Laser Ablation Molecular Isotopic Spectrometry for Rare Isotopes of the Light Elements

Laser ablation molecular isotopic spectrometry (LAMIS) involves measuring isotope-resolved molecular emission. Measurements of several key isotopes (hydrogen, boron, carbon, nitrogen, oxygen, and chlorine) in laser ablation plumes were demonstrated. Requirements for spectral resolution of the optical detection system could be significantly relaxed when the isotopic ratio was determined using chemometric regression models. Multiple applications of LAMIS are anticipated in the nuclear power industry, medical diagnostics and therapies, forensics, carbon sequestration, and agronomy studies.

Laser ablation molecular isotopic spectrometry (LAMIS) is a technique that uses optical spectra of transient molecular species produced in ablation plumes in air or buffer gases (Ar, Ne, He, and N₂) for rapid isotopic analysis of solid samples (1–8). This technique is similar to laser-induced breakdown spectroscopy (LIBS), but optical emission spectra in LAMIS are measured at longer delays after an ablation pulse than what is used in LIBS. Molecular emissions yield relatively easily detectable isotopic information. Therefore, LAMIS adds a supplementary function of isotopic measurements to the well-established benefits of LIBS (real-time elemental analysis at atmospheric pressure, minimal sample preparation, chemical mapping and depth profiling at high spatial definition, and laboratory and field operation possible at a standoff distance to the sample). The LIBS and LAMIS techniques can be accomplished on the same instrument.

Molecular radicals are generated effectively when the ablation plume cools down, resulting in an increase of the molecular emission in the plasma afterglow. Several mechanisms contribute to the formation of molecules in the plume such as radiative association of neutral atoms, associative excitation or ionization, recombination of molecular ions, fragmentation of polyatomic clusters or nanoparticles, and vaporization of intact molecules from the ablated surface. Molecular spectra are advantageous for isotopic analysis because the isotopic shifts in molecular emission are considerably larger than in atomic spectra. The difference in isotopic masses has only a small effect on electronic transitions in atoms, but significantly affects the vibrational and rotational energy levels in molecules (1). A compact spectrometer can be sufficient to resolve molecular isotopic spectra; therefore LAMIS measurements can be performed in the field (4,5). The ability to measure isotope abundance using a portable spectrometer with modest spectral resolution is a significant merit of LAMIS, along with no sample preparation and data collection at ambient pressure.

The qualitative and quantitative studies of isotopes of the light elements are both very important for their use in many fields and for their very extensive scope, with many applications in different fields of their chemistries and materials science. This article describes measurements of isotopes of hydrogen, boron, carbon, nitrogen, oxygen, and chlorine. All of these elements are ubiquitous in nature as well as in practice, with carbon being the basis of many millions of compounds. Variation of the natural isotopic ratio ¹³C/¹²C in different materials ranges from 0.96% to 1.15% (9). These variations can be measured using LAMIS. Fractionation in stable isotopes of C, N, and O can be particularly indicative of a range of diverse biotic and abiotic processes relevant to the ecology, biosphere, and geochemistry, both organic (fossil) and inorganic. Fertilizers enriched typically above 40% in ¹⁵N relative to ¹⁴N are broadly used to track the efficiency of plant uptakes, fertilizer losses, and nitrogen turnover in soil.

High neutron absorbing capacity of the ¹⁰B isotope led to the development of multiple boron-loaded materials for neutron shielding in nuclear reactors and spent fuel storage pools, as well as for screens and curtains in nuclear medicine centers. Enrichment from 50% up to 99% in ¹⁰B relative to ¹¹B is often used. Many neutron detection devices are also based on highly borated (¹⁰B) and deuterated (²H) scintillator materials. Heavy water nuclear reactors require deuterium enrichment to 99.85%. Water-¹⁸O is used as a precursor in the radiopharmaceutical industry. Deuterated drugs and stable isotopic markers are increasingly used in the fields of medicine, pharmacology, nutrition, and physiology for tracing biochemical processes. Medical multinuclear magnetic resonance imaging (MRI) scanners require compounds enriched in nuclear magnetic resonance (NMR)-active isotopes (¹³C, ¹⁵N, and ¹⁷O). To assess quality, specifications, or aging of these materials, their isotopic homogeneity or distribution, and degree of isotopic enrichment can be rapidly tested by LAMIS in open air or inert gas flush without using laborious and expensive mass spectrometry (MS). The feasibility of standoff LAMIS analysis is particularly important for the nuclear industry.

Natural geochemical heterogeneity in isotopic composition reflects the complex history of our planet. Isotopic ratios of B, C, O, and Cl are particularly variable in nature because of chemical reactivity, high solubility, and volatility associated with the compounds consisting of these elements. Boron and chlorine isotopic fractionation is used to trace the records of evolution and weathering reactions because of interactions of rocks, soil, and sediment materials with water. Biomediated alteration of the ³⁷Cl/³⁵Cl ratio is distinctive in microbial reduction of anthropogenic perchlorates, biphenyls, halocarbons, and other chlorinated compounds. Simultaneous measurement of multiple stable isotopes is also increasingly used to constrain the carbon cycling in ecosystems better. LAMIS can facilitate the isotopic analysis directly in the field. Numerous applications of LAMIS are anticipated in the nuclear power industry, medical diagnostics and therapies, forensics, carbon sequestration, and agronomy studies.

Experimental

A flash-lamp pumped Nd:YAG laser operating at 1064 nm with a pulse energy of 50–150 mJ and a pulse duration of 4 ns was used to ablate the samples. The laser beam was focused onto the sample with a fused-silica lens to a spot diameter of ~100 μm, either in open air or in a cylindrical quartz chamber flushed with inert gas. Inert gas flow removed ambient air when the rare isotopes of nitrogen and oxygen were to be measured. A second lens was used to collect the laser-induced plasma emission in the direction perpendicular to the ablating laser beam. Two spectrometric configurations were interchangeably used. In one of them, collected light was focused onto a fiber-optic cable coupled to a compact echelle spectrograph (EMU-65, Catalina Scientific) fitted with an electron multiplying charge-coupled device (EMCCD). In the other configuration, collected light was focused onto the entrance slit of a Czerny-Turner spectrograph fitted with an intensified charge-coupled device (ICCD). The spectrum acquisition gate widths and delays were optimized to detect molecular emission, while minimizing intensity of atomic and continuum radiation. The collected spectra were averaged over several tens or hundreds of laser shots (see further text and notes in the figure captions). The spectral averaging, background correction, normalization, and final chemometric regression analysis of spectral data were performed using proprietary software designed at Applied Spectra for multivariate calibration and quantification of LIBS spectra.

Substances with known isotopic content were obtained from commercial sources in the form of powder, then mixed in appropriate proportions and pressed by a 7-ton press into 1-cm-diameter pellets. Isotopically enriched water was frozen into ice tables, which were kept cold by a Peltier thermoelectric plate during ablation for the LAMIS analysis. Samples of borated aluminum materials were provided by Ceradyne, Inc.

Results and Discussion

Our previous measurements of boron isotopes ¹⁰B and ¹¹B were performed using three chemically pure substances, isotope-enriched ¹⁰B₂O₃, ¹¹B₂O₃, and natural BN (2,7). The results indicated that LAMIS can be calibrated using multivariate partial least squares regression (PLSR) for quantitative measurements of isotopic ratio with high precision represented by relative standard deviation of only ±0.45% (2σ = 9‰), and this can be further improved (7). Abundances of individual isotopes in material composition can be measured at least down to ~1% (¹³C and ¹⁵N can be detected in subnatural concentrations, below 0.3%). High spectral resolution was found unnecessary for isotopic determination (2). However, an ambiguity remained about whether similar results could be obtained on practical samples with possible spectral interferences from other chemical elements.

The pertinence of LAMIS measurements to industrial applications was tested on neutron absorber materials. Two samples of aluminum alloyed composites with different boron isotope ratios were ablated and analyzed using a 320-mm IsoPlane spectrograph fitted with a PI-MAX4 ICCD detector (Acton/Princeton Instruments). One of these samples was a Boral core (Ceradyne) that consolidated aluminum alloy with 56% boron carbide of natural boron isotope abundance. The other sample was BorAluminum (Ceradyne), a regular unalloyed aluminum (grade 1100) with 4.5% added elemental boron, enriched in the lighter isotope ¹⁰B up to 95%. The ablation spectra of these samples recorded in the 591–608 nm spectral interval are presented in Figure 1 (continuum background subtracted). They are different because the isotopic composition is different. These spectra belong to moderately resolved rotational lines of boron monoxide in the vibrational band (0-3) of the A²∏_i→X²Σ⁺ emission system. The spectra were acquired using an 1800-grooves/mm (gr/mm) grating with the ICCD delay of 10 μs and a gate width of 100 μs. Spectral intensity was accumulated over 1000 laser pulses (analysis time 100 s at 10 Hz).

Figure 1: Isotope-specific molecular spectra of BO (A→X; 0-3) emission formed during ablation of two aluminum-based composite materials containing different 11B/10B ratios. The upper spectrum (BorAluminum) is shifted up for clarity. Lower spectrum: Boral.

In previous work on LAMIS, the BO A²∏_i→X²Σ⁺ emission was used at both the (0-2) band (4,7) and the (0-3) band (2,7,10). A choice of the spectral interval was driven by quality of the spectra for multivariate quantification with different chemometric procedures. The data in Figure 1 illustrate a significant isotopic shift between spectra of Boral (80% ¹¹B) and BorAluminum (95% ¹⁰B), requiring only modest spectral resolution. Another important factor is that this spectral interval does not have any visible interferences from atomic aluminum or chemical impurities in the samples. When the ablation spectrum was recorded at the usual acquisition delays for LIBS (<1 μs), several spectral lines of Fe, Ti, and Mn atoms coming from impurities in borated aluminum were observed within this range. Provided that the interfering lines are sparse, they do not necessarily preclude the isotopic analyses. Thus, time-gated attenuation of atomic emission is deemed essential for LAMIS measurements in practical analyses.

A possibility of LAMIS applications for geological analysis was examined using two tourmaline samples of different color (greenish "blue grass" and black minerals). The regular LIBS spectra of these samples indicated that a spectral region from 572 to 585 nm had the least amount of atomic interferences. Aimed at further decreasing atomic emission, the acquisition gate delay was set at 3 μs for LAMIS measurements. However, two spectral lines of atomic iron were persistent in the spectrum of one of the tourmaline samples. The positions of these interfering lines could be masked from the spectrum to determine the isotopic composition using multivariate PLSR described in the literature (2,7). Further quantitative work is needed to determine the limits of isotopic precision and sensitivity in LAMIS of geological samples. However, measuring boron isotopes in real geological samples using LAMIS has already been established as a credible prospect. Our previously attained precision of 9‰ (7) is reasonably adequate because natural variation of the ¹⁰B/¹¹B ratio in the terrestrial environment is more than 90‰ (9).

The ability to measure the ¹⁰B and ¹¹B content by LAMIS will be useful in other applications. For instance, the ¹¹B isotope is added to semiconductor grade silicon as a doping agent that must be controlled. LAMIS can monitor nuclear facilities remotely, delivering a laser beam through a small window and collecting optical emission either by a telescope or an optical fiber cable.

Carbon and nitrogen isotopes can be determined simultaneously using spectra of CN molecules in LAMIS measurements. This possibility was briefly mentioned in our earlier publications (1,4,5). The carbon isotope ¹³C is the most important one for studying biochemistry, but measuring only one isotope is often insufficient. Several key isotopes can follow the biological processes better. To demonstrate simultaneous isotopic measurements, we ablated natural graphite and several solid substances enriched in ¹³C and ¹⁵N. The spectra presented in Figure 2 include the B²Σ⁺→X²Σ⁺ vibrational band progression: (0-1), (1-2), (2-3), (3-4) of the CN radical. The rotational structure of these bands is barely resolved in our measurements, but isotopic shifts between isotopologues of ¹²C¹⁴N, ¹³C¹⁴N, and ¹²C¹⁵N are clearly observed. Two pellets of isotope-enriched benzamide with ¹³C and ¹⁵N atomic fractions 14% and 99%, respectively, were used as ablation targets. The latter was ablated in a helium flow to preclude nitrogen entrainment from air. Another sample was 99% enriched with ¹³C amorphous carbon. The spectra were acquired using a 10-μs timing gate delayed after the laser pulse only by 1.5 μs because CN molecules are promptly formed in laser ablation.

Figure 2: Emission spectra of CN vibrational band progression (B→X; Δν = +1) formed during ablation of benzamide pellets enriched in 13C and 15N isotopes. Spectra were averaged over 100 laser pulses.

Isotopic shifts in the bands with changing vibrational quantum number Δν = +1 shown in Figure 2 are all larger than 0.3 nm and are easily resolvable using a compact optical spectrograph. These shifts are about 10 times greater than the isotopic shift in the (0-0) band head of the same CN B²Σ⁺→X²Σ⁺ transition at ~388 nm (1). The opposite change in a vibrational quantum number Δν = –1 produces another CN band progression around 359 nm (Figure 3). The isotopic differences in these bands are well resolved, but their rotational structure is not. An advantage of using the bands (1-0), (2-1), and (3-2) near 359 nm shown in Figure 3 compared to those shown in Figure 2 is that the spectral peaks with Δν = –1 attributed to the minor isotopes ¹³C and ¹⁵N occur on the other side of and apart from the sharp edge of the ¹²C¹⁴N (1-0) band head. This is important for quantitative calibration. The progression with Δν = +1 requires multivariate calibration because the bands from minor isotopes can be hidden within unresolved rotational lines of the major ¹²C¹⁴N bands. In contrast, spectral data of the progression with Δν = –1 can be quantified using either univariate or multivariate calibration.

Figure 3: CN vibrational bands (B→X; Δν = –1) formed during ablation of natural graphite and enriched benzamide pellets. Spectra were averaged over 100 laser pulses.

The results of calibrating LAMIS spectra of a CN molecule using a multivariate PLSR procedure for quantitative determination of ¹³C content in solid organic samples are presented in Figure 4. Artificial reference samples were prepared by physically blending powders of natural and ¹³C-enriched decanoic acid. A natural ¹³C atomic fraction was assumed as 1.07%, and an additional five blended samples were made nominally at 1.2%, 1.5%, 2.0%, 5.0%, and 9.9%. The CN B→X (Δν = –1) spectra were collected in the interval 353–362 nm, similar to those shown Figure 3. The error bars in Figure 4 correspond to standard deviation of a single measurement within 10 replicates, while each measurement was made by accumulating spectra from 100 laser pulses. The natural ¹³ C atomic fraction was recalculated from these measurements as 1.09 ± 0.14%. Averaging over 10 measurements (each one made of 100-pulse accumulations) reduces a random error of the average down to ±0.044%. It is obvious from data presented in Figure 4 that LAMIS can measure subnatural abundances of ¹³C with sufficient precision. However, the decanoic acid powders used to prepare our reference samples consisted of microcrystals and it was difficult to fully homogenize a blended material. Better homogenized standards will improve precision and accuracy of LAMIS measurements.

Figure 4: Calibration plot with a linear fit and 95% confidence interval for the 13C isotopic content calculated using the PLSR model.

We tested LAMIS in analyzing the soil samples, in which the total carbon content was 12.26%, and the nitrogen content was 0.27%. For initial experiments, we spiked this soil with ¹⁵N-enriched benzamide and ¹³C-enriched decanoic acid in different proportions. Several spectral lines of iron appear as interferences in the region of CN B→X Δν = –1) bands, but iron lines were well resolved from carbon features. An algorithm that automatically fits a Lorentzian profile to every interfering spectral line and digitally subtracts them from the spectrum can be applied as a remedy. Such a procedure can work relatively well if the interfering lines are no more than three times stronger than the main spectrum.

A similar multivariate PLSR calibration was performed for measurements of the ¹⁵N isotope using reference samples prepared by blending powders of natural and ¹⁵N-enriched benzamide. The ¹⁵N atomic fractions in prepared reference samples were 0.37%, 1.0%, 2.0%, 5.0%, 10%, 20%, and 99%. In these experiments, the CN B→X (Δν = +1) spectra were collected within the range 414–422 nm, similar to those shown Figure 2. The results are displayed in Figure 5 with the error bars corresponding to standard deviation of a single measurement, each made by signal accumulation from 100 laser pulses. Again, quantification precision and accuracy were limited by microscale isotopic inhomogeneity of the reference samples that were prepared. Inaccuracies in preparing these reference mixtures to be equal to the nominal isotopic content also increased the calibration errors.

Figure 5: Calibration plot with a linear fit and 95% confidence interval for the 15N isotopic content calculated using the PLSR model.

During laser ablation of carbon-containing samples, both CN and C₂ radicals are often generated in comparable quantities. Hence, both of these species can be used for measuring the carbon isotopic ratio. Utilization of diatomic carbon molecules C₂ for LAMIS measurements was described previously (1,5,8). Rapid spatial and temporal evolution of the plasma plumes during laser ablation may cause kinetic fractionation effects altering stoichiometric and isotopic equilibrium between the plasma and the sample. Kinetic isotope fractionation was indeed observed at early stages of plume evolution (gate delays <2 μs) in C₂ emission (8). At longer delays, the measured isotopic ratio was close to the real ratio in the sample. These results suggest that carbon isotopes ¹²C and ¹³C can behave differently in laser ablation plumes, and the change in ratio is related to the time and spatial distance above the sample. For the purpose of isotopic analysis, it is important that the isotopic ratio measured at longer delays (>2 μs) equilibrates and approaches the real ratio in the sample roughly at the same time when molecular emission is also reaching its maximum.

To demonstrate LAMIS for hydrogen and oxygen isotopic measurements, we ablated ordinary water ice (H₂O), heavy water ice (D₂O), and ice with 90% atomic fraction of ¹⁸O. As in earlier work (1,6), we recorded emission of hydroxyl radicals at the A²Σ⁺→X²∏_i (0-0) transition in the ablation plumes. The major features of OH emission are R₁₁ and R₂₂ branch heads near 306 nm, the Q₁₁ branch head at ~308 nm, and the Q₂₂ branch head at ~309 nm. These rotational branch heads are suitable for the D/H ratio determination because the isotopic shifts between OH and OD features are large and can be measured at low resolution. However, isotopic shifts between ¹⁶OH and ¹⁸OH are smaller, especially for the lines with low rotational quantum numbers. The isotopic shifts increase with increasing quantum numbers in the band tails.

The hydroxyl isotopologue spectra recorded in wavelength region 311.8–314.0 nm using an IsoPlane spectrograph (Princeton Instruments) with a 3600-gr/mm grating are presented in Figure 6. The ICCD gate duration was 50 μs and delayed 5 μs after the laser pulse. Argon flow was applied to displace air. The most intense features contributing to the shown spectra belong to OH rotational lines of Q₁₁ and Q₂₂ branches with quantum numbers from J = 13.5 to J = 18.5. Smaller contributions come from P₁₁, P₂₂, and other minor branches of the (0-0) band, and also from the second vibrational band (1-1). These results demonstrate that hydrogen and oxygen isotopes can be determined simultaneously using spectra of hydroxyl isotopologues ¹⁶OH, ¹⁸OH, and ¹⁶OD recorded with a compact spectrograph. Quantitative isotopic calibration for ice analysis can be accomplished using multivariate PLSR as previously realized for water vapor (6).

Figure 6: Spectra of 16OH, 18OH, and 16OD molecules in the A→X (0-0) band formed during laser ablation of ice. Spectra were averaged over a series of 10 replicates, each made of 100-pulse accumulations.

Isotopic analysis of ice is the main tool in paleoclimatology and glaciology studies, but measuring ¹⁸O/¹⁶O ratios in rocks and minerals is important for geochemists as well. Aluminum oxide is a common constituent in many minerals. As a result, molecular spectra of AlO are easily detectable in laser ablation of rocks. We observed five band progressions of AlO B²Σ⁺→X²Σ⁺ (Δν = 0,±1, ±2) emission within a broad interval 445–545 nm from several rocks including feldspar, tourmaline, and rock salt. However, this spectral range is generally contaminated by many other spectral lines, and each particular experiment will require choosing a smaller specific segment of the spectrum. We ablated natural and isotope-enriched Al₂O₃ pellets, flushed with helium flow to illustrate that Al¹⁸O and Al¹⁶O can be resolved in LAMIS. The spectra of AlO B→X (Δν = +1) are presented in Figure 7, indicating that AlO emission can be used to measure the oxygen isotopic ratio ¹⁸O/¹⁶O in rock samples.

Figure 7: Emission spectra of AlO vibrational band progression (B→X; Δν = +1) formed during ablation of Al2O3 natural and 18O-enriched pellets. Spectra were acquired with a 20-μs delay and averaged over 100 laser pulses.

Calcium oxide is another common constituent of rocks and minerals. Several emission bands of CaO A¹Σ⁺→X¹Σ⁺ (2-0, 1-0, 0-0, 0-1) were clearly observed within 765–960 nm during ablation of calcite (CaCO₃) and calcium chloride (CaCl₂) in air. Molecular structure and spectrophysical properties of CaO are similar to those of SrO. Emission of the latter was previously used for LAMIS detection of three strontium isotopes (⁸⁸Sr, ⁸⁷Sr, and ⁸⁶Sr) (3–5). Our numerical simulation of the Ca¹⁸O and Ca¹⁶O spectra predicted that CaO emission can be used to infer oxygen isotopic information. Additionally, laser ablation of CaCl₂ produced emission of the CaCl B²Σ⁺→X²Σ⁺ band progression with Δν = –1. We recorded the spectra of CaCl B→X system using two spectrographs, one of which was 1250-mm Czerny-Turner spectrograph (Horiba JY) and the other one was an echelle spectrograph EMU-65. The results are displayed in Figure 8. The natural isotopic ratio of chlorine ³⁷Cl/³⁵Cl is 24.2% to 75.8%, respectively. The isotopic shifts because of these two isotopes are apparent in the spectra shown in Figure 8. While the large high-resolution spectrograph recorded a superior spectrum, the compact echelle spectrograph yielded sufficient resolution for LAMIS measurements.

Figure 8: Spectra of CaCl band progression (B→X; Δν = –1) formed during ablation of natural CaCl2 pellets. Spectra were acquired with a 20-μs delay and averaged over 20 laser pulses.

Conclusion

We demonstrated several possibilities of using LAMIS for rapid optical analysis of rare isotopes of hydrogen, boron, carbon, nitrogen, oxygen, and chlorine. A significant advantage of this technique is that solid samples can be analyzed in their original unaltered condition, without preparation. Precleaning of the sample surface can be completed by laser ablation before or during the analysis, which may be considered as depth profiling. Isotopic two- and three-dimensional mapping is possible. Quantitative measurements can be realized using multivariate regression models that will relate the spectral intensities of the isotope-specific molecular spectra to the original abundances of isotopes in the sample. Quantification is based on measuring spectra of known reference samples. Calibrating the LAMIS response for quantitative determination of C and N isotopes is presented here, while calibration for H and B isotopes was described earlier (2,6,7). Accuracy and precision of determination depends on isotopic homogeneity of the reference standards, but the analyzed samples can be heterogeneous. LAMIS can yield a map of the isotopic distribution.

References

(1) R.E. Russo, A.A. Bol'shakov, X. Mao, C.P. McKay, D.L. Perry, and O. Sorkhabi, Spectrochim. Acta, Part B 66, 99–104 (2011).

(2) X. Mao, A.A. Bol'shakov, D.L. Perry, O. Sorkhabi, and R.E. Russo, Spectrochim. Acta, Part B 66, 604–609 (2011).

(3) X. Mao, A.A. Bol'shakov, I. Choi, C.P. McKay, D.L. Perry, O. Sorkhabi, and R.E. Russo, Spectrochim. Acta, Part B 66, 767–775 (2011).

(4) A.A. Bol'shakov, X. Mao, C.P. McKay, and R.E. Russo, Proc. SPIE 8385, paper 83850C (2012).

(5) A.A. Bol'shakov, NASA Tech Briefs, Photonics, 36(11), Ia–3a (2012).

(6) A. Sarkar, X. Mao, G.C.-Y. Chan, V. Zorba, and R.E. Russo, Spectrochim. Acta, Part B 88, 46–53 (2013).

(7) A. Sarkar, X. Mao, and R.E. Russo, Spectrochim. Acta, Part B 92, 42–50 (2014).

(8) M. Dong, X. Mao, J.J. Gonzalez, J. Lu, and R.E. Russo, Anal. Chem. 85, 2899–2906 (2013).

(9) T.B. Coplen et al., US Geological Survey, Water-Resources Investigations Report 01-4222 (2002).

(10) B. Yee, K.C. Hartig, P. Ko, J. McNutt, and I. Jovanovic, Spectrochim. Acta, Part B 79–80, 72–76 (2013).

Alexander A. Bol'shakov is with Applied Spectra, Inc., in Fremont, California. Richard E. Russo is with Applied Spectra, Inc., and the Lawrence Berkeley National Laboratory at the University of California in Berkeley, California. Xianglei Mao and Dale L. Perry are with the Lawrence Berkeley National Laboratory at the University of California. Direct correspondence to: rerusso@lbl.gov