CN115420730B - A method for quantitative elemental analysis in longitudinal depth based on laser-induced breakdown spectroscopy - Google Patents

Abstract

The invention discloses a longitudinal depth element quantitative analysis method based on laser-induced breakdown spectroscopy, which comprises the following steps of S1, training ablation depth data of LIBS by using a partial least square method to obtain a depth prediction model; s2, analyzing quantitative depth distribution of the sample element to be detected; the method specifically comprises the following substeps: s21, respectively ablating the sample to be detected by using LIBS technology for different pulse times, and collecting spectra; s22, predicting an ablation depth value of the sample to be detected through a spectrum and depth prediction model; s3, analyzing the element content in the sample to be tested through a calibration method or a non-calibration method; s4, quantitatively analyzing the element depth distribution of the sample to be tested through the depth value predicted in S22 and the element content calculated in S3: and (3) mapping the ablation depth value predicted in the step (S22) and the content of the target element calculated in the step (S3) as an abscissa and an ordinate respectively to obtain the content depth distribution of the target element in the sample to be detected.

Description

A method for quantitative elemental analysis in longitudinal depth based on laser-induced breakdown spectroscopy

Technical Field

The invention belongs to the field of spectroscopy, and in particular relates to a longitudinal depth element quantitative analysis method based on laser induced breakdown spectroscopy.

Background Art

At present, the detection of element content and distribution of solid samples is mostly based on the detection of sample surface. The detection of elements inside the sample generally requires the complete destruction of the sample, such as making it into powder; or the detection of elements in the longitudinal depth of the sample requires the longitudinal cutting of the sample and the detection on the longitudinal section. These detection methods limit the real-time and in-situ detection of samples, especially for some samples that cannot be destroyed. Laser-induced breakdown spectroscopy (LIBS) technology is a qualitative and quantitative analysis technology based on atomic excitation. Since this technology generates high-temperature plasma by laser ablation of samples and collects the characteristic spectrum generated by high-temperature plasma to analyze the sample, LIBS technology does not need to destroy or cut the sample, but only needs to perform quantitative element analysis from the surface to the depth. This provides convenience for in-situ and real-time element detection. For example, in steel smelting, the use of LIBS can realize real-time in-situ detection of steel smelting purity during the smelting process; in the ITER project, the depth and distribution of impurity ion deposition in the first wall material of the tokamak device can be detected in real time.

At present, the element depth analysis based on LIBS can only use the number of laser pulses to characterize the ablation depth value, and cannot give the actual depth value; testing the specific depth requires other depth measurement instruments as auxiliary equipment (such as confocal microscopes, etc.) to specifically determine the specific depth value of the sample, and the test sample needs to be taken out for sample preparation or cutting to meet the detection requirements of other equipment, which undoubtedly increases the cost of detection and the inability to achieve in-situ detection. All these limit the practical application of LIBS deep element quantitative analysis. Therefore, it is a technical problem to be solved urgently to simultaneously detect the specific depth value of the sample and the element content at different depths through LIBS.

Summary of the invention

The purpose of the present invention is to overcome the shortcomings of the prior art and provide a method for longitudinal depth element quantitative analysis based on laser induced breakdown spectroscopy, which first predicts the ablation depth by using an ablation depth prediction model based on partial least squares combined with LIBS, then calculates the sample content by a calibration method, and finally analyzes the depth distribution of the target element content in the sample to be tested. This method improves the feasibility of the application of LIBS instruments in the prediction of deep content distribution. After the training process is completed, LIBS can be used only to perform deep quantitative analysis of the target element in the sample to be tested without the help of other instruments. At the same time, this method has simple steps, is easy to operate, and has high practical value.

The objective of the present invention is achieved through the following technical solution: A longitudinal depth element quantitative analysis method based on laser induced breakdown spectroscopy comprises the following steps:

S1. Use partial least squares method to train the LIBS ablation depth data to obtain a depth prediction model; specifically, it includes the following sub-steps:

S11. Preparation of standard samples: design a series of element content ratios of standard samples according to the elemental composition of the sample to be tested, mix a series of powder raw materials with different element ratios evenly, and press them into thin sheets by a manual tablet press to obtain a series of standard samples with different element contents;

S12, LIBS spectrum acquisition: The standard samples were ablated with different pulse times using LIBS technology, and the spectrum was acquired;

S13, obtaining ablation depth values: using a confocal microscope to measure ablation depth values of the S11 standard sample at different pulse times;

S14, establishing an ablation depth prediction model: selecting multiple spectral lines of the element with the highest content in the LIBS spectrum collected in S12, and using the vector composed of the intensities of the multiple spectral lines as input data for the partial least squares method; using the ablation depth value obtained in S13 as a label for the partial least squares method; and then fitting the input data and the label to obtain a depth prediction model;

S2, analyzing the quantitative depth distribution of elements in the sample to be tested; specifically including the following sub-steps:

S21, using LIBS technology to ablate the samples with different pulse times, and collect spectra;

S22, predicting the ablation depth value of the sample to be tested by using the spectrum collected in S21 and the depth prediction model obtained in S14;

S3. Analyze the element content in the sample by calibration method or non-calibration method;

S4. Complete the quantitative analysis of the element depth distribution of the sample to be tested through the depth value predicted by S22 and the element content calculated by S3: Use the ablation depth value predicted in S22 and the target element content calculated in S3 as the horizontal axis and the vertical axis to make a graph, and obtain the depth distribution of the target element content in the sample to be tested.

Furthermore, the calibration method is specifically as follows: multiple spectral lines with clear outlines in the LIBS spectrum in S12 are selected, and the target element spectral line intensity is fitted with the target element content, or the target element intensity ratio is fitted with the target element content ratio, so as to fit different calibration curves; then the fitted calibration curve is verified by multiple cross-validation methods, and the calibration curve with the smallest average percentage deviation is selected as the optimal calibration curve; finally, the target element spectral line intensity or the target element intensity ratio to be calculated is substituted into the optimal calibration curve to obtain the target element content or the target element content ratio.

The beneficial effects of the present invention are as follows: the present invention provides an element depth quantitative analysis method based on a LIBS instrument, firstly predicting the ablation depth by using an ablation depth prediction model based on partial least squares combined with LIBS, then calculating the sample content by a calibration method, and finally analyzing the depth distribution of the target element content in the sample to be tested. Thereby improving the feasibility of the application of LIBS instruments in the prediction of deep content distribution. In general, the method can be performed without the help of other instruments after the training process is completed, and only LIBS is used to perform deep quantitative analysis of the target elements in the sample to be tested. At the same time, the method has simple steps, is easy to operate, has high practical value, and is worthy of promotion in the industry.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG1 is a flow chart of a method for quantitative element analysis in longitudinal depth based on laser induced breakdown spectroscopy of the present invention;

FIG2 is a regression curve of the depth prediction model in Experiment (I) of the present invention;

FIG3 is a depth prediction result of the tungsten and molybdenum samples to be tested in Experiment (I) of the present invention;

FIG4 is the optimal calibration curve selected in Experiment (I) of the present invention;

FIG5 is a depth distribution of the tungsten and molybdenum content in the sample to be tested in Experiment (I) of the present invention;

FIG6 is a depth prediction result of the tungsten and molybdenum samples to be tested in Experiment (II) of the present invention;

FIG. 7 is a depth distribution of the tungsten and molybdenum content in the sample to be tested in Experiment (II) of the present invention;

FIG8 is a regression curve of the depth prediction model in Experiment (III) of the present invention;

FIG9 is a depth prediction result of the sample to be tested in Experiment (III) of the present invention;

FIG10 is a depth distribution of tungsten and lithium content in the sample to be tested in Experiment (III) of the present invention;

FIG11 is a depth prediction result of the sample to be tested in Experiment (IV) of the present invention;

FIG. 12 is a diagram showing the depth distribution of the tungsten and lithium content in the sample to be tested in Experiment (IV) of the present invention.

DETAILED DESCRIPTION

The present invention is further described below in conjunction with the accompanying drawings and specific embodiments.

As shown in FIG1 , a method for quantitative element analysis based on laser induced breakdown spectroscopy of the present invention comprises the following steps:

S1. Use partial least squares method to train the LIBS ablation depth data to obtain a depth prediction model; specifically, it includes the following sub-steps:

S11. Preparation of standard samples: design a series of element content ratios of standard samples according to the elemental composition of the sample to be tested, mix a series of powder raw materials with different element ratios evenly, and press them into thin sheets by a manual tablet press to obtain a series of standard samples with different element contents;

S12, LIBS spectrum acquisition: The standard samples were ablated with different pulse times using LIBS technology, and the spectrum was acquired;

S13, obtaining ablation depth values: using a confocal microscope to measure ablation depth values of the S11 standard sample at different pulse times;

S14, establishing an ablation depth prediction model: selecting multiple spectral lines of the element with the highest content in the LIBS spectrum collected in S12 (usually selecting spectral lines with more than 10 wavelength positions), and using the vector composed of the intensities of the multiple spectral lines as input data for the partial least squares method; using the ablation depth value obtained in S13 as a label for the partial least squares method; and then fitting the input data and the label to obtain a depth prediction model;

S2, analyzing the quantitative depth distribution of elements in the sample to be tested; specifically including the following sub-steps:

S21, using LIBS technology to ablate the samples with different pulse times, and collect spectra;

S22, predicting the ablation depth value of the sample to be tested by using the spectrum collected in S21 and the depth prediction model obtained in S14;

S3, analyze the element content in the sample to be tested by calibration method or non-calibration method; the calibration method is specifically: select multiple spectral lines with clear outlines in the LIBS spectrum in S12, use the target element spectral line intensity and the target element content to fit, or the target element intensity ratio and the target element content ratio to fit, so as to fit different calibration curves; then use multiple cross-validation methods to verify the fitted calibration curve, and select the calibration curve with the smallest average percentage deviation as the optimal calibration curve; finally, bring the target element spectral line intensity or the target element intensity ratio to be calculated into the optimal calibration curve to obtain the target element content or the target element content ratio. The non-calibration method is a commonly used method in this field, and specific reference can be made to "A.Ciucci, M.Corsi, V.Palleschi, S.Rastelli, A.Salvetti, E.Tognoni. "New procedure for quantitative elemental analysis by laser-induced plasmaspectroscopy", Applied Spectroscopy. 1999, 53 (8): 960-964.".

S4. Complete the quantitative analysis of the element depth distribution of the sample to be tested through the depth value predicted by S22 and the element content calculated by S3: Use the ablation depth value predicted in S22 and the target element content calculated in S3 as the horizontal axis and the vertical axis to make a graph, and obtain the depth distribution of the target element content in the sample to be tested.

The technical effect of the present invention is further verified by specific experiments below.

(A) A longitudinal depth element quantitative analysis method based on laser induced breakdown spectroscopy with different depth distributions of tungsten and molybdenum content.

Ten different proportions of tungsten and molybdenum powders were uniformly mixed, with the proportions of tungsten and molybdenum being 72.4% and 27.6%; 74.3% and 25.7%; 76.4% and 23.6%; 78.9% and 21.1%; 81.4% and 18.6%; 83.8% and 16.2%; 86.7% and 13.3%; 89.7% and 10.3%; 92.9% and 7.1%; 96.2% and 3.8%, respectively, and pressed into thin sheets by a manual tablet press to prepare a series of standard samples. Then, LIBS was used to collect spectra of different ablation times, with a delay time of 0.7 microseconds and a pulse width of 3.0 microseconds. Finally, a laser confocal microscope was used to test the depth of the sample ablation pit. The obtained depth prediction model is shown in Figure 2. The spectrum of the sample to be tested was collected and the ablation depth value was predicted. Figure 3 is the depth prediction result of the tungsten and molybdenum sample to be tested in this embodiment.

Then, the LIBS spectrum of the sample to be tested is analyzed using the calibration method to obtain the element content of the sample to be tested; in this embodiment, 20 molybdenum spectral lines and 21 tungsten spectral lines are selected, a calibration curve is prepared, and the optimal calibration curve is screened out through five cross-validation methods. The optimal calibration curve is y=7.200x+1.025, where y is the ratio of the intensity of the molybdenum 557.0 nm spectral line to the tungsten 498.2 nm spectral line, and x is the ratio of the molybdenum content to the tungsten content. The screened calibration curve is shown in Figure 4. The depth distribution of the tungsten and molybdenum content finally obtained in this embodiment is shown in Figure 5.

(ii) A longitudinal depth element quantitative analysis method based on laser induced breakdown spectroscopy with different depth distributions of tungsten and molybdenum content.

Ten different proportions of tungsten and molybdenum powders were uniformly mixed, with the proportions of tungsten and molybdenum being 72.4%, 27.6%; 74.3%, 25.7%; 76.4%, 23.6%; 78.9%, 21.1%; 81.4%, 18.6%; 83.8%, 6.2%; 86.7%, 13.3%; 89.7%, 10.3%; 92.9%, 7.1%; 96.2%, 3.8%, and pressed into thin sheets respectively. Then LIBS was used to collect spectra of different ablation times, and finally a laser confocal microscope was used to test the depth of the sample ablation pit. The depth prediction model obtained is the same as (a).

Figure 6 shows the depth prediction results of tungsten and molybdenum samples in another sample to be tested with different element distribution content. 20 molybdenum spectral lines and 21 tungsten spectral lines were selected, calibration curves were prepared, and the optimal calibration curve was screened out through five cross-validation methods. The optimal calibration curve is y=7.200x+1.025, y is the intensity ratio of the 557.0 nm spectral line of the molybdenum element and the 498.2 nm spectral line of the tungsten element, and x is the ratio of the molybdenum element content to the tungsten element content. The optimal calibration curve and the LIBS spectrum collected in step S21 are used to calculate the corresponding tungsten and molybdenum element contents; finally, the ablation depth prediction results and the tungsten and molybdenum element content calculation results are used to predict the depth distribution of the molybdenum element content. As shown in Figure 7, the depth distribution of the tungsten and molybdenum element contents in the sample to be tested in this embodiment is shown.

(III) A longitudinal depth element quantitative analysis method based on laser induced breakdown spectroscopy of tungsten samples infiltrated by lithium diffusion.

LIBS was used to collect spectra of pure tungsten samples with different ablation times, with a delay time of 0.7 μs and a pulse width of 3.0 μs. A confocal microscope was used to test the depth values of the ablation pits under different times of laser pulses on the standard sample. The depth was fitted and the regression curve of the depth prediction model was shown in FIG8 .

LIBS was used to collect spectra of the sample under test with different numbers of laser pulses, and then the depth prediction model was used to predict the ablation depth corresponding to the LIBS spectrum of the sample under test; the prediction results were shown in Figure 9. The non-calibrated method was then used to quantitatively calculate and analyze the element content of the LIBS spectrum collected by the sample under test, and the predicted ablation depth and the target element content were used as the horizontal and vertical coordinates, respectively, to predict the depth distribution of the target element content in the sample under test. Figure 10 shows the depth distribution of tungsten and lithium element contents.

(iv) Longitudinal depth elemental quantitative analysis method based on laser induced breakdown spectroscopy of tungsten samples infiltrated by lithium diffusion.

LIBS was used to collect spectra of pure tungsten samples at different ablation depths, with a delay time of 0.7 μs and a pulse width of 3.0 μs. Finally, a laser confocal microscope was used to test the depth of the sample ablation pit. The depth prediction model obtained by fitting was the same as that in (III).

The depth prediction model is used to predict the ablation depth of another tungsten sample to be tested that has been diffused and infiltrated by lithium, and the prediction results are shown in Figure 11. The tungsten and lithium content in the sample to be tested is measured using the non-calibrated method. Finally, the depth distribution of the element content is predicted using the ablation depth prediction results and the tungsten and lithium element content calculation results, and the prediction results are shown in Figure 12.

Those skilled in the art will appreciate that the embodiments described herein are intended to help readers understand the principles of the present invention, and should be understood that the protection scope of the present invention is not limited to such specific statements and embodiments. Those skilled in the art can make various other specific variations and combinations that do not deviate from the essence of the present invention based on the technical revelations disclosed by the present invention, and these variations and combinations are still within the protection scope of the present invention.

Claims (2)

1. The longitudinal depth element quantitative analysis method based on the laser-induced breakdown spectroscopy is characterized by comprising the following steps of:

S1, training ablation depth data of LIBS corresponding to a standard sample by using a partial least square method to obtain a depth prediction model; the method specifically comprises the following substeps:

S11, preparing a standard sample: designing a series of element content ratios of a standard sample according to element components of the sample to be tested, uniformly mixing a series of powder raw materials with different element proportions, and respectively pressing the powder raw materials into slices through a manual tablet press to obtain a series of standard samples with different element contents;

s12, LIBS spectrum acquisition: ablation of different pulse times is carried out on the standard sample by using LIBS technology, and spectrum acquisition is carried out;

S13, obtaining an ablation depth value: determining ablation depth values of the S11 standard sample under different pulse times by using a confocal microscope;

s14, establishing an ablation depth prediction model: selecting a plurality of spectral lines of the element with the highest content in the LIBS spectrum acquired in the S12, and taking a vector formed by the intensity of the plurality of spectral lines as input data of a partial least square method; taking the ablation depth value obtained in the step S13 as a label of a partial least square method; fitting the input data and the label to obtain a depth prediction model;

S2, analyzing quantitative depth distribution of the sample element to be detected; the method specifically comprises the following substeps:

S21, respectively ablating the sample to be detected by using LIBS technology for different pulse times, and collecting spectra;

S22, predicting an ablation depth value of the sample to be detected through the spectrum acquired in the S21 and the depth prediction model obtained in the S14;

S3, analyzing the element content in the sample to be tested through a calibration method or a non-calibration method;

S4, quantitatively analyzing the element depth distribution of the sample to be tested through the depth value predicted in S22 and the element content calculated in S3: and (3) mapping the ablation depth value predicted in the step (S22) and the content of the target element calculated in the step (S3) as an abscissa and an ordinate respectively to obtain the content depth distribution of the target element in the sample to be detected.

2. The quantitative analysis method of longitudinal depth elements based on laser induced breakdown spectroscopy according to claim 1, wherein the scaling method specifically comprises: selecting a plurality of spectral lines with clear outlines in the LIBS spectrum in S12, and fitting by using the spectral line intensity of the target element and the content of the target element, or fitting by using the intensity ratio of the target element and the content ratio of the target element, so as to fit different calibration curves; then, verifying the fitted calibration curve by adopting a multiple-time cross verification method, and selecting the calibration curve with the minimum average percentage deviation as an optimal calibration curve; and finally, the spectral line intensity of the target element to be calculated or the intensity ratio of the target element is brought into an optimal calibration curve to obtain the content of the target element or the content ratio of the target element.