Study On Coaxial Remote Laser Induced Breakdown Spectroscopy

Since the beginning of the 21 st century,with the advancement of science and technology,the exploration of mineral resources and the detection of hazardous materials have achieved unprecedented development.Laser-induced breakdown spectroscopy(LIBS),an emerging technology of material element analysis,has attracted more and more attention due to the advantages of rapid,non-destructive,sensitive,and real-time analysis.With the continuous in-depth research of material detection technology,people pay more attention to the practicability of detection methods in a variety of complex environments.Besides,they request more portable and more accurate detection methods.At present,most of the research on LIBS technology is the short-distance measurement,which is carried out by the laboratory.The samples need to be brought back to the laboratory for analysis,which cannot fully reflect the non-contact and real-time measurement of LIBS.Especially in industrial production,samples often have a high temperature of thousands of degrees,and it is difficult for operators to bring samples back to the laboratory;When detecting explosives,the samples may explode at any time during transportation,causing potential safety hazards.These situations may cause sampling failure or obtain limited number of samples,which will adversely affect the safety of related operators.Remote LIBS technology can take full advantages of long-distance non-contact measurement,real-time analysis of LIBS.It can complete the analysis of substances without sampling,and can quickly analysis material in extreme environments.The work of this article mainly includes the following three aspects.1.Development of remote LIBS device.In this paper,a set of coaxial remote LIBS device with continuous zoom is developed.The device uses two 1064 nm Nd:YAG solidstate lasers as laser sources.In order to measure samples at different distances,a 3-lens zoom system is designed to achieve 1-40 m continuous zoom.A set of Cassegrain reflection telescope is used to collect the spectrum.2.Single pulse remote LIBS.In the experiment,in order to restore real environments,this paper does not do any surface treatment for the experimental samples.In order to measure the detection ability of a single pulse,the metal standard samples and titanium alloy samples are detected with the same laser energy at a distance of 5 m,10 m,15 m and 20 m.The relationship of the signal strength and distance is studied in the long-distance measurement.For different grades of titanium alloy samples,their LIBS spectrals are so similar that we cannot effectively identify and classify them.In order to obtain a higher classification accuracy,it is necessary to combine machine learning algorithms to classify the spectrum.For alloy materials,the concentration of different elements directly affects the physical properties and chemical properties of the material.In quantitative analysis,the jitter of operating conditions will cause the experimental results to deteriorate.The internal standard method can eliminate errors caused by operating conditions.Therefore,the internal standard method is selected to analyze the concentration of different elements in the samples.3.Double pulse LIBS experiment.In the long-distance measurement,the power density of laser at the focal point will decrease,and the plasma collection solid angle will also decrease.Therefore,the intensity of collected spectral lines is low and they are mixed with the background,which cannot be effectively distinguished.In this paper,a collinear double pulse technology is used to enhance the spectrum without increasing the laser energy.In order to obtain the strongest spectrum under the double pulses,parameters are optimized so that the two light pulses are recombined at the focal point in the optimal delay.Comparing with the quantitative analysis curves of single pulse at the same distance,the relative standard deviation(RSD)of double pulse LIBS is lower and the detection limit is also lower.When classifying the spectrum of titanium alloy collected at 20 m,the intensity of the spectrum fluctuates greatly due to laser fluctuations,resulting in a low classification accuracy.In order to obtain a higher classification accuracy,before classification,the PCA algorithm is used to reduce the dimensionality of the spectrum,and the KNN algorithm is used to classify the processed data.When the number of principal components is selected as6,99% of the classification accuracy is achieved.In summary,the device is not only suitable for online analysis of industrial metallurgy,but also can remotely detect organics and explosives.