1)深圳大学物理与光电工程学院,光电子器件与系统教育部/广东省重点实验室,广东深圳 518060; 2)中国科学院深圳先进技术研究院光电工程技术中心,广东深圳 518055
1)College of Physics and Optoelectronic Engineering, Key Laboratory of Optoelectronic Devices and Systems of Ministry of Education and Guangdong Province, Shenzhen University, Shenzhen 518060, Guangdong Province, P.R.China2)Center of Optoelectronic Engineering Technologies, Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences, Shenzhen 518055, Guangdong Province, P.R.China
optical engineering; laser induced breakdown spectroscopy(LIBS); Raman spectroscopy(RS); remote detection; 2-dimensional mapping; mineral composition analysis
DOI: 10.3724/SP.J.1249.2019.05538
发展用于远程探测物质成分的激光诱导击穿光谱仪(laser induced breakdown spectroscopy, LIBS)和拉曼光谱复合检测系统,其能实现以2 mm的扫描精度对二维区域进行扫描,并在30 m的最大距离处检测样品中的元素含量及其表面分布. 通过二维扫描系统分析30 m距离处硅灰石表面的Fe元素分布. 根据LIBS与拉曼信号的时间差异,采用门控增强型CCD相机分别采集LIBS和拉曼光谱信号. 实验结果表明,该系统可用于对铝合金的分类识别,并可用于合金分选过程,从而节约大量的资源和能源. 该系统还可用于远程矿物识别,矿物和岩石可以通过拉曼光谱分析得出分子信息,如C—O、S—O和Si—O的拉伸模式,从而区分出矿物和岩石的种类,即碳酸盐、硫酸盐和硅酸盐等.
We develop a laser induced breakdown spectroscopy(LIBS)and Raman spectroscopy(RS)combined detection system for remote detection of substance composition. It can scan a 2-dimensional region with scanning accuracy of 2 mm and detect element contents and elementary surface distribution in samples at the maximum distance of 30 m away from the detector. Using our 2-dimensional scanning system, we analyze Fe element distribution on the surface of wollastonite at a distance of 30 m. According to the time difference between LIBS signal and Raman signal, both LIBS and Raman spectrum signals are collected by a gated intensified CCD(ICCD)camera. The experimental result shows that the system can be applied to identify the type of aluminum alloys and also be used in alloy sorting process to save resources and energy. The system can also be applied in remote mineral identification. Molecular information such as C—O, S—O and Si—O stretching patterns can be obtained from Raman spectroscopy analysis of minerals and rocks, thus distinguishing minerals and rocks, i.e. carbonates, sulphates and silicates.
Laser induced breakdown spectroscopy(LIBS)is becoming a very popular method for analyzing elemental composition in last few decades[1-2], while Raman spectroscopy(RS)has already been well known as a powerful tool for chemical analysis[3]. These two techniques both utilize high-intensity laser to excite samples and analyze spectroscopic information during sample illumination. The advantages of these two techniques, such as remote measurement capability, simple or no sample preparation and rapid analysis, have led to their wide-spread applications in different fields[3-6].
The remote sensing potential of LIBS and RS is probably a major driving force for using them in military applications in detecting explosives, chemical and biological weapons, etc. Other important applications, in which elemental as well as molecular information is needed, are remote sensing of minerals and analysis of art objects[4-6]. For example, LIBS spectrum of a copper ore reveals the presence of Ti, while information about polymorphic forms of TiO2 is provided by Raman measurement[5]. Stand-off option for measurement is exceptionally useful for dangerous situations, fragile objects or environments difficult to access.
However, Raman scattering signal is weak, and stand-off distance exacerbates the process of signal acquisition. Increasing the sensitivity and detection power is the principal goal in any analytical system, and therefore, a substantial amount of effort was devoted to achieving it for LIBS and RS. In terms of LIBS signal optimization, two major points should be noticed. One is the selection of wavelength, the other is the use of photo-detector coupled to a spectrometer. The choice of wavelength is usually determined by the material[7]. For example, ultraviolet(UV)emission is the best for ceramics and metals, while visible light would be better for water solution because of its strong absorption in the UV and infrared spectra. Therefore, small and compact Q-switched Nd:YAG laser with millijoule output and capability of changing wavelength from UV to visible or even infrared(IR)range is a good choice as LIBS excitation source. Detectors should be affordable to provide measurements in real-time or in single-shot mode within broadband spectral range. Although non-intensified CCD can meet the requirement, use of intensified CCDs is beneficial to maximize the signal-to-noise ratio in LIBS spectra[8] due to the capability of selecting time interval and delay of measurement(gated ICCD option).
Two-dimensional scanning function is essential to a remote detection system. HOEHSE et al.[5] showed that LIBS-Raman mapping of heterogeneous materials is beneficial for a comprehensive material characterization. Two-dimensional mapping remote detection system can also be applied in other high-hazard operation circumstances, such as high temperature furnace monitoring and maintenance of high voltage insulator.
In this paper, we demonstrate the capability to combine remote LIBS and RS at stand-off distance. We use a 2-dimensional scanning system to analyze Fe element distribution on the wollastonite surface at a distance of 30 m. To achieve optimal sensing in this dual-mode technique, we design the setup based on the component with advanced parameters. To enhance technical potential of our setup, we use a laser source which operates at two fixed wavelengths, i.e. IR(1 064 nm)for LIBS mode and visible(532 nm)for RS mode. The wavelength switching is designed to provide optimal excitation for both LIBS and RS signals. Finally, to maximize the signal-to-noise ratio of LIBS signal and to increase the contrast between emission lines and continuum, a gated ICCD is used as the detector.
/pic57.jpg)
Fig.1 (Color online)The coaxial excitation, collection optical path and 2-dimensional scanning for the LIBS and RS combined remote detection system
Our LIBS and RS combined remote detection system mainly consists of three parts, the coaxial excitation and collection optical path, the 2-dimensional scanning system and the portable spectral analysis and processing system. The coaxial excitation and collection optical path contains a nanosecond laser, a 5 times beam expander and a Cassegrain telescope. The laser is an Nd:YAG laser(Nimma- 600, Beamtech)operating at 532 nm(300 mJ)and 1 064 nm(600 mJ)respectively with pulse duration of approximately 10 ns and repetition rate of 10 Hz. The beam is guided into the coaxial excitation and collection optical path through three 1 064/532 nm HR mirrors as shown in figure 1. A customized 5× beam expander is used to focus the laser to a remote distance. By adjusting the optical length of the beam expander, we acquire a 5 to 30 m focusing distance. LIBS and RS signals are collected by a Schmidt-Cassegrain telescope(C6-A-XLT-CG-5, Celestron)with diameter of 5 inches and delivered by a UV/VIS optical fiber to the imaging spectrometer(MS3504i, SOL Instruments)through the long pass filters for rejection of Rayleigh lights.
The two-dimensional scanning system is fixed on two rotating displacement tables, one is for horizontal rotation and the other is for vertical rotation. During a mapping procedure, wherever the rotating displacement table goes, the laser excitation path is coaxial with the spectrum collection path.
The portable spectral analysis and processing system contains a spectrometer, an ICCD, a laptop and a portable case. The spectrometer is equipped by turret with 4 gratings(300 and 600 grooves/mm with 400 nm blaze, 1 200 and 1 800 grooves/mm with 500 nm blaze). This option is designed for aligning the spectrometer to convenient wavelength range with proper spectral resolution. A gated ICCD camera(iStarDH320T, Andor)with the delay control from 0 to 10 s in 10 ps steps is attached to a photo-detector part of the imaging spectrograph.
Figure 2 shows the time scale difference between Raman scattering and LIBS. Based on such a difference, RS and LIBS signals could be separated by a gated ICCD camera. With Rayleigh scattering set at the initial time point, the camera receives Raman scattering signal in the first 30 ns, while from 1.5 to 3.5 μs, the camera acquires the LIBS signal. As shown in figure 2, dolomite is selected as a model to demonstrate the mechanism of separately detecting Raman signal and LIBS signal.
/pic58.jpg)
Fig.2 (Color online)Raman spectra and LIBS acquisition through time gate controlled by ICCD
The 2-dimensional scanning system contains two high-precision rotating machinery with a minimum rotating angle of 0.003 8°(2 mm step length at a distance of 30 m), both horizontally and vertically(as shown in figure 1). The laser is set up on the horizontal rotating machinery and the laser is imported to the vertical rotating machinery by its rotating shaft, so that the excitation and collection optical paths can automatically overlap and pass through the coaxial optical path for each 2-dimensional detection spot. Thus 2-dimensional scanning can be performed without adjusting the optical overlap at every detection spot.
To enable the LIBS and RS combined remote detection system for applications in different environments, the chassis is designed to be mobile. The diagram and the prototype of the whole system are shown in figure 3.
/pic59.jpg)
Fig.3 (Color online)Diagram of the combined remote LIBS and RS detection system and photo of prototype
We use the 2-dimensional scanning system to analyze material composition of wollastonite as shown in figure 4(a). The scanning area is 12×12 mm2, with 2 mm scanning accuracy. The detection distance is 30 m. Each excitation point leaves a 1 mm diameter spot on the surface, as shown in figure 4(a).The sample surface is basically green and white, which indicates element distributes heterogeneously. We detect a white spot(X10Y8)and a green spot(X6Y12)to obtain LIBS signals(figure 4(b)). Both spots contain element Ca, Si, Sr and Mn, while only green spot of the sample contains element Fe and Al. We adopt a spectral intensity separation method to emphasize the difference of the element contents: LIBS(X6Y12)is divided by LIBS(X10Y8)(figure 4(c)). As element Si is distributed uniformly in wollastonite, Si spectral intensity at 390.5 nm is selected as the denominator by which Fe spectral intensity at 404.5 nm is divided. Surface distribution mapping of element Fe is then achieved by interpolating from a single detection spot(figure 4(d)). In the 2-dimensional plot, white area indicates low content of element Fe, while green area indicates high content of Fe element. According to the surface mapping, element Fe is mostly distributed in the central part of the scanning area, which is consistent with the green color distribution in the photo(figure 4(a)).
/pic60.jpg)
Fig.4 (Color online)2D scanning experiment of remote LIBS
According to this experiment, a two-dimensional mapping of elementary distribution is completed by the LIBS and Raman spectroscopy combined remote detection system with a minimum spatial resolution of 2 mm at a distance of 30 m. Two-dimensional mapping remote detection system can also be applied in other high-hazard operation circumstances, such as high temperature furnace monitoring and maintenance of high voltage insulator.
In current alloy recycling process, alloys are not sorted by type, which causes massive waste of energy and resources. The LIBS detection system can be applied in alloy recycling business. Al alloy 3004, 5052, 5182, 6061, 6063 and 7010 are randomly placed on a board 15 m away. Through the 2-dimentional scanning detection, a series of Al alloy LIBS are obtained, as shown in figure 5. Analysis result is shown in table 1. Al alloy 3004 is known as Al & Mn alloy, Al alloy 5052 and 5182 are Al & Mg alloy, Al alloy 6061 and 6063 are Al & Mg & Si alloy, and Al alloy 7010 is Al & Mg & Zn alloy. All LIBS spectra peak assignments are citing from the NIST:atomic spectra database[17]. According to the results in table 1, Al alloy 3004 is identified by the high percentage of Mn and low percentage of Mg. Al alloy 5052 and 5182 are identified by the significantly high percentage of Mg. Al alloy 6061 and 6063 are identified by the medium percentage of Mg and Mn, and Al alloy 7010 is identified by the high percentage of Zn. The only flaw of these results is that LIBS does not identify the presence of element Si, which is a direct evidence of identification of series Al alloy 6061 and 6063.
/pic61.jpg)
Fig.5 (Color online)LIBS spectra of a series of Al alloys, detected at a distance of 15 m
/pic62.jpg)
Table 1 Detected elements of series of Al alloys and the spectral intensity ratios of Mg/Al and Mn/Al
Figure 6(a)shows Raman spectra of minerals and rocks which are all detected at a distance of 15 m. All Raman spectrum peaks are classified in table 2. The RS peak of apatite is at 1 071 cm-1, which is originated from the stretching mode of carbonate, v1(CO3)2-[5]. Dolomite and calcite are the most common carbonate minerals. Their strong peaks are located at 1 098 cm-1and 1 070 cm-1 respectively, which are all assigned as v1(CO3)2-[9-10]. Calcite has a trigonal structure with two molecules per unit cell, and dolomite has a hexagonal structure. This is more likely to cause the splitting and distorting of the carbonate groups. Another cause for the difference is the cation substituting for Mg in the dolomite mineral[10]. RS peaks of plagioclase and diopside are assigned as Si—O stretching modes[11-12]. In the spectra of gypsum and barite, peaks present around 440 cm-1, 600 cm-1, 970 cm-1 and 1 140 cm-1, which are originated from ν2(SO4)2-, ν4(SO4)2-, ν1(SO4)2- and ν3(SO4)2-(symmetric stretching of SO4 tetrahedra)respectively[11,14]. 299 cm-1 is identified as Mg—O stretch mode[16]. 455 cm-1 is identified as T—O(T=Al or Si)[15]. 620 cm-1 is identified as Si—Obr stretch mode[16]. 266 cm-1 is identified as translatory oscillations of(CO3)[16]. Therefore, minerals and rocks can be easily distinguished as carbonates, sulphates or silicates by Raman spectra. Because RS provides molecular information such as stretching modes of C—O, S—O and Si—O, RS can also contribute to the identification of non-metallic elements, which is relatively difficult for LIBS.
/pic63.jpg)
Table 2 Raman spectra peak positions of minerals and rocks
/pic64.jpg)
Fig.6 (Color online)Testings of remote Raman spectra and LIBS detection system
Figure 6(b)shows the LIBS spectra of minerals and rocks which are all detected at a distance of 15 m. All the LIBS peaks are assigned according to the NIST:atomic spectra database[17]. With combined information on LIBS and RS, the mineral components are identified, as shown in table 3. LIBS provides the atomic composition information, while the RS provides the molecular vibration and rotation information. For instance, RS of both barite and gypsum shows the(Si—O)stretch mode. It is difficult to distinguish one from the other. While LIBS can easily identify the Ba element for barite and the Ca element for gypsum. Both the LIBS and RS are essential for the remote detection system.
/pic65.jpg)
Table 3 Combined LIBS and Raman spectra of minerals
The complementary natures of LIBS and RS mentioned above validate the necessity of combining LIBS and RS in one single remote detection system. The results show that with the combined system, we can perform identification of both metallic and non-metallic stand-off samples using LIBS and RS simultaneously, and the system can also be applied for analysis in different environments.
A LIBS and RS combined remote detection system is developed, which is capable of scanning detection of elemental and molecular information of samples at a maximum distance of 30 m away from the detector in a 2-dimensional area with 2 mm scanning accuracy. The detection distance can be adjusted from 5 to 30 m. A gated ICCD camera is applied to acquire both LIBS signal and Raman spectrum signal according to their time difference. The system has been applied to analyze different samples in different environments. Two examples have shown that the combined system can not only be used to identify the type of Al alloys in the alloy sorting process for saving abundant resources and energy, but also be applied in mineral identification from a stand-off distance. The component analysis of minerals and rocks can easily be detected by the remote combined LIBS and Raman spectra detection system. RS provides molecular information and helps the identification of non-metallic elements, serving as a complement for LIBS.
深圳大学学报理工版
JOURNAL OF SHENZHEN UNIVERSITY SCIENCE AND ENGINEERING
（1984年创刊 双月刊）
主 管 深圳大学
主 办 深圳大学
编辑出版 深圳大学学报理工版编辑部
主 编 李清泉
国内发行 深圳市邮电局
国外发行 中国国际图书贸易集团有限公司（北京399信箱）
地 址 北京东黄城根北街16号
邮 编 100717
电 话 0755-26732266
0755-26538306
Email journal@szu.edu.cn
标准刊号 ISSN 1000-2618
CN 44-1401/N
/qkfm.jpg)