Method for rapidly detecting content of bound acrylonitrile in nitrile rubber cement

Abstract

The invention discloses a method for rapidly detecting the content of bound acrylonitrile in nitrile rubber cement, which comprises the following steps of: acquiring a near infrared spectrum of the rubber cement by using a near infrared spectrum analyzer, measuring the content of bound acrylonitrile in the rubber cement, and performing spectrum pretreatment on infrared spectrum information; establishing a mathematical model of the processed near infrared spectrum and the corresponding bound acrylonitrile content, evaluating the mathematical model to obtain an accurate near infrared prediction model, and measuring the bound acrylonitrile content in the to-be-measured rubber cement by using the near infrared prediction model. The method disclosed by the invention is based on a near infrared spectroscopic analysis technology, can realize rapid nondestructive measurement of the content of bound acrylonitrile, is simple and convenient to operate and is an environment-friendly analysis method, and an operator does not need to contact a chemical reagent.

Description

Technical Field

[0001] The invention belongs to the technical field of chemical substance content determination methods and relates to a method for quickly detecting the content of combined acrylonitrile in nitrile rubber mortar. Background Art

[0002] Acrylonitrile-butadiene rubber, referred to as nitrile rubber (NBR), is a synthetic rubber made by copolymerization of acrylonitrile and butadiene. It has strong polarity due to its -CN content, and has the characteristics of good oil resistance, wear resistance, heat resistance and barrier properties. The amount of bound acrylonitrile affects the difficulty of rotation within the rubber molecular chain and the intermolecular force, thereby affecting the flexibility of the molecular chain and the product's glass transition temperature, solubility parameter, oil resistance, wear resistance, cold resistance and other physical and mechanical properties. Therefore, accurate determination of bound acrylonitrile content is particularly important in the synthetic rubber industry.

[0003] Currently, the methods for testing the bound acrylonitrile content in NBR mainly include combustion method, Kjeldahl nitrogen determination method, infrared spectroscopy, thermogravimetric and differential thermogravimetric analysis, and pyrolysis chromatography.

[0004] Combustion method: It is necessary to prepare the sample, extract it with anhydrous ethanol, dry it and cool it, and then use an element analyzer to measure its nitrogen content, and calculate the combined acrylonitrile content according to the formula. Because this method measures the sum of all nitrogen-containing substances in the sample, the test results are easily affected by free acrylonitrile.

[0005] Kjeldahl method: The sample is pretreated and then alkalized, and the total nitrogen content is calculated. This method has been widely used as a classic method, but the catalyst used is highly toxic, not environmentally friendly, and the experimental process is time-consuming.

[0006] Infrared spectroscopy: After the sample is processed and prepared into a sheet, it is heated and dissolved, and the filtrate is applied to a salt sheet and dried to form a film, and its infrared spectrum is measured. This method has a fast analysis speed and simple operation; it can analyze the combined acrylonitrile content of NBR raw rubber, vulcanized rubber, and NBR blended with other polymers. With the advent of Fourier transform infrared spectroscopy attenuated total reflection ATR-FTIR technology, raw rubber can be directly tested without sample pretreatment such as dissolution or high-temperature pyrolysis. However, infrared spectroscopy also has some shortcomings: the relative error of the test is large; a working curve needs to be established, and the working curve is affected by factors such as instrument resolution and sample preparation method. ATR can only directly test raw rubber samples. For vulcanized rubber samples, high-temperature pyrolysis method is required to prepare samples, so the test results are affected by the cracking temperature and cracking time. In summary, the infrared method is generally used to test raw rubber or vulcanized rubber samples, and cannot be used to detect the glue of the control material.

[0007] Pyrolysis chromatography: The premise of this method is that the relative production rate of acrylonitrile and butadiene in the pyrolysis product has a good linear relationship with the ratio of the two monomers in the corresponding NBR. It has the advantages of high sensitivity, fast analysis speed, small sample dosage, no pretreatment, and low price. However, it has poor repeatability and requires high control of the pyrolysis temperature, otherwise it will affect the accuracy.

[0008] Thermogravimetric and differential thermogravimetric analysis method: Thermogravimetric and differential thermogravimetric curves of NBR are tested under nitrogen and air conditions respectively, and the amount of carbon residue is obtained from the obtained curves, thereby establishing a working curve of carbon residue and acrylonitrile content.

[0009] From the existing methods for testing acrylonitrile content, most of them require sample pretreatment or appropriate preparation. The advantages and disadvantages of various methods are shown in Table 1.

[0010] Table 1 Summary of advantages and disadvantages of existing test methods

[0011]

[0012] Summary of the invention

[0013] The invention aims to provide a method for quickly detecting the content of bound acrylonitrile in nitrile rubber mortar. The method is based on near-infrared spectroscopy analysis technology, can realize rapid and non-destructive measurement of the content of bound acrylonitrile, and is simple to operate. Operators do not need to contact chemical reagents, and the method is an environmentally friendly analysis method.

[0014] The technical scheme adopted by the present invention is a method for quickly detecting the bound acrylonitrile content in nitrile rubber mortar, using a near-infrared spectrometer to obtain a near-infrared spectrum of the mortar, measuring the bound acrylonitrile content in the mortar, performing spectral preprocessing on infrared spectral information, establishing a mathematical model of the processed near-infrared spectrum and the corresponding bound acrylonitrile content, evaluating the mathematical model to obtain an accurate near-infrared prediction model, and using the near-infrared prediction model to measure the bound acrylonitrile content in the mortar to be tested.

[0015] The present invention is also characterized in that:

[0016] The method for rapidly detecting the content of bound acrylonitrile in nitrile rubber mortar is specifically implemented according to the following steps:

[0017] Step 1, collecting near infrared spectrum of the glue;

[0018] Step 2, measuring the bound acrylonitrile content of the mortar in step 1;

[0019] Step 3, performing spectral preprocessing on the near infrared spectral information obtained in step 1;

[0020] Step 4, dividing the data obtained in step 3 into a calibration set and a validation set, and then performing regression modeling on the combined acrylonitrile content data obtained in step 2 and the spectral information data of the calibration set;

[0021] Step 5: Use the validation set data in step 4 to evaluate the obtained model and obtain a near infrared prediction model;

[0022] Step 6: collect the near infrared spectrum of the tested mortar, substitute the spectrum into the near infrared prediction model evaluated in step 5, and obtain the bound acrylonitrile content of the tested mortar.

[0023] The acquisition conditions of the near-infrared spectrum are as follows: spectral scanning range 780nm-2500nm; transmission mode; sample pool 1mm-2mm optical path; each sample is collected 64 times to take the average spectrum.

[0024] The methods of spectral preprocessing include: first-order derivative method, second-order derivative method, SG smoothing, NORRIS smoothing, combination of first-order derivative and SG smoothing, combination of first-order derivative and NORRIS smoothing, combination of second-order derivative and SG smoothing, and combination of second-order derivative and NORRIS smoothing.

[0025] The partial least squares method is used in modeling, the optimal wavelength range is selected, and abnormal data of chemical values ​​and spectral values ​​are eliminated.

[0026] The model evaluation was performed based on the cross-validation standard deviation and the cross-validation correlation coefficient, and finally the model with the correlation coefficient closest to 1 and the smallest cross-validation standard deviation was selected as the near-infrared prediction model.

[0027] The beneficial effects of the present invention are:

[0028] (1) The method of the present invention does not require sample pretreatment and can achieve non-destructive measurement of samples;

[0029] (2) Compared with the 11-hour experimental time of the traditional method, the method of the present invention has a fast measurement speed, convenient and quick operation, and the bound acrylonitrile content data can be obtained in 3 minutes;

[0030] (3) The entire operation process of the method of the present invention does not require contact with any chemical reagents, and is an environmentally friendly rapid analysis method;

[0031] (4) The object of measurement of the method of the present invention is the intermediate material in the rubber production process, namely, the rubber mortar. More than 95% of the rubber mortar is water. The existing detection method requires the preparation or pretreatment of the rubber mortar first, and there will be varying degrees of loss or contamination during this process. The method of the present invention realizes non-destructive and direct detection of the rubber mortar, thus avoiding this problem.

[0032] (5) Since more than 95% of the glue is water, it is difficult to directly obtain the near-infrared spectrum. The existing testing methods generally perform pre-treatment operations such as curing and tableting on the glue. The method of the present invention is based on the purpose of fast measurement and convenient operation. It does not pre-treat the glue, but directly scans the spectrum, uses transmission, and selects a sample cell with an optical path of 1-2mm to collect the near-infrared spectrum of the glue. Although the obtained spectral peak is not obvious and the peak contains too much useless information and interference information, useful information can be extracted through spectral pre-processing, and a mathematical relationship between acrylonitrile and the spectrum can be established. Subsequent work verification proves that the accuracy of the mathematical relationship is very good;

[0033] (6) The present invention does not require the preparation or correction of a working curve. When the production process does not fluctuate much (generally, rubber production is relatively stable, which is also a factor that near-infrared spectroscopy analysis technology is particularly suitable for industrial analysis), there is no need to calibrate and improve the analysis model. In addition, the larger the database of the method of the present invention is (i.e., the longer it has been in use), the more stable it is and the higher the accuracy is.

[0034] (7) The method of the present invention is not only in the laboratory research and development stage, but has also been applied in industry. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 is a near infrared spectrum of the glue measured in Example 1 of the present invention;

[0036] Figure 2 is a correlation diagram of the measured value and the predicted value of the adhesive bonded to acrylonitrile in Example 1 of the present invention;

[0037] Figure 3 is a near infrared spectrum of the glue measured in Example 2 of the present invention;

[0038] Figure 4 It is a correlation diagram of the measured value and the predicted value of the adhesive bonded acrylonitrile in Example 2 of the present invention. DETAILED DESCRIPTION

[0039] The present invention is described in detail below with reference to the accompanying drawings and specific embodiments.

[0040] The present invention discloses a method for rapidly detecting the bound acrylonitrile content in nitrile rubber mortar based on near infrared spectroscopy analysis technology, wherein the near infrared spectrum of the mortar is obtained by using a near infrared spectroscopy analyzer, a near infrared spectrum information library of the mortar and a corresponding bound acrylonitrile database are established, and on this basis, a mathematical relationship between the bound acrylonitrile content and the near infrared spectrum of the mortar is found, thereby calculating the bound acrylonitrile content in the mortar to be detected. The method is specifically implemented according to the following steps:

[0041] Step 1, collecting the near infrared spectrum of the glue, the collection conditions are: spectral scanning range 780nm-2500nm; transmission mode; sample pool 1mm-2mm optical path; each sample is collected 64 times to obtain the average spectrum;

[0042] Step 2: Measure the bound acrylonitrile content of the rubber in step 1 by using the Kjeldahl method, and collect the data. The reference standard for the Kjeldahl method is: SH / T 1157.2-2015, "Determination of bound acrylonitrile content in raw rubber acrylonitrile-butadiene rubber (NBR)";

[0043] Step 3, performing spectral preprocessing on the near-infrared spectrum information obtained in step 1, wherein the preprocessing methods include first-order derivative method, second-order derivative method, SG smoothing, NORRIS smoothing, first-order derivative and SG smoothing combination method, first-order derivative and NORRIS smoothing combination method, second-order derivative and SG smoothing combination method, and second-order derivative and NORRIS smoothing combination method to obtain near-infrared spectrum data with noise interference eliminated;

[0044] Step 4: Divide the data obtained in step 3 into a calibration set and a validation set, and then perform regression modeling on the combined acrylonitrile content data obtained in step 2 and the spectral information data of the calibration set. The partial least squares method is used in modeling, the optimal wavelength range is selected, and after eliminating abnormal data of chemical values ​​and spectral values, a near-infrared prediction model for combined acrylonitrile content is established;

[0045] Step 5: Use the validation set data in step 4 to evaluate the near infrared prediction model. The evaluation is performed based on the cross-validation standard deviation and the cross-validation correlation coefficient. Finally, the model with the correlation coefficient closest to 1 and the smallest cross-validation standard deviation is selected as the near infrared prediction model.

[0046] Step 6: collect the near infrared spectrum of the tested mortar, substitute the spectrum into the near infrared prediction model evaluated in step 5, and obtain the bound acrylonitrile content of the tested mortar.

[0047] Example 1

[0048] The near infrared spectrometer used in this example is Thermofisher Antris II, and the modeling software is TQ 9.

[0049] The mortar samples come from the nitrile rubber mortar produced in real time by the nitrile rubber unit of Lanzhou Petrochemical Company.

[0050] Step 1: Collect 104 batches of glue samples, shake them evenly, put them into a sample pool with an optical path of 1-2 mm, and put them into a near-infrared spectrometer for scanning in transmission mode. The scanning results are as follows: Figure 1 As shown;

[0051] Step 2: Determine the bound acrylonitrile content of the mortar in step 1 by Kjeldahl nitrogen determination method. The test results are shown in Table 2.

[0052] Table 2 Measured data of acrylonitrile content in adhesive

[0053]

[0054]

[0055] Step 3, performing spectral preprocessing on the near-infrared spectrum information obtained in step 1. The preprocessing method used in this embodiment is a method combining first-order derivative and Norris smoothing;

[0056] Step 4: Enter the near infrared spectra of 104 batches of samples and the corresponding combined acrylonitrile content data into the modeling software to start building the near infrared prediction model: (1) divide the calibration set and the validation set; (2) the optimal wavelength range is 7513-6549 cm -1 ; (3) Eliminate abnormal data of chemical values ​​and spectral values; (4) Use partial least squares method for regression modeling.

[0057] Step 5: The established near infrared prediction model is evaluated by the interactive validation standard deviation secv, the calibration set standard deviation sec, the validation set standard deviation sep, the correlation coefficient R, and the optimal number of principal factors P. In this embodiment, sec=0.613, sep=0.746, rmsecv=0.424, R=0.7129, and P=6. The correlation is as follows: Figure 2 shown.

[0058] Step 6: Collect near-infrared spectral information of 14 batches of mortar samples to be tested (samples other than batch 104), substitute into the near-infrared prediction model obtained in step 5, and obtain the predicted values ​​of the combined acrylonitrile content of the 14 batches of samples; then use the Kjeldahl method to measure the combined acrylonitrile content of the 14 batches of samples to obtain the measured values, and examine the accuracy of the model. The results are shown in Table 3.

[0059] Table 3 Comparison of measured and predicted values ​​of acrylonitrile content in adhesive

[0060] Serial number Predicted value / % Measured value / % Absolute Deviation 1 33.81 33.69 -0.12 2 33.79 33.75 -0.04 3 33.99 33.85 -0.14 4 34.01 33.66 -0.35 5 33.92 34.16 0.24 6 33.82 33.93 0.11 7 33.72 33.53 -0.19 8 33.71 33.46 -0.25 9 33.18 32.93 -0.25 10 33.16 33.51 0.35 11 33.2 33.21 0.01 12 33.13 32.89 -0.24 13 33.21 33.20 -0.01 14 33.31 33.15 -0.16

[0061] From the data in Table 3, it can be seen that the rapid detection method of the present invention has a high accuracy, the maximum absolute deviation is only 0.35, and the standard deviation SD of the deviation is 0.20%.

[0062] Example 2

[0063] The near infrared spectrometer used in this embodiment is Thermofisher Antris II, and the modeling software is TQ 9. The mortar sample comes from No. 2 nitrile rubber mortar produced in real time by the nitrile rubber unit of Lanzhou Petrochemical Company.

[0064] Step 1: Collect 167 batches of glue samples, shake them evenly, put them into a sample pool with an optical path of 1-2 mm, and put them into a near-infrared spectrometer for scanning in transmission mode. The scanning results are as follows: Figure 3 As shown;

[0065] Step 2: Determine the bound acrylonitrile content of the mortar in step 1 by Kjeldahl method.

[0066] Step 3, performing spectral preprocessing on the near-infrared spectrum information obtained in step 1, wherein the preprocessing method used in this embodiment is the first-order derivative;

[0067] Step 4: Enter the near infrared spectra of 167 batches of samples and the corresponding combined acrylonitrile content data into the modeling software to start building the near infrared prediction model: (1) divide the calibration set and the validation set; (2) the optimal wavelength range is 8319-5869cm -1 ; (3) Eliminate abnormal data of chemical values ​​and spectral values; (4) Use partial least squares method for regression modeling.

[0068] Step 5: The established near infrared prediction model is evaluated by the interactive validation standard deviation secv, the calibration set standard deviation sec, the validation set standard deviation sep, the correlation coefficient R, and the optimal number of principal factors P. In this embodiment, sec=0.543, sep=0.664, rmsecv=0.536, R=0.799, and P=5. The correlation is as follows: Figure 4 shown.

[0069] Step 6: Collect near-infrared spectral information of 20 batches of mortar samples to be tested (samples other than batch 167), substitute into the near-infrared prediction model obtained in step 5, and obtain the predicted value of bound acrylonitrile content; then use the Kjeldahl method to measure the bound acrylonitrile content of the 20 batches of samples to obtain the measured value, and examine the accuracy of the model. The results are shown in Table 4.

[0070] Table 4 Comparison of measured and predicted values ​​of acrylonitrile content in adhesive

[0071]

[0072]

[0073] From the data in Table 4, it can be seen that the rapid detection method of the present invention has a high accuracy, the maximum absolute deviation is only 0.47, and the standard deviation SD of the deviation is 0.23%.

[0074] Example 3

[0075] The data used for modeling in this embodiment are the same as those in embodiment 2, except that the preprocessing method used for the near infrared spectrum is the second-order derivative. The near infrared prediction model obtained in this embodiment is used to measure the 20 batches of mortar samples to be tested in embodiment 2 to obtain the predicted values, which are then compared with the actual measured values. The maximum absolute deviation is 0.42, and the standard deviation of the deviation is SD=0.33%.

Claims

1. A method for rapidly detecting the content of bound acrylonitrile in nitrile rubber mortar, characterized in that: A near-infrared spectrum of the glue is obtained by using a near-infrared spectrometer, the bound acrylonitrile content in the glue is measured, the infrared spectrum information is spectrally preprocessed, a mathematical model of the processed near-infrared spectrum and the corresponding bound acrylonitrile content is established, the mathematical model is evaluated to obtain an accurate near-infrared prediction model, and the near-infrared prediction model is used to measure the bound acrylonitrile content in the glue to be tested.

2. The method for rapidly detecting the content of bound acrylonitrile in nitrile rubber mortar according to claim 1, characterized in that: Follow the steps below to implement it: Step 1, collecting near infrared spectrum of the glue; Step 2, measuring the bound acrylonitrile content of the mortar in step 1; Step 3, performing spectral preprocessing on the near infrared spectral information obtained in step 1; Step 4, dividing the data obtained in step 3 into a calibration set and a validation set, and then performing regression modeling on the combined acrylonitrile content data obtained in step 2 and the spectral information data of the calibration set; Step 5: Use the validation set data in step 4 to evaluate the obtained model and obtain a near infrared prediction model; Step 6: collect the near infrared spectrum of the tested mortar, substitute the spectrum into the near infrared prediction model evaluated in step 5, and obtain the bound acrylonitrile content of the tested mortar.

3. The method for rapidly detecting the content of bound acrylonitrile in nitrile rubber mortar according to claim 1 or 2, characterized in that: The acquisition conditions of the near infrared spectrum are as follows: spectral scanning range 780nm-2500nm; transmission mode; sample pool 1mm-2mm optical path; each sample is collected 64 times to obtain the average spectrum.

4. The method for rapidly detecting the content of bound acrylonitrile in nitrile rubber mortar according to claim 1 or 2, characterized in that: The methods of spectral preprocessing include: first-order derivative method, second-order derivative method, SG smoothing, NORRIS smoothing, combination of first-order derivative and SG smoothing, combination of first-order derivative and NORRIS smoothing, combination of second-order derivative and SG smoothing, and combination of second-order derivative and NORRIS smoothing.

5. The method for rapidly detecting the content of bound acrylonitrile in nitrile rubber mortar according to claim 1 or 2, characterized in that: The partial least squares method is used in modeling, the optimal wavelength range is selected, and abnormal data of chemical values ​​and spectral values ​​are eliminated.

6. The method for rapidly detecting the content of bound acrylonitrile in nitrile rubber mortar according to claim 1 or 2, characterized in that: The model evaluation was performed based on the cross-validation standard deviation and the cross-validation correlation coefficient, and finally the model with the correlation coefficient closest to 1 and the smallest cross-validation standard deviation was selected as the near-infrared prediction model.

Images