Effects of Plume Hydrodynamics and Oxidation on the Composition of

Feb 1, 2018 - New type of protein interaction observed. Biochemists have long thought that biomolecules must have or develop well-defined structures t...
0 downloads 11 Views 3MB Size
Download PDF
Article Cite This: J. Phys. Chem. A 2018, 122, 1584−1591

pubs.acs.org/JPCA

Effects of Plume Hydrodynamics and Oxidation on the Composition of a Condensing Laser-Induced Plasma David G. Weisz,*,† Jonathan C. Crowhurst,† Mikhail S. Finko,‡ Timothy P. Rose,† Batikan Koroglu,† Reto Trappitsch,† Harry B. Radousky,† Wigbert J. Siekhaus,† Michael R. Armstrong,† Brett H. Isselhardt,† Magdi Azer,§ and Davide Curreli‡ †

Lawrence Livermore National Laboratory, 7000 East Avenue, Livermore, California 94550, United States Department of Nuclear, Plasma, and Radiological Engineering, University of Illinois at Urbana−Champaign, Champaign, Illinois 61801, United States § Illinois Applied Research Institute, University of Illinois at Urbana−Champaign, Champaign, Illinois 61820, United States ‡

S Supporting Information *

ABSTRACT: High-temperature chemistry in laser ablation plumes leads to vaporphase speciation, which can induce chemical fractionation during condensation. Using emission spectroscopy acquired after ablation of a SrZrO3 target, we have experimentally observed the formation of multiple molecular species (ZrO and SrO) as a function of time as the laser ablation plume evolves. Although the stable oxides SrO and ZrO2 are both refractory, we observed emission from the ZrO intermediate at earlier times than SrO. We deduced the time-scale of oxygen entrainment into the laser ablation plume using an 18O2 environment by observing the in-growth of Zr18O in the emission spectra relative to Zr16O, which was formed by reaction of Zr with 16O from the target itself. Using temporally resolved plumeimaging, we determined that ZrO formed more readily at early times, volumetrically in the plume, while SrO formed later in time, around the periphery. Using a simple temperature-dependent reaction model, we have illustrated that the formation sequence of these oxides subsequent to ablation is predictable to first order.

1. INTRODUCTION Laser ablation is a broadly used tool with application to a wide array of fields, including analytical chemistry, electronics engineering, and biology.1 Despite its widespread use, the physicochemical processes that control plume dynamics are multifaceted and not completely defined.2 High-temperature chemical interactions in multicomponent (e.g., elements and molecules) laser ablation plumes may be difficult to interpret, due to the formation of multiple molecular species during plume expansion.1,2 These high-temperature chemical processes have implications for the precision of sample analyses that employ laser ablation as a probe, such as laser-ablation inductively coupled plasma mass spectrometry (LA-ICPMS),3−5 and may produce unfavorable thin-film compositions in pulsed-laser deposition.6,7 Pulsed laser deposition of perovskite materials for thin film deposition is an active area of research for epitaxial film growth; these films are of interest to a broad range of industries including photovoltaics,8 fuel cells,9 and semiconductor development.10 Understanding thin film stoichiometry is critical, given its effect on the material properties that are exploited by these applications. Pulsed laser deposition typically imparts near-stoichiometric transfer of material, making it an ideal method for thin film deposition.11 However, the ambient gas composition and pressure, laser parameters, and target © 2018 American Chemical Society

composition can induce molecular formation in the expanding laser-induced plasma plume, which can change the resultant stoichiometry.7,12,13 Numerous studies have used time-resolved emission spectroscopy to study the formation of molecules in laser ablation plumes,14−17 and some have begun to explore the spatial distribution of such molecules.18−20 Some studies have compared atomic and molecular emission features using narrow bandpass filters to analyze plume hydrodynamics and composition,18,19 while others have investigated diatomic oxide formation via laser-induced fluorescence (LIF) as a function of spatial position.6,7,20 These studies tracked the evolution of ions and atoms into diatomic oxide species showing that molecules preferentially form at the cool peripheral edge of the expanding laser ablation plume.2,18 It is theorized that this may be due to resultant shockwaves18 or thermal gradients in the ablation plume.2 In multicomponent plumes expanding into ambient oxygen, vapor-phase oxidation may impart a time-dependent molecular formation process that could result in condensation-induced chemical fractionation. Received: December 5, 2017 Revised: January 16, 2018 Published: February 1, 2018 1584

DOI: 10.1021/acs.jpca.7b11994 J. Phys. Chem. A 2018, 122, 1584−1591

Article

The Journal of Physical Chemistry A

ambient gas into the ablation plume, by observing the isotopic spectral shift between Zr16O and Zr18O molecular emission bands after ablation of the SrZrO3 target. To measure this isotopic spectral shift, spectra were acquired after the ablation of a ZrO2 target (prepared in the same way as the SrZrO3 target), in environments of pure 16O2 and 18O2 (97.4% 18O2, 1.7% 17O2). These spectra were acquired using an 1800 lines/ mm grating at 4050 ns after ablation, by which time the atomic emission was much reduced. Spectra were then acquired from the ablation of the SrZrO3 target in the of 20% 18O2 and 80% N2 environment at 300, 1050, 2050, and 3050 ns after triggering by the laser. Plume images (1024 × 1024 pixel, grayscale TIF images) were collected using an Andor iSTAR intensified CCD (ICCD) at 500 ns delay increments from 50 to 9050 ns, with a 1000 ns gate-width. A 1% transmission neutral density filter was used for imaging full plume emission, to avoid saturation of the ICCD. Narrow bandpass filters (centered at 570 and 600 nm, respectively, with 10 nm fwhm) were used to selectively record emission from SrO and ZrO molecules during plume image acquisitions. The 5× Mitutoyo apochromatic objective was again placed orthogonal to the path of the ablation laser, and a planoconvex lens (178 mm focal length) was used to produce an image on the ICCD at a magnification of ∼4.5×. The intensifier gain was adjusted to avoid saturation of the ICCD. Image contrast at different delay increments are not directly comparable, as all images were contrast-adjusted to elucidate features of interest in the plume, although the images within a given delay increment were adjusted in an identical fashion. Images taken through the narrow bandpass filters were falsecolorized to show the spatiotemporal distribution of SrO and ZrO throughout the plume. The spatial scales of the plume images were calibrated by comparison with a 150-μm standard pinhole placed at the focus of the optical system. Full details on the plume image acquisition parameters and image processing for each image are given in the Supporting Information.

Here, we illustrate that the spatiotemporal distribution of molecules formed in a SrZrO3 laser ablation plume is dependent on both plume hydrodynamics and the hightemperature chemical properties of the plasma constituents. The SrZrO3 target material is relevant to perovskite thin film production,10,21,22 and its stoichiometry ensures a homogeneous distribution of Sr and Zr in the target. Key emission features of SrO and ZrO lie in approximately the same spectral range, allowing their in-growth to be observed simultaneously in a single, moderate-resolution spectral window. Using narrow bandpass filters, we obtained images of the laser ablation plume as a function of time, revealing the distribution and evolution of these molecular species. Combining these data with spectra acquired in ambient 18O2, we show spatial partitioning of SrO and ZrO in laser ablation plumes and temporally correlate their formation with gas entrainment in the plume. Finally, by modeling the chemical kinetics of oxide formation in a simple, 3-element system, we can predict how oxidation conditions affect rates of molecular formation and the resultant vapor composition.

2. EXPERIMENTAL METHODS Sample targets were prepared by mixing 100 mesh SrZrO3 powder in poly(vinyl alcohol) to form a paste and pressing the material into a 1 cm cylindrical pellet 3 mm in thickness. The pellet was dried at 800 °C for 24 h, resulting in a hardened, brittle target. We ablated the SrZrO3 target surface in air (1 atm), using a 1.4 mJ, 1064 nm wavelength, ∼5 ns laser pulse from a New Wave Polaris II Nd:YAG laser. Given an output beam diameter of ∼2.75 mm, the pulses were focused using a 6.4 cm focal length planoconvex lens to achieve ablation of the surface with a theoretical spot-size limit of 31 μm (180 J/cm 2 ) corresponding to an intensity of 3.6 × 1010 W/cm2. A 5× Mitutoyo apochromatic objective placed orthogonal to the path of the ablation laser was used to collect emitted light. Space-integrated spectra from the plumes were acquired using a Shamrock SR-303i spectrometer equipped with 300, 1200, and 1800 lines/mm gratings and coupled to an Andor iSTAR intensified CCD (1024 × 1024 pixels, 13 μm pixel size), that was triggered for acquisition by the ablation laser introducing a ∼50 ns latency between ablation and acquisition. We ablated a Zr-metal target and a SrO target (prepared in the same way as the SrZrO3 target) under the same conditions to acquire data that would allow us to deconvolve Zr- and Sr-related emission features in the spectra acquired from the ablation of the SrZrO3 target (see Supporting Information). For both the SrZrO3 and the Zr-metal samples, a series of spectra were collected using the 1200 lines/mm grating at 550 ns, and subsequently in 1000 ns increments from 1050 to 20050 ns (10050 ns for the Zr-metal sample), after triggering by the laser taking into account the 50 ns latency between the laser firing and the triggering of the camera acquisition. The spectra were summed over 200 shots at 20 Hz, with a spectral window of 30 nm (∼570−600 nm), with a 500 ns gate-width for each acquisition, at each given delay. The wavelength was calibrated using a Ne standard lamp, and the Ne atomic emission lines had a measured fwhm of 0.18 nm at the 1200 lines/mm grating and a fwhm of 0.11 nm at the 1800 lines/mm grating. Measured spectral intensities were corrected for the instrumental response using a standard lamp. Ablation experiments were then conducted in an environment of 20% 18O2 and 80% N2 (1 atm), to track the influx of

3. RESULTS Time-Resolved Emission Spectroscopy of the SrZrO3 Target. In Figure 1, we show the normalized emission spectra from SrZrO3 ablation at 550, 3050, 5050, and 9050 ns postablation. The emission spectra are normalized to their minimum and maximum intensities, to emphasize the relative intensities of features that are prominent as the ablation plume evolves over time. Spectral data acquired 550 ns after ablation is dominated by atomic emission lines, seen as narrow, well-resolved peaks in Figure 1. While molecular emission may be visible at 550 ns, these features are significantly more intense at later times. ZrO emission bands from the c3Π2−a3Δ3 (0,0) ZrO transition23 are clear in the 3050 and 5050 ns spectra, in the region highlighted in blue. The SrO orange band is evident as a broad, doublehumped feature at 5050 ns (shown highlighted in red) and is likely the convolution of the emission from multiple SrO transitions, including at least two identified 3Π-3Π transitions.24 By 9050 ns, the ZrO features are barely distinguishable from the background, while the SrO orange band features are prominent. The temporal variation in the emission intensities of ZrO and SrO was quantified by comparing the average emission intensity of the ZrO c3Π2−a3Δ3 (0−0) transition region (571− 573 nm) to the SrO orange band region (592−600 nm) as a function of time after ablation, shown in Figure 2. Average 1585

DOI: 10.1021/acs.jpca.7b11994 J. Phys. Chem. A 2018, 122, 1584−1591

Article

The Journal of Physical Chemistry A

Figure 2. Blue line represents the emission intensity of the ZrO c3Π2− a3Δ3 band (integrated intensity from 571 to 573 nm) and the red line represents the emission intensity of the SrO “orange band” system (integrated intensity from 592 to 600 nm) as a function of time postablation, indicated in arbitrary units (a.u.). The ZrO band becomes dominant between 2 and 5 μs, after which the SrO band is dominant. At early times (5 μs), they are dwarfed by the ZrO and SrO emission features.

Figure 1. Emission spectra acquired at 550, 3050, 5050, and 9050 ns after ablation of the SrZrO3 target in air. At 550 ns, the spectrum is dominated by Zr atomic emission, though features assigned to ZrO are present. The peak at ∼585 nm marked with an asterisk may either be a ZrO emission band or atomic emission from an unidentified impurity. The feature at ∼597 nm marked with an asterisk is likely from SrO as it persists at late times and appears readily in the ablation of SrO (Supporting Information), though it lies close to a strong Sr atomic emission line. At 3050 ns postablation, ZrO emission bands are clearly evident, which give way to the prominent SrO “orange band” at later times. Highly intense sodium D1 and D2 lines (589.6 and 589.0 nm, respectively) persist at 9050 ns.

positioned at 574.84 nm.25 The second spectrum in Figure 3 (blue) shows the Zr18O c3Π2−a3Δ3 band, with the (1,1) band notably shifted by 0.15 nm relative to the Zr16O band and is positioned at 574.69 nm. The bottom four spectra in Figure 3 were acquired from ablation of the SrZrO3 target in the presence of 20% 18O2 and 80% N2, to simulate atmospheric composition making these spectra directly comparable to the spectra in Figure 1. Given that the oxygen in the target is predominantly 16O, the ingrowth of the Zr18O c3Π2−a3Δ3 (1,1) band relative to the Zr16O c3Π2−a3Δ3 (1,1) band was used to track the influx of ambient oxygen into the ablation plume. As shown in Figure 4, at early times (300 ns) after the ablation of the SrZrO3 target, the emission spectrum shows clear atomic emission lines, as well as a prominent peak at 574.84 nm, indicative of Zr16O formation. Between 1 and 3 μs, there is notable in-growth of the Zr18O band at 574.69 nm. By roughly 3 μs, the Zr18O c3Π2−a3Δ3 (1,1) band is the dominant band, although there is still some contribution from the Zr16O peak at 574.48 nm. These results are of interest to our understanding of oxidation rates in ablation plumes (see Discussion). Time-Resolved Plume Imaging. Plume images are presented as a time-series from 50 to 10050 ns (Figure 4). The left side of each frame shows the ablation plume with no bandpass filter, and the right side shows superimposed plume images acquired through the 570 nm (blue) and 600 nm (red) narrow bandpass filters, to capture emission from the c3Π2−

emission intensity at each delay time was calculated as the integral of emission intensity in each transition region, divided by the interval of integration. Further, the average emission intensity of the SrO orange band region was corrected for contribution from Zr-related emission features (see Supporting Information). Using this approach, we observe that ZrO emission is most intense between approximately 2 and 5 μs after the ablation pulse, while SrO emission becomes most intense about 6 μs after the ablation pulse (Figure 2). At ∼9 μs, ZrO emission from the c3Π2−a3Δ3 (0,0) transition region is virtually gone, while SrO emission is observable until about 20 μs. Isotope Substitution in an 18O2 Atmosphere. To monitor the influx of ambient gas into the ablation plume, we acquired emission spectra from ablation of the SrZrO3 target (which contains predominantly 16O) in an 18O enriched ambient gas. To do this, we first measured the isotopic shift between the Zr16O and Zr18O c3Π2−a3Δ3 (1,1) band, which was determined from the ablation of a ZrO2 target in the presence of 16O2 and 18O2. The top spectrum in Figure 3 (red) shows the Zr16O c3Π2−a3Δ3 band, with the (1,1) band 1586

DOI: 10.1021/acs.jpca.7b11994 J. Phys. Chem. A 2018, 122, 1584−1591

Article

The Journal of Physical Chemistry A

a3Δ3 ZrO band and the SrO orange band, respectively. Each frame in Figure 4 displays imaging acquired from a single ablation, each acquired from a new position on the target (∼1 mm spacing), with their intercomparison assuming shot-to-shot variation is minimal for identical acquisition parameters. As seen in the left-hand panes of Figure 4 (no bandpass filter), the plume rapidly expands hemispherically from the target surface. The region of highest emission intensity follows the leading edge of the plume as it expands into the sample chamber. At later times (5−10 μs), the unfiltered emission intensity distribution is horseshoe-shaped, delineating the periphery of the plume, though emission is still evident volumetrically even at ∼10 μs. The emission imaged through the 570 nm bandpass filter (blue) at early and intermediate times (1−3 μs) appears volumetric in the expanding plume, with a distinct gradient from the interior to the leading edge of the plume. However, atomic emission lines are present in this region (Figure 1), and may contribute to imaged emission in this spectral region particularly at early times. At later times (5−7 μs), the emission is confined to the leading edge of the plume, and by 10 μs it is no longer clearly visible. The emission imaged through the 600 nm bandpass filter (red) is concentrated near the center of the plume in the 1050 and 3050 ns frames. At roughly 5 μs, the emission observed through this filter begins to show the highest intensity along the leading edge, approximately colocated with the intensity distribution observed through the 570 nm bandpass filter. At 10060 ns, this distribution persists and is co-located with the region of highest intensity emission observed in the left-hand images (no bandpass filter). Model Results. To assist our interpretation of the Sr and Zr oxidation chemistry data, we developed a simple chemical kinetics model to describe reversible reaction rates between oxygen and strontium and between oxygen and zirconium (Table 1). Forward reaction rates were acquired from prior experimental studies26,27 or from first-order transition state theory (TST) calculations,28 while reverse reaction rates were calculated using the principle of detailed balance using thermodynamic information retrieved from the NIST-JANAF

Figure 3. Top two spectra were acquired ∼4 μs post-ablation of a ZrO2 target in pure 16O2 (red) and pure 18O2 (blue) environments, illustrating the isotopic shift of the ZrO vibronic band at approximately 574.75 nm. The bottom four spectra were acquired 300, 1050, 2050, and 3050 ns post-ablation of a SrZrO3 target in a 20% 18O2, 80% N2 environment. At early times post-ablation Zr16O is the dominant form, indicating reaction with oxygen from the target, while at later times (∼3 μs) Zr18O is the dominant form, indicating infiltration of atmosphere 18O2 into the ablation plume. Tick marks are provided on the ordinate, for comparison of the relative intensities of features within individual spectra.

Figure 4. Laser ablation plume imaging as a function of time post-ablation. The left-hand images show the plume growth through a 0.01 transmission neutral density filter. The right-hand images show the plume growth through two narrow bandpass filters (10 nm fwhm) centered at 570 and 600 nm, selected to image emission from the ZrO c3Π2−a3Δ3 band (∼571−573 nm) and SrO orange band system (∼592−600 nm), respectively. At early times, ZrO emission dominates the images and appears to be present volumetrically in the ablation plume. After 5 μs, ZrO and SrO emission is concentrated at the edge of the plume, while at late times (>10 μs), SrO emission dominates the image around the edge of the ablation plume. 1587

DOI: 10.1021/acs.jpca.7b11994 J. Phys. Chem. A 2018, 122, 1584−1591

Article

The Journal of Physical Chemistry A Table 1. Reactions and Rate Coefficient for the Sr + Zr + O System no.

reaction

1 2 3 4 5

Sr + O2 ⇌ SrO + O Zr + O2 ⇌ ZrO + O Zr + O2 ⇌ ZrO2 SrO ⇌ Sr + O ZrO ⇌ Zr + O

rate coefficient 3.2 1.3 1.3 1.0 1.0

× × × × ×

10−10 10−10 10−10 10−15 10−15

exp(−7192.3/Tg) exp(−745.7/Tg) exp(−745.7/Tg) exp(−53543/Tg) exp(−99777/Tg)

thermochemical tables.29 The ablation plume temperature was derived from Zr atomic emission intensities measured between ∼0−5 μs, which was used as the time-dependent model temperature input (∼8000 K cooling to ∼5500 K). Initial number densities of Sr, Zr, and O for the model were (3.0 ± 1.5) × 1019/cm3 for Sr and Zr, and (9.0 ± 3.0) × 1019/cm3 for O, estimated from the mass excavated from the target, assuming complete dissociation. The number density uncertainties had 1 μs (when there was appreciable ZrO formation), and negligible effects on SrO formation at any time. See Supporting Information for more details on temperature calculations and post-ablation target profilometry. In Figure 5, we show the modeled change in number density of ZrO, SrO, and ZrO2 as a function of time, as determined from the simple reaction model (note the change in scale between the ZrO plot, and the SrO and ZrO2 plots). The solid lines represent the evolution of the number density assuming an oxygen-rich environment (9 × 1019 O atoms/cm3 from the sample; 2.5 × 1019 O2 molecules/cm3 from the ambient gas, assuming atmospheric O2 concentration at 298.15 K) and the dashed lines represent the evolution of the number density for each species assuming an oxygen-poor scenario (9 × 1019 O atoms/cm3, assuming the only oxygen available is from the target itself). In both scenarios, the rate of ZrO production is faster than that of SrO, as anticipated from the experimental data. In the oxygen-rich scenario, the ZrO number density sharply increases between 1 and 2 μs, while in the oxygen-poor scenario, it increases more gradually, between 1 and 3 μs. For SrO, the number density reaches a maximum between 4 and 5 μs in both scenarios, although it is suppressed by roughly 20% in the oxygen-limited scenario. The maximum ZrO and SrO number densities modeled here are separated by approximately 2 μs in both scenarios. While the production of ZrO2 follows a similar time-scale as SrO, its production rate is notably faster, particularly in the oxygen-rich scenario. In the oxygen-limited scenario, the absolute ZrO2 production is modeled to be suppressed by roughly 35%.

Figure 5. 0-D reaction modeling of Zr, Sr, and O post-ablation, assuming full dissociation. Model parameters were adjusted to illustrate molecular formation in an oxygen-rich environment, and an environment where the only oxygen available came from the SrZrO3 sample itself to illustrate the effect of sample oxygen contribution. As is consistent with our emission data, ZrO forms more rapidly than SrO at early times, with a substantial time-offset in the formation of both molecules in an oxygen-limited environment. Note the change in scale between ZrO, SrO and ZrO2.

ZrO2) will also form more readily at higher temperatures than SrO. Modeling the chemical kinetics of the vapor phase reactions offers additional insight into the oxidation sequence. In Figure 5, the modeled ZrO number density reaches a plateau roughly 2−3 μs before SrO, consistent with the experimental variation in ZrO and SrO emission intensities shown in Figure 2. The offset in absolute time of about 2 μs between the experimental and modeling results is likely due to transport effects that are not included in the 0-D kinetics model. Between 4 and 6 μs, the experimental data show a sharp decline in ZrO emission, concurrent with a rise in SrO emission (Figure 2). Model results suggest ZrO2 forms more slowly than ZrO (Figure 5) and is still increasing in number density 5 μs post-ablation, after the ZrO number density has plateaued. While this decline in ZrO emission may also be controlled by the fluorescence lifetime for the c3Π2−a3Δ3 transition of ZrO, modeled variation of the fluorescence lifetime was not found to have a significant effect. It is further noted that output from the kinetic model

4. DISCUSSION The variations in ZrO and SrO emission intensities shown in Figure 2 imply a distinct sequence of molecular formation during expansion and cooling of the ablation plume. Dissociated Zr atoms start to combine with oxygen to form ZrO first, with a peak emission intensity at about 4 μs. In comparison, Sr atoms combine with oxygen at later times achieving a maximum intensity roughly 2 μs later than that of ZrO. The observed sequence of formation is consistent with the thermodynamic properties of the metal oxides. ZrO has a bond dissociation energy of 753 kJ/mol30 as compared to 470 kJ/mol for SrO;31 we therefore expect ZrO (an intermediate to 1588

DOI: 10.1021/acs.jpca.7b11994 J. Phys. Chem. A 2018, 122, 1584−1591

Article

The Journal of Physical Chemistry A

combination of emission spectroscopy and reaction modeling, we have begun to elucidate high-temperature molecular formation sequences. By being able to predict the resultant chemistry in laser ablation plumes, we can begin to model elemental fractionation in LA-ICP-MS, and the conditions which control stoichiometric thin film production. It is noteworthy that the molecular formation sequence observed in these laser ablation plumes have some striking similarities to the condensation patterns in a nuclear fireball. Early work by Freiling et al.34,35 describes the role of oxide vapor pressures in determining the distribution of fission products between the vapor phase and condensed phase in cooling fireballs. At a given temperature, differences in the vapor pressures provide an index of thermodynamic volatility that can be used to predict chemical fractionation patterns in fallout. For example, at 2500 K and 1 atm O2, the vapor pressure of SrO is more than 3 orders of magnitude higher than that of ZrO2 (increasing to more than 4 orders of magnitude at 1500 K),36 and ZrO2 is predicted to condense before SrO. This general pattern is consistent with our emission experiments and modeling, where we observe that ZrO (a ZrO2 precursor) forms sooner and much more readily. This suggests that laser ablation experiments could be used to investigate the relative condensation histories of metal oxides for materials with homogeneous starting compositions.

matched well with chemical equilibrium model output (see Supporting Information). We further used the kinetics model to compare the rate of oxide formation for two scenarios: (1) ablation in an oxygenrich environment and (2) ablation in an inert atmosphere. In the latter case, oxygen is only available from the sample, and the rate of production for ZrO, SrO, and ZrO2 are all suppressed compared to ablation in air, particularly for ZrO2. From this, we infer that the rate of oxidation, and ultimately the composition of the condensing vapor component, will change depending on the availability of oxygen from the target and the ambient gas. In Figure 3, the emission spectra we acquired from ablation of the SrZrO3 target in the presence of a 20% 18O2 ambient gas show that there is, in fact, interaction with both target oxygen and the ambient gas. This observation is consistent with another isotopic substitution study, which shows late molecular formation from interaction with ambient gas constituents.32 From our isotopic substitution experiments, we found that the influx of the ambient gas occurs between 1 and 3 μs, and that interactions with the ambient oxygen surpass that of the sample oxygen after 3 μs. This effect may be related to shockwave dissipation; it has previously been shown that the post-ablation shockwave may prevent ambient oxygen from entering the plume initially.18 The sequence of molecular formation and ambient oxygen influx we observed in the spectroscopic studies is substantiated by plume imaging through bandpass filters (Figure 4). At early times (1−3 μs), oxidation reactions due to mixing with oxygen from the surrounding gas are confined to the margins of the expanding plume. In comparison, oxygen derived from the sample is volumetrically distributed throughout the plume and available to react once the plume has cooled to temperatures favorable for chemical bonds to form. The fact that ZrO emission is volumetrically distributed throughout the plume during early times suggests the matrix was an important source of oxygen during this time, though it should be noted that there is non-negligible contribution from atomic emission in this spectral region at early times (5 μs), the absence of a volumetric emission contribution in the plume images (Figure 4) is consistent with the consumption of oxygen that originated in the target, while the ambient oxygen influx at the leading edge allows for oxidation of Sr and Zr in this region. Further, the presence of SrO emission around the periphery of the plume persisting at very late times (>10 μs) is consistent with a recent flame emission spectroscopy experiment, where SrO emission (at >700 nm) was observed at temperatures