Way-out for laser-induced breakdown spectroscopy

Ever since its creation in 1963 [1], laser-induced breakdown spectroscopy (LIBS) has gained considerable attention due to its unique capability for real-time, in situ or online analysis [2, 3]. The future world is heading into the age of artificial intelligence (AI), and data would be the most valuable asset for human society [4]. In terms of real-time, all-elements (a simultaneous analysis feature of LIBS), in addition to the suitability for solid, liquid, and gas analysis, plus remote sensing, LIBS, to some extent, is a back-window in exploring and improving our world by easily providing a tremendous amount of elemental spectral big-data and is able to increasingly contribute more with technological and societal development. That is, LIBS may play a more important role than we expect depending on the quantification performance of LIBS. The higher the performance, the more important the technology to future human society.

In 2000, the 1st International Conference of LIBS was held in Pisa, Italy, symbolizing the coming of high-speed development for LIBS. The great potential of LIBS attracted Asian researchers as well. The 1st Chinese Symposium on LIBS (CSLIBS 2011), hosted in Qingdao, and the 8th International Symposium on LIBS (LIBS 2014), held in Beijing, greatly stimulated the establishment of the Asian LIBS community and Asian Symposium on LIBS (ASLIBS). In 2015, the 1st Asian Symposium on LIBS was hosted in Wuhan, marking the birth of the Asian LIBS community. With the rapid development of the Asian LIBS community, we have successfully become one of the three major LIBS communities in the world.

From the beginning of the ASLIBS meeting, it has become a tradition to publish a special issue to witness the development of the Asian LIBS community and technology [5, 6]. Plasma Science and Technology started the connection with ASLIBS2017 which was held in Tokushima, Japan [6]. The present issue collected 15 articles for the 3rd Asian Symposium of LIBS (ASLIBS2019) which was held in Jeju, Republic of Korea. These articles, together with other reports in the ASLIBS2019, provide a picture of the current research interests from the Asian LIBS community. The work published in this collection was summarized as below:

There are four articles on LIBS fundamentals. Two articles investigated the fundamentals of underwater LIBS, one studied the water pressure effects on LIBS [7] and the other studied the water temperature effects [8]. The other two articles on fundamentals investigated the influence of target temperature on LIBS [9] and the plasma evolution of hydrogen retention in tantalum [10], respectively. In this issue, an introduction work of a mobile prototype fiber-optic LIBS (FO-LIBS) is included [11]. This prototype was applied to rapidly online analyze the trace elements in large diameter steel tube. Results demonstrated that the performance of this mobile FO-LIBS prototype can be compared with conventional LIBS systems. In addition, there are ten articles regarding data processing and different applications. A series of different new data processing methods were proposed, such as variable selection [12–14], principal component analysis Mahalanobis distance [15], combination of support vector machine and partial least square [16], as well as the classification method based on the elemental intensity ratio [17]. It is worth mentioning that variable selection is an important way to improve the quantification performance and has attracted more interest [12–14]. The application not only included steel classification [17], aluminum alloy analysis [14], coal analysis [12, 16], geographical authenticity evaluation [18], and plants classification [15], but also included some new or uncommon application fields such as uranium measurements [19], mercury measurements [20], soil pH measurements [13], and rock salt analysis [21]. Overall, in ASLIBS, there was a wide range of LIBS applications presented, which was quite reasonable considering large market potentials exist in Asian countries, especially in China [22]. There are also some highlight works including uncertainty generation understanding [23], self-absorption correction/reduction [24, 25], and long-short double pulse LIBS [26], the machining learning quantification method [27, 28] and so on, attracting much attention. The Asian LIBS community has made solid and rapid progress in pushing LIBS forward.

However, the time for wide commercial application for LIBS has not come yet. It is well recognized that the main obstacle is its quantification performance, which was called as the Achilles' heel for LIBS [3]. To fully develop the technology from the so called 'future superstar' into a real superstar, there are many things that must be done, which may take a long time without a good strategy and organization. From the editors' point of view, the only way-out for LIBS's development is to unify more researchers to focus on key problems and make key breakthroughs. Here are some key issues we want to propose for further discussion:

• (1)  
  Relatively low repeatability was the number one problem causing low quantification performance for LIBS. There is no way-out for LIBS wide applications without solving this problem. Understanding the uncertainty generation mechanism is, of course, a key way-out. For normal nano-second LIBS, plasma experiences an extremely quick temperature increase and decease with spatial inhomogeneity since in the plasma evolution, the plasma goes through an ultra-fast expansion and mixes with surrounding gases, accompanied with the generation of a shockwave. All these quickly varied but uncontrollable processes affected the plasma evolution and LIBS spectra in a great and complicated manner; making it challenging to understand the mechanism of signal uncertainty generation. It has been pointed out in our previous work that the contribution to signal uncertainty mainly comes from plasma morphology fluctuation [23], and what we need to investigate next is the mechanism leading to the fluctuation of plasma morphology. It was believed that with deeper understanding to uncertainty generation mechanisms, more ways would be proposed to produce more repeatable laser-induced plasma.
• (2)  
  It has been widely accepted that matrix effects are the key to affect LIBS measurement accuracy, but how the matrix effect works and how to control matrix effects remain obscure. Concentrated work should be conducted to explain the key mechanism or key process that leads to severe matrix effects. We believe that the existence of large measurement uncertainty has a huge effect on measurement accuracy, and uncertainty generation would be possibly one of the key mechanisms for matrix effects. The correlation between plasma evolution and matrix effects certainly needs much more work to clarify.
• (3)  
  Modern AI algorithms such as deep learning and transfer learning have shown great advantages in image processing, speech recognition, and so on. LIBS was regarded as future technology for the big-data age, and naturally, AI technology should be introduced into LIBS quantifications. To achieve better results, it would be better to combine the data-driven AI technology with physical principle based traditional models.