Laser-induced breakdown spectroscopy (LIBS) - Geochemistry ...

16 downloads 6465 Views 2MB Size Report
field-portable sensor technology for real-time, in-situ geochemical and environmental ... 3Ocean Optics Inc., 4202 Metric Drive, Winter Park, FL 32792, USA.
Laser-induced breakdown spectroscopy (LIBS) – an emerging field-portable sensor technology for real-time, in-situ geochemical and environmental analysis Russell. S. Harmon1, Frank C. De Lucia2, Andrzej W. Miziolek2, Kevin L. McNesby2, Roy A. Walters3 & Patrick D. French4 1

US Army Research Office, PO Box 12211, Research Triangle Park, NC 27709, USA (e-mail: [email protected]) 2 US Army Research Laboratory, Aberdeen Proving Ground, MD 21005, USA 3 Ocean Optics Inc., 4202 Metric Drive, Winter Park, FL 32792, USA 4 ADA Technologies, Inc., 8100 Schaffer Parkway, Suite 130, Littleton, CO 80127, USA ABSTRACT: Laser-induced breakdown spectroscopy (LIBS) is a simple spark spectrochemical sensor technology in which a laser beam is directed at a sample to create a high-temperature microplasma. A spectrometer/array detector is used to disperse the light emission and detect its intensity at specific wavelengths. LIBS has many attributes that make it an attractive tool for chemical analysis. A recent breakthrough in component development, the commercial launching of a small, high-resolution spectrometer, has greatly expanded the utility of LIBS and resulted in a new potential for field-portable broadband LIBS because the technique is now sensitive simultaneously to all chemical elements due to detector response in the 200 to 980 nm range with 0.1 nm spectral resolution. Other attributes include: (a) small size and weight; (b) technologically mature, inherently rugged, and affordable components; (c) in-situ analysis with no sample preparation required; (d) inherent high sensitivity; (e) real-time response; and (f) point sensing or standoff detection. LIBS sensor systems can be used to detect and analyse target samples by identifying all constituent elements and by determining either their relative or absolute abundances. KEYWORDS: laser-induced breakdown spectroscopy, field-portable elemental analysis, Pb in soil

INTRODUCTION

LASER-INDUCED BREAKDOWN SPECTROSCOPY

Laser-based spectroscopic techniques are beginning to emerge as important tools for chemical analysis because of the prospect they offer for the selective, minimally destructive, and high sensitivity detection and analysis of solid, liquid, aerosol, and gaseous materials in real time. Laser-induced breakdown spectroscopy (LIBS) is one such technique. LIBS is not a new technique: early laser-induced breakdown studies go back to the early 1960s and important application studies date from the 1980s with the work of Radziemski (1983), Cremers & Radziemski (1983), and Cremers et al. (1984). A comprehensive review of LIBS development and applications through the mid-1990s was produced by Rusak et al. (1997). Recently, LIBS has received renewed attention because of its simple and direct nature, which make it an optimal technology for use as a real-time, field-portable sensor. The objective of this paper is to illustrate the potential of field-portable LIBS analysis for applications related to geochemical surveying and environmental monitoring. The focus of the discussion is restricted to the toxic metal Pb in soil, but a broad array of other potential applications involving both single particle detection and bulk material analysis are feasible.

LIBS is a simple spark spectrochemical technique that uses a pulsed laser to create the spark. The technique has many attributes that make it an attractive tool for chemical analysis, particularly as regards its potential as a field-portable sensor for geochemical analysis. LIBS is relatively simple and straightforward, so skilled analysts are not required. Little to no sample preparation is required, which eliminates the possibility of adulteration of the sample through improper handling or storage or cross-contamination during sample preparation. LIBS provides a real-time response and simultaneous multielement detection and analysis. The laser plasma is formed over a very limited spatial area, so that only a very small amount of sample (picograms to nanograms) is engaged in each laser microplasma event. All components of the instrument can be made small and rugged for field use and LIBS sensors can be operated either as a point sensor or in a standoff detection mode. The detection limits of LIBS are in the low hundreds to tens of ppm range for most common elements, so field-portable LIBS should be capable of field surveying and screening for the geochemical exploration and environmental remediation applications envisaged.

Geochemistry: Exploration, Environment, Analysis, Vol. 5 2005, pp. 21–28

1467-7873/05/$15.00  2005 AAG/ Geological Society of London

22

R.S. Harmon et al.

Fig. 1. Schematic drawing of a typical analytical system for laserinduced breakdown spectroscopy (LIBS) consisting of a pulsed laser, optics for focusing the laser energy onto a sample surface and optics for collecting the light produced during the LIBS reaction and delivering it to a detector, a spectrometer for resolution of the light spectrum, and a computer for system control and data processing and analysis.

A typical LIBS system (Fig. 1) consists of a pulsed laser, optics for focusing the laser energy onto a sample surface and for delivering the light produced during the LIBS event to a detector and spectrometer for resolution of the light spectrum, and a computer for system control and data processing and analysis. The foundation for LIBS is a short-pulsed, Q-switched solid-state laser that is optically focused to rapidly heat the surface of a target sample material to the point of volatilization and material ablation, which results in the generation of a high-temperature plasma on the surface of a sample. Upon cooling, the excited atomic, ionic, and molecular fragments produced within the plasma emit radiation that is characteristic of the elemental composition of the elements within the volatilized material. Fibre optic technology, which offers the potential for designing portable LIBS analysers, can be used to collect the light signal and deliver it to a detector/spectrometer that is capable of resolving part or all of the 200 to 980 nm spectral region. All elements emit within this region, so that both the detection of elemental, ionic and molecular species on the basis of spectral line presence and the quantitative analysis based upon differences in spectral line intensity can be achieved. The LIBS technique has proven capable of detecting many metals of environmental concern in both natural and anthropogenic materials. Because sample preparation and off-site analysis are unnecessary with field-portable LIBS, the measurement complexity is greatly reduced and there is no chance for sample loss or cross-contamination during transport or complicated preparations for laboratory analysis. Additional LIBS advantages include the high intrinsic sensitivity of the method, the small sample size (100 m (e.g. Cremers et al. 1995; 2002a; Whitehouse et al. 2001) which allows access to difficult or contaminated locations, and the analysis for certain elements that are outside the capability of other current field portable techniques such as X-ray fluorescence. LIBS has been applied under laboratory conditions to the analysis of water and ice (Knopp et al. 1996; Arca et al. 1998; Caceres et al. 2001), atmospheric particulates such as aerosols and pollen (Radziemski et al. 1983; Hahn 1998; Neuhauser et al. 1999; Carranza et al. 2001; Panne et al. 2001), metal species in plants and coals (Otteson et al. 1991; Wallis et al. 2000; Body & Chadwick 2001; Tozzi et al. 2002), soils and other natural

materials (Wisbrun et al. 1994; Ciucci et al. 1996; Eppler et al. 1996; Yamamoto et al. 1996; Barbini et al. 1999; Lazic et al. 2001; Hilbk-Kortenbruck et al. 2001; Maravelaki-Kalaitzaki et al. 2001; Cremers et al. 2002b), and anthropogenic materials such as concrete and paint (Pakhomov et al. 1996; Yamamoto et al. 1996). However, field applications have been limited because of the large size of typical LIBS laboratory systems. Some progress was made toward the development of LIBS systems for field deployment during the 1990s. Cremers (1987) and Cremers et al. (1995) used a fibre optic probe to both deliver laser light and collect plasma emission to detect trace elements in soil. Theriault et al. (1998) and Miles & Cortes (1998) described a LIBS system developed for deployment on a cone penetrometer system for assessing toxic metal contamination in the shallow subsurface. Wainner et al. (2001) described a prototype fieldportable LIBS system that was used for the detection of Pb in soil. LIBS systems have been installed at ore processing plants for automated quality assessment of ore during beneficiation (Barrette & Turmel 2001; Rossenwasser et al. 2001). Recently, LIBS technology has matured through advances in, and miniaturization of, its component parts to the point where rugged and affordable prototype field-portable and field-deployable LIBS instruments have been developed and are beginning to be applied to environmental and industrial analysis (Yamamoto et al. 1996; Wainner et al. 2001; Moskol & Hahn 2002; Palanco et al. 2003; Walters & Rose 2003). THE LABORATORY AND FIELD-PORTABLE LIBS SYSTEMS Many LIBS applications use multiple laser shots to accumulate spectra. This approach, which was utilized in our preliminary fieldwork at the Sierra Army Depot, has the advantage of increasing the cumulative magnitude of the signal when one is interested in elements of low abundance in a homogeneous sample matrix and when not limited by the amount of sample available for analysis. By contrast, the focus of our environmental work in the Army Research Laboratory is single-shot LIBS in soil, a typically heterogeneous material, where the complete 200 to 980 nm spectrum is obtained with a single laser shot. The data presented in this paper were acquired using two different LIBS systems. A small amount of preliminary survey data was collected in the field using a narrow-band prototype portable LIBS system, whereas the majority of data were collected on the broadband bench-top LIBS system at the Army Research Laboratory. In both cases, petrochemical analysis is based on collecting and measuring the intensity of light generated by the laser spark on a soil sample. The laboratory LIBS system The bench-top system at the Army Research Laboratory (ARL) is a broadband system designed to capture the full 200 to 980 nm spectral range in a single laser shot. A single c. 10 ns, 30 mJ pulse from an actively Q-switched Ultra Big Sky Nd-YAG laser is focused by a 50 mm convex lens onto a soil sample. A bundle of seven 600 µm diameter optical fibres collects the light generated by the plasma spark on the sample surface and delivers this light to a high-resolution (0.1 nm FWHM) Ocean Optics LIBS2000+ spectrometer using seven 2048-element linear silicon charge-coupled diode array detectors. A defocusing lens is placed in front of the fibre optic bundle to ensure that each fibre collects the same light emission. Plasma continuum effects were moderated by collecting LIBS spectra for a time of 2 ms following a 1.5 µs delay after generation of the laser spark. Typically, the LIBS system is ready

Laser-induced breakdown spectroscopy

23

Fig. 2. LIBS spectrum for air between 725 and 975 nm with peaks for N and O identified.

to collect data in 2–3 min after the laser unit is switched on, so qualitative analysis is rapid. If it is necessary to develop calibration curves for a specific quantitative analysis problem, then an additional hour or two are a necessary part of the preparation process. The data collection and processing software used were those provided with the Ocean Optics spectrometer. A field-portable LIBS system based upon that in use at the ARL laboratory is currently under development, with the prototype system to be fielded in early 2004. Laboratory results Many different types of geological materials (rocks, minerals and soils) have been analysed at the ARL. A few examples are presented to illustrate the potential of broadband LIBS for different geochemical applications such as geochemical surveying for mineral exploration or environmental cleanup. In such applications, one may take the approach of examining a narrow spectral window for specific elements of interest or instead examine the full 200 to 980 nm spectrum. The former approach can be used if the analytical focus is a single element, whereas the latter is preferable either when one is interested in sensing all of the constituent elements of an unknown material as well as their relative abundances or one wants to compare the LIBS spectrum of an unknown with a library of reference spectra. In our laboratory work, which used an Ocean Optics 2000+LIBS spectrometer with a resolution of c. 0.1 nm (FWHM) and a 50 mJ Big Sky Ultra laser, the laser spark was formed under normal atmospheric conditions, as would be the case in the applications envisaged for field-portable LIBS. This being the case, each LIBS spectrum contains the peaks for N and O between 725 and 975 nm due to their presence in air (Fig. 2). Figure 3 shows a full LIBS spectrum for quartz (SiO2), with the most prominent peaks identified for Si (252.8, 251.6 and 288.1 nm), O (777.1 nm), and H (656.3 nm). The same N and O peaks that are present in the air spectrum (Fig. 2) are observed in the quartz spectrum (Fig. 3), but the O peaks are enhanced because oxygen comprises approximately 50% of the quartz sample. The hydrogen peak in the LIBS spectrum for quartz reflects the few per cent of structural water typically present in quartz. Figure 4 shows several metals of geological interest, illustrating the distinct character of each spectrum. Each metal is characterized by a unique broadband LIBS spectrum, suggesting that a statistical matching approach against a predetermined and assembled spectral library for materials of interest may be an effective approach for geochemical exploration using a field-portable LIBS system.

Fig. 3. Broadband LIBS spectrum for quartz with the peaks for Si (from quartz), O (from quartz plus air), and N (from air) identified.

Fig. 4. Broadband LIBS spectra for eight metals: Pb, Ni, Ag, Mo, Fe, Cu, Au and Zn. Each metal is characterized by a unique broadband LIBS spectrum, suggesting that a statistical matching approach against a spectral library may be an effective approach for geochemical exploration using a field-portable LIBS system.

Figure 5a shows full LIBS spectra for pure, reagent-grade NaCl, with the most prominent peaks for sodium at 588.9, 589.9, 568.9, 330.2, 330.3 and 616.1 nm identified. Figure 5b is a LIBS spectrum for a natural halite sample produced through evaporation of seawater. The more complex spectrum of Figure 5b compared to Figure 5a is due to the presence of the trace constituents Ca (393.4 nm, 396.8 nm, 317.9 nm, 316.8 nm), Mg (518.4 nm, 280.3 nm, 279.6 nm, 383.8 nm, 383.2 nm), K (766.5 nm, 769.8 nm), and Li (6707.8 nm, 610.3 nm, 323.3 nm) that were incorporated into the halite during its formation from the evaporation seawater. A LIBS spectrum for a pink tourmaline specimen is shown in Figure 6. Tourmaline is a complex aluminosilicate mineral typically associated with granites, granite pegmatites, and pneumatolytic veins characterized by the general chemical formula NaR3Al6[Si6O18](BO3)3(OH,F)4, where R = Mg and Fe or Mn for the end-member minerals dravite and schorl or Al plus Li for the mineral elbaite. The presence of strong Li and Al peaks and only a minor Fe peak in Figure 6 indicates that the tourmaline sample analysed is of the elbaite family. The examples cited above illustrate the potential for geochemical exploration via spectral matching based upon a predetermined and assembled spectral library for materials of interest. These two examples illustrate capability of LIBS to

24

R.S. Harmon et al.

Fig. 7. The prototype field-portable LIBS system used for the soil lead survey at US Army Sierra Army Depot, CA.

Fig. 5. Broadband LIBS spectra for (a) pure NaCl and (b) natural halite. Note the presence of the cations Ca, Mg, K and Li that were incorporated into the halite during its formation from the evaporation of seawater.

Fig. 6. Broadband LIBS spectrum for a laboratory specimen of pink tourmaline. The strong spectral lines for Li, Na and Al indicate that tourmaline belongs to the elbaite family.

detect trace constituents in natural materials, either in the laboratory or in the field. The ADA field-portable LIBS system The field-portable prototype LIBS instrument was produced in the mid- to late 1990s by ADA Technologies, Inc. As configured, this instrument (Fig. 7) consists of two parts: (i) a sample probe containing the laser and an optical fibre for signal detection; and (ii) a central detector/analyser unit that houses the spectrometer/detector, timing, power, and data acquisition and analysis equipment. The laser is a passively Q-switched

Kigre laser that provides a nominal energy of 15 mJ per pulse and can be fired once every 4 seconds. Upon focusing by a 45 mm focal length lens, the laser spot size is 60 µm. The instrument design utilizes a 0.5 mm round bundle to linear array fibre optic to transport the light from the plasma to the spectrograph. No focusing lens is used in front of the fibre optic to collect the light and the width of the individual fibres (100 µm) defines the aperture for the light input to the spectrograph. The LIBS analyser is run off an independent clock circuit, which triggers the laser, the detector and the data collection system. The detectors are read a few milliseconds after the plasma event and these readings form the basis for the spectral analysis. The laser power cables and fibre optic connect the two units. The complete LIBS instrument is contained within a 23 ( 51  38 cm aluminium case. This system has been operated from both a standard 12 V snowmobile battery and 115 V AC current. The fibre optic cable collects and transmits the light to a small spectrograph with a thermoelectrically cooled, 250  12-element charge-coupled diode detector. This design yields a 20 nm spectral range that can be adjusted to capture the elemental peaks of interest. Data are acquired and processed in a palmtop-type personal computer. Field results The soil samples analysed in this study come from a Pb-contaminated site at Sierra Army Depot, CA. The Sierra Army Depot (SIAD) is located in an arid setting in northeastern California, east of the Sierra Nevada mountains in the western portion of the basin-and-range physiographic province. The geological history of the region is complex, with recent Tertiary-age volcanism, block faulting, basin-fill sedimentation, and Quaternary lake and alluvial fan development superimposed on an older basement. The soil at SIAD consists of a thin surficial cover of highly permeable, windblown sand that overlies unconsolidated and low-permeability carbonate-rich lacustrine and fluvial lakebed sediments and alluvial fan deposits of sand, silt and clay. The study area at SIAD is an area known as the ‘Old Popping Furnace’ (OPF) site (Fig. 8). Here, a furnace was used from the 1940s to mid-1950s for the demilitarization of small arms ammunition by burning. Residual metal casings and Pb were recovered from the furnace, which was operated without air pollution controls, and the ash and solid furnace residues were buried locally at shallow depths in the soil around the furnace. The furnace was dismantled and removed some time after operations ceased, but the concrete pad on which the furnace was situated remains at the site. There is also physical evidence that small amounts of small arms ammunition also may have

Laser-induced breakdown spectroscopy

Fig. 8. Map of the Old Popping Furnace site at Sierra Army Deport, CA, showing measured total Pb in surface soils (modified from Harding Lawson Associates 2000). The shaded zone surrounding the concrete pad is the area in which surface soil Pb concentrations exceed 6000 ppm.

been burned on the surface in the area immediately surrounding the furnace. As a result, Pb contamination in the soils of the OPF site is most likely particulate material in the form of PbO from furnace ash and solid furnace residues. Bulk surface soil and samples from the Old Popping Furnace site were analysed for concentrations of 23 trace metals including Pb during an environmental survey conducted by Harding Lawson Associates under contract to SIAD (Harding Lawson Associates 2000). Eleven subsurface samples from five boreholes were also analysed for their Pb content. This survey recognized distinctly elevated and highly variable concentrations of several metals such as Pb, Cu, and Zn across the OPF site compared to background soil levels. Lead is the most prominent contaminant in soils of the Old Popping Furnace site, extending up to 180 000 mg kg1 Pb compared with background soil Pb levels that typically are in the range of 2 to 3 mg kg1. In general, Pb concentrations were highest to the east of the furnace site in the prevailing downwind direction and decreased with distance from the furnace site (Fig. 8), an observation which is consistent with the deposition of Pb on the soil surface from furnace stack emissions. However, the highest soil Pb levels were observed in an area c. 250 m SE of concrete pad, suggesting that this location may have been a site of furnace ash and residue disposal. The subsurface soil samples exhibited much lower Pb values, with concentrations declining to local ambient soil concentrations below 2 m depth. The portable LIBS system was employed to conduct a Pb survey at the SIAD OPF site. Five locations were sampled. Individual soil samples were collected over an area c. 5 cm and to a depth of c. 1 cm, homogenized, and sieved to remove large particles. A split of each sample was formed into a firm pellet in an aluminium dish using a small hydraulic press. The head of the LIBS analyser was placed on the soil pellet for analysis, the laser fired, and the laser head then moved incrementally by hand to a new spot on the pellet for the next shot. LIBS emission spectra from 400 to 410 nm for SIAD OPF soils, with each spectrum being the sum of the individual spectra for 10 laser shots of each soil sample, are shown in Figure 9. These data illustrate the potential of LIBS as a field sensor for qualitative use in geochemical applications such as mineral exploration and environmental contamination survey-

25

Fig. 9. LIBS spectra (10-shot averages) from 400 to 410 nm for Pb-contaminated soils along a 120 m traverse southeast from the Old Popping Furnace concrete pad (Fig. 8). Note the variable intensity of the 405.8 nm Pb emission line along the traverse compared to the peaks for the 404.4 nm K and 407.7 nm Sr emission lines. The relative intensities of the Pb peak along the traverse from OPF-11 to OPF-15 mimic the soil Pb concentration trend shown in Figure 8.

ing. The strong Pb enrichments that characterize samples OPF12SS and OPF13SS are clearly detected by the portable LIBS system. As illustrated by the two samples from the OPF13SS location, which were collected