Anal. Chem. 2003, 75, 1972-1977
A 10-Fold Improvement in the Precision of Boron Isotopic Analysis by Negative Thermal Ionization Mass Spectrometry Jason Jiun-San Shen*,† and Chen-Feng You‡
Institute of Earth Sciences, Academia Sinica, P.O. Box 1-55, Nankang, Taipei, Taiwan, R.O.C. and Department of Earth Sciences, National Cheng Kung University, Tainan, Taiwan, R.O.C.
Boron isotopes are potentially very important to cosmochemistry, geochemistry, and paleoceanography. However, the application has been hampered by the large sample required for positive thermal ionization mass spectrometry (PTIMS), and high mass fractionation for negative-TIMS (NTIMS). Running as BO2-, NTIMS is very sensitive and requires only nanogram sized samples, but it has rather poor precision (∼0.7-2.0 ‰) as a result of the larger mass fractionation associated with the relatively light ion. In contrast, running as the much heavier molecule of Cs2BO2+, PTIMS usually achieves better precision around 0.1-0.4 ‰. Moreover, there is a consistent 10 ‰ offset in the 11B/10B ratio for NIST SRM 951 standard boric acid between the NTIMS and the certified value, but the cause of this offset is unclear. In this paper, we have adapted a technique we developed earlier to measure the 138La/139La using LaO+ 1 to improve the NTIMS technique for BO2. We were able to correct for instrumental fractionation by measuring BO2species not only at masses of 42 and 43, but also at 45, which enabled us to normalize 45BO2/43BO2 to an empirical 18O/16O value. We found that both I45/I42 ) (11B16O18O/10B16O16O) and (I43/I42)C ) (11B16O16O/ 10B16O16O) vary linearly with (I /I ) 45 43 C × 0.5 ) (11B16O18O/11B16O16O) × 0.5 ) 18O/16O. In addition, different activators and different chemical forms of B yield different slopes for the fractionation lines. After normalizing 11B16O18O/11B16O16O × 0.5 to a fixed 18O/16O value, we obtained a mean 11B/10B value of NIST SRM 951 that matches the NIST certified value at 4.0430 ( 0.0015 ((0.36‰, n ) 11). As a result, our technique can achieve precision and accuracy comparable to that of PTIMS with only 1‰ of the sample required. This new NTIMS technique for B isotopes is critical to the studies of early solids in the solar system and individual foraminifera in sediments that require both high sensitivity and precision. Natural boron consists of two stable isotopes, 10B and 11B, with abundances of 19.9 and 80.1%, respectively. The high mobility of boron, together with the large relative mass difference between † ‡
Academia Sinica. National Cheng Kung University.
1972 Analytical Chemistry, Vol. 75, No. 9, May 1, 2003
the two isotopes, leads to significant 11B/10B fractionation in nature of more than 100 ‰.2 Although B has been used as a tracer in a wide variety of geochemical and oceanographic studies (e.g., ore deposit, industrial pollution, water/rock interaction, slab-derived fluids in the subduction zones, paleo-SSTs, and pH), it is still difficult to measure micrometer-sized samples because of numerous analytical difficulties. For example, benthic forams are rare in the sediment cores, and it is necessary to examine the B isotope for different parts of the shell for individual benthic-forams in order to understand the relationship between the pH in the oceans and the CO2 concentration in the atmosphere.3,4 Moreover, excess 10B was found to be correlated with the Be content in various minerals of calcium-aluminum-rich inclusions (CAI) in chondrites using secondary ionization mass spectrometry (SIMS) and which was thought to have reflected the presence of a now-extinct radionuclide, 10Be (t1/2 ∼ 1.5 My), in the early solar system.5 These data were in situ SIMS measurements, however, that are rather imprecise because of the small overall signals (inherent to SIMS), and as a consequence, only minerals with high Be/B ratios, thus with large excess 10B, were studied. It is, therefore, necessary to improve the B isotopic analysis not only for sensitivity but also for precision. Two distinct thermal ionization mass spectrometry methods, positive TIMS6,7 and negative TIMS.8,9 have been developed for B isotopic analysis in the past few decades. PTIMS uses either Na2BO2+ or Cs2BO2+ molecules, while NTIMS uses BO2-. The former technique requires a few micrograms of B for each measurement in order to overcome the low ionization efficiency for either Na2BO2+ or Cs2BO2+ ions, and of the two methods, Cs2BO2+ can achieve a much better precision, up to 0.1‰10-14 (1) Shen, J. J.; Lee, T.; Chang, C. T. Anal. Chem. 1992, 64, 2216-2220. (2) Barth, S. Geol. Rundsch. 1993, 82, 640-651. (3) Palmer, M. R.; Pearson, P. N.; Cobb, S. J. Science 1998, 282, 1468-1471. (4) Pearson, P. N.; Palmer, M. R. Science 1999, 284, 1824-1826. (5) McKeegan, K. D.; Chaussidon, M.; Robert, F. Science 2000, 289, 13341337. (6) Palmer, G. H. J. Nucl. Energy 1958, 7, 1-12. (7) Ramakumar, K. L.; Parab, A. R.; Khodade, P. S.; Almaula, A. L.; Chitambar, S. A.; Jain, H. C. J. Radioanal. Nucl. Chem. Lett. 1985, 94, 53-62. (8) Papic, P. B.; Ciric, M. M.; Zmbov, K. F. Glas. Hem. Drus. Beograd. 1979, 44, 195-201. (9) Zeininger, H.; Heumann, K. G. Int. J. Mass Spectrom. Ion Phys. 1983, 48, 377-380. (10) Spivack, A. J.; Edmond, J. M. Anal. Chem. 1986, 58, 31-35. (11) Xiao, Y.; Beary, E. S.; Fassett, J. D. Int. J. Mass Spectrom. Ion Processes 1988, 85, 203-213. 10.1021/ac020589f CCC: $25.00
© 2003 American Chemical Society Published on Web 04/03/2003
mainlyas a result of the higher relative mass. In contrast, the BO2NTIMS method is characterized by a considerably higher sensitivity and, thus, allows the analysis of samples with low B abundance at nanogram level. However, because of the natural instrumental fractionation that is impossible to correct for B, BO2- NTIMS measurement suffers from a relatively large analytical uncertainty of 0.7-2.0 ‰,9,15-17 which has prevented the NTIMS technique from being applied to studies that require high spatial resolution (e.g., foraminiferal shells and corals). To minimize instrument fractionation,16,18-20 NTIMS studies routinely monitor several things very carefully. These include (1) sample size; (2) loading temperature, while the sample evaporates to dryness on the filament; (3) pH of the loading sample; (4) warming up process to reach the optimal ionization temperature; (5) the working temperature range; and (6) data acquisition. Despite these efforts, several serious analytical pitfalls are difficult to overcome in the traditional BO2- NTIMS measurement. These include (1) the large offset in the measured value for NIST SRM 951 B standard, as compared to the certified value that is confirmed by PTIMS results; (2) the lack of mass fractionation correction procedure; (3) irreproducible procedures to maintain a constant degree of fractionation during each analysis; (4) subjective rejection of unwanted portions of data during data acquisition. In this paper, we have modified our previous method that was set up to correct for mass fractionation, specifically for elements with only two isotopes and measured as oxides (LaO), by normalizing to the oxygen isotopic ratio. Combining the two isotopes of B with three oxygen isotopes to form BO2-, we expect six isotopic species for BO2- between 42 and 47 amu, three of which can be measured precisely because of the higher abundance. This enables the possibility of normalizing the ratio for oxide ions of the same metal isotope to the ratio for the oxygen isotopes,11 for example, 11B16O18O/11B16O16O can be normalized to 18O/16O. Here, we report the results of the NIST SRM 951 standard boric acid to examine in detail the ionization behavior of BO2- and to evaluate the possibility of using 18O/16O normalization to monitor the instrumental mass fractionation effects. EXPERIMENTAL PROCEDURES Ten milligrams of coral sample was dissolved in 3.3 mL of 0.12 N HCl. Only 0.1 mL of this solution was passed through a small PFA column (i.d. 1.5 mm and 3 cm in length) with 0.05 mL of Bio-Rad AG1×8 (200-400 mesh) anion resin to remove Ca, because it may interact with the activator loaded on the filament and, thus, disturb the mass fractionation. Boron passed through directly, but Ca and other trace elements remained in the column. (12) Ishikawa, T.; Nakamura, E. Anal. Chem. 1990, 62, 2612-2616. (13) Leeman, W. P.; Vocke, R. D.; Beary, E. S.; Paulsen, P. J. Geochim. Cosmochim. Acta 1991, 55, 3901-3907. (14) Swihart, G. H.; McBay, E. H.; Smith, D. H.; Siefke, J. W. Chem. Geol. 1996, 127, 241-250. (15) Duchateau, N. L.; Verbruggen, A.; Hendrickx, F.; De Bievre, P. Talanta 1986, 33, 291-294. (16) Vengosh, A.; Chivas, A. R.; McCulloch, M. T. Chem. Geol. 1989, 79, 333343. (17) Hemming, N. G.; Hanson, G. N. Geochim. Cosmochim. Acta 1992, 56, 537543. (18) Klo ¨tzli, U. S. Chem. Geol. 1992, 101, 111-122. (19) Hemming, N. G.; Hanson, G. N. Chem. Geol. 1994, 114, 147-156. (20) Heumann, K. G.; Zeininger, H. Int. J. Mass Spectrom. Ion Processes 1985, 67, 237-252.
The elutent was collected along with an additional 100 µL of 0.05 N HCl added to the column to wash out the remaining B. The boron-specific resin (Amberlite 743)21 was used to extract boron out of seawater, and the remaining seawater was served as the B-free seawater activator. The coral sample was then evaporated at 60 °C to dryness and then mixed with 0.3 µL of B-free seawater and loaded onto a Pt filament. To burn off the organic at mass 42 (CNO-) coming from acid-dissolved resin, the filament current was raised slowly until it turned dark red, and the current was maintained for 3 s. Our loading blank for B was
