Anal. Chem. 2006, 78, 1883-1889
Direct Determination of Rare Earth Elements at the Subpicogram per Gram Level in Antarctic Ice by ICP-SFMS Using a Desolvation System Paolo Gabrielli,†,‡ Carlo Barbante,*,†,‡ Clara Turetta,† Alexandrine Marteel,‡,§,| Claude Boutron,⊥,§ Giulio Cozzi,†,‡ Warren Cairns,† Christophe Ferrari,§,# and Paolo Cescon†,‡
Institute for the Dynamics of Environmental Processes-CNR, University of Venice, Ca’ Foscari, I-30123 Venice, Italy, Department of Environmental Sciences, University of Venice, Ca’ Foscari, I-30123 Venice, Italy, Laboratoire de Glaciologie et Ge´ ophysique de l’Environnement du CNRS 54, rue Molie` re, B.P. 96, 38402 St. Martin d’Heres Cedex, France, Department of Geological Sciences, University of Siena, Via del Laterino 8, 53100 Siena, Italy, Observatoire des Sciences de l’Univers et Unite´ de Formation et de Recherche de Physique (Institut Universitaire de France), Universite´ Joseph Fourier, Domaine Universitaire, B.P. 68 38041 Grenoble, France, and EÅ cole Polytechnique Universitaire de Grenoble (Institut Universitaire de France), Universite´ Joseph Fourier, 28 avenue Benoıˆt Frachon, B.P. 53, 38041 Grenoble, France
A method, based on inductively coupled plasma sector field mass spectrometry coupled with a microflow nebulizer and a desolvation system, has been developed for the direct determination of rare earth elements (REE) (La, Ce, Pr, Nd, Sm, Eu, Gd, Tb, Dy, Ho, Er, Tm, Yb, Lu) down to the subpicogram per gram level (1 pg/g ) 10-12 g g-1) in ∼1 mL of molten Antarctic ice. Contamination problems were carefully taken into account by adopting ultraclean procedures during the sample pretreatment phases. The use of a desolvation system for sample introduction during the analysis greatly reduced spectral interferences from oxide formation; the residual interfering contributions were calculated and subtracted whenever necessary. A matched calibration curve method was used for the quantification of the analytes. Instrumental detection limits ranged from 0.001 pg/g for Ho, Tm, and Lu to 0.03 pg/g for Gd. The precision, in terms of relative standard deviation on 10 replicates, ranged from 2% for La, Ce, Pr, and Lu, up to 10% for Er, Tm, and Yb. This methodology allowed the direct determination of REE in a 1-mL sample of ancient Antarctic ice with concentration ranges between 0.006 and 0.4 pg/g for Tm and 0.9-60 pg/g for Ce.
reasons. First, REE are fractionated in the environment due to small but systematic differences in their chemical properties, resulting from their identical trivalent charge and the concomitant systematic decrease in ionic radius with increasing atomic number (the so-called lanthanide contraction). Second, REE show a conservative behavior in the environment, due primarily to their low solubility and relative immobility in the terrestrial crust. As they are mostly transported in the atmosphere in the particulate phase, the REE content of atmospheric particulate matter will reflect the characteristics of the original site of provenance,1 so they are an extraordinarily useful geochemical tool for source fingerprinting studies. The successively dated snow and ice layers that are deposited in permanent glaciers in polar regions have proven to be valuable archives for studying the biogeochemical cycles of trace elements in the Earth’s system.2-6 However, until now, the reconstruction of past atmospheric REE changes was restricted to an ombrotrophic peat bog profile.7 So far, REE determination has never been extensively attempted in glacial paleoclimatic archives such as polar ice cores, essentially because of the extremely low REE concentration and because of the limited volume of the samples. Previously, REE have been determined only in relatively high amounts of dried dust (∼2-4 mg) obtained from few extremely
Rare earth elements (REE) have been widely adopted as excellent proxies for several geochemical processes in cosmochemistry, igneous petrology, and sedimentology. Their importance as a robust and powerful tracer is essentially due to two
(1) Henderson, P., Ed. Rare Earth Element Geochemistry, 2nd ed.; Elsevier Science: Amsterdam, 1984. (2) Hong, S.; Boutron, C. F.; Gabrielli, P.; Barbante, C.; Ferrari, C.; Petit, J. R.; Lee, K.; Lipenkov, V. Y. Geophys. Res. Lett. 2004, 31. Doi: 10.1029/ 2004GL021075. (3) Gabrielli, P.; Barbante, C.; Plane, J. M. C.; Varga, A.; Hong, S.; Cozzi, G.; Gaspari, V.; Planchon, F.; Cairns, W.; Ferrari, C.; Crutzen, P.; Cescon, P.; Boutron, C. F. Nature 2004, 432, 1011-1014. (4) Vallelonga, P.; Gabrielli, P.; Rosman, K.; Barbante, C.; Boutron, C. F. Geophys. Res. Lett. 2005, 32. Doi: 10.1029/2004GL021449. (5) Gabrielli, P.; Planchon, F.; Hong, S.; Lee, K.; Hur, S. D.; Barbante, C.; Ferrari, C.; Petit, J. R.; Lipenkov, V. Y.; Cescon, P.; Boutron, C. F. Earth Planet. Sci. Lett. 2005, 234/1-2, 249-259. (6) Gabrielli, P.; Barbante, C.; Boutron, C. F.; Cozzi, G.; Gaspari, V.; Planchon, F.; Ferrari, C.; Cescon, P. Atmos. Environ. 2005, 39, 6420-6429. (7) Krachler, M.; Mohl, C.; Emons, H.; Shotyk, W. J. Environ. Monit. 2003, 5, 111-121.
* Corresponding author. Phone: +39-041-2348942. Fax: +39-041-2348549. E-mail:
[email protected]. † Institute for the Dynamics of Environmental Processes-CNR, University of Venice. ‡ Department of Environmental Sciences, University of Venice. § Laboratoire de Glaciologie et Ge ´ ophysique de l’Environnement du CNRS. | University of Siena. ⊥ Observatoire des Sciences de l’Univers et Unite ´ de Formation et de Recherche de Physique (Institut Universitaire de France). # E Å cole Polytechnique Universitaire de Grenoble (Institut Universitaire de France). 10.1021/ac0518957 CCC: $33.50 Published on Web 02/03/2006
© 2006 American Chemical Society
Analytical Chemistry, Vol. 78, No. 6, March 15, 2006 1883
large masses (∼8 kg) of polar ice.8 Only light REE isotopes, such as the isotopes of Nd, have been studied in polar ice cores, because of their higher mean crustal abundance and their consequently easier detection in dust particles conserved in ancient ice.8,9 Inductively coupled plasma mass spectrometry (ICPMS) is one of the most commonly used techniques for the determination of REE in natural waters because of the capability for a rapid multielement detection over a wide concentration range with relatively low detection limits. Until now, extensive REE determinations were commonly performed only on seawater, lakes, and riverine waters10-12 but not in ice cores. However, even in natural waters, REE determination was exceedingly difficult to achieve because of the complex sample matrix and the very low REE concentrations. Thus, REE preconcentration and separation from the matrix was required for the determination of REE by ICPMS in natural waters.13-15 Most of the previously developed analytical methods used high sample take-up rates (2-4 mL/min of analysis)16 or involved many sample manipulations, which may introduce contamination or produce the partial loss of the analytes. Previously, there has only been one attempt at determining REE concentrations in polar surface snow by ICPMS. In this case, a preconcentration step by evaporation was used in order to try to measure the naturally less abundant heavy REE,17 but the attempt in this case was largely unsuccessful. Analytical difficulties may have also arisen from the fact that, although snow is a relatively simple matrix, spectroscopic interferences can still occur because of the dust content, and thus, its elemental composition is magnified in preconcentrated samples.18 Major difficulties in the determination of REE in natural waters are caused by the interferences from Ba compounds formed in the plasma and the reciprocal oxide interferences that REE generate on each other. Here we present a direct ultrasensitive method for the determination of REE in less than 1 mL of molten ice by inductively coupled plasma sector field mass spectrometry (ICPSFMS) without the need of any kind of preconcentration. Spectroscopic interferences were completely eliminated or greatly reduced by using a desolvating sample introduction system. However, whenever necessary (e.g., in the case of Gd), a mathematical correction was used to resolve the excess of the signal caused by interferences. (8) Basile-Doelsch, I.; Grousset, F. E.; Revel-Rolland, M.; Petit, J. R.; Biscaye, P. E.; Barkov, N. I. Earth Planet. Sci. Lett. 1997, 146, 573-589. (9) Delmonte, B.; Basile-Doelsch, I.; Petit, J. R.; Maggi, V.; Revel-Rolland, M.; Michard, A.; Jagoutz, E.; Grousset, F. E. Earth-Sci. Rev. 2004, 66, 63-87. (10) Sotto Alibo, D.; Nozaki, Y. Geochim. Cosmochim. Acta 1999, 63, 363-372. (11) Nozaki, Y.; Lerche, D.; Sotto Alibo, D.; Snidvongs, A. Geochim. Cosmochim. Acta 2000, 64, 3983-3994. (12) De Carlo, E. H.; Green, W. J. Geochim. Cosmochim. Acta 2002, 66, 13231333. (13) Murty, D. S. R.; Chakrapani, G. J. Anal. At. Spectrom. 1996, 11, 815-820. (14) Halicz, L.; Gavrieli, I.; Dorfman, E. J. Anal. At. Spectrom. 1996, 11, 811814. (15) Zhang, T.-h.; Shan, X.-q.; Liu, R.-x.; Tang, H.-x.; Zhang, S.-z. Anal. Chem. 1998, 70, 3964-3968. (16) Yamasaky, S.-i.; Tsumura, A.; Takaku, Y. Microchem. J. 1994, 49, 305318. (17) Ikegawa, M.; Kimura, M.; Honda, K.; Akabane, I.; Makita, K.; Motoyama, H.; Fujii, Y.; Itokawa, Y. Atmos. Environ. 1999, 33, 1457-1467. (18) Gabrielli, P.; Varga, A.; Barbante, C.; Boutron, C. F.; Cozzi, G.; Gaspari, V.; Planchon, F.; Cairns, W.; Hong, S.; Ferrari, C.; Capodaglio, G. J. Anal. At. Spectrom. 2004, 19, 831-837.
1884
Analytical Chemistry, Vol. 78, No. 6, March 15, 2006
EXPERIMENTAL SECTION Reagents and Materials. At the Laboratory of Glaciology and Geophysics of the Environment (LGGE),19 a mixed bed of ionexchange resins (maximum flow rate was 2 L h-1) from Maxy (La Garde, France)20 was used to produce the ultrapure water used in the different steps of the method. At the Department of Environmental Sciences (DES)21 in Venice, the ultrapure water was produced by coupling a Milli-RO (Millipore, Bedford, MA) water system with a Purelab Ultra system (Elga, High Wycombe, U.K.). Chloroform (Merck, Darmstadt, Germany), and Suprapur grade HNO3 (65% Merck) were used in the initial phases of the washing method. Ultrapure double-distilled HNO3, produced at Curtin University of Technology (CUT) in Perth, Australia, was used in the final phases of the cleaning procedure, for sample acidification and standard preparation. Standards were prepared from a 10 mg L-1 REE multielement ICPMS stock solution (Claritas, Spex). The low-density polyethylene (LDPE) bottles used to contain the samples (from Nalgene Corp., Rochester, NY) and the stainless steel tools for the decontamination of the ice cores were washed on a clean bench (class 100) located in a clean room (class 10000) at the LGGE.20 Sample Collection and Processing. For this study, an electromechanical drill was used to obtain the ice cores, which are cylindrical samples, with a diameter of ∼10 cm, of the deep ice layers in the glaciers. The ice core sections were retrieved from Antarctica at Dome C (75°06′ S; 123°21′ E, 3233 m, mean annual temperature -54 °C) within the framework of the European Project for Ice Coring in Antarctica (EPICA).22 Only a part of the cross section (∼30%) was available because, in the EPICA program, a given core section (typically 55 cm long) is cut longitudinally into several parts, which are then used for different kinds of measurements. A drawback of deep electromechanical drilling is that ice cores are contaminated on their outside by several heavy metals originating from the wall retaining fluid, which is used in the field to avoid premature closure of the drilling borehole. The contamination of the samples, which may also derive from handling of the core, was carefully evaluated. Following this, a decontamination procedure, successfully adopted for other trace elements,18,23 was performed by chiseling the external ice core layers under a class 100 laminar flow clean bench located in the cold laboratory (t ) -15 °C) at the LGGE. The obtained uncontaminated inner core was cut into two parts and placed in two separate ultraclean 1-L LDPE bottles. The ice samples were then melted at room temperature under a class 100 clean bench, and an aliquot of ∼5 mL was transferred to a 15-mL ultraclean LDPE bottle and kept frozen until analysis. Instrumentation. The analytical measurements were conducted at DES by ICP-SFMS (Element2, Thermo Finnigan MAT, Bremen, Germany). The instrumental conditions and measure(19) Ferrari, C.; Moreau, A. L.; Boutron, C. F. Fresenius J. Anal. Chem. 2000, 366, 433-437. (20) Boutron, C. F. Fresenius J. Anal. Chem. 1990, 337, 482-491. (21) Barbante, C.; Cozzi, G.; Capodaglio, G.; Van de Velde, K.; Ferrari, C.; Veysseyre, A.; Boutron, C. F.; Scarponi, G.; Cescon, P. Anal. Chem. 1999, 71, 4125-4133. (22) EPICA community members. Nature 2004, 429, 623-628. (23) Candelone, J. P.; Hong, S.; Boutron, C. F. Anal. Chim. Acta 1994, 299, 9-16.
Table 1. Instrumental Conditions and Measurement Parameters for the ICP-SFMS and the Desolvation Unit forward power (W) gas flow rates Cool (L min-1) Auxiliary (L min-1) Nebulizer (L min-1) Sweep gas (L min-1) Nitrogen flow rate (mL min-1) Membrane temperature (°C) spray chamber temperature (°C) sample uptake (µL min-1) acquisition mode selected isotopes
resolution adopted no. of scans dwell time per acquisition point (ms) no. of acquisition points per mass segment (sample per peak) acquisition window (%) search window (%) integration window (%) a
1250 15.50 1.80 0.8-1.1a 3.40-4.15a 1-18a 175 95 100 E-scan: over a small mass range 139La, 140Ce, 141Pr, 144Nd, 151Eu, 152Sm, 155Gd, 159Tb, 164Dy, 165Ho, 166Er, 169Tm, 174Yb, 175Lu low (m∆m-1) ≈ 400 50 10 30 100 100 70
Optimized to obtain maximum signal intensity.
ment parameters are reported in Table 1. The instrument is equipped with a class 100 clean bench as a clean sample introduction area. The samples were melted under this clean bench and were carefully handled with disposable powderless gloves and suitable clean room clothing. To maximize ion transmission, the low-resolution mode (LRM) with a nominal resolution setting of m∆m-1 ≈ 400 was used to determine the REE. To minimize spectral interferences, a microflow nebulization/desolvation sample introduction system24 (Aridus, Cetac Technologies, Omaha, NE) was used. This device is composed of a microflow (
