Carbon quantum dot, preparation method, detection method and application of XGBoost in identification of Tongxikang components

Abstract

The invention relates to a carbon quantum dot, a preparation method, a detection method, and an application of XGBoost in identification of Tongxingkang components. The quantum dots provided by the invention can be subjected to fluorescence quenching through the Tongxikang class, fluorescence intensities corresponding to two excitation wavelengths of the quantum dots are reduced in an unequal proportion, and the type and concentration of the Tongxikang class can be determined to a certain extent through the ratio of the fluorescence intensities corresponding to the two excitation wavelengths. Based on the quantum dots provided by the invention, the XGBoost can effectively identify the specific types of the Tongxi Kang type. On the basis, based on the quantum dots provided by the invention, the CNN can further accurately identify the concentration of the Tongzhikang class. On the basis of cooperation of the two, low-cost and high-precision detection of the Tonxiakang class is realized.

Description

Technical Field

[0001] The present invention relates to the field of detection of piroxicam-like substances, and particularly to a carbon quantum dot, a preparation method, a detection method, and the application of XGBoost in the identification of piroxicam-like components. Background Art

[0002] As a non-steroidal anti-inflammatory drug, piroxicam-like substances can inhibit the activity of cyclooxygenase and play an anti-inflammatory and analgesic role, so they are widely used in inflammation and postoperative management. However, the long-term use of piroxicam-like substances will also cause damage to the hematopoietic system, leading to adverse reactions such as aplastic anemia and coagulation disorders. To prevent the abuse of piroxicam-like substances, the detection of piroxicam-like substances is of great significance. However, the existing detection methods for piroxicam-like substances have problems such as low detection accuracy, difficult quantification, poor accuracy, and high cost. Summary of the Invention

[0003] Based on this, in view of the problems of poor detection accuracy and high cost of piroxicam-like substance detection, it is necessary to provide a carbon quantum dot, a preparation method, a detection method, and the application of XGBoost in the identification of piroxicam-like components.

[0004] A carbon quantum dot, whose chemical formula is: .

[0005] A preparation method of a carbon quantum dot, which comprises dissolving fleroxacin in ethanol and reacting in a reaction kettle at 180°C - 220°C to obtain the carbon quantum dot.

[0006] An application of the carbon quantum dot in detecting piroxicam-like substances.

[0007] A detection method for piroxicam-like substances, comprising: Before the carbon quantum dot is mixed with piroxicam-like substances, irradiate the carbon quantum dot with excitation lights of 340 nm and 280 nm, and respectively obtain the light intensities f 340 and f 280 of the 440-nm fluorescence emitted by the carbon quantum dot; After the carbon quantum dot is mixed with piroxicam-like substances, irradiate the carbon quantum dot with excitation lights of 340 nm and 280 nm, and respectively obtain the light intensities F 340 and F 280 of the 440-nm fluorescence emitted by the carbon quantum dot; Based on at least F 280 / F 340 and f 280 / f 340 , determine the concentration and type of piroxicam-like substances.

[0008] A training set, the construction method of which comprises: Configure a mixed solution of the carbon quantum dot and different types of piroxicam-like substances at different concentrations; Detect the PLE spectrum corresponding to each mixed solution; Combine the PLE spectrum and the corresponding piroxicam class combination to form a first training array; Combine at least part of the first training arrays into a training set.

[0009] An XGBoost trained with the above-mentioned training set.

[0010] An application of the above-mentioned XGBoost in the identification of piroxicam components.

[0011] A training set, the construction method includes: Configure a solution in which one type of piroxicam is mixed with carbon quantum dots at different concentrations; Irradiate the solution with excitation lights of 280nm and 340nm respectively and take pictures to obtain a first picture and a second picture; Obtain the RGB values of the pixel points in the first picture and the second picture; Combine the piroxicam concentration, the RGB values of the pixel points in the first picture corresponding to this concentration, and the RGB values of the pixel points in the second picture corresponding to this concentration to form a second training array; Combine at least part of the second training arrays into a training set.

[0012] A CNN trained with the above-mentioned training set.

[0013] An application of the above-mentioned CNN in detecting the concentration of piroxicam.

[0014] A detection method for piroxicam, including: Add the carbon quantum dots to the solution to be tested, detect the PLE spectrum of the solution to be tested, input the test data of the PLE spectrum into the above-mentioned XGBoost, and output the type of piroxicam in the solution to be tested by the XGBoost; Select the corresponding CNN according to the type of piroxicam output by the XGBoost; Irradiate the solution to be tested with excitation lights of 280nm and 340nm respectively and take pictures, input the RGB values of the pixel points in the two obtained pictures into the CNN, and output the corresponding concentration of piroxicam by the CNN.

[0015] A kind of carbon quantum dots, characterized in that the chemical formula is: .

[0016] A preparation method of carbon quantum dots, dissolve resorcinol and tris(hydroxymethyl)aminomethane in ethanol, and react in a reaction kettle at 140°C - 180°C to obtain the carbon quantum dots.

[0017] Piroxicam can quench the fluorescence of the quantum dots provided by the present invention and cause non-proportional decreases in the fluorescence intensities corresponding to its two excitation wavelengths. To a certain extent, the type and concentration of piroxicam can be determined by the ratio of the fluorescence intensities corresponding to the two excitation wavelengths.

[0018] Based on the quantum dots provided by the present invention, XGBoost can effectively identify the specific type of piroxicam. On this basis, based on the quantum dots provided by the present invention, CNN can further accurately identify the concentration of piroxicam. With the cooperation of the two, low-cost and high-precision detection of piroxicam is achieved. Description of the Drawings

[0019] Figure 1 TEM image of the carbon quantum dots in Example 1 of the present invention; Figure 2 XRD pattern of the carbon quantum dots in Example 1 of the present invention; Figure 3 Raman spectrum of the carbon quantum dots in Example 1 of the present invention; Figure 4 FT-IR spectrum (Fourier transform infrared spectrum) of the carbon quantum dots in Example 1 of the present invention; Figure 5 XPS spectrum of the carbon quantum dots in Example 1 of the present invention; Figure 6 High-resolution XPS spectrum of C 1s in Example 1 of the present invention; Figure 7 High-resolution XPS spectrum of N 1s in Example 1 of the present invention; Figure 8 High-resolution XPS spectrum of O 1s in Example 1 of the present invention; Figure 9 High-resolution XPS spectrum of F 1s in Example 1 of the present invention; Figure 10 Ultraviolet-visible absorption spectrum (Abs), PL spectrum, and PLE spectrum of the carbon quantum dots in Example 1 of the present invention; Figure 11 Fourier emission spectrum of the carbon quantum dots in Example 1 of the present invention under illumination at different wavelengths; Figure 12 Excitation-emission matrix (EEM) image of the carbon quantum dots in Example 1 of the present invention; Figure 13 Relative fluorescence intensity of the 440 nm fluorescence emitted by the carbon quantum dots in Example 1 of the present invention after ultraviolet illumination for different durations; Figure 14The relative fluorescence intensity of the 440 nm fluorescence emitted by the carbon quantum dots in Example 1 of the present invention after being exposed to sunlight for different durations; Figure 15 The relative fluorescence intensity of the 440 nm fluorescence emitted by the carbon quantum dots in Example 1 of the present invention in NaCl solutions with different concentrations; Figure 16 The fluorescence intensity of the 440 nm fluorescence emitted by the carbon quantum dots in Example 1 of the present invention in solutions with different pH values; Figure 17 The EEM diagrams of the carbon quantum dots before and after mixing with MLX in Example 2 of the present invention; Figure 18 The PLE spectra of the carbon quantum dots after mixing with different concentrations of MLX in Example 2 of the present invention; Figure 19 For Example 2 of the present invention, F 280 / F 340 Variation curve with MLX concentration; Figure 20 For Example 2 of the present invention, F 280 / F 340 Variation curve one with LNX concentration; Figure 21 For Example 2 of the present invention, F 280 / F 340 Variation curve two with LNX concentration; Figure 22 For Example 2 of the present invention, F 280 / F 340 Variation curve one with PRX concentration; Figure 23 For Example 2 of the present invention, F 280 / F 340 Variation curve two with PRX concentration; Figure 24 For Example 2 of the present invention, F 280 / F 340 Variation curve one with TNX concentration; Figure 25 For Example 2 of the present invention, F 280 / F 340 Variation curve two with TNX concentration; Figure 26 The ultraviolet-visible absorption spectra of MLX, carbon quantum dots, and the mixture of MLX and carbon quantum dots; Figure 27 The ultraviolet-visible absorption spectrum of MLX, the PL spectrum, and the PLE spectrum of carbon quantum dots; Figure 28Fluorescence decay curves of carbon quantum dots before and after mixing with MLX under 280 nm light illumination and 340 nm light illumination in Example 2 of the present invention; Figure 29 Typical scoring diagrams when MLX, LNX, PRX, and TNX are 100 μM in Example 3 of the present invention; Figure 30 Typical scoring diagrams when MLX, LNX, PRX, and TNX are 30 μM in Example 3 of the present invention; Figure 31 Test results of XGBoost in the identification of pain-relieving and anti-inflammatory components in Example 3 of the present invention; Figure 32 Test results of random forest in the identification of pain-relieving and anti-inflammatory components in Example 3 of the present invention; Figure 33 Test results of logistic regression in the identification of pain-relieving and anti-inflammatory components in Example 3 of the present invention; Figure 34 Deviation diagram between the predicted result of CNN for MLX concentration and the actual concentration of MLX in Example 4 of the present invention; Figure 35a Deviation diagram between the predicted result of CNN for LNX concentration and the actual concentration in Example 4 of the present invention; Figure 35b Deviation diagram between the predicted result of CNN for PRX concentration and the actual concentration in Example 4 of the present invention; Figure 35c Deviation diagram between the predicted result of CNN for TNX concentration and the actual concentration in Example 4 of the present invention; Figure 36 FT-IR spectrum of carbon quantum dots in Example 5 of the present invention; Figure 37 XPS spectrum of carbon quantum dots in Example 5 of the present invention; Figure 38 EEM spectrum of carbon quantum dots in Example 5 of the present invention; Figure 39 I 425 / I 533 Numerical bar schematic diagram after carbon quantum dots are mixed with different substances in Example 5 of the present invention. Detailed implementation manners

[0020] To make the above objects, features, and advantages of the present invention more obvious and understandable, the following provides a detailed description of the specific implementation manners of the present invention. Many specific details are set forth in the following description to fully understand the present invention. However, the present invention can be implemented in many other ways different from those described herein, and those skilled in the art can make similar improvements without departing from the connotation of the present invention. Therefore, the present invention is not limited by the specific embodiments disclosed below.

[0021] Example 1: This example provides a kind of carbon quantum dots, whose chemical formula is: .

[0022] It is not difficult to see that N and F are doped in the carbon quantum dots.

[0023] The preparation method of the carbon quantum dots includes the following steps: Add 1 mmol fleroxacin into 30 mL ethanol and ultrasonicate for 5 min. After complete dispersion, transfer it to a 50 mL volume reaction kettle and react at 180 °C - 220 °C (200 °C in this example) for 8 h; After the reaction is completed, cool it, first filter with a filter screen with a pore diameter of 0.22 μm, and then dialyze with a dialysis membrane with a cut-off molecular weight of 500 Da at room temperature for 12 h to obtain the above carbon quantum dots.

[0024] The TEM image of the carbon quantum dots is as shown in Figure 1 , which shows that the carbon quantum dots of this example have a good spherical structure and good dispersibility. After further detection, the particle size of the carbon quantum dots of this example is mainly distributed between 3 nm and 7 nm, with an average of about 5 nm. In addition, based on Figure 1 the high-resolution TEM image shown in the upper right corner, it can be seen that the carbon quantum dots have an obvious lattice structure, and the lattice plane spacing is 0.21 nm.

[0025] Refer to Figure 2 , the carbon quantum dots of this example have a relatively wide peak at 2θ = 23.1°, indicating that the carbon quantum dots have a layered structure. Further combined with Figure 3 , the intensity ratio of the D peak (1365 cm -1 ) and the G peak (1600 cm -1 ) in the Raman spectrum is 0.91, further proving that the above carbon quantum dots have a highly graphitized phenomenon.

[0026] Refer to Figure 4 , the carbon quantum dots of this example have a relatively large absorption peak between 3340 cm -1 and 3590 cm -1 , corresponding to O-H and N-H inside. The characteristic peaks of the carbon quantum dots between 2830 cm -1 and 2972 cm -1 correspond to C-H. The characteristic peaks at 1042 cm -1 , 1480 cm -1 , 1624 cm -1 , 1725 cm -1 correspond to the groups C-F, C-N, C=N, and C=O respectively. Combined with Figure 5, the XPS spectrum of the carbon quantum dots has characteristic peaks at 284.4 eV, 398.1 eV, 530.5 eV, and 685.5 eV, corresponding to C 1s, N 1s, O 1s, and F 1s, respectively. Finally, the atomic concentration ratios of C, N, O, and F in the carbon quantum dots are calculated to be 54.90%, 18.29%, 20.74%, and 6.07%, respectively.

[0027] See further Figures 6 - 9 . Specifically, Figure 6 shows three peaks, located at 287.4 eV (corresponding to C-N / C-F), 284.8 eV (corresponding to C=O / C-O), and 283.3 eV (corresponding to C=C / C-C), respectively; Figure 7 shows two peaks, located at 398.1 eV (corresponding to pyridine N) and 399.6 eV (corresponding to graphitic N), respectively; Figure 8 shows two peaks, located at 530.5 eV (corresponding to C-O) and 531.7 eV (corresponding to C=O), respectively; Figure 9 shows two peaks, located at 684.2 eV (corresponding to semi-ionized C-F) and 685.5 eV (corresponding to covalent C-F), respectively.

[0028] By Figures 4 - 9 , it can be effectively proved that the carbon quantum dots of the aforementioned chemical formula can be obtained based on the preparation method of the carbon quantum dots of this embodiment.

[0029] As Figure 10 shown, the ultraviolet-visible absorption spectrum of the above carbon quantum dots has a peak at 210 nm, corresponding to the π-π* transition of C=C. In addition, there is also a peak at 283 nm, corresponding to the n-π* transition of C=O. There is also a weak absorption peak between 300 nm and 350 nm, corresponding to the n-π* transition of C-F and C-N. The emission wavelength in the PL spectrum is 440 nm. Further, in the PLE spectrum, the emission wavelength of 440 nm corresponds to two excitation wavelengths of 280 nm and 340 nm. As Figure 10 shown in the inset of

[0030] See further Figure 11 and Figure 12 , the emission wavelength of the carbon quantum dots of this embodiment is constantly maintained at 440 nm and does not change with the change of the excitation light wavelength.

[0031] Taking quinine sulfate as the standard, the fluorescence quantum yield of the carbon quantum dots of this embodiment is 43%.

[0032] As Figure 13As shown, the relative fluorescence intensity on the vertical axis represents the ratio of the fluorescence intensity of the 440-nm fluorescence emitted by the carbon quantum dots after a certain period of ultraviolet light irradiation to the fluorescence intensity of the 440-nm fluorescence emitted by the carbon quantum dots without ultraviolet light irradiation. The horizontal axis is the duration of ultraviolet light irradiation to which the carbon quantum dots are subjected. Based on Figure 13 It can be seen that even after the carbon quantum dots in this example are subjected to 24 hours of ultraviolet light irradiation, the fluorescence intensity generated by them hardly changes.

[0033] As Figure 14 shown, the relative fluorescence intensity on the vertical axis represents the ratio of the fluorescence intensity of the 440-nm fluorescence emitted by the carbon quantum dots after a certain period of sunlight exposure to the fluorescence intensity of the 440-nm fluorescence emitted by the carbon quantum dots without sunlight exposure. The horizontal axis is the duration of sunlight exposure to which the carbon quantum dots are subjected. Based on Figure 14 It can be seen that even after the carbon quantum dots in this example are subjected to 30 days of sunlight exposure, the fluorescence intensity generated by them hardly changes.

[0034] As Figure 15 shown, the relative fluorescence intensity on the vertical axis represents the ratio of the fluorescence intensity of the 440-nm fluorescence emitted by the carbon quantum dots in a specific concentration of NaCl solution to the fluorescence intensity of the 440-nm fluorescence emitted by the carbon quantum dots in deionized water. The horizontal axis is the concentration of NaCl. Based on Figure 15 It can be seen that even in a 2.1 mol / L concentrated NaCl solution, the fluorescence intensity of the carbon quantum dots in this example hardly changes.

[0035] As Figure 16 shown, the vertical axis represents the fluorescence intensity of the 440-nm fluorescence emitted by the carbon quantum dots. When the excitation light wavelength is 280 nm, as the pH value of the solution in which the carbon quantum dots are located increases, the fluorescence intensity of the 440-nm fluorescence emitted by them only decreases slightly. Similarly, when the excitation light wavelength is 320 nm, as the pH value of the solution in which the carbon quantum dots are located increases, the fluorescence intensity of the 440-nm fluorescence emitted by them also only decreases slightly.

[0036] Based on Figures 14 - 16 It can be seen that the carbon quantum dots in this example have extremely strong light resistance, salt resistance, and acid and alkali resistance characteristics, showing extremely strong stability.

[0037] Example 2: This embodiment provides an application of carbon quantum dots as described in Embodiment 1, which can be used to detect the residues of oxicams. Here, oxicams at least include meloxicam (abbreviated as MLX), lornoxicam (abbreviated as LNX), piroxicam (abbreviated as PRX), and tenoxicam (abbreviated as TNX).

[0038] Taking MLX as an example, as Figure 17 and Figure 18 shown, after the carbon quantum dots are mixed with MLX, regardless of whether the carbon quantum dots are excited with excitation light at 280 nm or 340 nm, the fluorescence intensity at 440 nm generated by the carbon quantum dots decreases, and the higher the concentration of MLX, the smaller the fluorescence intensity at 440 nm. This provides a basis for the quantitative detection of MLX by carbon quantum dots.

[0039] In Figures 19 - 25 where F 280 represents the fluorescence intensity at 440 nm emitted by the carbon quantum dots when irradiated with excitation light at 280 nm, and F 340 represents the fluorescence intensity at 440 nm emitted by the carbon quantum dots when irradiated with excitation light at 340 nm, where the excitation light at 280 nm and 340 nm excites the carbon quantum dots with the same light intensity (the excitation source voltage is unified as 650 V in this embodiment and subsequent embodiments).

[0040] From Figure 19 it can be seen that after the carbon quantum dots are mixed with MLX, as the concentration of MLX increases, there is an obvious positive correlation between F 280 / F 340 and the concentration of MLX. In other words, based on the value of F 280 / F 340 the concentration of MLX mixed in the carbon quantum dots can be determined. Especially when the concentration of MLX is in the range of 0 - 25 μM, there is a good linear relationship between the concentration of MLX and F 280 / F 340 (relatively accurate quantitative detection can be achieved within this concentration range), and at the same time, the detection limit of the carbon quantum dots for the concentration of MLX is 97 nM.

[0041] On the contrary, as Figure 20 and Figure 21 shown, after the carbon quantum dots are mixed with LNX, there is an inverse correlation between F 280 / F 340 and the concentration of LNX. Of course, from this, the concentration of LNX can also be determined based on F 280 / F 340The value determines the concentration of LNX mixed in the carbon quantum dots. Among them, when the LNX concentration is in the range of 0 - 40 μM, there is a good linear relationship between the LNX concentration and F 280 / F 340 The lower detection limit of the carbon quantum dots for the LNX concentration is 75 nM.

[0042] Similar to MLX, see Figure 22 and Figure 23 After the carbon quantum dots are mixed with PRX, there is an obvious positive correlation between F 280 / F 340 and the PRX concentration. Based on the value of F 280 / F 340 The concentration of PRX mixed in the carbon quantum dots can be determined. Among them, when the PRX concentration is in the range of 0 - 30 μM, there is a good linear relationship between the PRX concentration and F 280 / F 340 The lower detection limit of the carbon quantum dots for the PRX concentration is 54 nM.

[0043] Similar to LNX, see Figure 24 and Figure 25 After the carbon quantum dots are mixed with TNX, there is an obvious negative correlation between F 280 / F 340 indicating that the concentration of TNX mixed in the carbon quantum dots can be determined based on the value of F 280 / F 340 Among them, when the TNX concentration is in the range of 0 - 21 μM, there is a good linear relationship between the TNX concentration and F 280 / F 340 The lower detection limit of the carbon quantum dots for the TNX concentration is 102 nM.

[0044] Based on this, this embodiment provides a detection method for tolmetin-like substances, including the following steps: Step 101: Before the carbon quantum dots are mixed with tolmetin-like substances, irradiate the carbon quantum dots with 340 nm excitation light to obtain the light intensity f 340 of the 440 nm fluorescence emitted by the carbon quantum dots; Step 102: Before the carbon quantum dots are mixed with tolmetin-like substances, irradiate the carbon quantum dots with 280 nm excitation light to obtain the light intensity f 280 of the 440 nm fluorescence emitted by the carbon quantum dots; Step 103: After the carbon quantum dots are mixed with tolmetin-like substances, irradiate the carbon quantum dots with 340 nm excitation light to obtain the light intensity F 340 of the 440 nm fluorescence emitted by the carbon quantum dots; Step 104: After mixing carbon quantum dots with piroxicam-like substances, irradiate the carbon quantum dots with an excitation light of 280 nm to obtain the light intensity F of the 440 nm fluorescence emitted by the carbon quantum dots. 280 ; Step 105: Determine the concentration and type of piroxicam-like substances based on at least F 280 / F 340 and f 280 / f 340 .

[0045] To further explore the detection mechanism of carbon quantum dots for piroxicam-like substances, as Figure 26 shown, taking MLX as an example, when MLX is mixed with carbon quantum dots, no new peak is generated in the mixture relative to MLX and carbon quantum dots alone, which proves that no new compound is generated between MLX and carbon quantum dots.

[0046] Furthermore, as Figure 27 shown, there is a large overlap between the PL spectrum and PLE spectrum of carbon quantum dots and the UV-vis absorption spectrum of MLX, which proves that carbon quantum dots detect MLX based on both IEE (inner filter effect) and FRET (fluorescence resonance energy transfer). Further referring to Figure 28 , before and after mixing carbon quantum dots with MLX, under the excitation of 280 nm light, the fluorescence lifetime of carbon quantum dots decreases from 2.98 ns to 2.88 ns, with a small decrease, indicating that the detection of MLX by carbon quantum dots is mainly dominated by IEE at this time. Under the excitation of 340 nm light, the fluorescence lifetime of carbon quantum dots decreases from 3.79 ns to 3.06 ns, with a large decrease, indicating that the degree of FRET increases during the detection of MLX by carbon quantum dots at this time.

[0047] Furthermore, E obsd represents the percentage of fluorescence quenching at 440 nm after mixing carbon quantum dots with piroxicam-like substances, and E cor represents the percentage of fluorescence quenching after removing IEE from E obsd .

[0048] When the piroxicam-like substance is MLX and the excitation wavelength is 280 nm, E obsd = 27.3%, E cor = 1.9%; when the piroxicam-like substance is MLX and the excitation wavelength is 340 nm, E obsd = 29.1%, E cor = 8.2%.

[0049] Similarly, before and after the mixing of carbon quantum dots and LNX, under the excitation of 280 nm light, the fluorescence lifetime of the carbon quantum dots decreased from 2.98 ns to 2.91 ns, with a relatively small decrease. Under the excitation of 340 nm light, the fluorescence lifetime of the carbon quantum dots decreased from 3.79 ns to 2.99 ns, with a relatively large decrease. When the pain relief and anti-inflammatory drug is LNX and the excitation wavelength is 280 nm, E obsd = 38.1%, E cor = 0.6%; when the pain relief and anti-inflammatory drug is LNX and the excitation wavelength is 340 nm, E obsd = 24.9%, E cor = 2.6%.

[0050] Before and after the mixing of carbon quantum dots and PRX, under the excitation of 280 nm light, the fluorescence lifetime of the carbon quantum dots decreased from 2.98 ns to 2.91 ns, with a relatively small decrease. Under the excitation of 340 nm light, the fluorescence lifetime of the carbon quantum dots decreased from 3.79 ns to 3.19 ns, with a relatively large decrease. When the pain relief and anti-inflammatory drug is PRX and the excitation wavelength is 280 nm, E obsd = 27.6%, E cor = 3.5%; when the pain relief and anti-inflammatory drug is PRX and the excitation wavelength is 340 nm, E obsd = 61%, E cor = 41.9%.

[0051] Before and after the mixing of carbon quantum dots and TNX, under the excitation of 280 nm light, the fluorescence lifetime of the carbon quantum dots decreased from 2.98 ns to 2.89 ns, with a relatively small decrease. Under the excitation of 340 nm light, the fluorescence lifetime of the carbon quantum dots decreased from 3.79 ns to 3.11 ns, with a relatively large decrease. When the pain relief and anti-inflammatory drug is TNX and the excitation wavelength is 280 nm, E obsd = 40.8%, E cor = 7.6%; when the pain relief and anti-inflammatory drug is TNX and the excitation wavelength is 340 nm, E obsd = 41.8%, E cor = 6.8%.

[0052] Based on the above analysis, it can be seen that for the detection of pain relief and anti-inflammatory drugs by carbon quantum dots, when the excitation light is 280 nm, IFE plays a dominant role; when the excitation light is 340 nm, for MLX, LNX and TNX, IFE still plays a dominant role, but the effect of FRET increases. Only for PRX, FRET plays a dominant role. Different types of pain relief and anti-inflammatory drugs cause different degrees of IEE and FRET, which also leads to different corresponding relationships between the concentrations and F 280 / F 340 values of different types of pain relief and anti-inflammatory drugs, which provides a basis for the carbon quantum dots to distinguish the types of pain relief and anti-inflammatory drugs.

[0053] Example 3: Based on the analysis in Example 2, when the painkillers are MLX and PRX, there is a positive correlation between F 280 / F 340 and the concentration of the painkillers. When the painkillers are LNX and TNX, there is a negative correlation between F 280 / F 340 and the concentration of the painkillers. Considering the experimental error and the influence of the parameter settings, based on the detection method provided in Example 2, it is impossible to effectively distinguish MLX and PRX, or LNX and TNX only by manual recognition.

[0054] Based on this, in this example, we first attempt to perform linear discriminant analysis based on F 280 and F 340 corresponding to the mixing of carbon quantum dots with different types of painkillers, and try to distinguish the painkillers through machine learning. As Figure 29 shown, when the concentrations of MLX, LNX, PRX, and TNX in the carbon quantum dots are all 100 μM, the spectral clustering of MLX, LNX, PRX, and TNX can be completely distinguished. As Figure 30 shown, when the concentrations of MLX, LNX, PRX, and TNX in the carbon quantum dots are reduced to 30 μM, the spectral clustering spacing of MLX, LNX, PRX, and TNX decreases, some eigenvalues are close, and the difficulty of distinction increases, especially it is difficult to distinguish LNX and TNX. It is not difficult to understand that when the concentration of the painkillers in the carbon quantum dots is further reduced and approaches the lower limit of the detection of the concentration of the painkillers by the carbon quantum dots, it is very likely that the linear discriminant analysis will be completely unable to distinguish between MLX, LNX, PRX, and TNX.

[0055] Based on this problem, this example provides an application of XGBoost (eXtreme Gradient Boosting) in the identification of painkiller components. Different from other machine learning methods, when the concentration of the painkillers is lower than 30 μM, XGBoost can effectively distinguish MLX, LNX, PRX, and TNX.

[0056] In order to achieve the above effect, the construction method of the first training set (for the convenience of distinguishing from the subsequent examples, the training set is renamed the first training set here) required in the training process of XGBoost is as follows: Step 201: Configure the mixed solutions of carbon quantum dots and different types of painkillers at different concentrations.

[0057] Specifically, 36 first mixed solutions of MLX and carbon quantum dots with different concentrations are prepared. The concentrations of MLX in the 36 first mixed solutions are 0.1 μM, 0.2 μM,......, 3.5 μM respectively; 36 second mixed solutions of LNX and carbon quantum dots with different concentrations are prepared. The concentrations of LNX in the 36 second mixed solutions are 0.1 μM, 0.2 μM,......, 3.5 μM respectively; 36 third mixed solutions of PRX and carbon quantum dots with different concentrations are prepared. The concentrations of PRX in the 36 third mixed solutions are 0.1 μM, 0.2 μM,......, 3.5 μM respectively; 36 fourth mixed solutions of TNX and carbon quantum dots with different concentrations are prepared. The concentrations of TNX in the 36 fourth mixed solutions are 0.1 μM, 0.2 μM,......, 3.5 μM respectively. The concentration of carbon quantum dots in the first mixed solution, the second mixed solution, the third mixed solution and the fourth mixed solution is 100 μg / nL.

[0058] Step 202: Detect the PLE spectra of at least part of the first mixed solution, at least part of the second mixed solution, at least part of the third mixed solution, and at least part of the fourth mixed solution.

[0059] Specifically, when detecting the PLE spectrum in this embodiment, the excitation light wavelength scanning range is 200 nm - 800 nm, and the scanning step is 1 nm. Therefore, each PLE spectrum has 601 data points, and each data point contains the excitation light wavelength and the 440 nm emission light intensity information. In this embodiment, the 601 data points included in the same PLE spectrum are used as a whole.

[0060] Preferably, in this embodiment, the PLE spectra of all the first mixed solutions, all the second mixed solutions, all the third mixed solutions, and all the fourth mixed solutions are detected.

[0061] Further preferably, in order to reduce experimental errors, in this embodiment, the PLE spectra of all the first mixed solutions, all the second mixed solutions, all the third mixed solutions, and all the fourth mixed solutions are detected 10 times.

[0062] Step 203: Combine each PLE spectrum and the corresponding type combination of painkillers to form a first training array. The first training array does not include the painkiller concentration information corresponding to the PLE spectrum, and combine all the first training arrays to form the first training set.

[0063] In this embodiment, 80% of the first training arrays in the first training set are used to train XGBoost, and 20% of the first training arrays are used to test the training results of XGBoost. The test results are as Figure 31As shown, the XGBoost performs discriminative detection on MLX, LNX, PRX, and TNX under low-concentration (not higher than 3.5 μM) conditions, with an accuracy rate, recall rate, and F1 value all being 100%.

[0064] For comparison, the random forest and logistic regression are also trained and tested using the above-mentioned first training set, and the test results are as Figure 32 and Figure 33 shown. Different from XGBoost, the random forest and logistic regression only have good recognition capabilities for LNX and MLX, and have very low recognition capabilities for PRX and TNX. Especially for the random forest, it is completely unable to recognize PRX.

[0065] This proves that in order to distinguish MLX, LNX, PRX, and TNX using machine learning under low-concentration conditions, not only a specific first training set is required, but also a specific machine learning model needs to be relied on to achieve this.

[0066] Example 4: Based on the application of XGBoost in the recognition of pain-relieving and anti-inflammatory drug components in Example 3, this example further provides an application of a CNN (Convolutional Neural Network) in the detection of pain-relieving and anti-inflammatory drug concentrations, thereby eliminating the detection of f 280 and f 340 in Example 2.

[0067] It should be noted that since the emission wavelength of the carbon quantum dots in this example is 440 nm, which belongs to visible light, this example provides a method for constructing a second training set (named the second training set here for easy distinction from the previous examples) required during the training process of the CNN, including the following steps: Step 301: Configure a fifth mixed solution of different concentrations of pain-relieving and anti-inflammatory drugs of a specific type and carbon quantum dots.

[0068] Specifically, in this example, the pain-relieving and anti-inflammatory drug is MLX, and there are a total of 100 portions of the fifth mixed solution. The corresponding MLX concentrations are 1 μM, 2 μM,..., 100 μM in sequence, and the carbon quantum dot concentration is 100 μg / nL.

[0069] Step 302: Irradiate the fifth mixed solution with 280 nm excitation light and take a photo of the light generated by the carbon quantum dots to form a first photo; irradiate the fifth mixed solution with 340 nm excitation light and take a photo of the light generated by the carbon quantum dots to form a second photo; thus, obtain the first photo and the second photo corresponding to all the fifth mixed solutions, and at the same time obtain the RGB values of all pixel points in all the first photos and all the second photos.

[0070] Step 303: Combine the concentration of Tongxikang class, the RGB values of all pixel points in the first photo corresponding to this concentration, and the RGB values of all pixel points in the second photo corresponding to this concentration to form a second training array.

[0071] Step 304: Preferably, repeat Step 302 and Step 303 ten times, so that each fifth mixed solution obtains 10 groups of corresponding second training arrays.

[0072] Step 305: Combine all the second training arrays into a second training set.

[0073] Use 80% of the second training arrays to train the CNN. The CNN uses 16 convolutional layers, 5 average pooling layers, and 3 fully connected layers during the training process. Use the remaining 20% of the second training arrays to test the trained CNN. The specific test results are as Figure 34 shown. It can be seen that the deviation between the result of each MLX concentration prediction by the CNN and the actual concentration is very small. The average deviation value obtained by averaging the deviations between all prediction results and the actual concentration is only 124 nM, which is much lower than the minimum difference of 1 μM in the MLX concentration between different fifth mixed solutions. Therefore, it can be considered that the CNN trained with the second training set in this embodiment has a 100% prediction accuracy for the MLX concentration.

[0074] In some other embodiments, the MLX in the fifth mixed solution can be replaced with TNX, LNX, or PRX, but the corresponding concentration of Tongxikang class in 100 parts of the fifth mixed solution is still 1 μM, 2 μM,..., 100 μM. Correspondingly, after the CNN training is finally completed, the concentration prediction and judgment can be performed for one of TNX, LNX, and PRX. In other words, after the CNN training is completed, it can only perform the concentration judgment for the specific type of Tongxikang class corresponding to the second training array used in its training process. Therefore, in order to perform the concentration judgment for MLX, TNX, LNX, and PRX respectively, four different CNNs need to be trained separately. As Figure 35c shown, for the CNN trained for TNX, the average deviation between the predicted TNX concentration value output by it and the actual TNX value is 145 nM; as Figure 35a shown, for the CNN trained for LNX, the average deviation between the predicted LNX concentration value output by it and the actual LNX value is 97 nM; as Figure 35b shown, for the CNN trained for PRX, the average deviation between the predicted PRX concentration value output by it and the actual PRX value is 116 nM. The above average deviation values are all much less than 1 μM and can be attributed to the error range in the actual use process. Therefore, it can also be considered that their concentration recognition accuracy is 100%.

[0075] Based on this, this embodiment further provides a method for detecting tongxikang, including the following steps: Step 401: Detect the PLE spectrum of the solution to be measured, input the test data of the PLE spectrum into the XGBoost after training as described in Embodiment 3, and output the types of tongxikang in the solution to be measured by XGBoost; Step 403: According to the types of tongxikang output by XGBoost, select the CNN after training corresponding to this embodiment; Step 404: Irradiate the solution to be measured with excitation lights of 280nm and 340nm respectively and take pictures, input the RGB values of all pixel points in the two obtained pictures into the CNN selected in Step 403, and output the corresponding concentration of tongxikang by the CNN.

[0076] Embodiment 5: This embodiment further provides a method for preparing carbon quantum dots, including the following steps: Dissolve 1 mmol of naphthoresorcinol and 1 mmol of tris (hydroxymethyl) aminomethane in 20 ml of ethanol, ultrasonically treat for 15 min, then transfer to a reaction kettle and heat at 160 °C for 6 h. After the reaction is completed and cooled to room temperature, filter using a filter screen with a pore diameter of 0.22 μm, then dialyze the filtrate using a dialysis tube with a molecular weight cut-off of 500 Da for 24 h, and then perform lyophilization to obtain the carbon quantum dots.

[0077] Specifically as Figure 36 shown, the carbon quantum dots of this embodiment have four stretching vibration peaks, located at 1067 cm -1 , 1476 cm -1 , 1595 cm -1 , 1719 cm -1 respectively. The four stretching vibration peaks correspond to C-N, C-C, C=O, and C=N respectively. In addition, the carbon quantum dots of this embodiment also have an absorption band between 2900 cm -1 and 3400 cm -1 , corresponding to the stretching vibration of -NH2 / -OH. Further combined with Figure 37 , the C content in the carbon quantum dots of this embodiment is 56.26%, the N content is 14.23%, and the O content is 29.51%. Based on this, the chemical formula of the carbon quantum dots of this embodiment is .

[0078] As Figure 38As shown, the carbon quantum dots of this embodiment have two fluorescence emission centers. The excitation wavelengths corresponding to the two fluorescence emission centers are 374 nm and 470 nm respectively, and the corresponding emission wavelengths are 425 nm and 533 nm respectively. For convenience of description, the carbon quantum dots of this embodiment are irradiated with 374-nm excitation light, and the fluorescence intensity of 425 nm emitted by the carbon quantum dots is I 425 ; when the carbon quantum dots of the embodiment are irradiated with 470-nm excitation light, the fluorescence intensity of 533 nm emitted by the carbon quantum dots is I 533 . The gains corresponding to the 374-nm excitation light and the 470-nm excitation light in this embodiment are both 550 V.

[0079] Furthermore, as shown in Figure 39 , after the carbon quantum dots of this embodiment (the concentration of carbon quantum dots is 50 ng / μL) are mixed with different substances (the concentration is uniformly 100 μM), only TNX, LNX, PRX, and MLX can make I 425 / I 533 change significantly. This proves that the carbon quantum dots of this embodiment can also be used to detect drugs of the tongxikang type. In addition, since Na + , K + , Ag + , Mg 2+ , Zn 2+ , Co 2+ , Ca 2+ , Ce 2+ , Pb 2+ , Hg 2+ , Cd 2+ , Cu 2+ , Mn 2+ , Al 3+ , Cl - , SO4 2- , NO3 - , SO3 2- , glucose (Glu), ascorbic acid (AA), urea, uric acid (UA), L-asparagine (L-asp), L-serine (L-ser), L-tryptophan (L-try), L-proline (L-pro), L-alanine (L-ala), L-cysteine (L-cys), D-alanine (D-ala) cannot make I 425 / I 533 change significantly, further proving the specificity of the carbon quantum dots of this embodiment for detecting drugs of the tongxikang type.

[0080] The technical features of the above-described embodiments can be combined arbitrarily. For the sake of concise description, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered as falling within the scope described in this specification.

[0081] The above-described embodiments merely represent several implementation manners of the present invention, and the description thereof is relatively specific and detailed. However, it should not be construed as a limitation on the scope of the invention patent. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present invention, several modifications and improvements can still be made, and these all belong to the protection scope of the present invention. Therefore, the protection scope of the invention patent should be subject to the appended claims.

Claims

1. A carbon quantum dot, characterized in that, The chemical formula is: .

2. A preparation method of carbon quantum dots, characterized in that, Dissolve fleroxacin in ethanol and react in a reaction kettle at 180 °C - 220 °C to obtain the carbon quantum dots as described in claim 1.

3. Application of the carbon quantum dots as described in claim 1 in the detection of phenylbutazone-like substances.

4. A detection method for tongxikang, characterized in that, It includes: Before the carbon quantum dots and piroxicam - like substances described in claim 1 are mixed, the carbon quantum dots are irradiated with excitation lights of 340 nm and 280 nm, and the light intensities f 340 and f 280 ; After the carbon quantum dots described in claim 1 are mixed with piroxicam-like substances, the carbon quantum dots are irradiated with excitation lights of 340 nm and 280 nm, and the light intensities F 340 and F 280 ; Based on at least F 280 / F 340 and f 280 / f 340 , determine the concentration and type of tongxikang 5. A training set, characterized in that, The construction method includes: Prepare a mixed solution of the carbon quantum dots as described in claim 1 and different types of phenylbutazone-like substances at different concentrations; Detect the PLE spectrum corresponding to each mixed solution; Combine the PLE spectrum and the type combination of the phenylbutazone-like substance corresponding to the PLE spectrum to form a first training array; Combine at least part of the first training arrays into a training set.

6. An XGBoost, characterized in that, Train with the training set as described in claim 5.

7. Application of XGBoost as described in claim 6 in the identification of phenylbutazone-like components.

8. A carbon quantum dot, characterized in that, The chemical formula is: .

9. A preparation method of carbon quantum dots, characterized in that, Dissolve resorcinol and tris(hydroxymethyl)aminomethane in ethanol and then react in a reaction kettle under the condition of 140 °C - 180 °C to obtain the carbon quantum dots as described in claim 8.

10. Application of the carbon quantum dots as described in claim 8 in the detection of phenylbutazone-like substances.

Images