Anal. Chem. 1996, 68, 1101-1109
Time-Resolved Inductively Coupled Plasma Mass Spectrometry Measurements with Individual, Monodisperse Drop Sample Introduction Michael P. Dziewatkoski, Lizla B. Daniels, and John W. Olesik*
Laboratory for Plasma Spectrochemistry, Laser Spectroscopy and Mass Spectrometry, Department of Geological Sciences, The Ohio State University, 125 South Oval Mall, Columbus, Ohio 43210
Individual ion clouds, each produced in the ICP from a single drop of sample, were monitored using timeresolved mass spectrometry and optical emission spectrometry simultaneously. The widths of the ion clouds in the plasma as a function of distance from the point of initial desolvated particle vaporization in the ICP were estimated. The Li+ cloud width (full width at halfmaximum) varied from 85 to 272 µs at 3 and 10 mm from the apparent vaporization point, respectively. The Sr+ cloud width varied from 97 to 142 µs at 5 and 10 mm from the apparent vaporization point, respectively. The delays between optical and mass spectrometry signals were used to measure gas velocities in the ICP. The velocity data could then be used to convert ion cloud peak widths in time to cloud sizes in the ICP. Li+ clouds varied from 2.1 to 6.6 mm (full width at half-maximum) and Sr+ clouds varied from 2.4 to 3.5 mm at the locations specified above. Diffusion coefficients were estimated from experimental data to be 88, 44, and 24 cm2/s for Li+, Mg+, and Sr+, respectively. The flight time of ions from the sampling orifice of the mass spectrometer to the detector were mass dependent and varied from 13 to 21 µs for Mg+ to 93 to 115 µs for Pb+. The processes that convert sample aerosol into ICPMS signals include: droplet desolvation, particle vaporization, atomization, ionization, diffusion within the ICP, and physical transport of ions from the ICP to the MS detector. Quantitatively understanding how each of these processes are affected by plasma conditions and sample matrix is key to unraveling the origins of matrix effects and polyatomic species that can degrade accuracy in ICPMS. Here, initial results on diffusion and transport of ions from the ICP to the MS detector will be presented. Diffusion results in a loss of signal, as the concentration of ions in the center of the ICP decreases with time, and in the case of sample aerosols with high number densities, a mixing of ions originating from different sample drops. In order to convert the time-resolved ICPMS signals due to individual ion clouds into physical size of the ion clouds, gas velocities in the ICP must be measured. The transport of ions from the ICP to the MS detector is very inefficient. Information on the mass-dependent flight time of ions will allow models of ion transport to be tested and refined. It may also lead to further insight into where the positive ion beam forms from the neutral plasma sampled from the ICP. This is a key process related to space charge effects and ion transmission loss. 0003-2700/96/0368-1101$12.00/0
© 1996 American Chemical Society
It is difficult to separately investigate each of the processes that control ICPMS signals, discussed above, when a conventional nebulizer is used for liquid sample introduction. The aerosol entering the plasma consists of many thousands of polydisperse drops per milliliter of gas. Desolvation, vaporization, and atom and ion formation will occur at different places in the plasma (times in the plasma), depending on the drop size. Furthermore, because of the high drop number density, the observed ICPMS signals (at any interval in time) have contributions from analyte produced from many drops (that may be in different stages in the conversion from liquid drops to vaporized particles to free atoms and ions). If sample was introduced into the ICP as a reproducible train of individual, monodisperse droplets of selectable size, new insight into fundamental processes and chemical matrix effects in ICPMS could be gained that was not previously attainable. Each drop would complete desolvation at the same point in the plasma. Particle vaporization would begin at a fixed point in the ICP. Location in the ICP can be directly related to time by measuring the plasma gas velocity. The kinetics of drop desolvation, particle vaporization, and ionization could be measured because these processes would be at least partially separated in time and space. Diffusion would be directly observable. By observing ICPMS signals due to individual ion clouds, the flight time and ion cloud expansion can be directly monitored. Previously, researchers have shown that unique fundamental information on the desolvation, vaporization, atomization, ionization, and diffusion processes in large, laminar flow flames could be obtained by introducing monodisperse droplets of sample.1-7 Chemical matrix effects due to shifts in ionization equilibrium and changes in vaporization resulting in lateral diffusion effects could be observed.3 Until recently, attempts to inject individual monodisperse drops into ICPs from the side or bottom, using monodisperse drop generators similar to those used with flames,8 failed because of the rapidly expanding ICP gas, extensive plasma heterogeneity and small size of the ICP compared to large, laminar flow flames. (1) Hieftje, G. M.; Malmstadt, H. V. Anal. Chem. 1968, 40, 1860-1867. (2) Bastiaans, G. J.; Hieftje, G. M. Anal. Chem. 1974, 46, 901-910. (3) Boss, C. B.; Hieftje, G. M. Anal. Chem. 1979, 51, 1897-1905. (4) Clampitt, N. C.; Hieftje, G. M. Anal. Chem. 1972, 44, 1211-1219. (5) Bleasdell, B. D.; Wittig, E. P.; Hieftje, G. M. Spectrochim. Acta 1981, 36B, 205-213. (6) Childers, A. G.; Hieftje, G. M. Anal.Chem. 1993, 65, 2761-2765. (7) Hieftje, G. M.; Miller, R. M.; Pak, Y.; Wittig, E. P. Anal. Chem. 1987, 59, 2861-2872. (8) Russo, R. E.; Withnell, R.; Hieftje, G. M. Appl. Spectrosc. 1981, 35, 531536.
Analytical Chemistry, Vol. 68, No. 7, April 1, 1996 1101
French, Etkin, and Jong9 designed a laminar flow furnace with an on-demand, piezoelectric-based drop generator [called the Monodisperse Dried Microparticulate Injector (MDMI)] that allowed individual drops to be reproducibly produced, desolvated and transported into the ICP. French et al.9 interfaced the MDMI to an ICPMS and made time-integrated measurements of ion signals from both analyte and molecular oxide species. A reduction in oxide signals (relative to elemental ion signals) and a decrease in MS signal noise was observed when using the MDMI rather than a conventional cross-flow nebulizer and spray chamber. Olesik and Hobbs10 described both spatial imaging and timeresolved ICPOES measurements using the MDMI for sample introduction. The data showed that monodisperse drops of selectable size were reproducibly transported into the ICP. The location in the plasma where desolvation and vaporization occurs could be controlled by varying drop size by adjusting either the gas flow rate through the furnace or the furnace temperature. Atom emission intensity peaked low in the plasma while ion emission peaked higher in the plasma. Compared to a conventional nebulizer, which produces a polydisperse aerosol, the vertical emission profiles showing the transition from atom to ion occurred in a much more sharply defined region when the MDMI was used for sample introduction. The time scale for analyte particle vaporization was less than 100 µs. The time in the ICP required for drops of particular sizes to be completely desolvated was estimated from the location of initially detected atom emission intensity10 and typical gas velocities in the ICP. Under the plasma conditions used, drops that were initially 5 µm in diameter required roughly 1 ms to desolvate, while drops that were initially 20 µs in diameter required roughly 2 ms to desolvate. Applied to ICPMS, the MDMI has the potential to provide information on the ion-sampling process, ion transport through the mass spectrometer, matrix effects, and the formation of interfering species within the plasma and mass spectrometer from time-resolved measurements on a microsecond time scale. Time resolved optical emission,11-15 mass spectrometry16 and laser-induced fluorescence14 measurements have been used in the past to monitor the effect of individual, incompletely desolvated drops or vaporizing particles on fundamental processes occurring within the ICP. The effects of individual desolvating droplets and vaporizing particles on excitation, ion number densities, and the transmission efficiency of ions from the ICP to the MS detector were investigated.14,16 Cicerone and Farnsworth15 measured plasma gas velocities by monitoring time-resolved emission from a single atom cloud at two different heights, a known distance apart, in an ICP. Only limited information on the desolvation of sample drops and analyte particle vaporization could be gleaned from timeresolved experiments that observed individual, incompletely desolvated drops and vaporizing particles. The initial size of drops that were incompletely desolvated in the ICP was estimated on (9) French, J. B.; Etkin, B.; Jong, R. Anal. Chem. 1994, 66, 685-691. (10) Olesik, J. W.; Hobbs, S. E. Anal. Chem. 1994, 66, 3371-3378. (11) Olesik, J. W.; Smith, L. T.; Williamsen, E. J. Anal. Chem. 1989, 61, 20022008. (12) Olesik, J. W.; Fister, J. C., III. Spectrochim. Acta 1991, 46B, 851-868. (13) Fister, J. C., III; Olesik, J. W. Spectrochim. Acta 1991, 46B, 869-883. (14) Hobbs, S. E.; Olesik, J. W. Spectrochim. Acta 1993, 48B, 817-833. (15) Cicerone, M. T.; Farnsworth, P. B. Spectrochim. Acta 1989, 44B, 897907. (16) Hobbs, S. E.; Olesik, J. W. Anal. Chem. 1992, 64, 275-283.
1102 Analytical Chemistry, Vol. 68, No. 7, April 1, 1996
Table 1. Instrumentation and Experimental Parameters ICPMS (with Laboratory-Built Interface) plasma power supply Plasma-Therm HFL-1500G, 40 MHz impedance matcher Plasma-Therm AMNPS-2 plasma applied power 1100 W reflected power
