Functional group classification-based chemical ionization mass spectrometry ambient humidity correction method

Abstract

The application discloses a chemical ionization mass spectrometry environment humidity correction method based on functional group classification, and comprises the following steps: establishing a humidity-mass spectrometry sensitivity dataset of different functional group species; based on the humidity-mass spectrometry sensitivity dataset, a humidity-mass spectrometry sensitivity data correction equation based on machine learning functional group classification is established; and the actual mass spectrometry sensitivity of a target species under current environmental humidity conditions is calculated according to the humidity-mass spectrometry sensitivity data correction equation. The application acquires the mass spectrometry sensitivity response characteristics of different functional group species under different environmental humidity conditions, constructs a classified humidity-mass spectrometry sensitivity data correction equation, effectively reduces the uncertainty caused by humidity changes on the mass spectrometry quantitative results of semi-volatile organic substances, improves the accuracy and stability of quantitative analysis under different humidity conditions, and solves the problem that the humidity influence is not effectively considered or only a single humidity correction method is adopted in the prior art, resulting in insufficient applicability to multiple functional group species.

Description

Technical Field

[0001] This invention belongs to the field of analytical chemistry and environmental monitoring technology, specifically relating to a chemical ionization mass spectrometry method for environmental humidity correction based on functional group classification. Background Technology

[0002] With the increasing severity of secondary air pollution, ozone (O3) and fine particulate matter (PM2.5) are becoming increasingly prominent. 2.5 Pollution has become a major challenge in global environmental governance. Moderate volatile organic compounds (S / IVOCs) in the atmosphere... x As an important photochemical precursor, it reacts with nitrogen oxides (NOx) under the action of oxidants such as hydroxyl radicals. x The free radical-driven oxidation reaction occurs with respect to O3 and PM2.5. 2.5 The formation of atmospheric pollutants (S / IVOCs) has a significant impact, and achieving highly accurate online measurement of S / IVOCs is a prerequisite for understanding the mechanisms of secondary atmospheric pollution. However, due to the characteristics of S / IVOCs, such as low concentration, short lifetime, and multiple functional groups, accurate measurement of them is quite difficult. Therefore, solving the problem of accurate measurement of S / IVOCs is of great scientific value for accurately assessing atmospheric chemical processes and improving the predictive ability of air quality models.

[0003] In recent years, the development of chemical ionization mass spectrometry (CIMS) has provided new methods for measuring S / IVOCs. Among them, iodine ion source chemical ionization mass spectrometry (I-CIMS) can separate the analyte species from iodine ions (I- ions). - I-CIMS combines with iodide adducts to form iodide-adducts, exhibiting low fragmentation and good detection capability for S / IVOCs containing hydroxyl, carboxyl, peroxy, and multifunctional groups. It is commonly used in field and laboratory measurements of S / IVOCs. However, I-CIMS measurements of S / IVOCs are significantly affected by ambient humidity. When humidity increases, the sensitivity of the analyte increases if the introduction of water vapor stabilizes the iodide-adduct. Conversely, the introduction of water vapor may attract iodide ions and combine to form iodide water ions (H₂OI₂). - When humidity is high, the sensitivity of the tested species decreases. Studies have shown that the effect of humidity on I-CIMS is related to the type and number of polar functional groups in the iodide-adduct, and I-CIMS exhibits significant differences in sensitivity and humidity response to species with different functional group types.

[0004] In existing technologies, relevant quantitative methods either fail to consider environmental humidity factors, such as calibrating by heating and evaporating prepared liquid standards onto the filter membrane of a gas and aerosol filter inlet (FIGAERO). Such methods are difficult to introduce and control environmental humidity conditions during calibration, thus failing to effectively correct for humidity effects. Alternatively, while considering the impact of humidity on mass spectrometry sensitivity, they often use a single humidity correction factor to uniformly correct the measurement results. This approach typically assumes that different compounds have similar response behaviors to humidity changes, making it difficult to apply to systems of multiple compounds with different functional group structures. Especially under complex and variable external humidity conditions, it easily introduces systematic quantitative errors. This results in significant uncertainties in current I-CIMS techniques for the qualitative and quantitative detection of low- and medium-volatile organic compounds.

[0005] In summary, existing I-CIMS technology is significantly affected by changes in ambient humidity in practical applications, and the effect of humidity on different measured species exhibits obvious functional group differences, making it difficult to achieve accurate quantification of S / IVOCs under varying humidity conditions. Therefore, there is an urgent need to develop a novel chemical ionization mass spectrometry method for environmental humidity correction, capable of systematically revealing the humidity response patterns of species with different functional groups and implementing a universal technical solution for classified humidity correction. This would improve the applicability to species with different functional groups and enable accurate measurement of S / IVOCs in the atmospheric environment, which is of significant practical importance for a deeper understanding of secondary atmospheric pollution processes. Summary of the Invention

[0006] To address the aforementioned shortcomings in existing technologies, this invention provides a chemical ionization mass spectrometry environmental humidity correction method based on functional group classification. This method acquires the mass spectrometry sensitivity response characteristics of different functional group species under different environmental humidity conditions and uses machine learning methods to automatically classify the humidity response curves, constructing a categorized humidity-mass spectrometry sensitivity data correction equation. This effectively reduces the uncertainty caused by humidity changes in the quantitative results of semi-volatile organic mass spectrometry, significantly improves the accuracy and stability of quantitative analysis under different humidity conditions, and solves the problem that existing technologies either do not effectively consider the influence of humidity or only use a single humidity correction method, resulting in insufficient applicability to multiple functional group species.

[0007] To achieve the aforementioned objectives, the technical solution adopted by this invention is: a chemical ionization mass spectrometry environmental humidity correction method based on functional group classification, comprising the following steps:

[0008] S1. Establish a humidity-mass spectrometry sensitivity dataset for species with different functional groups;

[0009] S2. Based on the humidity-mass spectrometry sensitivity dataset, establish a humidity-mass spectrometry sensitivity data correction equation based on machine learning functional group classification;

[0010] S3. Calculate the actual mass spectrometry sensitivity of the target species under the current environmental humidity conditions based on the humidity-mass spectrometry sensitivity data correction equation.

[0011] Furthermore, S1 includes the following sub-steps:

[0012] S11. Select compounds with multiple functional group types as dataset standards. The compound types include monoacids, diacids, monophenols, polyphenols, ketones, keto acids, ketone phenols, phenols, and nitro compounds.

[0013] S12. An environmental humidity control unit is set up in the standard gas calibration system. The environmental humidity control unit is connected to the output end of the permeameter or equivalent standard gas generator. It is used to adjust and control the humidity of the output gas without changing the stability of the target compound concentration.

[0014] S13. After adjusting the humidity in the humidity control unit, perform mass spectrometry sensitivity calibration for different compounds within the ambient absolute humidity range, normalize the mass spectrometry sensitivity data of all compounds, and establish a humidity-mass spectrometry sensitivity dataset based on the normalized mass spectrometry sensitivity of all compounds.

[0015] Furthermore: In S12, the environmental humidity control unit is equipped with a temperature and humidity sensor to calculate and control the absolute humidity of the gas entering the mass spectrometer, wherein the absolute humidity... The specific expression is:

[0016]

[0017] In the formula, Relative humidity, For temperature, This is the actual partial pressure of water vapor. The saturated water vapor pressure Standard atmospheric pressure.

[0018] Furthermore: In S13, mass spectrometry sensitivity calibration of different compounds was performed under the ambient absolute humidity range. The calibrated compounds... Mass spectrometry sensitivity data The specific expression is:

[0019]

[0020] In the formula, For compounds The mass spectrometry standard signal, For compounds The known gas concentration, Humidity conditions;

[0021] After normalization, the humidity-mass spectrometry sensitivity response curve Under absolute humidity conditions Below, compounds Normalized mass spectrometry sensitivity The specific expression is:

[0022]

[0023] In the formula, Humidity-mass spectrometry sensitivity response curve In a dry state Compounds detected by chemical ionization mass spectrometry Mass spectrometry sensitivity.

[0024] Furthermore, S2 includes the following sub-steps:

[0025] S21. Based on the humidity-mass spectrometry sensitivity dataset, the humidity-mass spectrometry sensitivity response curves are divided into multiple clusters using a hierarchical clustering method.

[0026] S22. Establish a humidity-mass spectrometry sensitivity correction equation based on the humidity-mass spectrometry sensitivity response curve of each cluster using machine learning methods.

[0027] Furthermore: S21 includes the following sub-steps:

[0028] S211. Calculate the similarity between the humidity-mass spectrometry sensitivity response curves, including calculating the similarity between the compound and the mass spectrometry sensitivity response curves. With compounds The distance between the humidity-mass spectrometry sensitivity response curves The specific expression is:

[0029]

[0030] In the formula, For compounds Normalized mass spectrometry sensitivity, The number of humidity condition features in the humidity-mass spectrometry sensitivity response curve;

[0031] S212. Calculate the centroid of the cluster. The centroid measures the concentration of data points within a cluster. Multiple clusters are generated through clustering. For each cluster... Its center of mass The specific expression is:

[0032]

[0033] In the formula, For clusters Each humidity-mass spectrometry sensitivity response curve Normalized mass spectrometry sensitivity, For clusters The number of data points in the data;

[0034] S213. Calculate the sum of squared errors for each cluster, and select the closest clusters to merge. For each cluster... Its sum of squared errors The specific expression is:

[0035]

[0036] In the formula, This is the norm symbol.

[0037] Furthermore: S22 specifically refers to:

[0038] Analyzing the scatter plots of the humidity-mass spectrometry sensitivity response curves for each cluster confirmed a significant nonlinear dependence between mass spectrometry sensitivity and humidity. Based on the specific characteristics of this relationship, a matching function form was selected from a mathematical function library as the fitting model. The specific expression of the fitting model is as follows:

[0039]

[0040] In the formula, For the absolute humidity of each cluster, For the normalized mass spectrometry sensitivity in each cluster, P1 to P n To fit the model Undetermined model parameters;

[0041] Based on the fitting model, a machine learning fitting algorithm was used to perform regression analysis on the normalized mass spectrometry sensitivity of all compounds. The undetermined model parameters were obtained, and the humidity-mass spectrometry sensitivity correction equation corresponding to the humidity-mass spectrometry sensitivity response curve for each cluster was acquired. This equation, based on the relationship between the mass spectrometry signal and sensitivity under known humidity conditions, performs corresponding humidity correction calculations. The humidity-corrected mass spectrometry sensitivity... The specific expression is:

[0042]

[0043] In the formula, This represents the measured mass spectrometry sensitivity. For clusters The fitting model is used for absolute humidity conditions. The relationship between mass spectrometry sensitivity and mass spectrometry sensitivity.

[0044] Furthermore: S3 specifically refers to:

[0045] Obtain the actual humidity data in the target environment, input it into the humidity-mass spectrometry sensitivity data correction equation, and calculate the actual mass spectrometry sensitivity of the target species under the current environmental humidity conditions.

[0046] The beneficial effects of this invention are as follows:

[0047] (1) The method of the present invention achieves system calibration of species with different functional groups in a wide humidity range by introducing a controllable humidification system into the calibration system.

[0048] (2) The method of the present invention normalizes the data of the humidity-mass spectrometry sensitivity response curve, so that the humidity response behavior of different compounds is comparable.

[0049] (3) The method of the present invention introduces machine learning classification to realize automatic classification of humidity response curves, avoids the subjectivity brought about by manually setting classification rules, and improves the objectivity and stability of classification results.

[0050] (4) The method of the present invention combines machine learning classification results with functional group features to reveal the humidity response law dominated by different functional groups.

[0051] (5) The classification humidity correction equation constructed by the method of the present invention is applicable to long-term online observation under complex humidity conditions in the field. Attached Figure Description

[0052] Figure 1 This is a flowchart of the environmental humidity correction method for chemical ionization mass spectrometry based on functional group classification according to the present invention.

[0053] Figure 2 This is a flowchart illustrating the technical process of the present invention.

[0054] Figure 3 A quantitative dataset of "humidity-sensitivity" for species with different functional groups in I-CIMS mass spectrometry.

[0055] Figure 4 This is a contour analysis diagram of the hierarchical clustering effect of the "humidity-mass spectrometry sensitivity" curve.

[0056] Figure 5 This is a graph showing the hierarchical clustering criterion values ​​of the "humidity-mass spectrometry sensitivity" curve.

[0057] Figure 6 A schematic diagram of the "humidity-mass spectrometry sensitivity" correction equation established based on machine learning functional group classification. Detailed Implementation

[0058] The specific embodiments of the present invention are described below to enable those skilled in the art to understand the present invention. However, it should be understood that the present invention is not limited to the scope of the specific embodiments. For those skilled in the art, various changes are obvious as long as they are within the spirit and scope of the present invention as defined and determined by the appended claims. All inventions utilizing the concept of the present invention are protected.

[0059] like Figure 1 As shown, in one embodiment of the present invention, the chemical ionization mass spectrometry environmental humidity correction method based on functional group classification includes the following steps:

[0060] S1. Establish a humidity-mass spectrometry sensitivity dataset for species with different functional groups;

[0061] S2. Based on the humidity-mass spectrometry sensitivity dataset, establish a humidity-mass spectrometry sensitivity data correction equation based on machine learning functional group classification;

[0062] S3. Calculate the actual mass spectrometry sensitivity of the target species under the current environmental humidity conditions based on the humidity-mass spectrometry sensitivity data correction equation.

[0063] S1 includes the following steps:

[0064] S11. Select compounds with multiple functional group types as dataset standards. The compound types include monoacids, diacids, monophenols, polyphenols, ketones, keto acids, ketone phenols, phenols, and nitro compounds.

[0065] In this embodiment, based on the excellent detection capability of I-CIMS technology for S / IVOCs containing functional groups such as ketones, hydroxyl groups, carboxyl groups, and nitro groups, this invention selected representative compounds as dataset standards. These compounds include, but are not limited to: monoacids, diacids, monophenols, polyphenols, ketones, keto acids, ketols, phenols, and nitro compounds. By selecting these compounds, this invention ensures the diversity and coverage of the dataset, effectively reflecting the mass spectrometry response characteristics of different functional group structures under various humidity conditions.

[0066] S12. An environmental humidity control unit is set up in the standard gas calibration system. The environmental humidity control unit is connected to the output end of the permeameter or equivalent standard gas generator. It is used to adjust and control the humidity of the output gas without changing the stability of the target compound concentration.

[0067] In this embodiment, an ambient humidity control unit is set in the standard gas calibration system. The ambient humidity control unit is connected to the output end of the permeameter or equivalent standard gas generator. It is used to adjust and control the humidity of the output gas without changing the concentration stability of the target compound, so as to realize the mass spectrometry signal measurement of compounds with different functional group structures under multiple humidity conditions.

[0068] S13. After adjusting the humidity in the humidity control unit, perform mass spectrometry sensitivity calibration for different compounds within the ambient absolute humidity range, normalize the mass spectrometry sensitivity data of all compounds, and establish a humidity-mass spectrometry sensitivity dataset based on the normalized mass spectrometry sensitivity of all compounds.

[0069] In S12, the environmental humidity control unit is equipped with a temperature and humidity sensor to calculate and control the absolute humidity of the gas entering the mass spectrometer. The absolute humidity... The specific expression is:

[0070]

[0071] In the formula, absolute humidity The unit is mmol / mol. Relative humidity, expressed as a percentage (%). For temperature, This is the actual partial pressure of water vapor, expressed in Pa. This is the saturated water vapor pressure, expressed in Pa. It is the standard atmospheric pressure, with a value of 101325 Pa.

[0072] During humidity control, the absolute humidity range of the output gas is maintained between 0 and 23.00 mmol / mol. Considering that the operating environment temperature of mass spectrometry equipment is generally 20°C, at this temperature, the humidity control unit can adjust the relative humidity range from 0% to 100%, and correspondingly adjust the absolute humidity range to maintain it between 0 and 23.00 mmol / mol, so as to ensure that the mass spectrometry signal is stable under different humidity conditions and accurately reflects the concentration of the target substance.

[0073] In S13, mass spectrometry sensitivity calibration of different compounds was performed under ambient absolute humidity conditions. The calibrated compounds... Mass spectrometry sensitivity data The specific expression is:

[0074]

[0075] In the formula, For compounds The mass spectrometry standard signal, For compounds The known gas concentration, Humidity conditions; in this embodiment, the compound mass spectrometry standard signal This refers to adjusting the mass spectrometry signal (such as ion intensity or ion count) of a compound to a ratio signal relative to the total reagent ions. This method reduces the influence of ion source effects, equipment performance fluctuations and other variables on the signal by comparing the signal of the analyte compound with the signal of the reagent ions.

[0076] To ensure comparability of response data under different humidity conditions, all measured humidity-mass spectrometry sensitivity data were normalized. The normalized humidity-mass spectrometry sensitivity response curves are then obtained. Under absolute humidity conditions Below, compounds Normalized mass spectrometry sensitivity The specific expression is:

[0077]

[0078] In the formula, Humidity-mass spectrometry sensitivity response curve In a dry state Compounds detected by chemical ionization mass spectrometry The mass spectrometry sensitivity was determined. Through this normalization process, the final humidity-mass spectrometry sensitivity dataset, as shown in Table 1, will provide a consistent data foundation for subsequent machine learning classification and humidity correction model development.

[0079] Table 1 Normalized Humidity-Mass Spectrometry Sensitivity Dataset

[0080]

[0081] S2 includes the following steps:

[0082] S21. Based on the humidity-mass spectrometry sensitivity dataset, the humidity-mass spectrometry sensitivity response curves are divided into multiple clusters using a hierarchical clustering method.

[0083] S22. Establish a humidity-mass spectrometry sensitivity correction equation based on the humidity-mass spectrometry sensitivity response curve of each cluster using machine learning methods.

[0084] In S21, various machine learning methods can be used to classify the humidity-mass spectrometry sensitivity response curve, such as hierarchical clustering, random forest, K-means clustering, and support vector machine (SVM). A humidity-mass spectrometry sensitivity data correction equation can be established based on the response features of each class.

[0085] S21 includes the following steps:

[0086] S211. Calculate the similarity between the humidity-mass spectrometry sensitivity response curves, including calculating the similarity between the compound and the mass spectrometry sensitivity response curves. With compounds The distance between the humidity-mass spectrometry sensitivity response curves The specific expression is:

[0087]

[0088] In the formula, For compounds Normalized mass spectrometry sensitivity, The number of humidity condition features in the humidity-mass spectrometry sensitivity response curve;

[0089] S212. Calculate the centroid of the cluster. The centroid measures the concentration of data points within a cluster. Multiple clusters are generated through clustering. For each cluster... Its center of mass The specific expression is:

[0090]

[0091] In the formula, For clusters Each humidity-mass spectrometry sensitivity response curve Normalized mass spectrometry sensitivity, For clusters The number of data points in the data;

[0092] S213. Calculate the sum of squared errors for each cluster, and select the closest clusters to merge. For each cluster... Its sum of squared errors The specific expression is:

[0093]

[0094] In the formula, This is the norm symbol.

[0095] In hierarchical clustering, the closest (most similar) clusters are selected for merging, and the change in the sum of squared errors (SSE) of the merged clusters is evaluated. The evaluation process primarily involves calculating the sum of squared errors (SSE) of the new clusters after merging to determine whether to proceed with the merge. The change in the sum of squared errors is measured when merging two clusters A and B. Calculate using the following formula:

[0096]

[0097] In the formula, This represents the sum of squared errors of the merged clusters. Let A be the sum of squared errors of cluster A. Let be the sum of squared errors for cluster B.

[0098] S22 specifically refers to:

[0099] Analyzing the scatter plots of the humidity-mass spectrometry sensitivity response curves for each cluster confirmed a significant nonlinear dependence between mass spectrometry sensitivity and humidity. Based on the specific characteristics of this relationship, a matching function form was selected from a mathematical function library as the fitting model. The specific expression of the fitting model is as follows:

[0100]

[0101] In the formula, For the absolute humidity of each cluster, For the normalized mass spectrometry sensitivity in each cluster, P1 to P n To fit the model Undetermined model parameters;

[0102] Based on the fitting model, a machine learning fitting algorithm is used to perform regression analysis on the normalized mass spectrometry sensitivity of all compounds, solving for and obtaining the undetermined model parameters of the fitting model. For example, if the relationship between sensitivity and humidity exhibits an "S"-shaped saturation change characteristic, the Boltzmann function can be used as the basic form of the fitting equation, and its general expression is as follows:

[0103]

[0104] In the formula, Initial sensitivity, For saturation sensitivity, Humidity at the inflection point The humidity change rate is then used. Subsequently, a machine learning fitting algorithm is employed to perform regression analysis on the original data pairs, solve for and obtain the specific parameters of the Boltzmann function, thereby establishing the final humidity-mass spectrometry sensitivity correction equation.

[0105] Obtain the humidity-mass spectrometry sensitivity correction equation corresponding to the humidity-mass spectrometry sensitivity response curve for each cluster. This equation performs corresponding humidity correction calculations based on the relationship between the mass spectrometry signal and sensitivity under known humidity conditions. The humidity-corrected mass spectrometry sensitivity is then calculated. The specific expression is:

[0106]

[0107] In the formula, This represents the measured mass spectrometry sensitivity. For clusters The fitting model is used for absolute humidity conditions. The relationship between mass spectrometry sensitivity and mass spectrometry sensitivity.

[0108] S3 specifically refers to:

[0109] Obtain the actual humidity data in the target environment, input it into the humidity-mass spectrometry sensitivity data correction equation, and calculate the actual mass spectrometry sensitivity of the target species under the current environmental humidity conditions.

[0110] To verify the effectiveness of the present invention, this embodiment provides the following experimental cases, and the technical flowchart is as follows. Figure 2 As shown:

[0111] Step 1: Establish a quantitative dataset of "humidity-mass spectrometry signal" for species with different functional groups:

[0112] (1) Data preparation:

[0113] This embodiment selects representative low- to medium-volatile organic compounds as standard substances, classifying them according to their main functional group types, to study the chemical ionization mass spectrometry response characteristics of iodine ion sources under various humidity conditions for different functional group structures. The selected compounds include:

[0114] Monophenolic compounds: m-cresol;

[0115] Polyphenolic compounds: 2,4-Dihydroxytoluene, 2,4,6-Trihydroxytoluene;

[0116] Monocarboxylic acid compounds: Formic acid, Propionic acid, Acrylic acid, Butyric acid, n-Pentanoic acid;

[0117] Dicarboxylic acid compounds: oxalic acid and malonic acid.

[0118] Ketoacids: Pinonic acid.

[0119] (2) Humidity control and mass spectrometry measurement:

[0120] An ambient humidity control unit is incorporated into the standard gas calibration system to precisely regulate the ambient humidity of the gas to be measured. This humidity control unit includes one standard sample drying carrier gas channel and one humidifying carrier gas channel. The humidifying carrier gas channel generates a water vapor-saturated gas flow through a water vapor generator. The gas flows from the drying and humidifying carrier gas channels are mixed in a mixing module according to a preset ratio and then output to the mass spectrometry injection system. This allows for continuous and adjustable control of the output gas ambient humidity while maintaining a constant target compound concentration.

[0121] Under different humidity conditions, the mixed standard gases were sequentially introduced into the I-CIMS for online measurement. Under stable operation, the mass spectrometry system continuously acquired the iodide ion addition signals of each target compound and recorded the mass spectrometry signal intensity under the corresponding humidity conditions. By adjusting the dry-to-wet flow rate ratio of the humidification carrier gas channel, multiple sets of mass spectrometry response data covering the low-to-high humidity range were obtained while maintaining a constant concentration. This data were used to characterize the sensitivity changes of compounds with different functional group types under different environmental humidity conditions.

[0122] (3) Humidity-mass spectrometry sensitivity data relationship and normalization:

[0123] After obtaining the mass spectrometry signals of each target compound under different environmental humidity conditions, the collected mass spectrometry data were preprocessed and sensitivity calculated. The sensitivity was defined as the intensity of the mass spectrometry signal produced by a unit concentration of the target compound under the corresponding humidity conditions. The mass spectrometry sensitivity data of each target compound under different humidity conditions were obtained by calculating the ratio of the mass spectrometry signal intensity to the corresponding calibrated concentration.

[0124] To eliminate the influence of initial sensitivity differences between different compounds on subsequent analyses, the sensitivity data were normalized. Specifically, the sensitivity under dry conditions or a selected reference humidity condition was used as a baseline value, and the ratio of the sensitivity under different humidity conditions to this baseline value was calculated to obtain the relative sensitivity response curves of each compound as a function of ambient humidity.

[0125] Through the above normalization process, the humidity response characteristics of different compounds are uniformly mapped to a dimensionless space, thereby weakening the influence of absolute sensitivity differences and highlighting the response patterns of compounds with different functional group types to changes in environmental humidity. This provides standardized input data for subsequent functional group-based cluster analysis and the establishment of humidity correction models. The quantitative "humidity-sensitivity" dataset of different functional group species in I-CIMS mass spectrometry constructed based on the above method is shown below. Figure 3 As shown, Figure 3 In China, normalized mass spectrometry sensitivity does not have a unified unit and is considered a dimensionless parameter.

[0126] Step 2: Establishment of the humidity-mass spectrometry signal correction equation based on machine learning functional group classification:

[0127] (1) Hierarchical clustering analysis of humidity response curves based on machine learning:

[0128] The normalized humidity-mass spectrometry sensitivity (HMS) response curves of different compounds obtained in Step 1 are used as input features to characterize the mass spectrometry response behavior of each target compound under varying environmental humidity conditions. Based on this, an unsupervised machine learning method is introduced to perform similarity analysis on the humidity response curves of different compounds, and hierarchical clustering is implemented based on distance metrics to classify compounds with similar humidity response characteristics into several candidate categories. The contour analysis results of the hierarchical clustering effect of the humidity-mass spectrometry sensitivity curves are shown below. Figure 4 As shown, this is used to evaluate the reasonableness and stability of the clustering results. Figure 4 In this context, distance is a dimensionless parameter.

[0129] (2) Determining the number of clusters and classifying functional group responses:

[0130] Based on the hierarchical clustering analysis, the optimal number of clusters was determined by analyzing the relationship between the clustering criterion values ​​and the number of clusters, thereby obtaining stable and physically meaningful functional group humidity response categories. The hierarchical clustering criterion value analysis results of the "humidity-mass spectrometry sensitivity" curve are as follows: Figure 5 As shown, with the increase of the number of clusters, the clustering criterion value generally shows a trend of first increasing and then stabilizing. When the number of clusters is 4, the clustering criterion value reaches a high level, indicating that there is good distinguishability between clusters and high consistency within clusters. However, when the number of clusters continues to increase, the criterion value improves only slightly, and in some cases even fluctuates, indicating that further subdivision does not significantly improve the clustering effect, but may even introduce the problem of oversegmentation. Therefore, considering the stability, interpretability, and physical rationality of functional group responses of the clustering effect, a cluster size of 4 is selected as the optimal classification result.

[0131] By analyzing the functional group characteristics of compounds in each cluster, these four cluster species can be divided into:

[0132] Small molecule acid compounds: These compounds are mainly represented by formic acid, with molecules containing no more than 5 atoms. Their mass spectrometry sensitivity initially increases and then decreases with increasing humidity.

[0133] Single-functional-group compounds: These compounds are characterized by containing a single carboxyl or acid group. They respond dramatically to changes in humidity, and their mass spectrometry sensitivity decreases significantly with increasing humidity. They mainly include m-cresol, propionic acid, acrylic acid, butyric acid, and n-valeric acid.

[0134] Compounds with multiple active functional groups: These compounds are characterized by having multiple carboxyl functional groups or one ketone group and one carboxyl group. Their mass spectrometry sensitivity shows a slow decreasing trend with increasing humidity. They mainly include oxalic acid, malonic acid and pinocembranoic acid.

[0135] Polyhydroxyphenolic compounds: These compounds are characterized by having multiple carboxyl groups and benzene ring structures. They have a very strong ability to bind to iodide ions and are almost unaffected by humidity. They mainly include 2,4-dihydroxytoluene and 2,4,6-trihydroxytoluene.

[0136] Based on the analysis results, different compounds were classified into corresponding functional group response categories, providing a classification basis for the subsequent construction of humidity correction models.

[0137] (3) Establishment of the humidity-mass spectrometry sensitivity correction equation based on functional group classification:

[0138] After classifying functional group humidity responses, a corresponding humidity-mass spectrometry sensitivity correction equation was established for each class of compounds with similar humidity response characteristics. This correction equation uses absolute ambient humidity as the independent variable and relative mass spectrometry sensitivity as the dependent variable to describe the changes in mass spectrometry response of this functional group species under different humidity conditions. The humidity-mass spectrometry sensitivity correction equation established based on machine learning functional group classification is shown below. Figure 6 As shown.

[0139] The results show that the humidity correction equation for small molecule acid compounds is an asymmetric double S-function, and the correlation R of the fitted curve is... 2 The value is 0.998. The humidity correction equation for single-functionalized compounds conforms to the Boltzmann function, and the correlation R of the fitted curve is 0.998. 2 The coefficient reached 0.98. The humidity correction equation for multi-functional compound compounds conforms to the Boltzmann function, and the correlation R of the fitted curve is [value missing]. 2 The value reached 0.856. The mass spectrometry sensitivity of polyhydroxyphenols does not change with humidity, therefore the correction factor is 1.

[0140] Step 3: Field application of environmental humidity correction method for chemical ionization mass spectrometry:

[0141] In actual atmospheric environmental field observations, chemical ionization mass spectrometry (CISMS) using an iodine ion source is used to perform online measurements of target species, simultaneously acquiring environmental humidity and corresponding mass spectrometry signal intensity data to obtain the raw mass spectrometry response of the target species under actual environmental conditions. Based on the humidity-mass spectrometry sensitivity correction equation established in step two using machine learning functional group classification, the target species is first assigned to the corresponding functional group humidity response category. Then, based on the measured environmental humidity conditions, the corresponding humidity-sensitivity correction coefficient for that category is found or calculated. Subsequently, the measured mass spectrometry signal is corrected for humidity effects using the correction coefficient, thereby eliminating the influence of environmental humidity changes on mass spectrometry sensitivity and obtaining the corrected target species concentration under actual environmental humidity conditions. This achieves accurate online quantification of semi-volatile organic compounds under complex humidity conditions in the field.

[0142] In the description of this invention, the above are merely preferred embodiments and are not intended to limit the scope of protection of this invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this invention should be included within the scope of protection of this invention.

Claims

1. A chemical ionization mass spectrometry environmental humidity correction method based on functional group classification, characterized in that, Includes the following steps: S1. Establish a humidity-mass spectrometry sensitivity dataset for species with different functional groups; S2. Based on the humidity-mass spectrometry sensitivity dataset, establish a humidity-mass spectrometry sensitivity data correction equation based on machine learning functional group classification; S3. Calculate the actual mass spectrometry sensitivity of the target species under the current environmental humidity conditions based on the humidity-mass spectrometry sensitivity data correction equation. S1 includes the following steps: S11. Select compounds with multiple functional group types as dataset standards. The compound types include monoacids, diacids, monophenols, polyphenols, ketones, keto acids, ketone phenols, phenols, and nitro compounds. S12. An environmental humidity control unit is set up in the standard gas calibration system. The environmental humidity control unit is connected to the output end of the permeameter or equivalent standard gas generator. It is used to adjust and control the humidity of the output gas without changing the stability of the target compound concentration. S13. After adjusting the humidity in the humidity control unit, perform mass spectrometry sensitivity calibration of different compounds within the absolute humidity range of the environment, normalize the mass spectrometry sensitivity data of all compounds, and establish a humidity-mass spectrometry sensitivity dataset based on the normalized mass spectrometry sensitivity of all compounds. S2 includes the following steps: S21. Based on the humidity-mass spectrometry sensitivity dataset, the humidity-mass spectrometry sensitivity response curves are divided into multiple clusters using a hierarchical clustering method. S22. Establish a humidity-mass spectrometry sensitivity correction equation based on the humidity-mass spectrometry sensitivity response curve of each cluster using machine learning methods. S21 includes the following steps: S211. Calculate the similarity between the humidity-mass spectrometry sensitivity response curves, including calculating the similarity between the compound and the mass spectrometry sensitivity response curves. With compounds The distance between the humidity-mass spectrometry sensitivity response curves The specific expression is: In the formula, For compounds Normalized mass spectrometry sensitivity, The number of humidity condition features in the humidity-mass spectrometry sensitivity response curve; S212. Calculate the centroid of the cluster. The centroid measures the concentration of data points within a cluster. Multiple clusters are generated through clustering. For each cluster... Its center of mass The specific expression is: In the formula, For clusters Each humidity-mass spectrometry sensitivity response curve Normalized mass spectrometry sensitivity, For clusters The number of data points in the data; S213. Calculate the sum of squared errors for each cluster, and select the closest clusters to merge. For each cluster... Its sum of squared errors The specific expression is: In the formula, Norm symbol; S22 specifically refers to: Analyzing the scatter plots of the humidity-mass spectrometry sensitivity response curves for each cluster confirmed a significant nonlinear dependence between mass spectrometry sensitivity and humidity. Based on the specific characteristics of this relationship, a matching function form was selected from a mathematical function library as the fitting model. The specific expression of the fitting model is as follows: In the formula, For the absolute humidity of each cluster, For the normalized mass spectrometry sensitivity in each cluster, P1 to P n To fit the model Undetermined model parameters; Based on the fitting model, a machine learning fitting algorithm was used to perform regression analysis on the normalized mass spectrometry sensitivity of all compounds. The undetermined model parameters were obtained, and the humidity-mass spectrometry sensitivity correction equation corresponding to the humidity-mass spectrometry sensitivity response curve for each cluster was acquired. This equation, based on the relationship between the mass spectrometry signal and sensitivity under known humidity conditions, performs corresponding humidity correction calculations. The humidity-corrected mass spectrometry sensitivity... The specific expression is: In the formula, This represents the measured mass spectrometry sensitivity. For clusters The fitting model is used for absolute humidity conditions. The relationship between mass spectrometry sensitivity and mass spectrometry sensitivity.

2. The method for environmental humidity correction of chemical ionization mass spectrometry based on functional group classification according to claim 1, characterized in that, In S12, the environmental humidity control unit is equipped with a temperature and humidity sensor to calculate and control the absolute humidity of the gas entering the mass spectrometer. The absolute humidity... The specific expression is: In the formula, Relative humidity, For temperature, This is the actual partial pressure of water vapor. The saturated water vapor pressure Standard atmospheric pressure.

3. The method for environmental humidity correction of chemical ionization mass spectrometry based on functional group classification according to claim 2, characterized in that, In S13, mass spectrometry sensitivity calibration of different compounds was performed under ambient absolute humidity conditions. The calibrated compounds... Mass spectrometry sensitivity data The specific expression is: In the formula, For compounds The mass spectrometry standard signal, For compounds The known gas concentration, Humidity conditions; After normalization, the humidity-mass spectrometry sensitivity response curve Under absolute humidity conditions Below, compounds Normalized mass spectrometry sensitivity The specific expression is: In the formula, Humidity-mass spectrometry sensitivity response curve In a dry state Compounds detected by chemical ionization mass spectrometry Mass spectrometry sensitivity.

4. The method for environmental humidity correction of chemical ionization mass spectrometry based on functional group classification according to claim 1, characterized in that, S3 specifically refers to: Obtain the actual humidity data in the target environment, input it into the humidity-mass spectrometry sensitivity data correction equation, and calculate the actual mass spectrometry sensitivity of the target species under the current environmental humidity conditions.

Images