# Tandem laser-induced breakdown spectroscopy laser-ablation inductively-coupled plasma mass spectrometry analysis of high-purity alumina powder

• Lee, Yonghoon (Department of Chemistry, Mokpo National University) ;
• Kim, Hyang (Department of Chemistry, Mokpo National University)
• Accepted : 2019.08.01
• Published : 2019.08.25

#### Abstract

Alumina is one of the most important ceramic materials because of its useful physical and chemical properties. Recently, high-purity alumina has been used in various industrial fields. This leads to increasing demand for reliable elemental analysis of impurities in alumina samples. However, the chemical inertness of alumina makes the sample preparation for conventional elemental analysis a tremendously difficult task. Herein, we demonstrated the feasibility of laser ablation for effective sampling of alumina powder. Laser ablation performs sampling rapidly without any chemical reagents and also allows simultaneous optical emission spectroscopy and mass spectrometry analyses. For six alumina samples including certified reference materials and commercial products, laser-induced breakdown spectroscopy (LIBS) and laser-ablation inductively-coupled plasma mass spectrometry (LA-ICP-MS) analyses were performed simultaneously based on a common laser ablation sampling. LIBS was found to be useful to quantify alkali and alkaline earth metals with limits-of-detection (LODs) around 1 ppm. LA-ICP-MS could quantify transition metals such as Ti, Cu, Zn, and Zr with LODs in the range from a few tens to hundreds ppb.

# 1. Introduction

Alumina (Al2O3) is used not only as a raw material for aluminum production but also as an anticorrosive material owing to its properties of wear resistance and minimal chemical reactivity. In addition, because itis also an excellent insulator against heat and electricity, alumina has been used as a source material in various sectors that need materials with such properties.1,2 Recently, there is a growing demand for alumina powder with a high purity of ≥ 99.99 % and micrometer-level particle size in advanced industrial sectors, including display, energy material, automotive, and semiconductor sectors. In particular, sapphire ingots and wafers produced using high-purity aluminaas a raw material are essential materials required for producing light-emitting diodes. Here, impurities in alumina may alter its optical properties. Accordingly, alumina impurity analysis is critical for quality control of the source material.

The most difficult process in impurity analysis of high-purity alumina could be viewed as sample pre-treatment. Alumina is physically very hard and chemically inert and thus makes the dissolution process of the solid-phase matrix complicated and time-consuming (10-20 h). Sample pre-treatment is typically conducted by dissolving alumina powder in an aqueous acid solution under a high-temperature and high-pressure condition.3-6 Moreover, because the likelihood of the sample being exposed to contamination increases with the length and complexity of the sample preparation process, simplifying the sample pre-treatment process is very important foranalyzing ultra-trace quantities of impurities in high-purity materials. Therefore, dry methods are moresuitable for sampling materials, such as alumina, thatare difficult to prepare in the solution phase, and the most commonly used dry sampling method for such purposes is the glow discharge method.7-12 With drysampling using glow discharge, the direct current (DC) discharge mode is typically used.13 Therefore, whenglow discharge is used, conductive samples can act as an electrode to allow effective sampling for mass spectrometry or optical emission spectroscopy; however, when analyzing nonconducting samples, such as alumina, separate electrode composed of expensivehigh-purity tantalum or indium must be used.

For dry sampling of nonconducting samples, laserablation can be an effective alternative to glowdischarge.14 For laser ablation sampling, a femtosecondor nanosecond pulse laser is typically used. Whenthe laser beam with a fluence of 10-100 GW/cm2 is focused on the sample surface, the part of the sample where the laser energy is focused becomes atomized and ionized to produce plasma. This plasma has alifestime of approximately several tens of microseconds and is extinguished as it emits light. By recording the optical emission spectrum, quantitative as well asqualitative analysis for the sample material can be performed. Moreover, the ablated particles can bedirectly sent to the mass spectrometer for elementalanalysis. The elemental analysis method using the optical emission spectrum of laser-generated plasma istermed laser-induced breakdown spectroscopy (LIBS), 15, 16 while the elemental analysis methodusing inductively-coupled plasma mass spectrometry on a solid sample atomized via laser ablation istermed laser-ablation inductively-coupled plasmamass spectrometry (LA-ICP-MS).17 Accordingly, because both LIBS and LA-ICP-MS laser-induced plasma, they have the characteristic of being able to share the sampling process. Moreover, information regarding the elemental composition of the sample obtainable from these two analytical methods ismostly non-duplicative and complimentary. Generally, LIBS can analyze metallic elements with a limit of detection (LoD) of a few to several hundred ppmand nonmetallic elements with an LoD of several hundred ppm to several %.18 The LoD performance of LIBS is even better for alkali and alkaline earthmetals, sometimes reaching several hundred ppb. Forrelatively light elements, it is difficult to obtain reliableresults from mass spectrometry using ICP-MS because of mass interference caused by molecules generated in the mass spectrometer and other isotopes with the same mass as the element being analyzed. However, this method has the advantage of offering excellent LoD performance for heavy elements. The LoD performance of LA-ICP-MS using laser ablation as the sampling method and ICP-MS as the detection method is generally superior to that of LIBS. Foranalysis of impurities contained in high-purity aluminasamples, the LoD performance of LIBS is insufficient for most impurity elements as compared to that of LA-ICP-MS. However, LIBS may be an alternative incases where reliable results for light alkali and alkaline earth metals, including Na, K, Ca, and Mg, cannot be obtained using ICP-MS because of the mass interference. It is widely known that 39K, 40Ca, and 24Mg, the main isotopes of K, Ca, and Mg, respectively, are susceptible to mass interference by 38Ar1H+, 40Ar+, and 12C2+ generated in the ICP-MS spectrometers using Ar plasma.19-20 Moreover,23Na, the only isotope of Na, can be found as abackground signal even when a sample has not beeninjected because of accumulated contamination insidethe mass spectrometer.21

This paper reports that elemental analysis of lightmetallic elements and heavy elements in high-purity alumina can be effectively performed using tandemLIBS/LA-ICP-MS using laser ablation. The feasibility of quantitative analysis was investigated using the intensity of the emission lines of the correspondingelements that appear in the LIBS spectra for light alkali and alkaline earth elements, such as Na, K, Ca, and Mg, and signal intensities of 49Ti, 65Cu, 66Zn, and 90Zr mass channels in ICP-MS for the heavier elements. As alumina samples, four certified reference materials (CRMs) and two commercial source materials were used. LIBS analysis results for Na, K, Ca, and Mg showed LoD values of 0.64, 2.4, 0.63, and 1.5 ppm, respectively. LA-ICP-MS analysis results for Ti, Cu, Zn, and Zr showed LoD values of 55, 58, 550, and 260 ppb, respectively.

# 2. Materials and Methods

## 2.1. Alumina samples

The samples used in the study consisted of three CRMs (R034, R035, and R035) prepared by the Ceramic Society of Japan, one CRM (8007-a) prepared by the National Metrology Institute of Japan, onecommercial alumina product from Sumitomo Chemical (AKP-3000), and one commercial alumina product(HGH) from POS-HiAL. Table 1 shows the Na, K, Mg, Ca, Ti, Cu, Zn, and Zr concentrations chosen as the elements to be analyzed in this study for each sample. Among these, this concentrations of K, Ca, Ti, and Zn in the APK-3000 sample and Na, K, Mg, Ca, Ti, Cu, Zn, and Zr in the HGH sample wereseparately analyzed using an acid digestion method, as described in the following, prior to LIBS/LA-ICP-MS analysis. The sample (1.0 g) was placed in a 25-mL polytetrafluoroethylene (PTFE) container, which was then placed in a high-pressure container (Parr, Part No.4746) for sample pre-treatment. After adding 15 mL of aqueous H2SO4 solution (Aldrich, 99.999 %) prepared to a volume ratio of 25 % to the container, the high-pressure container lid was tightly sealed and the container placed in an oven (at 230 °C) for 16 h. The sample pre-treated was then placed in a 100-mL volumetric flask and diluted until the H2SO4concentration reached a volume ratio of 2 %. Theliquid sample prepared in this manner was analyzed using inductively-coupled plasma atomic emissionspectroscopy (ICP-AES) and ICP-MS. For the ICP-AES and ICP-MS analyses, Vision from Spectro and ELEMENT XR from Thermo Fisher Scientific were used, respectively. Standards from AccuStandard were used to obtain the concentration calibration curves forquantitative analysis.

Table 1. Certified concentrations of the elements contained in the samples. The unit for the tabulated values is ppm in terms of weight ratio

For the LIBS/LA-ICP-MS analysis using laserablation sampling, alumina powder was pressurized to form pellets. Aluminum powder was placed in apulverizer (Pulverisette 23, FRITSCH) together withaluminum oxide balls of a diameter of 10 mm and pulverized for 10 min of a rotational speed of 40 rpm. From the resulting powder, 5 g was taken and placed in a stainless-steel cup, to which a pressure of 1.3 kN/cm2 was applied to form disc-shaped pellets with a diameter of 30 mm.

## 2.2. Experimental methods

A commercial LIBS instrument (J200 LIBSInstrument) from Applied Spectra, Inc. was used forlaser ablation sampling. LIBS analysis could be performed in open air without a separate samplechamber; however, because tandem LIBS/LA-ICP-MS analysis was to be performed in this study, thealumina pellets were placed in a sample chamber, as shown in Fig. 1, and laser beam was focused on the sample surface with a 100 micrometer diameter spot. A flash-lamp-pumped Q-switched Nd:YAG laserwas used as the ablation laser. The wavelength, pulse width, and energy per pulse of this laser were set to 266 nm, 10 ns, and 15 mJ, respectively. The laserrepetition rate was set to 10 Hz, and while the sample was being ablated, the stage holding the sample linearlymoved at a speed of 0.2 mm/s. Accumulation of signal sperformed by linearly scanning a length of 2 mm on the sample surface under the aforementioned conditions led to a single measurement. The time required for the singlemeasurement was 10 s. During this, 100 laser pulses were focused on the 2-mm scan line.

Fig. 1. A schematic diagram of the laser-ablation chamber. The inset is a picture of the alumina sample pellet

The light emitted from the plasma was received through two plano-convex lenses and introduced into a 6-channel spectrometer through an optical fiber. The spectrometer's wavelength coverage was 186-1050 nm, and the wavelength resolution was ~0.1 nm. The photodetector unit of the spectrometer was acharge-coupled device (CCD). From the laser Q-switching to the moment the CCD photodetector unitof the spectrometer began accumulating the lightintensity, a difference of 500 ns in time was provided to avoid strong background continuum emission and broadening of atomic and ionic emission lines because of the high electron density of the initial plasma. The CCD was exposed to the light emitted from the plasma generated by each laser pulse for 1.05 ms. While recording the LIBS spectrum, particles ablated from the sample surface were sent to the mass spectrometer using helium gas to simultaneously record the signal intensity profiles at the mass-to-charge ratios m/z = 48(49Ti), 63(65Cu), 64(66Zn), and 90(90Zr). The flow rate of helium gas transporting the ablated particles from the sample chamber to the mass spectrometer (quadrupole mass spectrometer, Plasma Quant MS Elite, Analytik Jena) was 0.7 L/min.

# 3. Results and Discussion

## 3.1. LIBS

Fig. 2(a) shows the LIBS spectrum of the R035 sample in the wavelength range from 18 to 1050 nm. The strong emission lines observed in this spectrum were mostly because of Al and O. In addition to Al and O emission lines, Na I, K I, Ca II,  and Mg II emission lines were also observed near 589, 766, 393, and 280 nm, respectively. Each of theseemission lines is marked in Fig. 2(b)-2(e), which showthe spectra of the six alumina samples. Moreover, the emission line that showed the strongest intensity foreach element is marked with an “*”, and the intensity value of this emission line was used to obtain the concentration calibration curve. The emission line intensity value was calculated by integrating the areabelow the emission line after removing the baseline.

Fig. 2. (a) LIBS spectrum of the sample R035 in the wavelength region between 186 nm and 1050 nm and the expanded spectra around the strongest (b) Na I, (c) K I, (d) Ca II and (e) Mg II emission lines that are denoted by “*”.

Fig. 3 shows the concentration calibration curves of Na, K, Mg, and Ca obtained via LIBS analysis. For the concentrations of three CRMs prepared by the Ceramic Society of Japan (R034, R035, and R035) and one CRM prepared by the National Metrology Institute of Japan (8007-a), the certified values provided by the institutions were used. Whenever the upperlimit value was provided, it was used to obtain the concentration calibration curve. For AKP-3000, thealumina product from Sumitomo Chemical, the impurity concentration data provided by the manufacturer were used, and for the impurity elements with no provided data, the values were obtained using the aforementioned acid digestion method. For HGH, the commercial alumina product from POS-HiAL, all elemental concentrations were obtained using the acid digestion method and were used to obtain the concentration calibration curves. The concentration calibration curves were derived by linear function fitting of the emission line intensities appearing in the LIBS spectrum for the certified concentration values or the analysis values using theacid digestion method. The R2 values of linear fits for Na, K, Mg, and Ca were 0.9979, 0.8489, 0.9115, and 0.9978, respectively. Na and Ca showed the R2 values close to 1, which indicated that the changes inemission line intensity appearing in the LIBS spectraaccording to the concentrations showed a greaterlinearity than those of K and Mg. The R2 values for K and Mg showed a relatively large deviation from 1, which could be explained by the concentration range of the samples used for the concentration calibration curves being relatively narrower thanthose of Na and Ca and not because of poor accuracy in the elemental analysis. The sample concentration ranges for Na and Ca were 0-1650 and 0.4-173 ppm, respectively, whereas the ranges for K and Mg were relatively narrower at 0-17 and 0.3-7.8 ppm, respectively. The data used to obtain the concentration calibration curve for each element were used tocalculate the root mean square error (RMSE) values for comparison. RMSE was calculated using Eq. (1) as follows:

$\operatorname{RMSE}=\sqrt{\frac{\sum_{i=1}^{n}\left(\hat{y}_{i}-y_{i}\right)^{2}}{n}}$       (1)

In Eq. (1), $\hat{y}_{i}$is the certified concentration value or the concentration value obtained from the wetanalysis; yi is the predicted concentration value based on emission line intensity observed in the LIBS spectra and calibration curves; n is the number of LIBS spectra obtained from all samples; and i is the number assigned to each LIBS spectrum. Using Eq. (1), the RMSE values of Na, K, Mg, and Ca werecalculated to be 44, 2.6, 1.3, and 4.3 ppm, respectively.

Fig. 3. Intensity-to-concentration calibration curves of (a) Na, (b) K, (c) Mg, and (d) Ca constructed on the basis of LIBS spectra.​​​​​​​

The large RMSE found in the Na concentration calibration curve could be explained by considering two factors: self-absorption and measurement precision. For quantitative analysis using LIBS, the sensitivity, in other words the concentration calibration curveslope, generally decreases with increasing concentration. Such a decreasing phenomenon in sensitivity according to concentration is known to occur because of self-absorption.22 Self-absorption occurs because the temperature of the laser-induced plasma is spatiallynonuniform. The outer part of the plasma that contacts the relatively cold air or ambient gas has a temperature that is relatively lower than that at the center. Generally, the light-receiving optical system in the LIBS spectrometer is aligned to focus on the plasma center where the signal is strongly measured. Under such conditions, light emitted from the plasma centerpasses through the outer part of plasma as it is directed toward the light-receiving optical system. There is a relatively large number of particles in the outer part of plasma, which are in the low-energy state because of the lower temperature, and as a result, the light emitted by the particles in the plasma centeris reabsorbed. When self-absorption occurs in laser-induced plasma, its absorbance follows the Beer-Lambert law (Eq. (2)) as follows:

$A=\log \frac{I_{0}}{I}=\varepsilon \cdot c \cdot l$       (2)

In Eq. (2), A, I0, I, e, c, and l represent the absorbance, incident light intensity, transmitted lightintensity, extinction coefficient, concentration, and the distance for interaction between light and object, respectively. Therefore, having a broader concentration range for obtaining the concentration calibration curvemakes it more difficult to obtain a linear calibrationcurve for the entire concentration range. The Beer-Lambert law indicates that the self-absorption intensity may vary according to not only the concentrations of the elements being analyzed but also the spectroscopic characteristics of the emission lines. In the liquid phase, e can represent the molar coefficient of extinction and c can represent the molar concentration. In plasma, these parameters can represent the Einsteincoefficient for the light absorption process (Blower-upper) and particle number density. In this case, self-absorptionincreases when Blower-upper is larger, where Blower-upper is proportional to the Einstein coefficient for thespontaneous emission process (Aupper-lower), as shown in Eq. (3). Therefore, even between emission lines from the same element, the emission line with a strong intensity shows more severe self-absorption phenomenon.

$B_{\text {lower-upper }}=\frac{g_{\text {upper }}}{g_{\text {lower }}} \cdot \frac{c^{3}}{8 \pi h \nu^{3}} \cdot A_{\text {upper-lower }}$       (3)

In Eq. (3), gupper and glower are the statistical weights (degrees of degeneracy) of the higher and lowerenergy states involving a certain transition, respectively, while c, h, and ν are the speed of light, Planck ' sconstant, and the frequency of absorbed or emitted light, respectively. Lastly, the characteristics of the lower energy state involved during the light emission process also affect the self-absorption intensity. The lower energy state of the emission line may be in the ground state or still in the excited state of whichenergy is lower than that of the initial excited state. Considering that the two energy states involved inself-absorption must be the same as the energy statesinvolved in light emission, except that only the energy state during the initial and final stages changes, if the lower energy state of the emission line is still in anexcited state and not the ground state, then the likelihood of self-absorption is significantly reduced because of the subsequent relaxation process. Therefore, self-absorption more severely appears when the lowerenergy state involved in the emission line is in the ground state. In the case of Na, the results can beexplained by the fact that the calibration standardscontain Na in the broader concentration range incomparison to K, Mg, and Ca; the lower energy stateinvolved in the emission line was in the ground state; and changes in sensitivity due to the self-absorptioneffect readily appear when Aupper-lower is large. Table 2 shows the spectroscopic constants of the emissionlines used to obtain the concentration calibrationcurves of Na, K, Mg, and Ca. These constants were extracted from the National Institute of Standards and Technology Atomic Spectra Database.23 The NaI emission line at 588.995 nm was the strongestamong all Na I emission lines observed in the LIBS spectrum, which had the ground state as the lowerenergy state involved in the emission line at 588.995nm. Therefore, because the Na concentration is 1603 ppm, which is much higher than those of otherelements, there is a significant change in sensitivity between the low- and high-concentration regions. Accordingly, if a single linear function fitting isperformed based on an assumption that there is nochange in sensitivity over the entire concentration range, then the data obtained can significantly deviate from the fitting function, resulting in a large RMSEvalue. Fig. 4 shows the results of linear fitting with the concentration range divided into a low-concentration region (0-13 ppm) and a high-concentration region (13-1603 ppm). Compared to that of the low-concentration region, the slope of the calibration curve for the high concentration region is significantly smaller. Such a finding is highly consistent with the influence expected because of the self-absorption effect. The RMSEvalue calculated using the linear function obtained by fitting the low Na concentration region (0-13ppm) was 1.0 ppm, which was similar to or actually lower than those of K, Mg, and Ca.

Table 2. Spectroscopic parameters of Na I, K I, Mg II, and Ca II lines used for the calibration curves

Fig. 4. (a) Intensity-to-concentration calibration curve of Na in the concentration range between 0 and 13 and (b) that between 13 and 1603 ppm.​​​​​​​

In the case of Na, K, Mg, and Ca, the relativestandard deviations (RSDs), i.e., the ratio of the standard deviation of measured signal intensities to their average were 8, 10, 6, and 4 %, respectively. The RSD can be used as an indicator of analytical precision. Therefore, it can be viewed that the level of precision for the analysis conditions used on the alumina powder pellet samples in the present study was 4-10 % RSD. This precision performanceshowed an RSD of 2 % in the R035 sample withan Na concentration of 1603 ppm and an RSD of 10 % in the AKP-3000 sample with an Na concentration of 0.4 ppm. Accordingly, it was determined that the RSD is independent of the concentration at the given concentration range and the standard deviation (SD) increased at a rate similar to the increase insignal intensity or concentration. Under such conditions where RSD is independent of concentration, using samples with relatively high concentrations as the standards for obtaining the concentration calibrationcurves results in an increased RMSE, which is anindicator of accuracy. Therefore, the RMSE value that was especially large in Na could be explained by the influence of the RSD independent of concentration or signal intensity (the SD increased at a similar rate as that of the concentration orsignal intensity) and self-absorption when the Naconcentration range was much broader than that of the other elements in the samples used to obtain the concentration calibration curves.

The LoDs of the LIBS analysis for Na, K, Mg, and Ca were predicted under the given conditions. The LoDs were calculated using Eq. (4) as follows:

$\mathrm{LOD}=\frac{3 \sigma}{s}$       (4)

In Eq. (4), s is the SD of the signal measured in the sample with low concentrations of elements being analyzed and s is the concentration calibration curveslope. The LoDs of Na, K, Mg, and Ca predicted using this method were 0.64, 2.4, 0.63, and 1.5 ppm, respectively.

## 3.2. LA-ICP-MS

Fig. 5 shows the mass signal intensity profiles of the R034, R035, R036, 8007-a, AKP-3000, and HGHalumina samples recorded at the m/z = 49 channel. The results show that the background signal intensity was approximately 103. These mass signal intensityprofiles were simultaneously obtained with the LIBS spectrum. For simultaneous measurement using LIBS and LA-ICP-MS (single round), a linear scan was performed over a length of 2 mm above the sample surface while the sample stage rotated at aspeed of 0.2 mm/s. Accordingly, the time required torecord the mass signal was 10 s. Among the masssignal intensity profiles of the samples shown in Fig. 5, the sample with the highest Ti concentration was R036 with 19 ppm, the second highest was R035 with 17 ppm, and the third highest was HGH with 1.7 ppm. The other samples showed a Ti concentration less than 1 ppm. The measured mass signal intensity clearly showed correlations with the sample concentrations.

Fig. 5. Mass signal intensity profiles recorded at m/z = 49 ( 49Ti) for the six alumina samples.

The mass signal intensity values used for the concentration calibration curves were obtained by removing the background signal from the mass signalintensity profile and integrating the area below the profile. Fig. 6 shows the concentration calibration curves obtained using the mass signal intensity values of 49Ti, 65Cu, 66Zn, and 90Zr and the concentrations of Ti, Cu, Zn, and Zr listed in Table 1. The concentration calibration curves are linear fits of the mass signalintensity values. After fitting, R2 values were from 0.95 to 0.98. Similar to the LIBS analysis, Eq. (4) was used to predict the LoDs of Ti, Cu, Zn, and Zr, which are 55,58, 550, and 260 ppb, respectively.

Fig. 6. Intensity-to-concentration calibration curves of (a) Ti, (b) Cu, (c) Zn, and (d) Zr constructed on the basis of LA-ICPMS analysis.

# 4. Conclusions

The present study investigated the feasibility ofusing a laser ablation sampling method for effective analysis of alumina powder impurities. Laser ablation is not only effective for sampling materials that arephysically hard and chemically inert but it also has the advantage of LIBS analysis being simultaneouslyperformed with mass spectrometry by sending the particles generated from the laser-ablated samples. This study performed tandem LIBS/LA-ICP-MS analysis on six samples including alumina CRMsand commercial products. LIBS analysis results showed that, along with the Al and O emission lines, which are the main elements comprising the matrix, those of Na, K, Mg, and Ca were observed. The intensities of these emission lines were then used to obtain the concentration calibration curves. In the case of Na, the RMSE, which could be viewed as anindicator of analytical accuracy, was larger than that of all the other elements, and accordingly, the reasons for such a difference were discussed. The LIBS analysis results for Na, K, Ca, and Mg showed a LoD of 0.64, 2.4, 0.63, and 1.5 ppm, respectively. Along with the LIBS spectrum, mass signal intensityprofiles for the m/z = 49, 65, 66, and 90 channels weresimultaneously obtained via LA-ICP-MS analysis. The results were then used to obtain the concentration calibration curves of Ti, Cu, Zn, and Zr. Under the given LA-ICP-MS conditions, the LoDs of Ti, Cu, Zn, and Zr were 55, 58, 550, and 260 ppb, respectively. Based on the findings in the present study, it was determined that by using laser ablation sampling, a LIBS analysis with a LoD performance of approximately 1 ppm is possible for alkali and alkalineearth metals while simultaneous analysis at a level of several tens to several hundred ppb for transitionelements is possible. The LIBS spectrum presented in this study was recorded using a linear CCD photodetector. Therefore, it is believed that by using an intensified CCD, a photodetector with superior LoD performance, and increasing the number of laser pulsesused for laser ablation to perform a single round of measurements to a number higher than 100, the performance of the tandem LIBS/LA-ICP-MS analysis could be even further enhanced.

# Acknowledgements

This research was supported by a research grant from Mokpo National University in 2017.

#### Acknowledgement

Supported by : Mokpo National University

#### References

1. P. Atkins and L. Jones, 'Chemical Principles The Quest for Insight', W. H. Freeman and Company, New York, 2008.
2. Y. L. Wu, J. Hong, D. Peterson, J. Zhou, T. S. Cho, and D. N. Ruzic, Surf. Coat. Technol., 237, 369-378 (2013). https://doi.org/10.1016/j.surfcoat.2013.06.043
3. G. Molnar, J. Borossay, Z. B. Varga, M. Ballok, and A. Bartha, Mikrochim. Acta, 134, 193-197 (2000). https://doi.org/10.1007/s006040070037
4. H. Matusiewicz, Mikrochim. Acta, 111, 71-82 (1993). https://doi.org/10.1007/BF01240169
5. M. T. Larrea, I. Gomez-pinilla, and A. C. Farinas, J. Anal. At. Spectrom., 12, 1323-1332 (1997). https://doi.org/10.1039/A702875J
6. H.A. Foner, Anal. Chem., 56, 856-859 (1984). https://doi.org/10.1021/ac00268a072
7. S. Jung, S. Kim, and J. Hinrichs, Spectrochim. Acta B, 122, 45-51 (2016).
8. N. Jakubowski, T. Prohaska, L. Rottmann, and F. Vanhaecke, J. Anal. At. Spectrom., 26, 693-726 (2011). https://doi.org/10.1039/c0ja00161a
9. N. Jakubowski, T. Prohaska, F. Vanhaecke, P. H. Roos, and T. Lindemann, J. Anal. At. Spectrom., 26, 727-757 (2011). https://doi.org/10.1039/c0ja00007h
10. W. Schelles, S. D. Gendt, V. Muller, and R. V. Grieken, Appl. Spectrosc., 49, 939-944 (1995). https://doi.org/10.1366/0003702953964741
11. V. Hoffmann, M. Kasik, P. K. Robinson, and C. Venzago, Anal. Bioanal. Chem., 381, 173-188 (2005). https://doi.org/10.1007/s00216-004-2933-2
12. M. Kasik, C. Venzago, and R. Dorka, J. Anal. At. Spectrom., 18, 603-611 (2003). https://doi.org/10.1039/b300025g
13. J. C. Woo, N. Jakubowski, and D. Stuewer, J. Anal. At. Spectrom., 8, 881-889 (1993). https://doi.org/10.1039/ja9930800881
14. R. E. Russo, X. Mao, J. J. Gonzalez, V. Zorba, and J. Yoo, Anal. Chem., 85, 6162-6177 (2013). https://doi.org/10.1021/ac4005327
15. F. J. Fortes, J. Moros, P. Lucena, L. M. Cabalin, and J. J. Laserna, Anal. Chem., 85, 640-669 (2013). https://doi.org/10.1021/ac303220r
16. D. W. Hahn and N. Omenetto, Appl. Spectrosc., 66, 347-419 (2012). https://doi.org/10.1366/11-06574
17. N. L. LaHaye, M. C. Phillips, A. M. Duffin, G. C. Eiden, and S. S. Harilal, J. Anal. At. Spectrom., 31, 515-522 (2016). https://doi.org/10.1039/C5JA00317B
18. A. W. Miziolek, V. Palleschi, and I. Schechter, 'Laser-Induced Breakdown Spectroscopy (LIBS) Fundamentals and Applications', Cambridge University Press, Cambridge, 2006.
19. S. H. Tan, and G. Horlick, Appl. Spectrosc., 40, 445-460 (1986). https://doi.org/10.1366/0003702864508944
20. E. Albalat, P. Telouk, V. Balter, T. Fujii, V. P. Bondanese, M.-L. Plissonnier, V. Vlaeminck-Guillem, J. Baccheta, N. Thiam, P. Miossec, F. Zoulim, A. Puisieux, and F. Albarede, J. Anal. At. Spectrom., 31, 1002-1001 (2016). https://doi.org/10.1039/C5JA00489F
21. S.-H. Nam, H. Chung, J.-J. Kim, and Y.-I. Lee, Bull. Korean Chem. Soc., 29, 2237-2240 (2008). https://doi.org/10.5012/bkcs.2008.29.11.2237
22. J. P. Singh and S. N. Thakur, 'Laser-Induced Breakdown Spectroscopy', Elsevier Science, Amsterdam, 2007.
23. NIST Atomic Spectra Database, https://www.nist.gov/pml/atomic-spectra-database.