Development of Liquid Sampling and Correlation Method for Laser-induced Breakdown Spectroscopy

• Main Authors: Jer-Shing Huang, 黃哲勳
• Other Authors: King-Chuen Lin
• Format: Others
• Language: en_US
• Published: 2004
• Online Access: http://ndltd.ncl.edu.tw/handle/35425054105535239698

博士 === 國立臺灣大學 === 化學研究所 === 92 === Laser-induced breakdown (LIB) technique is applied to detect trace metal contained in liquid. The sample in liquid droplets is generated with an electrospray ionization needle. The microdroplets interact with an impinging laser pulse at 2 mm down stream from the needle tip. A sequence of single-shot time-resolved LIB emission signals of Na, K, and Al is detected, respectively. The signal intensity integrated within a gate is linearly correlated with plasma current obtained simultaneously on a single-shot basis. The correlation plot exhibits a straight line, of which the slope increases with the sample concentration, but appear to be irrespective of different matrix salts up to a 2000 mg/l concentration. Given the calibration curves, standard deviations of the blank measurement, and the focused cross sections of the incident pulse beam, the detection limits may reach 0.63 ± 0.02 (0.3pg), and 1.15 ± 0.04 (0.5pg), and 43 ± 5mg/l (21pg) for Na, K, and Al, respectively. The salt matrix effect on detection limits is studied with matrix salts having the same K+ cation but different anions. The current normalization might have probably taken into account the ablated amount of the sample and the plasma temperature. Accordingly, the LIB/current correlated analysis becomes efficient to suppress the signal fluctuation, improve the LOD determination, and concurrently correct the matrix effect. In addition, the plasma continuum background (CB) emission signal is found to hold the same fluctuation pattern with the LIB emission signal, and therefore can be utilized to normalize the LIB signal fluctuation. The effects of laser pulse energy on the detection limits of two normalization methods, i.e. current correlation and continuum emission correlation, are examined as well. The correlation between continuum emission signal and LIB emission signal no longer exits if laser pulse energy is tuned away from the optimum. Nevertheless, correlation between LIB emission and current seems not sensitive to the change in laser pulse energy. The influence of laser pulse energy is discussed as well. Given a laser radiation emitting at 355 nm with the energy fixed at 23 mJ, the current normalization method may achieve a LOD of 1.0 mg/l for the Na analysis, while the CB emission normalization reaches only 12 mg/l. The current normalization probably has taken into account the ablated amount of the sample and the plasma temperature such that the obtained linearity of the correlation plot may extend to a much wider range of laser energy. In contrast, the CB emission normalization is sensitive to the plasma temperature, yielding a correlation linearity limited in a small range of laser energy. Assuming a local thermal equilibrium (LTE) in the plasma, Saha-Boltzmann equation is applied and plasma temperature is estimated by Boltzmann plot method. Two-line ratio from OMA spectrum is considered as a factor of plasma temperature and is taken into account to explain the deviation from linearity in the correlation plot of LIB emission versus CB emission. Dramatic improvement in the linearity of correlation between the LIB emissions versus CB emission is obtained in the Ca and Sr experiment under the consideration of plasma temperature factor. On the contrary, correlation between Ca emission and current signal is usually perfectly linear and needs no further consideration of plasma temperature. We conclude that the laser-induced current signal is an independent signal, which correlates well with the emission signal and is capable of normalizing the emission signal fluctuation without further consideration of plasma temperature. With electrospray sampling, we benefit not only from the current normalization of signal fluctuation but also its ability to connect to an FIA system. LIBS coupled with FIA system is firstly demonstrated in this work. With the a simple home made preconcentration column and FI manifold, solutions containing dilute Al(III) are successfully detected and the LOD is more than one order improved. The Al(III) cation is first absorbed by immobilized Chromotrope 2B, which serves as a strongly chelating agent to Al(III). A 0.5M HCl solution in half and half methanol-water mixture is then pump into the column to wash out the retained Al(III). Simple FI manifold is configured to facilitate washing, original sample detection, sample loading, and eluating the concentrated sample out. Detection limit with FI preconcentration is reduced to 2.9 mg/l. Compared with LOD without FI treatment, it’s more than one order improved. Though it’s yet not that low to detect trace Al in bio-material, we first time provide the possibility to connect LIBS with a simple on-line preconcentration system.