Carboxyl-functionalized magnetic cof for quantitative detection of organic pollutants in water

Abstract

The application discloses a carboxyl functionalized magnetic COF water organic pollutant quantitative detection method and belongs to the technical field of water organic pollutant detection. The method constructs a carboxyl functionalized magnetic COF composite material as a solid-phase extraction adsorbent, adds the adsorbent into a water sample, separates the loaded adsorbent through an external magnetic field after oscillation adsorption, elutes the adsorbent through an eluent and collects an eluate; the eluate is introduced into a spectral detection channel, and a water sample without adsorption is introduced into a reference detection channel, a dynamic response surface is constructed according to the spectral signal difference of the two channels, the characteristic peak shift and peak intensity attenuation ratio at the adsorption equilibrium point are extracted on the dynamic response surface, input into a pre-trained competitive alternative quantitative model, and the initial concentration of the organic pollutant in the water is output. The method utilizes a two-channel differential dynamic response and a competitive alternative modeling, effectively overcomes complex matrix interference, and realizes high-accuracy quantification of the organic pollutant.

Description

Technical Field

[0001] This invention relates to the field of water organic pollutant detection technology, specifically a method for quantitative detection of carboxyl-functionalized magnetic COF water organic pollutants. Background Technology

[0002] Accurate quantitative detection of organic pollutants in water is a core requirement of environmental monitoring, and solid-phase extraction combined with spectral analysis is a commonly used technical approach. Existing detection schemes typically disperse magnetic solid-phase extraction materials in water samples to capture target pollutants. After magnetic separation and elution, the absorbance signal of the eluent at a characteristic wavelength is collected, and the signal is converted into pollutant concentration using a standard curve model. However, real water bodies contain a large amount of dissolved organic matter, inorganic salts, and other coexisting matrices. These interfering components may be partially co-extracted during elution, directly participating in spectral absorption and generating significant background spurious signals in single-channel detection mode. Simultaneously, the competitive occupancy effect of active sites on the adsorbent surface causes the actual adsorption behavior of the target pollutant to deviate from the ideal state, leading to the invalidation of the linear correspondence between the spectral response and the initial concentration, and resulting in systematic biases in the quantitative results of the standard curve method. Furthermore, existing methods largely rely on a single static spectral reading at the adsorption endpoint, failing to characterize the dynamic evolution information of the entire adsorption process, losing multidimensional criteria for determining the adsorption equilibrium state, and making it difficult to separate highly concentration-sensitive characteristic parameters from the spectral signal, thus limiting the improvement of detection sensitivity and selectivity in complex water samples. Therefore, the key to improving the reliability of rapid on-site quantification of organic pollutants in water bodies lies in how to eliminate background interference from coexisting matrices and extract multidimensional feature information of adsorption equilibrium points from dynamic adsorption spectra, as well as how to construct a quantitative conversion method that can adapt to competitive adsorption behavior and does not rely on ideal standard curves. Summary of the Invention

[0003] This paper provides a quantitative detection method for organic pollutants in water using carboxyl-functionalized magnetic COF. The method extracts the characteristic peak shift and peak intensity attenuation ratio of the adsorption equilibrium state through dual-channel spectral differential dynamic response surface, and uses a competitive substitution quantitative model to accurately estimate the initial concentration. This method solves the problems of quantitative deviation caused by complex water matrix interference and the insufficient adaptability of traditional standard curve models.

[0004] The objective of this invention can be achieved through the following technical solutions:

[0005] This invention provides a method for quantitative detection of organic pollutants in water using carboxyl-functionalized magnetic COF composite materials. This method uses carboxyl-functionalized magnetic covalent organic framework composite materials as solid-phase extraction adsorbents and achieves direct and highly sensitive quantitative analysis of organic pollutants in water through a dual-channel dynamic differential and competitive substitution quantitative model.

[0006] The carboxyl-functionalized magnetic COF composite material, used as a solid-phase extraction adsorbent, is constructed by uniformly dispersing magnetite nanoparticles in an organic solvent to form a magnetic substrate dispersion. A bifunctional ligand containing aldehyde and carboxyl groups, along with melamine monomer, is then added to the dispersion. A carboxyl-functionalized covalent organic framework shell is grown in situ on the surface of the magnetite particles using a solvothermal reaction. After the reaction, the precipitate is collected using an external magnetic field, washed sequentially with organic solvent and deionized water, and then vacuum dried to obtain a composite material with high specific surface area, abundant carboxyl binding sites, and rapid magnetic response. The carboxyl groups on the adsorbent surface form multiple interactions with organic pollutant molecules, including hydrogen bonds and electrostatic interactions, enabling efficient capture and enrichment of organic pollutants in water.

[0007] During testing, the solid-phase extraction adsorbent is added to the water sample according to a preset ratio of adsorbent mass to water sample volume. Under isothermal oscillation conditions, the sample is oscillated for a predetermined time to ensure sufficient capture of organic pollutants. Subsequently, the container is placed above a permanent magnet, and an external magnetic field is used to rapidly concentrate the adsorbent at the bottom of the container. After removing the supernatant, the loaded adsorbent containing organic pollutants is obtained. This magnetic separation process avoids cumbersome centrifugation or filtration steps, significantly reducing pretreatment time.

[0008] A predetermined volume of organic eluent is added to the container containing the above-mentioned loaded adsorbent. Preferably, the organic eluent is at least one of acetonitrile, methanol, or an acetonitrile-ethyl acetate mixture. The organic pollutants are desorbed from the adsorbent and transferred into the eluent by vortexing. An external magnetic field is then applied to cause the eluted magnetic adsorbent to settle, and the supernatant is removed as the eluent. The elution step is repeated twice, and the eluents are combined and concentrated to a fixed volume by nitrogen purging for subsequent detection, ensuring sufficient recovery of the analyte and the strength of the detection signal.

[0009] To eliminate background interference from the water sample itself, a dual-channel spectral detection mode was employed. The eluent was pumped into the first cuvette at a constant flow rate, and the spectrometer continuously acquired the first absorbance sequence within the characteristic absorption wavelength range. Simultaneously, the untreated water sample was pumped into the second cuvette at the same constant flow rate, and the second absorbance sequence within the same characteristic absorption wavelength range was acquired. The absorbance values ​​at corresponding wavelengths in the second absorbance sequence were subtracted point-by-point from the first absorbance sequence to obtain the net absorbance difference sequence. The net absorbance difference sequences acquired at multiple time points were then stitched together to form a dynamic response surface, with time as the horizontal axis, wavelength as the vertical axis, and the net absorbance difference as the vertical axis. This surface comprehensively records the real-time changes in the characteristic spectral signals of organic pollutants during the adsorption-detection process, effectively eliminating interference from the sample matrix.

[0010] After the dynamic response surface is constructed and before feature extraction, background noise subtraction preprocessing can be performed on the dynamic response surface: the baseline spectral difference sequence when no adsorption occurs is collected within the initial time window, the baseline average value at each wavelength position is calculated as the background noise baseline, the net absorbance difference of each wavelength on each time slice is subtracted from the background noise baseline of the corresponding wavelength, and then the moving window smoothing algorithm is used to perform moving average filtering along the time axis to obtain the noise-reduced dynamic response surface, so as to further improve the accuracy of feature extraction.

[0011] The characteristic peak shift and peak intensity attenuation ratio at the adsorption equilibrium point are extracted from the dynamic response surface. Specifically, the reference wavelength position of the standard characteristic peak of the organic pollutant is determined on the surface, and the real-time wavelength position of the actual peak in the spectral curve corresponding to each time slice is tracked. The difference between the real-time wavelength position and the reference wavelength position is calculated as the characteristic peak shift. The real-time absorbance peak value at this real-time wavelength position is extracted, and the ratio of the real-time absorbance peak value to the absorbance peak value of the standard characteristic peak is calculated as the peak intensity attenuation ratio. The adsorption equilibrium time point is automatically determined by monitoring that the rate of change of the characteristic peak shift over time is less than a preset threshold. The characteristic peak shift and peak intensity attenuation ratio corresponding to this time point are selected as input parameters. These two physical quantities reflect the electronic energy level perturbation caused by the interaction between the pollutant and the adsorbent and the signal attenuation caused by the concentration change, respectively, and have clear spectroscopic physical significance for quantitative analysis.

[0012] The extracted characteristic peak offset and peak intensity attenuation ratio are combined into a two-dimensional input feature vector, which is then input into a pre-trained competitive substitution quantitative model to output the initial concentration of organic pollutants in the water. This competitive substitution quantitative model is a nonlinear mapping model based on a radial basis function neural network, with the number of hidden layer nodes set to 1.5 to 2 times the number of standard pollutant concentration gradients. In the hidden layer, the competitive substitution quantitative model performs a dot product operation between the two-dimensional input feature vector and multiple preset standard pollutant concentration corresponding to standard offset-standard attenuation ratio vector pairs to generate multiple competitive scores. The competitive scores are converted into a probability distribution using a normalized exponential function, and the standard pollutant concentration corresponding to the highest probability is selected as the initial estimated concentration. Interpolation correction is then performed based on the difference between this concentration and the competitive scores corresponding to adjacent concentrations to obtain the final output initial concentration of organic pollutants. The model pre-prepares a series of organic pollutant standard solutions of varying concentrations, sequentially performing oscillatory adsorption, magnetic separation, elution, and dual-channel spectral acquisition. The standard dynamic response surface corresponding to each concentration and the standard characteristic peak offset and standard peak intensity attenuation ratio at the equilibrium point are extracted, forming standard offset-standard attenuation ratio vector pairs, which are stored as a model reference lookup table. This model eliminates the reliance on linear calibration based on a single peak area or absorbance value in traditional methods. It achieves precise quantification of trace organics in complex competitive adsorption systems through a competitive matching mechanism and nonlinear mapping, thereby improving the accuracy and anti-interference capability of the method.

[0013] As a technical solution of the present invention, the method further includes regenerating the carboxyl-functionalized magnetic COF composite material after adsorption-elution. The separated, used composite material is added to a regeneration eluent and ultrasonically vibrated for a predetermined time to remove residual organic contaminants. After collection under an external magnetic field, it is repeatedly washed with deionized water until neutral, vacuum dried, and weighed to calculate the mass recovery rate. If the mass recovery rate is lower than a preset recovery rate threshold, the batch of material is determined not to be reused. This regeneration step ensures the reusability of the adsorbent, reduces detection costs, and meets the requirements of green analytical chemistry.

[0014] The beneficial effects of this invention are:

[0015] The spectral signal is simultaneously acquired by the eluent spectral detection channel and the reference detection channel of the unadsorbed water sample. The absorbance of the two channels at corresponding wavelengths is calculated in real time to generate a net absorbance difference sequence, which is then stitched together to form a dynamic response surface. This dual-channel differential measurement method can remove common-mode noise such as inherent background absorption and light source drift in the water at each time slice. The dynamic response surface completely preserves the spectral position and intensity evolution information during the adsorption process. The reference wavelength position of the standard characteristic peak of organic pollutants is located on the dynamic response surface. The real-time wavelength position and real-time absorbance peak value of the actual peak in each time slice are tracked, and the characteristic peak shift and peak intensity decay ratio are calculated. The value at the adsorption equilibrium point is used as the output. The characteristic peak shift reflects the electronic energy level perturbation caused by the interaction between the adsorbate and the adsorbent, while the peak intensity decay ratio corresponds to the distribution ratio of substances between the adsorbed and unadsorbed states. Together, they form a multidimensional feature vector that is highly sensitive to the initial concentration and is stripped of background interference, significantly improving the quantitative resolution over a wide concentration range.

[0016] A two-dimensional input feature vector, composed of characteristic peak offset and peak intensity attenuation ratio, is input into the competitive substitution quantitative model. This model uses a series of standard offset-standard attenuation ratio vector pairs corresponding to standard solutions of varying concentrations as a reference lookup table. In the hidden layer, the competition score between the input vector and each standard vector pair is calculated through dot product operations. The competition score is then converted into a probability distribution using a normalized exponential function. The initial concentration of organic pollutants is output after interpolation correction based on the standard concentration corresponding to the highest probability. The competitive substitution mechanism does not rely on a pre-defined linear equation between spectral intensity and concentration. It matches target samples through vector similarity competition, naturally accommodating the nonlinear response caused by competitive adsorption. Interpolation correction further weakens the discrete errors caused by concentration gradient intervals, ensuring robust accuracy of the quantitative results even under cross-interference from coexisting organic compounds. Attached Figure Description

[0017] The invention will now be further described with reference to the accompanying drawings.

[0018] Figure 1 This is a flowchart of a method for quantitative detection of organic pollutants in water using carboxyl-functionalized magnetic COF.

[0019] Figure 2 This is a schematic diagram of the solid-phase extraction process of carboxyl-functionalized magnetic COF composite materials;

[0020] Figure 3 This is a flowchart of the organic pollutant elution and eluent concentration process. Detailed Implementation

[0021] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0022] See Figure 1 This invention provides a method for quantitative detection of organic pollutants in water using carboxyl-functionalized magnetic COF composite material, comprising: constructing a carboxyl-functionalized magnetic COF composite material as a solid-phase extraction adsorbent; adding the solid-phase extraction adsorbent to a water sample for oscillatory adsorption, and separating the loaded adsorbent after organic pollutant loading by applying an external magnetic field; eluting the loaded adsorbent with an eluent and collecting the eluent; introducing the eluent into a spectral detection channel, and simultaneously introducing an unadsorbed water sample into a reference detection channel, constructing a dynamic response surface based on the difference between the two-channel spectral signals; extracting the characteristic peak shift and peak intensity attenuation ratio at the adsorption equilibrium point from the dynamic response surface, and inputting the characteristic peak shift and peak intensity attenuation ratio into a pre-trained competitive substitution quantitative model to output the initial concentration of organic pollutants in the water.

[0023] In practice, the process of constructing carboxyl-functionalized magnetic COF composite materials as solid-phase extraction adsorbents begins with the preparation of magnetite magnetic nanoparticles. The magnetite magnetic nanoparticles are synthesized using a chemical coprecipitation method. Ferric chloride and ferrous chloride are dissolved in deionized water at a ferric ion molar ratio of 2:1. Under a nitrogen atmosphere, ammonia is added dropwise with continuous stirring, generating a black precipitate. The precipitate is then collected using an external permanent magnet. The precipitate is washed alternately with deionized water and anhydrous ethanol until the washing solution is neutral. After vacuum drying, the magnetite magnetic nanoparticles are obtained.

[0024] In the specific implementation, magnetite nanoparticles are uniformly dispersed in an organic solvent to form a magnetic substrate dispersion. The amount of magnetite nanoparticles used is 0.1 g per 20 ml of organic solvent. The organic solvent used is N,N-dimethylformamide. The magnetite nanoparticles are added to a round-bottom flask containing N,N-dimethylformamide and ultrasonically dispersed for 30 minutes to form a uniform magnetic substrate dispersion.

[0025] In the specific implementation, a bifunctional ligand containing aldehyde and carboxyl groups and a melamine monomer are added to the magnetic substrate dispersion. The bifunctional ligand containing aldehyde and carboxyl groups is 4-formylbenzoic acid, and the amount of 4-formylbenzoic acid added is 0.5 mmol; the amount of melamine monomer added is 0.33 mmol. 4-Formylbenzoic acid and melamine monomer are added to the magnetic substrate dispersion sequentially, and ultrasonic dispersion is continued for 15 minutes to ensure that 4-formylbenzoic acid and melamine monomer are completely dissolved and uniformly dispersed in the magnetic substrate dispersion.

[0026] In the specific implementation, a carboxyl-functionalized covalent organic framework shell was grown in situ on the surface of magnetite (Fe3O4) magnetic nanoparticles via a solvothermal reaction. A round-bottom flask containing the mixed dispersion was transferred to an oil bath, heated to 120°C under nitrogen protection, and stirred at this temperature for 24 hours. During the solvothermal reaction, the aldehyde group of 4-formylbenzoic acid undergoes a Schiff base condensation reaction with the primary amino group of melamine monomer. The resulting covalent organic framework oligomer gradually deposits and crosslinks on the surface of the magnetite (Fe3O4) magnetic nanoparticles, forming a carboxyl-functionalized covalent organic framework shell encapsulating the magnetite (Fe3O4) magnetic nanoparticles. After the reaction, the mixture was allowed to cool naturally to room temperature.

[0027] In practice, after the reaction is complete, the precipitate is collected using an external magnetic field. The cooled reaction solution is transferred to a beaker, and a neodymium iron boron permanent magnet is attached to the bottom of the beaker's outer wall. The beaker is left to stand for 5 minutes, allowing the magnetically responsive carboxyl-functionalized magnetic COF composite material particles to aggregate at the bottom of the beaker under the influence of the external magnetic field. The supernatant is then discarded, retaining the precipitate that has accumulated at the bottom.

[0028] In practice, the precipitate was washed sequentially with an organic solvent and deionized water. The organic solvent was anhydrous ethanol. First, the precipitate was redispersed with anhydrous ethanol, then sonicated for 5 minutes. The precipitate was then collected again using an external magnetic field, the washing liquid was discarded, and the washing with anhydrous ethanol was repeated three times. The precipitate was then washed three times with deionized water following the same procedure, using the same ultrasonic dispersion and external magnetic field collection methods during the deionized water washing process.

[0029] In practice, carboxyl-functionalized magnetic COF composite materials are obtained after vacuum drying. The washed precipitate is placed in a vacuum drying oven and dried at 60 degrees Celsius for 12 hours to obtain a black powdery solid, which is the carboxyl-functionalized magnetic COF composite material. The thickness of the carboxyl-functionalized covalent organic framework shell of the carboxyl-functionalized magnetic COF composite material is controlled by the synergistic effect of the feed ratio of 4-formylbenzoic acid to melamine monomer and the solvothermal reaction time. Under the above material ratio and reaction conditions, the carboxyl-functionalized covalent organic framework shell formed is uniform and dense, and the magnetite nanoparticles are completely coated by the carboxyl-functionalized covalent organic framework shell.

[0030] In specific implementation, please refer to Figure 2 The process involves adding a solid-phase extraction (SPE) adsorbent to a water sample and performing a shaking adsorption, establishing a predetermined ratio of adsorbent mass to water sample volume. Carboxyl-functionalized magnetic COF composite material is weighed as the SPE adsorbent, with a predetermined ratio of 1 mg SPE adsorbent to 1 mL water sample. When the water sample volume is 100 mL, 100 mg of carboxyl-functionalized magnetic COF composite material is weighed as the SPE adsorbent. The water sample is surface water or industrial wastewater containing the organic pollutants to be tested. Before adding the SPE adsorbent, the water sample is filtered through a 0.45 μm filter membrane to remove suspended particulate matter.

[0031] In the specific implementation, the solid-phase extraction adsorbent is added to the water sample. 100 mg of the weighed carboxyl-functionalized magnetic COF composite material is transferred to a 150 mL stoppered conical flask, which already contains 100 mL of filtered water sample. After sealing the flask, it is placed in a temperature-controlled shaker.

[0032] In the specific implementation, the temperature of the isothermal shaker was set to 25 degrees Celsius, the shaker speed was set to 150 revolutions per minute, and the first predetermined time was set to 30 minutes. The isothermal shaker continuously oscillated the stoppered conical flask at a speed of 150 revolutions per minute for 30 minutes, maintaining a constant temperature of 25 degrees Celsius during the oscillation process. During the oscillation process, the carboxyl-functionalized magnetic COF composite material was uniformly dispersed and kept in suspension in the water sample. Organic pollutant molecules in the water sample came into contact with the carboxyl-functionalized magnetic COF composite material. The organic pollutant molecules were captured by the solid-phase extraction adsorbent through π-π stacking, electrostatic interaction, and the size selectivity effect provided by the pores of the carboxyl-functionalized magnetic COF composite material, and adsorbed on the inner surface of the pores of the carboxyl-functionalized covalent organic framework shell and on the carboxyl active sites.

[0033] In practice, after the shaking is complete, the container holding the water sample is placed above the permanent magnet and left to stand for a second predetermined time. After the stoppered conical flask is removed from the thermostatic shaker, it is placed above a neodymium iron boron (NdFeB) permanent magnet with a surface magnetic induction intensity of not less than 0.4 Tesla. The second predetermined time is set to 3 minutes. The stoppered conical flask is left to stand above the NdFeB permanent magnet for 3 minutes, allowing the carboxyl-functionalized magnetic COF composite material loaded with organic pollutants to aggregate at the bottom of the stoppered conical flask near the permanent magnet under the influence of the applied magnetic field, forming a clear interface between the supernatant of the water sample and the aggregated solid-phase extraction adsorbent.

[0034] In practice, the loaded adsorbent with organic pollutants is obtained after removing the supernatant. The supernatant of the water sample is slowly aspirated along the inner wall of a stoppered conical flask using a pipette, with the pipette tip always positioned 2 mm below the liquid surface to avoid disturbing the loaded adsorbent accumulated at the bottom. The pipetting operation is stopped when approximately 2 ml of supernatant remains. The remaining small amount of supernatant is carefully removed by decantation. The black solid remaining at the bottom of the stoppered conical flask is the loaded adsorbent with organic pollutants. The loaded adsorbent can be used directly in subsequent elution operations without additional drying treatment.

[0035] In specific implementation, please refer to Figure 3 A predetermined volume of organic eluent was added to a container holding the loaded adsorbent containing organic pollutants. The loaded adsorbent was a carboxyl-functionalized magnetic COF composite material loaded with organic pollutants, obtained through the treatment described in the above embodiment. The predetermined volume was set to 2 ml. Acetonitrile was used as the organic eluent. 2 ml of acetonitrile was added to a stoppered conical flask containing the loaded adsorbent, ensuring the acetonitrile completely submerged the loaded adsorbent.

[0036] In the specific implementation, the organic pollutants are desorbed from the supported adsorbent and transferred into the eluent by oscillating in a vortex oscillator for a predetermined third time. A stoppered conical flask is placed on the oscillation platform of the vortex oscillator, with the oscillation speed set to 2500 rpm and the third predetermined time set to 2 minutes. The vortex oscillator oscillates continuously at 2500 rpm for 2 minutes. During this process, the supported adsorbent is fully dispersed in acetonitrile, and the carboxyl-functionalized covalent organic framework shell on the surface of the supported adsorbent comes into contact with acetonitrile. Acetonitrile molecules competitively occupy the adsorption sites on the carboxyl-functionalized covalent organic framework shell, causing the organic pollutant molecules to desorb from the adsorption sites and dissolve in acetonitrile. The organic pollutants are thus transferred from the supported adsorbent to the acetonitrile solution.

[0037] In the specific implementation, the container is placed above the applied magnetic field again and left to stand for a fourth predetermined time. After the eluted magnetic adsorbent has completely settled, the supernatant is transferred to the sample vial as the eluent. The stoppered conical flask after vortex oscillation is removed from the vortex oscillator and placed above a NdFeB permanent magnet with a surface magnetic induction intensity of not less than 0.4 Tesla. The fourth predetermined time is set to 2 minutes. The stoppered conical flask is left to stand above the NdFeB permanent magnet for 2 minutes. Under the action of the applied magnetic field, the eluted carboxyl-functionalized magnetic COF composite material gathers at the bottom of the stoppered conical flask, and a clear interface is formed between the eluent supernatant and the gathered carboxyl-functionalized magnetic COF composite material. The supernatant is transferred entirely to a clean 10 mL volumetric flask using a pipette. During the pipetting process, the tip of the pipette tip should always avoid contact with the gathered carboxyl-functionalized magnetic COF composite material at the bottom. The transferred supernatant is used as the eluent obtained from the first elution.

[0038] In practice, the elution step was repeated twice, and the eluents obtained from each step were combined. After the first eluent collection, 2 mL of acetonitrile was added again to the remaining carboxyl-functionalized magnetic COF composite material at the bottom of the stoppered conical flask. The mixture was vortexed at 2500 rpm for 2 minutes under the same conditions, and then allowed to stand for 2 minutes under the same external magnetic field. The supernatant obtained from the second elution was then combined with the supernatant in the same 10 mL volumetric flask using a pipette. A third elution was then performed, with another 2 mL of acetonitrile added. The same vortexing, external magnetic field separation, and supernatant transfer procedures were repeated, and the supernatant obtained from the third elution was combined with the supernatant in the 10 mL volumetric flask. After three elutions, the total volume of the combined eluent in the 10 mL volumetric flask was 6 mL.

[0039] In practice, nitrogen purging is used to concentrate the eluent to a fixed volume for subsequent detection. A 10 mL volumetric flask containing 6 mL of combined eluent is transferred to the sample position of the nitrogen purging apparatus. The height of the nitrogen purging needle is adjusted so that the tip is 5 mm above the surface of the eluent. The nitrogen purity is not less than 99.9%, the nitrogen purging pressure is set to 0.05 MPa, and the water bath temperature is set to 40°C. Nitrogen purging is initiated, and a continuous stream of high-purity nitrogen purges the surface of the eluent. Acetonitrile gradually evaporates under the action of the nitrogen stream, and the volume of the eluent continuously decreases. When the eluent volume is concentrated to 1 mL, nitrogen purging is stopped, the volumetric flask is removed, and the volume is precisely diluted to the 1 mL mark with acetonitrile to obtain the concentrated eluent. The concentrated eluent is directly used for subsequent sample injection analysis in the spectral detection channel. The fixed volume is set to 1 mL. After concentrating to this fixed volume, the concentration of organic pollutants in the eluent increases by 6 times compared to before elution, thereby improving detection sensitivity.

[0040] In practice, the process of introducing the eluent into the spectral detection channel is as follows: The concentrated eluent from the above embodiment is pumped into the first cuvette at a constant flow rate. One end of the pump tube of the constant-flow peristaltic pump is inserted into the sample vial containing the concentrated eluent, and the other end of the pump tube is connected to the inlet of a quartz flow-through cuvette with an optical path length of 10 mm. The constant flow rate of the peristaltic pump is set to 1.5 mL per minute. Driven by the constant-flow peristaltic pump, the concentrated eluent continuously flows through the first cuvette at a constant flow rate of 1.5 mL per minute. The outlet of the first cuvette is connected to a waste collection bottle.

[0041] In practice, a spectrometer continuously acquires the first absorbance sequence of the eluent in the first cuvette within the characteristic absorption wavelength range. The spectrometer is equipped with a deuterium-tungsten lamp composite light source and a CCD linear array detector. The characteristic absorption wavelength range is set to 200 nm to 400 nm, the spectral acquisition interval is set to 0.5 seconds per frame, and the spectral resolution is set to 0.5 nm. The spectrometer's optical path is aligned with the center of the first cuvette. When the concentrated eluent begins to flow through the first cuvette, the spectrometer simultaneously starts continuous acquisition, recording the absorbance value of each frame of the spectrum within the 200 nm to 400 nm wavelength range, and arranging them in chronological order of acquisition to form the first absorbance sequence.

[0042] In the specific implementation, the unadsorbed water sample is synchronously pumped into the second cuvette at the same constant flow rate. The spectrometer simultaneously acquires the second absorbance sequence of the water sample in the second cuvette within the same characteristic absorption wavelength range. The second cuvette is a quartz flow-through cuvette with the same optical path as the first cuvette, with a path length of 10 mm. A constant-flow peristaltic pump is used, with its constant flow rate also set to 1.5 mL / min. One end of the pump tube is inserted into the sample bottle containing the unadsorbed original water sample, and the other end is connected to the inlet of the second cuvette. The unadsorbed water sample refers to the original water sample from the same batch as used in the previous embodiment, but without the addition of carboxyl-functionalized magnetic COF composite material for oscillatory adsorption. The spectrometer alternately acquires the spectral signals of the first and second cuvettes via an optical path switching device. The optical path switching frequency is 1 Hz, meaning that after acquiring one frame of spectral data from the first cuvette, the spectrometer immediately switches the optical path to acquire one frame of spectral data from the second cuvette. The spectral data collected in the second cuvette are arranged in chronological order of collection time to form the second absorbance sequence.

[0043] In practice, the operation of subtracting the absorbance values ​​of the corresponding wavelengths in the second absorbance sequence point by point from the first absorbance sequence to obtain the net absorbance difference sequence is performed by the data processing unit. The data processing unit is a computer workstation connected to the spectrometer, which receives all spectral data frames from both the first and second absorbance sequences. For each acquisition time point, the spectral curve corresponding to that acquisition time point in the first absorbance sequence is plotted at wavelength... The absorbance value at that point is denoted as The spectral curves corresponding to the same acquisition time point in the second absorbance sequence are plotted at wavelength. The absorbance value at that point is denoted as The net absorbance difference is calculated using the following formula:

[0044]

[0045] in, Represents the first characteristic absorption wavelength range Each sampling wavelength point, The value range is 1 to 401, corresponding to all wavelength sampling points in the wavelength range of 200 nm to 400 nm with an interval of 0.5 nm; Indicates the first Each data collection time point The value range is 1 to , This refers to the total number of frames acquired from the start of sample injection until all the eluent has flowed through the first cuvette. The specific value is calculated by dividing the total volume of the eluent (1 ml) by the constant flow rate (1.5 ml / min) and then multiplying by the acquisition interval of 0.5 seconds per frame. The value is 80; Indicates the first Each acquisition time point and wavelength The absorbance of the eluent in the first cuvette was measured. Indicates the first Each acquisition time point and wavelength The absorbance measurement of the unadsorbed water sample in the second cuvette; Indicates the first Each acquisition time point and wavelength The net absorbance difference at each of the 80 acquisition time points and 401 wavelength sampling points was calculated using a computer workstation to obtain a sequence of net absorbance difference values.

[0046] In practice, a dynamic response surface is formed by stitching together the net absorbance difference sequences from multiple consecutively acquired time points, with time as the horizontal axis, wavelength as the vertical axis, and net absorbance difference as the vertical axis. The computer workstation then summarizes the net absorbance difference sequences from all 80 acquisition time points into a matrix with dimensions of 80 rows × 401 columns. The rows of the matrix correspond to the time axis, and the row indices are the acquisition time point numbers. The columns of the matrix correspond to the wavelength axis, and the column indices are the wavelength sampling point numbers. The first in the matrix Line number The element value of the column is The computer workstation invokes the 3D surface rendering function, in time... The horizontal axis represents the wavelength. The vertical axis represents the difference in net absorbance. Using the vertical axis, all element values ​​in the matrix are plotted as a three-dimensional surface, which is the dynamic response surface.

[0047] In practice, after constructing the dynamic response surface (RTS), background noise subtraction is performed on the RTS preprocessing. Baseline spectral difference sequences are acquired within the initial time window of the RTS when no adsorption occurs. The initial time window is set from the first to the tenth acquisition time point, corresponding to the period from sample injection to before the eluent front enters the first cuvette. During this period, both the first and second cuvettes contain background solutions with no signal change. The computer workstation reads all net absorbance difference data corresponding to the 10 acquisition time points within the initial time window, and performs analysis on each wavelength. Calculate the average of the net absorbance differences at 10 acquisition time points, and use the average value as the wavelength. The background noise baseline at that location is denoted as .

[0048] In practice, the net absorbance difference for each wavelength at each time slice in the dynamic response surface is subtracted from the background noise baseline for the corresponding wavelength to obtain the background-subtracted dynamic response surface. This is done for each of the 80 acquisition time points. and each wavelength in all 401 wavelength sampling points Computer workstations perform calculations ,in, This represents the net absorbance difference after deducting the background. This represents the difference in original net absorbance. Indicates wavelength The background noise baseline is determined. The computer workstation then uses the net absorbance difference after removing all background values ​​to form a new matrix, generating a dynamic response surface after background subtraction.

[0049] In practice, a moving window smoothing algorithm is used to apply a moving average filter along the time axis to the background-subtracted dynamic response surface to obtain the denoised dynamic response surface. The window width of the moving window smoothing algorithm is set to 5 acquisition time points. For each fixed wavelength in the background-subtracted dynamic response surface... Take the current acquisition time point along the time axis. The net absorbance difference after background subtraction at five consecutive time points centered on the sample, including , , , , The arithmetic mean of the five net absorbance differences after background subtraction was calculated as... The net absorbance difference after filtering at each time point. This is calculated for the start and end points of the time axis, i.e., when... When the window cannot be filled at the 1st, 2nd, or 79th and 80th acquisition time points, use the [missing information - likely a specific method or technique] respectively. and the Two points or the first and the The two-point bilateral padding completes the moving average calculation. The computer workstation performs moving window moving average filtering along the time axis for all 401 wavelength trajectories, generating a denoised dynamic response surface for subsequent feature extraction.

[0050] In practice, the characteristic peak shift and peak intensity attenuation ratio at the adsorption equilibrium point are extracted from the denoised dynamic response surface. The reference wavelength position of the organic pollutant standard characteristic peak is located on the denoised dynamic response surface. This reference wavelength position is determined by the pre-measured peak position of the absorption spectrum of the organic pollutant standard within the characteristic absorption wavelength range. The reference wavelength position is denoted as... The unit is nanometers. The value is taken as the wavelength corresponding to the maximum absorption peak of the organic pollutant standard in acetonitrile solution.

[0051] In practice, the real-time wavelength position of the actual peak in the spectral curve corresponding to each time slice on the denoised dynamic response surface is tracked. Each collection time point The corresponding spectral curve, with the computer workstation at the reference wavelength position. The wavelength position of the maximum absorbance is searched within a range of ±10 nanometers, and the wavelength position corresponding to the maximum absorbance is determined as... The actual peak real-time wavelength position at time is denoted as The difference between the real-time wavelength position and the reference wavelength position is calculated as the characteristic peak offset. The calculation method is as follows .

[0052] In practice, the real-time absorbance peak value corresponding to the real-time wavelength position is extracted and recorded as... The standard absorbance peak value of the standard characteristic peak of organic pollutants is a fixed value obtained from a pre-determined concentration. Organic pollutant standards in The absorbance at a given point is determined, and the standard absorbance peak is denoted as . Calculate the real-time absorbance peak. Compared with standard absorbance peak The ratio between them is used as the peak intensity attenuation ratio. The calculation method is as follows .

[0053] In practice, the characteristic peak shift and peak intensity attenuation ratio corresponding to the adsorption equilibrium time point are selected as the output values. The adsorption equilibrium time point is automatically determined by monitoring the rate of change of the characteristic peak shift over time on the noise-reduced dynamic response surface, which is less than a preset threshold. The computer workstation calculates the rate of change of the characteristic peak shift between two adjacent acquisition time points. The rate of change is calculated by dividing the absolute value of the difference between the characteristic peak shifts of the two adjacent time points by the acquisition interval of 0.5 seconds. When the rate of change of the characteristic peak shift over time is less than the preset threshold for five consecutive acquisition time points, the first acquisition time point among the five consecutive acquisition time points is determined as the adsorption equilibrium time point, denoted as... The preset threshold is set to 0.02 nanometers per second. This threshold is based on the spectrometer's wavelength resolution of 0.5 nanometers. A change rate of 0.02 nanometers per second corresponds to a wavelength change of 0.01 nanometers between adjacent acquisition time points. This value is below the minimum wavelength change that the spectrometer can resolve, indicating that the characteristic peak shift has stabilized. (Computer workstation output) characteristic peak shift at time Peak intensity attenuation ratio As the extraction result.

[0054] In specific implementation, the characteristic peak shift and peak intensity attenuation ratio at the adsorption equilibrium time point output in the above embodiment are combined into a two-dimensional input feature vector. The characteristic peak shift is denoted as... The unit is nanometers, and the peak intensity attenuation ratio is denoted as . , which is a dimensionless ratio. and Arranged in order, they form a vector containing two elements, denoted as . ,in, This represents the transpose of a vector. Two-dimensional input feature vectors. Input data as a pre-trained, competitive alternative quantitative model.

[0055] In practical implementation, the quantitative model for competitive substitution is a nonlinear mapping model based on a radial basis function neural network. The input layer of the radial basis function neural network consists of two nodes, which respectively receive... and The hidden layers of the radial basis function neural network store multiple preset standard offset and standard attenuation ratio vector pairs corresponding to standard pollutant concentrations. The number of hidden layer nodes is related to the number of standard pollutant concentration gradients, and the number of hidden layer nodes is set to twice the number of standard pollutant concentration gradients.

[0056] In practical implementation, the number of standard pollutant concentration gradients is set to 12, and the number of hidden layer nodes is set to 24. The 12 standard pollutant concentration gradients are determined by a series of prepared organic pollutant standard solutions with varying concentrations. The concentration values ​​of the organic pollutant standard solutions are expressed in micrograms per liter, and the 12 concentration gradient values ​​are 0.5, 1.0, 2.0, 5.0, 10.0, 20.0, 50.0, 100.0, 200.0, 500.0, 1000.0, and 2000.0. These 12 standard pollutant concentrations are denoted as... Furthermore, the standard offset and standard attenuation ratio vector pairs corresponding to the 12 standard pollutant concentrations are copied and expanded to fill the 24 nodes of the hidden layer with twice the number of gradients. Each concentration gradient corresponds to 2 hidden layer nodes, and these 2 hidden layer nodes store the same standard offset and standard attenuation ratio vector pairs to enhance model stability.

[0057] In specific implementation, the standard offset and standard attenuation ratio vector pairs corresponding to the preset standard pollutant concentrations are obtained as follows: Twelve graded concentrations of organic pollutant standard solutions are prepared. For each concentration standard solution, oscillation adsorption, external magnetic field separation, eluent collection, and dual-channel spectral acquisition processing (spectral detection channel and reference detection channel) are performed sequentially. The specific operation steps are the same as those for the water sample processing in some embodiments. The standard dynamic response surface corresponding to each concentration standard solution is extracted, and the standard characteristic peak offset and standard peak intensity attenuation ratio at the equilibrium point are extracted from each standard dynamic response surface. For the first... Standard concentration The extracted standard feature peak shift is denoted as The standard peak intensity attenuation ratio is denoted as The standard offset standard attenuation ratio vector pair The concentrations of all 12 standard pollutants were... and their corresponding standard offset standard attenuation ratio vector pairs A reference lookup table for the competitive alternative quantitative model is stored in a database file on the computer workstation.

[0058] In practice, the parameters of the 24 hidden layer nodes are set as follows: The first 12 hidden layer nodes store 12 pairs of standard offset and standard attenuation ratio vectors, respectively. The center vector of each hidden layer node is set as follows: , The last 12 hidden layer nodes store these 12 standard offset standard attenuation ratio vector pairs again in the same order. The center vector of each hidden layer node is also set to... The radial basis functions of the hidden layer nodes all adopt the form of Gaussian kernel functions, but the nonlinear mapping of the radial basis function neural network uses dot product operation to replace the distance metric to calculate the similarity between the input vector and the center vector.

[0059] In practical implementation, the two-dimensional input feature vector is used in the hidden layer of the competitive substitution quantitative model. Paired with multiple preset standard offset standard attenuation ratio vectors Perform a dot product operation to calculate multiple competition scores. For the... Hidden layer nodes, competition score The calculation formula is:

[0060]

[0061] in, This is the index number of the hidden layer node. The value of is an integer ranging from 1 to 24; For the first The competition score calculated by each hidden layer node. The larger the value, the more significant the two-dimensional input feature vector. With the The standard offset and standard attenuation ratio vectors stored in each node The higher the degree of matching; The shift of the characteristic peak extracted at the adsorption equilibrium time point is expressed in nanometers. The peak intensity decay ratio extracted at the adsorption equilibrium time point is a dimensionless value. For the first The standard feature peak offset corresponding to each hidden layer node, in nanometers; For the first The standard peak intensity attenuation ratio corresponding to each hidden layer node is a dimensionless value. 24 competition scores. After all calculations are completed, the data is passed to the output processing stage.

[0062] In practice, multiple competition scores are converted into probability distributions using a normalized exponential function. The normalized exponential function takes the form of a softmax function, and for the ... The probability of the standard concentration corresponding to a hidden layer node being selected. The calculation method is as follows: ,in, Represented by natural constant As the base, based on competitive scoring This is an exponentiation operation, with the denominator being the sum of the exponential function values ​​of all 24 competition scores. Probability distribution. satisfy The standard pollutant concentration with the highest probability is selected as the initial predicted concentration; that is, the concentration that makes the prediction most likely is found. Largest index Initial estimated concentration The value is the index. The corresponding standard pollutant concentration, if ,but ;like ,Pick .

[0063] In practice, the initial estimated concentration is interpolated and corrected based on the difference in competition scores between the initial estimated concentration and the concentrations of two adjacent standard pollutants to obtain the final output initial concentration of organic pollutants. (Initial estimated concentration) The sequence number in the standard concentration series is denoted as... ,satisfy .when When there is no adjacent lower bound, take And the corresponding competition score Set to 0; when When there is no adjacent upper bound, take And the corresponding competition score Set as Standard concentration , and The corresponding competition score is determined by the score of the corresponding hidden layer node. If the same standard concentration corresponds to two hidden layer nodes, the average of the competition scores of the two nodes is taken as the representative competition score for that standard concentration, denoted as […]. , and The interpolation correction uses a weighted average method, and the final output is the initial concentration of organic pollutants. for:

[0064]

[0065] in This represents the difference in competition scores between two adjacent standard concentrations. Used for normalization, This is the half-interval between adjacent standard concentrations. When the calculated... When the concentration exceeds the lower limit of the standard range, The concentration was 0.5 micrograms per liter; when When the concentration exceeds the upper limit of the standard range, The limit is 2000.0 micrograms per liter.

[0066] In specific implementation, the process also includes a step of regenerating the adsorbed and eluted carboxyl-functionalized magnetic COF composite material. The used carboxyl-functionalized magnetic COF composite material, collected after elution and separated by an external magnetic field in the above embodiment, is transferred from a stoppered conical flask to a 10 mL screw-top glass bottle. A regeneration eluent, a 1:1 methanol-acetic acid mixture (volume ratio 1:1), is added to the screw-top glass bottle in a volume of 5 mL. The bottle is then placed in an ultrasonic cleaner with an ultrasonic frequency of 40 kHz, an ultrasonic power of 100 W, and a predetermined ultrasonic oscillation time of 15 minutes. During ultrasonic oscillation, the carboxyl-functionalized magnetic COF composite material is fully dispersed in the regeneration eluent, and residual organic contaminants desorb and dissolve from the carboxyl-functionalized covalent organic framework shell into the regeneration eluent.

[0067] In the specific implementation, after ultrasonic oscillation, a screw-top glass bottle is placed above a neodymium iron boron permanent magnet with a surface magnetic induction intensity of not less than 0.4 Tesla and left to stand for 2 minutes. The regenerated carboxyl-functionalized magnetic COF composite material is collected by applying an external magnetic field, and the supernatant is discarded. The collected carboxyl-functionalized magnetic COF composite material is repeatedly washed with deionized water, with 5 ml of deionized water added each time and ultrasonically dispersed for 1 minute. The precipitate is then collected by applying an external magnetic field, and the washing liquid is discarded. The washing with deionized water is repeated until the washing liquid is neutral when tested with pH paper. The washed regenerated carboxyl-functionalized magnetic COF composite material is placed in a vacuum drying oven and vacuum dried at 60 degrees Celsius for 6 hours.

[0068] In practice, after vacuum drying, the material is weighed, and the mass recovery rate is calculated. The mass recovery rate is defined as the ratio of the mass of the regenerated carboxyl-functionalized magnetic COF composite material to the mass of the carboxyl-functionalized magnetic COF composite material used before regeneration, expressed as a percentage. If the mass recovery rate is lower than a preset recovery rate threshold, the batch of carboxyl-functionalized magnetic COF composite material is determined not to be reused. The preset recovery rate threshold is set at 85%. This threshold is set based on the fact that when the mass recovery rate is lower than 85%, the adsorption capacity of the carboxyl-functionalized magnetic COF composite material has decreased by more than 15% in multiple reuses, and the accuracy of quantitative detection results will decrease due to insufficient adsorbent capacity. Therefore, the batch is determined not to be recycled. If the mass recovery rate is not lower than 85%, the regenerated carboxyl-functionalized magnetic COF composite material can be reused as a solid-phase extraction adsorbent in the next round of detection.

[0069] The foregoing has provided a detailed description of one embodiment of the present invention, but this description is merely a preferred embodiment and should not be construed as limiting the scope of the invention. All equivalent variations and modifications made within the scope of the claims of this invention should still fall within the patent coverage of this invention.

Claims

1. A method for quantitative detection of organic pollutants in water using carboxyl-functionalized magnetic COF, characterized in that, The method includes: Constructing carboxyl-functionalized magnetic COF composite materials as solid-phase extraction adsorbents; The solid-phase extraction adsorbent is added to a water sample for oscillation adsorption, and the loaded adsorbent after loading organic pollutants is separated by an external magnetic field. The loaded adsorbent is eluted with an eluent, and the eluent is collected. The eluent is introduced into the spectral detection channel, while an unadsorbed water sample is introduced into the reference detection channel. A dynamic response surface is constructed based on the difference in the dual-channel spectral signals, specifically including: The eluent is pumped into the first cuvette at a constant flow rate, and the first absorbance sequence of the eluent in the first cuvette in the characteristic absorption wavelength range is continuously collected by a spectrometer. Unadsorbed water samples are simultaneously pumped into the second cuvette at the same constant flow rate, and the spectrometer simultaneously collects the second absorbance sequence of the water samples in the second cuvette within the same characteristic absorption wavelength range. Subtract the absorbance values ​​of the corresponding wavelengths in the second absorbance sequence point by point from the first absorbance sequence to obtain the net absorbance difference sequence; The dynamic response surface is formed by stitching together the net absorbance difference sequence of multiple time points collected continuously, with time as the horizontal axis, wavelength as the vertical axis, and net absorbance difference as the vertical axis. The characteristic peak shift and peak intensity attenuation ratio at the adsorption equilibrium point are extracted from the dynamic response surface. These are then input into a pre-trained competitive substitution quantitative model to output the initial concentration of organic pollutants in the water body, specifically including: The characteristic peak offset and peak intensity attenuation ratio are combined into a two-dimensional input feature vector, and the two-dimensional input feature vector is passed to the input layer of the competitive alternative quantitative model. In the hidden layer of the competitive substitution quantitative model, the two-dimensional input feature vector is multiplied by the standard offset and standard attenuation ratio vectors corresponding to multiple preset standard pollutant concentrations to calculate multiple competitive scores. The multiple competition scores are converted into a probability distribution using a normalized exponential function, and the standard pollutant concentration corresponding to the highest probability is selected as the initial estimated concentration. The initial estimated concentration is interpolated and corrected based on the difference in competition scores between the initial estimated concentration and the concentrations of two adjacent standard pollutants to obtain the final output initial concentration of organic pollutants.

2. The method for quantitative detection of organic pollutants in water by carboxyl-functionalized magnetic COF as described in claim 1, characterized in that, The constructed carboxyl-functionalized magnetic COF composite material, used as a solid-phase extraction adsorbent, specifically includes: Magnetic nanoparticles of iron oxide are uniformly dispersed in an organic solvent to form a magnetic substrate dispersion. A bifunctional ligand containing aldehyde and carboxyl groups and a melamine monomer were added to the magnetic substrate dispersion, and a carboxyl-functionalized covalent organic framework shell was grown in situ on the surface of the magnetite magnetic nanoparticles by a solvothermal reaction. After the reaction was completed, the precipitate was collected by an external magnetic field. The precipitate was washed with organic solvent and deionized water in sequence, and then dried under vacuum to obtain the carboxyl-functionalized magnetic COF composite material.

3. The method for quantitative detection of organic pollutants in water by carboxyl-functionalized magnetic COF as described in claim 1, characterized in that, The process of adding the solid-phase extraction adsorbent to a water sample for oscillatory adsorption and separating the loaded adsorbent after organic pollutants are added using an external magnetic field specifically includes: The solid-phase extraction adsorbent is added to the water sample according to the preset ratio of adsorbent mass to water sample volume, and the sample is oscillated in a constant temperature shaker at a set speed for a first predetermined time, so that the organic pollutants in the water sample are captured by the solid-phase extraction adsorbent. After the oscillation is completed, the container holding the water sample is placed above the permanent magnet and left to stand for a second predetermined time, so that the solid phase extraction adsorbent gathers at the bottom of the container near the permanent magnet under the action of the magnetic field. After removing the supernatant, the loaded adsorbent containing organic pollutants is obtained.

4. The method for quantitative detection of organic pollutants in water by carboxyl-functionalized magnetic COF as described in claim 1, characterized in that, The step of eluting the supported adsorbent with an eluent and collecting the eluent specifically includes: A predetermined volume of organic eluent is added to a container containing the loaded adsorbent, and the container is oscillated for a third predetermined time using a vortex oscillator to desorb organic pollutants from the loaded adsorbent into the eluent. The container was placed above the external magnetic field again and left to stand for a fourth predetermined time. After the eluted magnetic adsorbent had completely settled, the supernatant was transferred to the sample bottle as the eluent. After repeating the elution step twice, the eluents obtained each time were combined and concentrated to a fixed volume using nitrogen purging for subsequent detection.

5. The method for quantitative detection of organic pollutants in water by carboxyl-functionalized magnetic COF according to claim 4, characterized in that, The organic eluent is at least one of acetonitrile, methanol, or a mixture of acetonitrile and ethyl acetate.

6. The method for quantitative detection of organic pollutants in water by carboxyl-functionalized magnetic COF as described in claim 5, characterized in that, The extraction of the characteristic peak shift and peak intensity attenuation ratio at the adsorption equilibrium point on the dynamic response surface specifically includes: Locate the reference wavelength position of the standard characteristic peak of organic pollutants on the dynamic response surface; Track the real-time wavelength position of the actual peak in the spectral curve corresponding to each time slice on the dynamic response surface, and calculate the difference between the real-time wavelength position and the reference wavelength position as the characteristic peak offset; Extract the real-time absorbance peak value corresponding to the real-time wavelength position, and calculate the ratio between the real-time absorbance peak value and the standard absorbance peak value of the standard characteristic peak of the organic pollutant as the peak intensity attenuation ratio; The characteristic peak shift and peak intensity decay ratio corresponding to the adsorption equilibrium time point are selected as the output values.

7. The method for quantitative detection of organic pollutants in water by carboxyl-functionalized magnetic COF as described in claim 6, characterized in that, The adsorption equilibrium time point is automatically determined when the rate of change of the characteristic peak offset on the dynamic response surface over time is less than a preset threshold.

8. The method for quantitative detection of organic pollutants in water by carboxyl-functionalized magnetic COF as described in claim 1, characterized in that, The competitive substitution quantitative model is a nonlinear mapping model based on a radial basis function neural network, with the number of hidden layer nodes set to 1.5-2 times the number of standard pollutant concentration gradients.

Images