Anal. Chem. 2003, 75, 747-753
Elemental Fractionation Studies in Laser Ablation Inductively Coupled Plasma Mass Spectrometry on Laser-Induced Brass Aerosols Hans-Rudolf Kuhn and Detlef Gu 1 nther*
Swiss Federal Institute of Technology (ETH), Laboratory for Inorganic Chemistry, HCI Ho¨nggerberg, G113, CH-8093 Zurich, Switzerland
Previous investigations on laser-induced aerosols of brass samples showed that preferential vaporization of zinc occurs during the ablation process leading to elemental fractionation and limited possibilities for non matrix matched calibration. In a variety of experiments carried out within this study it is shown that multiple effects are complicating the quantification of brass using LA-ICPMS. It is shown that the ablated copper and zinc is not homogeneously distributed within the laser-produced aerosol. Copper was found enriched up to 100% in particles larger than 100 nm as shown from EDX measurements (electron excited) on individual particles, and zinc was enriched by over 40% in the particles smaller than the lowest measurable particle size (below 100 nm or in the vapor phase). Solution nebulization analysis on digested filter-collected aerosols results in a higher Cu/ Zn ratio than the certified value for the brass sample. ESEM pictures with analysis of the electron excited X-rays measured on the filter-collected material support the results showing copper enrichment. However, online LAICPMS measurements carried out under the same operating conditions as for filtering show a copper depletion within the ICP, which leads to the conclusion of partial vaporization and ionization of the aerosol particles in the ICP. The larger particles containing more or exclusively copper are not completely ionized. Within this study, three sources of elemental fractionation can be distinguished and described: (A) The ablation process leads to no measurable copper enrichment at the ablation crater rim. (B) Zinc deposition between the ablation site and the aerosol collection on filters leads to an up to 37% higher Cu/Zn ratio on the filter in comparison to the certified value. (C) On-line laser ablation aerosols measured within the ICPMS lead to significantly lower Cu/Zn ratios in comparison to the certified value. (D) Combination of the various studied sources of fractionation can finally lead to an agreement between measured and certified values due to inverse overlapping of various fractionation sources. Laser ablation inductively coupled plasma mass spectrometry (LA-ICPMS) is a widely used technique for the analysis of various * Corresponding author: (e-mail)
[email protected]. 10.1021/ac0259919 CCC: $25.00 Published on Web 01/22/2003
© 2003 American Chemical Society
samples such as glass, minerals, metals, and plastic. Precision and accuracy of LA-ICPMS analysis of glass and minerals for the major components to the sub-ppm concentrations of ultratrace elements are established (ref 1 and references in there). The analysis of metal samples seems to be more difficult2-5 due to the different melting and boiling points of different components of an alloy, where the sampling time of the laser can have a larger influence on the ablation product. However, successful applications have been demonstrated for various metal samples.4,5 Among the different alloys, brass is one of the most difficult samples to analyze.6 The differences between melting (Cu, 1083 °C; Zn, 420 °C) and boiling points (Cu, 2567 °C; Zn, 907 °C) of copper and zinc are larger than between other metals such as chromium, nickel, and iron in common steel alloys. The melting points of brass alloys are very low, ranging from 1080 to 900 °C for up to 32% zinc.7 For alloys containing more zinc, a higher rate of ablation is obtained for nanosecond lasers.8 Calibration curves recorded for various brass alloys seem to be nonlinear for copper and zinc in ICP-AES9 and in ICPMS.6 Therefore, a non matrix matched calibration for the determination of brass composition has not been successful. Liquid nebulization calibration has been tried,10 but there was always an overestimation of zinc compared to copper. In laser ablation, the Cu/Zn ratio is influenced by several parameters such as pulse length and irradiance.7,11 Pulse energies significantly higher than the ablation threshold value for a given laser system (for example, higher than 1.8 J cm-2 for a Nd:YAG 266-nm laser with 6-ns pulse length11) result in an almost constant copper-to-zinc ratio in ICPMS and ICP-AES, independent of the crater diameters. With laser energies close to the ablation (1) Gu ¨nther, D.; Jackson, S. E.; Longerich, H. P. Spectrochim. Acta, Part B 1999, 54, 381-409. (2) Arrowsmith, P. Anal. Chem. 1987, 59, 1437-44. (3) Leach, A. M.; Hieftje, G. M. Anal. Chem. 2001, 73, 2959-67. (4) Mochizuki, T.; Sakashita, A.; Tsuji, T.; Iwata, H.; Ishibashi, Y.; Gunji, N. Anal. Sci. 1991, 7, 479-81. (5) Mao, X. L.; Chan, W. T.; Russo, R. E. Appl. Spectrosc. 1997, 51, 1047-54. (6) Borisov, O. V.; Mao, X. L.; Fernandez, A.; Caetano, M.; Russo, R. E. Spectrochim. Acta, Part B 1999, 54, 1351-65. (7) Margetic, V.; Niemax, K.; Hergenro¨der, R. Spectrochim. Acta Part B 2001, 56, 1003-10. (8) Salle, B.; Chaleard, C.; Detalle, V.; Lacour, J. L.; Mauchien, P.; Nouvellon, C.; Semerok, A. Appl. Surf. Sci. 1999, 139, 302-5. (9) Gagean, M.; Mermet, J. M. Spectrochim. Acta, Part B 1998, 53, 581-91. (10) Mao, X. L.; Russo, R. E. J. Anal. At. Spectrom. 1997, 12, 177-82. (11) Mao, X. L.; Ciocan, A. C.; Russo, R. E. Appl. Spectrosc. 1998, 52, 913-8.
Analytical Chemistry, Vol. 75, No. 4, February 15, 2003 747
Figure 1. (a) Volume distribution of MBH B26 in eight size fractions (Lasair 1001). The volume was calculated based on the assumption of spherical particles according to Figure 4. (b) Volume distribution of MBH B26 after the particle separator. The particle separator removed the particles in the two largest channels very efficient so that no particle was detected within the Lasair particle measurement system. The volume was calculated expecting spherical particles according to Figure 4. The effective cutoff point of the impactor was determined to be at 0.45 µm.
threshold, the sample is heated instead of being ablated and preferential vaporization of zinc takes place. The conclusion of the work carried out on brass samples is that, with several selected lasers, while using various energies and pulse lengths, the Cu/Zn ratio measured by the ICPMS is too low. Only one exception using a 35-ps laser and pulse energy of 5-40 GW cm-2 has been reported.11 Linear calibration graphs for zinc were obtained with laser-induced breakdown spectroscopy using copper as internal standard,7 indicating that preferential vaporization can be avoided using a femtosecond laser. The aim of this study was to investigate the sources of elemental fractionation of brass in LA-ICPMS. Except for some experiments using the direct laser-induced aerosol introduction into the ICPMS, most experiments involved the measurement of the aerosol composition. There are various techniques, such as collection of the aerosol on filters followed by the analysis of the digested material using solution nebulization ICPMS, scanning electron imaging, and electron probe X-ray analysis of the filtercollected material. In most studies, energies far above the ablation threshold have been used due to the fact that conditions close to threshold condition led to laser-induced elemental fractionation. Therefore, all laser fluencies used in this study were adjusted above 10 J cm-2 to avoid any of the reported effects. EXPERIMENTAL SECTION Instrumentation. A beam homogenized high-power Nd:YAG 266-nm laser ablation system previously described12 and a 193(12) Guillong, M.; Horn, I.; Gu ¨ nther, D. J. Anal. At. Spectrom. 2002, 17, 8-14.
748 Analytical Chemistry, Vol. 75, No. 4, February 15, 2003
Figure 2. (a) Schematic setup of the online analysis using the Aridus desolvating sample introduction system for standardization. (b) Scheme of the instrumental setup of all experiments.
nm ArF GeoLas M excimer laser ablation system (MicroLas, Go¨ttingen, Germany) described in ref 13 were used. In all experiments, the energy densities were adjusted to 14 J cm-2 for the 266-nm Nd:YAG laser and to 20 J cm-2 for the 193-nm ArF excimer laser. The pulse repetition rate was set to 4 Hz to minimize possible heating of the sample. A constant cell gas flow of 0.7 L min-1 for argon or helium was used for all aerosol filter experiments. Equally sized ablation cells were used for both laser systems at a volume of ∼30 cm3. An Elan 6000 ICPMS (Perkin-Elmer, Norwalk, CT) combined with the Nd:YAG 266-nm laser and an Elan 6100 DRC + ICPMS operating in normal mode (Perkin-Elmer) coupled to the 193-nm ArF excimer laser ablation system were used. ICPs and ablation cells were connected with a 1-m polyethylene tube. For calibration of the LA-ICPMS analysis, an Aridus desolvating sample introduction system (Cetac Technologies, Omaha, NB) was used as an alternative source for introducing liquid samples. The aerosol coming from the Aridus (1.2-1.4 L min-1 Ar) was mixed with the cell gas containing the laser-induced aerosol before entering the ICP as described elsewhere14 (Figure 2a). However, due to the combination of argon and helium, both systems (laser carrier, liquid carrier) were operated under optimum conditions. The argon flow was a combination of 0.9 L min-1 nebulizer gas (13) Gu ¨ nther, D.; Frischknecht, R.; Heinrich, C. A.; Kahlert, H. J. J. Anal. At. Spectrom. 1997, 12, 939-44. (14) Gu ¨ nther, D.; Cousin, H.; Magyar, B.; Leopold, I. J. Anal. At. Spectrom. 1997, 12, 165-70.
Table 1. Operating Conditions of the ICPMS Instruments for Liquid and Dry Aerosol Sample Introduction ICPMS Elan 6100 (laser) cell gas (He/Ar, L/min) makeup gas (Ar, L/min) Aridus neb gas (Ar, L/min) Aridus sweep gas (Ar, L/min) total flow from Aridus (L/min) autolens ICP rf power (W) auxiliary gas flow (L/min) plasma gas flow (L/min) detector mode oxide rate
Elan 6000 (laser)
0.7/0.7
1.05/1.0
0.8 3/2.5 (He/Ar cell gas) 1.3 on 1500 1 15 dual
