Improving Precision in Resonance Ionization Mass Spectrometry

Mar 16, 2011 - Influence of Laser Bandwidth in Uranium Isotope Ratio Measurements ... pulsed broadband dye laser (0.63-0.1 nm bandwidth) tuned...
1 downloads 0 Views 1MB Size
Download PDF
ARTICLE pubs.acs.org/ac

Improving Precision in Resonance Ionization Mass Spectrometry: Influence of Laser Bandwidth in Uranium Isotope Ratio Measurements B. H. Isselhardt,*,†,‡ M. R. Savina,§ K. B. Knight,‡ M. J. Pellin,§ I. D. Hutcheon,‡ and S. G. Prussin† †

Nuclear Engineering Department, University of California at Berkeley, Berkeley, California, United States Glenn Seaborg Institute, Lawrence Livermore National Laboratory, Livermore, California, United States § Materials Science Division, Argonne National Laboratory, Argonne, Illinois, United States ‡

ABSTRACT: The use of broad bandwidth lasers with automated feedback control of wavelength was applied to the measurement of 235 U/238U ratios by resonance ionization mass spectrometry (RIMS) to decrease laser-induced isotopic fractionation. By broadening the bandwidth of the first laser in a three-color, three-photon ionization process from a bandwidth of 1.8 GHz to about 10 GHz, the variation in sequential relative isotope abundance measurements decreased from 10% to less than 0.5%. This procedure was demonstrated for the direct interrogation of uranium oxide targets with essentially no sample preparation.

R

elative isotope abundances of uranium in nuclear materials are diagnostic of production and/or irradiation histories. For example, the 235U/238U ratio is useful for identifying intended use,1 the 234U/235U ratio can be indicative of the production process,1 and the ratios 236U/235U/238U can reveal a history of neutron irradiation.2 Measurements of these isotope abundances by mass spectrometry are common but typically require dissolution and chemical separation to ensure that contamination from interfering species (e.g., 238Pu in the case of 238U) is reduced or eliminated.3 In situations where time is a critical parameter, or when sample size is limited and background correction for chemical blanks influences detection limits and accuracy, alternative approaches may be useful. Resonance ionization mass spectrometry (RIMS) can distinguish isotopes of interest from isobaric interferences without the need for chemical processing. RIMS uses photons tuned to atomic transitions to selectively excite and ionize neutral atoms generated from a sample while leaving other atomic species largely un-ionized.4,5 RIMS is now routinely used for high-sensitivity and high-selectivity measurements of trace element and isotope composition in environmental6-8 and extraterrestrial samples.9,10 RIMS has been shown to be a potential tool for quantifying uranium isotope ratios.11-15 Early approaches used a single pulsed broadband dye laser (0.63-0.1 nm bandwidth) tuned to a one-color (591 nm), three-photon ionization scheme. Donohue et al.11 obtained RIMS spectra from U adsorbed onto ion-exchange beads covered with colloidal graphite; they reported a relative standard deviation (RSD) of 0.65% in the 235 U/238U ratio and an elemental selectivity against Pu of 3400 (Uþ ions/Puþ ions) when the laser was tuned for uranium. When the laser was tuned for plutonium, a Pu/U selectivity of only 150 was found. Green and Sopchyshyn,12 under very similar experimental conditions, obtained RIMS spectra from metallic r 2011 American Chemical Society

uranium and UO2(NO3)2 coated with colloidal graphite (to aid in the formation of neutral U atoms during thermal desorption); they demonstrated that the Uþ ion signal from uranyl nitrate was a combination of both resonantly ionized U atoms and nonresonant U ions attributed to the photodissociation ofþ UOx. Green and Sopchyshyn did not observe a resonant U signal from U3O8 deposited from a nitric acid slurry, possibly due to the lack of a colloidal graphite coating. They reported a RSD of 0.4% for the 235U/238U ratio for nonresonant ionization of UOx. Erdmann et al.13 combined resonant and nonresonant ionization in the analysis of depleted uranium to obtain a precision in the 235 U/238U ratio of 4.6% from a submicrometer particle containing less than 6  106 atoms of 235U. Goeringer et al.14 studied the relative yields of secondary and resonance Uþ ions using Arþ ion sputtering of uranium metal and oxides; they observed that the sputtered neutral atom yield depends strongly on the sample matrix composition. They did not observe Uþ photoions above background levels from U3O8. More recent RIMS studies have used chemical purification and isotopically selective ionization schemes with narrow bandwidth (1-3 MHz) continuous-wave lasers that are sequentially tuned to each isotope of interest. This technique is well-suited for measurements requiring a large dynamic range (>106) and provides 235U/238U ratios with precision ranging from 2% to 7%,6,15 limited mainly by the time dependence of sample atomization and laser parameters such as power and pointing stability. The potential of RIMS for measuring uranium isotope ratios with high accuracy in samples of unknown composition without the need for chemical

Received: September 30, 2010 Accepted: February 21, 2011 Published: March 16, 2011 2469

dx.doi.org/10.1021/ac102586v | Anal. Chem. 2011, 83, 2469–2475

Analytical Chemistry

ARTICLE

significantly improved internal precision in the measured isotope ratios for 235U/238U and 234U/238U. We discuss the advantages of the technique for discriminating between atomic uranium ions and possible isobaric interferences. Finally, we discuss the feasibility for developing this method into a technique useful for routine analysis of U-bearing samples.

Figure 1. A schematic representation of the first resonance transition for 234U, 235U, and 238U in wavelength, including an estimated Doppler broadening, is shown. Also shown are two ideal laser spectral distributions depicting the increase in superposition for a 5 pm bandwidth laser with multiple isotope resonances compared to a 1 pm laser with similar power. The lasers are shown at a mean wavelength of 415.5105 nm as in our ionization scheme, while the resonances for 235U and 238U are at 415.507 and 415.514 nm, respectively.

preparation, however, has not been demonstrated by any of these prior studies. The challenge of simultaneous isotope measurement by resonance ionization is to minimize the laser-induced isotope bias that can be severe and must be moderated.16 Levine et al.10 reported success measuring Cr isotope ratios to ∼1% precision using power broadening to overcome small isotope shifts (∼1 pm), the effects of Doppler broadening, and small pulseto-pulse fluctuations in wavelength. However, the much larger isotope shift between 235U and 238U and the need to maximize the ratio of resonant to nonresonant Uþ ions have prevented this approach from providing high precision in the measurement of isotope ratios from uranium oxides. Another option for simultaneous isotope measurements in the presence of large isotope shifts is the application of spectrally broadened lasers as initially demonstrated by Donohue et al.11 An idealized schematic, shown in Figure 1, locates the first resonances for 234U, 235U, and 238U excited in the present work and represents them as Gaussian distributions to approximate the spread in energy due to Doppler broadening (∼1 pm full width at half-maximum, fwhm) of the natural width (∼1 fm). Note that the separation in wavelength between the resonance centroids of 235 U and 238U is 7 pm. Superimposed on these resonances are two Gaussian models for the spectral distribution of the resonance laser with full width at half maxima of 1 and 5 pm, both of the same total intensity. The amplitude of the 5 pm bandwidth laser was chosen to demonstrate the overlap across an energy region sufficient for excitation of both 235U and 238U. Clearly, the 5 pm wide laser provides significantly improved overlap with the two U resonances. Table 1 lists the isotope shifts for the two resonance steps in our scheme, adapted from Schumann et al.18 The third resonance, to an autoionizing state,17 has a natural line width of about 52 pm, many times larger than the isotope shift.18 In the present work, we report the effect of laser bandwidth on the precision in measuring the relative abundances of U isotopes sputtered directly from unprocessed uranium oxide samples. The results show a marked reduction in the response of the measured isotope ratios to the wavelength of the first resonance laser when the bandwidth is increased from 1 to 5 pm, along with

’ EXPERIMENTAL SECTION Experiments were carried out with the CHARISMA instrument at Argonne National Laboratory, described in detail elsewhere.9 Solid uranium samples were sputtered with a 25 keV pulsed Gaþ ion beam with a spot size of ∼4 μm rastered over a 20  20 μm area. Immediately after the 300 ns Gaþ pulse, a potential of þ4 kV was applied to the sample to sweep charged species away from the ionization volume. The neutral sputtered atoms freely expanding into this volume were then irradiated 600 ns after the end of the Gaþ pulse by the three laser beams tuned to sequentially excite the two bound states given in Table 1 and the autoionizing state. The Uþ photoions were accelerated by a þ2 kV potential into a double-focusing time-of-flight mass spectrometer. Ions were detected with a microchannel plate detector at a sufficiently low rate [∼(1-3)  10-2 ions per pulse in a single 2 ns digitizer time bin] to avoid large dead time corrections.19 The entire system operated at 1 kHz. A typical acquisition sequence consisted of 105 laser pulses and required approximately 2.3 min. The resonance ionization scheme is a modification of a threecolor, three-photon ionization scheme described by Schumann et al.18 (Figure 2). The wavelengths of the first two lasers were tuned to the midpoints of the respective 235U and 238U resonances as shown schematically for the first resonance in Figure 1. All wavelengths are reported for vacuum. The lasers were Ti: sapphire cavities pumped with Nd:YLF lasers, producing 20 ns pulses of 1 mJ each in the wavelength range from 700 to 1000 nm. The first resonance beam of 415.5105 nm with a maximum energy of 250 μJ was obtained by frequency-doubling with an LBO crystal. The other lasers with wavelengths of 829.089 and 722.200 nm operated at energies of 600-780 and 500-650 μJ, respectively. At these pulse energies we are able to saturate the first two transitions (to the bound excited states) but not the autoionizing transition. The beams were focused to diameters of ∼1 mm and aligned in a nearly collinear geometry ∼0.5 mm above and parallel to the sample surface. In our implementation these systems nominally produce bandwidths (fwhm) of 1.3 GHz (3 pm in the fundamental range), which equates to approximately 1.8 GHz (1 pm) in the frequency-doubled beam. A broader bandwidth was generated by illuminating fewer lines on the tuning grating (by lowering the magnification of the prism pair beam expanders from 40 to 4) to give a frequencydoubled bandwidth of 9-13 GHz (5-7.5 pm), the broadest bandwidth we could generate stably. The bandwidths of the other two lasers were maintained at ∼3 pm for all experiments. Note that the bandwidth of the laser responsible for exciting the second transition (3 pm) is just smaller than the isotope shift for that transition (4 pm); broadening the bandwidth of this laser could also be explored as a source for improving the precision of uranium isotope ratio measurements. The isotope shift of the third transition is not accurately known, but it is negligible compared to the width of the autoionizing state. The experiments utilized the improvements in wavelength stabilization, pulse timing, and pointing stability described by 2470

dx.doi.org/10.1021/ac102586v |Anal. Chem. 2011, 83, 2469–2475

Analytical Chemistry

ARTICLE

Table 1. Isotope Shifts of Uranium for First and Second Resonance Transitionsa transition

U upper level energy (cm-1)

238

238

U wavelength (nm)

Δ234U (nm)

Δ235U (nm)

first resonance

5

L6 f 5L7

24 066.565

415.514

-0.008

-0.007

second resonance

5

L7 f J = 8

36 127.962

829.091

-0.0045

-0.004

ionization

J=8fJ=8

49 974.544

722.200

unknown

unknown

The transitions used to excite 238U and the energy of the upper excited level are shown. The isotope shifts for 234U and 235U are given as λ23X - λ238 and are from Schumann et al.18 All wavelengths are reported for vacuum. For the second resonance and ionization levels, only the total angular momenta are assigned. a

Figure 2. The three-color, three-photon resonance ionization scheme used in this work (adapted from Schumann et al.18) displays the difference in atomic transitions due to the isotope shifts for 235U and 238 U at each resonance, along with the corresponding average wavelength. For the high-lying excited states, only the angular momentum and parity are known.

Levine et al.10 Wavelengths were continuously monitored by either a High Finesse W/S-7 or W/S-6 wavelength meter. The output from the wavemeters was used to adjust the individual laser tuning gratings to maintain the desired wavelengths. The long-term drift in wavelength was controlled within (1 pm in the fundamental range ((0.5 pm in the second harmonic range), comparable to the pulse-to-pulse fluctuations (∼1.5 pm). At broad bandwidth, however, the laser wavelength is not accurately known as the laser spectrum is composed of multiple longitudinal modes with comparable amplitude, and it is not amenable to accurate measurement by the wavemeter. Observation of the laser spectral distribution for individual laser pulses on the wavelength meter and manual averaging over a number of laser pulses (∼50) provided an estimate of the mean wavelength with an accuracy of 1-2 pm. The timing of the laser pulses was monitored via photodiodes, and a programmable logic module was used to adjust the timing of the Nd:YLF pump laser triggers to synchronize the Ti:sapphire output pulses. The logic module ensures that the laser pulses arrive simultaneously in the ionization volume within the pulse-to-pulse timing fluctuations of the laser cavities, which was measured to be 14 ns on average. Finally, the pointing stability was maintained through use of regulated infrared heating of the laser cavities and beamline optics. Taken together, these enhancements drastically reduced long-term (∼hours) fluctuations in measured isotope ratios, enabling longer periods of stable analytical conditions with improved reproducibility.10 Isotope ratio measurements were performed on the wellcharacterized standards (235U/238U, composition) SRM 960 (0.0073, U metal with oxide coating; now known as CRM 112-A), CRM 125-A (0.04, UO2), and CRM U500 (0.50,

Figure 3. Three individual mass spectra in the U mass region from the CRM U500 show detected ions in the m/z region of uranium isotopes generated by 105 sputtering events for three different ionization conditions. One ion count has been added to the baseline of each trace to make it visible on the logarithmic scale and the higher traces have been multiplied by 4 and 20, respectively. The upper trace, “On Resonance”, was obtained via the resonance ionization scheme described in the text with the first resonance laser line width at 5 pm. The “Off Resonance” spectrum was collected under identical conditions but with the first resonance laser detuned by 0.05 to 415.561 nm. The “No Lasers” trace was collected by blocking the resonance lasers to detect only incompletely suppressed secondary ions produced by Ga+ sputtering.

U3O8). The metal standard was mounted on an aluminum stub with conductive epoxy, and the oxide standards were mounted simply by pressing into an indium metal foil. The data presented here represent four different experiments using these three samples.

’ RESULTS Resonant, Nonresonant, and Background Signals. The Uþ

signals from resonant and nonresonant ionization were distinguished from one another and from unsuppressed secondary ions produced by the Gaþ beam in three experiments using the CRM U500 U3O8 target. For resonant ionization, the excitation scheme was applied with 75 μJ/pulse in the first resonance laser at a bandwidth of ∼5 pm. The pulse energies of the second and third lasers were 650 and 510 μJ, respectively. Off-resonance spectra were obtained by detuning the first laser by 0.050 nm (to 415.561 nm). Finally, a residual secondary ion spectrum was obtained by blocking the lasers. The spectra obtained in these experiments are shown in Figure 3. The RIMS spectrum (labeled “On Resonance”) shows the expected peaks at m/z of 234, 235, 236, and 238, although the tailing of the m/z = 235 peak precludes clean separation of the contribution from 236U to the peak at m/z = 236. CRM U500 is known to contain 0.52 at. % 234U and 0.08 at. % 236U. The spectrum labeled 2471

dx.doi.org/10.1021/ac102586v |Anal. Chem. 2011, 83, 2469–2475

Analytical Chemistry

Figure 4. 235U/238U isotope ratio is shown as a function of the wavelength of the first resonance laser for two different bandwidths, measured from a natural uranium metal standard (SRM 960). The wavelengths of the second and third resonance lasers were held fixed at 829.089 and 722.344 nm, respectively. The vertical error bars are not shown, but the errors due to counting statistics are less than 1% relative for all data points.

“Off-Resonance” is much weaker and represents all sources of ions other than resonance ionization, and thus it provides a direct measure of the total background. Sources of ions in the offresonance spectrum include unsuppressed secondary ions (see below), ions generated by inadvertent grazing of the target by the edges of the resonance lasers (vanishingly few with proper laser alignment), nonresonant multiphoton ionization of neutral atoms and/or molecules,4 and photofragmentation of UO and/or UO2 into Uþ as discussed by Green and Sopchyshyn.12 The ratios of the peak intensities (resonant to nonresonant excitation) in these two spectra are 104 ( 10 and 111 ( 12 for masses 235 and 238, respectively. These ratios are clearly dependent on the details of the experimental conditions. They are also strongly dependent on the target composition and for the CRM U500 target were found to vary between 0.5% and 2.5% of the total U signal. Finally, the secondary ion spectrum (labeled “No Lasers”) shows nearly complete suppression of ions created directly by the Gaþ beam. The secondary ion intensity in the uranium mass region is ∼1.4 counts 3 amu-1 per 105 pulses, compared to an intensity of 235U of ∼150 per 105 pulses for nonresonant photoions and 15 600 for resonant photoions. Effect of Laser Bandwidth. In order to avoid isotopic fractionation in the simultaneous measurement of multiple uranium isotopes without the aid of power broadening, the approach of increasing the bandwidth of the resonance lasers was explored.20 We measured the 235U/238U isotope ratio from a natural uranium metal (SRM 960) target with a certified ratio of 0.0073 as a function of the wavelength of the first excitation laser for bandwidths of 1 and 5 pm. The wavelengths of the second and third resonance lasers were held fixed at 829.089 and 722.344 nm, respectively. The energies of the lasers in excitation order were 150, 900, and 540 μJ for the 1 pm bandwidth experiment versus 75, 900, and 600 μJ for the 5 pm bandwidth experiment. The isotope ratios determined in these experiments are shown in Figure 4. At a bandwidth of 1 pm, the 235U/238U ratio varies by over 2 orders of magnitude as the wavelength of the first resonance laser is scanned over the resonances of 235U (415.507 nm) and 238U (415.514 nm). When the bandwidth was increased to 5 pm, the variation in isotope ratio over the same wavelength range was reduced to a factor of 3. The dependence of the isotope ratio on the bandwidth is obviously quite large; for

ARTICLE

Figure 5. RIMS measurements were made of uranium sputtered from three U-rich materials of varying uranium isotopic composition by use of a 25 kV Gaþ primary beam. The spectra are normalized to the mass 238 integral. One ion count has been added to each data trace to make the baseline visible on a logarithmic scale. The scale is for the SRM 960 spectrum; the CRM 125A and CRM U500 spectra are multiplied by 4 and 20, respectively. The mass spectrometer was tuned for maximum sensitivity instead of mass resolution or abundance sensitivity and the tuning was not identical for each measurement, resulting in different peak shapes.

Table 2. Measured and Certified Values of Uranium Isotope Ratios on Standard Materialsa 235

U/238U

material

certified

measured

234

U/238U

certified

measured

SRM 960

0.0073(1)

0.0086(8)

CRM 125-A

0.041(1)

0.044(8)

0.000 39(1)

0.000 34(30)

CRM U500

0.9997(10)

1.155(10)

0.0104(1)

0.0118(20)

a

Certified uranium isotope ratios from several standards are shown with values as measured in Figure 5. Estimates of 2σ errors are shown based on counting statistics from single measurements and reported uncertainties in the isotope compositions of the standards.

1 pm bandwidth the deviation near the midpoint between the resonances is -220 ( 30% 3 pm-1, while for 5 pm bandwidth and the deviation is -13.5 ( 3% 3 pm-1. Although the sensitivity to wavelength is greatly reduced, it seems clear that a laser bandwidth well in excess of 5 pm is necessary to reduce the sensitivity to wavelength variation to levels of 100). Our enhanced signal-to-noise is derived from single-ion detection sensitivity and the use of a three-color resonance ionization scheme in which we optimize the laser power to minimize nonresonance ionization while preserving a high ionization probability for U atoms. The work of Goeringer et al.,14however, does underline the reduction in sputtering yield for U atoms from uranium oxides compared to uranium metal. Uranium oxides are challenging matrices for RIMS due to the strong tendency for U to desorb as molecular UOx, and the efficiency of desorbing neutral ground state atoms has not been optimized. Several groups have demonstrated increased efficiency for liberating neutral U and Pu atoms from oxide materials by using Ti as a reducing medium.22,23 There is also evidence to suggest that laser desorption methods may be more efficient than ion sputtering for creating neutral atoms.24 In addition to the three-color, three-photon ionization process described in Figure 2, other, more complex ionization pathways exist. Most possible three-photon combinations, other than the preferred route, are too low in energy to ionize atoms of U out of the ground state but there is at least one other significant twocolor, three-photon process in Figure 2 in which a photon from the first laser ionizes an atom out of the second excited state. In situations where the ionization by the third laser is unsaturated, this process may compete with the preferred process and contribute ions of a different isotope bias to the resonant ion signal. The involvement of this mechanism, and other possible but less likely ionization pathways, increase the importance of ensuring nearly complete ionization for available U atoms and the need for stability in laser power throughout an analytical session. The dynamic range of these measurements on the CHARISMA instrument is mainly limited by the abundance sensitivity of the spectrometer and secondary ion interferences. In this work, instrumental parameters were tuned for maximum sensitivity and not for maximum mass resolution or abundance sensitivity. The limitation imposed by the presence of incompletely suppressed ions from all nonresonant processes and the uncertainty in determining the background will limit the minimum detectable concentration. In the case of 236U this limit will likely be set by the tailing of ions from the 235 mass peak, the intensity of which will be a function of the instrumental abundance sensitivity. This instrumental constraint is improved in existing instrumental RIMS designs25 and could be further developed in an instrument designed to make these measurements routinely.

’ CONCLUSION Initial investigations using broad bandwidth laser resonance schemes for simultaneous ionization of multiple isotopes of U directly from uranium oxides have proved successful. These results indicate that it is feasible to use this approach for measuring relative isotope abundances for 234U, 235U, 238U, and likely 236U. We believe that this approach, with a properly calibrated wavelength control, will allow the measurement of relative abundances of any two uranium isotopes separated by 4

ARTICLE

amu or less, given sufficient abundance, virtually free of laserinduced bias. The empirical data from these studies are being used to develop an analytical model to define the laser system requirements to achieve a desired precision and accuracy in the measured isotope ratio. We are continuing to apply this approach to samples containing several different enrichments of U3O8 as well as other materials of interest; efforts have already resulted in measurement of U from U-bearing silicates without chemical preparation.26 This approach relies on reproducibility of laser performance, protection against drift in wavelength over time, and measurement of standards for correction of total analytical bias. In these experiments long-term drift was successfully controlled by the previously described wavelength feedback based on the High Finesse W/S 7 wavelength meter. The inaccuracy in the measured ratios, such as the 15% error in the measurement on CRM U500, can largely be explained by the difference between the intended and actual wavelength of the first resonance laser during this measurement and is the subject of current investigation. This experiment used a first-order laser modification to generate a broad wavelength distribution. Future options for improved stability and reproducibility include broader single-mode (or well-defined multimode) spectral distributions or rapid scanning of laser wavelength at constant frequency with a narrow and well-defined spectral distribution across the range of isotope resonances to create a constant average distribution of photon fluence across the resonances. The latter technique may prove useful for measuring isotopes of very low abundance by dwelling on or near specific resonances for improving detection efficiency.

’ AUTHOR INFORMATION Corresponding Author

*Address: 4155 Etcheverry Hall, MC 1730, Berkeley, CA 94720. E-mail: [email protected].

’ ACKNOWLEDGMENT This work was performed under the auspices of the U.S. Department of Energy by Lawrence Livermore National Laboratory under Contract DE-AC52-07NA27344. This work was funded by the Laboratory Directed Research and Development Program at LLNL under Project 10-SI-016, as well as with support from the Department of Energy Office of Nonproliferation Research and Development and the Department of Homeland Security. The CHARISMA facility at Argonne National Laboratory is funded by the U.S. Department of Energy, Basic Energy Sciences, Division of Material Sciences and Engineering under Award DEAC02-06CH11357.LLNL-JRNL-458116. ’ REFERENCES (1) Moody, K.; Hutcheon, I.; Grant, P. Nuclear forensic analysis; CRC Press: Boca Raton, FL, 2005. (2) Boulyga, S.; Becker, J. J. Anal. At. Spectrom. 2002, 17, 1143–1147. (3) Hou, X.; Roos, P. Anal. Chim. Acta 2007, 608, 105–139. (4) Hurst, G.; Payne, M.; Kramer, S.; Young, J. Rev. Mod. Phys. 1979, 51, 767–819. (5) Young, J.; Shaw, R.; Smith, D. Anal. Chem. 1989, 61, 1271–1279. (6) Raeder, S.; Fies, S.; Tomita, H.; Wendt, K. D. A. 4th Int. Conf. Laser Probing—LAP 2008, AIP Conf. Proc. 2009, 1104, 96–101. (7) Wendt, K.; Trautmann, N. Int. J. Mass Spectrom. 2005, 242, 161–168. 2474

dx.doi.org/10.1021/ac102586v |Anal. Chem. 2011, 83, 2469–2475

Analytical Chemistry

ARTICLE

(8) Erdmann, N.; Kratz, J.-V.; Trautmann, N.; Passler, G. Anal. Bioanal. Chem. 2009, 395, 1911–1918. (9) Savina, M. R.; Pellin, M. J.; Tripa, C. E.; Veryovkin, I. V.; Calaway, W. F.; Davis, A. M. Geochim. Cosmochim. Acta 2003, 67, 3215–3225. (10) Levine, J.; Savina, M.; Stephan, T.; Dauphas, N.; Davis, A.; Knight, K.; Pellin, M. Int. J. Mass Spectrom. 2009, 288, 36–43. (11) Donohue, D.; Smith, D.; Young, J.; McKown, H.; Pritchard, C. Anal. Chem. 1984, 56, 379–381. (12) Green, L.; Sopchyshyn, F. Int. J. Mass Spectrom. Ion Processes 1989, 89, 81–95. (13) Erdmann, N.; Betti, M.; Kollmer, F.; Benninghoven, A.; Gruning, C.; Philipsen, V.; Lievens, P.; Silverans, R. E.; Vandeweert, E. Anal. Chem. 2003, 75, 3175–3181. (14) Goeringer, D.; Christie, W.; Valiga, R. Anal. Chem. 1988, 60, 345–349. (15) Ziegler, S. L.; Bushaw, B. A. Anal. Chem. 2008, 80, 6029–6033. (16) Wunderlich, R.; Hutcheon, I.; Wasserburg, G.; Blake, G. Int. J. Mass Spectrom. Ion Processes 1992, 115, 123–155. (17) Lambropoulos, P.; Zoller, P. Phys. Rev. A 1981, 24, 379–397. (18) Schumann, P.; Wendt, K.; Bushaw, B. Spectrochim. Acta, Part B 2005, 60, 1402–1411. (19) Stephan, T.; Zehnpfenning, J.; Benninghoven, A. J. Vac. Sci. Technol., A 1994, 12, 405. (20) Payne, M.; Allman, S.; Parks, J. Spectrochim. Acta, Part B 1991, 46, 1439–1457. (21) Wunderlich, R.; Wasserburg, G.; Hutcheon, I.; Blake, G. Conf. Ser.—Inst. Phys. 1992, 128, 127–127. (22) Nunnemann, M.; Erdmann, N.; Hasse, H.-U.; Huber, G.; Kratz, J. V.; Kunz, P.; Mansel, A.; Passler, G.; Stetzer, O.; Trautmann, N.; Waldek, A. J. Alloys Compd. 1998, 271, 45–48. (23) Eichler, R.; Hubener, S.; Erdmann, N.; Eberhardt, K.; Funk, H.; Herrmann, G.; Kohler, S.; Trautmann, N.; Passler, G.; Urban, F. Radiochim. Acta 1997, 79, 221–233. (24) Nicolussi, G.; Pellin, M.; Lykke, K.; Trevor, J.; Mencer, D.; Davis, A. Surf. Interface Anal. 1996, 24, 363–370. (25) Veryovkin, I. V.; Calaway, W. F.; Moore, J. F.; Pellin, M. J.; Burnett, D. S. Nucl. Instrum. Methods Phys. Res., Sect. B 2004, 219220, 473–479. (26) Knight, K. B.; Savina, M. R.; Isselhardt, B. H.; Hutcheon, I. D.; Prussin, S. G.; Pellin, M. J. Unpublished work, 2010.

2475

dx.doi.org/10.1021/ac102586v |Anal. Chem. 2011, 83, 2469–2475