Carbon dot synthesis method based on machine learning guidance and detection application thereof

Abstract

The invention discloses preparation of high-quantum-yield carbon dots based on machine learning guidance and detection application of a hydrogel kit of the high-quantum-yield carbon dots, and relates to the technical field of fluorescent nano material preparation and sensors. The method comprises the following steps: inputting reaction parameter data of synthesis of various carbon dots into a machine learning model, obtaining an optimal influence factor according to a prediction result, and realizing synthesis of the high-quantum-yield carbon dots. The carbon dot is in a sphere-like shape, the average particle size is 2.12 nm, the carbon dot has high fluorescence intensity, and an adopted precursor material is waste paper which is wide in source and low in cost. The synthesized high-quantum-yield carbon dots have the advantages of being rapid and sensitive in detection and the like, and mercury ions in Chinese herbal medicines can be rapidly detected according to fluorescence signal quenching. Meanwhile, the preparation of the carbon dot hydrogel kit can further realize colorimetric sensing, and the carbon dot hydrogel kit has good application value and prospect in the fields of environment, medicine, food and the like.

Description

Technical Field

[0001] The present invention belongs to the technical field of fluorescent nanomaterial preparation and sensor, and specifically relates to a preparation method of high quantum yield carbon dots guided by machine learning and the detection application of a hydrogel kit thereof. Background Art

[0002] Chinese herbal medicines are increasingly used in the prevention and treatment of diseases. However, problems such as excessive heavy metals in the cultivation and processing of Chinese herbal medicines not only affect the quality of Chinese herbal medicines, but also limit the development of their applications. For example, mercury is toxic to the kidneys, brain, endocrine system and central nervous system of humans and animals through enzyme inhibition and induction of oxidative stress. It is considered to be one of the most harmful heavy metal pollutants. Therefore, mercury ions (Hg 2+ ) has become one of the mandatory inspection items for quality control of traditional Chinese medicines. Traditional methods such as atomic absorption spectroscopy, ion chromatography, and electrochemical methods usually require large instruments, professionals, high detection costs, and time-consuming procedures, which greatly limits their portability and applicability for on-site monitoring. As an innovative alternative, fluorescent sensors and hydrogel kits are receiving increasing attention in real-time, on-site monitoring.

[0003] Carbon dots (CDs) are highly regarded for their eco-friendliness, cost-effectiveness, excellent optical stability, and multifunctional applications. In the field of analytical chemistry, carbon dots are particularly effective in detecting heavy metal ions due to their high sensitivity and selectivity. In recent studies, a large number of CDs for the detection of Hg 2+ Fluorescent carbon dots. Despite the great progress made in this field, the synthesis of carbon dots still faces great challenges. The selection of precursors, solvents, and reaction conditions often varies greatly. Many experimental procedures rely heavily on empirical judgments obtained through extensive trial and error methods, resulting in poor reproducibility. At present, the most common method to improve the performance of carbon dots is to optimize the synthesis parameters. However, this method is time-consuming and costly, and it does not guarantee the acquisition of carbon dots with ideal luminescence properties (such as high quantum yield and long Stokes shift). In recent years, artificial intelligence has played an increasingly important role in fields such as materials chemistry. Machine learning (ML), as one of the core technologies of artificial intelligence, can use the autonomous learning ability of computers or machines to deeply explore the regular information hidden in the data and continuously improve the performance of a certain aspect. Therefore, machine learning can be used as a reliable tool to mine hidden information and extract effective information from massive data. Therefore, it is crucial to develop efficient and controllable methods for the synthesis of carbon dots. Standardized and reproducible synthesis protocols can not only achieve large-scale production of carbon dots, but also pave the way for advanced applications of carbon dots in analytical detection, biomedical fields, and other related fields. Summary of the invention

[0004] The purpose of the present invention is to provide a synthesis of high quantum yield carbon dots (P-CDs) guided by machine learning and its hydrogel kit detection application. The carbon quantum dots have a high quantum yield and the source of raw material waste paper is wide and cheap. The synthesized fluorescent carbon quantum dots have the advantages of rapid detection and sensitivity, and can be used as a fluorescent sensor to detect Hg in Chinese herbal medicine. 2+ Another purpose is to provide a method for constructing a carbon dot-based hydrogel kit to broaden its applications.

[0005] In order to achieve the above object, the present invention adopts the following technical solutions:

[0006] The synthesis of high quantum yield carbon dots guided by machine learning provided by the present invention realizes the detection of Hg in Chinese herbal medicine. 2+ The carbon dots were then fabricated into a portable hydrogel kit for colorimetric detection.

[0007] The first aspect of the present invention provides a method for synthesizing high quantum yield carbon dots based on machine learning guidance, comprising the following steps:

[0008] (1) Using machine learning models to predict the most important parameters affecting the synthesis method:

[0009] The application of carbon dots in Hg 2+ 102 English literatures on carbon dot synthesis methods in fluorescence detection were collected, and the literature on carbon dot synthesis methods was extracted to extract the synthesis parameters of reaction precursors, reaction solvents, reaction temperature, and reaction time and the corresponding quantum yield data; with reaction precursors, reaction solvents, reaction temperature, and reaction time as input values and the corresponding quantum yield as output values, different machine learning models were trained, and the machine learning model with the most accurate prediction was selected based on error measurement indicators and trend measurement indicators; the selected machine learning model was used to analyze the importance of the above synthesis parameters on high quantum yield, and the synthesis parameters with the most important impact on quantum yield were selected;

[0010] (2) using waste paper as a reaction precursor, extracting cellulose, and removing lignin and hemicellulose; using the extracted cellulose as a raw material and water as a reaction solvent, reacting a mixed solution of the two at a certain heating reaction temperature and heating reaction time;

[0011] (3) According to the method of step (2), the most important influencing parameter obtained in step (1) is reacted at different values, and the one with the highest quantum yield is used as the optimal synthesis parameter of step (2);

[0012] (4) completing the reaction of step (2) according to the optimal synthesis parameters to obtain a reaction product.

[0013] In the above technical solution, further, the trained machine learning models include Decision Tree (DT), Random Forest (RF), Adaptive Boosting (AdaBoost), and Gradient Boosting Decision Tree (GBDT); the measurement metrics are Mean Square Error (MSE), Mean Absolute Error (MAE), and R-Square (R 2 ); the smaller the values of MSE and MAE, the closer the value of R 2 is to 1.

[0014] In the above technical solution, further, in step (1), the synthesis parameter that has the most important influence on the quantum yield is the reaction temperature, and the assigned range of the reaction temperature is 150 - 200 °C.

[0015] In the above technical solution, further, in step (2), the method for extracting cellulose from waste paper is as follows: waste paper scraps are added to a mixed solution of nitric acid, acetic acid, and distilled water, and after stirring, the mixture is heated and reacted at 100 - 150 °C.

[0016] In the above technical solution, further, the mass of the waste paper scraps is 2 - 3 g; 30 - 50 mL of a mixed solution of HNO3, CH3COOH, and distilled water (80 / 10 / 10, v / v / v) is added, and magnetic stirring is carried out for 10 - 15 min to fully mix. The fully mixed mixture is transferred to a reaction kettle and heated at 100 - 150 °C for 30 min - 40 min to extract cellulose.

[0017] In the above technical solution, further, in step (4), the cellulose addition amount is 0.2 - 0.5 g, and the water is 15 - 20 mL; the autoclave is a 100 mL stainless - steel autoclave with a polytetrafluoroethylene lining; the heating temperature is 150 - 200 °C, and the heating time is 2 - 5 h.

[0018] In the above technical solution, further, the reaction product obtained in step (4) is filtered through a filter membrane and dialyzed; the filter membrane is a 0.22 μm microporous filter membrane, and the dialysis parameter is 300 - 800 Da.

[0019] The second aspect of the present invention provides carbon dots prepared by any of the foregoing methods.

[0020] The third aspect of the present invention provides a carbon dot hydrogel, which is prepared by mixing the foregoing carbon dots with an agarose solution and cooling and solidifying.

[0021] The fourth aspect of the present invention provides the aforementioned carbon dots or the aforementioned carbon dot hydrogel for detecting Hg 2+ applications.

[0022] In the above technical solution, further, the aforementioned carbon dots or the aforementioned carbon dot hydrogel are used for detecting Hg in Chinese herbal medicines 2+ content applications.

[0023] The detection method is as follows: Prepare standard solutions of Hg with different concentrations in sequence, then add 400 μL of the carbon dot stock solution to 1600 μL of the Hg 2+ standard solution and incubate for 15 min, and measure the fluorescence intensity at λ 2+ = 360 nm and λ ex = 453 nm. All detections are carried out in parallel 3 times. In the actual sample detection, Panax notoginseng, Paeonia lactiflora, and Ligusticum chuanxiong are selected as representatives of Chinese herbal medicines. Taking Panax notoginseng as an example, add 5 - 10 ml of concentrated nitric acid to 0.1 - 0.5 g of Panax notoginseng powder, heat at 150 - 200 °C for 50 - 100 min, then cool to room temperature at the end of the reaction, and adjust the pH value of the filtrate to neutral. Transfer the above solution to a volumetric flask and prepare Hg em standard solutions with low, medium, and high concentrations (0.1, 50, and 80 μM) for the detection of actual samples. 2+ standard solution for the detection of actual samples.

[0024] In the above detection method, the parameter settings of the fluorescence spectrophotometer used are: scanning speed (1000 nm / min), excitation bandwidth (10 nm), emission bandwidth (10 nm), gain (medium, 650 V).

[0025] The carbon dot solution shows blue fluorescence under the excitation of an ultraviolet lamp; when Hg 2+ is mixed with the carbon dot solution, since Hg 2+ can bind to the amino and carboxyl groups on the surface of the carbon dots, resulting in a decrease in the fluorescence response of the carbon dots. Within a certain concentration range, the fluorescence intensity is linearly correlated with the concentration of Hg 2+ . And the carbon dots have good selectivity and sensitivity to Hg 2+ , and other metal ions have little interference on this detection system. Therefore, the present invention can effectively detect Hg in Chinese herbal medicines through carbon dots 2+ .

[0026] The fifth aspect of the present invention provides an application of a portable hydrogel kit for detecting Hg 2+ , and the hydrogel kit is based on the aforementioned carbon dots or the aforementioned carbon dot hydrogel.

[0027] In the above technical solution, further, the preparation method of the hydrogel kit includes:

[0028] Agarose is completely dissolved in water, and then the high quantum yield carbon dot stock solution is added to the agarose solution and quickly mixed evenly, and then the mixed solution is quickly transferred to a 12-well plate, and the mixture is cooled and solidified to obtain a hydrogel kit; the mass of the agarose is 0.1-0.4 g; 15-30 mL of deionized water is added, and magnetic stirring is performed for 10-15 minutes to fully mix.

[0029] In the above technical solution, further, the hydrogel kit detects Hg 2+ for:

[0030] The different concentrations of Hg 2+ The solution was dripped into the hydrogel kit, and the color image of the hydrogel was captured by a smartphone after the reaction. The color intensity RGB value was analyzed using Image J software to achieve Hg 2+ Colorimetric quantitative analysis of Hg in hydrogel kit 2+ The detection conditions are 20-30°C and the incubation time is 30-50 minutes.

[0031] Compared with the prior art, the present invention has the following beneficial effects:

[0032] 1) Use machine learning technology to build multiple models, extract and analyze key synthesis parameters from a large amount of data to optimize the preparation process of carbon dots with high quantum yield. Improve the synthesis efficiency and repeatability, and overcome the limitations of traditional experiments that rely on experience.

[0033] 2) The innovative use of cellulose extracted from waste paper as the biomass precursor material for carbon dots achieves high-value recycling of waste resources, while reducing production costs and providing support for environmental protection.

[0034] 3) The prepared high quantum yield carbon dots achieve highly selective and sensitive fluorescence detection with a wide linear range and low detection limit, and exhibit excellent anti-interference ability.

[0035] 4) A portable carbon dot hydrogel kit based on smartphone imaging analysis was developed. This tool can detect Hg 2+ It can be used for rapid detection, with good operability and practical application prospects. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 A general flow chart for the preparation and detection of high quantum yield carbon dots guided by machine learning;

[0037] Figure 2 A is a comparison chart of mean square error (MSE) of machine learning models; Figure 2 B is a comparison chart of the mean absolute error (MAE) of machine learning models; Figure 2C is a comparison graph of the coefficient of determination (R 2 ) of the machine learning model; Figure 2 D is a graph of the feature importance analysis of the decision tree model;

[0038] Figure 3 A is the transmission electron microscopy (TEM) spectrum of P-CDs; Figure 3 B is the particle size distribution graph of P-CDs;

[0039] Figure 4 is the X-ray diffraction analysis (XRD) spectrogram of P-CDs;

[0040] Figure 5 is the Fourier transform infrared (FT-IR) spectrogram of P-CDs;

[0041] Figure 6 is the fluorescence spectrogram of P-CDs at different excitation wavelengths;

[0042] Figure 7 is the fluorescence spectrogram of P-CDs at different dilution multiples;

[0043] Figure 8 is the fluorescence emission spectrogram of P-CDs at different Hg 2+ concentrations;

[0044] Figure 9 is the relationship graph between different Hg 2+ concentrations and the fluorescence intensity of the P-CDs solution; F0 is the fluorescence intensity of P-CDs without adding Hg 2+ and F is the fluorescence intensity of P-CDs after adding Hg 2+ ;

[0045] Figure 10 is the comparison graph of the influence of other metal ions on the fluorescence intensity of P-CDs;

[0046] Figure 11 is the comparison graph of the influence of other metal ions on the fluorescence intensity of the P-CDs / Hg 2+ system;

[0047] Figure 12 is the fluorescence lifetime graph of P-CDs and the P-CDs / Hg 2+ system;

[0048] Figure 13 is the ultraviolet-visible spectrogram of Hg 2+ , P-CDs and the P-CDs / Hg 2+ system;

[0049] Figure 14 is the Zeta potential graph of the P-CDs and the P-CDs / Hg 2+ system;

[0050] Figure 15 Schematic diagram of the preparation process and detection application of the carbon dot hydrogel kit;

[0051] Figure 16 For different Hg 2+ Color comparison diagram of the hydrogel kit at different concentrations;

[0052] Figure 17 For different Hg 2+ Relationship diagram between concentration and RGB value of the color intensity of the hydrogel kit;

[0053] Figure 18 Investigation diagram of the cytotoxicity of the P-CDs solution. Detailed implementation manners

[0054] The technical solutions of the present invention will be further described in detail below in conjunction with the accompanying drawings and specific implementation manners, but it does not constitute any limitation to the present invention. Any limited modifications made within the protection scope of the claims of the present invention are still within the protection scope of the claims of the present invention.

[0055] Instrument software parameters and reagent details:

[0056] Machine learning calculation software and versions (Python 3.8, Scikit-learn 0.24.1, Pandas 1.2.3, NumPy 1.19.5, Matplotlib 3.3.4). F97 fluorescence spectrophotometer (Shanghai Lingguang Technology Co., Ltd., China), UV-5500PC ultraviolet-visible spectrophotometer (Shanghai Yuanxi Instruments Co., Ltd.), KQ-5200 ultrasonic cleaner (Jiangsu Kunshan Ultrasonic Instruments Co., Ltd., China), Thermo Fisher Nicolet iS20 Fourier transform infrared spectrometer (Thermo Fisher Company, USA), LC-400 low-speed centrifuge (Anhui Keda Innovation Co., Ltd.), DF-101S collector type constant temperature heating magnetic stirrer (Gongyi Yuhua Instruments Co., Ltd.), JEM-2100 transmission electron microscope (JEOL Company, Japan), X-ray diffraction (XRD) was measured by using a D8 Advance diffractometer (Bruker, Germany). Waste paper scraps are from waste printing paper (Shanghai Chenguang Stationery Co., Ltd.), mercury chloride (Shenyang Yuwang Chemical Glass Instruments Co., Ltd.), deionized water (Wahaha Group Co., Ltd.), sodium hydroxide and absolute ethanol (Tianjin Hengxing Chemical Reagent Manufacturing Co., Ltd.), agarose (Shanghai Macklin Reagents Co., Ltd.).

[0057] Example 1

[0058] Preparation and detection application of a machine learning-guided carbon dots and its hydrogel kit. The overall process is as follows Figure 1 shown. Cellulose is extracted from waste paper as a reaction precursor. Based on machine learning guidance, the most important influencing parameters of the synthesis method are screened, and finally the carbon dots P-CDs are prepared. The carbon dots are mixed with agarose to prepare a hydrogel, forming a detection kit. The carbon dots or the kit are used to detect Hg 2+ .

[0059] The preparation method of machine learning-guided carbon dots with high quantum yield is as follows:

[0060] (1) Collection of carbon dot synthesis parameters and construction of machine learning model

[0061] Collected 102 English literature on the synthesis methods of carbon dots in Hg 2+ fluorescence detection in the past decade and extracted the corresponding synthesis reaction parameters. The key parameters for the synthesis of high quantum yield carbon dots were analyzed using a machine learning model. The precursor, solvent, reaction time, and reaction temperature were used as input values, and the corresponding quantum yield was used as the output value. Using these data, we trained multiple machine learning models to determine which parameters are the most important for synthesizing high quantum yields. The trained machine learning models include Decision Tree (DT), Random Forest (RF), Adaptive Boosting (AdaBoost), and Gradient Boosting Decision Tree (GBDT). When selecting a model, several key indicators are usually considered, mainly including the interpretability and prediction accuracy of the model. The mean squared error (MSE) and mean absolute error (MAE) measure the average difference between the predicted value and the actual value of the model. The smaller the value, the higher the prediction accuracy of the model. Among them, the value of DT is the smallest, indicating that its prediction accuracy is the highest. At the same time, the coefficient of determination (R 2 ) measures the ability of the model to explain the data. The closer its value is to 1, the better the ability of the model to explain the data. The R 2 of DT is the highest, indicating that it has the best ability to explain the data. In summary, the DT model was selected to evaluate the importance of each feature on the high quantum yield, with the reaction precursor, reaction solvent, reaction temperature, and reaction time as input values and the corresponding quantum yield as the output value. The results show that the reaction temperature is the most important factor affecting the synthesis of high quantum yield carbon dots for Hg 2+ detection( Figure 2 ).

[0062] (2) Synthesis of waste paper-derived high quantum yield carbon dots (P-CDs)

[0063] Since the reaction temperature is the most important factor for the synthesis of carbon dots with high quantum yield, and in order to achieve the detection of Hg 2+ for the synthesis of carbon dots, we selected cellulose extracted from waste paper as the reaction precursor, water as the reaction solvent, the reaction time was 4 h, and the reaction temperatures were selected as 160 °C, 180 °C and 200 °C respectively. The comparison after synthesis showed (Table 1) that the quantum yield was the highest at the reaction temperature of 180 °C. In summary, all the reaction synthesis parameters of P-CDs were determined.

[0064] Table 1 Quantum yields of P-CDs at different reaction temperatures

[0065]

[0066] The specific method is as follows: Add 2.0 g of waste paper scraps to a beaker, and at the same time add 40 mL of a mixture of nitric acid (65%), acetic acid (99.5%) and distilled water (80 / 10 / 10, v / v / v), and stir magnetically for 10 min for mixing. Transfer the well-mixed mixture to a reaction kettle and heat it at 120 °C for 30 min to remove hemicellulose and lignin and retain cellulose. Centrifuge and wash the solid cellulose, and then dry it in an oven at 60 °C. Add 0.4 g of cellulose and 20 mL of water to the reaction kettle and carry out the synthesis at the reaction temperature of 180 °C. After natural cooling, centrifuge the reaction product and extract the supernatant. The supernatant is filtered through a 0.22 μm microporous filter to obtain a cellulose carbon dot solution with high quantum yield.

[0067] Characterization and testing:

[0068] ① Morphology characterization of the prepared P-CDs

[0069] Characterize the size and particle size distribution of the material by transmission electron microscopy (TEM). P-CDs have a spherical nanoparticle shape, with uniform size and high dispersion. The particle size of P-CDs is mainly distributed between 0.5 nm and 5 nm, and the average diameter is 2.12 nm ( Figure 3 ).

[0070] ② Optical property characterization and testing of P-CDs

[0071] Characterize and test P-CDs by Fourier transform infrared spectroscopy (FTIR), X-ray diffraction analysis (XRD), fluorescence optimal excitation wavelength and optimal dilution multiple experiments.

[0072] The X-ray diffraction characterization pattern of P-CDs shows a strong diffraction peak at 23.5°, corresponding to the (002) crystal plane of graphite ( Figure 4 ). The functional groups on the surface of P-CDs were evaluated by Fourier transform infrared spectroscopy (FTIR). AsFigure 5 As shown, the absorption peak at 3210 cm -1 is the stretching vibration of O-H or N-H, corresponding to the hydroxyl or amino groups on the surface of P-CDs. The absorption peak at 1721 cm -1 corresponds to the presence of C=O. The absorption peaks at 1383 cm -1 , 1204 cm -1 , 1166 cm -1 and 1090 cm -1 belong to the stretching vibration of C-O, corresponding to the hydroxyl or carboxyl functional groups. The absorption peak at 803 cm -1 is usually attributed to the stretching vibration of C-N, corresponding to the amine or amide groups on the surface of carbon dots. Through the above characterization results, the successful synthesis of P-CDs was effectively confirmed.

[0073] In addition, the optimal excitation wavelength and optimal dilution ratio of the carbon dots were further investigated. As Figure 6 shown, the intensity and peak position of fluorescence emission are determined by the excitation wavelength, showing the characteristics of excitation light dependence. The results show that the emission wavelength increases with the increase of the excitation wavelength. When the excitation wavelength changes from 300 nm to 460 nm, the emission maximum gradually shifts to a longer wavelength, showing a red shift, and the intensity peak decreases. The results show that the optimal excitation wavelength of P-CDs is 360 nm, and the maximum emission wavelength at the optimal excitation wavelength is 453 nm. 360 nm was selected as the excitation wavelength for subsequent experiments. At the same time, as Figure 7 shown, the optimal dilution of the synthesized P-CDs stock solution was carried out, and the results showed that the fluorescence intensity of P-CDs was the highest when diluted 5 times.

[0074] Example 2

[0075] The prepared high quantum yield carbon dots P-CDs were used for the detection of Hg 2+ , and the specific operation is as follows:

[0076] (1) Detect the fluorescence intensity of P-CDs at different Hg 2+ concentrations

[0077] In order to better realize the fluorescence detection of Hg 2+ , the experimental conditions were optimized. The optimal experimental conditions were pH 7, incubation time 15 min, and incubation temperature 25 °C. Under the optimal experimental conditions, the ability of P-CDs to detect Hg 2+ was further evaluated. 0.4 mL of the P-CDs stock solution was mixed with 1.6 mL of Hg 2+ standard solutions with different concentrations to obtain the P-CDs / Hg 2+ system. At the fluorescence excitation wavelength of 360 nm, the influence of Hg 2+ solutions with different concentrations on the fluorescence intensity of P-CDs was studied. AsFigure 8 As shown, the concentration of Hg 2+ varies between 0.05 and 100 μM, and the fluorescence intensity value of P-CDs decreases correspondingly with the increase in the concentration of Hg 2+ . In the range of 0.05 - 100 μM, the concentration of Hg 2+ has a linear relationship with the quenching degree (F0 - F / F0), and the linear equation is y = 0.0061x + 0.0346, with the correlation coefficient r 2 = 0.9956( Figure 9 ).

[0078] (2) Selectivity and anti-interference experiments of P-CDs for metal ions

[0079] The selectivity and anti-interference of the fluorescence probe based on P-CDs for detecting Hg 2+ are key issues. Only Hg 2+ results in a significant decrease in the fluorescence emission of P-CDs, while different metal ions, such as Na + , K + , Mg 2+ , Zn 2+ , Ba 2+ , Co 2+ , Cu 2+ , Pb 2+ , Fe 2+ , Mn 2+ , Al 3+ , Fe 3+ , Cr 3+ , even when their concentration is 5 times that of Hg 2+ , the fluorescence emission shows little change( Figure 10 ). This excellent selectivity can be attributed to the presence of carboxyl and hydroxyl groups on the surface of P-CDs, which leads to the interaction between the ligand and Hg 2+ . Similarly, through anti-interference experiments, it is found that the presence of other metal ions has little effect on the detection of Hg 2+ , showing good anti-interference performance( Figure 11 ).

[0080] (3) Mechanism analysis of P-CDs fluorescence detection of Hg 2+ Mechanism dissection

[0081] The mechanism for the sensitive detection of Hg 2+ by P-CDs was explored using fluorescence lifetime testing, UV-visible spectroscopy analysis, and Zeta potential testing. The fluorescence lifetime of P-CDs is as Figure 12 shown. After adding Hg 2+ , the fluorescence lifetime of P-CDs and after adding Hg 2+The fluorescence lifetime before is not significantly changed. In the case of dynamic quenching, the fluorescence lifetime of carbon dots changes significantly with the presence of added ions, so the mechanism of dynamic quenching may be excluded. Hg 2+ 、P-CDs and P-CDs / Hg 2+ The UV-visible absorption spectra of Figure 13 are shown as follows. It can be found that with the addition of Hg 2+ , the absorption peaks in the spectra decrease significantly. Therefore, it is preliminarily determined that the interaction between P-CDs and Hg 2+ results in the formation of a complex, leading to fluorescence quenching, which is a static quenching phenomenon. Figure 14 The Zeta potential of P-CDs / Hg 2+ in 2+ is 27.3 mV, which is significantly increased compared with P-CDs. It may be due to the strong affinity between functional groups such as carboxyl and hydroxyl groups and Hg 2+ , and it can be further shown that the coordination between Hg

[0082] (4) Fluorescence detection of Hg in traditional Chinese herbal medicines by P-CDs 2+

[0083] Taking traditional Chinese medicine Panax notoginseng as an actual sample, the feasibility of the established method was investigated. The accuracy and precision of the analytical method were evaluated by measuring the recovery rate and RSD value of the spiked serum samples. Table 2 shows that the detection recovery rate of Hg 2+ in the actual traditional Chinese herbal medicine samples is between 95.3% and 100.6%, and the RSD is not more than 3.2%, indicating that the prepared P-CDs can be used to detect Hg 2+ in actual traditional Chinese herbal medicine samples.

[0084] Table 2 Fluorescence detection of Hg 2+ in Panax notoginseng samples (n = 3)

[0085]

[0086] Example 3

[0087] Preparation and detection application of a hydrogel kit based on carbon dots P-CDs:

[0088] The high quantum yield carbon dots P-CDs prepared above were used to prepare a hydrogel kit and realize colorimetric detection application. The specific operation is as follows:

[0089] First, 0.2 g of agarose was completely dissolved in 20 mL of water at 60 °C. Then, 5 mL of the P-CDs stock solution was added to the above agarose solution, and the mixture was quickly mixed evenly. Then, the mixed solution was quickly transferred to a 12-well plate and cooled to solidify to obtain a P-CDs hydrogel kit.

[0090] Add 1 mL of Hg standard solutions with different concentrations (0.5 - 120 μM) into the hydrogel kit, and react at 25 °C for 40 min. After the reaction, capture the color images of the hydrogel with a smartphone, analyze the RGB values of the color intensity using Image J software, and perform colorimetric quantitative analysis on Hg 2+ 2+ Figure 15 ) Figure 16 It can be seen that the Hg concentration in the range of 0.5 - 120 μM has a good linear relationship with the RGB color intensity, and the RGB value = -0.1457C 2+ + 130.8 (r Hg2+ = 0.998). At the same time, it can be seen that as the Hg 2 concentration increases (0.5 μM - 120 μM), the blue color presented by the hydrogel kit gradually changes from bright blue to dark blue 2+ ) Figure 16 , achieving semi - quantitative detection.

[0091] Example 4

[0092] Detect the cytotoxicity of the prepared P - CDs solution:

[0093] To evaluate the cytotoxicity of the prepared fluorescent carbon dots of P - CDs, a cytotoxicity experiment was carried out. The CCK - 8 (Cell Counting Kit - 8) method was used to detect the biocompatibility of P - CDs. The survival rates of BRL - 3A (rat hepatocytes) and HT - 22 (mouse hippocampal neuron cells) in the presence of P - CDs were measured.

[0094] Experimental procedure: Select mouse hippocampal neuron cells (HT - 22) and rat normal hepatocytes (BRL - 3A) as model cells, inoculate them in a 96 - well plate at a density of 5×10 3 cells / mL, and culture them in an incubator with 5% carbon dioxide for 24 hours. Aspirate the culture medium in each well, add 100 μL of culture medium with different concentrations of P - CDs (0, 12.5, 25, 50, 100 μg / mL) into each well, and intervene for 24 h. Each concentration is repeated 6 groups. Subsequently, add 100 μL of 10% (v / v) CCK - 8 reagent into each well and incubate for 24 h. Measure the optical density (OD) at 450 nm with an enzyme - labeler. The well with the addition of an equal volume of PBS serves as the negative control well. The well with cells added but without material stimulation serves as the positive control well. The calculation method of cell viability is: Cell viability = (OD value of the material - stimulated well - OD value of the negative control well) / (OD value of the positive control well - OD value of the negative control well) × 100%. Observe the cell morphology with a fluorescence inverted microscope. Prepare 3 replicate samples for measurement during the experiment.​​

[0095] Experimental results: As Figure 18 shown, based on the above experimental method, the effects of P-CDs solutions at different concentrations (0 - 100 μg / mL) on cell viability were observed. After 24 hours of incubation, even under the action of the highest concentration of P-CDs (100 μg / ml), the cell viability was still higher than 95%. It was proved by inspection that there was no significant difference between the P-CDs group and the control group. This is sufficient to show that the P-CDs prepared in this study have low toxicity and good biocompatibility.

[0096] The above are only the preferred embodiments of the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A machine learning-guided method for synthesizing carbon dots, characterized in that, The method includes the following steps: (1) Using a machine learning model to predict the most important influencing parameters of the synthesis method: Collecting literature on carbon dot synthesis methods, extracting synthesis parameters such as reaction precursors, reaction solvents, reaction temperatures, and reaction times, and corresponding quantum yield data; using reaction precursors, reaction solvents, reaction temperatures, and reaction times as input values, and the corresponding quantum yields as output values, training different machine learning models, and screening out the machine learning model with the most accurate prediction according to error measurement indicators and trend measurement indicators; applying the screened machine learning model to analyze the importance of the above synthesis parameters on high quantum yield, and selecting the synthesis parameter that has the most important influence on the quantum yield; (2) Using waste paper as a reaction precursor, extracting cellulose, and removing lignin and hemicellulose; using the extracted cellulose as a raw material and water as a reaction solvent, reacting the mixed solution of the two at a certain heating reaction temperature and heating reaction time; (3) According to the method in step (2), reacting the most important influencing parameters obtained in step (1) at different values, and the one with the highest quantum yield is used as the optimal synthesis parameter for step (2); (4) Completing the reaction in step (2) according to the optimal synthesis parameters to obtain a reaction product.

2. The method according to claim 1, characterized in that, The trained machine learning models include Decision Tree (DT), Random Forest (RF), Adaptive Boosting (AdaBoost), and Gradient Boosting Decision Tree (GBDT); The measurement indicators are the mean squared error (MSE), mean absolute error (MAE), and coefficient of determination (R 2 )); the smaller the values of the MSE and MAE, the closer the value of R 2 is to 1.

3. The method according to claim 1, characterized in that, In step (1), the synthesis parameter that has the most important influence on the quantum yield is the reaction temperature, and the value range of the reaction temperature assignment is 150-200 °C.

4. The method according to claim 1, wherein In step (2), the method for extracting cellulose from waste paper is: adding waste paper scraps to a mixed solution of nitric acid, acetic acid, and distilled water, and heating and reacting the mixture at 100-150 °C after stirring.

5. The method according to claim 4, characterized in that The ratio of the waste paper scraps to the mixed solution is 2-3 g: 30-50 mL; the volume ratio of nitric acid, acetic acid, and distilled water in the mixed solution is 8:1:

1.

6. The method according to claim 1, wherein The reaction product obtained in step (4) is filtered through a filter membrane and dialyzed; the filter membrane is a 0.22 μm microporous filter membrane, and the dialysis parameter is 300-800 Da.

7. Carbon dots prepared by the method according to any one of claims 1-6.

8. A carbon dot hydrogel, characterized in that, Mixing the carbon dots according to claim 7 with an agarose solution and cooling and solidifying.

9. Use of the carbon dots according to claim 7 or the carbon dot hydrogel according to claim 8 in detecting Hg 2+ ; preferably, use in detecting Hg 2+ content in Chinese herbal medicines.

10. Application of a hydrogel kit in detecting Hg 2+ Preferably, the application is as follows: dropping standard solutions of Hg 2+ with different concentrations into the hydrogel kit, reacting at 25 °C for 40 min, after the reaction ends, capturing the color image of the hydrogel with a smart phone, analyzing the RGB values of the color intensity with Image J software, and performing colorimetric quantitative analysis on Hg 2+ .

Images