Arc-plasma-assisted Laser-induced Breakdown Spectroscopy (AP-LIBS): A Study on Signal Enhancement and Spatiotemporal Distribution

Abstract

This study investigated the fundamental aspects of signal enhancement in arc-plasma-assisted laser-induced breakdown spectroscopy (AP-LIBS), as a crucial step towards its potential application for enhanced real-time compositional analysis in electric arc furnaces (EAF). By superimposing a sustained arc discharge with nanosecond laser pulses on molten iron, AP-LIBS achieved significant signal enhancement compared with conventional LIBS. Spatiotemporal characterizations revealed that the enhancement was most pronounced in the peripheral plasma region, characterized by larger plasma size and longer lifetime in AP-LIBS setups. The enhancement factor η, defined as the ratio of AP-LIBS signal intensity to the sum of individual arc and laser-induced plasma intensities, exceeds 10 for most emission species. Spatial distribution analyses show increased emission intensities at greater distances from the laser spot in AP-LIBS, in contrast to the decay observed in standard LIBS. Temporal analysis demonstrated extended high-intensity periods for AP-LIBS compared to the rapid decay in conventional LIBS techniques. The spatiotemporal behavior of the enhancement factor varies significantly among the emission species, thereby providing insights into complex plasma dynamics. Elements with low vapor pressure and ionic species generally exhibited higher enhancement, whereas elements with high vapor pressure exhibited limited enhancement, indicating minimal additional evaporation effects for high vapor pressure element. These findings provide valuable insights into plasma generation and maintenance mechanisms in AP-LIBS, suggesting its potential for improved sensitivity in elemental analysis for electric arc furnace applications.

1. Introduction

The transition of the steel industry toward carbon neutrality is driving the shift from blast furnaces to electric arc furnaces (EAFs). This change demands a higher processing capacity and advanced compositional control in EAFs, particularly when handling new iron sources, such as hydrogen-reduced iron. Consequently, real-time compositional analysis^1,2,3,4,5,6) and modeling^7,8,9,10) techniques for EAFs have become increasingly crucial.

Current composition analysis techniques for molten metal and slag, such as spark or glow discharge optical emission spectrometry and X-ray fluorescence (XRF), are relatively accurate compared with real-time measurements that involve time-consuming sampling and preparation processes. However, these methods cannot provide the rapid in situ analysis required for the real-time process control of EAFs. Studies have explored arc emission spectroscopy by leveraging a continuous high-energy arc plasma in EAFs.^1,2,3,4) However, this approach faces challenges such as spectral interference, signal distortions from slag foaming, and arc plasma instability.

Laser-induced breakdown spectroscopy (LIBS) has emerged as a promising technique for elemental analysis in the steel industry.^{11,12,13,14,15,16,17,18,19,20,21)} The technique offers quantitative analysis of the emission spectra generated by laser-induced plasma;²²⁾ its simple configurations and remote measurement capabilities render it attractive for EAF applications. However, conventional LIBS faces limitations in maintaining consistent plasma conditions and achieving sufficient sensitivity.²²⁾

Various techniques have been developed to enhance the performance of LIBS, including double-pulse LIBS (DP-LIBS),^23,24,25) spark-discharge-assisted LIBS (SD-LIBS),^26,27,28,29) and microwave-assisted LIBS (MW-LIBS).^30,31,32) These methods share the common approach of superimposing other energy sources on the main laser to improve surface cleaning, preliminary vaporization, plasma energization, and plasma lifetime extension.

Building on these advancements and considering the unique environment of EAFs, we propose a technique called arc-plasma-assisted LIBS (AP-LIBS). This method combines the continuous high-energy arc plasma, which is inherently present in EAFs, with the precise and pulsed nature of laser-induced plasma. AP-LIBS aims to leverage the synergies between these two energy sources to overcome the limitations of conventional arc spectroscopy and conventional LIBS.

The proposed AP-LIBS technique shares similarities with some of the advanced LIBS methods mentioned earlier. For instance, SD-LIBS utilizes a laser-induced spark discharge, which is similar to our proposed approach of combining a laser with a discharge plasma. SD-LIBS has reported signal enhancement factors of up to 223 times,²⁹⁾ although it cannot provide effects such as pre-laser surface cleaning. Similarly, in DP-LIBS, the long-short double-pulse LIBS (LS-DP-LIBS) extends the plasma lifetime by overlapping a longer-pulse-width laser with a nanosecond-order laser pulse.^{33,34,35,36,37)} This concept of superimposing energy sources of different timescales, as employed in LS-DP-LIBS, is analogous to the approach adopted in the proposed AP-LIBS technique.

In AP-LIBS, the continuous arc discharge provides a stable background and surface cleaning, and the laser pulse offers precise, localized excitation. This synergy can lead to more consistent plasma conditions, improved measurement reliability, and a reduced matrix effect. The superposition of the arc discharge, which operates on a longer timescale than the spark discharge, is expected to extend the plasma lifetime beyond that of SD-LIBS, similar to the extension observed in LS-DP-LIBS.^33,34) This extended lifetime enabled longer integration times and potentially improved the signal-to-noise ratios and quantitative accuracy. Furthermore, the combination of arc and laser excitation can provide a higher overall energy input, which potentially improves the sensitivity and lowers the detection limits for elements that are typically difficult to detect in a complex EAF environment.

This study aimed to explore the feasibility of utilizing AP-LIBS for the compositional analysis of EAFs by elucidating the signal enhancement effect and its spatiotemporal distribution during arc superposition.

2. Experimental Setup

2.1. Sample Preparation and Furnace Configuration

An experimental setup was constructed to melt raw materials in a graphite crucible inside a horizontal tube furnace with resistance heating. This setup generates an arc on the molten steel surface while simultaneously irradiating it with a nanosecond pulsed laser. Figure 1 shows a schematic of the experimental apparatus.

Fig. 1. Schematic of the experimental setup. (Online version in color.)

For the molten iron raw material, representative tramp element Cu and Sn plates (Nilaco Corp.) were placed under a stainless-steel (JIS-SUS303) plate. A graphite reagent was added to suppress melting of the graphite crucible. The total mass of the sample was 8.0 g. Table 1 lists the expected concentrations of the elements in the sample. It should be noted that these concentrations are estimated values based on the initial composition of the materials used, not the result of a detailed analysis. This study focused on signal enhancement in AP-LIBS rather than precise composition estimation; therefore, a detailed compositional analysis was not performed. The samples were heated and melted at 1773 K in an argon atmosphere and left to stand for more than 1 h before conducting the experiment.

Table 1. Expected compositions of molten sample (wt%).

Fe	Cr	Ni	C	Mn	Si	P	S	Cu	Sn
65.0	16.0	8.2	4.0	0.9	0.5	0.1	0.1	4.6	0.6

2.2. Arc Generation

During arc generation, Ar (99.99%) was supplied at 1.0 SLM (Standard Liters per Minute at 298 K, 0.1 MPa) into the horizontal tube. A direct current of 12 A was applied using the upper graphite electrode as the cathode and molten iron as the anode. The voltage was approximately 20 V, resulting in an average arc power consumption of 240 ± 10 W. The arc length was set to 4 mm approximately.

The arc was ignited by focusing the pulsed laser onto the molten steel surface to generate a laser-induced plasma. In arc-only experiments, the laser pulse was stopped after ignition, which resulted in a sustained arc.

2.3. Laser System

A Q-switched frequency-doubled Nd:YAG laser (Amplitude, Surelite III) with a wavelength of 532 nm was used as the pulse laser. It was operated with a pulse width of 5 ns and a repetition rate of 10 Hz. The laser power was adjusted to 50 ± 5 mJ/pulse immediately prior to the final mirror.

2.4. Optical Detection System

Measurements were performed using a bundle optical fiber (quartz core, 27 cores, 200 μm diameter, 2 m length). The fiber end face was imaged at approximately 0.8 times magnification using a lens. The opposite side of the optical fiber was spectrally resolved using an aberration-corrected Czerny-Turner spectrometer (Princeton, IsoPlane 320) and detected using an ICCD (intensified charge-coupled device) camera (Princeton, PI-MAX).

A flat quartz plate was placed at a 45-degree angle along the focal path of the lens. The reflected light was imaged onto a CMOS (complementary metal-oxide semiconductor) camera sensor for synchronized photography with the spectroscopic measurements. This configuration facilitated the determination of the spatial position of the spectroscopic measurements.

The focusing lens, a part of the bundle optical fiber, and CMOS camera were configured to move in unison, enabling the measurement position to be changed by moving the entire assembly. Figure 2 shows a sample of the CMOS camera images.

Fig. 2. Typical photographs of (a)–(c) laser-induced plasma and (d)–(f) superimposed plasma of arc and laser irradiation for different laser shots. (Online version in color.)

The center of the laser was set as the origin of the spatial distribution of the emission spectrum and the vertical direction was evaluated as the Z-axis. The temporal distribution of the emission spectrum was evaluated by varying the delay time τ, with τ = 0 set at the moment of laser emission, by adjusting the ICCD gate delay time.

3. Results and Discussion

3.1. Plasma Appearance and Positional Dependence of Signal Intensities

Figures 2(a)–2(c) and 2(d)–2(f) show the images of the plasma in conventional LIBS, where only a single laser pulse was used, and in AP-LIBS, respectively. Video S1 in the Supplementary Material includes a video comprising 90 photographs.

In conventional LIBS, a cone-shaped plasma, approximately 1 mm width and height, was formed; a slightly brighter region extended by approximately 2 mm in the Z direction. Their sizes were generally stable.

In contrast, in AP-LIBS, in addition to a cone-shaped plasma of similar size, the plasma extended toward the graphite electrode. The brightness of the extended part was not uniform, and a strong emission was observed in the region away from the laser plasma. The size of the cone-shaped plasma (z < 2 mm) was almost identical for both conventional LIBS and AP-LIBS. The primary difference between them was more pronounced in the z > 2 mm region. In the z < 2 mm region, the bright image was dominated by the laser-induced plasma, which emitted an intense light for a short duration. Conversely, the z > 2 mm region was originally dominated by arc plasma, which continuously emitted considerably weaker light.

Figure 3 shows the Z-dependence of the signal intensity per unit time for conventional LIBS (I_L), arc plasma (I_AP), and AP-LIBS (I_AP+L) at the wavelength of the Cr I line (520 nm). I_AP+I_L is also plotted. This emission line belongs to Cr, which is a representative element contained in the sample used in this study. Among the Cr emission lines, it is the most clearly observable due to its favorable relationship between emission intensity and spectral overlap. It should be noted that, as mentioned in Section 3.3, there is no significant difference in the trends described below for other elements. I_AP of the arc plasma does not depend on the time delay τ; hence, the same value is plotted for all τ. Owing to large plasma fluctuations, spectroscopic measurements were performed with 100 accumulations. As τ increased, both I_AP+L and I_L decayed. Although there was no significant difference between I_AP+L (laser and arc superposition) and I_AP+I_L (simple sum of laser and arc) in the region where τ was small, a difference between the two emerged at large Z when τ > 1 μs. Regarding the Z-dependence of each intensity, I_L peaked at Z = 1–2 mm and decreased as Z increased further. Although I_AP remained almost constant regardless of Z, I_AP+L showed an increasing trend as Z increased. That is, while the intensity was maintained or weakened as Z increased when the laser or arc was used alone, the emission intensity of the line spectrum increased with the distance from the laser spot when both were superimposed, which is an interesting result.

Fig. 3. Vertical position dependence of signal intensities for the Cr I (520 nm) line in AP-LIBS (I_AP+L), conventional LIBS (I_L), arc plasma (I_AP), and I_AP+I_L at delay times τ of (a) 100 ns, (b) 1 μs, (c) 5 μs, (d) 10 μs, (e) 50 μs, and (f) 250 μs. (Online version in color.)

An apparent contradiction arises when the spatial distributions shown in Figs. 2 and 3 are compared. As shown in Fig. 2, the plasma became darker with increasing Z, whereas Fig. 3 shows an increase in I_AP+L with increasing Z. This discrepancy can be attributed to the different natures of the measurements. The images in Fig. 2 capture the total light emission, including the intense continuous spectrum (thermal radiation and bremsstrahlung), immediately after laser irradiation, because of the long exposure time used in imaging. In contrast, Fig. 3 shows the intensity of the Cr I spectral line. The continuous spectrum, which dominates the visual appearance shown in Fig. 2, decayed rapidly with time and distance from the laser spot. However, the Cr I line emission benefits from the extended plasma lifetime in AP-LIBS, thereby resulting in the observed intensity increase with Z. This difference highlights the complex spatiotemporal dynamics of AP-LIBS plasma and the importance of spectral and time-resolved measurements for accurate characterization.

3.2. Delay Time Dependence of the Signal Intensities and Enhancement Factor

This section explores the evolution of the signal intensities and enhancement factors in AP-LIBS over time, providing insights into the temporal dynamics of plasma, and compares its performance with that of conventional LIBS.

Figure 4 shows the time delay dependence of the signal intensities of AP-LIBS, conventional LIBS, and arc plasma for the Cr I (520 nm) line. The plots represent the average values of multiple measurement positions at Z = 0.4–0.6 mm. Averaging the results from multiple measurement points helps reduce the impact of fluctuations. This Z region is relatively close to the spatial center of the laser-induced plasma. We note that similar trends are observed in other Z regions and for other elements.

Fig. 4. (a) Delay time dependence of average signal intensities at Z = 0.4–0.6 mm for the Cr I (520 nm) line in AP-LIBS (I_AP+L), conventional LIBS(I_L), arc plasma (I_AP), and the sum of I_AP and I_L. (b) Delay time dependence of various enhancement factors calculated from the data in (a): I_AP+L/I_AP, I_AP+L/I_L, and I_AP+L/(I_AP + I_L). (Online version in color.)

Figure 4(a) shows the time-delay dependences of I_AP+L, I_L, I_AP, and the sum of I_AP and I_L. While I_L monotonically decreased as the time delay increased with almost constant gradient, I_AP remained constant and independent of the time delay. In the small time-delay region (a few microseconds or less), I_AP+L and I_AP+I_L were almost identical. However, as the time delay increased, in the range of several microseconds to several tens of microseconds, the slope of I_AP+L becomes smaller. During this time period, I_AP+L shows higher values than I_AP+I_L. At even longer elapsed times, as the influence of laser irradiation diminishes, both I_AP+I_L and I_AP+L approach values close to I_AP.

Figure 4(b) shows that the time dependence of the I_AP+L/I_AP, I_AP+L/I_L, and I_AP+L/(I_AP+I_L). I_AP+L/I_AP exhibited large values, reaching more than 100 when the time delay was small (< 1 μs). In contrast, I_AP+L/I_L exhibited large values, reaching more than 100 as the time delay increased (> 100 μs). In previous studies on SD-LIBS with a superimposed spark discharge, I_AP+L/I_L was often used as an indicator of the enhancement factor. Using this indicator, it is possible to achieve enhancement factors of hundreds by measuring at large delay times τ. However, comparing the intensities when the laser-induced plasma has completely decayed is not meaningful, and the essential parameter is I_AP+L/(I_AP+I_L) for AP-LIBS. Therefore, in this study, the enhancement for for AP-LIBS, η, is defined as I_AP+L/(I_AP+I_L).

Notably, η peaks from a few μs to a few tens of μs, which corresponds to the time delay when I_AP and I_L become almost equal in intensity. Under the condition of Z = 0.4–0.6 mm, the peak value of η is approximately 10. While the amplification rates differ for other elements, they also become larger at similar timings. The effect of the interaction between the arc plasma and laser-induced plasma in AP-LIBS became more pronounced when the time delay was increased, and the laser-induced plasma decayed to an intensity comparable to that of the background arc plasma. The primary reason for this phenomenon is likely the extended lifetime of active species generated by laser-induced plasma, attributed to energy supply from the arc plasma. After nanosecond pulse laser irradiation, the laser-induced plasma decays, but we think its lifetime is extended due to energy supply from the surrounding arc plasma. In Fig. 3, the difference between I_AP+L and I_AP+I_L increased with larger Z values. The data suggest that the amplification rate increases in the peripheral region of the plasma, which corresponds to the spatial boundary region. This increase, along with the expansion of the plasma region, can be attributed to the energy supply from the surrounding arc. This mechanism is a trend also observed in other LIBS techniques that superimpose different energy sources, such as LS-DP-LIBS,^33,34) and it is thought that a similar effect occurs with arc plasma as well.

That is, clearly, in AP-LIBS, the enhancement factor is maintained at a high level for specific time delays. This demonstrates that in AP-LIBS, the interaction between the arc plasma and laser-induced plasma results in effects that can be up to an order of magnitude larger than their simple sum. In addition, while this study only measured the time delay at sparse intervals for large Z conditions owing to experimental constraints, η has been found to increase up to several tens of times for large Z. Thus, further experimental exploration is needed to determine the maximum amplification rate.

3.3. Enhancement Factor of Various Emission Species

Figure 5 shows the enhancement factor η for various emission species. The vertical position dependence at τ = 10 μs is presented in Figs. 5(a)–5(c), whereas the delay time dependence at Z = 0.80 mm and Z = 3.52 mm is shown in Figs. 5(d)–5(f). The emissions were categorized into three groups: (a, d) atomic emissions from elements with relatively high vapor pressures (>10 Pa at 1773 K), (b, e) atomic emissions from elements with relatively low vapor pressures (<10 Pa at 1773 K), and (c, f) ionic emissions. To provide context for this categorization, we calculated the vapor pressures of Mn, Cu, Cr, Fe, Ni, and Si at 1773 K using values from Ref. 38. The resulting vapor pressures were 2.3 × 10³ Pa, 48 Pa, 4.7 Pa, 2.0 Pa, 0.72 Pa, and 0.15 Pa, respectively.

Fig. 5. Enhancement factor η for various emission species: (a)–(c) Vertical position dependence at τ = 10 μs and (d)–(f) delay time dependence at Z = 0.80 and 3.52 mm. (a, d) Atomic emissions from elements with relatively high vapor pressures, (b, e) atomic emissions with relatively low vapor pressures, and (c, f) ionic emissions. (Online version in color.)

From Figs. 5(a)–5(c), all emission species exhibited a trend of higher η in the peripheral plasma region (Z > ~2 mm), probably because of the expansion of the plasma region and suppression of spatial decay that resulted from the superposition with the arc plasma.

For atomic emissions with relatively high vapor pressures, shown in Fig. 5(d) (Cu I: 515 nm and Mn I: 294 nm), η is relatively small, not exceeding 10 times. This suggests that the further evaporation of high vapor pressure elements has a limited effect. In contrast, atomic emissions with relatively low vapor pressures, shown in Fig. 5(e) (Si I: 865 nm, Cr I: 520 nm, Fe I: 358.6 nm, and Ni I: 338 nm), show larger η, sometimes exceeding 20 times. This indicates that further evaporation significantly contributes to the enhancement factor of these elements.

Ionic emissions, shown in Fig. 5(f) (Fe II: 275 nm, Cr II: 286 nm, and Sn II: 791 nm), demonstrate moderate η (10–20 times). The amplification of high-energy ion emissions suggests energization of the excitation species.

The temporal behaviors of these species also differ. In Figs. 5(d)–5(f) for Z = 0.80 mm, high vapor pressure elements showed limited enhancement, whereas low vapor pressure elements exhibited a significant increase in η after 20 μs when LIBS decayed. Ionic emissions increased in η from a few μs after laser irradiation, earlier than low vapor pressure elements. At Z = 3.52 mm, both low vapor pressure elements and ionic species maintained high η for at least several tens of μs after laser irradiation, whereas high vapor pressure elements exhibited more moderate enhancement.

These observations indicate that η varies significantly depending on the emission species. The extended plasma lifetime and further evaporation contribute to higher η for low vapor pressure elements and ionic species. Although the current data show considerable variability, these findings provide valuable insights for optimizing AP-LIBS conditions in analytical applications.

4. Conclusions

This study investigated the signal enhancement effect and spatiotemporal distribution using arc plasma-assisted laser-induced breakdown spectroscopy (AP-LIBS). The key findings are as follows:

(1) AP-LIBS demonstrated significant signal enhancement, particularly in the spatiotemporal peripheral plasma region (Z > ~2 mm), which is characterized by a larger plasma size and longer lifetime than those of conventional LIBS.

(2) The enhancement factor η, defined as the ratio of AP-LIBS signal intensity to the sum of individual arc and laser-induced plasma intensities, exceeded 10 for most emission species. It is important to note that this definition differs from the signal enhancement ratios reported in other techniques such as SD-LIBS, precluding direct comparisons.

(3) Low-vapor-pressure elements and ionic species showed greater enhancement, which was attributed to the extended plasma lifetime and promoted evaporation.

(4) Temporal analysis revealed that η peaks when the laser-induced plasma decayed to an intensity comparable to the background arc plasma, with the interaction effect up to an order of magnitude larger than a simple sum.

(5) The spatiotemporal behavior of η varied significantly among emission species, thereby providing insights into complex plasma dynamics.

These findings provide valuable insights into the plasma generation and maintenance mechanisms in AP-LIBS, which suggests its potential for improved sensitivity in elemental analysis for electric arc furnace applications. It has become clear that this signal amplification occurs in the spatiotemporal boundary region of the laser-induced plasma. While the appropriate delay time and spatial position are likely to vary significantly depending on the conditions, these results suggest that tuning the temporal and spatial measurement conditions to explore such regions can be effective. Future work should investigate the dependence of signal enhancement on various arc and laser parameters and perform a quantitative analysis of trace elements to further demonstrate the capabilities of AP-LIBS.

Statement for Conflict of Interest

The authors declare no conflict of interest.

Supporting Information

A video file is available as Supporting Information.

Movie S1. Comparative analysis of conventional LIBS and AP-LIBS using a sequence of 90 photographs. This video demonstrates the visual differences between conventional laser-induced breakdown spectroscopy (LIBS) and arc-plasma-assisted LIBS (AP-LIBS) for molten iron.

This material is available on the Journal website at https://doi.org/10.2355/isijinternational.ISIJINT-2024-221.

Acknowledgments

This work was supported in part by the 31st ISIJ Research Promotion Grant. The authors would like to express their gratitude to Mr. Wataru Sato for his assistance in constructing the experimental apparatus.

Nomenclature (Notation Example)

I_AP+L: Signal intensity of AP-LIBS (arb. unit)

I_L: Signal intensity of conventional LIBS (arb. unit)

I_AP: Signal intensity of arc plasma (arb. unit)

Z: Vertical position from laser spot (mm)

τ: Delay time (μs)

η: Enhancement factor (–)

References

• 1)   M.  Aula,  T.  Demus,  T.  Echterhof,  M.  Huttula,  H.  Pfeifer and  T.  Fabritius: ISIJ Int., 57 (2017), 478. https://doi.org/10.2355/isijinternational.ISIJINT-2015-677
• 2)   M.  Aula,  A.  Mäkinen and  T.  Fabritius: Appl. Spectrosc., 68 (2014), 26. https://doi.org/10.1366/13-07079
• 3)   H.  Pauna,  M.  Aula,  J.  Seehausen,  J.-S.  Klung,  M.  Huttula and  T.  Fabritius: Steel Res. Int., 91 (2020), 2000051. https://doi.org/10.1002/srin.202000051
• 4)   H.  Pauna,  A.  Tuomela,  M.  Aula,  P.  Turunen,  V.  Pankratov,  M.  Huttula and  T.  Fabritius: Metall. Mater. Trans. B, 53 (2022), 454. https://doi.org/10.1007/s11663-021-02382-5
• 5)   A.  Rautioaho,  H.  Pauna,  V.-V.  Visuri,  M.  Huttula and  T.  Fabritius: IOP Conf. Ser. Mater. Sci. Eng., 1309 (2024), 012001. https://doi.org/10.1088/1757-899x/1309/1/012001
• 6)   B.  Mitas,  H.  Pauna,  J.  Feldbacher and  J.  Schenk: IOP Conf. Ser. Mater. Sci. Eng., 1309 (2024), 012002. https://doi.org/10.1088/1757-899x/1309/1/012002
• 7)   R. D.  Morales,  A. N.  Conejo and  H. H.  Rodriguez: Metall. Mater. Trans. B, 33 (2002), 187. https://doi.org/10.1007/s11663-002-0004-7
• 8)   T.  Hay,  A.  Reimann and  T.  Echterhof: Metall. Mater. Trans. B, 50 (2019), 2377. https://doi.org/10.1007/s11663-019-01632-x
• 9)   T.  Hay,  V.-V.  Visuri,  M.  Aula and  T.  Echterhof: Steel Res. Int., 92 (2021), 2000395. https://doi.org/10.1002/srin.202000395
• 10)   B.  Kleimt,  W.  Krieger,  D.  Mier Vasallo,  A. A.  Ayarza and  I. U.  Iriondo: Metals (Basel), 13 (2023), 1332. https://doi.org/10.3390/met13081332
• 11)   J.  Gruber,  J.  Heitz,  H.  Strasser,  D.  Bäuerle and  N.  Ramaseder: Spectrochim. Acta Part B At. Spectrosc., 56 (2001), 685. https://doi.org/10.1016/S0584-8547(01)00182-3
• 12)   G.  Hubmer,  R.  Kitzberger and  K.  Mörwald: Anal. Bioanal. Chem., 385 (2006), 219. https://doi.org/10.1007/s00216-006-0321-9
• 13)   L.  Sun,  H.  Yu,  Z.  Cong,  Y.  Xin,  Y.  Li and  L.  Qi: Spectrochim. Acta Part B At. Spectrosc., 112 (2015), 40. https://doi.org/10.1016/j.sab.2015.08.008
• 14)   V.  Sturm,  R.  Fleige,  M.  de Kanter,  R.  Leitner,  K.  Pilz,  D.  Fischer,  G.  Hubmer and  R.  Noll: Anal. Chem., 86 (2014), 9687. https://doi.org/10.1021/ac5022425
• 15)   M. O.  Vieitez,  J.  Hedberg,  O.  Launila and  L.-E.  Berg: Spectrochim. Acta Part B At. Spectrosc., 60 (2005), 920. https://doi.org/10.1016/j.sab.2005.05.024
• 16)   T.  Zhang,  L.  Liang,  K.  Wang,  H.  Tang and  X.  Yang: J. Anal. At. Spectrom., 29 (2014), 2323. https://doi.org/10.1039/C4JA00217B
• 17)   R.  Noll,  H.  Bette,  A.  Brysch,  M.  Kraushaar,  I.  Mönch,  L.  Peter and  V.  Sturm: Spectrochim. Acta Part B At. Spectrosc., 56 (2001), 637. https://doi.org/10.1016/S0584-8547(01)00214-2
• 18)   Z.  Wang,  Y.  Deguchi,  F.  Shiou,  S.  Tanaka,  M.  Cui,  K.  Rong and  J.  Yan: ISIJ Int., 60 (2020), 971. https://doi.org/10.2355/isijinternational.ISIJINT-2019-317
• 19)   S.  Imashuku and  K.  Wagatsuma: ISIJ Int., 62 (2022), 811. https://doi.org/10.2355/isijinternational.isijint-2021-393
• 20)   P.  Chen,  H.  Luo,  M.  Cui,  Z.  Wang,  Y.  Deguchi and  J.  Yan: ISIJ Int., 62 (2022), 875. https://doi.org/10.2355/isijinternational.isijint-2021-417
• 21)   J.  Thomas and  H.  Chandra Joshi: Appl. Spectrosc. Rev., (2023), 1. https://doi.org/10.1080/05704928.2023.2187817
• 22)   J. P.  Singh and  S. N.  Thakur: Laser-Induced Breakdown Spectroscopy, Elsevier, Amsterdam, (2020), 23. ISBN: 9780128188293
• 23)   P. A.  Benedetti,  G.  Cristoforetti,  S.  Legnaioli,  V.  Palleschi,  L.  Pardini,  A.  Salvetti and  E.  Tognoni: Spectrochim. Acta Part B At. Spectrosc., 60 (2005), 1392. https://doi.org/10.1016/j.sab.2005.08.007
• 24)   V. I.  Babushok,  F. C.  DeLucia,  J. L.  Gottfried,  C. A.  Munson and  A. W.  Miziolek: Spectrochim. Acta Part B At. Spectrosc., 61 (2006), 999. https://doi.org/10.1016/j.sab.2006.09.003
• 25)   J.  Scaffidi,  S. M.  Angel and  D. A.  Cremers: Anal. Chem., 78 (2006), 24. https://doi.org/10.1021/ac069342z
• 26)   A.  Robledo-Martinez,  H.  Sobral and  A.  Garcia-Villarreal: Spectrochim. Acta Part B At. Spectrosc., 144 (2018), 7. https://doi.org/10.1016/j.sab.2018.03.002
• 27)   D. N.  Stratis,  K. L.  Eland and  S. M.  Angel: Appl. Spectrosc., 54 (2000), 1270. https://doi.org/10.1366/0003702001951174
• 28)   O. A.  Nassef and  H. E.  Elsayed-Ali: Spectrochim. Acta Part B At. Spectrosc., 60 (2005), 1564. https://doi.org/10.1016/j.sab.2005.10.010
• 29)   K.  Li,  W.  Zhou,  Q.  Shen,  Z.  Ren and  B.  Peng: J. Anal. At. Spectrom., 25 (2010), 1475. https://doi.org/10.1039/b922187e
• 30)   Y.  Liu,  M.  Baudelet and  M.  Richardson: J. Anal. At. Spectrom., 25 (2010), 1316. https://doi.org/10.1039/C003304A
• 31)   Y.  Ikeda,  J. K.  Soriano,  H.  Ohba and  I.  Wakaida: Sci. Rep., 13 (2023), 13901. https://doi.org/10.1038/s41598-023-41208-z
• 32)   J.  Viljanen,  Z.  Sun and  Z. T.  Alwahabi: Spectrochim. Acta Part B At. Spectrosc., 118 (2016), 29. https://doi.org/10.1016/j.sab.2016.02.002
• 33)   J.  Register,  J.  Scaffidi and  S. M.  Angel: Appl. Spectrosc., 66 (2012), 869. https://doi.org/10.1366/11-06531
• 34)   Z.  Wang,  Y.  Deguchi,  R.  Liu,  A.  Ikutomo,  Z.  Zhang,  D.  Chong,  J.  Yan,  J.  Liu and  F.-J.  Shiou: Appl. Spectrosc., 71 (2017), 2187. https://doi.org/10.1177/0003702817693239
• 35)   I. Y.  Elnasharty,  F. R.  Doucet,  J.-F. Y.  Gravel,  P.  Bouchard and  M.  Sabsabi: J. Anal. At. Spectrom., 29 (2014), 1660. https://doi.org/10.1039/C4JA00099D
• 36)   M.  Cui,  H.  Guo,  Y.  Chi,  L.  Tan,  C.  Yao,  D.  Zhang and  Y.  Deguchi: Spectrochim. Acta Part B At. Spectrosc., 191 (2022), 106398. https://doi.org/10.1016/j.sab.2022.106398
• 37)   M.  Cui,  Y.  Deguchi,  Z.  Wang,  Y.  Fujita,  R.  Liu,  F.-J.  Shiou and  S.  Zhao: Spectrochim. Acta Part B At. Spectrosc., 142 (2018), 14. https://doi.org/10.1016/j.sab.2018.02.002
• 38)   B.  Mondal,  T.  Mukherjee,  N. W.  Finch,  A.  Saha,  M. Z.  Gao,  T. A.  Palmer and T. DebRoy: Materials, 16 (2022), 50. https://doi.org/10.3390/ma16010050