CMC-based MIM/NIMs or by adding 1% EPC (to interact with sulfates SO\u003csub\u003e3\u003c/sub\u003e⁻ [\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e]) for SCNC-based MIM/NIM.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec6\" class=\"Section3\"\u003e\u003ch2\u003e1.2.2. Elimination of NSA from Cs-based MIM\u003c/h2\u003e\u003cp\u003eTo remove NSA from Cs-MIM, electrostatic (non-covalent) modification was performed, testing various agents, including 1% Na\u003csub\u003e2\u003c/sub\u003eSO\u003csub\u003e4\u003c/sub\u003e, 1% STPP, 1% SDS, 1% Cu\u0026sup2;⁺ (known for their affinity to NH₃⁺ [\u003cspan additionalcitationids=\"CR29 CR30\" citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e]). Covalent crosslinking using 1% GA (to interact with NH₃⁺ [\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e]).\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec7\" class=\"Section3\"\u003e\u003ch2\u003e1.2.3. Elimination of NSA from Gel and CA-based MIMs\u003c/h2\u003e\u003cp\u003eFor Gel-MIM, non-covalent modification was tested using 1% Na\u003csub\u003e2\u003c/sub\u003eSO\u003csub\u003e4\u003c/sub\u003e, 1% STPP, or 1% SDS, for interactions with NH\u003csub\u003e3\u003c/sub\u003e⁺ groups, and 1% SDS to interact with COO\u003csup\u003e-\u003c/sup\u003e groups.\u003c/p\u003e\u003cp\u003eFor covalent crosslinking, 1% GA was employed to interact with amine groups in Gel by using 0.2 M EDC/NHS by activating the carboxylic groups of Gel, followed by reaction with amine groups (present in Gel) via self-crosslinking.\u003c/p\u003e\u003cp\u003eFor CA-MIM, 1% GA and 1% EPC were used as crosslinkers for OH groups.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003eTested covalent and non-covalent modification mechanisms for NSA elimination in prepared bio-based MIMs.\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"4\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u003cp\u003eBio-based MIM\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003eCrosslinking and modification agent\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u003cp\u003eMechanism\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c4\"\u003e\u003cp\u003eFunction\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003eSA and CMC\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e0.5% Ca\u0026sup2;⁺ + 1 mM PB\u003c/p\u003e\u003cp\u003e1% CTAB\u003c/p\u003e\u003cp\u003e1% Cu\u0026sup2;⁺\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eElectrostatic (non-covalent)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003eCOO\u003csup\u003e-\u003c/sup\u003e\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e0.2 M EDC/NHS/ 1% EDA\u003c/p\u003e\u003cp\u003e(or 1% PDA)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eCovalent\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003eSCNC\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e0.5% Ca\u0026sup2;⁺ + 1 mM PB\u003c/p\u003e\u003cp\u003e1% CTAB\u003c/p\u003e\u003cp\u003e1% Cu\u0026sup2;⁺\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eElectrostatic (non-covalent)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003eSO\u003csub\u003e3\u003c/sub\u003e\u003csup\u003e-\u003c/sup\u003e\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e1% EPC\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eCovalent\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003eCs\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e1% Na\u003csub\u003e2\u003c/sub\u003eSO\u003csub\u003e4\u003c/sub\u003e\u003c/p\u003e\u003cp\u003e1% STPP\u003c/p\u003e\u003cp\u003e1% SDS\u003c/p\u003e\u003cp\u003e1% Cu\u0026sup2;⁺\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eNon-covalent\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003eNH₃⁺\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e1% GA\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eCovalent\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\" morerows=\"3\" rowspan=\"4\"\u003e\u003cp\u003eGel\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e1% Na\u003csub\u003e2\u003c/sub\u003eSO\u003csub\u003e4\u003c/sub\u003e\u003c/p\u003e\u003cp\u003e1% STPP\u003c/p\u003e\u003cp\u003e1% SDS\u003c/p\u003e\u003cp\u003e1% Cu\u0026sup2;⁺\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eNon-covalent\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eNH₃⁺\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e1% CTAB\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eNon-covalent\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eCOO\u003csup\u003e-\u003c/sup\u003e\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e1% GA\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eCovalent\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eNH\u003csub\u003e3\u003c/sub\u003e\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e0.2 M EDC/NHS\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eCovalent\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eCOOH and NH\u003csub\u003e3\u003c/sub\u003e\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eCA\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e1% GA\u003c/p\u003e\u003cp\u003e1% EPC\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eCovalent\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eOH\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec8\" class=\"Section3\"\u003e\u003ch2\u003e1.2.4. Template removal\u003c/h2\u003e\u003cp\u003eAfter the preparation of bio-based MIMs/NIMs, they were cut into 0.5 cm diameter discs using a perforator and then mounted in a membrane holder integrated into a 10 mL syringe. The membranes were initially rinsed with 1 mL of methanol to remove unreacted reagents. Template removal was then carried out using 10 mL of 0.5 N Folin-Ciocalteu (FC)/0.1% NaOH. In a continued way to ensure complete template removal (Scheme \u003cspan refid=\"Sch2\" class=\"InternalRef\"\u003e2\u003c/span\u003eA). The developed approach of extraction was compared to the conventional approach using methanol and methanol: acetic acid (9:1) mixture as extraction solvents. In the conventional approach, the extraction solvents were added, and the process continued until complete removal of the target analytes was achieved. Afterward, the membranes were thoroughly washed with distilled water to eliminate residual molecules such as FC. Template removal was monitored visually and also assessed by UV-Vis spectroscopy.\u003c/p\u003e\u003c/div\u003e\u003c/div\u003e\u003cdiv id=\"Sec9\" class=\"Section2\"\u003e\u003ch2\u003e1.3. Binding study using the prepared bio-based MIMs\u003c/h2\u003e\u003cp\u003eThe binding performance of the bio-based MIMs and NIMs prepared from each biopolymer was assessed by evaluating their binding capacity for templates (TC and Caf) using MIM/NIM disc (0.5 cm, 4 mg) as sorbents in 10 mL of ethanol: water (1:1) template solution at different concentrations (0.5-8 \u0026micro;g/mL for Caf and 0.5-5 \u0026micro;g/mL for TC ) within 15 minutes.\u003c/p\u003e\u003cp\u003eTo further investigate the adsorption behavior of the membranes, the adsorption capacity (Q(mg/g)) was determined through isotherm studies by analyzing the uptake capacities of MIMs and NIMs across varying template concentrations. The adsorption capacity was calculated using the following equation (\u003cb\u003eEq.\u0026nbsp;1\u003c/b\u003e):\u003c/p\u003e\u003c/div\u003e\n\u003cdiv class=\"Heading\"\u003e\u003cb\u003eEq.\u0026nbsp;1\u003c/b\u003e \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\:\\:\\:Q\\:\\left(mg/g\\right)=\\frac{({C}_{initial}-{C}_{final})}{m}\\times\\:V\\)\u003c/span\u003e\u003c/span\u003e\u003c/div\u003e\u003cp\u003eWhere: C\u003csub\u003einitial\u003c/sub\u003e​ (\u0026micro;g/mL) is the initial template concentration, C\u003csub\u003efinal\u003c/sub\u003e (\u0026micro;g/mL) is the template concentration remaining in the solution after adsorption, m(mg) is the weight of the membrane (MIM or NIM), and V (mL) is the loading volume.\u003c/p\u003e\u003cp\u003eImprinting factor (IF) represents the efficiency of each bio-based MIM compared to its corresponding NIM. An IF\u0026thinsp;\u0026gt;\u0026thinsp;1 indicates that the selective sites of the template were successfully created in MIM (the higher the IF, the more performant and selective the MIM is). IF is calculated according to the following equation (\u003cb\u003eEq.\u0026nbsp;2\u003c/b\u003e):\u003c/p\u003e\n\u003cdiv class=\"Heading\"\u003e\u003cb\u003eEq.\u0026nbsp;2\u003c/b\u003e \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\:\\:\\:\\:\\:IF=Q\\left(MIM\\right)/Q\\left(NIM\\right)\\:\\:\\)\u003c/span\u003e\u003c/span\u003e\u003c/div\u003e\u003cp\u003eWhere: Q (MIM) and Q (NIM) were the adsorption capacities of MIM and NIM, respectively.\u003c/p\u003e\u003cdiv id=\"Sec12\" class=\"Section2\"\u003e\u003ch2\u003e1.4. Stability and reusability study\u003c/h2\u003e\u003cp\u003eStability tests were conducted by storing the membranes (Cs-MIM, Gel-MIM, SA-MIM) at room temperature for 90 days, followed by evaluating their adsorption efficiency for 5 mg/L Caf, tested on Cs-MIM (GA) and Cs-MIM (Cu\u003csup\u003e2+\u003c/sup\u003e) or TC, tested on SA-MIM (EDC/NHS/PDA) and Gel-MIM (EDC/NHS). For reusability, these bio-based MIMs underwent 10 consecutive adsorption/desorption cycles: adsorption in water (5 mg/L Caf or TC) and desorption using FC extracting agent. Recoveries are quantified via the developed smartphone-based colorimetric detection.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec13\" class=\"Section2\"\u003e\u003ch2\u003e1.5. Smartphone-based colorimetric detection of phenolic-based compounds using MIMs sensor and FC reagent\u003c/h2\u003e\u003cp\u003eFor both Caf and TC detection, bio-based MIMs (Cs-MIM (GA) for Caf and SA-MIM (EDC/NHS/PDA) for TC), each measuring 0.5 cm in diameter, were placed in the wells of a microplate. Then, 100 \u0026micro;L of the analytes (added in four equal drops of 25 \u0026micro;L, with a 5-minute interval between additions) were introduced at concentrations ranging from 0.1 to 5 \u0026micro;g/mL for TC and 0.05 to 8 \u0026micro;g/mL for Caf. Subsequently, 10 \u0026micro;L of the FC reagent and 100 \u0026micro;L of 0.6 M Na\u003csub\u003e2\u003c/sub\u003eCO\u003csub\u003e3\u003c/sub\u003e were added. The resulting blue coloration was measured using an ELISA reader at 750 nm and analyzed via smartphone imaging, processed with ImageJ software (Scheme \u003cspan refid=\"Sch2\" class=\"InternalRef\"\u003e2\u003c/span\u003eA).\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec14\" class=\"Section2\"\u003e\u003ch2\u003e1.6. MIMs-based sensor arrays for the smartphone detection of Caf in fruits using machine learning\u003c/h2\u003e\u003cdiv id=\"Sec15\" class=\"Section3\"\u003e\u003ch2\u003e1.6.1. Sensor arrays preparation\u003c/h2\u003e\u003cp\u003eFor the detection of Caf in fruit samples, four bio-based MIMs were selected based on their superior affinity and selectivity. These MIMs were Cs-MIM (Cu\u0026sup2;⁺), Cs-MIM (GA), CMC-MIM (EDC/NHS/PDA), and CMC-MIM (Ca\u0026sup2;⁺/PB). The selected MIMs were tested at three concentrations of Caf: 0.1 \u0026micro;g/mL, 0.5 \u0026micro;g/mL, and 1.5 \u0026micro;g/mL.\u003c/p\u003e\u003cp\u003eThe fruit samples were prepared as follows. Fresh fruits were ground using a garlic press to extract their juice, which was then filtered using a nylon filter with a pore size of 0.45 \u0026micro;m. The resulting filtrate was diluted 10-fold with distilled water to ensure compatibility with the detection system. The experimental sensing process followed the same detection protocol as previously established to maintain consistency across all tests.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec16\" class=\"Section3\"\u003e\u003ch2\u003e1.6.2. Calibration and validation using PLS and ANN approaches\u003c/h2\u003e\u003cp\u003eFor calibration and validation data set collection, the [Caf] response for each further sample was measured using a smartphone integrated with image processing software (ImageJ), which analyzed the RGB color space. Each RGB channel was presented as a spectrum ranging from 0 to 255 pixels, resulting in 265 \u0026times; 3 channels\u0026thinsp;=\u0026thinsp;795 spectral variables per sample. Consequently, the size of the calibration dataset was calculated as 4 MIMs \u0026times; 3 added [Caf] \u0026times; 4 replicates\u0026thinsp;=\u0026thinsp;48 samples; and for each sample, 795 variables were collected.\u003c/p\u003e\u003cp\u003eThe RGB data collected from ImageJ were processed and analyzed using Unscrambler software and JMP-Pro. Two multivariate approaches, PLS and ANN, were employed for the regression. PLS was used for modeling relationships between independent variables and dependent variables, reducing dimensionality while maximizing variance [\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e]. The validation involved a 6-fold cross-validation approach. ANN was also employed to model the relationship between input variables and a continuous output, effectively capturing complex patterns and dependencies within the data [\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e]. The ANN model implemented in this study consisted of three hidden layers, allowing for the extraction of complex patterns from RGB data to enhance predictive accuracy. The model was trained using boosting with 100 iterations and a 0.1 learning rate, an iterative technique that improves predictions by continuously adjusting weights based on previous errors.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec17\" class=\"Section3\"\u003e\u003ch2\u003e1.6.3. Prediction of [Caf] in fruit samples\u003c/h2\u003e\u003cp\u003eThe obtained PLS and ANN models acquired from the calibration were applied for the prediction of [Caf] in fruit samples, including plum, pear, and apple. The size of the prediction dataset was 3 fruits \u0026times; 4 MIMs\u0026thinsp;=\u0026thinsp;12 samples (without adding Caf), with 795 variables per sample collected similarly from ImageJ software.\u003c/p\u003e\u003c/div\u003e\u003c/div\u003e"},{"header":"2. Results and discussions","content":"\u003cdiv id=\"Sec19\" class=\"Section2\"\u003e\u003ch2\u003e2.1. Bio-based MIMs as an alternative approach in imprinting technology\u003c/h2\u003e\u003cp\u003eIn the present study, a generic strategy of bio-based MIMs preparation was proposed to enhance both the synthesis performance and the selectivity. The conventional synthesis of MIPs faces several challenges, including long synthesis times (\u0026gt;\u0026thinsp;24 hours), high energy inputs (thermal, ultrasonic, microwave irradiation, or UV exposure [\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e]), and the use of toxic and expensive reagents [\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e]. To overcome these limitations, our approach utilized readily available biopolymers as functional \u0026sup2;polymers for MIMs preparation. These biopolymers are abundant, cost-effective, and green. They serve as ideal candidates for the preparation of MIMs. Additionally, biopolymers with diverse functional groups are tested, offering a versatile strategy for imprinting a wide range of templates, thus enhancing the scope of MIM applications.\u003c/p\u003e\u003cp\u003eThe bio-based MIMs were synthesized in less than 2h without the need for toxic or expensive reagents. The results (summarized in Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e) showed that before modification, Qs(MIM) was higher than that of Qs(NIM), confirming the creation of TC cavities in all bio-based MIMs registering IF values between 1.2 and 1.7, confirming the successful creation of template cavities in all bio-based MIMs. The imprinting performance of these MIMs was further enhanced using both covalent and electrostatic crosslinking approaches.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec20\" class=\"Section2\"\u003e\u003ch2\u003e2.2. NSA elimination from the prepared bio-based MIMs\u003c/h2\u003e\u003cp\u003eThe crosslinking in bio-based MIMs, conducted in the presence of the template, not only facilitates the formation of a stable 3D network around the template, thereby enhancing the recognition cavities by making them complementary in both functional groups and spatial configuration, but also plays a crucial role in eliminating NSA. Specifically, this strategy enables the selective crosslinking of polymer chains in the functional groups not involved in reaction with the template, effectively occupying and blocking free non-specific sites that would otherwise contribute to undesired analyte retention during the rebinding step. By blocking these groups before the template is removed, the resulting material presents minimal accessible non-imprinted sites, ensuring that specific interactions within the imprinted cavities primarily drive recognition. Thus, negligible adsorption in NIM signifies the successful NSA elimination.\u003c/p\u003e\u003cp\u003eBased on Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e, before crosslinking, all bio-based MIMs exhibited IF ranging from 1.3 to 1.5 (\u0026gt;\u0026thinsp;1), indicating moderate imprinting performance and confirming the successful formation of template sites via the proposed approach. The use of biopolymers as both functional polymers and membrane matrices, combined with borax-mediated crosslinking, facilitated the initial creation of a 3D network around the template by occupying the diol functions from the two chains of biopolymers. Despite this, the IF values remained moderate, and the NIMs still displayed high affinity for the template due to the NSA, highlighting the need for further optimization of the system.\u003c/p\u003e\u003cp\u003eFurthermore, the mild conditions applied in this study (room temperature, minimal mechanical stress) ensured that the possible dynamic boronate ester bonds remained intact, preserving the integrity of the imprinted cavities. These findings align with our further experiments, which demonstrated the effectiveness and robustness of the MIMs, particularly after additional covalent or non-covalent crosslinking steps to reinforce the imprinting architecture and eliminate NSA.\u003c/p\u003e\u003cp\u003eFor SA- and CMC-based MIMs (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eA, B), after modification with CTAB or Cu\u0026sup2;⁺, a noticeable reduction in Q(NIM) was observed, suggesting partial elimination of NSA. Crosslinking with Ca\u0026sup2;⁺/PB or EDC/NHS/(EDA or PDA) resulted in the lowest Q(NIM) with IFs of 20\u0026ndash;30 and 67\u0026ndash;80, respectively, indicating effective imprinting enhancement. For SCNC-based MIM (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eC), CTAB or Cu\u0026sup2;⁺ reduced significantly Q(NIM). EPC was the most effective in enhancing the imprinting performance (IF\u0026thinsp;=\u0026thinsp;45). For Cs-based MIM (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eD), Na\u003csub\u003e2\u003c/sub\u003eSO\u003csub\u003e4\u003c/sub\u003e, STPP, and SDS reduced the Q(NIM) significantly but left moderate NSA. Cu\u0026sup2;⁺ and GA, with IF values of 90 and 80, respectively, provided optimal NSA elimination. For Gel-based MIM (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eE), Na\u003csub\u003e2\u003c/sub\u003eSO\u003csub\u003e4\u003c/sub\u003e, STPP, SDS, CTAB, GA, and Cu\u003csup\u003e2+\u003c/sup\u003e provided lower Q(NIM). Self-crosslinking with EDC/NHS further enhanced the imprinting performance (IF\u0026thinsp;=\u0026thinsp;10).\u003c/p\u003e\u003cp\u003eDespite the improvement of CA-based MIM (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eF), NIM adsorption capacity could not be completely eliminated due to the inherent difficulty in crosslinking acetyl functions in CA without decomposing them, which limits the complete suppression of NSA.\u003c/p\u003e\u003cp\u003eThe previously discussed modifications for NSA elimination were also tested on different days to evaluate the variability of the generic strategy. Consistent results were obtained, demonstrating the robustness of the method with a coefficient of variation below 5%.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec21\" class=\"Section2\"\u003e\u003ch2\u003e2.3. Template removal\u003c/h2\u003e\u003cp\u003eThe removal of templates from MIMs is a critical step that significantly influences their selectivity and performance. Conventional extraction methods often require prolonged processing times due to the high affinity of the templates for the imprinted matrix, and the use of toxic, harsh solvents can damage the cavity structures, ultimately reducing the material\u0026rsquo;s selectivity. Therefore, developing a green, rapid, and simple extraction approach is of high importance. In this work, template removal was achieved via complexation with the FC reagent. The FC is well-known for its high affinity to phenolic compounds, forming a stable blue complex that is structurally incompatible with the original cavity architecture. This prevents re-adsorption and facilitates the template release while preserving the integrity of the imprinted sites. In addition, the developed MIMs were designed such that template binding occurred predominantly at or near the surface of the membrane, as is typical in membrane-imprinted systems. This superficial binding significantly reduces the diffusion path length, making template accessibility feasible even in the presence of covalent crosslinking (e.g., via GA in Cs-MIM or EDC/NHS in SA-MIM ).\u003c/p\u003e\u003cp\u003eA comparative investigation was conducted to examine the removal of TC and Caf using two approaches: the complexation reaction with FC versus conventional elution with methanol: Acetic acid (9:1) and pure methanol. As depicted in \u003cb\u003eTable \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e\u003c/b\u003e, the conventional methods required a time from 6 h to 10 h for extraction, whereas the FC-based complexation achieved complete template removal within only 20 min. This high extraction efficiency is attributed to the superior complexation capability of the FC reagent. Moreover, the continuous extraction strategy used here offers a distinct advantage over traditional methods that involve soaking the MIM in an elution solution, which often face the risk of template re-adsorption and release equilibrium. The use of a syringe-coupled MIM scaffold further enhances the process by enabling extraction under controlled flux, thereby accelerating continuous extraction and ensuring effective template removal. This highlights the effectiveness of the proposed removal method as a rapid and environmentally friendly method for phenolic templates removal in MIMs.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec22\" class=\"Section2\"\u003e\u003ch2\u003e2.4. Binding study of the bio-based MIMs for TC and Caf templates\u003c/h2\u003e\u003cp\u003eAfter selecting the optimum modifications for each bio-based MIM, the proposed generic strategy was applied to two different templates: Caf and TC. The results presented in Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e highlight the Q(mg/g) of Caf and TC using bio-based MIMs and their corresponding NIMs. The significant differences in Q(MIM) and Q(NIM) confirm the successful creation of template cavities in the MIMs. However, the affinity of each template for each bio-based MIM varies due to differences in the interactions between the functional groups in the bio-based MIMs and the templates, as well as the effectiveness of the crosslinking strategies. Despite the difference in the affinity of the bio-based MIMs for each template, NSA elimination using the optimal crosslinking agents showed similar improvements in imprinting performance. Therefore, this generic strategy can be applied regardless of the template.\u003c/p\u003e\u003cp\u003eFor Caf (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eA), the highest Q(mg/g) was observed in Cs-based MIMs. This is attributed to the abundant amino groups in Cs, which enable strong hydrogen interactions with phenolic and carboxylic groups of Caf. Then, CMC and SA-based MIMs demonstrated lower, but still significant, affinity. In these cases, the carboxyl groups in CMC and SA facilitated moderate interactions with Caf, though their affinity was slightly lower than that of Cs-based MIMs. This difference in affinity may be attributed to the absence of amino groups, which are present in the Cs-based MIMs and likely contribute to stronger interactions. Gel-based MIMs, despite having abundant amine functions, showed reduced affinity due to limited spatial accessibility for interactions compared to Cs. CA and SCNC-based MIMs exhibited the lowest adsorption capacities, as acetyl groups and sulfate have limited interaction potential with phenols and acids.\u003c/p\u003e\u003cp\u003eIn the case of TC (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eB), Gel-, SA-, and CMC-based MIMs showed the highest affinity. This is due to the multiple carboxyl groups in SA and CMC, which interact favorably with the amino and amide groups in TC, forming strong hydrogen bonds, as well as the amide and carboxyl groups in Gel. Cs-based MIMs demonstrated moderate adsorption, while CA- and SCNC-based MIMs exhibited the weakest binding.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec23\" class=\"Section2\"\u003e\u003ch2\u003e2.5. Isothermal study of Caf and TC using the MIMs\u003c/h2\u003e\u003cp\u003eAfter selecting the optimum bio-based MIMs for each template, the isothermal study for Caf and TC, conducted over specific concentration ranges (0.5-8 mg/L for Caf and TC), highlights the adsorption behavior of the selected MIM and NIM.\u003c/p\u003e\u003cp\u003eFor Caf (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eC), the Q increases progressively with concentration of Caf for both Cs-MIM (Cu\u003csup\u003e2+\u003c/sup\u003e) and Cs-MIM (GA). The Q of both MIMs far exceeds that of their respective NIMs, confirming the successful creation of the cavity and elimination of NSA.\u003c/p\u003e\u003cp\u003eFor TC (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eD\u003cb\u003e)\u003c/b\u003e, Gel-MIM (EDC/NHS), SA-MIM (Ca\u003csup\u003e2+\u003c/sup\u003e/PB), and SA-MIM (EDC/NHS/PDA) demonstrate a significant increase in Q(MIMs). The corresponding NIMs display non-adsorption across all concentrations, confirming the elimination of NSA. For Gel-NIM (EDC/NHS), NSA was not observed at lower concentrations except at 5 mg/L, though it remained insignificant compared to its corresponding Gel-MIM.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec24\" class=\"Section2\"\u003e\u003ch2\u003e2.6. Characterizations\u003c/h2\u003e\u003cdiv id=\"Sec25\" class=\"Section3\"\u003e\u003ch2\u003e2.6.1. FT-IR results\u003c/h2\u003e\u003cp\u003eIn Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eA, the FT-IR spectrum for Cs-MIM showed the prominent peaks of Cs, including a broad band at 3200 cm⁻\u0026sup1; attributed to O\u0026ndash;H and NH₂ groups, a peak at 2880 cm⁻\u0026sup1; for CH vibrations, a peak at 1560 cm⁻\u0026sup1; for NH₂ stretching, and a peak at 1030 cm⁻\u0026sup1; corresponding to C\u0026ndash;O vibrations from the pyranose ring. After modification with Cu\u0026sup2;⁺, the 2880 cm⁻\u0026sup1; peak disappears, and the NH₂ stretching peak shifts to 1650 cm⁻\u0026sup1; due to Cu\u003csup\u003e2+\u003c/sup\u003e-induced crosslinking that alters the chemical environment of the amino groups; concurrently, the pyranose ring peak becomes sharper and more intense, suggesting a more ordered structure. After modification with GA, significant spectral changes are observed: the appearance of a new peak at 1660 cm⁻\u0026sup1;, indicative of C\u0026thinsp;=\u0026thinsp;N stretching and confirming Schiff base formation, an increase in C\u0026ndash;H stretching intensity from the additional alkyl groups in GA, notable shifts and enhanced intensity of the C\u0026ndash;O related peak at 1050 cm⁻\u0026sup1; reflecting alterations in the polysaccharide backbone, also a new peak at 950 cm⁻\u0026sup1; attributed to changes in glycosidic linkages after crosslinking.\u003c/p\u003e\u003cp\u003eFor Gel-MIM (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eB), before modification exhibited characteristic peaks including a broad band around 3300 cm⁻\u0026sup1; for O\u0026ndash;H and N\u0026ndash;H stretching, and distinct peaks around 1650 cm⁻\u0026sup1; and 1550 cm⁻\u0026sup1; for amide groups, with additional bands between 1100 and 1500 cm⁻\u0026sup1; corresponding to C-O and C-N free functional groups. After self-crosslinking with EDC/NHS, there is a noticeable reduction in the intensity of these primary peaks and the disappearance of the bands in the 1100\u0026ndash;1500 cm\u003csup\u003e-\u003c/sup\u003e\u0026sup1; range. This indicates that the free amine and carboxyl groups have been linked, leading to a crosslinked Gel structure.\u003c/p\u003e\u003cp\u003eIn Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eC, the FTIR spectrum of SA-MIM displays the characteristic peaks of SA: a broad band at 3270 cm⁻\u0026sup1; due to OH stretching, peaks at 1580 cm⁻\u0026sup1; and 1400 cm⁻\u0026sup1; corresponding to the asymmetric and symmetric stretching vibrations of \u0026ndash;COO⁻ groups, and a peak at 1020 cm⁻\u0026sup1; from C\u0026ndash;O stretching in the polysaccharide backbone. Upon crosslinking with Ca\u0026sup2;⁺/PB, the peak at 1580 cm⁻\u0026sup1; shifts to 1600 cm⁻\u0026sup1; and that at 1400 cm⁻\u0026sup1; shifts to 1430 cm⁻\u0026sup1;, suggesting the formation of ionic bonds between Ca\u0026sup2;⁺ and the \u0026ndash;COO⁻ groups, with the appearance of a distinct peak around 950 cm⁻\u0026sup1; further indicating the establishment of a crosslinked network. Additionally, modification with EDC/NHS/PDA causes a shift from 1580 cm⁻\u0026sup1; to 1607 cm⁻\u0026sup1;, which is attributed to the conversion of the \u0026ndash;COO⁻ into amide bonds after activation by EDC/NHS and reaction with PDA, thereby confirming the successful modification of SA-MIM.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec26\" class=\"Section3\"\u003e\u003ch2\u003e2.6.2. Water contact angle results\u003c/h2\u003e\u003cp\u003eAs shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e \u003cb\u003e(D, E\u003c/b\u003e, and \u003cb\u003eF)\u003c/b\u003e, Cs-MIM, Gel-MIM, and SA-MIM exhibited remarkable hydrophilicity before crosslinking with contact angles of 43\u0026deg;, 39\u0026deg;, and 27\u0026deg;, respectively. After non-covalent crosslinking, with Cu\u0026sup2;⁺ for Cs-MIM and Ca\u0026sup2;⁺/PB for SA-MIM, the contact angles increased to 75\u0026deg; and 74\u0026deg;, respectively. Covalent crosslinking further decreased hydrophilicity, with contact angles rising to 94\u0026deg; for Cs-MIM (GA), 78\u0026deg; for Gel-MIM (EDC/NHS), and 74\u0026deg; for SA-MIM (EDC/NHS/PDA); this is due to the formation of stable covalent bonds that consume free hydrophilic groups and introduce more hydrophobic moieties.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003c/div\u003e\u003c/div\u003e\u003cdiv id=\"Sec27\" class=\"Section2\"\u003e\u003ch2\u003e2.7. Stability and reusability performance\u003c/h2\u003e\u003cp\u003eAs shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eA, the stability over 90 days demonstrates that all membranes exhibit excellent long-term performance, with recovery rates consistently above 80%. While Cs-MIM (Cu\u0026sup2;⁺) shows slightly lower values, its overall performance remains remarkable.\u003c/p\u003e\u003cp\u003eSimilarly, Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eB illustrates the reusability over 10 cycles. Cs-MIM (GA) maintains reliable performance for Caf adsorption/desorption, retaining high recovery rates even after repeated cycles (\u0026gt;\u0026thinsp;80%). Likewise, SA-MIM (EDC/NHS/PDA) and Gel-MIM (EDC/NHS) exhibit stable reusability for TC adsorption/desorption, underscoring the practical suitability of bio-based MIMs for repeated applications due to efficient covalent crosslinking, which enhances their mechanical properties. Cs-MIM (Cu\u003csup\u003e2+\u003c/sup\u003e) and SA-MIM (Ca\u003csup\u003e2+\u003c/sup\u003e/PB) showed as well as remarkable recoveries up to cycle 8. After that, its regeneration efficiency reduces due to the limited stability of non-covalent crosslinking compared to covalent bonding.\u003c/p\u003e\u003cp\u003eOverall, the MIMs developed via electrostatic crosslinking effectively reduce NSA but exhibit limited stability. To overcome this limitation, covalent crosslinking was introduced to enhance imprinting performance and expand the applicability of these MIMs, making these MIMs more suitable for complex matrices, even under challenging conditions.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec28\" class=\"Section2\"\u003e\u003ch2\u003e2.8. The smartphone-based colorimetric detection using MIMs and FC reagent\u003c/h2\u003e\u003cp\u003eFC reagent is widely recognized for its ability to interact with phenols, such as Caf, It reacts with phenolic hydroxyl groups under alkaline conditions, producing a blue chromophore that absorbs at 750 nm [\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e]. This reaction occurs due to the interaction of the FC reagent with the phenol group, making the method suitable for the colorimetric detection of any molecule containing phenolic groups. In this study, the FC-based colorimetric method applied for the first time for TC detection, to the best of our knowledge.\u003c/p\u003e\u003cp\u003eAfter the development of blue coloration, images of the resulting color intensities were captured using a smartphone and analyzed with ImageJ software. The RGB (red-green-blue) intensities were measured and correlated with analyte concentrations. After subtracting the blank values, significant increases in all three channels (R, G, and B) were observed with rising analyte concentrations, as illustrated in Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e5\u003c/span\u003e.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eSubsequently, specific channels were selected to construct calibration curves. For Caf (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e5\u003c/span\u003eA), a remarkable R\u0026sup2; value of 0.985, a limit of detection (LOD) of 0.003 mg/L, and a limit of quantification (LOQ) of 0.015 mg/L were obtained. These limits were calculated based on calibration curve approach as follows (\u003cb\u003eEq.\u0026nbsp;3\u003c/b\u003e):\u003c/p\u003e\u003c/div\u003e\n\u003cdiv class=\"Heading\"\u003e\u003cb\u003eEq.\u0026nbsp;3\u003c/b\u003e \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:LOQ=10\\times\\:Blank\\:standard\\:deviation/Slope\\:\\)\u003c/span\u003e\u003c/span\u003e\u003c/div\u003e\u003cp\u003eSimilarly, for TC (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e5\u003c/span\u003eB), R\u0026sup2; value of 0.983, LOD of 0.03 mg/L, and LOQ of 0.1 mg/L were obtained. These results confirm the effectiveness of smartphone-based colorimetric detection and bio-based MIM binding performance in quantifying the concentrations of both analytes with high sensitivity and reliability.\u003c/p\u003e\u003cdiv id=\"Sec30\" class=\"Section2\"\u003e\u003ch2\u003e2.9. Application of MIMs in sensor arrays for the detection of Caf in fruits\u003c/h2\u003e\u003cdiv id=\"Sec31\" class=\"Section3\"\u003e\u003ch2\u003e2.9.1. PLS regression model\u003c/h2\u003e\u003cp\u003eThe PLS regression results for Caf detection are detailed in the \u003cb\u003eSupplementary material\u003c/b\u003e.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec32\" class=\"Section3\"\u003e\u003ch2\u003e2.9.2. ANN regression model\u003c/h2\u003e\u003cp\u003eTo assess the accuracy of the ANN model, the correlation between measured and predicted values was analyzed. A strong correlation was observed in the calibration, with R\u003csup\u003e2\u003c/sup\u003e\u0026thinsp;=\u0026thinsp;0.989, R\u003csup\u003e2\u003c/sup\u003eadj\u0026thinsp;=\u0026thinsp;0.974, and a low Root Mean Square Error of Calibration (RMSEC) of 0.01, indicating excellent model fit and accuracy. Similarly, in validation, the ANN model exhibited good performance, achieving R\u003csup\u003e2\u003c/sup\u003e\u0026thinsp;=\u0026thinsp;0.979, R\u003csup\u003e2\u003c/sup\u003eadj\u0026thinsp;=\u0026thinsp;0.958, and a RMSEV\u0026thinsp;=\u0026thinsp;0.02 (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e5\u003c/span\u003eC). Furthermore, the prediction results (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e5\u003c/span\u003eD) showed R\u003csup\u003e2\u003c/sup\u003e\u0026thinsp;=\u0026thinsp;0.970, R\u003csup\u003e2\u003c/sup\u003eadj\u0026thinsp;=\u0026thinsp;0.967, and a RMSEP\u0026thinsp;=\u0026thinsp;0.05. Additionally, the notably lower RMSE values and higher R\u003csup\u003e2\u003c/sup\u003e compared to PLS further highlight the superior performance of ANN. Among the tested samples based on Cs-MIM (Cu\u003csup\u003e2+\u003c/sup\u003e), plum exhibited the highest Caf concentration (1.9 mg/L), followed by pear (1.3 mg/L), while apple had the lowest concentration (0.8 mg/L) using the ANN model.\u003c/p\u003e\u003cp\u003eIn addition to the machine learning efficiency in sensing, the remarkable results are attributed to the high imprinting performance of the developed bio-based MIMs, which enable selective detection of Caf in complex matrices, as well as the reliability of the smartphone-based detection method. Together, contributed to the overall effectiveness and robustness of the proposed sensing strategy.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003c/div\u003e\u003c/div\u003e"},{"header":"Conclusions","content":"\u003cp\u003eThis study demonstrates the successful development of bio-based MIMs as a promising alternative to conventional synthetic antibodies. By utilizing various biopolymers with distinct functional groups, including Cs, SA, CMC, SCNC, CA, and Gel, we have shown the potential of the proposed strategy to address common limitations of synthetic antibodies. These MIMs were prepared using a rapid, green, and effective strategy. Also, the selectivity was significantly enhanced. The IF was improved considerably through covalent and non-covalent crosslinking with specific agents, increasing from initial values around 1.5 (for 5 mg/L) before NSA elimination to much higher values. For SA and CMC, crosslinking with Ca\u0026sup2;⁺/phosphate achieved an IF of 30, and EDC/NHS/diamine crosslinking improved the IF to 80. SCNC showed an IF of 45 with EPC, while Cs achieved an IF of 90 with Cu\u0026sup2;⁺ and 80 with GA. For Gel, crosslinking with EDC/NHS resulted in an IF of 10. These enhancements were observed for both TC and Caf templates.\u003c/p\u003e\u003cp\u003eIn addition, a sensor array using MIMs for smartphone-based detection of Caf coupled with PLS and ANN models was developed. The ANN model demonstrated strong predictive performance, with high values (0.989\u0026ndash;0.970) and low RMSE (0.01\u0026ndash;0.05) across calibration, validation, and prediction. These results confirm its accuracy, reliability, and potential for rapid Caf detection in complex fruit matrices.\u003c/p\u003e\u003cp\u003eOverall, this work not only enhances the selectivity and performance of bio-based MIMs but also paves the way for sustainable, cost-effective, and efficient alternatives to synthetic antibodies.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eCRediT authorship contribution statement\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eOuarda El Hani\u003c/strong\u003e: Investigation; Conceptualization; Data curation; Formal analysis; Methodology; Software; Writing-original draft\u003cstrong\u003e. Khalid Digua:\u0026nbsp;\u003c/strong\u003eSupervision; Writing \u0026ndash; review \u0026amp; editing; Validation\u003cstrong\u003e. Aziz Amine:\u0026nbsp;\u003c/strong\u003eSupervision; Conceptualization; Resources; Writing \u0026ndash; review \u0026amp; editing; Validation.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAcknowledgements\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe thank Prof. Nabil El Mo\u0026ccedil;ayd from Mohammed VI Polytechnic University (Morocco) for his valuable remarks and recommendations regarding machine learning models. We also acknowledge the support of the PRIMA project, \u0026ldquo;Agro Food Waste Recovery: New Processing Technologies for Food Safety and Packaging (FoWRSaP)\u0026rdquo;.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eEthics approval\u003c/strong\u003e: Not applicable.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eSource of biological material:\u003c/strong\u003e Not applicable.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eStatement on animal welfare\u003c/strong\u003e: Not applicable.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConflict of interest:\u0026nbsp;\u003c/strong\u003eThe authors have no financial or non-financial conflict of interest to disclose.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting of interest:\u0026nbsp;\u003c/strong\u003eThe authors have no financial or non-financial competing interests to disclose.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e: No funding was received to assist with the preparation of this manuscript.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eBelBruno JJ (2019) Molecularly Imprinted Polymers. Chem Rev 119:94\u0026ndash;119. https://doi.org/10.1021/acs.chemrev.8b00171\u003c/li\u003e\n\u003cli\u003eTeixeira SPB, Reis RL, Peppas NA, et al (2021) Epitope-imprinted polymers: Design principles of synthetic binding partners for natural biomacromolecules. Sci Adv 7:eabi9884. https://doi.org/10.1126/sciadv.abi9884\u003c/li\u003e\n\u003cli\u003eSmolinska-Kempisty K, Guerreiro A, Canfarotta F, et al (2016) A comparison of the performance of molecularly imprinted polymer nanoparticles for small molecule targets and antibodies in the ELISA format. Sci Rep 6:37638. https://doi.org/10.1038/srep37638\u003c/li\u003e\n\u003cli\u003eSaadatidizaji Z, Sohrabi N, Mohammadi R (2024) Development of a simple polymer-based sensor for detection of the Pirimicarb pesticide. Sci Rep 14:10293. https://doi.org/10.1038/s41598-024-60748-6\u003c/li\u003e\n\u003cli\u003eRutkowska M, Płotka-Wasylka J, Morrison C, et al (2018) Application of molecularly imprinted polymers in analytical chiral separations and analysis. TrAC Trends in Analytical Chemistry 102:91\u0026ndash;102. https://doi.org/10.1016/j.trac.2018.01.011\u003c/li\u003e\n\u003cli\u003eRatnaningsih E, Kadja GTM, Putri RM, et al (2022) Molecularly Imprinted Affinity Membrane: A Review. ACS Omega 7:23009\u0026ndash;23026. https://doi.org/10.1021/acsomega.2c02158\u003c/li\u003e\n\u003cli\u003eMartins RO, Bernardo RA, Machado LS, et al (2023) Greener molecularly imprinted polymers: Strategies and applications in separation and mass spectrometry methods. TrAC Trends in Analytical Chemistry 168:117285. https://doi.org/10.1016/j.trac.2023.117285\u003c/li\u003e\n\u003cli\u003eEl Hani O, Garc\u0026iacute;a-Guzm\u0026aacute;n JJ, Palacios-Santander JM, et al (2024) Development of a molecularly imprinted membrane for selective, high-sensitive, and on-site detection of antibiotics in waters and drugs: Application for sulfamethoxazole. Chemosphere 350:141039. https://doi.org/10.1016/j.chemosphere.2023.141039\u003c/li\u003e\n\u003cli\u003eEl Hani O, Garc\u0026iacute;a-Guzm\u0026aacute;n JJ, Palacios-Santander JM, et al (2024) Molecularly imprinted alginate-PVP membrane for the selective fluorescence detection of flavonoids in foods: A practical case for rutin. Sensors and Actuators B: Chemical 403:135232. https://doi.org/10.1016/j.snb.2023.135232\u003c/li\u003e\n\u003cli\u003eGaliano F, Brice\u0026ntilde;o K, Marino T, et al (2018) Advances in biopolymer-based membrane preparation and applications. Journal of Membrane Science 564:562\u0026ndash;586. https://doi.org/10.1016/j.memsci.2018.07.059\u003c/li\u003e\n\u003cli\u003eWang M, Cet\u0026oacute; X, Del Valle M (2022) A novel electronic tongue using electropolymerized molecularly imprinted polymers for the simultaneous determination of active pharmaceutical ingredients. Biosensors and Bioelectronics 198:113807. https://doi.org/10.1016/j.bios.2021.113807\u003c/li\u003e\n\u003cli\u003eYang M, Zhang M, Jia M (2023) Optical sensor arrays for the detection and discrimination of natural products. Nat Prod Rep 40:628\u0026ndash;645. https://doi.org/10.1039/D2NP00065B\u003c/li\u003e\n\u003cli\u003eYang J, Lu S, Chen B, et al (2023) Machine learning-assisted optical nano-sensor arrays in microorganism analysis. TrAC Trends in Analytical Chemistry 159:116945. https://doi.org/10.1016/j.trac.2023.116945\u003c/li\u003e\n\u003cli\u003eSaeed R, Glamuzina B, Tuyet Nga MT, et al (2025) Supervised learning-based artificial senses for non-destructive fish quality classification. Biosensors and Bioelectronics 267:116770. https://doi.org/10.1016/j.bios.2024.116770\u003c/li\u003e\n\u003cli\u003eLamaoui A, Palacios-Santander JM, Amine A, Cubillana-Aguilera L (2021) Molecularly imprinted polymers based on polydopamine: Assessment of non-specific adsorption. Microchemical Journal 164:106043. https://doi.org/10.1016/j.microc.2021.106043\u003c/li\u003e\n\u003cli\u003eKarrat A, Amine A (2024) Innovative approaches to suppress non-specific adsorption in molecularly imprinted polymers for sensing applications. Biosensors and Bioelectronics 250:116053. https://doi.org/10.1016/j.bios.2024.116053\u003c/li\u003e\n\u003cli\u003eLi X, Zhang B, Tian L, et al (2015) Improvement of recognition specificity of surface protein-imprinted magnetic microspheres by reducing nonspecific adsorption of competitors using 2-methacryloyloxyethyl phosphorylcholine. Sensors and Actuators B: Chemical 208:559\u0026ndash;568. https://doi.org/10.1016/j.snb.2014.11.045\u003c/li\u003e\n\u003cli\u003eHani OE, Digua K, Amine A (2025) Elimination of non-specific adsorption in the molecularly imprinted membrane: application for tetracycline detection. Anal Bioanal Chem. https://doi.org/10.1007/s00216-025-05804-4\u003c/li\u003e\n\u003cli\u003eKarrat A, Amine A (2024) Innovative approaches to suppress non-specific adsorption in molecularly imprinted polymers for sensing applications. Biosensors and Bioelectronics 250:116053. https://doi.org/10.1016/j.bios.2024.116053\u003c/li\u003e\n\u003cli\u003eAlenazi NA, Lai EPC, Manthorpe JM (2014) Enhanced selectivity of a molecularly imprinted polymer toward the target molecule via esterification of non‐specific binding sites with diazomethane. J of Molecular Recognition 27:755\u0026ndash;762. https://doi.org/10.1002/jmr.2402\u003c/li\u003e\n\u003cli\u003eTanpichai S, Phoothong F, Boonmahitthisud A (2022) Superabsorbent cellulose-based hydrogels cross-liked with borax. Sci Rep 12:8920. https://doi.org/10.1038/s41598-022-12688-2\u003c/li\u003e\n\u003cli\u003eMirtič J, Ila\u0026scaron; J, Kristl J (2018) Influence of different classes of crosslinkers on alginate polyelectrolyte nanoparticle formation, thermodynamics and characteristics. Carbohydrate Polymers 181:93\u0026ndash;102. https://doi.org/10.1016/j.carbpol.2017.10.040\u003c/li\u003e\n\u003cli\u003eZhou Q, Kang H, Bielec M, et al (2018) Influence of different divalent ions cross-linking sodium alginate-polyacrylamide hydrogels on antibacterial properties and wound healing. Carbohydrate Polymers 197:292\u0026ndash;304. https://doi.org/10.1016/j.carbpol.2018.05.078\u003c/li\u003e\n\u003cli\u003eHung W-S, Tsou C-H, De Guzman M, et al (2014) Cross-Linking with Diamine Monomers To Prepare Composite Graphene Oxide-Framework Membranes with Varying \u003cem\u003ed\u003c/em\u003e -Spacing. Chem Mater 26:2983\u0026ndash;2990. https://doi.org/10.1021/cm5007873\u003c/li\u003e\n\u003cli\u003eReddy N, Reddy R, Jiang Q (2015) Crosslinking biopolymers for biomedical applications. Trends in Biotechnology 33:362\u0026ndash;369. https://doi.org/10.1016/j.tibtech.2015.03.008\u003c/li\u003e\n\u003cli\u003eYan M, An X, Jiang Z, et al (2022) Effects of cross‐linking with EDC/NHS and genipin on characterizations of self‐assembled fibrillar gel prepared from tilapia collagen and alginate. Polymer Degradation and Stability 200:109929. https://doi.org/10.1016/j.polymdegradstab.2022.109929\u003c/li\u003e\n\u003cli\u003eHashem A, Aniagor CO, Mohamed LA, et al (2025) Hydroxypropyl sulfonated starch and Asperlligus oryzae biomass for cationic dye adsorption: characterization, mechanism, sorption modelling. Biomass Conv Bioref 15:1211\u0026ndash;1227. https://doi.org/10.1007/s13399-023-05248-9\u003c/li\u003e\n\u003cli\u003eBhatia S, Shah YA, Al-Harrasi A, et al (2024) Tuning the structure and physiochemical properties of sodium alginate and chitosan composite films through sodium tripolyphosphate (STPP) crosslinking. International Journal of Biological Macromolecules 264:130463. https://doi.org/10.1016/j.ijbiomac.2024.130463\u003c/li\u003e\n\u003cli\u003eJătariu AN, Danu M, Peptu CA, et al (2013) Ionically and Covalently Cross-Linked Hydrogels Based on Gelatin and Chitosan. Soft Materials 11:45\u0026ndash;54. https://doi.org/10.1080/1539445X.2011.580409\u003c/li\u003e\n\u003cli\u003eJiang S, Qiao C, Wang X, et al (2022) Structure and properties of chitosan/sodium dodecyl sulfate composite films. RSC Adv 12:3969\u0026ndash;3978. https://doi.org/10.1039/D1RA08218C\u003c/li\u003e\n\u003cli\u003eWu M, Zhang X, Zhao Y, et al (2022) A high-performance hydroxide exchange membrane enabled by Cu2+-crosslinked chitosan. Nat Nanotechnol 17:629\u0026ndash;636. https://doi.org/10.1038/s41565-022-01112-5\u003c/li\u003e\n\u003cli\u003eBuhus G, Popa M, Desbrieres J (2009) Hydrogels Based on Carboxymethylcellulose and Gelatin for Inclusion and Release of Chloramphenicol. Journal of Bioactive and Compatible Polymers 24:525\u0026ndash;545. https://doi.org/10.1177/0883911509349687\u003c/li\u003e\n\u003cli\u003eNi W, Brown SD, Man R (2009) Stacked partial least squares regression analysis for spectral calibration and prediction. Journal of Chemometrics 23:505\u0026ndash;517. https://doi.org/10.1002/cem.1246\u003c/li\u003e\n\u003cli\u003eWu Y, Feng J (2018) Development and Application of Artificial Neural Network. Wireless Pers Commun 102:1645\u0026ndash;1656. https://doi.org/10.1007/s11277-017-5224-x\u003c/li\u003e\n\u003cli\u003eLamaoui A, Garc\u0026iacute;a-Guzm\u0026aacute;n JJ, Amine A, et al (2021) Synthesis techniques of molecularly imprinted polymer composites. In: Molecularly Imprinted Polymer Composites. Elsevier, pp 49\u0026ndash;91\u003c/li\u003e\n\u003cli\u003eLamuela-Ravent\u0026oacute;s RM (2017) Folin-Ciocalteu method for the measurement of total phenolic content and antioxidant capacity. In: Apak R, Capanoglu E, Shahidi F (eds) Measurement of Antioxidant Activity \u0026amp; Capacity. John Wiley \u0026amp; Sons, Ltd, Chichester, UK, pp 107\u0026ndash;115\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"microchimica-acta","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"miac","sideBox":"Learn more about [Microchimica Acta](https://link.springer.com/journal/604)","snPcode":"604","submissionUrl":"https://submission.springernature.com/new-submission/604/3","title":"Microchimica Acta","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"Molecularly imprinted membrane, biopolymer, non-specific adsorption, crosslinking, sensor array, machine learning","lastPublishedDoi":"10.21203/rs.3.rs-7609402/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7609402/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eMolecularly imprinted polymers offer exceptional sensing capabilities; however, their use is limited by the complex synthesis process, which relies on toxic and costly reagents. Additionally, issues with non-specific adsorption (NSA) further diminish their selectivity, limiting their overall effectiveness. Thus, a rapid, green, and cost-effective strategy was proposed for the preparation of bio-based molecularly imprinted membranes (MIMs) by studying diverse functional biopolymers, including chitosan, sodium alginate, carboxymethyl cellulose, sulfonated cellulose nanocrystalline, cellulose acetate, and gelatin. These MIMs were tested for tetracycline and caffeic acid (Caf) templates. The MIMs were prepared in less than 2h, without the need for complex synthesis, toxic, or expensive reagents. Furthermore, various approaches were introduced to eliminate the NSA for the first time, using covalent and non-covalent crosslinking. After appropriate selection of the biopolymer and crosslinker, the affinity of the non-imprinted membranes, free of cavities, toward the templates became negligible, thereby confirming the improvement in imprinting performance and the suppression of NSA. Coupling these NSA-free MIMs with smartphone colorimetric detection offers rapid, cost-effective, and on-site sensing. Sensor arrays were developed for the detection of Caf in pears, plums, and apples. The RGB spectral data was processed using machine learning. The artificial neural network model showed excellent regression performance, with high R\u003csup\u003e2\u003c/sup\u003e (0.989\u0026ndash;0.970) and low RMSE (0.01\u0026ndash;0.05), confirming the strategy's precision and the MIMs selectivity efficiency in complex matrices. This work provides a pathway to transition from conventional synthetic to sustainable artificial antibodies for an ultra-selective, green, cost-effective, and efficient future for imprinting technologies.\u003c/p\u003e","manuscriptTitle":"Generic strategy for high-performance bio-based imprinted membranes free of non-specific adsorption: machine learning-assisted sensing","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-10-09 12:10:07","doi":"10.21203/rs.3.rs-7609402/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2026-01-25T17:13:01+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-01-22T08:30:26+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"44836825027044777374345652603020946995","date":"2026-01-19T08:52:24+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"189555711781670950481134950549946749239","date":"2025-09-28T11:06:35+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2025-09-27T05:39:05+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2025-09-16T23:47:12+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2025-09-16T23:46:49+00:00","index":"","fulltext":""},{"type":"submitted","content":"Microchimica Acta","date":"2025-09-13T20:32:07+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"microchimica-acta","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"miac","sideBox":"Learn more about [Microchimica Acta](https://link.springer.com/journal/604)","snPcode":"604","submissionUrl":"https://submission.springernature.com/new-submission/604/3","title":"Microchimica Acta","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"9ea2d7b4-150a-4bda-979d-e332df5dc8b1","owner":[],"postedDate":"October 9th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[],"tags":[],"updatedAt":"2026-03-23T16:01:39+00:00","versionOfRecord":{"articleIdentity":"rs-7609402","link":"https://doi.org/10.1007/s00604-026-07986-9","journal":{"identity":"microchimica-acta","isVorOnly":false,"title":"Microchimica Acta"},"publishedOn":"2026-03-21 15:58:21","publishedOnDateReadable":"March 21st, 2026"},"versionCreatedAt":"2025-10-09 12:10:07","video":"","vorDoi":"10.1007/s00604-026-07986-9","vorDoiUrl":"https://doi.org/10.1007/s00604-026-07986-9","workflowStages":[]},"version":"v1","identity":"rs-7609402","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7609402","identity":"rs-7609402","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}