5 There are currently two types of laser ablation systems being used for U-Pb geochronology: 1) nanosecond (ns) lasers (common); and 2) femtosecond (fs) lasers (currently rare).

6 The first type includes 266 nm, 213 nm, and 193 nm lasers. The exact mechanism used to generate ns-laser pulses of a specific wavelength and energy is discussed elsewhere (e.g. Basting and Marowsky 2005). It is worth noting, however, that 193 nm excimer (i.e. ‘excited-dimer’) lasers are generated using an ArF gas (the ‘premix’), which, when pumped above its ground state results in photon emission at a 193 nm fundamental frequency. This means that in contrast to solid-state Nd:YAG laser sources, no complex optical quintupling methods are necessary and this greatly simplifies the optical design and maintenance of 193 nm ArF excimer laser systems. Excimer laser pulses have higher energy density compared to 266 nm and 213 nm UV lasers and by varying laser operating conditions and using optical attenuators, 193 nm excimer lasers can deliver a wide range of fluences (<1 to 40 J/cm²) to the target. This wide dynamic range is necessary because ablation thresholds depend strongly on the composition of the target (i.e. absorption, electrical conductivity). For example, controlled ablation of translucent apatite typically requires fluence >10 J/cm² whereas zircon, monazite, and titanite can be ablated over a range of fluences (2 to 8 J/cm²)that can be fine tuned to achieve an ablation rate (e.g. 0.5 μm/s) suitable for in situ analyses in polished thin sections.

7 Whatever the ns-laser source, the energy contained in each pulse is transferred into the target lattice over ns-timescales. When the energy of the laser pulses exceeds a target’s ablation threshold, the solid’s binding energy is overcome causing it to thermally decompose into a mixture of atoms, ions, and fragments. The particle size distribution generated by ns-LA-ICPMS peaks in the range of 60 to 100 nm, and 193 nm ArF excimer lasers typically generate fewer μm-scale fragments, compared to 213 nm and 266 nm wavelengths (Guillong et al. 2003). The short timescales for interaction of the incoming pulse (e.g. between 4 ns and 20 ns for excimer lasers) allows most of the energy to be absorbed within the target area. Nonetheless, the target area is subject to zone heating that can cause preferential evaporation of volatile elements, such as Pb. This ablated material rapidly expands into a transient plasma plume above the target area. This plume interacts with subsequent laser pulses in a process known as plasma shielding which further alters the size and composition of the aerosol. Outer-shell electron transitions in the plasma plume also release photons with wavelengths characteristic of the elements present in the plume. This is the basis of laser-induced breakdown spectroscopy (LIBS).

8 The net result of these thermally activated laser-target interaction processes is ‘non-stoichiometric’ ablation and the occurrence of elemental fractionation. For U-Pb geochronology applications, this element fractionation is manifested as increasing Pb/U as craters deepen. Eggins et al. (1998) have elegantly demonstrated and described the processes occurring at the ablation site that give rise to this laser-induced elemental fractionation (LIEF). This study revealed the presence of a volatile-element-enriched condensate blanket surrounding ablation craters and related its formation to condensation of volatile elements (e.g. Pb) out of the expanding plasma plume. Figure 1 shows a typical condensate blanket surrounding craters in zircon analysed by the UNB laser ablation system. This observation suggests that heavier elements with greater inertia are actually preferentially entrained in the carrier gas so that some other process must account for the observed increase in Pb/U as a function of time during ablation of craters. Careful inspection of craters by Eggins et al. (1998) revealed, however, that heavier refractory elements (e.g. U) appear to be preferentially deposited (e.g. by vapour deposition) on the inner walls of the craters as the plasma plume expands from progressively deeper into the material. Their results also suggest that whereas the ablation process is quantitative for shallow craters (i.e. most of the laser interaction volume is converted to aerosol), deep craters (i.e. depth/diameter >>1) are subject to plasma re-sampling and re-deposition that prevents a large fraction of the ablated volume from becoming aerosolized and entrained in the carrier gas.

9 These are some of the characteristics of ns-laser ablations systems that cause Pb/U fractionation resulting from laser–target interactions in drilling mode. The following movies available on YouTube demonstrate how three laser operating conditions (19 μm crater @ 10 Hz, 26 μm crater @ 5 Hz, and 45 μm crater @ 4 Hz) can influence the time-dependent increase in ²⁰⁶Pb/²³⁸U during ablation of zircon 91500 standards:

10 In contrast, femtosecond laser pulses transfer energy to a material in a fundamentally different way. Rather than being a thermal decomposition process, femtosecond laser energy is believed to be absorbed primarily by electrons (e.g. photon–electron interactions). This energy absorption and transfer takes place over picosecond timescales such that bonds are broken without heating the lattice (i.e. a disequilibrium process). Binding energies are thus overcome without significant heating of the lattice: the material ‘explodes’ into an aerosol composed of a mixture of atoms, ions, and small molecular fragments. As a result, inter-element fractionation from laser-target interactions is predicted to be negligible. This has obvious advantages for U-Pb geochronology. The significantly higher cost (currently) of turn-key femtosecond LA-ICP-MS systems (Photon Machines FS198.G2, NewWave Research NWR-femto, Allied Spectra J100) has limited their deployment to a relatively small cohort of institutions. In North America, there are only a handful of femtosecond LA-ICP-MS systems operated by Earth Sciences departments (University of Windsor, University of Wisconsin, UC Santa Barbara, and UC Berkeley) with other experimental systems housed and operated by physics and materials science departments.

11 There are, however, many other sources of inter-element fractionation within a LA-ICP-MS system independent of the laser instrumentation. These include: 1) deposition of material on the inner surfaces of the cell (e.g. upper window or walls); 2) fractionation in the sample tubing (e.g. gravitational settling and electrostatic processes); and 3) mass-fractionation in the ICP-MS plasma and the ion-optical system. Some of these processes are not time-dependent and so can be compensated for by matrix-matched external standardization.

12 Changes to quadrupole-ICP-MS ion optical system and interface designs recently introduced by Agilent (e.g. 7700x), Thermo (X-SeriesII), Perkin-Elmer (NexION 300), and Bruker Daltonics (Aurora M90) have resulted in instruments with inherently higher sensitivity than previous generations. Additional steps can be taken to further enhance the sensitivity of LAICP-MS by changing the conditions of: 1) the ‘dry’ plasma, and 2) the vacuum in the interface region between the cones and the extraction lens. First, adding a small amount of high-purity H₂ or N₂ to the aerosol carrier gas can be used to significantly increase the temperature of the plasma and thereby enhance its ionizing efficiency (Guil-long and Heinrich 2007). Other studies have recently shown that mixtures of high-purity CH₄ + Ar and CH₃ OH (methanol) + Ar may have even larger effects on sensitivity enhancement (Fliegel et al. 2011), but this approach can also generate unwanted polyatomic interferences. Using N₂ has merit as it can be obtained at purities of the same level as the He carrier gas (e.g. 99.999%) and is inert but results in significantly increased nitride interferences. The amount of N ₂ or other gas mixture required to maximize sensitivity varies from instrument to instrument. Typical flows for N ₂ are between 2 to 5 mL/min added to the He + Ar carrier gas. Figure 2 shows the effect on ²⁰⁷ Pb intensity of adding 2.8 mL/min N ₂ to an Agilent 7700x. These data were acquired under the same analytical conditions used for UPb geochronology of zircon. It should be noted that sensitivity on NIST610 could be significantly increased simply by increasing the laser fluence (e.g. to 8 J/cm ² ), but this would compromise the goal of keeping ablation craters shallow for in situ work in thin sections. Notice that the robustness of the plasma (monitored as oxide production and U/Th ratio ≈ 1) remains high even under maximum sensitivity conditions. The much greater ionization efficiency under these conditions does enhance the production of doubly-charged species (more ions reach their second ionization potential), and may contribute to the formation of molecular ion interferences (e.g. possible ¹⁷⁶Hf^14N14N interference on ²⁰⁴Pb in zircon).

13 Another technique that can enhance sensitivity is the use of a larger interface rotary pump, or multiple rotary pumps, to increase the vacuum in the region between the sample cone and the first ion extraction lens. Thermo Scientific have introduced a similar upgrade for their magnetic-sector ICPMS line called the ‘Jet Interface’ that features a larger ‘dry’ vacuum pump and redesigned cones (see, for example, Newman 2012). We use a manually controlled second external rotary pump that can be turned on to increase sensitivity of Pb and U. In Figure 2, the effect of turning the second rotary pump on is illustrated by the sharp jump in ²⁰⁷Pb intensity. The effect is instantaneous because operating the second pump immediately lowers the pressure in the interface region from 2.3 × 10^-2 kPa to 1.9 × 10^-2 kPa on our Agilent 7700x ICP-MS. The combination of N₂ and second rotary pump allows us to obtain sensitivity as measured on the NIST610 silicate glass of ~2700 icps/ppm for ²⁰⁷Pb at a spot diameter of 33 μm, pulse energy of 4 J/cm², and repetition rate of 5 Hz. The zircon 91500 standard contains only about 1 ppm ²⁰⁷Pb such that under these analytical conditions we can achieve a net intensity of ~2500 icps for ²⁰⁷Pb.

14 For U-Pb geochronology of young accessory minerals, such as titanite, rutile, and allanite, which can contain a significant common-Pb component (e.g. ²⁰⁷Pb*/²⁰⁴Pb^c<100), it may be necessary to quantitatively subtract the common-Pb component from each of the radiogenic daughter products. This is most accurately accomplished by direct measurement of the ²⁰⁴Pb signal during ablation. The presence of trace-level impurities of Hg in the carrier gases (with ²⁰⁴Hg isobarically interfering with ²⁰⁴Pb) requires that Hg be either removed from the gas using impurity traps (e.g. gold-coated sand traps) or peak-stripped from the net mass 204 peak by monitoring ²⁰²Hg and assuming an instrumental ²⁰⁴Hg/²⁰²Hg value. Even so, some authors (Simon-etti et al. 2006) suspect the existence of additional polymolecular interferences on ²⁰⁴Pb, casting doubt on the reliability of this approach. Alternatively, many labs simply ignore ²⁰⁴Pb and rely on common-Pb corrections, such as the method proposed by Andersen (2002), which is based on the measured ²⁰⁸Pb signal, U/Th ratio, an estimate of the age of the target, and the model common-Pb composition (e.g. Kramers and Tolstikhin 1997). In the UNB lab, we use dedicated high-capacity (up to 20 L/min) Hg traps manufacturered by Vici Metronics on all of the gas lines that enter the laser and ICP-MS. These ensure that ²⁰⁴Hg gas backgrounds remain <150 icps under the highest sensitivity conditions.

15 The sample cell is arguably one of the most important components of any laser ablation system. The cell holds the samples and standards in a controlled atmosphere and presents the targets to the focused laser beam for ablation. Ideally, all ambient air must be removed from the cell (e.g. by repeated evacuation) and then continuously flushed with an ablation carrier gas (typically pure He or He + Ar). Aerosols generated during ablation should be fine grained, uniform in particle size distribution, and ideally be transported to the ICP-MS as quickly and efficiently as possible (fast cell washout and high sensitivity) with minimal contact with the cell walls or roof (which can cause additional inter-element fractionation). The conditions of aerosol generation and transport should also be identical anywhere in the cell to maximize its useable area. Whereas numerous custom-designed cells have been described in the literature, which are capable of fast washout or high sensitivity, commercially available cells come in two main configurations: 1) open cylindrical designs, and 2) two-volume cells. It is now widely documented that large-volume open cells can achieve fast washout and high sensitivity only for parts of the cell that lie between the inlet and outlet for the carrier gas. Transport of aerosol is efficient within this central channel, but inefficient and variable away from it, which limits the useful working area within these cells (Fisher et al. 2011; Koch and Gunther 2011). To circumvent the problem of maintaining constant washout and sensitivity throughout the cell, while also providing a large enough area to accommodate geologic samples (e.g. thin sections, polished slabs) and standards, two-volume low-volume (2VLV) cells have been developed. The most commonly used, the Australian National University (ANU) Helex design (Eggins et al. 1998) has been adopted by Resonetics (in collaboration within Laurin Technic Pty) and more recently by Photon Machines (e.g. Analyte G2 in collaboration with ANU Research School of Earth Sciences). A similar ‘large format cell’ was also developed by New Wave Research, but deployment of this cell has been limited. One of the principle benefits of the 2VLV cell is that ablation takes place within a low-volume (<1 cm³) funnel, which can move around inside the cell (in the ANU Helex and Laurin Technic M-50 designs the cell actually moves around the funnel which is fixed). This ensures that ablation conditions, washout times, and transport efficiency are identical anywhere in the cell. The low-volume of the 2VLV design also ensures that aerosol washout is extremely fast (several orders of magnitude in <1 second). This means that background levels are re-established only a few seconds after the cessation of ablation, so that the next target in the sequence can be started almost immediately. This innovation significantly increases the number of analyses that can be conducted per hour. In a recent detrital zircon study using a Resonetics M-50-LR laser ablation system equipped with a Laurin Technic Pty 2VLV cell, 97 analyses (76 unknown zircons, 18 zircon standards, and 3 NIST glasses), each comprising a 15-second ablation followed by a 20-second washout period, were completed in just over 1 hour. The detrital zircons occur in a single, heavy-mineral layer in a thin section of tourmalinized metaquartzite from the lower Aldridge member of the Belt-Purcell basin west of Kimberley, BC. The results of this ablation sequence are shown in Figure 3.