Emission Spectrometric Analysis Using an Okamoto-cavity Microwave-induced Plasma with Nitrogen-Oxygen Mixed Gas

Abstract

Microwave-induced plasma optical emission spectrometry (MIP-OES) using Okamoto-cavity has a unique feature being suitable for in-situ analysis in steel-making industry. The Okamoto-cavity MIP can be directly loaded with organic solvents including sample solution as well as fine particles of a solid sample in the nitrogen-oxygen mixed gas plasma; furthermore, their emission intensities are much elevated by adding oxygen of up to 10% to the nitrogen matrix gas, thus contributing to better sensitivity in the MIP-OES. The reason for the intensity enhancement was investigated with a spectroscopic method that the spatial distribution of the emission intensity from the plasma was estimated with a two-dimensionally imaging spectrograph. Emission lines of chromium, which was the most important alloying element in steel materials, as well as band heads of nitrogen molecule were observed, indicating that the emission intensity of atomic chromium lines was drastically elevated whereas the intensities of ionic chromium lines and the nitrogen band heads were commonly reduced when oxygen gas was added to the nitrogen plasma. This result implies that the ionization of chromium, which dominantly occurs through collisions with nitrogen excited species, can be suppressed because the nitrogen excited species would be consumed through collisions with oxygen molecules to cause their dissociation. Optimization of the measuring parameters in the Okamoto-cavity MIP-OES was conducted to determine chromium contents with good precision, and finally the analytical performance in the MIP was compared with that in a conventional ICP-OES.

1. Introduction

A microwave-induced plasma (MIP) has been employed as an excitation source in optical emission spectrometry (OES); however, the analytical application appears only in a particular field such as a detector for gas chromatography.¹⁾ The major reason for this is that a conventional MIP excited with a Beenaker cavity²⁾ or a Surfatron³⁾ has a limited tolerance to aqueous aerosol solution because they should be maintained at relatively low powers and cannot produce an annular-shaped plasma like an inductively-coupled radio-frequency-induced plasma (ICP) having the central channel where sample aerosols can be easily introduced with the carrier gas. This drawback has hindered extensive applications of MIP-OES in spite of the low running cost.

A surface-wave-excited non-resonant cavity, developed by Okamoto (Okamoto-cavity),⁴⁾ has a unique feature beyond the conventional MIPs. It produces an annular flame-like plasma where the electric field of microwave is concentrated at the surrounding portion of the plasma, having a doughnut-like structure so that sample aerosols can be easily introduced through a central tube of the plasma torch,^4,5) as similar to ICP. The Okamoto-cavity can work at high powers up to 1.5 kW, to make this plasma source highly tolerant to loading of wet aerosols not only in an excitation source in OES⁵⁾ but also in an ionization source for mass spectrometry.⁶⁾ Furthermore, the Okamoto-cavity MIP has several features beyond a conventional ICP as follows: (1) it has lower background emission intensity compared to the Ar-ICP being suitable for the excitation source in OES, (2) it can be sustained with various gases, such as nitrogen, nitrogen-oxygen, and helium,^7,8) (3) it can be directly loaded with organic solvents when using nitrogen-oxygen mixed gas.^8,9,10,11) Concerning the feature (2), the Okamoto-cavity MIP can be operated by using an air compressor¹²⁾ and thus works under the conditions without any cylinder gas. The Okamoto-cavity using pure nitrogen gas is also suitable for an ionization source in mass spectrometry due to its low mass interferences.⁶⁾ Concerning the feature (3), the analytical procedure in a solvent extraction method can be simplified by direct loading of the organic solvent, such as 4-methyl-2-pentanone (MIBK).^9,13)

Several researchers have reported on the excitation characteristics and the analytical performance of the Okamoto-cavity MIP.^7,8,14,15) This paper focuses on large variations in the emission intensities from the Okamoto-cavity MIP when oxygen gas is added to the nitrogen plasma, and discusses the excitation process to cause the intensity variation. Because the MIP yields the inhomogeneous distribution of the emission signal, spatial images of the emission intensity are estimated with a two-dimensionally imaging spectrograph to clarify the excitation mechanism as well as to obtain an optimum measuring condition for the analytical application. The analytical performance in the Okamoto-cavity MIP is compared with that in a conventional argon ICP.

2. Experimental

The apparatus and the principle of operation have been described in detail by Okamoto.^4,16) This paper introduces the explanation briefly. Figure 1 is a block diagram of the measuring system employed.⁹⁾ The MIP is generated with a high-voltage power supply (KN-153-3T-LR-PS, Nippon Koushuha Ltd., Japan), a 2.45-GHz microwave generator (MKN-153-LA-OSC, Nippon Koushuha Ltd., Japan), a wave guide equipped with a three-stub tuner, and a mode transformer (cavity). The mode transformer, called Okamoto-cavity,⁴⁾ consists of an inner conductor and an outer cylindrical conductor terminated by a front plate. By adjusting the geometry of the wave guide and the mode transformer appropriately, a progressive surface wave of the microwave can be excited at the gap between the inner and the outer conductors, where the electric field becomes the maximum and can be coupled with the plasma gas most efficiently, resulting in an annular-shaped plasma like ICP. The plasma torch (300-8352, Hitachi Corp., Japan) comprises a duplex quartz tube having two individual gas flows: one is a tangentially-introduced gas flow as the plasma gas and the other is introduced through the inner tube to carry the sample aerosol to the plasma. In this study, nitrogen-oxygen mixed gas was employed as the plasma gas under the conditions that the flow rate of nitrogen was fixed at 14.0 dm³/min and that of the oxygen was varied in the range of 0–1.5 dm³/min, and pure argon of 0.5 dm³/min was introduced as the carrier gas. The forward power of the microwave was adjusted in the range of 0.8–1.0 kW.

Fig. 1.

 A block diagram of the measuring system.

The emission signal was focused with a biconvex lens on the entrance slit of a scanning spectrometer (P-5200, Hitachi Corp., Japan), comprising a modified Czerny-Turner mounting monochromator and a photomultiplier tube (R955, Hamamatsu Photonics Corp., Japan), and then dispersed to estimate the averaged intensity of a particular emission line over duplicate measurements. The focal length is 0.75 m and the grating has 3600 grooves/mm at a blaze wavelength of 200 nm. In addition, the emitted radiation was taken from the axial direction of the plasma, and the radial distribution of the emission intensity was observed by using a two-dimensionally (2D) imaging spectrograph. The imaging spectrograph system comprised a collimator optics, an image spectrograph and a charge-couple device (CCD) detector.¹⁷⁾ The emission signal from the excitation source was introduced through the collimator onto the entrance slit of the spectrograph (Model 12580, BunkoKeiki Corp., Japan), dispersed at a certain wavelength, and then detected on the CCD detector (SensiCam QE Model, PCO Imaging Corp., Germany), where the 2D image of a particular emission line could be observed in the radial direction of the plasma. The optical alignment between the excitation source and the spectrograph was adjusted by using zero-order diffraction light, so that an image of the source could be observed most clearly. It was determined by measuring an image of a scale that a 2D image having 10 × 10 pixels approximately corresponded to an actual sample area of 0.20 × 0.20 mm². The data were accumulated and averaged on a personal computer to reduce the intensity fluctuation, and the plasma images were finally recorded.

A stock solution of chromium (10 g/dm³) was prepared by dissolving a high-purity chromium metal (99.9%) with 6-M/dm³ hydrochloric acid. Test solutions for estimating the emission characteristics, having the chromium concentration of 200 mg/dm³ or 1000 mg/dm³, and sample solutions for obtaining the calibration curves containing 0–20 mg/dm³ were each prepared by diluting the stock solution with deionized water.

For comparison, the emission signal from an argon ICP (P-5200, Hitachi Corp., Japan) was also measured by using the same detection system and the same sample solutions as in the MIP. The ICP was driven by a 27.12-MHz radio-frequency generator at a forward power of 1.0 kW. The other plasma and measuring conditions were optimized for each analytical line as described later.

3. Results and Discussion

3.1. Analytical Emission Lines

Three neutral atomic lines and two singly-ionized lines of chromium were selected as the analytical lines. The atomic lines have relatively low excitation energies of 2.89–3.46 eV and the ionic lines have excitation energies of 5.86–6.03 eV. The details are summarized in Table 1, which also includes their relative emission intensities. Their assignments are based on an energy level table complied by Suger and Corliss.¹⁸⁾ Because the first ionization potential of chromium is 6.72 eV, the ionization/excitation of these ionic lines needs an energy of more than 12.6 eV. In addition to the chromium emission lines, two band heads of nitrogen molecule and nitrogen molecule ion were observed to investigate the excitation of plasma gas itself, also as indicated in Table 1. Their excitation energies are cited from a reference.¹⁹⁾

Table 1. Measured emission lines of chromium and band heads of nitrogen molecule.

Wavelength (nm)	Excitation energy (eV)	Relative intensity (arb. unit)
Cr I 428.972	2.89	12000
Cr I 425.435	2.91	20000
Cr I 360.533	3.44	14000
Cr I 357.869	3.46	19000
Cr II 284.983	5.86	1100
Cr II 284.324	5.88	1700
Cr II 276.653	6.03	980
N2 337.1	11.05
N2+ 391.4	17.66

3.2. Effect of Oxygen Addition to Nitrogen Plasma

Figure 2 shows variations in the emission intensity for the atomic chromium lines (Cr I) when oxygen gas is added to a nitrogen MIP. In this measurement, the flow rate of nitrogen gas and the forward power of microwave were fixed at 14.0 dm³/min and 800 W, respectively. The emission signal was observed at a plasma position of 10 mm above the front plate of the cavity. The emission intensities were dominantly elevated with an increase in the oxygen flow rate added to the nitrogen plasma, and increased by a factor of 4 when the flow rate of oxygen gas became 1.5 dm³/min (ca. 10 vol.%). Figure 3 shows variations in the emission intensity for the ionic chromium lines (Cr II) as a function of the oxygen flow rate. The plasma parameters and the measuring conditions were the same as those in Fig. 2. Differing from the result of the Cr I lines, their emission intensities were largely reduced by adding oxygen gas to the nitrogen plasma. It can be considered from Figs. 2 and 3 that the ionization of chromium is hindered by coexistence of oxygen species to increase the number density of neutral chromium atoms in the plasma.

Fig. 2.

Variations in the emission intensity of chromium atomic lines: Cr I 425.435 nm (circle), Cr I 428.972 nm (inverted triangle), Cr I 357.869 nm (triangle), and Cr I 360.533 nm (square), as a function of oxygen added to a nitrogen MIP. Sample: Cr 200 mg/dm³; microwave power: 800 W; plasma gas: N₂ 14.0 dm³/min + O₂; carrier gas: Ar 0.5 dm³/min.

Fig. 3.

Variations in the emission intensity of chromium ionic lines: Cr II 284.324 nm (circle), Cr II 284.983 nm (square), and Cr II 276.653 nm (triangle), as a function of oxygen added to a nitrogen MIP. The measuring conditions are the same as in Fig. 2.

3.3. Emission Images of Chromium Line and Nitrogen Molecule Bands

Figure 4 shows 2D emission images for the Cr I 425.435-nm line at several oxygen flow rates added to the nitrogen plasma, whose intensities are expressed by mapping with several colors, such that the intensity becomes smaller from red to blue. In this measurement, the flow rate of nitrogen gas and the forward power of microwave were fixed at 14.0 dm³/min and 800 W, respectively. The oxygen gas was added to be 0.0 (pure nitrogen), 0.8, and 1.2 dm³/min. By comparing the result of the pure nitrogen plasma, these two images in the nitrogen-oxygen plasma clearly indicate that the emission intensity of Cr I 425.435 nm is much enhanced by introducing oxygen gas as the whole, and that the increase drastically appears at the central portion of the plasma and extends towards the axial direction. The reason for this is probably that oxygen species hinder the ionization of chromium and thus enhance the number density of the excited state of neutral chromium atom especially at the central channel of the plasma. The most intense zone is observed at a height of ca. 10 mm above the front plate of the cavity, which is an optimum observation height for the analytical application.

Fig. 4.

Two-dimensional images of the intensity of the Cr I 425.435-nm line in nitrogen-oxygen mixed gas MIP when the content of oxygen is varied. Sample: Cr 1000 mg/dm³; microwave power: 800 W; plasma gas: N₂ 14.0 dm³/min + O₂ 0.0 (a), 0.8 (b), and 1.2 dm³/min (c); carrier gas: Ar 0.5 dm³/min.

Figure 5 shows 2D emission images for a band head of nitrogen molecule at 337.1 nm at several oxygen flow rates added to the nitrogen plasma. In this case, pure water was sprayed into the plasma. The other measuring conditions and the plasma parameters were the same as those in Fig. 4. In contrast to the result of the chromium atomic line (Fig. 4), the emission intensity of the band head is drastically reduced by adding oxygen gas to the nitrogen plasma, as the plasma looks to shrink as a whole. Similarly, 2D emission images for a band head of nitrogen molecule ion at 391.4 nm are shown in Fig. 6, indicating that the intensity decrease by the oxygen addition occurs more drastically than that of the neutral species (see Fig. 5). It occurs more obviously at the central portion of the plasma. These effects are derived from a large decrease in the number density of the corresponding excited species of nitrogen molecule when oxygen gas is added to the nitrogen plasma. It should be noted that the variation in the emission image of the chromium line would conversely correspond to that of the nitrogen band heads, implying that these species would be closely related to each other with respect to their excitation processes.

Fig. 5.

Two-dimensional images of the intensity of a band head of nitrogen molecule ion at 391.4 nm. Pure water is introduced into the plasma and the other measuring conditions are the same as in Fig. 4.

Fig. 6.

Two-dimensional images of the intensity of a band head of nitrogen molecule at 337.1 nm. Pure water is introduced into the plasma and the measuring conditions are the same as in Fig. 4.

3.4. Excitation Mechanism

It is considered that chromium atom is ionized through a collision with fast electron, as depicted in Eq. (1) , or through collisions with excited species of nitrogen molecules, as depicted in Eqs. (2) and (2)’.   

$$\mathrm{C}\mathrm{r}+{\mathrm{e}}^{–}\left(\mathrm{f}\mathrm{a}\mathrm{s}\mathrm{t}\right)\to {\left(\mathrm{C}{\mathrm{r}}^{+}\right)}^{\text{*}}+{\mathrm{e}}^{–}\left(\mathrm{s}\mathrm{l}\mathrm{o}\mathrm{w}\right)+{\mathrm{e}}^{–}\text{\hspace{0.17em},}$$(1)

  

$$\mathrm{C}\mathrm{r}+{\mathrm{N}}_2^{\text{*}}\to {\left(\mathrm{C}{\mathrm{r}}^{+}\right)}^{\text{*}}+{\mathrm{N}}_2+{\mathrm{e}}^{–}\text{\hspace{0.17em},}$$(2)

  

$$\mathrm{C}\mathrm{r}+{\mathrm{N}}_2^{+}\to {\left(\mathrm{C}{\mathrm{r}}^{+}\right)}^{\text{*}}+{\mathrm{N}}_2\text{\hspace{0.17em},}$$(2)’

where the asterisk denotes an excited state of a chromium ion or nitrogen molecule. The excited chromium ion falls down to the ground state with emitting the characteristic radiation. The former reaction, called a collision of the first kind,²⁰⁾ is caused by transfer of the kinetic energy of the electron, and the latter reaction, called a collision of the second kind,²⁰⁾ by transfer of the internal energy of the excited species of nitrogen molecule. The A ³Σ_u and the B ³Π_g excited states of nitrogen molecule have internal energies of ca. 6.2 eV and ca. 7.4 eV in each lowest vibrational level,¹⁹⁾ and the ²Σ_g ground state of nitrogen molecule ion has an internal energy of ca. 14.5 eV in the ground vibrational level,¹⁹⁾ which is enough energy for chromium atom to be ionized. These collisions lead to an increase in the number density of singly-ionized chromium, and eventually an increase in the emission intensity of Cr II lines.

A possible reason for the intensity decrease of the Cr II lines, as shown in Fig. 3, is that the kinetic energy of electrons is reduced through collisions with oxygen molecule in the mixed gas plasma, which results in a shift towards the left-hand side in Eq. (1) decreasing the number density of the ionic species. Moreover, it can be considered from the emission images that the major reason is the drastic decreases in the number density of the nitrogen excited species, which results in a shift towards the left-hand side in Eq. (2). In the nitrogen-oxygen plasma, the internal energy of the nitrogen excited species would be given to dissociate oxygen radicals from oxygen molecule through a collision between the gas species, as denoted in Eq. (3).   

$${\mathrm{N}}_2^{\mathrm{\hspace{0.17em}}\text{*}}+{\mathrm{O}}_2\to {\mathrm{N}}_2+\mathrm{O}+\mathrm{O}$$(3)

The dissociation energy of oxygen molecule was reported to be ca. 5.1 eV when the atomic oxygen has the ³P electron configuration;¹⁹⁾ therefore, the excited states of nitrogen molecule have enough energies to promote this reaction. In the case of a collision with nitrogen molecule ion, a possible reaction producing oxygen radicals is denoted in Eq. (4).   

$${\left({\mathrm{N}}_2{\mathrm{\hspace{0.17em}}}^{+}\right)}^{\text{*}}+{\mathrm{O}}_2\to {\mathrm{N}}_2+{\mathrm{O}}^{+}+\mathrm{O}$$(4)

This reaction may contribute to a decrease in the number density of the nitrogen molecule ion. Another collision should be taken into account: a stepwise ionization from the excited state of nitrogen molecule, as depicted in Eq. (5).   

$${\mathrm{N}}_2^{\mathrm{\hspace{0.17em}}\text{*}}+{\mathrm{e}}^{–}\left(\mathrm{f}\mathrm{a}\mathrm{s}\mathrm{t}\right)\to {\mathrm{N}}_2^{+}\mathrm{\hspace{0.17em}}+{\mathrm{e}}^{–}\left(\mathrm{s}\mathrm{l}\mathrm{o}\mathrm{w}\right)+{\mathrm{e}}^{–}$$(5)

This collision would occur less effectively when the excited nitrogen molecules are reduced through collisions with -oxygen molecules, also leading to a decrease in the number density of the nitrogen molecule ion.

Because of several types of collisions with oxygen molecule, resultant radicals of oxygen atom make the plasma under very strong oxidizing atmosphere;^10,11) however, the excited species of nitrogen molecule have smaller number density in the nitrogen-oxygen plasma than in the pure nitrogen plasma. This effect could cause a shift in the ionization of chromium from the ionic to atomic species, which could generally enhance the emission intensity of atomic lines of chromium, when oxygen gas is added to the nitrogen plasma.

3.5. Analytical Performance

Figure 7 shows calibration curves of the Cr I 357.869-nm line, using four standard solutions containing chromium of 0.0, 5.0, 10.0, and 20.0 mg/dm³, at oxygen flow rates of 0.0, 0.8 and 1.5 dm³/min in a nitrogen-oxygen MIP. In this measurement, the flow rate of nitrogen gas and the forward power of microwave were fixed at 14.0 dm³/min and 1020 W, respectively. As shown in Fig. 4, an intense emission zone of chromium lines appeared at heights of 0–15 mm; therefore, the observation height for the calibration curves was selected to be 10 mm, where the background continuum gradually decreased. It can be seen from Fig. 7 that larger values of the slope are obtained at lager flow rates of oxygen, which contributes to improving the detection limit of chromium in the nitrogen-oxygen mixed gas plasma. Standard deviations of the blank signal at 357.88 nm were measured when pure water was sprayed into the plasma, in order to estimate the detection limit. The 3σ detection limits were obtained to be 0.21 mg/dm³ in the pure nitrogen plasma and 0.09 mg/dm³ in the mixed gas plasma containing oxygen of 1.5 dm³/min. The detection limit became improved by a factor of 2.3.

Fig. 7.

Calibration curves for Cr I 357.869 nm in a nitrogen-oxygen mixed gas MIP at oxygen flow rates of 0.0 dm³/min (circle), 0.8 dm³/min (triangle), and 1.5 dm³/min (square). Microwave power: 1020 W; plasma gas: N₂ 14.0 dm³/min + O₂; carrier gas: Ar 0.5 dm³/min; observation height: 10 mm.

We conducted a comparison in the analytical performance between the nitrogen-oxygen MIP and a conventional argon ICP: the signal-to-background ratio (SBR) in Fig. 8 and the calibration curve in Fig. 9, when the same measuring system and test solutions of chromium were employed. In the measurement of the ICP, two analytical lines of chromium, Cr I 357.869 nm and Cr II 284.324 nm, were measured under each optimum measuring condition to estimate the detection limit. The SBR values in MIP always become larger than those in ICP, implying that better analytical conditions are able to be obtained in MIP. All the calibration curves in Fig. 9 follow a linear relationship having the correlation coefficient (R²) of 0.9992–0.9997. The 3σ detection limit was computed to be 0.39 mg/dm³ for the Cr I 357.869-nm line, and the Cr II 284.324-nm line was less sensitive. These results indicate that the nitrogen-oxygen MIP is more suitable to the trace-level determination of chromium compared to the argon ICP.

Fig. 8.

SBR of Cr I 357.869 nm (square) in a nitrogen-oxygen MIP, and SBRs of Cr I 357.869 nm (circle) and Cr II 284.324 nm (triangle) in an argon ICP. The MIP works at an oxygen flow rate of 1.5 dm³/min and the other measuring conditions are the same as in Fig. 7. The ICP is operated at a forward r.f. power of 1000 W, at a plasma gas of Ar 10.5 dm³/min, at a carrier gas of Ar 0.55 dm³/min, and at observation heights of 14 mm for Cr II 284.324 nm and 35 mm for Cr I 357.869 nm.

Fig. 9.

Calibration curve for Cr I 357.869 nm (square) in a nitrogen-oxygen MIP, and calibration curves for Cr I 357.869 nm (circle) and Cr II 284.324 nm (triangle) in an argon ICP. The measuring conditions in MIP and in ICP are the same as those described in Fig. 8.

4. Conclusions

The emission characteristics of an Okamoto-cavity MIP were investigated using a sample solution of chromium when nitrogen-oxygen mixed gas was employed as the plasma gas. The emission intensities of Cr I lines having small excitation energies were much elevated by adding oxygen of up to 10% to the nitrogen matrix gas, whereas those of Cr II lines requiring lager excitation energies decreased in the mixed gas plasma. The intensities of nitrogen band heads were also reduced when oxygen gas was added to the nitrogen plasma. This result implies that the ionization of chromium, which dominantly occurs through collisions with nitrogen excited species, can be hindered because the nitrogen excited species would be consumed through collisions with oxygen molecules to cause their dissociation. The intensity enhancement of the Cr I lines could improve the detection limit in the MIP-OES. The spatial distribution of the emission intensity from the plasma was estimated with a two-dimensionally imaging spectrograph, in order to optimize the measuring parameters for the determination of chromium. The best observation position was 10–15 mm above the front plate of the cavity when nitrogen of 14.0 dm³/min and oxygen of 1.5 dm³/min were mixed as the plasma gas. Then, the 3σ detection limit was obtained to be 0.09 mg/dm³ for the Cr I 357.869-nm line, which was superior to the detection limit in a conventional argon ICP.

Acknowledgment

This research is supported by Grant-in-Aids from The Iron and Steel Institute of Japan.

References

• 1)   J. A. C.  Broekaert: Analytical Atomic Spectrometry with Flames and Plasmas, Wiley -VCH Verlag, Weinheim, (2001).
• 2)   C. I. M.  Beenakker: Spectrochim. Acta, 31B (1976), 483.
• 3)   J.  Hubert,  M.  Moisan and  A.  Ricard: Spectrochim. Acta, 34B (1979), 50.
• 4)   Y.  Okamoto: Anal. Sci., 7 (1991), 283.
• 5)   Y.  Okamoto,  M.  Yasuda and  S.  Murayama: Jpn. J. Appl. Phys., 29 (1990), L670.
• 6)   Y.  Okamoto: J. Anal. At. Spectrom., 9 (1994) 745.
• 7)   K.  Ogura,  M.  Yamada,  Y.  Sato and  Y.  Okamoto: Appl. Specrosc., 55 (2001), 114.
• 8)   Z.  Zhang and  K.  Wagatsuma: J. Anal. At. Spectrom., 17 (2002), 699.
• 9)   T.  Maeda and  K.  Wagatsuma: Microchem. J., 76 (2004), 53.
• 10)   T.  Maeda and  K.  Wagatsuma: Spectrochim. Acta, 60B (2005), 81.
• 11)   T.  Maeda and  K.  Wagatsuma: Tetsu-to-Hagané, 91 (2006), 35.
• 12)   Z.  Zhang and  K.  Wagatsuma: Spectrochim. Acta, 57B (2002), 1247.
• 13)   Y.  Nishimoto and  K.  Wagatsuma: Bunseki Kagaku, 58 (2009), 153.
• 14)   M.  Ohata and  N.  Furuta: J. Anal. At. Spectrom., 13 (1998), 447.
• 15)   T.  Nakahara: ISIJ Int., 42 (2002), s114.
• 16)   K.  Wagatsuma: Appl. Spectrosc. Rev., 40 (2005), 229.
• 17)   C.  Kitaoka and  K.  Wagatsuma: Anal. Sci., 23 (2007), 1261.
• 18)   C.  Corliss and  W. R.  Bozman: Experimental Transition Probabilities for Spectral Lines of Seventy Elements, NBS Monograph 53, NBS, Washington DC, (1962).
• 19)   A. G.  Gaydon: Dissociation Energies and Spectra of Diatomic Molecules, Chapman and Hill, London, (1968).
• 20)   A.  von Engel: Ionized Gases, Clarendon Press, Oxford, (1965).