An array material and method for detecting VOCs components in a gas

Abstract

The application discloses an array material and a method for detecting VOCs components in gas, and the array material is obtained through the following steps: step 1: each dye is respectively dissolved in an ethanol solution to obtain a dye reagent; the concentration of the ethanol solution is 50%; step 2: polyvinylidene fluoride film is selected as a substrate, the dye reagent obtained in step 1 is separately and individually added on the polyvinylidene fluoride film to form a plurality of dye detection points and form a detection array; then a sealing film is covered above and below the polyvinylidene fluoride film, the polyvinylidene fluoride film is sealed, and an air inlet hole and an air outlet hole are respectively formed above and below the sealing film and correspond to the positions of each dye detection point; nitrogen is slowly blown to the polyvinylidene fluoride film to obtain the array material. According to the method, the array material constructed in the application can clearly distinguish a plurality of volatile organic compounds through visual results or pattern recognition, and the recognition accuracy is high.

Description

Technical Field

[0001] This invention relates to the field of chemical detection technology, specifically to an array material and method for detecting VOCs components in a gas. Background Technology

[0002] Volatile organic compounds (VOCs), as potential biomarkers for various diseases, are typically released through intracellular metabolic processes, circulate in the blood to the lungs, and are eventually exhaled via gas / blood exchange. Different phenotypes of lung diseases produce identifiable VOC compositions. Therefore, breath analysis based on VOC compositional differences is gradually becoming an attractive, rapid, and non-invasive method for screening lung diseases. Existing VOC sensing technologies include gas chromatography-mass spectrometry (GC-MS), ion mass spectrometry, laser absorption spectroscopy, infrared spectroscopy, polymer-coated surface acoustic wave sensors, and monolayer-coated quartz crystal microbalance sensors. These methods mostly require sophisticated and expensive large-scale instruments, complex and time-consuming sample pretreatment methods, and specialized technical operators, limiting the widespread application of VOC detection.

[0003] In contrast, sensor arrays based on broad cross-response sensors offer a more ideal non-invasive screening method for rapid disease screening because their detection results can be obtained within minutes. They also offer advantages such as ease of preparation, high efficiency, portability, and result visualization. However, most traditional colorimetric arrays are composed of conductive polymers, solvochromic dyes or fluorophore-functionalized polymers, and metal oxides. Therefore, they typically rely on weak and low-specificity interactions, primarily van der Waals forces and physisorption between the sensor and the analyte. Consequently, previous arrays often suffer from poor selectivity, high response temperatures, and low stability to humidity changes or other disturbances. Summary of the Invention

[0004] In view of the above-mentioned shortcomings of the existing technology, the purpose of this invention is to provide an array material and method for detecting VOCs components in gas, so as to solve the problems of poor selectivity, high response temperature and low stability to humidity changes or other interferences in the traditional colorimetric array.

[0005] To solve the above-mentioned technical problems, the present invention adopts the following technical solution:

[0006] An array material for detecting VOCs components in a gas, said array material being obtained through the following steps:

[0007] Step 1: Dissolve each dye separately in an ethanol solution to obtain dye reagents; the concentration of the ethanol solution is 50%, and the concentration of the dye in the ethanol solution is 1 mg / 2 mL; the dyes include bromooxylenol blue, phenol red, rhodamine B, bromophenol blue, nitrorazine yellow, bromothymol blue, alizarin, metanil yellow, bromophenol red, bromocresol purpur, acridine orange, cresol red, Congo red, thymol blue, and brilliant yellow;

[0008] Step 2: Select a polyvinylidene fluoride (PVDF) film as a substrate, and individually drop the dye reagent obtained in Step 1 onto the PVDF film to form multiple dye detection points, thus forming a detection array; then cover the PVDF film with a sealing film above and below it to seal it, and make an air inlet and an air outlet on the sealing film above and below each dye detection point, respectively; slowly blow nitrogen gas into the PVDF film to obtain the array material.

[0009] Preferably, the dye includes pretreated dye and untreated dye. The pretreated dye is obtained by dissolving each dye separately in a 100 mM / L NaOH solution for modification, resulting in multiple pretreated dyes. The untreated dye is dye that is directly added to an ethanol solution.

[0010] Preferably, the pretreated dyes include bromophenol blue, phenol red, rhodamine B, bromophenol blue, nitromethamine yellow, bromothymol blue, and alizarin; the pretreated dyes include m-amine yellow, bromophenol red, bromocresol violet, acridine orange, cresol red, Congo red, thymol blue, brilliant yellow, and phenol red.

[0011] Preferably, a modifying material is also added to each dye; the modifying material is a mixture of graphene aerogel and copper metal-organic framework, with a mass ratio of graphene aerogel to copper metal-organic framework of 1:1. Specifically, GA and Cu-MOF are completely dissolved in 4 mL of 50% ethanol and sonicated for 1 hour.

[0012] Preferably, the graphene aerogel is obtained by the following method: adding graphene oxide to distilled water and ultrasonically treating it to form a uniform graphene oxide dispersion; then transferring the graphene oxide dispersion to a reactor and heating it at 180°C for 12 hours to obtain a black product; after cooling to room temperature, centrifuging and cleaning are performed to obtain the graphene aerogel.

[0013] Preferably, the copper metal-organic framework is obtained by the following method:

[0014] S1: Dissolve polyvinylpyrrolidone in DMF and ethanol, and label it as reagent 1; wherein the volume ratio of DMF to ethanol is 1:1, and the concentration of polyvinylpyrrolidone in the mixed solution is 0.025 g / ml;

[0015] S2: Dissolve copper nitrate trihydrate and 2-aminoterephthalic acid in DMF and label it as reagent 2; wherein the concentration of copper nitrate trihydrate in DMF is 0.025 mmol / ml and the concentration of 2-aminoterephthalic acid in DMF is 0.025 mmol / ml.

[0016] S3: Mix reagent 1 and reagent 2 under ultrasonic treatment for 30 min, transfer to a high-pressure reactor, and react at 100℃ for 8 h to obtain the product;

[0017] S4: The product obtained in S3 was washed with ethanol several times, and the precipitate was collected by centrifugation. After vacuum drying, a green powder was obtained, which was soluble in ethanol.

[0018] The present invention also provides a method for detecting VOCs components in a gas, specifically comprising the following steps:

[0019] Step 1: Place the array material in a transparent detection box, then scan the array and acquire images before the reaction, and obtain the RGB values ​​before the reaction;

[0020] Step 2: Select dry nitrogen as a control sample, pump dry nitrogen into the detection box and circulate it slowly to scan the array and obtain an image as the blank array image, and obtain the RGB values ​​of the blank array;

[0021] Step 3: Pump the gas to be detected into the detection box and circulate it slowly. Then scan the array and acquire the image after the reaction as the image of the array after the reaction, and obtain the RGB value of the array after the reaction.

[0022] Step 4: Using the RGB values ​​of the blank array as initial values, subtract the RGB values ​​before the reaction from the RGB values ​​of the array after the reaction to obtain ΔR, ΔG, and ΔB;

[0023] Step 5: Take the absolute value of the ΔRGB value to construct the fingerprint spectrum, and use the difference vector for statistical analysis to obtain the VOCs composition in the gas to be detected.

[0024] Preferably, the gas to be detected is either the gas exhaled by a human body or the gas to be detected.

[0025] Preferably, in step 3, the gas to be detected is pumped into the detection box and reacted for 15 to 30 minutes before scanning.

[0026] Preferably, the VOCs components include toluene, acetic anhydride, propionic acid, benzene, methacrolein, formic acid, acetonitrile, n-pentane, butyric acid, n-pentanal, formaldehyde, acetaldehyde, n-propanol, butyl acetate, acetone, hexadecane, 4-methyloctane, butyraldehyde, n-hexane, n-butylamine, 2,5-dimethylfuran, n-heptane, 1,2,4-trimethylbenzene, benzyl alcohol, nonanal, methyl phenylacetate, hexylcyclohexane, and isoprene.

[0027] Compared with the prior art, the present invention has the following beneficial effects:

[0028] 1. Based on chemically diverse cross-responsive dyes, this invention constructs a colorimetric sensing array to encompass various strong chemical interactions, aiming to enhance the interaction types and forces in traditional colorimetric arrays. It can successfully distinguish and identify 28 VOCs components, providing material reserves for the cross-responsive sensitive points of composite arrays.

[0029] 2. The method described in this invention uses the array material constructed in this invention. Through visualization results or pattern recognition, it can clearly distinguish a variety of volatile organic compounds with high recognition accuracy. Attached Figure Description

[0030] Figure 1 The UV-Vis absorption spectra are those of 16 dyes reacted with 3 self-made volatile analytical samples.

[0031] Figure 2 This is the encoding of the array image and each dye dot of the present invention.

[0032] Figure 3 This is a graph showing the optimization of array detection time.

[0033] Figure 4 Fingerprints (300 μM) of 28 volatile organic molecular markers.

[0034] Figure 5 Principal component cumulative plot.

[0035] Figure 6 The results show the identification of 28 VOCs under the first three principal component analyses.

[0036] Figure 7 Results of distinguishing 28 VOCs under HCA.

[0037] Figure 8 Results of distinguishing 28 VOCs under LDA.

[0038] Figure 9 Transmission electron microscopy (TEM) image and elemental analysis of GA / Cu MOF.

[0039] Figure 10 The UV-Vis absorption spectra, ΔRGB values, and difference spectra of dyes AP1~AP4, AP9~AP12 after doping with GA / Cu MOF are shown.

[0040] Figure 11 The UV-Vis absorption spectra, ΔRGB values, and difference spectra of dyes AP5~AP8, AP13~AP16 after doping with GA / Cu MOF are shown.

[0041] Figure 12 To optimize the array detection time after doping.

[0042] Figure 13 Fingerprints of 28 volatile organic molecules (A: 100 μM, B: 300 μM).

[0043] Figure 14 Results of PCA differentiation of 28 VOCs (100 μM and 300 μM).

[0044] Figure 15 Results of distinguishing 28 VOCs (100 μM and 300 μM) under HCA.

[0045] Figure 16 Results of distinguishing 28 VOCs (100 μM and 300 μM) under LDA. Detailed Implementation

[0046] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0047] I. An array material for detecting VOCs components in gas

[0048] Most traditional colorimetric arrays are composed of conductive polymers, solvochromic dyes or fluorophore-functionalized polymers, and metal oxides. Therefore, they typically rely on weak and low-specificity interactions, primarily van der Waals forces and physisorption between the sensor and the analyte. Consequently, traditional colorimetric arrays often suffer from poor selectivity, high response temperatures, and low stability to humidity changes or other interferences.

[0049] This invention addresses the technical limitations of traditional colorimetric arrays by constructing novel colorimetric arrays to encompass a wider range of stronger interactions, such as electrostatic interactions between ions, Brønsted acid-base interactions, ligand coordination, hydrogen bonding, charge transfer, π-π molecular complexation, and dipole or multipolar interactions. Therefore, this invention utilizes chemically diverse cross-responsive dyes to construct a colorimetric sensing array material that encompasses various strong chemical interactions. This aims to enhance the types and forces of interactions in traditional colorimetric arrays and provide a material reserve for the cross-responsive sensitive points of composite arrays.

[0050] 1. Construction of array materials

[0051] Step 1: Dissolve each dye separately in a 50% (w / w) ethanol solution to obtain multiple independent dye reagents, each with a concentration of 1 mg / 2 ml. The dyes include bromophenollenol Blue, phenol red, rhodamine B, bromophenol blue, nitrazine yellow, bromothymol blue, alizarin, metanil yellow, bromophenol red, bromocresol purpur, acridine orange, cresol red, Congo red, thymol blue, and brilliant yellow.

[0052] Some of the dyes were pretreated, meaning they were first dissolved in a 100 mM / L NaOH solution for modification, aiming to neutralize the VOCs and alter the dye properties. The remaining untreated dyes were directly dissolved in a 50% ethanol solution. The specific dyes used are shown in the table below:

[0053] Table 1

[0054] AP# dye AP# dye 1 Bromoxylenol Blue + NaOH 9 Cresol Red 2 Phenol red + NaOH 10 Nitrazine Yellow + NaOH 3 Rhodamine B + NaOH 11 Phenol Red 4 Metanil Yellow 12 Congo Red 5 Bromophenol Red 13 Thymol Blue 6 Bromophenol Blue + NaOH 14 Bromothymol Blue + NaOH 7 Bromkresolpurpur 15 Brilliant Yellow 8 Acridine Orange 16 Alizarin + NaOH

[0055] Step 2: Select polyvinylidene fluoride (PVDF) film as the substrate. Using a glass capillary tube, apply the dye reagent from Step 1 evenly to the PVDF film until each dot expands in diameter, forming multiple dye detection points, ultimately creating a detection array on the PVDF. PVDF is a hydrophobic substrate, which reduces humidity interference during detection. This is a traditional method for constructing colorimetric sensor arrays using chemically responsive dyes. The array is manually dotted using a capillary glass tube. The dotted array is then placed in a petri dish and sealed with plastic wrap. Two small holes (approximately 5 mm in diameter) are then made symmetrically on the sealing film. Nitrogen gas is slowly purged into the petri dish. After purging, the petri dish is sealed and protected from light, then placed in a dry, cool place to mature before subsequent sensing detection.

[0056] 2. Preparation of volatile organic compound gases

[0057] Volatile organic compounds (VOCs) have small molecular weights, low density, and low boiling points, allowing them to transform into a gaseous state at room temperature or under mild heating. A 0.1L gas bag is filled with 0.05L of nitrogen and a certain volume (calculated using Equation 1) of liquid VOCs. It is then incubated in a 37℃ oven for 2 hours until all the liquid has transformed into a gas. Nitrogen is then added again until the gas bag inflates. All gases are produced using… PVF air bag storage and utilization The PVF hose was used for transfer. A total of 28 volatile organic compounds were used in this experiment to evaluate the sensor array's discrimination and recognition performance; the gas information is shown in Table 2.

[0058] V=(V1*c*M) / ρ (Equation 1)

[0059] Where V1 is the gas bag volume of 0.1L, and c, M, and ρ are the required VOC concentration, molecular mass, and density, respectively.

[0060] Table 2

[0061]

[0062]

[0063]

[0064] In practical implementation, the relevant mixed VOCs formulations are as follows:

[0065] HVOS 1: Toluene (10mM), Acetic anhydride (10mM), Propionic acid (10mM), Benzene (10mM), Methacrolein (10mM), Acetone (10mM), Hexadecane (10mM), 2,5-Dimethylfuran (10mM), n-Heptane (10mM), Hexylcyclohexane (10mM);

[0066] HVOS 2: hexadecane (25mM), 4-methyloctane (25mM), butyraldehyde (25mM), n-hexane (25mM), n-butylamine (25mM), acetaldehyde (25mM), n-propanol (25mM), butyl acetate (25mM), butyric acid (25mM), formaldehyde (25mM);

[0067] HVOS 3: Toluene (50mM), Acetic anhydride (50mM), Formic acid propionate (50mM), Acetonitrile (50mM), n-Pentane (50mM), Butyric acid (50mM), n-Pentanol n-Butylamine (50mM), 2,5-Dimethylfuran (50mM), n-Heptane (50mM);

[0068] HVOS 4: 1,2,4-Trimethylbenzene (50mM), benzyl alcohol (50mM), nonanal (50mM), methyl phenylacetate (50mM), hexylcyclohexane (50mM), isoprene (50mM), n-propanol (50mM), butyl acetate (50mM), acetone (50mM), hexadecane (50mM).

[0069] 4. Synthesis and characterization of modified materials

[0070] The present invention also considers adding modifying materials to the dye to improve the response of the dye reagent on the array to VOCs components in the mixed gas.

[0071] Synthesis of GA (graphite aerogel): 45 mg of GO (graphene oxide) was dissolved in 30 mL of deionized water and sonicated for 1 hour to obtain a uniform dispersion. The GO dispersion was transferred to a polytetrafluoroethylene-lined autoclave and heated at 180 °C for 12 hours. Subsequently, the resulting product was centrifuged at 11,000 rpm for 15 minutes to collect the black precipitate, which was then washed three times with deionized water. After freeze-drying for 12 hours, porous GA was obtained.

[0072] Synthesis of Cu-MOF: 0.8 g PVP was dissolved in a mixture of 16 mL DMF and 16 mL ethanol. 0.0968 g CuNO3·3H2O and 0.0217 g NH2-BDC were dissolved in 16 mL DMF, then mixed with the PVP solution and sonicated for 30 min. The final mixture was then transferred to a polytetrafluoroethylene-lined autoclave and heated at 100 °C for 8 h. The green precipitate was collected according to the centrifugation parameters for “GA synthesis” and washed three times with ethanol. Finally, the pure green precipitate was vacuum dried at 60 °C for 6 h to obtain Cu-MOF.

[0073] GA / Cu MOF: The GA / Cu-MOF complex consists of 1 mg GA and 1 mg Cu-MOF, which are completely dissolved in 4 mL of 50% ethanol and sonicated for 1 hour.

[0074] II. A method for detecting VOCs components in gas

[0075] Step 1: Place the array material in a transparent detection box, then scan the array and acquire images before the reaction, and obtain the RGB values ​​before the reaction;

[0076] Step 2: Select dry nitrogen as a control sample, pump dry nitrogen into the detection box and circulate it slowly to scan the array and obtain an image as the blank array image, and obtain the RGB values ​​of the blank array;

[0077] Step 3: Pump the gas to be detected into the detection box and circulate it slowly. Then scan the array and acquire the image after the reaction as the image of the array after the reaction, and obtain the RGB value of the array after the reaction. The gas to be detected is pumped into the detection box and reacted for 30 minutes before scanning.

[0078] Step 4: Using the RGB values ​​of the blank array as initial values, subtract the RGB values ​​before the reaction from the RGB values ​​of the array after the reaction to obtain ΔR, ΔG, and ΔB;

[0079] Step 5: Take the absolute value of the ΔRGB value to construct the fingerprint spectrum, and use the difference vector for statistical analysis to obtain the VOCs composition in the gas to be detected.

[0080] The gas to be detected is either collected human exhaled gas or a collected gas to be tested. The method described in this invention can be used to detect the presence of VOCs in human exhaled gas, providing data for subsequent disease research, and can also be used to detect VOCs in other collected gases.

[0081] The VOCs components include toluene, acetic anhydride, propionic acid, benzene, methacrolein, formic acid, acetonitrile, n-pentane, butyric acid, n-pentanal, formaldehyde, acetaldehyde, n-propanol, butyl acetate, acetone, hexadecane, 4-methyloctane, butyraldehyde, n-hexane, n-butylamine, 2,5-dimethylfuran, n-heptane, 1,2,4-trimethylbenzene, benzyl alcohol, nonanal, methyl phenylacetate, hexylcyclohexane, and isoprene.

[0082] 1. Data extraction and calculation of VOCs samples before and after reaction

[0083] The array material described in this invention is placed in a transparent chamber. A smartphone is fixed above the transparent chamber for photography (without filters). The array is then scanned and images are acquired before and after the analyte exposure reaction. During the reaction, the analyte gas circulates slowly in the system under the action of a peristaltic pump. Dry nitrogen is used as a control. Multiple parallel experiments are performed for each analyte. After the gas is introduced, the gas circulates within the reaction chamber to ensure sufficient contact between the sensitive material and the gas, and images of the array are acquired after the reaction. Using the RGB values ​​of the blank array as initial RGB values, the RGB values ​​after the reaction are subtracted from the RGB values ​​before the reaction to obtain ΔR, ΔG, and ΔB. The array contains 16 array points, i.e., a total of 48 dimensions (16*3) of data. The absolute values ​​of the ΔRGB values ​​are used to construct a fingerprint spectrum, and the difference vector will be used for subsequent pattern recognition statistical analysis.

[0084] The optimization of response time in the experiment involved the calculation of Euclidean distance change (ED). The Euclidean distance evaluation of the array points involved in the colorimetric array was performed according to Equation 2.

[0085]

[0086] 2. Statistical Analysis

[0087] Deep data mining can extract valuable data from complex output data and build corresponding models. Currently, pattern recognition analysis is being applied to deep mining of sensor array-based data. The main methods of pattern recognition analysis are hierarchical clustering analysis (HCA), principal component analysis (PCA), and linear discriminant analysis (LDA). PCA reduces dimensionality by utilizing data characteristics to obtain orthogonal correlations between principal components, which can reduce background noise to some extent. However, PCA dimensionality reduction may result in the loss of important information. The basic idea of ​​HCA is to first treat n target variables as one class; then calculate the statistical distance between each pair of classes, and cluster the two targets with high similarity, thus obtaining n-1 classes; this process continues, and each measurement of statistical distance reduces one class of targets until all targets can be classified into one class. HCA, by judging the similarity of target sample items, is helpful in discovering the hierarchy and belonging relationships of different targets. LDA, by establishing a discriminant function, attempts to project the labeled training dataset onto a straight line, making the projected points of samples with the same label as close as possible and the projected points of samples with different labels as far apart as possible. During classification, new sample projections are introduced onto the model line, and the label attribute of the new sample is determined based on the position of its projection point. LDA uses supervised learning to measure sample discrepancies, accurately reflecting the differences between samples. A set of criteria is designed from these samples so that any sample taken from this category can be classified according to this set of criteria. Discriminant analysis typically involves establishing a discriminant function, which is then used for discrimination. To accurately evaluate the array's ability to recognize the analytes, all pattern recognition analyses were performed using SPSS 22.0 software.

[0088] 3. Construction of chemically responsive dye liquid arrays and detection of mixed marker gases

[0089] The prerequisite for constructing a solid paper-based array is that traditional dyes can exhibit cross-response to mixed gaseous markers. After extensive screening, 16 chemically responsive dyes were selected, and some of them were modified. Table 1 lists the chemically responsive dyes used for each sensitive array point (AP). The gaseous markers were detected using interionic electrostatic interactions, Brønsted acid-base interactions, ligand coordination, hydrogen bonding, charge transfer, π-π molecular complexation, and dipole or multipolar interactions between the dye molecules and the analytes. Figure 1The UV-Vis spectra of each chemical dye and the homemade volatile analytical sample (HVAS) after reaction with the chemically responsive dye in liquid are shown. The composition and content of HVAS are described above. Bromophenol Blue, Phenol Red, Nitrazine Yellow, Brilliant Yellow, and Alizarin are all pH indicators that can identify acidic or alkaline substances in VOCs. However, due to differences in color change range and coordination ability, they can produce different changes with various components in VOCs, thus affecting the UV-Vis absorption spectrum. Rhodamine B can specifically adsorb nonpolar organic compounds in VOCs, producing color changes. Metanil Yellow can react with oxidizing substances and polar organic compounds. Bromophenol Red can coordinate with negatively charged organic compounds in VOCs, producing color changes. Bromoxylenol Blue has weak oxidizing properties and can react with reducing targets in VOCs. Bromkresolpurpur can specifically identify dihydroxy acids in VOCs. The dimethylamino group on Acridine Orange can sensitively identify components with polar groups in VOCs and can also detect deoxyribonucleic acid in human exhaled breath. Cresol Red can react with amines to detect specific components in VOCs. Congo... Red dyes can specifically detect thiocyanate and thiophosphate in VOCs, and also respond to the polarity and pH of VOCs. Thymol Blue and Bromothymol Blue can specifically detect changes in the recognition of long-chain alkanes and polar groups in VOCs. Sixteen chemically responsive dyes produced obvious visual color changes and UV-Vis absorption spectral shifts after reacting with three self-prepared volatile analysis samples. This invention concludes that the screening of chemically responsive liquid dyes is successful.

[0090] A rectangular array substrate was fabricated using polyvinylidene fluoride (PVDF) film. Reagent was drawn using a 0.3 mm glass capillary and uniformly dotted onto the paper substrate until the diameter of each dot expanded. PVDF is a hydrophobic substrate, which can reduce humidity interference during detection, unlike the construction of traditional chemically responsive dye-based colorimetric sensor arrays. Figure 2 The image shows the paper-based array prepared according to this invention. Poor contact between the spotting droplets and the hydrophobic substrate can be observed, and some dyes fail to diffuse uniformly on the PVDF film. During the experimental fabrication of the paper-based array, due to imperfect liquid diffusion, excessive force can easily puncture the PVDF film, resulting in a low formation rate. Since the array detects human exhaled gas, the response time of the array was optimized at 37°C using toluene and tetrahydrofuran. Figure 3As shown, the array response reaches its maximum value and then tends to level off after 30 minutes; this time is selected as the array's response time for subsequent VOCs detection.

[0091] 4. VOCs Pattern Recognition Based on Difference Vector

[0092] First, visual fingerprints of 28 VOCs were collected. For example... Figure 4 As shown, the absolute values ​​of the array's ΔRGB values ​​are used to construct a fingerprint spectrum by performing simple digital subtraction on the images before and after the array exposure reaction. The RGB fingerprint spectrum, as a visualization result, will be very beneficial for applications such as intuitive differentiation and identification. Based on the visualized spectrum, these VOCs can be distinguished even with the naked eye. However, as can be seen from the figure, the overall response of the array to various analytes is generally weak, which may be due to the weak coordination between alkanes, alkenes, and alcohols and the array. Although the array described in this invention is specifically designed to encompass a wide variety of interactions, most analytes are either extremely inert or have low volatility, and these inactive analytes can only elicit low and unique responses on the array. More accurate statistical analysis methods, such as hierarchical cluster analysis (HCA), principal component analysis (PCA), and linear discriminant analysis (LDA), are needed to analyze the difference vector.

[0093] (1) Principal Component Analysis (PCA)

[0094] PCA is a dimensionality reduction technique that reduces dimensionality by removing redundant data, thereby assessing the contribution of each component to the experimental results. Figure 5 The cumulative contribution of each principal component to the data analysis is shown. A three-dimensional scatter plot can be used to display the scores of each substance on PCA1, PCA2, and PCA3. The first three principal components only provide 49.61% of the effective information for sensor array performance evaluation. Figure 6 The data showed multiple overlaps, making it impossible to identify all 28 VOCs. This is because the Kaiser-Meyer-Olkin (KMO) value of the data model was only 0.4446 (<0.750), indicating a low correlation between the data and each principal component. Principal component analysis would miss crucial data information, and PCA could not correctly distinguish and identify all samples. Other pattern recognition methods are needed to differentiate and identify VOCs.

[0095] (2) Hierarchical Cluster Analysis (HCA)

[0096] Hierarchical Cluster Analysis (HCA) is an unbiased clustering method well-suited for handling high-dimensional data. To distinguish between 28 VOCs, HCA with squared Euclidean distance was used to measure the similarity between samples, merging them according to similarity until all samples were interconnected. The dendrogram's connectivity and Euclidean distance respectively explained the similarity and degree of similarity between samples. In principle, in a hierarchical clustering tree, branches representing the same VOC will cluster into a single cluster. For samples with unknown composition, cluster analysis can also indicate which substances the unknown sample is most similar to. In parallel experiments, a total of 84 experiments were conducted on the 28 VOCs. Based on the resulting difference vectors, hierarchical cluster analysis showed clear and compact clustering results. Figure 7 The HCA results for 28 VOCs were presented. Six experimental data points from four substances—n-heptane, 1,2,4-trimethylbenzene, benzyl alcohol, and nonanal—were incorrectly grouped together, resulting in an accuracy of 92.9%. This indicates a higher error rate when introducing new samples. The HCA results demonstrate that the array has some ability to distinguish and identify VOCs, but further improvement is needed.

[0097] (3) Linear Discriminant Analysis (LDA)

[0098] LDA was used to analyze the ΔR, ΔG, and ΔB data of 28 VOCs reacting with a solid paper-based array. A two-dimensional scatter plot was plotted using the scores of the data on the first and second linear discriminant functions as the X and Y axes. Figure 8 Compared to principal component analysis, linear discriminant analysis primarily aims to construct optimal orthogonal dimensions to maximize the display of differences between analyte categories. Discriminant analysis should be the most effective pattern recognition method, but in this scatter plot, multiple samples show overlap: 2,5-dimethylfuran, benzyl alcohol, n-propanol, formaldehyde, and propionic acid are cross-contaminated, as are n-pentanal, 4-methyloctane, and n-heptane. This is likely related to the low summation of discriminant function eigenvalues. The first two discriminant functions provide only 36.0% of the information, but the omission of multiple key pieces of information fails to represent the typical correlations among various substances, indicating that this array needs further improvement.

[0099] A solid paper-based array was constructed based on the liquid dye array of this invention for the detection of mixed VOCs components. After evaluation by various pattern recognition methods such as PCA, HCA, and LDA, the array can identify and detect 28 VOCs to a certain extent. However, it still has some drawbacks, such as low success rate of the dot matrix process, long reaction time, and low accuracy of target identification. Therefore, the array needs to be modified to increase its detection capability.

[0100] 5. Construction of modified paper-based arrays

[0101] (1) Synthesis and characterization of GA / Cu MOF

[0102] TEM was used to observe the morphology and size of GA / Cu MOF. Figure 9 A shows that GA / Cu MOF has good dispersion and uniformity, which lays a stable foundation for the sensor response. Figure 9 B clearly shows the morphology and size of the GA / Cu MOF. The GA / Cu MOF is spherical with an average particle size of 240 nm. Its interior consists of a Cu·-MOF core, while the exterior is covered by multiple layers of GA. Figure 9 (CF) indicates that the main elements of GA / Cu-MOF are C, O, N, and Cu.

[0103] (2) Modification of dyes by GA / Cu MOF

[0104] GA and Cu MOF were mixed in equal proportions and then doped with dyes to construct a modified array. The modification of GA / Cu MOF with 16 chemically responsive dyes was analyzed using UV-Vis absorption spectra before and after reaction with HVOS 4 (GA / Cu MOF to dye volume ratio of 0.5:1), RGB difference vectors (GA / Cu MOF to dye volume ratios of 0:1, 0.5:1, 1:1, 1.5:1, 3 parallel experiments), and difference spectra constructed from the absolute values ​​of the RGB differences in 3 parallel experiments (GA / Cu MOF to dye volume ratios of 0:1, 0.5:1, 1:1, 1.5:1). Figure 10 and Figure 11 As shown.

[0105] After AP 1 was doped with GA / Cu MOF, the Abs peak shape changed. The increase in the dye absorption peak near 420 nm may be due to the superposition of the GA / Cu MOF peak, while the disappearance of the dye absorption peak near 610 nm may be due to the adsorption of the dye by GA / Cu MOF. The RGB difference vectors of AP 1 before and after the reaction with HVOS 4 with different volumes of GA / Cu MOF showed significant changes, greatly improving the detection sensitivity of AP 1 for HVOS 4. The difference spectrum also proved that doping with GA / Cu MOF can indeed improve the color change range of AP 1, providing support for the detection of various VOCs. The situation of AP 2 after doping with GA / Cu MOF is similar to that of AP 1. The Euclidean distance of the sample with a doping volume ratio of 0.5:1 and the sample with a doping volume ratio of 1.5:1 are close. Considering the uniformity of array construction, the latter was selected as the doping ratio. The RGB difference vector and difference spectrum of AP 3 doped with GA / Cu MOF showed significant changes in reaction with HVOS 4. Based on the Abs analysis, it is speculated that the doping state of AP 3 with GA / Cu MOF is simple physical adsorption, but the auxiliary pigment (-NH2) of GA / Cu MOF attaches to the chromophore of the dye, enhancing its chromogenic ability. A doping ratio of 1.5:1 was selected. The doping of AP 4, 5, and 6 with GA / Cu MOF showed similar results to AP 1. After doping AP 7 with GA / Cu MOF, its absorption peak near 600 nm was significantly reduced, possibly because AP 7 was coated by GA / Cu MOF. However, the RGB difference and difference spectrum analysis indicated that this doping did not increase the detection sensitivity of AP 7 with HVOS 4; therefore, doping was not chosen to modify AP 7. Doping AP 8 with GA / Cu MOF did not improve the response of AP 8 to HVOS 4; therefore, doping was not chosen to modify AP 8. While doping AP9 with GA / Cu MOF does not affect its absorption peak near 500 nm, it significantly increases the RGB difference vector and fingerprint spectrum difference after reacting with HVOS 4. A doping ratio of 1:1 was chosen as the modification method. Doping AP10 with GA / Cu MOF greatly reduces the difference between the array points and VOCs after reaction; therefore, doping was not chosen to modify AP10. The situation with AP11-16 doped with GA / Cu MOF is similar to that of AP1: the peak shape of the UV-Vis absorption spectrum does not change significantly, but the RGB difference vector changes significantly after reacting with HVOS 4, and the difference spectrum difference changes significantly, greatly increasing the color change range and ability of the array points after reaction.

[0106] Therefore, the dye formulation in the modified array material is as follows:

[0107] Table 3

[0108]

[0109] Considering that the array material of this invention can be used to detect exhaled gases, the response time of the array was optimized at 37°C using toluene and tetrahydrofuran, such as... Figure 12 As shown, the array response reaches its maximum and then gradually levels off after 15 minutes; this time is selected as the array's response time for subsequent VOCs detection.

[0110] (3) VOCs pattern recognition based on difference vector

[0111] The modified paper-based array was exposed to 28 VOCs for 15 min. Images before and after the reaction were collected using a mobile phone. The RGB values ​​before and after the reaction were calculated using simple digital silhouettes for subsequent pattern recognition and fingerprint construction. This section describes the identification and detection of small organic molecule markers at two concentrations (100 μM and 300 μM). A richer and more diverse color spectrum indicates a greater color change in the array before and after the reaction. In the visualized difference spectrum of the reaction between the modified array and VOCs, the difference spectrum exhibited a rich variety of colors and a wide color gamut. These VOCs could be distinguished visually at both concentrations, indicating that the modification of the solid paper-based array using Cu MOF was successful. The difference spectrum is shown below. Figure 13 As shown.

[0112] ① Principal Component Analysis (PCA)

[0113] Figure 14The PCA results of the modified paper-based array reacting with 28 VOCs are presented. Modifying the array using GA / Cu MOF significantly reduced the principal component dimension and increased the importance of the first principal component (≥79%), greatly improving its interpretability. A 2D scatter plot could be constructed using the first two principal components to distinguish and identify the 28 VOCs. At low concentrations (100 μM), the KMO value was 0.870 (>0.750), indicating its suitability for principal component analysis. PCA1 and PCA2 explained 79.89% and 6.60% of the total information, respectively, with the first two principal components explaining a combined 86.49% of the effective information. In the scatter plot, except for significant overlap between n-heptane and benzyl alcohol, and butyric acid and nonanal, all other samples could be distinguished and identified, demonstrating significantly higher efficiency than the original array. At high concentrations (300 μM), the KMO value of 0.885 (>0.750) indicates that this condition is suitable for principal component analysis. PCA1 and PCA2 can explain 82.05% and 4.65% of the total information, respectively, and the first principal components can explain 86.70% of the effective information. This scatter plot differs significantly from previous principal component analyses. Each substance is evenly distributed on the 2D plot, rather than locally clustered. This indicates that at higher concentrations, PCA1 and PCA2 can reflect the chemical activity space of VOCs. Except for n-hexane and butyl acetate, which show significant overlap, other samples can be distinguished and identified. The discrimination ability is significantly improved after array modification, and the accuracy is further improved, proving that the modification of the array by GA / Cu MOF is successful.

[0114] ② Hierarchical Cluster Analysis (HCA)

[0115] By calculating the Euclidean distance between two vectors, it can be used as a metric for HCA (Hybrid Care Aspect). In HCA constructed using Euclidean distance (… Figure 15The most similar samples were grouped together based on similarity, with 92.9% and 98.8% of the 28 VOCs correctly classified, demonstrating the good performance of this novel sensor array. The connectivity and Euclidean distance of the dendrogram explained the similarity and degree of similarity between samples, respectively. In 84 experimental samples, substances of the same class were tightly clustered into a single cluster due to smaller Euclidean distance differences and the most similar properties. For samples with unknown composition, cluster analysis could also indicate which substances were most similar to the unknown sample. At low concentrations, six data points for 4-methyloctane and hexylcyclohexane (two alkanes) were misclassified, with an error rate of 7.1%. At high concentrations, one acetonitrile sample was incorrectly classified into the upper-level sample library, with an error rate of 1.2%. The increased accuracy and the ability to classify substances relatively accurately indicate that the construction of this modified array was successful in HCA and is beneficial for analyzing the similarity between substances. Although the twenty analytes were not clustered entirely according to major chemical categories, the clear and accurate clustering results for individual VOCs demonstrate the array's powerful ability to identify VOCs in exhaled breath. Once a database of the array's responses to various VOCs is constructed through cluster analysis, the chemical category of unknown samples can be easily identified.

[0116] ③ Linear Discriminant Analysis (LDA)

[0117] Compared to principal component analysis, linear discriminant analysis (LDA) can reveal the differences between analyte categories to the greatest extent. LDA was used to analyze the ΔR, ΔG, and ΔB data after the reaction of 28 VOCs with a modified paper-based array. A two-dimensional scatter plot was plotted using the scores of the data on the first discriminant function (Function 1) and the second linear discriminant function (Function 2) as the X and Y axes, respectively. Figure 16 At low concentrations, LDA Function 1 provides 60.7% of the effective characteristic information, and Function 2 provides 26.3%. The information structure of the first two discriminant functions is more reasonable than that of principal component analysis, which should help distinguish and identify target analytes. In the two-dimensional scatter plot, all samples can be successfully classified into one class and distinguished from other substances, with a 100% accuracy rate for linear discriminant leave-one-out (LOO) analysis. At high concentrations, Function 1 provides 70.8% of the effective characteristic information, and Function 2 provides 10.0%. The first two discriminant functions provide 80.8% of the effective information. In the two-dimensional scatter plot, all samples can be successfully classified into one class and distinguished from other substances, with a 100% accuracy rate for linear discriminant leave-one-out (LOO) analysis. After array modification, its specific response to VOCs is greatly enhanced. LDA at both concentrations demonstrates that the modified array can successfully distinguish and identify 28 VOC components, indicating that the construction of the modified array is successful.

[0118] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit the technical solutions. Those skilled in the art should understand that any modifications or equivalent substitutions to the technical solutions of the present invention without departing from the spirit and scope of the present invention should be covered within the scope of the claims of the present invention.

Claims

1. An array material for detecting VOCs components in a gas, characterized in that, The array material is obtained through the following steps: Step 1: Dissolve each dye separately in an ethanol solution to obtain dye reagents; the concentration of the ethanol solution is 50%, and the concentration of the dye in the ethanol solution is 1 mg / 2 mL; the dyes include bromophenol blue, phenol red, rhodamine B, bromophenol blue, nitromethamine yellow, bromothymol blue, alizarin, m-amine yellow, bromophenol red, bromocresol violet, acridine orange, cresol red, Congo red, thymol blue, and brilliant yellow; Step 2: Select a polyvinylidene fluoride (PVDF) film as a substrate, and individually drop the dye reagent obtained in Step 1 onto the PVDF film to form multiple dye detection points, thus forming a detection array; then cover the PVDF film with a sealing film on the top and bottom to seal it, and make an air inlet and an air outlet on the sealing film above and below each dye detection point, respectively; slowly blow nitrogen gas into the PVDF film to obtain the array material; A modifying material was also added to each dye; the modifying material was a mixture of graphene aerogel and copper metal-organic framework, with a mass ratio of graphene aerogel to copper metal-organic framework of 1:

1.

2. The array material for detecting VOCs components in gas according to claim 1, characterized in that, The dyes include pretreated dyes and untreated dyes. The pretreated dyes are obtained by dissolving each dye separately in a 100 mM / L NaOH solution for modification. The untreated dyes are dyes that are directly added to an ethanol solution.

3. The array material for detecting VOCs components in gas according to claim 2, characterized in that, The pretreated dyes include bromophenol blue, phenol red, rhodamine B, bromophenol blue, nitromethamine yellow, bromothymol blue, and alizarin; the untreated dyes include m-amine yellow, bromophenol red, bromocresol violet, acridine orange, cresol red, Congo red, thymol blue, brilliant yellow, and phenol red.

4. The array material for detecting VOCs components in gas according to claim 1, characterized in that, The graphene aerogel was obtained by the following method: graphene oxide was added to distilled water and ultrasonically treated to form a uniform graphene oxide dispersion; then the graphene oxide dispersion was transferred to a reactor and heated at 180°C for 12 hours to obtain a black product. After cooling to room temperature, the product was centrifuged and clarified to obtain the graphene aerogel.

5. The array material for detecting VOCs components in gas according to claim 1, characterized in that, The copper metal-organic framework was obtained by the following method: S1: Dissolve polyvinylpyrrolidone in DMF and ethanol, and label it as reagent 1; wherein the volume ratio of DMF to ethanol is 1:1, and the concentration of polyvinylpyrrolidone in the mixed solution is 0.025 g / ml; S2: Dissolve copper nitrate trihydrate and 2-aminoterephthalic acid in DMF and label it as reagent 2; wherein the concentration of copper nitrate trihydrate in DMF is 0.025 mmol / ml and the concentration of 2-aminoterephthalic acid in DMF is 0.025 mmol / ml. S3: Mix reagent 1 and reagent 2 under ultrasonic treatment for 30 min, transfer to a high-pressure reactor, and react at 100℃ for 8 h to obtain the product; S4: The product obtained in S3 was washed with ethanol several times, and the precipitate was collected by centrifugation. After vacuum drying, a green powder was obtained, which was soluble in ethanol.

6. A method for detecting VOCs components in a gas, characterized in that, Specifically, the steps include the following: Step 1: Place the array material described in any one of claims 1 to 5 in a transparent, closed detection box, then scan the array and acquire images before the reaction, and obtain the RGB values ​​before the reaction; Step 2: Select dry nitrogen as a control sample, pump dry nitrogen into the detection box and circulate it slowly to scan the array and obtain an image as the blank array image, and obtain the RGB values ​​of the blank array; Step 3: Pump the gas to be detected into the detection box and circulate it slowly. Then scan the array and acquire the image after the reaction as the image of the array after the reaction, and obtain the RGB value of the array after the reaction. Step 4: Using the RGB values ​​of the blank array as initial values, subtract the RGB values ​​before the reaction from the RGB values ​​of the array after the reaction to obtain ΔR, ΔG, and ΔB; Step 5: Take the absolute value of the ΔRGB value to construct the fingerprint spectrum, and use the difference vector for statistical analysis to obtain the VOCs composition in the gas to be detected.

7. The method for detecting VOCs components in a gas according to claim 6, characterized in that, The gas to be detected is either the gas exhaled by a human body or the gas to be detected.

8. The method for detecting VOCs components in a gas according to claim 6, characterized in that, In step 3, the gas to be detected is pumped into the detection box and reacted for 15-30 minutes before scanning.

9. The method for detecting VOCs components in a gas according to claim 6, characterized in that, The VOCs components include toluene, acetic anhydride, propionic acid, benzene, methacrolein, formic acid, acetonitrile, n-pentane, butyric acid, n-pentanal, formaldehyde, acetaldehyde, n-propanol, butyl acetate, acetone, hexadecane, 4-methyloctane, butyraldehyde, n-hexane, n-butylamine, 2,5-dimethylfuran, n-heptane, 1,2,4-trimethylbenzene, benzyl alcohol, nonanal, methyl phenylacetate, hexylcyclohexane, and isoprene.

Images