Plasma Science and Technology, Vol.17, No.11, Nov. 2015

Signal Detection of Carbon in Iron-Based Alloy by Double-Pulse Laser-Induced Breakdown Spectroscopy∗ LIN Xiaomei (林晓梅), LI Han (李晗), YAO Qinghua (姚清华) Department of Electronics and Electrical Engineering, Changchun University of Technology, Changchun 130012, China

Abstract

Although single-pulse lasers are often used in traditional laser-induced breakdown spectroscopy (LIBS) measurements, their measurement outcomes are generally undesirable because of the low sensitivity of carbon in iron-based alloys. In this article, a double-pulse laser was applied to improve the signal intensity of carbon. Both the inter-pulse delay and the combination of laser wavelengths in double-pulse laser-induced breakdown spectroscopy (DP-LIBS) were optimized in our experiment. At the optimized inter-pulse delay, the combination of a first laser of 532 nm and a second laser of 1,064 nm achieved the highest signal enhancement. The properties of the target also played a role in determining the mass ablation enhancement in DP-LIBS configuration.

Keywords: DP-LIBS, carbon, signal detection, iron-based alloy

PACS: 52.38.Dx DOI: 10.1088/1009-0630/17/11/12 (Some figures may appear in colour only in the online journal)

1

Introduction

bon detection by SP-LIBS during the surface cleaning of materials, including polymer identification, cleaning monitoring, and carbon quantification. Unfortunately, SP-LIBS is difficult to use for quantifying carbon in steel samples because of its relatively low signal-noise ratio and weak sensitivity [15] . Double-pulse LIBS (DP-LIBS) presents several advantages over SP-LIBS, including signal enhancements of approximately 1-2 orders of magnitude and a lower limit of detection [16] . For example, via optimizing the experimental parameters, the signal intensity of DPLIBS was shown to be 350 times greater than that of SP-LIBS [17] . Over the last decade, efforts devoted toward optimizing and enhancing DP-LIBS have slowly increased; however, the mechanism of this improvement remains unclear [18] . In this article, the effect of ambient gas on carbon detection by DP-LIBS was studied. The fundamental (i.e., 1,064 nm) and second (i.e., 532 nm) harmonics of the pulse laser were used, and the effects of the two pulse lasers were compared. The inter-pulse delay between lasers was identified in a specific device. The role of the target’s properties in mass ablation was investigated by comparing the electron temperatures of the two samples.

Real-time detection and quantification are two important quality control factors in the production of iron-based alloys. Although inductively coupled plasma mass spectrometry and photoelectric direct reading spectrometry have been developed to analyze carbon content in iron-based alloys, these methods cannot be applied to real-time quantitation because of their low sensitivity and sample pretreatment requirements. Laser-induced breakdown spectroscopy (LIBS) has been reported to be an attractive technology for real-time analysis because of its many advantages, such as a short measurement time, minimal sample preparation, and simultaneously multi-element analyses. Several studies have been conducted to determine the effects of the plasma properties and the ambient environment on the experimental parameters and LIBS performance [1−11] . Most of these previous studies have applied single-pulse LIBS (SP-LIBS) because this technique is suitable for measuring metal elements, a process that requires high sensitivity. As carbon is less sensitive than metal elements, limited research on carbon analysis has been performed. Li et al. conducted a series of carbon measurements in coal samples and proved the suitability of SP-LIBS for quantitative analysis of samples where carbon is the major elemental component [12,13] . LIBS has also been applied to carbon content detection in soils; in particular, the concentration detected by SP-LIBS ranges from 0.16% to 4.32% [14] . Ctvetnickova et al. reported several applications of car-

2

Experiment

A double-pulse laser (Vlite-200, Beamtech Optronics Co., Ltd., Beijing, China) was used in this experiment.

∗ supported by National Natural Science Foundation of China (No. 51374040) and the National Key Scientific Instrument and Equipment Development Project of China (No. 2014YQ120351)

953

Plasma Science and Technology, Vol.17, No.11, Nov. 2015 The lasers in this system featured working wavelengths of 1,064 nm and 532 nm, pulse widths of 10 ns and 8 ns, and pulse energies of 300 mJ and 180 mJ. The detection system consisted of spectrometers (AvaSpec-ULS2048, Avantes, Apeldoom, Netherlands) that were equipped with a linear charge-coupled device (CCD) array containing 2,048 elements. Spectrometers and CCDs were used to separate element-specific spectral lines and convert the intensities to electrical signals. The wavelength region ranged from 180 nm to 310 nm at a resolution of 0.07 nm. The minimum integration of the spectrometer was 2 ms and the inherent time delay was 1.3 µs. Both of the laser beams were focused on the target through a quartz lens with a diameter of 200 mm. The radiation emitted by the plume was collected by fiber optics, with the collimating lens positioned 45o to the direction of the plume and then transmitted to the fiber optic spectrometer with its integrated software. The sample was placed on a rotating stage to provide a fresh surface after each laser pulse and to improve sampling uniformity. Inert gas was used to minimize interference from the ambient environment, e.g., oxygen in the air. In our experiment, C 193.09 nm was chosen for carbon analysis because strong interference from the Fe emission line at 247.95 nm may be observed in the carbon line at 247.86 nm [19] .

3 3.1

Fig.1 LIBS intensity of the carbon line at 193.09 nm versus the ambient gas pressure (inert gas) Table 1. Percentage of main elements in the samples used in the experiment Element (%) Sample Sample Sample Sample Sample Sample Sample Sample

1 2 3 4 5 6 7 8

Fe

Cr

C

34.28 35.29 34.21 34.77 33.77 33.59 34.61 34.11

56.93 55.51 56.91 55.91 56.60 56.02 56.61 56.96

7.62 7.92 7.77 7.50 7.43 7.27 7.36 7.72

Results and discussion DP-LIBS spectra were obtained by the two laser pulses at 1,064 nm with a pulse energy of 100 mJ. The region of inter-pulse delay was 0-21 µs and 100 spectra were collected for each sample. The other experimental parameters were held constant. Variations in the signal intensities of these lines as a function of the delay between pulses are shown in Fig. 2.

Inter-pulse delay optimization

The signal enhancement mechanism of the doublepulse laser is currently unclear. Laser shielding is believed to be the primary mechanism influencing the laser, material interaction efficiency, and instability [20] . The inter-pulse delay between the two lasers affects the ambient environment of the plasma plume, which has a crucial effect on plasma shielding. During optimization of experimental parameters, the delay time between the lasers was found to play an important role in carbon detection in an iron-based alloy. To determine the effect of ambient gas on carbon signal detection [21] , carbon emission intensity as a function of the pressure of the ambient inert gas was studied. As shown in Fig. 1, when the pressure was near 0.1 Torr, carbon emission increased as the pressure increased in the range of 0.1-50 Torr, and then decreased when the pressure exceeded 50 Torr. Carbon emission was weak at atmospheric pressure. The pressure must be within the range of 0-100 Torr [20] to achieve a compromise between the ablation mass and the plasma temperature. In the experiment, the optimized pressure was determined to be approximately 50 Torr, at which a proper compromise between the ablation mass and plasma temperature may be achieved. Eight high-carbon ferrochrome samples were used in the experiment. Table 1 shows the major elemental contents of our samples.

Fig.2 Signal intensity versus inter-pulse delay between the two laser pulses at 1,064 nm, each with 100 mJ pulse energy

The signal intensity is reported as an average of eight samples. When the inter-pulse delay was approximately 0.05 µs, the carbon signal intensity was much lower than the signal intensity of the metal elements and the spectrum intensity of the metal elements increased 954

LIN Xiaomei et al.: Signal Detection of Carbon in Iron-Based Alloy by Double-Pulse LIBS

3.2

faster in the initial stage of detection. Maximum signal intensity enhancement of carbon at 193.09 nm was observed at 7 µs inter-pulse delay between lasers. Strong background emissions were observed when the interpulse delay was approximately 7 µs. As the inter-pulse delay between the lasers increased, both the carbon emission line at 193.09 nm and the background emission began to degrade; however, the background signal drops much faster. The experiment was repeated using a single-pulse laser (532 nm, 150 mJ) and the second harmonic of the double-pulse laser (532 nm, 150 mJ), each with an energy of 75 mJ. A portion of the spectrum is shown in Fig. 3. The signal intensity increased by approximately 10-fold in the DP-LIBS system compared with that in SP-LIBS. The maximum signal intensity appeared at an inter-pulse delay of 7 µs between the lasers.

Effect of the wavelengths on the signal intensity

To clarify the effect of laser wavelengths on the signal intensity, different wavelength combinations were selected, as shown in Table 2. Table 2. Different wavelength combinations used in the experiment Time

First laser wavelength (nm)

Second laser wavelength (nm)

1st 2nd 3rd 4th

532 1,064 532 1,064

1,064 532 532 1,064

In the first part of the experiment, the first laser (532 nm, 100 mJ) and the second delayed laser (1,064 nm, 100 mJ) were used, and the variation of the signal intensity with inter-pulse delay was tested. The maximum signal intensity was observed at 7 µs interpulse delay. In the second part of the experiment, the laser wavelengths of the first and the second lasers were interchanged; that is, the first laser featured characteristics of 1,064 nm, 100 mJ, and the delayed laser featured characteristics of 532 nm, 100 mJ. The same laser wavelengths were used as the first (532 nm, 100 mJ) and the second (532 nm, 100 mJ) lasers in the third part of our experiment. The fourth part showed the same time delay between lasers as in the first part of the experiment. All of the inter-pulse delay times were 7 µs under the same experiment conditions. Fig. 4 shows the signal intensity variation with the laser wavelengths. When the first laser pulse was shorter, the observed emission intensity was higher than that produced by the first laser pulse.

Fig.3 Part of the LIBS spectrum with a laser pulse energy of 150 mJ at 532 nm. (a) DP-LIBS spectrum (black line) with inter-pulse delay of 7 µs and energy of 75 mJ, (b) SP-LIBS spectrum (blue line)

According to the principle of plasma shielding, the plasma absorbs laser energy through the inverse bremsstrahlung mechanism and the increased electron temperature hinders the remaining part of the laser travelling through the plasma to reach the sample. In DP-LIBS, the gas density may be lower than the ambient gas density inside the shockwave [22,23] . In this case, the effect of plasma shielding is weak and most of the second laser energy is able to reach the iron alloy and increase the ablation mass. In other words, laser ablation by the second pulse in a low-pressure environment is likely to enhance carbon emission. The signal was evidently enhanced when the first and the second lasers were fired at the same time (i.e., tdelay > toptim ), the first plasma relapsed toward ambient conditions and no dual-pulse ablation or emission enhancements can be seen. The signal enhancement mechanisms seem to have no effect; in fact, the potential cooperation in this mechanism would produce a much larger ablation mass than any single-pulse mechanism.

Fig.4 Comparison of the enhancement factor using part of the DP-LIBS spectrum

Maximum signal intensity was observed with a combination of 532 nm (the first laser) and 1,064 nm (the second laser). Because the wavelength was shorter than the fundamental emission (1,064 nm), the focus of the beam on the sample surface was tighter [24] . As the inverse bremsstrahlung absorption of the laser energy is proportional to λ3 , absorption of the first laser energy 955

Plasma Science and Technology, Vol.17, No.11, Nov. 2015 via inverse bremsstrahlung at the surface is lower and a laser pulse with a shorter wavelength produces a larger ablation but colder plasma with respect to the fundamental wavelength with the same energy [25] . In contrast, the plasma produced by the first pulse can absorb longer pulse wavelengths because of its high electron density and further increase its temperature. Therefore, in DP-LIBS, when the wavelength of the first laser is shorter than the fundamental emission, the spectrum signal is enhanced and the detection sensitivity of carbon in the iron-based alloy can be improved.

3.3

Target properties

Silicochromium and low chromium alloy were used to study the possible relationship between the target physical properties and the observed signal intensity. Table 3 presents the thermal and electronic properties of the main elements in the targets and Table 4 shows the percentage of elements in the targets. Table 3. targets Element Fe Cr Si

Fig.5 Plasma electron temperature at different inter-pulse delays

The high value of the electron temperature at 7 µs is due to re-excitation by the second laser-induced plasma. The plasma temperature produced by the lowchromium alloy is slightly higher than the one produced by silicochromium. The thermal properties (e.g., melting and boiling temperature, and melting and evaporation heats) are correlated with the molten and vaporized volume of the sample and with the breakdown threshold. The first ionization energy of the metal may be related to the plasma shielding capability of the laser beam. Besides the ambient gas, the ablation mechanism and temperature also affect DP-LIBS. Ablation enhancement is related not only to the ambient gas but also to the target’s properties. Electron temperature is an important parameter of the plasma for the carbon signal. Mass ablation affects the electron temperature. Although the effect of the target properties on the plasma temperature is not significant, it should not be ignored because it may affect signal detection in DP-LIBS configuration.

Thermal properties of the main elements in the Tm (K)

Tb (K)

Ev -heat λ Eion 2 −1 (kJ/mol) (cm · s ) (kJ/mol)

1808 3023 2130 2945 1687 3173

349.6 344.3 348.2

0.130 0.290 0.129

759 635 786

Tm , melting temperature; Tb , boiling point; Ev -heat, evaporation heat; λ, thermal diffusivity; Eion , first ionization energy. Table 4.

Percentage of elements in the targets

Element (%)

Cr

Si

C

P

Silicochromium

31.93

46.97

0.025

0.0269

Low-chromium

64.12

0.286

0.170

0.0327

In this case, the samples were induced by the doublepulse laser with two laser pulses at 1,064 nm with energies of approximately 100 mJ. Assuming that the ablation plasma is in local thermal equilibrium, the electron temperature was estimated by using the Boltzmann plot method from two spectral lines of carbon at 193.09 nm and 195.9952 nm. Population of the excited states follows the Boltzmann distribution. The relative spectral line intensity, Imn is given as à ! ¡ N (T ) ¢ λmn Imn Em ln =− + ln , (1) hcgm Amn κTe U (T )

4

Conclusion

Inter-pulse delay affects the ambient gas of the plasma plume and the plasma shielding effect produced by the first pulse. The optimal inter-pulse delay between the lasers for the detection of carbon was found to be approximately 7 µs. The signal intensity of DPLIBS was enhanced by approximately 10-fold compared with the SP-LIBS configuration. During DP-LIBS, the wavelength of the first laser was shorter than that of the second laser and the carbon detection was more accurate because of the limited plasma shielding effect. Carbon signal detection through the DP-LIBS configuration correlated not only with ambient gas but also with target properties. Our results can be used to improve the sensitivity of carbon detection in iron-based alloy by LIBS.

where, λmn , Amn , gm , h, and c are the wavelength, the transition probability, the statistical weight of the upper level, Planck’s constant, and the speed of light in vacuum, respectively. Em is the upper level energy, Te is the electron temperature, κ is the Boltzmann constant, U (T ) is the partition function, and N (T ) is the total number density of neutrals. In Fig. 5, the plasma temperature versus inter-pulse delays in the two types of alloys exhibits a similar variation trend. 956

LIN Xiaomei et al.: Signal Detection of Carbon in Iron-Based Alloy by Double-Pulse LIBS

References 1 2 3 4 5 6 7 8

9 10 11 12 13 14

15

Adrain R S, Watson J. 1984, J. Phys. D: Appl. Phys., 17: 1915 Zhang S H, Yu X L, Li F, et al. 2012, Optics and Lasers in Eng., 50: 877 Wang Z, Li L Z, West L, et al. 2012, Spectrochim. Acta Part B, 68: 58 Yao S C, Lu J D, Chen K, et al. 2011, Appl. Surf. Sci., 257: 3103 Ma Q L, Motto-Ros V, Lei W Q, et al. 2010, Spectrochim. Acta Part B, 65: 707 Suyanto H, Lie Z S, Niki H, et al. 2012, Anal. Chem., 84: 2224 Heilbrunner H, Huber N, Wolfmeir H, et al. 2012, Appl. Phys. A, 106: 15 Wolf-Rottke L, Radziemski J. 2006, Handbook of Laser-Induced Breakdown Spectroscopy. John Wiley & Sons Ltd., England, Chichester Vadillo J M, Laserna J J. 2004, Spectrochim. Acta Part B, 59: 147 Matiaske A M, Gornushkin I B, Panne U. 2012, Anal. Bioanal. Chem., 402: 2597 Nagli L, Gaft M, Gornushkin I. 2011, Anal. Bioanal. Chem., 400: 3207 Li X W, Wang Z, Fu Y T, et al. 2014, Spectrochim. Acta Part B, 99: 82 Li X W, Wang Z, Fu Y T, et al. 2014, Applied Spectroscopy, 68: 955 Maetin M Z, Wullschleger S D, Garten C T Jr, et al.

16

17 18 19 20 21 22

23 24 25

2007, Laser-Induced Breakdown Spectroscopy. Elsevier, Oxford Ctvrtnickova T, Mateo M P, Yanez A, et al. 2010, Journal of Optoelectronics and Advanced Materials, 12: 668 Maria Margaretha Suliyanti, Affnur Hidayah, Marincan Pardede, et al. 2012, Spectrochim. Acta Part B, 69: 56 Ahmed R, Baig M A. 2010, IEEE Trans. Plasma Sci., 38: 2052 Ahmed R and Baig M A. 2009, J. Appl. Phys., 106: 033307 Kunwar P S, Singh N, Pawar M S, et al. 2010, J. Glob. Pharma. Technol., 2: 133 Babushok V I, DeLucia F C, Gottfried J L, et al. 2006, Spectrochim. Acta Part B, 61: 999 Lida Y. 1990, Spectrochim. Acta Part B, 45: 1353 Zeldovic Y B, Raizer Y P. 1987, Physics of ShockWaves and High-Temperature Hydrodynamic Phenomena. Academic Press, New York Register J, Scaffidi J, Angel S M. 2012, Appl. Spectrosc., 66: 869 Sdorra W, Brust J, Niemax K. 1992, Mikrochim. Acta, 108: 1 Cristoforetti G, Lorenzetti G, Benedetti P A, et al. 2008, J. Phys. D: Appl. Phys., 42: 312

(Manuscript received 13 April 2015) (Manuscript accepted 17 June 2015) E-mail address of corresponding author YAO Qinghua: [email protected]

957