1. Introduction

Laser ablation spectroscopic methods such as laser induced breakdown spectroscopy (LIBS) have always been regarded as simple and dependable elemental analytical techniques and have become one of the leading technologies for direct solid sampling in analytical chemistry [1–3]. These methods are able to provide in situ, real-time or near-real-time analysis with no sample preparation prerequisites, and are very versatile in a wide variety of investigation environments including hostile, remote or physically-inaccessible locations. Also, since they affect only a tiny portion of the sample with commonly micron-scale ablation spots, they are much less destructive than many other analytical tools. Thus they are often the methods of choice for applications requiring rapid analysis (such as in-the-field geological prospecting and planetary exploration [4]), prompt material identification (such as metal and plastic sorting for recycling [5]), examinations of targets not readily accessible (such as nuclear reactors [6]), and standoff investigation of waste sites, explosives and other hazardous locations at several meters distance or more [7]. For example, among investigation methods, LIBS is the in situ analytical tool chosen for NASA’s current Mars exploration missions, not only for offering remote sensing capabilities and real-time analysis, but also for little or no sample preparation requirement, no consumables and no chemical waste by-products [8]. Chemical composition analysis is vital for NASA’s planetary and asteroid explorations, which provide crucial information on the origin and history of the planetary bodies and the solar system, the likelihood of harboring past and existing life, and the discovery and identification of various resources. Surface and interior composition of terrestrial planets provide data critical for understanding the formation and evolution of habitable worlds like our own. Organic materials and water on any astronomical object will not only be important for searching for extraterrestrial life forms but also essential for supporting future human space exploration. LIBS instruments in the ultraviolet/visible/near-infrared (UVN) spectral region, such as the one used in the Chemistry and Camera complex (ChemCam) on the Mars Curiosity rover, are able to rapidly assess the elemental composition of surface minerals [8]. In addition, the depth profiling capability of LIBS has been utilized for remote regolith formation studies beneath the dust coating [9] and provides valuable elemental composition information of the Martian surface. Laser ablation spectroscopy including LIBS provide easy, fast, and in situ chemical analysis with reasonable precision, detection limits, and cost.

However, despite being potent and agile analytical probes, as far as we know, all of the laser ablation spectroscopic techniques, including LIBS, analyze only the elemental emissions in the UVN spectral region with little molecular information of the sample materials [1–3]. They reveal only the elemental composition of the material, and provide no information on the chemical bonds between the atoms, which is important to determine molecular composition. To recover the molecular composition information, complicated analysis (e.g. Principle Component Analysis) of relative intensity ratios between various elemental emission features is often attempted. The interference with ambient air and aerosols, e.g. N₂ and O₂ on Earth, N₂ and CH₄ on Titan, and CO₂ on Mars, and temperature changes in the LIBS plasma from shot to shot can influence the intensity ratios and lead to possible errors for quantitative analysis of compounds that are rich in carbon, nitrogen, and oxygen such as organic materials [10]. Despite their great versatility, classification of the molecular as well as the elemental composition remains a great challenge for laser ablation spectroscopic techniques. While LIBS is an efficient technique for understanding transition metals, it is not as useful for understanding organics [11].

In laser ablation spectroscopy such as LIBS, a minuscule volume of the sample material is ablated and vaporized and further interacts with a trailing portion of the laser pulse to form a micro-plasma that contains breakdown species, such as free electrons, excited ions, atoms, and molecules [12,13]. The target furthermore interacts with the laser-induced micro-plasma as the plasma being evolving between its initiation, expansion, and decay. During the plasma evolution, ionic and atomic breakdown products, vapor, liquid, and solid phases of target materials are all present at the target sample surface [13]. The core of laser-induced plasma is too hot for molecular species to survive and the ablated sample species in this region are mostly found in their neutral atomic and ionized states [14]. However, it is known that the temperature of the laser-induced plasma is not uniform [14,15]. A temperature gradient exists in the laser-induced plasma with a colder zone located around the perimeter of the plasma plume where complex molecular species can survive as shown in Fig. 1. In a simulation of nanosecond laser-induced plasma on the surface of aluminum, an extensive region located at the perimeter of the plasma plume was found with electron temperature well below 5 eV [12] whereas most of the bond-dissociation energies of covalent bonding in molecules are between 1 to 6 eV [16]. For example dissociation of HO–H bond of a water molecule (H₂O) is > 5.11 eV [16], dissociation of the CO–H bond of isopropanol (iso-C₃H₇OH) is > 4.6 eV [16], and dissociation of the O-Si–O bond of a silica cluster (SiO₂) is > 4 eV [17]. Therefore, it is reasonable to expect intact complex (polyatomic) molecular species in the cooler zone or in the vicinity of the laser-induced plasma. It is also well known that molecules exhibit molecular spectroscopic signatures or “fingerprints” in the mid-wave IR (MWIR) to long wave infrared (LWIR) spectral regions that result from vibrational-rotational transitions. In the past decade, a series of pioneer infrared emission spectroscopic studies of laser-induced breakdown (LIBS) plasmas have been reported [18–25]. The use of fast and sensitive photo-conductive detectors such as InSb and HgCdTe (Mercury-Cadmium-Telluride, MCT) made it possible to detect emission signatures of complex molecules from laser-induced plasmas at the surface of various materials using relatively simple optical constructs.

 

Fig. 1 Conceptual diagram of temperature gradient in the laser-induced plasma plume with colder zone located nearer the perimeter of the plasma. The colored regions are schematically reproduced from simulations of nanosecond laser-induced plasma on the surface of aluminum [12] and characterized by different electron temperatures (in eV) in the plasma plume (see the color codes).

Download Full Size | PDF

Several emitting atomic and complex molecular species in laser-induced plasmas were identified in those pioneer mid- to long-wave infrared (2-12 µm) studies such as neutral metal atoms [23], oxygenated combustion molecular by-products (e.g. NO, CO₂, H₂O [21,24]), and intact inorganic and organic molecules from the condensed phase samples [12,18–20]. In addition to the molecules that result from the recombination of atomic species in the plasma [21], it was for the first time shown that abundant intact molecules from the sample targets not only survived the laser ablation, but were also thermally excited by the laser-induced plasmas [18–20]. During a LIBS event, sample molecules were excited into high vibrational states, presenting infrared signatures of intact molecules in the fingerprint region. Thus, LIBS emission in the long-wave infrared region (LWIR LIBS) is occasionally also referred to as laser-induced thermal emission (LITE) that provided direct information on the molecular composition of the sample substances. By focusing on infrared vibrational emissions from the periphery of the laser-induced plasma, the laser ablation spectroscopic technique can readily single out target molecules from contaminants, without the need to unscramble the spectral fingerprints of targets from the cluttered background.

An MCT linear array detection system that is capable of rapidly capturing (with sub-millisecond time resolution) a broad spectrum of atomic and molecular LIBS emissions in the LWIR was recently developed by which a broad band emission spectrum of condensed phase samples (in a spectral range between 4.7 to 10 µm) can be acquired from just a single nanosecond laser-induced micro-plasma [18,19]. This setup offers the capability of a simultaneous UVN + LWIR LIBS measurement from a single laser pulse-induced micro-plasma and has the capability of rapidly probing samples “as is” without the need of elaborate sample preparation. The simultaneous UVN + LWIR LIBS instrument can capture concurrently the emission signatures from atomic, and simple and complex molecular target species existing in the laser-induced plasma plume in both UVN and LWIR spectral regions within micro-seconds [18]. Our previous studies using this recently developed lab-based bench-top simultaneous UVN + LWIR LIBS spectrometer have proved the feasibility and benefits of simultaneous UVN + LWIR LIBS for rapid chemical identification [18]. The limit of detection (LOD) of this lab-based bench-top LWIR LIBS spectrometer setup is estimated to be < 60 µg/cm² in detecting organic explosives depositing on metals [25] which is comparable to the LOD of the conventional UVN LIBS (1 - 100 µg/cm²) in detecting similar samples [26–28].

The combination of atomic emission signatures derived from conventional UVN LIBS and the fingerprints of intact molecular entities determined from LWIR LIBS acquired in one single micro-second measurement suggests the combined measurements to be a powerful spectral technique for in situ chemical analysis. In situ standoff (≥ 1 meter) detection and identification of inorganic and organic compounds on regolith/metal substrates are essential in various applications both on Earth and on other extraterrestrial objects in the Solar system. LWIR LIBS has been used in surface contamination studies of organic high explosives and chemical warfare agent simulants on terrestrial surfaces [25]. LWIR LIBS has also been used in some astrophysics-related studies to probe the high temperature carbon emissions [21] and organic residues in Martian regolith simulant [22]. The distance between samples and collection optics in all those previous studies [18–25] were 10 cm. In the past year, the standoff detection range of the simultaneous UVN + LWIR LIBS detection systems was successfully extended from a merely 10 cm to ≥ 1 meter (up to 6 meters) by adopting a reflecting telescope collection scheme. To survey the feasibility of simultaneous UVN + LWIR LIBS as a new approach for in situ standoff chemical and biological composition analysis in geological, environmental survey, national security applications in threat identification, and solar planetary explorations, several relevant substances such as water, ice, JSC-1A Mars regolith simulants (Hawaiian weathered volcanic ash) [29] with and without organic residues, and several common terrestrial minerals were studied at standoff (1 meter) distance.

2. Experimental setup

The novel standoff (≥ 1 m) UVN + LWIR LIBS detection system consists of a nanosecond Nd:YAG laser, steering optics, a collection telescope, an UVN spectrometer with a CCD detector array, and IR grating spectrometer with a linear MCT detector array as shown in Fig. 2. The excitation laser was a flash lamp pumped, actively Q-switched Nd:YAG laser (by Quantel Laser). The wavelength of the ~10 ns exciting laser pulse was 1064 nm with pulse energy being adjustable between 67 to 340 mJ. Solid samples in this study were mounted vertically on a xyz translational sample stage that was positioned at 1 meter away from the front end of the collection reflecting telescope as shown in Fig. 2 and a plano-convex lens of 1.5-meter focal length focused the laser pulses onto the sample surface. In this study, the laser pulse energy used was about 120 mJ (except in the measurement of ice, a ~310 mJ pulse was used) which usually generates a spark size of one to a few millimeters. The collection optics for UVN spectral range was a 75 mm diameter plano-convex lens with focal length of 100 mm. Collected signal was focused into a 50 µm core multimode optical fiber that was connected to a Czerny Turner grating/CCD-based spectrometer covering 200 to 1000 nm range (Thorlabs CCS200). For the LWIR LIBS collection arm, the primary mirror of the LWIR reflecting-telescope collection optics was a custom-made 6-inch concave gold mirror. LWIR emissions from the laser induced plasma was collected by the primary mirror, redirected to a 1.5-inch flat mirror and then focused onto the entrance slit of a grating-based monochromator. The monochromator from Princeton Instruments had a 15 cm-focal-length collimating mirror and a 30 grooves/mm grating blazed at 8 µm. The entrance slit width in this study was set to 0.5 mm. The output slit was removed and replaced with an MCT linear array detector connected to cryogenic infrared integrated readout integrated circuits (ROIC). The MCT linear detector array had 332 photo diode pixels (dimension of each pixel was 50 µm × 50 µm) along the direction of dispersion with a cutoff wavelength around 10 µm (range: 2-10 µm). The original 15 cm-focal-length focusing mirror inside the monochromator was replaced with a flat mirror and a ZnSe lens with a 5 cm focal length to increase the spectral range covered by the MCT array. The resulting magnification of 1/3 reduced the spot size along the dispersion direction at the MCT array to ~0.16 mm (about 3.2 pixels) for a given wavelength. The signal spot size perpendicular to the dispersion direction at the MCT array was estimated to be > 0.6 mm when sample placed at 1 meter from collection optics. Therefore, this setup has capability to probe samples up to approximately 6 meters away with similar signal intensity and SNR to those of 1 meter distance. The calibration of the grating-based monochromator and MCT array detector readouts were described in detail in our previous studies [19]. The single line resolution limit of this LWIR LIBS detection system, the FWHM of a single narrow emission line, is around 76 nm.

 

Fig. 2 Conceptual Diagram of the Simultaneous UVN + LWIR LIBS Setup with Standoff Distance of ≥ 1 Meter (1000 cm).

Download Full Size | PDF

Broad spectra of emission signatures in both the UVN and LWIR spectral regions initiated by the same single laser-pulse-induced micro-plasma on the target surface could be recorded at different delay times after the laser pulse. The Nd:YAG laser driver triggered the active Q-switch and fired a nanosecond pulse at the sample. The same Q-switch trigger also triggered the control electronics of the detection system to initiate the acquisition sequence of both UVN and LWIR LIBS detectors. The control electronics subsequently triggered the UVN spectrometer and LWIR detector array separately according to the operation parameters of the UVN and LWIR LIBS detection components (delay time and integration time) set by a user interface of the computer controlling software connected to both detectors.

From our previous studies [16] the lifetimes of the molecular vibrational LIBS emissions are generally much longer (> 10 times) than those of the atomic LIBS emissions. Therefore, to optimize the signature features, in general the UVN spectrometer’s delay time was set to be 3 µs with 10 µs integration time while the LWIR detector array’s delay time was set to be 24 µs with 44 µs integration time for all but the time-resolved sequential measurements in this study. The 24 µs delay time lessened the influence of the plasma-induced thermal background emissions. The accuracy of delay and integration time is well within 0.1 µs. LWIR measurement of this 1-meter UVN + LWIR LIBS detection system results showed that the signal level is comparable to that of the bench top configuration with 10 cm standoff distance in our previous LWIR LIBS studies [18,19]. This is consistent with predicted performance, because the spot size on the MCT array detector pixel are larger than the pixel size in both cases, and numerical apertures remain the same as well.

The wavelength range of the broad LWIR LIBS spectrum of this simultaneous UVN + LWIR LIBS detection system can be easily set by adjusting the orientation of the grating in the monochromator and employing appropriate long-pass filter placed right in front of the entrance slit of the monochromator. When the spectral region of interest of a LWIR measurement was centered on 8 µm, the grating of the monochromator was oriented to direct the 8 µm light onto the middle of the linear array detector, pixel number 150 and a 5.5 µm long pass filter was placed right in front of the entrance slit of the monochromator to prevent higher order leakage. In this configuration, a broad band LWIR emission spectrum from 5.6 to 10 µm can be acquired from just a single nanosecond laser-induced micro-plasma. Similarly, when the spectral region of interest of a LWIR measurement was centered on 6 µm, the grating of the monochromator was orientated to direct the 6 µm light onto the pixel number 150 of the linear array detector, and a 4.6 µm long pass filter was placed right in front of the entrance slit of the monochromator to prevent higher order leakage. In this configuration, a broad band LWIR emission spectrum from 4.7 to 9 µm can be acquired from just a single nanosecond laser-induced micro-plasma.

Simultaneous UVN + LWIR LIBS spectra in this study were obtained by averaging results from four laser pulses sampling different spots on the sample surface for better signal-to-noise ratio over a single laser pulse. All the soil samples in this work were made from the Martian regolith simulant, JSC Mars-1A powders (Orbitec, particle size < 1 mm) pressed into solid sample tablets of ~2 cm in diameter and 7 mm in thickness using a high-pressure hydraulic press with approximately 10 tons of pressure. Unless a specific atmosphere (such as Ar and N₂) is stated, all the simultaneous UVN + LWIR LIBS emissions studied in this work were performed in ambient air atmosphere.

3. Results and discussion

Standoff detection and identification of water, organics, and minerals are essential in exploration missions on terrestrial planets and moons. The terrestrial planets and moons are composed primarily of rocks, metals, and metal oxides. Their surfaces also often have certain forms of water (e.g. ice). Mars regolith simulant JSC Mars-1A was recently studied at 1-meter standoff distance in ambient atmosphere using the simultaneous UVN + LWIR LIBS spectrometer. Similar to the Martian regolith, the main constituents of the Mars regolith simulant JSC Mars-1A are Fe₂O₃, Al₂O₃, and SiO₂ [29]. Elemental emission signatures from constituents of JSC Mars-1A could be readily observed in the UVN LIBS spectra (Fig. 3(a)) such as Si from SiO₂, Al from Al₂O₃, and Fe from Fe₂O₃ [30]. An intense and broad SiO₂ molecular vibrational emission feature around 9 and 9.5 µm [31] (Fig. 3(b)) could also be identified in the LWIR LIBS spectrum of JSC Mars-1A. The simultaneous UVN + LWIR LIBS spectra of a mock mixture sample disk made from uniformly mixed 18% Fe₂O₃, 30% Al₂O₃, and 52% SiO₂ powders, similar to the weight ratio of these three components in the JSC Mars-1A sample [29], were measured and shown in Fig. 3(c) and 3(d). The UVN LIBS spectra of both samples showed comparable emission signatures from Al, Fe, Si, and O [30]. It is sensible considering both samples mainly made of SiO₂, Al₂O₃, and Fe₂O₃. However, the LWIR LIBS spectra of these two samples showed significant differences. While intense and broad SiO₂ molecular vibrational emission feature around 9 and 9.5 µm being apparent in both spectra, the LWIR LIBS spectrum of the JSC Mars-1A was dominated by strong emission features around 6.6 μm that were absent in the mock mixture sample. Instead one can see clearly an Al atomic signature around 6.1 at 7.4 µm in the spectrum of the mock mixture sample.

 

Fig. 3 The simultaneous UVN (a) + LWIR (b) LIBS spectra of the Martian regolith simulant JSC Mars-1A and the simultaneous UVN (c) + LWIR (d) LIBS spectra of a mock mixture sample made with 18% Fe₂O₃, 30% Al₂O₃, and 52% SiO₂.

Download Full Size | PDF

The strong emission features from 6.6 to 8µm of JSC Mars-1A LWIR LIBS spectrum can be attributed to water molecules bound in the regolith matrix. Water has been observed in its gaseous form in various astronomical environments. Spectra of protoplanetary disks of seven stars in Taurus-Auriga star-forming region exhibited emission features at 6.6 µm due to the vibration-rotational bending mode of water vapor [32]. This water bending mode emission band centered at 6.3 µm was also observed in the flame emission spectrum (shown in Fig. 4) of hydrogen flame with abundant amount of oxygen in the ambient environment [33]. The freely rotating water molecule in the vapor phase is a strongly asymmetric top, its Ray asymmetry parameter κ is −0.4384 as calculated from its ground vibrational state geometry. Due to the selection rules of the symmetric bending vibration-rotation transitions of freely rotating water molecules, with their oscillating dipole moments along the axis of symmetry of the molecule (thus a B-type band), there was very little emission observed at the center of the band around 6.3 µm. The R branch of the water band of the hydrogen flame had maximum emission intensity around 5.4 µm and the P branch around 6.6 µm [33]. The rotational structure of the water vapor emission spectrum of the hydrogen flame was not well-resolved, but three strong and distinct lines could be seen between 7.5 to 8 µm which were clearly visible in both flame and LWIR LIBS emission spectra in Fig. 4. The close resemblance of the UVN + LWIR LIBS emission spectrum of JSC Mars-1A to the emission spectrum of hydrogen flame in oxygen between 5.6 to 8 µm, as shown in Fig. 4, suggested this prominent LWIR features in the UVN + LWIR LIBS spectrum could also be attributed to water vapor.

 

Fig. 4 The LWIR LIBS spectra between 5.6 to 10 µm of the Martian regolith simulant JSC Mars-1A (black curve) and the flame emission spectrum of hydrogen flame with abundant amount of oxygen in the ambient environment [33] (red curve).

Download Full Size | PDF

The simultaneous UVN + LWIR LIBS measurements of JSC Mars-1A (Fig. 5) were performed in both the ambient atmosphere and an argon atmosphere by gently blowing argon gas directed at the sample surface to exclude the ambient air completely from the plasma plume. The exclusion of interaction between the plasma plume and ambient air achieved by the argon flow was confirmed by the fact that there were no O and N atomic emission lines in the UVN LIBS spectrum of the reference graphite sample (Fig. 6) and in the UVN LIBS spectrum of the JSC Mars-1A sample (Fig. 5(a)) under argon atmosphere. However, we can clearly see the water vapor LWIR LIBS emission band of JSC Mars-1A sample (Fig. 5(b)) was not affected by the removal of the ambient oxygen. This tells us the origin of the vibration-rotational bending emission signatures of water vapor in the LWIR was not the by-product of combustion of sample H as those in the hydrogen flame but the intact water molecules from the sample regolith matrix. JSC Mars-1A is known for its moderate water content [29]. The water in the regolith matrix was vaporized and excited by the laser-induced plasma and consequently emitted strong and distinct vibrational-rotational signatures in the LWIR from likely the cooler region of the plasma.

 

Fig. 5 The simultaneous UVN (a) + LWIR (b) LIBS spectra of the Martian regolith simulant JSC Mars-1A in ambient air (red curves) and in Ar-flow atmosphere (black curves).

Download Full Size | PDF

 

Fig. 6 The UVN LIBS spectra of the reference graphite sample in Ar-flow atmosphere.

Download Full Size | PDF

Since the vibration-rotational bending mode of water vapor spans 5 to 8 µm, the collection optics of the LWIR LIBS spectrometer were adjusted and reoriented to acquire the spectrum between 4.7 to 9 µm. The LWIR LIBS spectrum of the Martian regolith simulant JSC Mars-1A between 4.7 to 9 µm is shown in Fig. 7(a) along with the water bending mode emission band in the flame spectrum of hydrogen. Both R and P branches of the water vapor bending bands were fully visible. These same features were also observed in the LWIR LIBS spectrum of ice (Fig. 7(b)). As seen in Fig. 3 and 7, due to the moderate water content of the Mars regolith simulant JSC Mars-1A, the LWIR LIBS spectrum in 5-8 μm was dominated by the water vapor bending bands while in its UVN LIBS spectrum an H atomic signature at 656 nm was clearly visible. Furthermore, dried Mars regolith simulant JSC Mars-1A sample disks were prepared by using the dried JSC Mars-1A powders baked in a 900 °C oven for two hours. The UVN + LWIR (4.7 to 9 μm) LIBS spectra of this dried JSC Mars-1A sample are shown in Fig. 8. The LWIR LIBS spectrum of dried JSC Mars-1A was very similar to that of the water-free mock mixtures (Fig. 3(d)) with Al atomic signature around 6.1 and 7.4 µm, and a very broad and intense SiO₂ molecular vibrational emission feature around 9 µm. The near absence of the H atomic signature at 656 nm in the UVN LIBS spectrum of dried JSC Mars-1A also indicated the successful removal of water from the simulant regolith. A dried JSC Mars-1A sample disk was then rehydrated by wetting the top surface with 0.1 ml of de-ionized (DI) water and UVN + LWIR (4.7 to 9 μm) LIBS spectra were measured, as shown by the red curve in Fig. 8, after the applied DI water completely infiltrated the sample disk. The prominent vapor water bending vibrational-rotation bands centered at 6.3 µm and a strong H atomic signatures at 656 nm were clearly visible again after rehydration. Similar to those of Martian regolith simulant JSC Mars-1A and ice, the center of the water vapor bending band of rehydrated Martian regolith simulant JSC Mars-1A was located at 6.3 µm. The R branch of the band had maximum emission intensity around 5.4 to 5.6 µm and the P branch around 6.6 to 6.7 µm.

 

Fig. 7 The LWIR LIBS spectra between 4.7 to 9 µm of (a) the Martian regolith simulant JSC Mars-1A and (b) ice. The water bending band in the emission spectrum of hydrogen flame [33] shown in (a) (red curve).

Download Full Size | PDF

 

Fig. 8 The simultaneous UVN (a) + LWIR (b) LIBS spectra of the oven-dried (black curves) and DI water re-hydrated (red curves) Mars regolith simulant JSC Mars-1A sample disks.

Download Full Size | PDF

As mentioned above, the vibrational-rotational band of water vapor recorded between 4.7 – 9µm is a bending vibrational mode, thus it is a B-type band which the change in dipole moment being along the axis with intermediate moment of inertia. The band spectrally looks similar to a parallel band for a linear molecule having two wings (R and P branches) and no central peak. This is indeed what we observed in LWIR LIBS spectra of the all of the water-containing samples. Water molecules in plasma surrounding are highly excited both vibrationally and rotationally. Emission from these vibration-rotation levels as well as from highly excited pure rotational levels occurs in the mid-infrared region. We used an application called HAPI (Hitran Application Programming Interface) that is based on the HITRAN/HITEMP database [34,35] for spectral simulations. HAPI is a free to use application. High temperature IR emission spectra of hot water vapors taken in the laboratory were previously published, e.g [33,36], and many papers utilized HITRAN for analyzing experimental data on water spectra at high temperatures. HAPI simulations assume thermodynamic equilibrium, i.e. the vibrational and rotational temperatures of the various vibrational-rotational transitions are taken to be equal to the thermal gas molecular temperature. Vibrational-rotational excitation in plasma environment is likely more complicated than in the usual thermally excited gases and we do not know whether any former application of HAPI to plasma environment gases has been published.

HAPI is written using the Python programming language and presently also available for Windows computers. In order to use HAPI the user needs only to download the water spectral lines from the HITRAN on-line portal (https://hitran.org) [35]. The data file is very large (between 4.7 – 9.1 µm spectral ranges there are 7467 vibration-rotation lines for the normal isotopologue, ¹H₂¹⁶O). While the spectral resolution used in high-resolution applications is between 0.01 and 0.001 cm⁻¹ (~0.04-0.004 nm), the present experimental LWIR LIBS spectra have much lower resolution, around 76 nm. Thus the spectral simulations were done by taking instrumental slit width into account. The other simulation parameters used were the original size of the wavenumber step size for the database lines (0.04 nm), called OmegaStep in HAPI, the partial gas pressure for water (taken to be 0.1 atm), the emission path length (taken to be 5 mm), the gas temperature in Kelvin units, and the convolution width identical to the experimental spectral resolution (76 nm).

It was found that the partial gas pressure value and the emission path length were not crucial simulation parameters, due to the fact that pressure broadening of the individual transitions is negligible at the present spectral resolution, and similarly the emission path length mainly influences the emission intensity. However the gas molecular temperature is an important parameter. In HAPI, the simulated spectrum is calculated by setting the gas molecular temperature T to be corresponding to the molecular vibrational and rotational excitation temperatures (T_vib and T_rot respectively). This molecular temperature is determined by assuming a Boltzmann thermal equilibrium between the rotational and vibrational modes (energy levels) of the molecules, i.e. T = T_vib = T_rot. Thermal equilibrium of vibrational-rotational levels does not mean that the temperature of free electrons in the plasma is necessarily identical to this molecular temperature T especially after long delays (> 24 µs). The electron temperatures derived from atomic line broadening (mainly Stark broadening) are often different from the vibrational-rotational temperatures of diatomic species in various LIBS studies of molecular bands in the visible [37–40]. The molecular vibrational and rotational energy levels were found to be remaining in thermal equilibrium in those studies.

Several water molecular temperature values were tested and it was found that the temperature was not perfectly determined by the simulations, i.e. the observed LWIR LIBS spectra deviated slightly with the HAPI simulations by any single-temperature simulation. HAPI simulations of vibrational-rotational emissions of water vapor at three temperatures (T): T = 1000 K, T = 1700 K and T = 2500 K shown in Fig. 9 demonstrate these small deviations. This may be due to deviations from the Boltzmann thermal equilibrium conditions implied by HAPI simulations. It may also be that the small spectral deviations in Fig. 9 are not due to non-thermal equilibrium conditions but to the variation of kinetic temperatures within the IR radiating parts of the plasma. In this case the observed IR emission comes from plume locations of different temperatures and the detected emission is averaged over a certain volume of the plasma plume. Therefore, the resulting spectrum is not assignable to a singular well-defined temperature value. The excitation mechanisms and conditions of vibrational-rotational levels in hot gas molecules embedded in plasmas are not yet known. The population of vibrational-rotational levels of the water molecules in the plasma may be governed by additional processes besides electron collisions (e.g. molecular recombination processes), thus further studies and more refined treatments are necessary to clarify the vibrational excitation conditions for such LWIR LIBS spectroscopy. As a comparison, excitation conditions in electron collision dominated plasmas are well described in [41]. Nevertheless, the present work shows that the observed emission bands are indeed due to water, since the spectral features are rather truly reproduced by the HAPI simulations and T = 2500 K simulation is better than the other two temperatures. Considering that water vapor infrared signatures were observed and identified from 3000 K temperature sunspots [42] and < 2900 K temperature hydrogen flame [33,43], the water molecule temperature derived from the LWIR LIBS spectrum appears sensible.

 

Fig. 9 The comparison of LWIR LIBS spectra of DI water re-hydrated Mars regolith simulant JSC Mars-1A sample disks (black curves) and HAPI simulations of vibration-rotation emissions of water vapor at three temperatures: T = 1000 K (red), T = 1700 K (green) and T = 2500 K (blue).

Download Full Size | PDF

The UVN (640-800 nm) + LWIR (4.7 to 9 µm) LIBS emission spectra at different delay times of JSC Mars-1A and ice are shown in Fig. 10 and 11. Within the 640-800 nm region, the prominent emission features were H (656 nm), N (747 nm), and O (777 nm) for both samples. The intensity of those atomic emission lines generally deceased with time and become insignificant after 20 μs, which were consistent with the atomic line emission timescale (0.1 to tens of microseconds) observed in UVN LIBS studies [13], while the water vapor molecular vibrational-rotational emission band from both samples decayed at a much slower rate and all the molecular emission features were still distinguishable even beyond 200 μs. Similar slow decay rates also were observed from other molecular LWIR LIBS emission features of vibrational transitions in solid samples (e.g. ammonium perchlorate, NH₄ClO₄) in LWIR region such as: NH₄ stretching at 7.5 μm and ClO₄ stretching and deformation bands around 9 μm in previous studies of inorganic energetic materials [13]. The time-resolved spectra of water vapor emissions showed that gas-phase complex molecular species in the laser-induced plasma plume lingered for a few hundred microseconds after the initiation of the plasma.

 

Fig. 10 The UVN (640-800 nm) + LWIR (4.7 to 9 µm) LIBS emission spectra of JSC Mars-1A at different delay times.

Download Full Size | PDF

 

Fig. 11 The UVN (640-800 nm) + LWIR (4.7 to 9 µm) LIBS emission spectra of ice at different delay times.

Download Full Size | PDF

Recently NASA’s LCROSS mission discovered both water and hydroxyls (OH) in the top millimeters of the lunar regolith [44]. One of the most plausible sources of the hydroxyl and water is the sputtering and implantation of the solar wind hydrogen (H) in the oxides of the lunar regolith. Solar wind protons (H⁺) strike the lunar soils with sufficient energy to penetrate to depths of 5– 10 nm. The extremely space-weathered grains of the top layer lunar regolith have many defects and dislocations. These surface and internal sites of the lunar soils are locations where solar wind protons could be trapped and bonded with the abundant oxygen, forming individual OH and HOH molecules to depths of 5–10 nm [45]. Therefore, the hydroxyls (OH) in the regolith would likely be in the form of MOH (M: Si, Al, Fe, C, etc). The water molecule is similar to hydroxyls because of the OH bonds. The broad and intense OH stretching bands of water and hydroxyls between 2.7 to 3 µm are very similar both in position and shape. Therefore, it is difficult to distinguish between water and the hydroxyl molecules based solely on the 2.8 µm OH stretching band. Fortunately, their spectra are very different in the finger-print region (6 to 20 µm) due to the bending modes of those molecules. The water band at 6.3 µm is due to the scissoring bending of the two OH bonds and is unique to water. The presence of this band as shown in the LWIR LIBS spectrum of the Martian regolith simulant JSC Mars-1A is a strong evidence for the presence of intact water molecules in/in the vicinity of the plasma plume, therefore a very strong evidence for the presence of water in the regolith sample.

The dried JSC Mars-1A sample disks mentioned above were used as a substrate for water and hydroxyl tests due to its space application relevance and lack of prominent LWIR LIBS spectral features between 5 to 8 μm. In an oxygen-free (or nearly free) atmospheric environment such as on the Moon, asteroids, N₂-rich Titan, Mars (thin CO₂), and Europa (very thin O₂), the interaction between laser-induced plasma and the atmospheric oxygen would be practically non-existent. Approximately 0.1 ml of DI water and isopropanol ((CH₃)₂CHOH) were deposited on the top surface of two separate dried JSC Mars-1A sample disks. After the liquid chemicals completely infiltrated the dried JSC Mars-1A sample disks, UVN + LWIR LIBS measurements of both samples (Fig. 12) were done in an oxygen-free atmosphere by gently blowing nitrogen gas onto the sample surface adjusted to eliminate the ambient oxygen completely from the plasma plume. Therefore, all the molecular emission UVN + LWIR LIBS features of OH and HOH would solely come from the hydroxyl and water molecules of the sample and not from combustion by-products. In an oxygen-free atmosphere the LWIR LIBS spectrum of water adsorbed in dried JSC Mars-1A sample disk was dominated by the 6.3 µm HOH bending bands of the water vapor of the sample thermally excited by the laser-induced plasma just as what we already saw from the original JSC Mars-1A sample disk. On the other hand, the LWIR LIBS spectrum of isopropanol absorbed in dried JSC Mars-1A sample disk in the oxygen-free atmosphere had distinct vibrational CH₂ and hydroxyl group COH bending bands of the sample isopropanol vapor at 7.26 µm [31]. The bending modes of other hydroxyl groups such as SiOH and AlOH in the LWIR signature spectral region are usually located between 9 and 11 µm [46]. Therefore, UVN + LWIR LIBS has a great potential for distinguishing between the water and hydroxyl molecules in planetary regolith.

 

Fig. 12 The LWIR LIBS spectra between 4.7 to 9 µm of water absorbed in dried Martian regolith simulant JSC Mars-1A (black curve) and isopropanol absorbed in dried Martian regolith simulant JSC Mars-1A (red curve) in an oxygen-free atmosphere.

Download Full Size | PDF

In additional to water and hydroxyls, standoff detection and identification of organic residues on the terrestrial surfaces are also of interest in many applications. JSC Mars-1A samples with organic methyl salicylate (MeS, wintergreen oil) residues were also probed using this standoff UVN + LWIR LIBS array detection system. At room temperature MeS is in liquid phase. About 0.1 ml of Methyl Salicylate (MeS) was deposited on the surface of JSC Mars-1A sample tablets and allowed to sit for about 5 minutes before the thin liquid film of MeS on sample substrate was mounted and measured. From Fig. 13(a), it was clear that the UVN LIBS spectra of the JSC Mars-1A samples with and without the organic residues looked qualitatively similar. However, besides the molecular spectral signature around 6.6 μm and 9.5 μm of Martian regolith simulant, molecular signature emissions from the surface organics (MeS) residues were clearly observed (as shown in Fig. 13(b) along with the FTIR absorption spectrum of MeS in gas phase from NIST database [31]) and readily identified from the LIBS emission spectra in the LWIR region, such as the C = O stretching band 5.9 µm, the C-CH₃O asymmetric deformation at 7.6 µm, the C-Phenyl-O stretching 7.9 µm, and the aromatic CH deformation at 8.2 µm of MeS.

 

Fig. 13 The simultaneous UVN (a) + LWIR (b) LIBS spectra of the Martian regolith simulant JSC Mars-1A with (black curves) and without (red curves) organic methyl salicylate (MeS) residues. The FTIR absorption spectrum of MeS in gas phase (blue curve) from NIST database [31] is shown in (b).

Download Full Size | PDF

To further assess the feasibility and versatility of this technique on standoff surveying the geological surfaces, several common types of terrestrial minerals were also probed in ambient air at standoff distance of 1 meter. Simultaneous UVN + LWIR LIBS spectra of Shale, Tufa, and Gypsum are shown in Fig. 14-16 respectively. All three terrestrial mineral samples were from a GSI 1050-00E mounted 50 specimen collection set and probed “as is” without any sample preparation. Shale is a fine-grained, clastic sedimentary rock containing mainly quartz (SiO₂) and calcite (CaCO₃). Several intense atomic emission features of Ca and Si between 400 to 900 nm and strong molecular vibrational features of CO₃ and SiO₂ at 7 and 9.5 µm [31] were readily observed in the UVN + LWIR LIBS spectra of Shale minerals. Tufa is a limestone that contains predominantly CaCO₃. Its UVN + LWIR LIBS spectra showed intense atomic emission features of Ca between 400 and 900 nm and a single distinct CO₃ vibrational feature at 7 µm [31]. Gypsum is a soft hydrous sulfate mineral composed of calcium sulfate dihydrate. In Gypsum, the bounded water molecules interact with the CaO and SO₄ groups [47] and lead to a distinct, sharp water bending band in the low temperature infrared emission spectrum around 6.3 µm [48]. Therefore, besides the strong molecular vibrational features of SO₄ at 8.9µm [46], the 6.3 µm vapor phase bending features of these bound water molecules in Gypsum could also be readily observed in the UVN + LWIR LIBS spectra of Gypsum. Despite the fact that the detection of Sulphur by typical optical emission spectroscopy generally presents difficulties due to the strongest lines being located in the UV region < 185 nm, using UVN + LWIR LIBS one can quickly determine the predominant ingredients of these three mineral samples to be CaSO₄ for Gypsum, CaCO₃ for Tufa and CaCO₃ + SiO₂ for Shale without having to perform more rigorous and time-consuming data processing.

 

Fig. 14 The simultaneous UVN (a) + LWIR (b) LIBS spectra of Shale.

Download Full Size | PDF

 

Fig. 15 The simultaneous UVN (a) + LWIR (b) LIBS spectra of Tufa.

Download Full Size | PDF

 

Fig. 16 The simultaneous UVN (a) + LWIR (b) LIBS spectra of Gypsum.

Download Full Size | PDF

4. Conclusion

The standoff detection range of a simultaneous UVN + LWIR LIBS detection system has been successfully extended from 10 cm to ≥ 1 meter (up to 6 meters) by adopting a reflecting telescope collection scheme and it acquired UVN + LWIR LIBS emission signatures in various atmospheres for a wide variety of inorganic materials: such as water/ice (H₂O), alumina (Al₂O₃), iron oxide (Fe₂O₃), silicate (SiO₂), sulfate (SO₄), and organic materials such as methyl salicylate (MeS). These measurements of UVN + LWIR LIBS signatures in different atmospheres have revealed the existence of plasma-generated molecular emitting species along with atomic ones which gave intense and distinct signature emissions in both UVN (conventional LIBS) and LWIR (LWIR LIBS, LITE) spectral regions. For example, water molecules bound in the solid soil and mineral samples were vaporized by the pumping laser and abundant hot vapor water molecules in/near the laser-induced plasma plume were detected as early as < 3 µs after the initiation of the plasma and lasted more than 250 µs. It worth mentioning that the close match between the LWIR LIBS emission spectrum of the target sample and its gas phase IR emission spectrum (Fig. 9) and gas phase FTIR absorption spectrum (Fig. 13(b)) indicates LWIR LIBS measurements provide spectral information of the sample molecules most likely in their unaltered state.

Laser ablation spectroscopy detection systems such as LIBS, laser-ablation molecular isotopic spectrometry (LAMIS), and ICP-optical emission spectroscopy (OES) detection have become prominent elemental spectroscopic technologies for the direct analysis of solids. There are increasing needs for extending the molecular identification capabilities of laser ablation spectroscopy. Raman spectroscopy is an optical molecular spectroscopic technique that has been bundled with the laser ablation spectroscopy to offer additional complex molecular analytical capabilities [49,50]. Despite great potential, several challenges remain. In general, optical and spectral conditions that favor Raman scattering tend to reduce LIBS emission and vice versa [51]. Furthermore, Raman spectroscopy is a weak scattering technique where signals are easily masked by fluorescent and continuum background emissions [51] and can only be used to analyze the non-ablated area of the sample [50]. On the other hand, simultaneous UVN + LWIR LIBS can concurrently capture the emission signatures from atomic, and simple or complex molecular target species existing in the same laser-induce plasma plume within micro-seconds. Using a LIBS spectrometer capable of collecting emission signatures simultaneously in both the UVN and LWIR spectral range from a single laser excited plasma pulse, information about both atomic compositions and molecular structures of the target analytes can be rapidly obtained and thus the in situ analysis capability of the spectrometer can be greatly enhanced. The results in this work showed that the combination of atomic emission signatures derived from conventional UVN LIBS and fingerprints of intact molecular entities determined from LWIR LIBS could be a powerful spectral probe for both inorganic and organic chemical identification and detection and well suited for rapid in situ detection and identification of chemical materials in soil and minerals that are essential in many geoscience, defense and security, and environmental applications as well as planetary missions of resource exploration or life-form (organic materials) searching. With all the advantages similar to the conventional UVN LIBS of NASA’s ChemCam, a small size, small weight, and small power consumption simultaneous UVN + LWIR LIBS spectrometer can be mounted on exploration vehicles and conducts rapid and in situ composition analysis in planetary and asteroid explorations as well as in many environmental and defense applications in our terrestrial world. The sensitivity of the LWIR LIBS for carbon- and oxygen-rich material identification will likely find important applications in the detection and identification of water and organic materials signatures on space missions.

Despite extensive studies over decades, laser ablation is still largely unexplored at the fundamental level [52]. The physics involved in laser-induced plasma generation and subsequent evolution on solid samples is very complex and includes surface heating, melting, vaporization, ejection of particles, and plasma creation and expansion. Few known studies exhaustively detail the nature of (complex) molecular species consisting of more than two atoms in the laser-induced plasma [21,53]. As this study has shown, LWIR LIBS measurements identified and revealed spectral information of water molecules in the vapor phase from the interrogated sample that likely exist in the cooler zone of the laser-induced plasma. A further improved high-performance UVN + LWIR LIBS spectrometer that can monitor the spatial and temporal distribution of both atomic and complex molecular plasma species would enable unprecedented experiments on fundamental mechanisms of laser-induced plasmas. In addition, LIBS background emission estimation and removal processes are very complicated and the subjects of many ongoing researches [54,55]. The origin of and possible corrections of both UVN and LWIR background emissions we plan to examine in our future work.