Plume dynamics and gas-phase molecular formation in transient laser-produced uranium plasmas

Reactive processes^1,2 which may occur in laser-produced plasmas (LPPs) are of general interest for understanding combustion physics^3,4 and general laser ablation (LA) physics, particularly those related to nanoparticle formation and kinetics,⁵ pulsed laser deposition,⁶ and analytical techniques such as laser-induced breakdown spectroscopy (LIBS), laser-induced fluorescence (LIF), and LA-inductively coupled plasma mass spectrometry (LA-ICP-MS).^7–12 In particular, uranium (U)-containing LPPs are of current interest in nuclear safeguards, security, and nonproliferation.^13–17 U-containing LPPs are suitable for rapid, in situ, all-optical, spectroscopic material characterization using well-established techniques including LIBS^18,19 and LIF. LPPs also provide laboratory-scale surrogates for studying fireballs and high-explosive detonation.^16,20

At the earliest times $(≲1 μs)$ in the evolution of the LPP, ions and atoms dominate, while molecules generally form at later times $(≳10 μs)$⁠, when the plasma is cooler due to atomic collisions with ambient gas particles and recombination.^21,22 In the later phase of the plasma lifecycle (∼ms), aerosols and nanoclusters form through a nucleation-condensation process.²³ Figure 1 summarizes the phases that occur during LA and plasma evolution with their approximate timescales, while also highlighting the most influential plasma parameters and environmental conditions. We note that the time-scales presented in Fig. 1 are generalized and may vary significantly depending on the ablation laser parameters as well as the environment.²⁴ Although these aerosols and clusters are routinely produced in LA applications, the dynamics and conditions in which they exist are not fully understood. The formation of particles and agglomerates in the LPP is influenced by internal plasma properties, temperature, and number density of atoms and molecules, but also by external conditions such as pressure and the nature of the ambient gas.

Several prior studies investigated the early-time plasma characteristics, emission features, plasma properties, and late-time particle nucleation and growth by ex-situ measurements, as discussed in the recent review by Kim et al.²⁵ In the present work, we focus on understanding the intermediate stages of U-containing nanosecond-LPP evolution (100 ns–100 μs), which may eventually lead to agglomeration and the growth and evolution of particulates. Optical emission spectroscopy (OES) is an attractive technique for high-fidelity, nonperturbing investigation of LPPs. OES of U-containing LPPs provides a practical approach for real-time, in situ, remote detection and characterization of U. However, the interpretation of OES data is often cumbersome. Numerous atomic and ionic U signatures congest the emission spectrum of U-containing LPPs. For example, the UV-visible spectral region contains nearly 100 000 atomic (U I) and ionic (U II) transitions from ∼1600 energy levels.²⁶ As a result, OES of U requires high-resolution spectroscopic instrumentation to differentiate among the fine U emission features. The task of OES data analysis is further complicated by the formation of gas-phase U oxides in oxygen-containing environments. Since the U oxide molecular emission is composed of fine structures due to a high density of states,²⁷ the resulting spectral features may overlap even when high-resolution spectroscopic instrumentation is used $(≲10 pm)$⁠. Line discrimination and identification are even more difficult at spectral resolutions $≳10 pm$ where a single broad peak may actually consist of multiple overlapping but unresolved atomic and molecular lines.

Detailed spectroscopic studies have been reported on uranium monoxide (UO) emission measured by LIF of supercooled, supersonically expanding laser-produced U plasmas at low temperatures (∼130 K) and from UO thermal emission in a high-temperature furnace environment.^28,29 These studies have identified numerous electronic transitions of UO in the UV and visible spectral regions. Despite this prior work, almost all emission studies in LA plumes focus on the emission band around 593.55 nm.^14–16 Until recently, other UO features in the visible region were not well studied,³⁰ and features from U species in higher oxidation states (U_xO_y, e.g., UO₂, UO₃, U₂O₂) are still largely unknown.³¹ 

LPPs exhibit steep temperature and particle density distributions which vary in space and time, incident laser parameters (e.g., laser pulse duration, wavelength, energy), and ambient environment.^32–34 Understanding of the plume dynamics in LPP is essential for all LA applications; it is well documented that the hydrodynamics of the plume is strongly influenced by the laser parameters and environment. However, few studies to date investigated the time-dependent morphology of LPPs comprised of heavier (high-Z) atomic constituents such as U. Since the plume dynamics largely depends on particle inertia, the greater atomic mass of U in these LPPs may result in interesting morphological behavior. Instabilities can form in the plume front when a flowing LPP interacts with a steady-state ambient medium. For example, several studies^35,36 reported plume features reminiscent of the Rayleigh-Taylor instability that occurs on the interface between two fluids, when the lighter fluid pushes on the heavier fluid. Hence, such oscillatory striations in the interface may be prominent when an LPP comprised of heavy particles, such as U species and compounds, interacts with an ambient composed of lighter particles (e.g., nitrogen and argon). Additionally, the molecules in LPPs are formed via various temperature- and density-dependent processes such as combustion through reactions with ambient oxygen, recombination with other elements present in the matrix in the gas phase, and fragmentation/dissociation of larger clusters/molecules.^21,22 The morphology of U-containing LPPs is especially interesting considering the strong affinity of U for oxygen species and likely combustion.¹ Redistribution of species in the plume via morphological expansion and confinement in the presence of an ambient gas affects the rates at which chemical reactions occur in the plasma.²² Although recent kinetic simulations have investigated the conditions in which U and uranium oxide species exist in nanosecond-¹ and femtosecond-LPPs,³⁷ the reaction pathways for the formation of uranium compounds in LPPs have not been validated through experiments. No systematic research effort has been committed to explore the influence of the particle dynamics and plume morphology on the spatial extent of uranium oxide formation.

We perform comprehensive studies of plume morphology and space- and time-resolved species tracking using OES in various environments in order to gain insight into the formation and dissociation of U compounds (oxides) in LPP. Comparing dynamics of atomic and molecular species in various ambient gases (inert and reactive) shows the contribution of gas-phase chemistry to plume hydrodynamics. We find that the presence of oxygen in the environment causes the formation of U oxides in the plasma and influences the rate at which the plume expands. A systematic study in which the ambient oxygen concentration in a background of inert argon is varied was conducted, and it reveals competing reaction pathways which favor the formation of UO at lower oxygen partial pressures and the formation of higher uranium oxide states at greater oxygen partial pressures.

Figure 2 shows the experimental setup used in this study. We employed a Nd:YAG (1064 nm, 6 ns) laser focused to a spot diameter (1/e²) of ∼1 mm at normal incidence to the target. The focusing lens had a 15-cm focal length. The laser fluence is estimated to be approximately 10 J cm⁻² on the target surface. The target was a U metal plate (natural isotopic concentration with dimensions 1.0 × 1.0 × 0.1 cm³) and was enclosed in a low-pressure vessel with a base pressure of ∼5 mTorr. The target was translated periodically to prevent excessive pitting by the laser. Prior to each measurement series, the U oxide layer on the surface was removed by several laser cleaning shots. LPP emission was imaged through the quartz window of the chamber with a collection telescope (1.35× magnification) onto a 0.5-m Czerny-Turner spectrograph (2400 lines/mm grating, ∼0.04 nm resolution) coupled to an intensified charge-coupled device (ICCD, minimum gate width ∼3 ns) and photomultiplier tube (PMT, rise time of ∼2 ns) coupled to a fast oscilloscope (50 Ω termination, 1 GHz). The broadband, low-resolution spectra were obtained using the 300-lines/mm grating (∼0.3-nm resolution). A diverting mirror within the spectrograph was used to select one of the two detectors. Measurements were synchronized to the laser pulse using the digital delay generator integrated with the ICCD. We conducted experiments at 100-Torr pressure with different fill gases (air, nitrogen, argon) as well as air at low pressure (∼5 mTorr). The 100-Torr pressure was experimentally determined to simultaneously yield the emission features of U and its oxides in the presence of oxygen. We performed further studies to investigate the effects of oxygen on U and U oxide signatures by varying the partial pressure of oxygen (99.99% pure) in a background of argon (99.99% pure). Oxygen concentrations were varied from ∼3.2 × 10¹³ (5 mTorr) to 7.2 × 10¹⁷ cm⁻³. We note that because we did not continuously flow the fill gas through the experimental chamber, we cannot predict with certainty the oxygen concentration in “pure” gas or low pressure conditions. We expect the ambient oxygen concentration to increase slowly from both target ablation and leaks in the chamber.

Spectroscopic measurements were performed with an entrance slit width of 30 μm and averaged over ten laser shots. PMT measurements used an entrance slit width of 50 μm and exit aperture width of 100 μm and were averaged over 16 laser shots. Horizontal translation of the imaging lens(es) allowed for scanning of the plasma image orthogonal to the slit. The vertically oriented slit aperture size was ∼2 mm and was centered at the laser axis for both spectroscopy and PMT measurements. The spatial resolution for spectroscopy is determined by the magnification of the plasma image (1.35×) and the entrance and exit apertures of the spectrometer (entrance slit 30 μm, ICCD pixel size 13 μm). Consequently, we expect the entrance slit to limit spatial resolution to ∼22 μm, which is greater than the spatial resolution of the ICCD. Similarly, the spatial resolution for PMT measurements is determined by the image magnification as well as the smaller of the two entrance and exit aperture sizes. The limiting spatial resolution for PMT measurements is estimated to be ∼37 μm limited by the entrance slit of the spectrometer. Imaging of the LPP emission was performed using a variable-magnification telescope (objective) coupled to the ICCD positioned at an angle of 90° with respect to the laser axis. Spectrally resolved images of the plasma emission along the laser axis were recorded using a Dove prism to rotate the image by 90°, so that the laser axis was aligned to the vertical slit.

We employed fast-gated imaging as well as time- and space-resolved emission spectroscopy in order to investigate the characteristic morphological behavior of the high-Z LPP as well as potential gas-phase U chemistry enabled by the presence of reactive oxygen in the background gas.

In order to gain insight into the collisional processes and morphological changes mediated by chemical reactions which may occur in the U LPP, we recorded time-resolved, single-shot, side-on images of the plasma emission using a gated ICCD at various pressures and with several fill gases. Figure 3 shows the LPP in air at low pressure (5 mTorr), 100 Torr nitrogen, 100 Torr air, and 100 Torr argon. Each image was recorded from a separate laser shot and is normalized to its own maximum intensity. We select three parameters to characterize an LPP image: (i) axial plume length, (ii) shape, and (iii) emission intensity distribution. We expect plume oscillations at 100 Torr to be caused by instabilities (Raleigh-Taylor, Richtmyer-Meshkov, etc.) when the expanding plume interacts with steady state ambient medium and partly due to plasma chemistry (in air). In such a scenario, a small change in plume temperature, pressure, and/or density may affect the structures seen in the plume periphery. However, by using a stationary fill gas and periodic sample translation to prevent excessive drilling, we provide the conditions to ensure minimal shot-to-shot fluctuations in the primary plume shape. The maximum variation which was observed in the plume length (along the laser axis) was ∼6% between five laser shots. The fluctuation of the maximum intensity in the image was observed to be the greatest ∼8% at the earliest times (<1 μs), similarly for five laser shots. At late times (∼8–10 μs), the fluctuation was observed to be ∼5%. The plasma expands rapidly with negligible evidence for confinement and collisional interactions at the low background gas pressure (5 mTorr), whereas the 100 Torr cases show evidence of confinement. The plume length is similar for all environments up to about 100 ns (Fig. 4). Beyond this point, the low-pressure case exhibits continued adiabatic expansion. The shape of the LPP in argon is noticeably different from that for air or nitrogen. Plume splitting, which is evident in air and nitrogen as shown more clearly by the intensity profile taken from the center of the laser axis in Fig. 4(a), is less pronounced in the case of argon. Figures 4(b)–4(d) show the length of the plasma along the laser axis as a function of time measured from the plume images. The plume length was measured from the target plane, which we determined from a reference image, to 5% of the maximum emission in the image along the laser axis. We fit the low-pressure case with a free-expansion linear model L ∝ t, where L is the plume length and t is time;³⁵ the model fits the data well with an adjusted R² = 0.957.

The visible emission of the plasma expands beyond 10 mm from the target in the case of low pressure. The 100 Torr cases show slowed expansion to a plume size of ∼8 mm in nitrogen and argon and ∼7 mm in air. The shock barrier weakens toward later times in the evolution of the LPP, causing an increased rate of chemical reaction because ambient species penetrate the plume more easily, for example between uranium and oxygen species as hinted by the apparent slowed expansion observed beyond ∼5 μs in air in Fig. 4. Plume expansion in the presence of an ambient gas is governed by collisional interactions between the plume and ambient species. At early times, the plume expands with the shockwave, and the expansion can be described by the Sedov blast model L ∝ t^0.4. At later times, after the plume detaches from the shock and plume-ambient collisions govern plume deceleration, expansion is better described by the drag model³⁵ L ∝ (1 − e^−βt). We fit the expansion data with both blast and drag models in Fig. 4(b) in order to estimate the type of expansion for each case. Early-time expansion in nitrogen and argon matches the blast model. However, the data deviate from both the blast and drag models at later times, exhibiting deceleration characteristics of neither a shock expansion nor a purely collisional deceleration. In air, early-time data deviate significantly from the blast model, while late-time data more closely follow the drag model. The expansion trends in nitrogen and argon suggest an interplay between different plume deceleration mechanisms at late times. The expansion trend in air suggests that there is a significant contribution of collisional interactions to the slowing of the plume at late times, and that the early-time behavior does not correspond to expansion with a shock. We note that quenching of the optical emission may be prevalent, especially in the case of air where inelastic U-O reactions are probable. Consequently, the insights on morphological behavior should be regarded with care considering the limitation of probing only the emitting excited state populations. Additional diagnostics, such as absorption-based methods, are necessary in order to more appropriately probe ground state populations in the plume.

We next explore the broad visible emission spectrum of the plasma in order to further investigate the interplay between plume hydrodynamics and plasma chemistry. Figure 5 shows the typical emission spectra in the visible wavelengths for low ambinent pressure, and higher pressures of nitrogen, air, and argon. The spectra were recorded at a delay of 1 μs with an integration time of 1 μs for 5-mTorr and 200 ns for 100-Torr background pressures, respectively. The strongest emission lines in the spectrum are the U I resonance lines at 436.20, 556.41, and 591.54 nm. We observe similar effects of confinement and increased collisional excitation evidenced by increased background and atomic line intensity for the larger, 100-Torr, pressures as compared to the 5-mTorr case in Fig. 5. We similarly observe an increased background in the case of air, which is different from both pure nitrogen and argon. Argon and nitrogen promote collisional excitation and plasma confinement with minimal plasma chemistry. In contrast, we expect the air background to enable gas-phase chemical reactions between U species from the target and oxygen from the environment. The continuum radiation (bremsstrahlung and recombination) from LPP is more pronounced at early times (<500 ns) and is not expected to contribute to the observed spectral features. Therefore, overlapping, unresolved spectral features from U oxides may represent a significant contribution to the background that is seen in the air spectrum when compared to the spectra recorded in nitrogen and argon background.^30,31 The difference in the background between air and the other gases provides further evidence for U combustion/oxide formation in the LPP via interactions with oxygen from the environment. From an analytical point of view, the overlapping lines may act as a source of background, severely limiting the diagnostic capabilities in U plasmas unless very high spectral resolution measurements are performed.

The morphological behavior of the U LPP in various gases as well as the emission spectra indicates improved plasma uniformity at low pressures and in argon, plume splitting at higher pressures in nitrogen and air, and additional chemical interactions in air. More detailed conclusions could be derived from selective imaging of individual plume species. We use a PMT to record the time profile of emission in a narrow spectral window selected by the rotation of the spectrometer grating. The spectral range observed by the PMT for U I and UO features is ∼0.05 nm (measured with 2400 l/mm grating). Figure 6 shows sample PMT profiles for a selected atomic U transition (U I 591.54 nm) at several distances from the target plane. PMT profiles at several distances are combined to form the contour maps for atomic species in the plasma and are shown in Fig. 6. By adopting this approach, both the spatial and temporal distributions of various species within the LPP are obtained. For the low-pressure case, we observe atomic uranium (U I 591.54 nm) in the LPP, as shown in Figs. 6(a) and 6(b). Negligible confinement in the low-pressure case allows atomic species to expand freely with little evidence of confinement. We show emission contours for the U I species also for the 100 Torr cases in Figs. 6(c)–6(e). The emission contours in nitrogen and air show evidence of plume splitting, similar to that observed in Fig. 3, whereas the emission contours in argon show a more uniform expansion of the U species in the plasma. The most evident difference between the contours in various gases is the short persistence of U I emission in air compared to both argon and nitrogen. This presents evidence that the presence of reactive species, of which the most prominent is oxygen, causes a more rapid depletion of the excited state U I population. Previous studies^34,38 that employed LIF and laser absorption spectroscopy, which probes the ground state U population, showed similar quenching of atomic U emission from the plasma generated from a low-concentration U sample (1.3% UO₂ in glass matrix) in oxygen-containing environments due to reaction pathways leading to the formation of U oxides. To further improve the understanding of these phenomena, we performed a systematic study in which we varied the oxygen concentration in a background of pure argon gas in order to isolate the interactions between U from the plasma and oxygen from the environment, focusing specifically on the spatiotemporal behavior of U and its oxides in the plasma.

We first investigate the consumption of atomic U species via reactions with oxygen leading to the formation of UO. We begin with higher-resolution spectroscopic measurements in order to identify the signatures of both atomic U and UO with a controlled 1% oxygen partial pressure at a total pressure of 100 Torr, with the background gas being inert argon. Spectrally resolved imaging of the LPP provides simultaneous information on the location of both U I and UO species, while time-resolved spectroscopy yields the time-dependent appearance of UO. Figure 7 shows the space- and time-dependent emission features of atomic U and UO. The prominent U I feature in this spectrum originates from the ground-state transition at 591.54 nm; the associated broad UO feature is centered at 593.55 nm. Figure 7(a) shows a spectrally resolved image of the plasma emission along the laser axis. A Dove prism was used to align the laser axis with the vertical slit, such that the vertical axis of the camera corresponds to a distance from the target. The image was time-integrated for 20 μs at a delay of 20 μs with respect to the incident laser pulse, so that simultaneous, strong emissions from both species can be observed. The UO emission is significant even with the small concentration of oxygen (∼10¹⁶ cm⁻³) in the environment. We observe no evidence for segregation of U I from UO, as seen by similar spatial distributions along the laser axis in Fig. 7(a). Similar species distribution implies that the forward and reverse reaction pathways between U and UO occur at similar rates during this late period of plasma plume lifecycle.

We perform time-resolved spectroscopy in the same spectral window. Figure 7(b) shows a series of time-resolved spectra at a distance of 3 mm from the target. The delayed appearance (after ∼10 μs) of the UO feature in Fig. 7(b) implies that the UO compound forms later in the lifetime of the plasma, when the plasma is colder, and primarily originates from reactions in the plasma. We perform species-, space-, and time-resolved tracking of U and UO with increasing oxygen concentration to further investigate the U combustion pathways leading to oxide formation in the plume.

Figure 8 shows contour maps for the emission of U I and UO species in the LPP for various oxygen partial pressures. For the pure argon case with trace oxygen (concentration ∼3.2 × 10¹³ cm⁻³), we observe long-lived emission from both U and UO species. Similarly as observed in Fig. 6, the expansion in argon leads to a relatively uniform species distribution, with no evidence for plume splitting or species segregation. These results indicate that a trace amount of oxygen is sufficient for gas-phase oxide formation. A notable observation is that increasing the oxygen concentration has a similar quenching effect on the persistence of emission from both U and UO. The lifetime of the emission from both species decreases nearly fourfold in 18% oxygen when compared to control (pure argon) environment. The decreasing persistence of emission from both species implies that both U and UO react similarly with ambient oxygen, depleting their excited state populations in the presence of oxygen at similar rates.

The rate of depletion of both U and UO emission that increases with oxygen concentration points to a possible mechanism, where higher oxidation state U species form (e.g., UO₂, UO₃, U₂O₂) at higher rates at higher ambient oxygen concentrations. We therefore investigate the emission spectrum for occurrence of features characteristic of higher U oxide species. In previous work, we identified the emission feature between 320 and 380 nm and suggested that it may originate from one or more higher oxide species³¹ due to its late-time prominence far from the target, as well as by comparison with the known electronic spectrum of uranium dioxide.^39–41 Representative spectra are shown in Fig. 9 at various oxygen partial pressures. Greater oxygen partial pressures (≥1%) diminish the fine atomic and ionic signatures, leaving prominent broad features corresponding to U oxides. Although early-time micrometer-scale ejecta may form from subsurface boiling, recoil ejection and exfoliation mechanisms⁴² could contribute to the observed broad emission via graybody radiation, we see no evidence of such macroscopic particulates in the plasma images. Additionally, the ejection and early-time formation of nanometer-scale particles, which may also emit broadband radiation, is also feasible.⁴³ However, further work should be performed in order to investigate early-time ejecta from targets composed of high-Z constituents, such as U. A plausible noninvasive technique for recording early-time micron-scale ejecta is shadowgraphy; all-optical Rayleigh scatter has also been previously proposed to observe nanoparticles.⁴⁴ Figure 10 compares the confluent decreasing persistence of U and UO with the increasing prominence of the potential higher U oxide feature. The persistence of U and UO emission shown in Figs. 10(a) and 10(b) is derived from Fig. 8. The persistence is measured up to the point in time (at any position) at which the emission decreases to 5% of the maximum in the contour. The persistence of atomic U decreases monotonically with increasing oxygen partial pressure; however, the persistence of UO peaks at a particular oxygen concentration and then begins to decrease along with atomic U persistence. Simultaneously, the prominence of the potential higher oxide feature increases. We define the prominence to be the ratio of the area beneath the spectral feature between 310 and 380 nm to the area under the spectrum between 310 and 640 nm, as shown in the inset of Fig. 10(c). The monotonic decrease in U persistence and the optimal condition for UO persistence suggest the existence of competing reaction pathways for varied oxygen concentrations, leading to the formation of different U oxide states. Lower oxygen concentrations promote reaction pathways which consume atomic U to form UO, while greater concentrations promote formation of higher oxides.

Although the expansion of the plume is highly dynamic in low pressure conditions, the expansion becomes increasingly complex in the presence of an ambient gas.^36,45 We observe the expected free expansion for the 5 mTorr case in Figs. 3, 4, and 6(b). The plume expands according to a linear free expansion model³⁵ as shown in Fig. 4(b). Previous studies showed that a transition regime from collisionless to collisional exists when the background gas pressure reaches ∼100 mTorr, indicated by enhanced emission from all species in the plume, as well as the generation of a sharp boundary between the plume and the background medium.³⁵ In order to estimate the average velocities of atomic U species, we determined the time at which the maximum intensity occurred in the profile measured by the PMT at a particular distance from the target. We then averaged the instantaneous velocities to determine the given particle velocities. The U atoms at 5 mTorr move with an average velocity of ∼1.1 × 10⁶ cm s⁻¹, which corresponds to a kinetic energy of 150 eV. At 100-Torr pressure level, the confinement of the plume is evident even at early times (∼100 ns), as shown in Figs. 3 and 4(b). The average velocity of U atoms is slower in the 100-Torr cases, down to ∼1.3 × 10⁵ cm s⁻¹ in air.

Larger background pressures typically exhibit interesting morphological behavior such as plume splitting and confinement caused by collisions between plasma species and the ambient. High-Z (tungsten) LPPs exhibit an increasingly complex turbulent structure with increasing background pressure, including plume splitting and plume sharpening. Sharpening is caused by a greater kinetic energy of target species in the target normal direction.³⁵ Plume sharpening is typically observed for pressures $≲1 Torr$ and is evidenced by a conical plume front in which the energetic species are ejected primarily in the normal direction from the target. We observe notable plume splitting with no evidence for sharpening with the U LPP in 100 Torr nitrogen, air, and mixed oxygen-argon environments. Plume splitting is influenced by the interaction between the expanding plasma and steady-state background gas. The strength of the shock front which forms at the plasma-ambient interface depends strongly on the number density of ambient gas species. Previous studies highlighted that the shock front persists for a very long time (>20 μs) for LA in atmospheric pressure conditions.⁴⁶ The strong shock waves may act as a barrier mitigating interactions between plasma and ambient species at earlier times;²¹ then weakening regions in the shock may allow ambient species to penetrate the plasma at later times affecting the morphology of the plume and perhaps causing splitting.

The presence of reactive species, namely oxygen, in the background gas enhances plasma-assisted chemical reactions,^47,48 which deplete the populations of excited and lower-energy state atoms in the plasma plume,^22,34 affecting the temperature and number densities in the plume. Buckley⁴⁹ reported that the addition of oxygen has a profound effect in reducing the emission of atomic lead. However, their results also showed different signal trends for atomic chromium and beryllium line emissions in comparison with lead. Gleason and Hahn⁵⁰ reported a dramatic reduction in the signal from a mercury plasma in air compared to a pure nitrogen background, which is similar to the results observed in the lead plasma by Buckley.⁴⁹ These previous results highlight that the oxygen-initiated plasma chemistry and its associated reduction of the excited state population depend strongly on the target material and its matrix. In this study, we find that the maximum persistence of U I emission in the inert nitrogen and argon fill gases is significantly greater than the persistence in air and, additionally, the plume expands more slowly and irregularly in air, as shown in Fig. 4. The formation of U oxides in large oxygen concentrations at the outer edges of the plume may quench the emission of the plasma in these regions, especially at late times. This may explain why the plume in air decelerates more quickly than in argon or nitrogen. In general, plume deceleration can be correlated with the plume-ambient pressure interplay, where the plume front is strongly compressed.^35,51

Emission signals and backgrounds, the essential observables for analysis, are influenced by the morphological characteristics of the LPPs in different ambient environments. Investigation of plume expansion with copper-containing plasmas suggests that confinement at ambient pressures on the order of 10–100 Torr increases the atomic line intensity as well as early-time continuum emission. At lower pressures (≤1 Torr), the intensity of atomic emission features increases slowly with increasing pressure, suggesting that the excitation may occur from collisions between plasma species and ambient gas particles.⁵² Additionally, the nature of the ambient gas plays a crucial role in determining the signal levels; for example, signal-to-background ratios of U emission features in argon gas increased fourfold compared to air in 100 Torr pressure.³⁸ Modeling studies also showed that the plasma generated in argon is typically characterized by higher plasma temperatures than in air and nitrogen.⁵³ A plausible mechanism for quenching atomic U emission is collisional interaction with ambient species. Nitrogen, having a greater density of states due to its rovibrational levels, may more quickly quench the emission from atomic U species via inelastic collisions explaining the shorter-lived atomic U emission in Fig. 6 between nitrogen and argon.⁵⁴ Additionally, molecular nitrogen is more readily ionized than argon since it has a lower ionization potential than argon; and notably the specific heat of nitrogen is approximately twice that of argon.⁵⁵ Recent research demonstrated that, under certain conditions, LPPs in argon can provide a uniform radiation source.⁵⁶ The greater plasma temperature may explain the prolonged persistence of the atomic U emission compared to nitrogen observed in Fig. 6. One contributor to the different temperatures in air and argon may be the absence of endothermic reactions between the LPP constituents and particles from the ambient in argon.

The contrast between the nitrogen and air cases implies that the presence of oxygen, among other species in the surrounding environment, causes depletion of the excited state atomic U population. Both U oxides and nitrides may form in air. However, the formation of U nitrides is considered insignificant due to the extremely high bond dissociation energy of nitrogen molecules in comparison with U reaction pathways with oxygen molecules.⁵⁷ This argument is consistent with negligible differences observed in persistence of U species in argon and nitrogen environments, as shown in Fig. 6. Furthermore, we see an increasing degree of plume splitting with the addition of oxygen into the pure argon background shown in Fig. 8. The UO emission in the split region farther from the target grows in prominence with increasing oxygen partial pressure when compared to the control (pure argon) case, where the species expands more uniformly. Table I shows the most probable thermochemical U ↔ UO ↔ U_xO_y reaction pathways summarized from Ref. 1. The forward reaction rate coefficients given in Table I show the temperature dependences for each pathway. The rate coefficients predict that heavier gas-phase U oxides (UO₂, U₂O₂, U₂O₃, and UO₃) tend to form at lower gas temperatures. Emission from U atoms is typically observed at greater plasma temperatures (∼4000–10 000 K) while UO features are prevalent at lower temperatures (∼3000–6000 K).³⁰ Polyatomic compounds (e.g., UO₂) are expected to form at temperatures below ∼4000 K.^1,29

In the results reported here, the background level decreases with time along with the disappearance of the atomic lines originating from the higher-lying lower energy levels in Fig. 7. Recent work suggests that a part of this background may originate from emission of U oxides,³⁰ which may explain why the background continues to decrease with time after the majority of atomic or ionic emission features have already diminished, as seen in Fig. 7(b). The late-time broad spectra in Fig. 9 show the appearance of broad spectral features of which the most prominent lies between 320 and 380 nm with increasing ambient oxygen concentration. In recent work, we identified this feature as possibly originating from higher U oxide species including uranium dioxide.³¹ Such unidentified broad spectral features from higher U oxides may contribute, especially at later times (10–100 μs), to the background in the optical emission spectrum. Late-time agglomeration tends to occur on the millisecond-time scale. Kim et al.²⁵ reviewed the formation of nanometer-sized agglomerates with silicon among other low-Z targets following nanosecond-LA. These late-time agglomerates may contribute graybody-like broad emission to the optical emission background. Amoruso et al.⁵⁸ reported nanoparticle temperatures of ∼2000 K following femtosecond-laser ablation of silicon, which corresponds to a peak in the blackbody emission in the near-infrared (∼1.5 μm). However, further work is necessary to characterize particulates which may form following nanosecond-LA of heavy, high-Z targets such as U. Several prior studies investigated both the early- and late-time particulates in nanosecond-LPPs. Early-time ejecta arise from phase explosion, recoil ejection, and exfoliation.⁴² These particles may range in size between nanometer⁴³ and micrometer.⁵⁹ Late-time agglomeration, on the other hand, produces nanometer-scale particulates and occurs as a result of condensation in low-temperature regions of the plasma remnants.⁶⁰ Recent works investigated the influence of reactive species in the environment on agglomeration kinetics with low-Z targets including aluminum,^5,61 titanium, and silver;⁶¹ however, current literature on agglomeration kinetics in LPPs composed of heavy atoms with atomic numbers greater than that of gold is lacking. Consequently, further work is necessary to investigate the late-stages of plasma evolution (>100 μs) with particular focus on agglomeration of U- and U oxide-containing nanoparticles.

The comparison of U and UO signals in Fig. 10 provides evidence that optimal concentrations of uranium and oxygen exist in the LPP for the formation of UO. The weaker temperature-dependence of the rate coefficient for Reaction (2) in Table I is consistent with the observation in Fig. 10 that the formation of UO depends sensitively on the concentration of oxygen. The depletion of UO to form UO₂ occurs at a similar rate over a larger range of temperatures while the most probable exothermic reaction which forms UO [Reaction (1)] has a steeper temperature dependence and is more likely to occur at greater gas temperatures. Our recent work⁶² investigated the time-dependent temperatures in varying oxygen fractions for U-LPPs by comparison of simulated and observed U I emission. The results show that greater oxygen fractions promote lower gas temperatures. The leveling off of the prominence of the higher U oxide feature in Fig. 10(c) implies that greater oxygen concentrations may lead to an increased reaction rate of the intermediate uranium oxides such as uranium dioxide into even higher gas-phase oxides, UO₃, U₂O₂, etc., “This observation is consistent with the temperature-dependences of the rate coefficients given in Table I for the formation of higher U oxides [Reactions (2)–(12)]; lower gas temperatures promote the formation of heaver U oxides.” However, current literature lacks information about the spectroscopic emission signatures of such higher U oxides. Further work is necessary to determine the plasma temperature and particle densities which correspond to the reaction pathways responsible for the formation of U oxides.

We investigated the morphology of laser-produced U plasmas in order to understand the processes which dictate the observable atomic and molecular emission signatures from the plasma, relevant to practical applications of spectroscopic techniques such as LIBS or LIF. Our results address the gaps in experimental validation of kinetics of high-Z atoms and compounds in LPP which may contribute to gas-phase chemistry in the presence of reactive species in the environment. The comprehensive measurements tracking the evolution of the plume and distributions of several species in the LPP provide evidence of complex hydrodynamics which result in confinement and plume splitting in both nitrogen and air. The inert argon background extends the persistence of atomic uranium emission compared to the nitrogen background.

The air environment exhibits significant reduction in the persistence of atomic uranium emission as well as an increased background compared to both nitrogen and argon, suggesting that combustion/oxidation of U with oxygen from the ambient occurs at a significant rate and leads to the formation of U oxides. The emission from U oxides may contribute to the increased background due to the unresolved, overlapping fine features among the strong emission of atomic and ionic U species at earlier times in the plasma. Further systematic studies investigating the effects of oxygen in an argon background suggest that chemical reactions between plasma species and ambient oxygen species contribute significantly to the formation of U oxides.

We observe a small degree of segregation between U and UO in the LPP, though the emission of both species behaves similarly beyond a given optimal oxygen concentration. It can be concluded that large background oxygen concentrations foster the formation of higher oxides, whereas lower concentrations instead promote reactions which form UO from atomic U. The results of this work that pertain to the morphology and chemistry of uranium in the LPP may be of more general interest by providing insight into the behavior of other heavy, reactive particles in plasmas.

This research was supported by the National Science Foundation Graduate Research Fellowships Program (NSF GRFP) (No. DGE 1256260), the Department of Energy (DOE) National Nuclear Security Administration (NNSA), Consortium for Verification Technology (CVT) (No. DE-NA0002534), and the Office of Defense Nuclear Nonproliferation (DNN) (NA 22). Pacific Northwest National Laboratory (PNNL) is operated for the U.S. DOE by the Battelle Memorial Institute (No. DE-AC05-76RLO1830). The authors would like to thank the PNNL Radiological Services Department.