1. Introduction

Laser-induced breakdown spectroscopy (LIBS) is an analytical technique based on atomic emission spectroscopy of laser plasma [1–3]. In the previous decades, LIBS has been developed as an extremely popular and useful elemental analytical tool. The technique has several interesting advantages, which include little or no sample preparation, nearly nondestructive, real-time analysis, simultaneous multi-elemental analysis, and simplicity of design for field utilization. Therefore, LIBS has been widely used for the direct elemental analysis of various materials, including solids, liquids, and gases.

Despite of the several advantages of LIBS for spectral analysis, the technique suffers from several limitations and drawbacks, which mainly involve a relatively poor limit of detection and signal intensity instability, which in turn restrict its further development and quantitative analytical applications. Several methods have been proposed to improve the LIBS analysis performance. One popular method involves the use of a pair or a train of laser pulses to excite the sample within a short time window, and this method has successfully achieved lower limits of detection for the determination of several elements [4–6]. Laser ablation fast pulse discharge plasma spectroscopy, which is analogous to double-pulse LIBS but utilizes periodic oscillating discharge plasma generation instead of the second laser beam, could effectively reheat the laser-induced plasma and thereby improve the sensitivity, precision, and S/N of the measurement of various major and trace elements in soil [7,8]. Another interesting and more sensitive method that can be applied to improve the LIBS performance is based on the resonance excitation of the analyzed atoms induced by an additional tunable laser source [9,10]. However, this method suffers from the loss of instrumental simplicity and only a selected element can be excited and analyzed in the plume generated at each laser pulse.

Apart from the mentioned multiple excitation approaches, the spatial confinement of laser plasma has been also studied to enhance the intensity of plasma emission in single-pulse LIBS. Corsi et al. [11] and Zeng et al. [12] have studied the confinement in the craters or cavities characterized by a extremely small size (significantly smaller than laser plume size). In contrast, the spatial confinement effects in cavities that are significantly larger than plasma size have been also studied, including a cylindrical pipe with diameter larger than 10 mm [13], an aluminum hemispherical cavity with diameter of 11.1 mm [14], or a pair of parallel Al wall [15] (10mm away from each other), which confines the laser plume from two sides. Alternatively, a brass cylindrical chamber (ϕ 4mm × 4mm) [16,17] or a polytetrafluoroethylene (PTFE) cylindrical cavity (ϕ 3 mm × 1.5 mm) [18,19] has been used to confine plasma within cavities that are comparable to plasma size. In all these confinement cases, together with the plasma generation, shock waves will be produced and be reflected back to the front plasma by an obstacle in front of the plasma expansion path. This back reflection leads to an increase in the number of collisions among particles and atoms in high-energy states in the plasma, which in turn enhances the emission intensity. Clearly, the plasma confinement effect will vary using different cavity shapes and sizes. Therefore, to know how the cavity size will affect the confined plasma emission is interesting. However, no directed comparative study on the cavity size effect has been performed so far.

To date, the reported studies on the confinement effect mainly focused on the intensity enhancement and its mechanism, but rarely on the signal stability of the LIBS measurement, which in fact is a key parameter for real LIBS analytical application. Only two of the studies to date have studied signal precision in terms of the relative standard deviation (RSD) when the cavity was introduced with LIBS. The study of Wang et al. [18,19] demonstrated that a 1.5 mm-high and 3 mm-diameter cylindrical PTFE cavity can always increase the signal repeatability with a reduced RSD. However, Popov et al. [17] contrarily showed worse signal repeatability with a higher RSD when a 4mm-high and 4mm-diameter brass cavity confinement was introduced. Evidently, the effect of plasma confinement on signal quality will not only enhance the signal intensity, but will also affect the signal repeatability for LIBS measurement. In addition, these effects will vary with different cavity shapes and sizes. Therefore, to clarify whether and how the cavity confinement and cavity size will affect the signal stability is necessary. Moreover, to obtain the optimized cavity size for LIBS signal quality improvement is desired. In the present study, laser-induced plasmas were produced and confined inside cylindrical cavities with varying diameters and heights. The optical emission spectra of the confined plasmas were used to evaluate the signal intensity enhancement and measurement precision of the plume emission. Based on these results, an optimized cylindrical cavity was produced.

2. Experimental apparatus

Fig. 1 shows the schematic diagram of the experimental setup. The LIBS system is mainly composed of a pulse Nd:YAG laser and an Avaspec-2048 fiber optic spectrometer, which has been described elsewhere [20,21]. Briefly, a 1064 nm pulsed YAG laser with a pulse width of 10 ns and repetition rate of 1 Hz was used to produce plasma. The laser beam was focused on the sample surface by a convex lens (f = 70 mm) at a normal angle with respect to the surface. The emission from the laser plasma was collected using a collimating lens at an angle of 30° relative to the laser beam, and the intersection point of laser beam axis with the collimating lens optical axis is about 0.8 mm above the sample surface. A multimode sampling optical fibers was used to deliver the collected light to the fiber optic spectrometer, which covered a spectral range from 200 nm to 500 nm and has a nominal resolution of 0.1 nm. The spectrometer detector was a charge-coupled device linear array with 2048 pixels, and can be externally trigged by a DG535 pulse generator to start signal integration. The gate delay and integration time of the detectors were fixed at 1 μs and 2 ms, respectively. Forty cylindrical aluminum cavities were prefabricated by drilling on a set of aluminum plates with different heights from 0.5 mm to 2.5 mm and with different diameters from 1.5 mm to 6 mm.

 

Fig. 1 Schematic diagram of the experimental setup.

Download Full Size | PDF

The brass sample was mounted on a motorized XY translation stage so the laser can ablate a fresh spot at each shot. The two height-adjustable poles held the aluminum plate, so the aluminum plate can be placed near the sample surface but about 100 μm higher than the sample surface, and be able to avoid movement during the translation of the sample. The laser beams ablated the sample through the center of the hole. The standard analysis brass sample from the National Institute of Metrology of China was used in the present study. All experiments were performed under the normal atmospheric condition. To increase the signal-to-noise ratio and improve the measurement precision for each experiment, 100 spectra were obtained and all the spectra presented were the averaged spectra of 20 accumulations. Such an averaged spectrum served as a single measurement; therefore, five such independent replicate measurements were carried out for each experiment.

3. Results and discussion

3.1Effect of cavity size on line intensity

To characterize the plasma emission, neutral atomic and ionic lines were measured for LIBS generated with the presence of cavities with different diameters d and heights h, and were at a fixed laser energy of 30 mJ. Figure 2 shows the varying intensities of the atomic and ionic lines (for major element Cu I 271.89 nm, Cu II 224.70 nm, and minor element Fe II 274.93 nm, and Fe II 275.57 nm) obtained in the presence of different cavities with heights of 1, 1.5, 2, and 2.5 mm, respectively. In addition, at each height, the cavity diameters were increased from 1.5 mm to 6.0 mm. We note, in our light collection configuration, the measured line intensity in Fig. 2 with large height to diameter (H/D) ratio, such as the first data point in each graph, the light might partly be blocked by cavity wall and cannot be fully collected. But from other lower H/D ratio data points in Fig. 2 (H = 1 mm, 1.5 mm and 2 mm cavity), where the block effect can be neglected, we can find that at a specific cavity height, the intensity of all the selected lines initially increased first and then decreased along with increased cavity diameter. The figure also shows that the maximum intensity was reached at diameters of about 3 mm to 3.5 mm, which depends on the cavity height and specific spectral line. While for cavity with larger height 2.5 mm, if take into account the block effect, smaller diameter cavity seems have higher intensity enhancement, different to that of small height cavity confinement. However, the intensity was probably still lower than that obtained with 1 mm height and 3 mm diameter cavity.

 

Fig. 2 Variation of the intensities of selected atomic and ionic spectral lines of plasma in cavities of different heights. At each height, the cavity diameter are varying from 1.5 to 6 mm.

Download Full Size | PDF

Fig. 3 shows the variations of the spectral line intensities of the cavity-confined plasma at fixed cavity diameters of 3 mm and 3.5 mm, but with varying cavity heights of 0.5, 1, 1.5, 2, and 2.5 mm. All four line intensities initially increased and then decreased with increased cavity heights, and showed the maximum line intensity at height of 1 mm for the 3 mm-diameter cavity and at a height of 1.5 mm for the 3.5 mm-diameter cavity, which is dependent on the cavity diameter. Overall, the best signal intensity can be obtained with a confinement cavity with 1 mm height and 3 mm diameter. Moreover, at the optimized cavity size, the Cu II 224.70 nm line intensity of the confined laser plasma was about 2.5 times higher than that without a cavity confinement. As the laser parameters and sample character will affect the breakdown behavior, the plasma character and expansion, and then to some extent will have some effect on the optimized cavity geometry. Therefore, for a specific cavity confined LIBS application, the optimized cavity geometry probably will be different and should be optimized base on the used experiment condition.

 

Fig. 3 Variation of spectral intensity vs. different cavity heights at diameter 3 and 3.5 mm.

Download Full Size | PDF

3.2 Effect of cavity size on signal stability

In LIBS analysis, signal stability is an important factor affecting the reliability of the measured elemental concentration and is actually related to measurement precision in terms of the RSD of the specific line intensities in repeatedly measured spectra. To know whether and how the cavity confinement affects the measurement repeatability is interesting. Therefore, the RSD of several spectral line intensities (Cu I 271.89 nm, Cu II 224.70 nm, Fe II 274.93 nm, and Fe II 275.57 nm) were derived and presented to estimate the signal stability.

Figure 4(a) shows the RSD of the four selected line intensities of the spectra recorded with cavity confinement with different cavity heights and a fixed diameter of 3 mm. The RSD was calculated using five replicate measurement spectra, as stated in experimental section. Generally, if a suitable cavity size was used, the RSD with cavity confinement would be lower than that without cavity. In addition, with the optimized cavity size (3 mm diameter and 1 mm height), from which the most intense signal was obtained, the value of RSD was lower. The result was consistent with that reported by Wang et al. Cavity confinement is also confirmed to be an effective approach to improving measurement repeatability, assuming that the confinement cavity size was carefully selected. However, if the cavity size was not properly selected, the confinement will not evidently improve the signal stability. In some case, the signal stability even worsened with a higher RSD, as shown in Fig. 4(a), when a cavity with height of 2 mm was selected. This result might explain case reported by Popov et al., where a worse signal repeatability was obtained with the presence of cavity confinement.

 

Fig. 4 (a) Variation of RSD vs. different cavity heights at diameter 3 mm. The point at height equals 0 in the figure means without cavity confinement. (b) Spectrum of brass sample obtained by LIBS with cylindrical cavity of height 1.0 mm and diameter 3.0 mm and (c) LIBS without cavity confinement.

Download Full Size | PDF

3.3 LIBS Spectrum with optimized cavity

The typical time-integrated emission spectra of the brass sample measured using cavity-confined LIBS and traditional LIBS are shown in Fig. 4(b) and Fig. 4(c) . The spectra were averaged over 20 laser pulses to increase the sensitivity of the system and reduce the standard deviation. Figure 4(b) is the spectrum for laser plasma in a cavity with height of 1 mm and diameter of 3 mm. Fig. 4(c) is the spectrum obtained using traditional LIBS technique. In both cases, the laser energy was 30 mJ. Several selected atomic and ionic line intensities, and RSD of five measurements are listed in Table1. Considering the confinement effect, a 2.5- to 3.5-fold line intensity enhancement was observed with the presence of optimized cylindrical cavity, which is dependent on the specific selected line. In addition, using this specifically optimized cavity with height of 1 mm and diameter of 3 mm, an improved signal quality with reduced RSD was observed. Therefore, cavity confinement did not only enhance the line intensity of the laser plasma emission (as suggested by other recent studies) but also effectively increased measurement repeatability assuming that cavity size was carefully selected.

Table 1. Improvement of Line Intensity and RSD of Several Selected Lines

View Table

4. Conclusion

The emission spectra of the confined laser-induced plasma of a brass sample with cavities of different heights and diameters were measured and compared with the emission spectra of laser plasma without a cavity. The effects of cavity size on the plasma emission intensity and measurement precision in terms of RSD were carefully investigated. As the cavity diameter increased from 1.5 mm to 6 mm, the line intensity initially increased and then decreased, and had a maximum value at diameter of about 3 mm to 3.5 mm, depending on the cavity height and specific spectral line. In addition, if a suitable cavity size was selected, the line intensity and stability of the confined plasma emission can be improved. Through the signal intensity enhancement and measurement repeatability, the optimized confinement cavity size was determined, which was at height of 1 mm and diameter of 3 mm. As the plasma volume and expansion are dependent on the experimental parameters, therefore, for a specific cavity confined LIBS application, the optimized cavity geometry probably will be different and should be optimized base on the used experiment condition.