Phase and Composition Changes of Titanite during Laser Ablation

Apr 27, 2010 - Hana Nováková , Markéta Holá , Michal Vojtíšek-Lom , Jakub Ondráček , Viktor Kanický. Spectrochimica Acta Part B: Atomic Spect...
0 downloads 0 Views 4MB Size
Download PDF
Anal. Chem. 2010, 82, 4272–4277

Phase and Composition Changes of Titanite during Laser Ablation Inductively Coupled Plasma Mass Spectrometry Analysis Daniel Fliegel,*,† Mariana Klementova,‡ and Jan Kosler† Centre for Geobiology, University of Bergen, 5007 Bergen, Norway, and Institute of Inorganic Chemistry of the ASCR, ˇ ezˇ 1001, Czech Republic v.v.i., 250 68 Husinec-R Changes in phase and chemical composition of the mineral titanite (monoclinic CaTiSiO5) during laser ablation and plasma-aerosol interaction were investigated using electron diffraction and electron microbeam X-ray analysis with transmission electron microscopy (TEM) and scanning electron microscopy (SEM). Laser ablation of titanite with a solid state 213 nm nanosecond laser generates a bimodal aerosol consisting of condensed nanoparticles and spherical micrometersized particles. The two types of particles produced by laser ablation are amorphous on the scale resolvable by the electron diffraction. The ratio of Ca-Si-Ti does not change significantly during laser ablation. Aerosol of titanite particles introduced into the ICP and collected after interaction with the plasma contains nanometer-sized particles of a condensate and spherical micrometer-sized particles with a molten surface. The condensed particles are enriched in silicon whereas the spherical micrometer-sized particles show a deficiency in Si relative to the titanite composition. During laser ablation-inductively coupled plasma mass spectrometry (LA-ICPMS) analysis of titanite, Si and Ti showed positive and negative fractionation trends relative to Ca, respectively. This is consistent with the observed chemical composition changes of the titanite aerosol within the ICP. This study links for the first time the chemical and phase changes of a sample within the ICP to the elemental fractionation during LA-ICPMS. Laser ablation inductively coupled plasma mass spectrometry (LA-ICPMS) is a sensitive and selective technique for analysis of major, minor, and trace elements in various sample matrixes. The effective spatial resolution of laser ablation is as good as a few micrometers and the depth resolution per laser pulse is at least 0.1 µm. The technique has been used widely for elemental and isotopic analysis of geological samples,1 biological samples,2 steels, * Corresponding author. E-mail: [email protected]. † University of Bergen. ‡ Institute of Inorganic Chemistry of the ASCR. (1) Sylvester, P. J. Geostand. Geoanal. Res. 2005, 29, 41–52. (2) Becker, J. S.; Zoriy, M.; Becker, J. S.; Dobrowolska, J.; Matusch, A. J. Anal. At. Spectrom. 2007, 22, 736–744.

4272

Analytical Chemistry, Vol. 82, No. 10, May 15, 2010

and alloys.3 In addition to determination of the concentration of analytes in various sample matrixes, in situ radiometric dating of natural minerals4,5 presents another important application field. Zircon is the most commonly used mineral phase in LA-ICPMS geochronology6 but other minerals, such as titanite and monazite,4 have also been dated using this technique. Absolute quantification in LA-ICPMS is often hampered by matrix and fractionation effects. It has been shown previously that the laser ablation process,7 transport, and ionization of ablated aerosol in the ICP8 can lead to nonaccurate analytical results. Matrix matched standards with well-known concentration of internal standard element are therefore needed for precise and accurate quantification by LA-ICPMS. The origin and nature of fractionation effects have been discussed in detail since the pioneering studies of A. Gray,9 B. J. Fryer,10,11 and D. Figg.12 Their work has indicated that laser ablation process induces a thermal stress on the sample which may lead to preferential release of volatile elements in the aerosol. Kosler et al. described phase decomposition during laser ablation leading to elemental fractionation.13 It has been also shown that laser wavelength and pulse duration are crucial parameters for stoichiometric ablation.7 This has led to laser developments toward deep UV wavelength and pulse duration at the femtosecond level.14 The aerosol transport and atomization/ionization in the ICP have also been shown to play important role in the fractionation process. Using emission traces of YO particles ablated from a Y2O3 pellet, Aeschliman et al. have shown that particulate matter is not (3) Fisher, A. S.; Goodall, P. S.; Hinds, M. W.; Penny, D. M. J. Anal. At. Spectrom. 2005, 20, 1398–1424. (4) Simonetti, A.; Heaman, L. M.; Chacko, T.; Banerjee, N. R. Int. J. Mass Spectrom. 2006, 253, 87–97. (5) Kosler, J. Proc. Geologists’ Assoc. 2007, 118, 19–24. (6) Slama, J.; Kosler, J.; Condon, D. J.; Crowley, J. L.; Gerdes, A.; Hanchar, J. M.; Horstwood, M. S. A.; Morris, G. A.; Nasdala, L.; Norberg, N.; Schaltegger, U.; Schoene, B.; Tubrett, M. N.; Whitehouse, M. J. Chem. Geol. 2008, 249, 1–35. (7) Gu ¨ nther, D.; Heinrich, C. A. J. Anal. At. Spectrom. 1999, 14, 1369–1374. (8) Kuhn, H. R.; Guillong, M.; Gu ¨ nther, D. Anal. Bioanal. Chem. 2004, 378, 1069–1074. (9) Gray, A. L. Analyst 1985, 110, 551–556. (10) Fryer, B. J.; Jackson, S. E.; Longerich, H. P. Chem. Geol. 1993, 109, 1–8. (11) Fryer, B. J.; Jackson, S. E.; Longerich, H. P. Can. Mineral. 1995, 33, 303– 312. (12) Figg, D. J.; Cross, J. B.; Brink, C. Appl. Surf. Sci. 1998, 129, 287–291. (13) Kosler, J.; Wiedenbeck, M.; Wirth, R.; Hovorka, J.; Sylvester, P.; Mikova, J. J. Anal. At. Spectrom. 2005, 20, 402–409. (14) Fernandez, B.; Claverie, F.; Pecheyran, C.; Donard, O. F. X. TrAC, Trends Anal. Chem. 2007, 26, 951–966. 10.1021/ac902284y  2010 American Chemical Society Published on Web 04/27/2010

completely vaporized in an ICP.15 Several studies have outlined the importance of the ICP for elemental fractionation. Preferential vaporization in combination with diffusion processes in the ICP have been proposed as sources of elemental fractionation based on the measured element signal ratios.8 Also, cooling of the ICP by high sample load has been proposed to be a cause for nonstoichiometric ionization in the ICP.16 So far, the evidence for elemental fractionation in the ICP has been almost exclusively based on indirect evidence from mass spectrometry measurements, and no data were available to document the phase and chemical changes of sample aerosol during the interaction with the ICP. The phase change of zircon during ablation and plasma-sample interaction has been investigated with the conclusion of a major influence of the laser ablation process on the fractionation. A better understanding of sample phase changes due to interaction with the ICP would be another important step toward a nonmatrix matched calibration strategy. This study thus investigates phase changes of titanite (monoclinic CaTiSiO5) during laser ablation to evaluate the mineralogical and crystallographic properties of the ablated aerosol. Further, the interaction of titanite aerosol with the ICP was studied to evaluate the phase and chemical changes during sample-plasma interaction. These measurements were then combined with observed fractionation effects during LA-ICPMS. Titanite was chosen for this study because (a) it is a geologically important mineral for U-Pb dating4,17,18 and any Pb-U fractionation effects during LA-ICPMS analysis might lead to inaccurate age determination and (b) due to its crystalline structure and potential phase transformation at higher temperatures, it is a suitable material for a study of laser- and ICP-induced phase changes. EXPERIMENTAL SECTION Laser ablation experiments were conducted using a commercially available 213 nm Nd:YAG laser ablation system (UP213 New Wave Research) at the University of Bergen. Applied fluence was as high as 4 J/cm2, and ablation was done using a 95 µm spot size, linear raster ablation (10 µm/s), and 10 Hz repetition rate. The laser raster mode was used to ensure a high rate of sample ablation and stable particle size distribution during the ablation experiment. The samples were placed in a custommade cylindrical ablation cell (dimensions, i.d. 58 mm; height 23 mm). The ablation gas was 1 L/min He, and the aerosol was collected on transmission electron microscopy (TEM) grids for subsequent mineralogical and chemical characterization. For the sample-plasma interaction part of the experiment, a fine milled titanite powder (grain size 1-10 µm) was introduced in a 1 L/min Ar carrier gas stream into the ICP (Iris, Thermo Scientific). Particles were collected past the ICP using an in house built water cooled interface (Figure 1) with a commercially available Ni sampler cone having an orifice diameter of 0.6 mm. (15) Aeschliman, D. B.; Bajic, S. J.; Baldwin, D. P.; Houk, R. S. J. J. Anal. At. Spectrom. 2003, 18, 1008–1014. (16) Kroslakova, I.; Gu ¨ nther, D. J. Anal. At. Spectrom. 2007, 22, 51–62. (17) Banerjee, N. R.; Simonetti, A.; Furnes, H.; Muehlenbachs, K.; Staudigel, H.; Heaman, L.; Van Kranendonk, M. J. Geology 2007, 35, 487–490. (18) Frost, B. R.; Chamberlain, K. R.; Schumacher, J. C. Chem. Geol. 2001, 172, 131–148.

Figure 1. Experimental setup for particle collection from the ICP.

The interface was positioned 25 mm above the torch. The aerosol was extracted from the ICP by a water jet vacuum pump attached to the interface and collected on a TEM grid (Cu, 3 mm diameter, 200 mesh, form ware coated) that was positioned in the extraction line. Minimum pressure within the extraction region did not exceed 60 mbar. This extraction pressure differs from pressure within the interface region of the ICP. However, the processes within the ICP under atmospheric pressure, such as melting and subsequent particle formation, should not be affected by this fact. The titanite powder was prepared from a natural titanite crystal (Gruber anorthosite, Otto v. Gruber Mountains, central Dronning Maud Land, Antarctica).19 The major element composition of the titanite determined by electron microprobe analysis was Al2O3 2.4 wt %, SiO2 29.3 wt %, CaO 27.9 wt %, TiO2 32.6 wt %, and FeO 2.6 wt %. TEM was carried out on a JEOL JEM-3010 microscope in the Institute of Inorganic Chemistry of the ASCR operated at 300 kV (LaB6 cathode, point resolution 1.7 Å) with an Oxford Instruments Energy Dispersive X-ray (EDX) detector attached. Images were recorded on a CCD camera with resolution 1024 × 1024 pixels using the Digital Micrograph software package. EDX analyses were acquired and treated by the INCA software package. The major element compositions of the bulk titanite, titanite powder, and the particles that passed the ICP were determined with a SEM-EDX (Zeiss supra 55VP SEM and Thermo Noran EDS) at the University of Bergen. The EDX analysis was done at 15 mm working distance with 15 keV accelerating voltage and an aperture size of 60 µm, whereas for the imaging part 10 µm aperture was used. All presented data for Ca, Si, and Ti were normalized to 100 atom %. Peaks resulting from carbon coating, the Al sample holder, or the copper TEM grid were omitted in the calculation. The fractionation index was calculated from single laser spot analyses (rather than laser rastering). This was done to enhance the fractionation effects because of the changing particle size distribution during single spot analysis. The 213 nm laser was coupled to an Element 2 (Thermo Scientific) single collector sector (19) Jacobs, J.; Fanning, C. M.; Henjes-Kunst, F.; Olesch, M.; Paech, H. J. J. Geol. 1998, 106, 385–406.

Analytical Chemistry, Vol. 82, No. 10, May 15, 2010

4273

field ICPMS at the University of Bergen. All measurements were done using He as the sample carrier gas mixed with Ar at the back end of the ICP torch. All isotopes were measured in low mass resolution using 1 point per mass peak and a dwell time of 10 ms. RESULTS AND DISCUSSION Morphology of the Aerosol Particles. Collection of aerosol particles produced by LA-ICPMS within the ICP interface region is difficult due to the low sample volume introduced and efficient (although not complete) digestion of even the large particles within the ICP. Accordingly, any model experiments have to be conducted with mass load into the plasma that is higher compared to normal LA-ICPMS analytical conditions. For the investigation of the phase and chemical changes during ICP-sample interaction, a powdered titanite was introduced into the ICP using ultrasonic nebulization. Plasma mass load introduced by this method was higher compared to laser ablation. However, reactions such as atomization, vaporization, and ionization of a sample mainly depend on the sample properties and only to a lesser extent on the plasma mass load. The powder nebulization used in the present study is therefore only valid as a qualitative model for major element composition of the aerosol introduced by LAICPMS. The powdered aerosol was generated by milling natural titanite in a ball mill. Subsequently the powder was suspended in water and the suspension was repeatedly decanted to separate fine and large particles. The suspension was then dried, resulting in a powder with particle size distribution in the range of 0.5-10 µm. This particle size overlaps with the size of aerosol particles produced by nanosecond laser ablation in moderate UV.22 No larger particles in the powder were observed using SEM. Because of the milling process, the particles were angular, rather than spherical. The aerosol collected on TEM grids shows a bimodal particle size distribution (Figure 2b). The aerosol consists of amorphous spherical particles 0.5-2 µm in diameter and a net structure of fine condensed particles. The primary particles of the net structure are below 10 nm in size. This observation is consistent with previously reported laser ablation experiments of various matrixes, such as glasses, metals, powders and alloys.20-22 This suggests that the experimental setup is capable of producing an aerosol comparable to laser ablation aerosols previously described in the literature. Phase and Chemical Composition of Ablated Aerosol. Electron diffraction study of ablated particles shows for both the spherical micrometer-sized particles and the net structure of nanometer-sized particles a diffuse diffraction pattern. Such a diffraction pattern (Figure 2C-II) and patterns on the highresolution micrograph (Figure 2C-III) are characteristic of amorphous material. In contrast, the diffraction pattern (Figure 2A-II) and the high-resolution micrograph (Figure 2A-III) of the bulk titanite prior to the ablation show distinct diffraction maxima. Accordingly, the phase change from crystalline titanite to amorphous aerosol must take place as a result of the ablation process. Not only the condensed nanometer-sized particles, which are thought to form by condensation of the vapor phase, are (20) Hergenroder, R. J. Anal. At. Spectrom. 2006, 21, 1016–26. (21) Hola, M.; Konecna, V.; Mikuska, P.; Kaiser, J.; Palenikova, K.; Prusa, S.; Hanzlikova, R.; Kanicky, V. J. Anal. At. Spectrom. 2008, 23, 1341–49. (22) Kuhn, H. R.; Gu ¨ nther, D. J. Anal. At. Spectrom. 2004, 19, 1158–64.

4274

Analytical Chemistry, Vol. 82, No. 10, May 15, 2010

Figure 2. TEM observations: (A) crushed titanite sample, part A-II shows an electron diffraction pattern and part A-III shows a highresolution TEM microcrograph consistent with titanite crystalline structure. (B) Laser ablated titanite with bimodal aerosol structure, parts B-II and B-III indicate an amorphous structure of CaTiSiO5 particles generated by laser ablation; the large spherical particle is coated by condensate of nanometer-sized particles, indicated by arrows. (C) Particle of CaTiSiO5collected past the ICP. Parts C-II and C-III show the amorphous structure of the titanite condensates; condensed nanometer-sized particles on the surface of a large spherical particle are shown in part C-IV.

amorphous, but so are the spherical particles. This supports the mechanism of ejecting the particles out of a melted phase in the crater. Ejection of particles from the laser crater due to Raleigh-Taylor instabilities has been described based on modeling and shadowgraphy experiments.23,24 The TEM-EDX analysis shows that the major element chemical composition of the particles does not change significantly during the laser ablation (and aerosol transport) process (Table 1). (23) Zhang, X.; Chu, S. S.; Ho, J. R.; Grigoropoulos, C. P. Appl. Phys. A: Mater. 1997, 64, 545–52. (24) Callies, G.; Berger, P.; Hugel, H. J. Phys. D: Appl. Phys. 1995, 28, 794– 806.

Table 1. Composition of Various Particle Phasesa phase

Si

Ca

Ti

initial titanite as bulkb initial titanite as aerosolb LA particles (spheres)b ICP particles (spheres)b ICP particles (condensate)b ICP particles (condensates on cone)b bulkc LA particles (spheres)c LA particles (condensates)c

37.1 ± 0.9 37.3 ± 6.4 38.4 ± 2.5 20.3 ± 13.6 64.7 ± 20.0 53.6 ± 6.1 42.1 ± 0.9 40.3 ± 0.4 61.7 ± 5.7

33.8 ± 0.9 35.7 ± 4.1 33.1 ± 1.7 44.1 ± 7.9 18.2 ± 14.6 28.5 ± 3.5 31.5 ± 1.3 33.7 ± 0.3 21.7 ± 2.3

29.2 ± 1.3 27.0 ± 4.0 28.5 ± 1.5 35.6 ± 6.8 17.2 ± 14.4 17.9 ± 3.4 26.1 ± 1.6 26.0 ± 0.1 16.6 ± 3.4

ICP particles (spheres)c ICP particles (condensate)c

24.4 ± 16.97 68.7 ± 1.48

42.8 ± 9.3 12.7 ± 0.5

32.9 ± 7.9 18.5 ± 1.0

a Data are normalized to 100% of Si + Ca + Ti, minor elements were not included in the analysis. LA-particles represent particles generated by laser ablation, ICP particles are particles which are sampled within/after the ICP. b SEM-EDX (atom % ± 1σ). c TEM-EDX (atom % ± 1σ).

Figure 3. (A) Titanite aerosol introduced into the ICP and (B) particles collected past the ICP on a TEM grid. Two different types of particles (condensates and spherical) are visible in part B; the surface of the spherical particle is covered by a condensate formed by nanometer-sized particles. Part C represents magnification of the dotted square in part A; the scale bar is 5 µm.

SEM Observations of Titanite Particles after Interaction with ICP. The effects of titanite particle interaction with the ICP were studied on particles deposited on the sampler cone and also on the aerosol extracted directly from the ICP and deposited on TEM grids. Both samples, the deposited aerosol on the cone as well as the aerosol collected from the ICP, showed a bimodal aerosol structure. Particulate matter that formed small nanometerscale spheres was agglomerated mainly on the TEM grids (i.e., extracted from the ICP) in net structures. These particles most likely formed as condensates out of a vapor phase since neither a netlike structure nor small spherical particles were observed in the aerosol introduced into the ICP. Also present on the TEM grid were spherical micrometer-sized particles. Larger particles with indication of continuous vapor deposition on their surface and melt aggregation were present as a deposit on the sampler cone (Figure 3). The primary aerosol structure introduced into the ICP contained micrometer-sized particles with sharp edges from the crushing process (Figures 3 and 2A-I). The surface of the spherical particles observed on the sampler cone is covered with the condensed phase. It appears that the phase which covers the sampler cone as condensed matter also covers the spherical particles. Spherical particles that interacted with the plasma and were deposited on the TEM grid show much less condensed matter on the surface. This is best explained by continuous deposition of vapor on the sampler cone during the sample introduction. Deposited particles can thus be coated by vaporized material as long as there is sample introduced into the ICP. In

contrast, particles extracted from the ICP via the sampler orifice had much less time to interact with the vapor phase since they were extracted immediately after the ICP, post their passage. Difference in the size and shape between particles deposited on the cone and on the TEM grid is also evident (Figure 2). Particles deposited on the sampler cone are significantly larger and irregularly shaped which might be due to the solidification-melting processes on the cone due to continuous and prolonged interaction with the ICP. On the contrary, particles deposited on the TEM grid were quenched in the gas stream and were not affected by the ICP once they escaped through the orifice of the sampler cone. Condensates past the ICP are amorphous as shown by the diffuse electron diffraction pattern (Figure 2C-II). Chemical Composition of Titanite Particles after Interaction with ICP. Condensed particles collected on the TEM grid, spherical particles on the TEM grid, and condensed matter on the sampler cone were analyzed using SEM-EDX and TEMEDX. The ratios of Ti-Ca-Si were evaluated for all these fractions, and the ratios were plotted in a three component Ti-Ca-Si diagram (Figure 4) together with the bulk particle composition, composition of the powder introduced in the ICP, and the composition of laser ablated aerosol. The data suggest that the major element composition of the sample does not change during the sample milling or laser ablation process. In contrast, the two aerosol fractions collected after their interaction with ICP (condensed nanometer-sized fraction and spherical micrometersized particles) show a significant change in the Ti-Ca-Si ratios. Notably, the large spherical particles are depleted in Si (Table 1 and Figure 5). In contrast, the agglomerated nanoparticles are enriched in Si and depleted in Ca and Ti. This experiment shows, for the first time, sample decomposition by ICP based on measurement of major element composition of solid particles surviving the thermal effects of the plasma. The condensate on the sampler cone also shows a similar enrichment in Si and depletion in Ca and Ti relative to the bulk titanite composition. Spherical particles deposited on the cone did not show depletion in Si. However, we observed condensation on the spherical particles in contrast to spherical particles deposited on the TEM grids. This is due to a continuous aerosol deposition. Therefore, the elemental analysis of spherical particles is a mixture between the composition of the particle itself and the condensates on the particle surface. Analytical Chemistry, Vol. 82, No. 10, May 15, 2010

4275

Figure 6. Fractionation index calculated for a single spot laser ablation analyses (n ) 3) of titanite (uncertainties are shown as 1σ).

Figure 4. Major element composition of primary titanite and particles before and after laser ablation and before and after the interaction with ICP. Particles collected past the ICP consist of two fractions: depleted in Si (circles) and enriched in Si (crosses and diamonds). Analyses were done by SEM-EDX; the mean values for particle composition are shown by large symbols and their compositions are given in Table 1.

Figure 5. Si and Ti composition (in percentages) of various particles relative to Ca. (1) refers to SEM-EDX analyzed particles, whereas (2) refers to TEM-EDX analyzed particles.

Fractionation Index. To link our observations of the change in chemical composition of titanite particles during plasma-sample interaction to a typical laser ablation ICPMS experiment, we analyzed the same titanite by LA-ICPMS using single spot ablation (Figure 6). The fractionation index for this analysis was calculated as proposed by Fryer et al.11 relative to Ca. Silicon shows enrichment (fractionation index > 1) and titanium shows a slight depletion relative to Ca during the LA-ICPMS analysis. This is consistent with the observed change in the chemical composition of titanite particles in the ICP. In addition, Pb and REEs in titanite 4276

Analytical Chemistry, Vol. 82, No. 10, May 15, 2010

both have their fractionation indexes above 1 whereas U and Th show fractionation below 1. DISCUSSION This study suggests a correlation between the effects of sample-plasma interaction and the element fractionation index in LA-ICPMS analysis of titanite. The fractionation index larger than 1 of silicon relative to Ca during LA-ICPMS and silicon enrichment in the fine condensed phase collected after the powdered aerosol passed through the ICP points to decomposition of titanite and preferential vaporization of silicon. A corresponding opposite trend for Ti resulted in the fractionation index below 1 for Ti relative to Ca during LA-ICPMS and enrichment in Ca and Ti in the powdered aerosol that passed through the ICP. Preferential vaporization of certain elements during LA-ICPMS analysis has been previously suggested in the literature.8 This study provides an evidence for preferential vaporization based on investigation of particles collected directly from the ICP and links the observed chemical and phase changes to the sample-plasma interaction. The fractionation behavior of Pb relative to Ca during LAICPMS and its affinity to Si can be compared with similar observations of Pb during ionization of the silicate matrix in thermal ionization mass spectrometry (TIMS). In TIMS analysis, the stable atomization and ionization of Pb can usually be achieved by loading the Pb sample on the evaporation filament together with silica gel25 because Pb is readily included in the silicate glass forming on the filament. In the present study, the preferential vaporization of silicon during melting of the titanite within the ICP is accompanied by preferential Pb evaporation. Consequently, Pb should be enriched in the condensed Si rich phase relative to its content in the aerosol that enters the ICP. Unfortunately, the present day analytical methods lack the sensitivity needed to determine trace element concentrations in single small objects such as the nanoparticles. Previously published phase equilibria in the Ca-Si-Ti oxide systems studied by high-temperature quenching experiments indicate a formation of CaTiO3 (perovskite) and a liquid phase (25) Cameron, A. E.; Smith, D. H.; Walker, R. L. Anal. Chem. 1969, 41, 525.

above 1380 °C.26 Additionally, studies of titanite as a glass matrix for storing nuclear waste also suggest decomposition of titanite to SiO2 and a perovskite phase (CaTiO3) during heating. During this phase separation, U from the dissolved nuclear fuel partitioned primary into the Ca-Ti rich phase in laboratory experiments.27 This observation is reflected by the fractionation index below 1 for U and Th (relative to Ca) during LA-ICPMS of titanite. The analogy of effects described in the literature (for TIMS Pb measurement and high-temperature phase and chemical changes of titanite) with the results obtained in this study suggests that phase separation during the interaction of laser ablation aerosol with the high-temperature ICP might be the cause for observed elemental fractionation during LA-ICPMS analysis of titanite. It is suggested that the elemental fractionation in LAICPMS may be controlled by phase separation of the ablated mineral phase and the partition coefficients of the analyte(s) in the newly formed mineral phases. These results might open a pathway to predicting the elemental fractionation effects during LA-ICPMS analysis from expected phase change reactions at high temperatures (above 5000 K) in the ICP and the crystallo-chemical properties of the newly forming phases. It should be also stressed that the laser ablation process changes the crystallographic properties of the sample and for certain matrixes it can even change the stoichiometric composition, as it has been previously shown for zircon ablation.13 The aerosol is mostly formed by amorphous phases produced during the laser ablation process. This can have implications for the aerosol properties during its interaction with the ICP. For example, the specific melting/ vaporization heat will be different for crystalline and amorphous phases of the same stoichiometric composition. This effect has

to be considered in modeling and prediction of sample-plasma interactions.

(26) Devries, R. C.; Roy, R.; Osborn, E. F. J. Am. Ceram. Soc. 1955, 38, 158– 71. (27) Hayward, P. J.; Vance, E. R.; Cann, C. D.; Doern, D. C. J. Am. Ceram. Soc. 1989, 72, 579–86.

Received for review October 9, 2009. Accepted April 9, 2010.

CONCLUSIONS This study links the fractionation effects observed in LA-ICPMS of titanite to chemical composition and sample phase changes in the ICP during sample-plasma interaction, and it is an important step toward better understanding the elemental fractionation processes during LA-ICPMS analysis. Results of this experiment can explain some differences in sensitivities of individual elements and their fractionation behavior during ablation of different sample matrixes using identical internal standards. The identification of the main source of elemental fractionation during titanite-plasma interaction is based on the nonstoichiometric vaporization processes in the ICP and it clearly outlines the importance of knowing/predicting the sample phase changes at elevated temperatures within the ICP. The data presented in this study suggest that, for the parameters used in the experiment, the ICP is unable to stoichiometrically digest titanite particles. Accordingly, a LAICPMS analysis that relies on nonmatrix matched calibration may suffer from differences in digestion efficiencies of different aerosol particles in the ICP. Corresponding chemical and isotopic composition changes during sample-plasma interaction are to be expected for minor and trace elements, and studies are under way to evaluate this process. ACKNOWLEDGMENT The funding from the Norwegian research council (Project No. 17063) is acknowledged. The authors wish to thank Josef Kusior and Ole Tumyr for the support during the construction of the interface and during the measurements and Joachim Jacobs for providing the titanite sample.

AC902284Y

Analytical Chemistry, Vol. 82, No. 10, May 15, 2010

4277