Photoelectrochemical sensor and biochemical polymolecule concentration detection method

Abstract

The invention provides a photoelectrochemical sensor and a biochemical polymolecule concentration detection method. The photoelectrochemical sensor includes: a working electrode; a photoelectrochemical converter; the molecularly imprinted polymer MIP membrane is prepared by mixing a first functional monomer and a second functional monomer, wherein the optical property of the first functional monomer meets the photoelectrochemical detection requirement, the electrical property of the second functional monomer meets the photoelectrochemical detection requirement, and the film-forming effect and the pore-forming effect of the first functional monomer and the second functional monomer respectively meet the preset requirements; wherein the biochemical molecules to be detected meet the requirement of being combined with the specific holes of the MIP membrane, so that the performance of the MIP membrane is changed, and the amplitude of light current generated in the photoelectrochemical detection process is changed; wherein the change degree of the light current amplitude meets the requirement for concentration detection of the biochemical molecules to be detected. The detection performance of the photoelectrochemical sensor based on the MIP film is improved.

Description

Technical Field

[0001] This application relates to the field of biochemical sensing and detection, specifically to a photoelectrochemical sensor and a method for detecting the concentration of multiple biochemical molecules. Background Technology

[0002] Molecularly imprinted polymer (MIP) membranes are functional polymer membrane materials based on a "template molecule recognition" mechanism. They achieve highly selective recognition of target molecules on the membrane surface by introducing target molecules (templates) during polymerization, forming recognition holes (imprinted sites) with "shape matching" and "functional group matching" within the polymer matrix. This is a key configuration of MIPs, and compared to bulk, particulate, and nanoparticle forms, MIP membranes exhibit greater interfacial activity and sensing compatibility.

[0003] In related technologies, MIP membranes are frequently used for biochemical molecular detection. Currently, biochemical molecular detection using MIP membranes includes three methods: optical, electrochemical, and photoelectrochemical methods, corresponding to three types of detection systems: optical (e.g., MIP-SPR), electrochemical (MIP-EC), and photoelectrochemical (MIP-PEC). Both electrochemical and optical detection systems based on MIP membranes inherit the disadvantages of their respective counterparts. Combining MIP membranes with electrical systems requires higher electrical performance, and combining them with optical systems requires higher optical performance. However, photoelectrochemical detection using MIP membranes can reduce the requirements for either electrical or optical performance. Even so, photoelectrochemical detection systems using MIP membranes often suffer from insufficient macromolecule detection performance, a limited range of macromolecules that can be detected, and cumbersome membrane preparation, failing to meet detection needs.

[0004] Therefore, this application provides a photoelectrochemical sensor and a method for detecting biochemical multi-molecule concentrations to solve one of the aforementioned technical problems. Summary of the Invention

[0005] The purpose of this application is to provide a photoelectrochemical sensor and a method for detecting the concentration of multiple biochemical molecules, which can solve at least one of the technical problems mentioned above. The specific solution is as follows: According to a specific embodiment of this application, in a first aspect, this application provides a photoelectrochemical sensor, comprising: A working electrode; a photoelectrochemical converter disposed on the surface of the working electrode; a MIP film disposed on the surface of the photoelectrochemical converter, formed by mixing a first functional monomer and a second functional monomer; wherein the optical properties of the first functional monomer meet the requirements for photoelectrochemical detection, the electrical properties of the second functional monomer meet the requirements for photoelectrochemical detection, and the film-forming and pore-forming effects of the first and second functional monomers meet preset requirements; wherein the biochemical molecule to be tested binds to the specific pores of the MIP film, thereby changing the performance of the MIP film and thus changing the photocurrent amplitude generated during the photoelectrochemical detection process; wherein the photocurrent amplitude is sufficient for the concentration detection of the biochemical molecule to be tested.

[0006] In one embodiment, the MIP membrane is configured to bind to the target biochemical molecule based on a charge adsorption sensing mechanism or a structure competition sensing mechanism.

[0007] In one embodiment, the MIP membrane is configured to bind to the target biochemical molecule based on a structure-competitive sensing mechanism; the first functional monomer is o-phenylenediamine (o-PD); and the second functional monomer is pyrrole.

[0008] In one embodiment, the MIP membrane is prepared based on a mixed solution containing template molecules, the first functional monomer, and the second functional monomer; the MIP membrane is disposed on the surface of the photoelectrochemical converter and the template molecules are eluted to obtain the specific pores; the template molecules are configured according to the biochemical molecules to be tested and are the same substance as the biochemical molecules to be tested.

[0009] According to a specific embodiment of this application, in a second aspect, this application provides a method for detecting biochemical multi-molecule concentrations, comprising: A photoelectrochemical sensor based on a MIP membrane is used to perform photoelectrochemical detection on a mixture containing at least two analyte biochemical molecules; wherein the MIP membrane is configured to specifically bind only to each of the analyte biochemical molecules; the dataset generated during the photoelectrochemical detection process is input into a pre-trained machine learning model to obtain the biochemical molecule concentration detection result output by the machine learning model.

[0010] In one embodiment, the MIP-based photoelectrochemical sensor is the photoelectrochemical sensor as described in claim 1.

[0011] In one embodiment, the dataset is a three-dimensional dataset including time data, bias voltage data, and photocurrent amplitude data, used to characterize the relationship between bias voltage data and photocurrent amplitude data over time; the machine learning model is pre-trained using the training datasets of each of the at least two biochemical molecules to be tested; wherein, the training dataset for each biochemical molecule to be tested includes: multiple photocurrent amplitude versus time curves corresponding to the biochemical molecule at different concentrations; wherein, in each of the photocurrent amplitude versus time curves, the curve data points are defined as those collected after applying one or more bias voltages in each time interval; wherein, one curve data point is collected after each bias voltage is applied.

[0012] In one implementation, the training dataset of the machine learning model is normally distributed, and the top 20% of data that deviates most from the center value are removed.

[0013] In one embodiment, the algorithm of the machine learning model is either a quadratic rational Gaussian process regression algorithm or a three-layer wide neural network algorithm, with the quadratic rational Gaussian process regression algorithm being preferred.

[0014] Compared with the prior art, the above-described solution of this application has at least the following beneficial effects: This application provides a photoelectrochemical sensor, including a working electrode, a photoelectrochemical converter disposed on the surface of the working electrode, and a MIP film disposed on the surface of the photoelectrochemical converter. The MIP film is made of a mixture of a first functional monomer and a second functional monomer. The first functional monomer has excellent optical properties, while the second functional monomer has good electrical properties. The combination of the two not only ensures the optical and electrical properties required for photoelectrochemical detection, but also provides excellent film formation and pore formation effects. This design allows the biochemical molecules to be tested to specifically bind to the pores on the MIP film, thereby changing the performance of the MIP film and causing a change in the photocurrent amplitude. This change can accurately reflect the concentration of biochemical molecules, greatly improving the accuracy and sensitivity of detection and the range of detection targets. Attached Figure Description

[0015] Figure 1 A block diagram of a photoelectrochemical sensor is shown. Figure 2 A schematic diagram of a process for preparing MIP films using a single template molecule is shown. Figure 3 A schematic diagram of a process for preparing MIP films using three template molecules is shown. Figure 4 A flowchart of the biochemical multi-molecule concentration detection method of this application is shown; Figure 5 A schematic diagram of the training data constructed for PD-L1 and CYFRA21-1 is shown; Figure 6 A schematic diagram of the process for training machine learning models based on the PD-L1 and CYFRA21-1 datasets is shown. Detailed Implementation

[0016] To make the objectives, technical solutions, and advantages of this application clearer, the application will be further described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0017] The terminology used in the embodiments of this application is for the purpose of describing particular embodiments only and is not intended to limit the application. The singular forms “a,” “said,” and “the” used in the embodiments of this application and the appended claims are also intended to include the plural forms, and “multiple” generally includes at least two unless the context clearly indicates otherwise.

[0018] It should be understood that the term "and / or" used in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, and B existing alone. Additionally, the character " / " in this article generally indicates that the preceding and following related objects have an "or" relationship.

[0019] It should be understood that although the terms first, second, third, etc., may be used in the embodiments of this application, these descriptions should not be limited to these terms. These terms are only used to distinguish the descriptions. For example, first may also be referred to as second without departing from the scope of the embodiments of this application, and similarly, second may also be referred to as first.

[0020] Depending on the context, the words “if” or “suppose” as used here can be interpreted as “when” or “in response to determination” or “in response to detection.” Similarly, depending on the context, the phrases “if determination” or “if detection (of the stated condition or event)” can be interpreted as “when determination” or “in response to determination” or “when detection (of the stated condition or event)” or “in response to detection (of the stated condition or event).”

[0021] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that an article or device that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such an article or device. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the article or device that includes said element.

[0022] It should be noted that any symbols and / or numbers present in the specification that are not marked in the accompanying drawings are not reference numerals.

[0023] Currently, photoelectrochemical sensors using MIP membranes in related technologies still face challenges such as insufficient detection performance for macromolecules, a limited range of macromolecules that can be detected, and cumbersome membrane preparation. The main reasons for this are as follows: 1. Poor porosity: In traditional monomeric material MIP membranes, biochemical macromolecules are not easy to form pores, and the matching between pores and molecular binding sites is low.

[0024] 2. Poor film-forming properties: The polymerization of monomer materials cannot simultaneously take into account their mechanical properties and biocompatibility, resulting in poor film-forming properties.

[0025] 3. Poor photoelectric properties: Traditional monomeric material MIP film is difficult to combine with PEC while simultaneously achieving both optical and electrical properties.

[0026] 4. Number of active functional groups: If the traditional monomer material MIP membrane has many active functional groups, its detection characteristics are poor; if it has few active functional groups, it is not easy to capture the detection target, resulting in low detection sensitivity.

[0027] 5. The preparation process of sensitive membranes and converters is complex: the formation of single-functional monomer films requires additional pore modification to improve porosity and selectivity.

[0028] In view of this, this application considers, on the one hand, improving the sensing performance and biochemical molecular range of macromolecule detection by changing the functional monomers of MIP membrane 103, and on the other hand, replacing the monofunctional monomers used in related technologies with bifunctional monomers to prepare MIP membrane 103, so as to reduce the difficulty of selecting functional monomer materials and provide MIP membrane 103 with good photoelectric and electrical properties at the same time, so that MIP membrane 103 can overcome the above-mentioned problems in related technologies.

[0029] The optional embodiments of this application are described in detail below with reference to the accompanying drawings.

[0030] Figure 1 A block diagram of a photoelectrochemical sensor 100 is shown.

[0031] For example, such as Figure 1 As shown, the photoelectrochemical sensor 100 includes a working electrode 101, a photoelectrochemical converter 102, and a MIP membrane 103.

[0032] The photoelectrochemical converter 102 is disposed on the surface of the working electrode 101, and the MIP film 103 is disposed on the surface of the photoelectrochemical converter 102.

[0033] For example, MIP membrane 103 is prepared by mixing a first functional monomer with a second functional monomer.

[0034] The optical properties of the first functional monomer meet the requirements for photoelectrochemical detection, the electrical properties of the second functional monomer meet the requirements for photoelectrochemical detection, and the film-forming and pore-forming effects of the first and second functional monomers meet the preset requirements.

[0035] In this application, a functional monomer whose optical properties meet the requirements of photoelectrochemical detection is selected as the first functional monomer. This means that the MIP film 103 composed of a single first functional monomer can enable the photoelectrochemical converter 102 to have excellent light absorption. Furthermore, when the MIP film 103 is composed of a bifunctional monomer including the first functional monomer, the MIP film 103 can still inherit this excellent optical property, so that the photoelectrochemical converter 102 can also have excellent light absorption when using the MIP film 103 based on the bifunctional monomer.

[0036] In this application, a functional monomer whose electrical properties meet the requirements of photoelectrochemical detection is selected as the second functional monomer. This means that the MIP membrane 103 composed of a single second functional monomer is itself easy to transport separated electrons and holes. Furthermore, when the MIP membrane 103 is composed of a bifunctional monomer including the second functional monomer, the MIP membrane 103 can still inherit this excellent electrical property, so that the photoelectrochemical converter 102 can also easily transport separated electrons and holes when using the MIP membrane 103 based on the bifunctional monomer.

[0037] In this application, a MIP sensitive membrane is fabricated by mixing a first functional monomer and a second functional monomer. Since both the first and second functional monomers possess good mechanical and biocompatibility, their mixing improves the film-forming and porosity of the MIP membrane 103. Furthermore, the first functional monomer's superior optical properties enhance the light absorption of the photoelectrochemical converter 102, while the second functional monomer's superior electrical properties facilitate the transport of separated electrons and holes in the MIP membrane 103. Therefore, mixing the two functional monomers improves both the optical and electrical properties of the MIP membrane 103.

[0038] For example, a functional monomer with a moderate amount of active functional groups can be selected as the first and second functional monomers. A moderate amount of active functional groups can ensure that there are fewer non-specific sites after polymerization to form the MIP membrane 103. At the same time, the pore-forming property of biochemical macromolecules is relatively good, which increases the recognition ability of macromolecule detection and enables the specific detection of biochemical macromolecules.

[0039] In this application, the target biochemical molecule binds to specific pores in the MIP membrane 103, causing a change in the tightness of the MIP membrane 103, which in turn changes the amplitude of the photocurrent generated during photoelectrochemical detection. This photocurrent amplitude is then used to detect the concentration of the target biochemical molecule. For example, after the target biochemical molecule binds to the MIP membrane 103 and the photocurrent amplitude changes, the concentration of the target biochemical molecule is determined by mapping the change in photocurrent amplitude to the molecular concentration.

[0040] For example, the biochemical molecule to be tested is captured by the pores on the MIP membrane 103, which increases the impedance and causes the photocurrent to decrease. The concentration of the substance to be tested is determined by detecting the magnitude of the decrease in photocurrent.

[0041] In this application, for the fabrication of the photoelectrochemical converter 102, an example is to use a gold electrode (AuE) as a substrate, and modify the surface of the AuE with MUA-AuNCs having photoelectric conversion properties through a covalent bond self-assembly method to form the photoelectrochemical converter 102 (AuE / AuNCs PEC electrode). Based on this, a MIP film 103 can be further polymerized on the surface of the photoelectrochemical converter 102.

[0042] In this application, considering the formation and detection process of the MIP film 103, from a macroscopic perspective, the sensing principle of the photoelectrochemical sensor 100 is as follows: the specific adsorption of target biochemical molecules on the MIP pores causes a change in the impedance of the photoelectrochemical sensor 100, which in turn causes a change in the output photocurrent, ultimately converting the biochemical signal into an electrical signal. Therefore, the specific adsorption of biochemical molecules plays a crucial role in this conversion process. From a microscopic perspective, the adsorption of biochemical molecules on the MIP can be broadly categorized into two types: sensing mechanisms based on charge attraction and sensing mechanisms based on structural competition. Therefore, the photoelectrochemical sensor 100 provided in this application can be configured based on either a charge attraction sensing mechanism or a structural competition sensing mechanism.

[0043] As one feasible implementation, the MIP membrane 103 is configured to bind to the target biochemical molecule based on a charge adsorption sensing mechanism.

[0044] As another feasible implementation, the MIP membrane 103 is configured to bind to the target biochemical molecule based on a structural competition sensing mechanism.

[0045] Figure 2 A schematic diagram of a process for preparing MIP films using a single template molecule is shown.

[0046] like Figure 2 As shown, in preparing the photoelectrochemical sensor AuE / AuNCs / MIP (PD-L1), firstly, a MIP film 103 solution is coated onto the surface of the photoelectrochemical converter 102 (AuE / AuNCs PEC electrode) to allow the MIP film 103 to adhere to the surface of the photoelectrochemical converter 102 (AuE / AuNCs PEC electrode), thus obtaining the photoelectrochemical sensor AuE / AuNCs / MIP (PD-L1). Then, the PD-L1 molecules used as template molecules in the MIP film 103 are eluted, leaving corresponding pores at the original positions of the PD-L1 molecules in the MIP film 103, resulting in a photoelectrochemical sensor AuE / AuNCs / MIP with specific pores. These pores have structural specificity consistent with PD-L1 molecules. During subsequent photoelectrochemical detection, other biochemical molecules besides PD-L1 molecules cannot specifically bind to the pores. This photoelectrochemical sensor is suitable for the concentration detection of PD-L1 molecules.

[0047] Furthermore, based on the same concept, MIP membranes 103 capable of specifically binding to various biochemical molecules can also be prepared using a variety of template molecules.

[0048] Figure 3 A schematic diagram of a process for preparing MIP membranes using three template molecules is shown.

[0049] like Figure 3As shown, taking the photoelectrochemical sensor AuE / AuNCs / rMIP as an example, three biochemical molecules—PD-L1, CYFRA21-1, and Cortisol—can be used as template molecules, and these three biochemical molecules are incorporated into the preparation of the MIP membrane 103 solution. Then, the MIP membrane 103 solution is coated onto the surface of the photoelectrochemical converter 102 (AuE / AuNCs PEC electrode) to obtain the photoelectrochemical sensor AuE / AuNCs / rMIP (PD-L1, CYFRA21-1, Cortisol). Further, the three template molecules are eluted, giving the MIP membrane 103 three corresponding specific pores. Subsequently, the photoelectrochemical sensor AuE / AuNCs / rMIP can be used to detect the concentration of one or more of the biochemical molecules PD-L1, CYFRA21-1, and Cortisol. Specifically, taking the detection of a mixed solution containing three biochemical molecules—PD-L1, CYFRA21-1, and Cortisol—as an example, during the photoelectrochemical detection process, the three pores specifically bind to their respective corresponding biochemical molecules, thereby causing changes in photocurrent. Based on this, in some further embodiments, the concentrations of the three biochemical molecules can be identified by analyzing the degree of influence of the three biochemical molecules on the photocurrent changes.

[0050] In this embodiment, the structural competition-based sensing mechanism utilizes the pores in the MIP membrane 103 of the template molecule to achieve specific binding with biochemical molecules. After the template molecule is eluted, the pores in the MIP membrane 103 completely retain the structural information of the template molecule, including its size, shape, and binding sites. Since different biochemical molecules have different structures, only biochemical molecules with the same size, shape, and surface group sites can specifically bind to the pores of the MIP membrane 103, thereby obtaining a specific biochemical signal. Size, shape, and surface group sites are all three factors that must be present simultaneously; none can be missing. Other molecules either have unsuitable sizes, cannot overcome shape limitations, or cannot induce recombination of binding sites, and therefore cannot be detected. These substances that fail to bind are easily detached under external stimuli, resulting in little or no change in impedance, or irregular changes. By using the target biochemical molecule as a template molecule to leave pores in the MIP membrane 103, ensuring that only the target biochemical molecule can match the corresponding pore in the MIP membrane 103, this sensing mechanism guarantees the specificity of the sensor.

[0051] In this application, to illustrate the above points, three photoelectrochemical sensors 100 with different pore sizes—AuE / AuNCs / rMIP (PD-L1), AuE / AuNCs / rMIP (CYFRA21-1), and AuE / AuNCs / rMIP (Cortisol)—were prepared based on biochemical molecules of different sizes. These sensors were used to test the ability of each of the three different pore sizes to capture PD-L1, CYFRA21-1, and Cortisol molecules, thereby determining the specific binding effect based on a structural competition mechanism. Here, AuE indicates that the substrate selected for the photoelectrochemical sensor 100 is a gold electrode; AuE / AuNCs / rMIP indicates the sensor structure obtained by modifying MUA-AuNCs onto the surface of AuE and then further polymerizing a MIP film 103; (X) indicates that the template molecule used by the photoelectrochemical sensor 100 is X, such as (PD-L1) indicating that the template molecule used by the photoelectrochemical sensor 100 is PD-L1.

[0052] For example, the template molecules used in the demonstration included the biomacromolecule CYFRA21-1 (~40 kDa), the biomacromolecule Cortisol (362.46 kDa), and PD-L1 (~26 kDa), which has a size between the two. Under the same conditions, cross-measurements were performed on the above three biomolecules using photoelectrochemical sensors 100 with different pore sizes to determine the binding capacity of each of the three different pore sizes for the three biomolecules. The following are the test results of the three different pore sizes of the photoelectrochemical sensors 100 for the three biomolecules.

[0053] For the measurement of CYFRA21-1, the AuE / AuNCs / rMIP (CYFRA21-1) electrode showed satisfactory linearity. The photocurrent decreased linearly with increasing CYFRA21-1 concentration, as expected. Due to the specific pores of CYFRA21-1 in the sensor, CYFRA21-1 in solution can completely bind to the pores on the MIP film 103, resulting in increased sensor impedance and reduced light absorption, thus lowering the sensor's photocurrent. When using AuE / AuNCs / rMIP (PD-L1) and AuE / AuNCs / rMIP (Cortisol) sensing electrodes to detect CYFRA21-1 with the same concentration gradient, the sensor's photocurrent remained essentially unchanged. This is primarily because the CYFRA21-1 is too large compared to PD-L1 and Cortisol. The CYFRA21-1 cannot fit into the holes of the AuE / AuNCs / rMIP (PD-L1) and AuE / AuNCs / rMIP (Cortisol) sensing electrodes, thus the sensor impedance is almost unaffected. Under these circumstances, it is logical that the CYFRA21-1 will not cause any response in the photocurrent of either sensor.

[0054] For PD-L1 measurement, the AuE / AuNCs / rMIP (PD-L1) electrode achieved satisfactory linearity in detecting PD-L1 solutions, with the photocurrent decreasing linearly with increasing PD-L1 concentration. This is because the MIP film 103 has specific pores for PD-L1, which bind to the measured PD-L1, leading to a decrease in photocurrent. When the AuE / AuNCs / rMIP (CYFRA21-1 pore) and AuE / AuNCs / rMIP (Cortisol pore) electrodes measured the same concentration range of PD-L1, the photocurrent did not show a linear relationship with the PD-L1 concentration. This indicates that when a sensor with larger pores (AuE / AuNCs / rMIP (CYFRA21-1)) measures smaller biomolecules (PD-L1), the sensor's photocurrent remains unaffected even if smaller PD-L1 molecules fall into the larger pores. This is because pore trapping biomolecules requires not only appropriate size but also suitable shape and binding sites. While PD-L1 meets the size requirement, other conditions remain mismatched, preventing the AuE / AuNCs / rMIP (CYFRA21-1) sensor from producing a linear photocurrent response to PD-L1. Larger biomolecules (PD-L1) cannot fall into smaller cavities, thus the impedance of the AuE / AuNCs / rMIP (Cortisol) sensor remains unchanged, and the photocurrent is unaffected. Therefore, PD-L1 does not produce a significant photocurrent response for either large or small cavities in the photoelectrochemical sensor 100. This demonstrates the satisfactory specificity and selectivity of the photoelectrochemical sensor 100.

[0055] For Cortisol measurements, when analyzing Cortisol solutions using AuE / AuNCs / rMIP (Cortisol) electrodes with Cortisol cavities, a significant linear relationship was observed between photocurrent changes and Cortisol concentration. When measuring Cortisol at the same concentration gradient using AuE / AuNCs / rMIP (CYFRA21-1) and AuE / AuNCs / rMIP (PD-L1) electrodes, the sensor photocurrent remained essentially unchanged. Although Cortisol is small enough to fill cavities of various sizes, it does not bind to the sites and therefore cannot cause changes in photocurrent.

[0056] In the above embodiments, the MIP membrane 103 is prepared based on a mixed solution containing a template molecule, a first functional monomer, and a second functional monomer. Specifically, approximately 10-30 mg of the first functional monomer and approximately 3-5 mg of the second functional monomer are mixed, and 10-20 mL of deionized water is added. The mixture is stirred at room temperature for 20-40 minutes to obtain a mixed solution of bifunctional monomers. Then, the template molecule is added to the mixed solution of bifunctional monomers, and the mixture is stirred at room temperature for 20-30 minutes to obtain a mixed MIP solution of the template molecule.

[0057] For the above embodiments, when the template molecule is selected as PD-L1, CYFRA21-1 or Cortisol, if the mass of PD-L1 or CYFRA21-1 in the mixed solution is 1-2 μg and the mass of Cortisol is 16-32 mg.

[0058] As some other examples, when the template molecule is selected as 4MPLA, the mass in the mixed solution is 16mg.

[0059] In some embodiments, the template molecules are configured according to the biochemical molecules to be tested and are the same substance as the biochemical molecules to be tested, so that the biochemical molecules to be tested can combine with the pores formed after the template molecules are eluted, thereby causing a change in photocurrent.

[0060] In this application, to bond the photoelectrochemical converter 102 with the MIP film 103, a 3 mL volume of mixed MIP solution can be dropped onto the surface of the photoelectrochemical converter 102, and a MIP film 103 containing template molecules can be formed on the surface of the photoelectrochemical converter 102 by electropolymerization. Taking the photoelectrochemical converter 102 as an AuE / AuNCsPEC electrode as an example, the obtained electrode is called AuE / AuNCs / MIP.

[0061] In the above embodiments, electropolymerization was performed using cyclic voltammetry at a scan rate of 40-50 mV / s, 12-14 scan cycles, and a scan potential of -0.2-1.2 V. Before use, template molecules were eluted from the AuE / AuNCs / MIP electrode using a mixed solution of ethanol and deionized water (volume ratio 1:1) for 40-60 minutes. After the MIP membrane 103 was deposited on the surface of the photoelectrochemical converter 102 and the template molecules were eluted, specific pores were obtained. The electrode was then cleaned with phosphate-buffered saline (PBS). The resulting electrode, referred to as AuE / AuNCs / rMIP, was used for subsequent actual detection. Therefore, photoelectrochemical sensors 100 with different pore sizes were prepared when different template molecules were applied.

[0062] In the above embodiments, the electropolymerization method for treating the MIP film is merely exemplary. In some other embodiments, methods such as spin coating, chemical oxidative polymerization, and drop coating can be used instead of electropolymerization to treat the MIP film.

[0063] In this application, theoretically, any functional monomer that meets the common requirements mentioned above, such as porosity, film-forming properties, and active functional groups, can be used as a candidate material for the first or second functional monomer. For example, among the candidate functional monomers with excellent porosity, excellent film-forming properties, and moderate active functional groups, a functional monomer with excellent optical properties can be selected as the first functional monomer, and a functional monomer with excellent electrical properties can be selected as the second functional monomer.

[0064] In some embodiments, functional monomers possessing excellent electrical properties include aniline, 3-thiopheneacetic acid, acrylamide, methacrylic acid, and dopamine; functional monomers possessing excellent optical properties include 4-vinylpyridine, acryloyloxyphenylboronic acid, styrene monomers, benzothiazole monomers, and azobenzene monomers. For example, one functional monomer can be selected from those possessing excellent electrical properties and another from those possessing excellent optical properties, thereby preparing a MIP membrane using the selected bifunctional monomer.

[0065] Furthermore, when selecting a bifunctional monomer, it is necessary to further consider the suitability of the first and second functional monomers based on the chosen sensing mechanism and the practical effect of the MIP membrane 103 after its formation. Ultimately, the bifunctional monomer that best meets the detection requirements should be used to prepare the MIP membrane 103. For example, when considering a structure-competitive sensing mechanism, the bifunctional monomer composed of o-PD and pyrrole in the above embodiment can meet the requirements.

[0066] As a specific embodiment, when the MIP membrane 103 is configured to bind to the target biochemical molecule based on a structure-competitive sensing mechanism, the first functional monomer can be o-phenylenediamine (o-PD), and the second functional monomer can be pyrrole. The MIP membrane 103 prepared using o-PD and pyrrole exhibits superior porosity, film-forming properties, electrical properties, and optical properties, while ensuring a suitable amount of active functional groups. Therefore, the MIP membrane 103 meets the detection requirements of photoelectrochemical methods.

[0067] In related technologies, the performance of the MIP membrane 103 cannot meet the requirements of PEC detection. This means that for the photoelectrochemical sensor 100 to measure macromolecules, either the sensitivity is poor or it cannot detect them at all, or complex modifications and alterations are needed to improve the pore quality and thus enhance sensing performance. The photoelectrochemical sensor 100 cannot measure the aforementioned molecules, especially the aforementioned biochemical macromolecules (such as CYFRA21-1 and PD-L1). In this application, a bifunctional monomer is used to configure the MIP membrane 103, thereby improving its performance. With appropriate pore design, the photoelectrochemical sensor 100 exhibits excellent sensing performance for biochemical molecules of different sizes.

[0068] The photoelectrochemical sensor 100 provided in this application has excellent recyclability, can be reused, and can be used in unattended field scenarios, avoiding the need for frequent sensor replacements. Furthermore, in addition to excellent sensitivity and detection limit, the photoelectrochemical sensor 100 provided in this application also exhibits strong stability and high selectivity.

[0069] This application also provides method embodiments that follow the above embodiments, for molecular detection using a variety of photoelectrochemical sensors, including the photoelectrochemical sensors of the above embodiments. The interpretation of the same names is the same as that of the above embodiments, and the same technical effects are achieved as those of the above embodiments, so they will not be repeated here.

[0070] Figure 4 A flowchart of the biochemical multi-molecule concentration detection method of this application is shown, such as... Figure 4 As shown, the procedure includes steps S301 to S302.

[0071] Step S301: The mixture containing at least two biochemical molecules to be tested is detected by photoelectrochemical sensor.

[0072] The MIP membrane is configured to specifically bind to each of the target biochemical molecules. For example, if a mixture contains two target biochemical molecules and one other molecule, the specific pores of the MIP membrane can bind to either of the two target biochemical molecules, while the other molecule cannot bind to the specific pores.

[0073] Step S302: Input the dataset generated during the photoelectrochemical detection process into the pre-trained machine learning model to obtain the biochemical molecule concentration detection results output by the machine learning model.

[0074] The biochemical molecule concentration detection results include the molecular concentration of each of the at least two types of biochemical molecules to be tested.

[0075] In some embodiments, the photoelectrochemical sensor may be selected as the photoelectrochemical sensor based on bifunctional monomers provided in the above embodiments of this application. In some other embodiments, the photoelectrochemical sensor may also be selected as the photoelectrochemical sensor based on monofunctional monomers in related technologies. In these embodiments, the MIP film of the photoelectrochemical sensor is configured to specifically bind only to the target biochemical molecule, thereby causing a change in the photocurrent amplitude.

[0076] In some embodiments, a mixed solution containing each type of biochemical molecule to be tested can be prepared according to the type of biochemical molecule to be tested, and then this solution can be used as a template molecule solution to prepare a MIP membrane. After the template molecules are eluted from the MIP membrane, the MIP membrane has the ability to specifically bind to each type of biochemical molecule to be tested.

[0077] In the method provided in this application, machine learning algorithms are used to model and analyze photoelectrochemical response signals. The multi-channel signals collected by the sensor are input as a three-dimensional dataset into a trained machine learning model, which outputs the corresponding concentrations of multiple analyte biochemical molecules. This method combines the data-driven capabilities of artificial intelligence with the physical sensing capabilities of a photoelectrochemical platform, significantly improving the efficiency and accuracy of multi-component identification and effectively addressing the limitations of photoelectrochemical sensors in terms of poor selectivity, signal overlap, and quantification difficulties.

[0078] The biochemical multi-molecule concentration detection method provided in this application not only meets the requirements for detecting the concentration of multiple biochemical molecules, but also takes into account the requirements for detecting the concentration of a single biochemical molecule. For example, when a photoelectrochemical sensor carries a single biochemical molecule, and this biochemical molecule is one of the molecules that specifically binds to the MIP membrane in the photoelectrochemical sensor, the change in the enhancement amplitude of the photocurrent generated by this single biochemical molecule can be input into a specific machine learning model, so that the machine learning model outputs the concentration detection result of this biochemical molecule.

[0079] Furthermore, the concentration detection results of a single biochemical molecule can be obtained through direct mapping; the methods for detecting the concentration of a single biochemical molecule are not detailed here. Based on this, the method for detecting the concentration of a single biochemical molecule can be combined with the multi-molecule biochemical concentration detection described in this application to achieve molecular concentration detection of one or more biochemical molecules in practical applications.

[0080] In some embodiments, the machine learning model is pre-trained using training datasets for at least two different biochemical molecules to be tested.

[0081] While it's feasible to determine the concentration of a single molecule by fitting a relationship, its anti-interference ability, stability, and accuracy are not excellent, and it cannot predict the concentration of multiple molecules using a single channel. To address this issue, this application proposes using a supervised machine learning algorithm to separate the current curves of mixed sensor data and identify the concentration of each molecule in the mixed data one by one.

[0082] In this application, feature values ​​are first selected based on the characteristics of the sensor output signal. Since the photocurrent amplitude changes differently over time when the photoelectrochemical sensor detects different concentrations of different biochemical molecules, i.e., the photocurrent amplitude change curve exhibits different shapes and amplitudes over time, this application uses time and different bias voltages as feature values ​​when constructing the training set. The dataset can be a three-dimensional dataset containing time data, bias voltage data, and photocurrent amplitude data, used to characterize the relationship between bias voltage data and photocurrent amplitude data over time. Secondly, the detection target, detection accuracy, and detection range are determined based on the actual application scenario. Then, a concentration gradient for the dataset is designed based on this. Finally, the data is cleaned to eliminate data redundancy, clutter, and incompleteness, thereby improving data quality and reliability. Specifically, data cleaning addresses outliers and missing values ​​in the original data, including deleting duplicate data and filling in missing data and erroneous data that does not conform to common understanding using the median or mean.

[0083] As some feasible implementations, the algorithm can be built on a three-dimensional dataset consisting of time, bias voltage, and photocurrent amplitude. The following uses PD-L1 and CYFRA21-1 as examples to illustrate the process of constructing training data and training machine learning models.

[0084] Figure 5 A schematic diagram of the training data constructed for PD-L1 and CYFRA21-1 is shown.

[0085] Figure 6 A schematic diagram of the process for training machine learning models based on the PD-L1 and CYFRA21-1 datasets is shown.

[0086] Exemplarily, when constructing the dataset, biochemical molecules at different concentrations can be collected separately. In other words, during a single data collection process, experiments are conducted with biochemical molecules at a fixed concentration, and the relationship between the photocurrent amplitude and time is observed during this period. At the same time, during the entire collection cycle, data can be collected at specific time intervals. On the one hand, within a time point after meeting the time interval, the influence of different bias voltages on the photocurrent amplitude can be explored by applying different bias voltages separately. On the other hand, using different data with the same bias voltage applied between consecutive time intervals, the relationship between the photocurrent amplitude and time when the bias voltage is fixed can be explored.

[0087] For example, as Figure 5 shown in the upper part, in this embodiment, for a single experiment, the interval testing can be carried out at a test interval of two minutes, and the time points for data collection at least include "0min", "2min", "4min", and "6min". Among them, for each test interval, the photocurrent is measured by means of photoelectrochemical detection respectively, and the light irradiation is stopped after the photocurrent amplitude reaches the peak value under this test, and the tables corresponding to "0min", "2min", "4min", and "6min" as shown in the upper part are obtained. Since the photocurrent amplitude is zero when there is no light irradiation, the photocurrent amplitude increases during light irradiation and slowly decays after reaching the peak value, and quickly returns to zero after the light irradiation is stopped, so the photocurrent amplitude data under a single test shows a "V" shape. For a single test, the peak value of the photocurrent amplitude during light irradiation can be used as the photocurrent amplitude data obtained from this test. Figure 5 Based on this, to explore the influence of different bias voltages on the photocurrent amplitude, it can be achieved by dividing each time interval into multiple time sub-intervals and conducting a single test within different time sub-intervals respectively. For example, when the time reaches "2min", taking ten seconds as the time sub-interval, three bias voltages are applied and three tests are conducted within two minutes to two minutes and ten seconds, two minutes and ten seconds to twenty seconds, and two minutes and twenty seconds to thirty seconds. Finally, the peak value of the photocurrent amplitude obtained from each test is used as the photocurrent amplitude value of a data point shown in the lower table. It should be noted that the application does not specifically limit the duration of the sub-interval setting and the number of sub-intervals.

[0088] Exemplarily, after completing the above data collection, the training dataset of each biochemical molecule to be measured includes multiple curves of the photocurrent amplitude varying with time corresponding to the biochemical molecule at different concentrations and / or different applied bias voltages. For example, as Figure 5 shown in the lower part of the table.

[0089] Exemplarily, after completing the above data collection, the training dataset of each biochemical molecule to be measured includes multiple curves of the photocurrent amplitude varying with time corresponding to the biochemical molecule at different concentrations and / or different applied bias voltages. For example, as Figure 5As shown, the lower left table represents the photocurrent amplitude versus time curves for PD-L1 concentrations of 1 pg / ml, 10 pg / ml, 20 pg / ml, 40 pg / ml, and 60 pg / ml under a single applied bias voltage. The lower right table represents the photocurrent amplitude versus time curves for CYFRA21-1 concentrations of 1 pg / ml, 10 pg / ml, 20 pg / ml, 40 pg / ml, and 60 pg / ml under a single applied bias voltage. For any given table, each data point in each curve was collected when the same bias voltage was applied. In other words, a single table contains multiple curves collected under a single bias voltage, with different curves corresponding to different molecular concentrations. In this application, a table containing photocurrent amplitude versus time curves at multiple concentrations can be generated for different bias voltages. This is equivalent to further constructing a three-dimensional data relationship on top of the two-dimensional linear relationship between photocurrent amplitude and time using different bias voltage values, and using this as the three-dimensional dataset used in this application. Based on this, the dataset covers photocurrent amplitudes corresponding to different molecules, different bias voltages, and different concentrations. Since the tables indicate the mapping relationship between the molecular concentration and the molecular concentration, the three-dimensional dataset can effectively train machine learning algorithms to recognize molecular concentrations.

[0090] Furthermore, such as Figure 6 As shown, multiple concentration combinations can be simulated based on the constructed dataset to generate thousands of data sets, which are then used to train the model until the algorithm converges. Based on this, the model can achieve single-molecule concentration detection by separating the molecules when faced with mixed sensing data containing PD-L1 and CYFRA21-1, according to the training set and pre-training results. For example, for a model trained using a dataset constructed with PD-L1 and CYFRA21-1, it can extract the concentration curve of a single molecule from the mixed sensing data provided by the photoelectrochemical sensor when faced with a mixed solution containing PD-L1 and CYFRA21-1, outputting the photocurrent amplitude versus time curves for PD-L1 and CYFRA21-1 respectively, and then combining the curve features with the corresponding molecular concentration.

[0091] As some other feasible embodiments, in addition to time data, bias voltage data, and photocurrent amplitude data, data such as temperature, light intensity, and humidity can also be used to construct the dataset. These data can replace time data, bias voltage data, and photocurrent amplitude data to construct different types of three-dimensional datasets, or serve as new data dimensions to construct more dimensional datasets based on the three-dimensional dataset. This application does not specifically limit the data dimensions of the dataset or the types of data used in the dataset.

[0092] In some embodiments, the training dataset for the machine learning model is normally distributed, and the top 20% of data that deviates most from the center value are removed. This method can improve the quality and reliability of the training data.

[0093] As some feasible embodiments, the algorithm of the machine learning model is selected as either the quadratic rational Gaussian process regression algorithm or the three-layer wide neural network algorithm, with the quadratic rational Gaussian process regression algorithm being preferred.

[0094] This application constructs a specific dataset and then uses machine learning algorithms to predict multi-molecule concentrations, a typical example of supervised regression learning. In preliminary experiments, the most suitable algorithms for predicting molecular concentrations using photoelectrochemical sensors were quadratic rational Gaussian process regression (QR-GPR) and three-layer wide neural networks. These two types of algorithms not only have high accuracy but are also suitable for handling nonlinear relationships. The appropriate algorithm is selected based on the dataset size, feature dimensionality, data type, whether it is labeled, the type of task, and the data distribution. This application ultimately chose QR-GPR. This is also related to the characteristics of QR-GPR. QR-GPR is an extension of Gaussian process regression, which uses a quadratic rational kernel to characterize the latent structure of the data. QR-GPR can capture more complex nonlinear relationships in the data, especially when there are proportional relationships or large relative changes between data points, exhibiting good fitting ability, flexibility, and scalability. Therefore, QR-GPR is suitable for datasets with complex nonlinear relationships, proportional changes in data, complex data distribution, irregular fluctuations, and small sample sizes. After training the model on the computer, we used MATLAB's Embedded Coder to convert the model and its trained built-in parameters into C for deployment on the detection instrument. Ultimately, the model achieved an error of ±5 pg / mL for single-molecule concentration detection.

[0095] Although the operations are described in a specific order in the accompanying drawings, this should not be construed as requiring these operations to be performed in the specific order or serial order shown, or requiring all of the operations shown to obtain the desired result. In certain environments, multitasking and parallel processing may be advantageous.

[0096] The methods, apparatus, devices, and storage media of this application can be implemented using standard programming techniques, and various method steps can be implemented using rule-based logic or other logic. It should also be noted that the terms "apparatus" and "module" as used herein and in the claims are intended to include implementations using one or more lines of software code and / or hardware implementations and / or devices for receiving input.

[0097] Any step, operation, or procedure described herein may be performed or implemented using one or more hardware or software modules, either alone or in combination with other devices. In one embodiment, the software module is implemented using a computer program product comprising a computer-readable medium containing computer program code, which is executable by a computer processor to perform any or all of the described steps, operations, or procedures.

[0098] The foregoing description of implementations of this application has been provided for illustrative and descriptive purposes. The foregoing description is not exhaustive and is not intended to limit this application to the exact forms disclosed. Various modifications and variations may exist in accordance with the foregoing teachings, or may arise from practice of this application. These embodiments were chosen and described to illustrate the principles of this application and its practical application, enabling those skilled in the art to utilize this application in various implementations and modifications to suit the specific purpose of the concept.

[0099] Regarding the apparatus in the above embodiments, the specific manner in which each module performs its operations has been described in detail in the embodiments related to the method, and will not be elaborated upon here.

[0100] It can be further understood that, unless otherwise specified, "connection" includes both direct connections where no other components exist between the two parties and indirect connections where other components exist between them.

[0101] It is further understood that although the operations are described in a specific order in the accompanying drawings in the embodiments of this application, this should not be construed as requiring these operations to be performed in the specific order or serial order shown, or requiring all the operations shown to obtain the desired result. In certain environments, multitasking and parallel processing may be advantageous.

[0102] Other embodiments of this application will readily occur to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. This application is intended to cover any variations, uses, or adaptations of this application that follow the general principles of this application and include common knowledge or customary techniques in the field of this application that are not disclosed herein. The specification and examples are to be considered exemplary only, and the true scope and spirit of this application are indicated by the following claims.

[0103] It should be understood that this application is not limited to the precise structure described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of this application is limited only by the appended claims.

[0104] The above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A photoelectrochemical sensor, characterized in that, include: Working electrode; A photoelectrochemical converter is disposed on the surface of the working electrode; A molecularly imprinted polymer (MIP) film, disposed on the surface of the photoelectrochemical converter, is prepared by mixing a first functional monomer and a second functional monomer. The optical properties of the first functional monomer meet the requirements for photoelectrochemical detection, the electrical properties of the second functional monomer meet the requirements for photoelectrochemical detection, and the film-forming effect and pore-forming effect of the first functional monomer and the second functional monomer respectively meet the preset requirements. The biochemical molecules to be tested bind to specific pores in the MIP membrane, thereby changing the properties of the MIP membrane and consequently altering the amplitude of the photocurrent generated during the photoelectrochemical detection process. The photocurrent amplitude is used to detect the concentration of the biochemical molecule to be tested.

2. The photoelectrochemical sensor according to claim 1, characterized in that, The MIP membrane is configured to bind to the target biochemical molecule based on a charge adsorption sensing mechanism or a structure competition sensing mechanism.

3. The photoelectrochemical sensor according to claim 2, characterized in that, The MIP membrane is configured to bind to the target biochemical molecule based on a structure competition sensing mechanism; The first functional monomer is o-phenylenediamine (o-PD); The second functional monomer is pyrrole.

4. The photoelectrochemical sensor according to claim 1 or 3, characterized in that, The MIP membrane is prepared based on a mixed solution containing template molecules, the first functional monomer, and the second functional monomer. The MIP film is disposed on the surface of the photoelectrochemical converter and the template molecules are eluted to obtain the specific pores; The template molecule is configured according to the biochemical molecule to be tested and is the same substance as the biochemical molecule to be tested.

5. A biochemical method for detecting the concentration of multiple molecules, characterized in that, The method includes: A photoelectrochemical detection of a mixture containing at least two analyte biochemical molecules is performed using a MIP membrane-based photoelectrochemical sensor; wherein the MIP membrane is configured to specifically bind only to each of the analyte biochemical molecules. The dataset generated during the photoelectrochemical detection process is input into a pre-trained machine learning model to obtain the biochemical molecule concentration detection results output by the machine learning model.

6. The method according to claim 5, characterized in that, The MIP-based photoelectrochemical sensor is the photoelectrochemical sensor described in claim 1.

7. The method according to claim 5, characterized in that, The dataset is a three-dimensional dataset containing time data, bias voltage data, and photocurrent amplitude data, used to characterize the relationship between bias voltage data and photocurrent amplitude data and time. The machine learning model is pre-trained using the training datasets of each of the at least two biochemical molecules to be tested. The training dataset for each type of biochemical molecule to be tested includes: Multiple photocurrent amplitude curves as a function of time for each biochemical molecule at different concentrations; In each of the aforementioned photocurrent amplitude versus time curves, the curve data points are defined as those acquired after applying one or more bias voltages in each time interval. Each time a bias voltage is applied, a curve data point is collected.

8. The method according to claim 5, characterized in that, The training dataset for the machine learning model is normally distributed, and the top 20% of data that deviates most from the center value are removed.

9. The method according to claim 5, characterized in that, The algorithm used in the machine learning model is either the quadratic rational Gaussian process regression algorithm or the three-layer wide neural network algorithm, with the quadratic rational Gaussian process regression algorithm being preferred.

Images