**2. LIBS system design with modern technology**

a. Laser as a breakdown energy source

LIBS uses pulsed-laser light and focuses it onto the sample surface to make a plasma plume that contains the highly excited species of the sample composition. For generating plasma, there is a threshold value of the energy density. The threshold level will depend on the absorption coefficient of the sample surface of the laser wavelength, which is highly different by the sample phase. Gas and liquid need more energy to make breakdown. Solids with a dark color surface easily make a strong breakdown compared to clear or highly reflective solids. Figure 2 shows the effect of laser energy to make breakdown by the relation of laser power and focusing. Starting with a laser beam as 1 cm diameter, this light beam can be condensed by a convex lens. The focused beam density becomes 160 J/cm2 as in Figure 2(a). Also, the laser is operating in the pulsed mode, assuming a10 nsec duration, total power per unit of time will be 16 GWatt/ cm2 as

Pulsed laser

analysis.

**Figure 1.** The conventional LIBS system configuration

Plasma

analytical purpose or application area.

a. Laser as a breakdown energy source

power per unit of time will be 16 GWatt/ cm2 as

**2. LIBS system design with modern technology** 

carrier gas or at least use a new column temperature cycle. The application of LIBS also needs case-by-case adjustment. Many new applications start with looking at the advantages of LIBS and choosing a LIBS setup, and it still needs a detailed investigation for successful

Computer Sample

Spectrometer

Mirror

This chapter describes how the LIBS system works and explains the major parts of LIBS to select specific functional requirements for its intended application. The three major parts: laser, sample and spectrometer are explained. The laser provides the breakdown energy and plasma generation. Analytical sample is the target of the laser shot and the source of emission species. The spectrometer comprises detection system with light detector and computer. Their disadvantages and limitations are discussed then suggesting how to select the equipment type and configuration to maximize the advantages of LIBS. This will provide a beginning inspiration of LIBS systems to install and apply the desired specific

LIBS uses pulsed-laser light and focuses it onto the sample surface to make a plasma plume that contains the highly excited species of the sample composition. For generating plasma, there is a threshold value of the energy density. The threshold level will depend on the absorption coefficient of the sample surface of the laser wavelength, which is highly different by the sample phase. Gas and liquid need more energy to make breakdown. Solids with a dark color surface easily make a strong breakdown compared to clear or highly reflective solids. Figure 2 shows the effect of laser energy to make breakdown by the relation of laser power and focusing. Starting with a laser beam as 1 cm diameter, this light beam can be condensed by a convex lens. The focused beam density becomes 160 J/cm2 as in Figure 2(a). Also, the laser is operating in the pulsed mode, assuming a10 nsec duration, total

**Figure 2.** Laser energy delivery for breakdown condition. (a) focusing effect, (b) pulsing effect

in Figure. 2(b). Most of breakdown needs a few GW (106 Watt) of energy density, indicating that 50 mJ of laser energy is sufficient to make breakdown and evaporate most of material.

At the early stage of LIBS development, several types of pulse laser were used to make laser-induced breakdown plasma. An eximer laser was an important pulse laser especially for the UV light pulse. XeCl-eximer with 308 nm was used in the LIBS to measure elemental distribution on the paper coating[25]. The laser energy of 0.2 mJ was focused and made a crater of 30 µm diameter. This energy is corresponding to 108 W/cm2. More than 90 % of ingredients in the paper coating are pigment, binder and other agents. The pigment's main component is usually aluminum oxide, silicon dioxide and calcium carbonate. The mass of coating material ablated by single laser pulse was estimated to be about 2 ng by a laser shot. A typical nitrogen laser has a wavelength 337.1 nm and a pulse duration of 10 nsec. Just like the eximer laser, the nitrogen laser LIBS configuration in Figure. 3 also includes discharge from the wide shape of the electrode. The laser beam is usually a few cm wide, so a tight focusing is needed. The surface of solar cell was measured by nitrogen laser breakdown and only a 40-nm-thick TiO2 layer was detected[26]. The very popular pulse laser is Nd:YAG laser because it has a solid laser oscillator in a small size and light weight. The fundamental wavelength is 1064 nm with a pulse duration of 10 nsec typically. The Nd:YAG laser does not require any gas supply. The laser model for LIBS size usually has a closed loop water cooling that excludes external connection. A typical LIBS setup was shown in an earlier paper[27] as in Figure 4. A 50 mm focal length convex lens makes a simple optics configuration to make plasma on the sample.

b. Optical arrangement for laser–induced breakdown spectroscopy

When a laser shoots on the sample surface, a plasma plume arises from the inner to the outer surface. The actual size of plasma plume made by a 100 mJ laser pulse will be few millimeters. During the plasma propagation from the sample surface, the time profile features can be observed. The very initial emission is generated at the bottom of the plasma

**Figure 3.** A LIBS setup with nitrogen laser.

**Figure 4.** Schematic diagram of the LIBS setup with ND:YAG laser

plume, and then expanded to the outer plume. Depending on the light collecting optics, plasma propagation is captured at the different time. At the initial LIBS design uses a sideview emission collection as in the Figure. 5 (a). The angle between the laser light path and the collecting optic can be any angle, but is typically 30-45 degrees. Some experiments use 90 degrees, which is a complete side view of the plasma and will lose some portion of emission by the shadow of the sample itself. This configuration is occasionally used for plasma physics study. The collateral view design Figure. 5 (b) is a useful optical configuration for non-fixed sample distances. Laser light path shares emission collection optics. A selective wavelength reflector or prism can be used to separate laser light and emission through the light path. This design has several optical parts and needs complicated adjustments for optimum light measurement. The collateral configuration has many advantages. Collecting optics looks at the plasma in front of the plasma (or top of the plasma)at every point in the light axis and in the focus cone, which means they capture every light emitting species during plasma propagation to the space. Because some elements have different propagation profile than others, propagation height changes the signal significantly at the angled collection. The next advantage is that the optical part can be integrated in the compact enclosure, and it allows the operator to move the optics (detector head of LIBS) more freely. Remote monitoring LIBS, hand held design LIBS, should be compact and have a mostly collateral optics configuration.

**Figure 5.** Side collection and (b) collateral collection configuration of plasma emission

#### c. Sampling technique

134 Advanced Aspects of Spectroscopy

**Figure 3.** A LIBS setup with nitrogen laser.

**Figure 4.** Schematic diagram of the LIBS setup with ND:YAG laser

plume, and then expanded to the outer plume. Depending on the light collecting optics, plasma propagation is captured at the different time. At the initial LIBS design uses a sideview emission collection as in the Figure. 5 (a). The angle between the laser light path and the collecting optic can be any angle, but is typically 30-45 degrees. Some experiments use 90 degrees, which is a complete side view of the plasma and will lose some portion of emission by the shadow of the sample itself. This configuration is occasionally used for plasma physics study. The collateral view design Figure. 5 (b) is a useful optical configuration for The first mentioned advantage of LIBS has been no-sampling step. In the very beginning review in the *Encyclopedia of analytical Chemisstry*, Yueh, Singh and Zhang described it as "LIBS uses a very small amount of samples, and no sample preparation is necessary. It has the ability to perform real-time analysis because it prepares and excites the sample in one step". They then consecutively mentioned, "The disadvantage of LIBS is that the plasma conditions vary with the environmental conditions as well as the laser energy fluctuation."We can infer from the description of LIBS that no-sampling is both an advantage and a disadvantage. Most of analytical the techniques need a certain sampling procedure to bring the sample to the technique (or machine). During the sampling procedure, like the acid digestion in the flame analysis, the sample is homogenized and their matrices become concordant. However, if LIBS analyzed the sample without any pretreatment process, then the irregular homogeneity is inevitable. As a result LIBS will include severe matrix effects at the real sample. It will mitigate the biggest advantage, i.e., nosample process. In other words, if the sample is measured as it is, the species in same concentration do not make a consistent signal, the analytical result will be severely diverted.

**Figure 6.** Solid sample and liquid sample under LIBS measurement

The fluctuation will be more serious because LIBS takes only a small amount of the sample, usually a micron sized spot. Two possible sample types are depicted in the Figure. 6. A solid sample is the most convenient sample type. Metal and ceramic samples include elements with strong atomic and ionic emission. Their emission spectra are measured at the range from UV to visible light, which is feasible by the most spectrometers. The spectra from many elements from the tool steel are shown by the Nd:YAG laser excitation[28]. In this research, the microscopic view of the ablated holes made on tool steel is about 10 microns in diameter. This resolution indicates that any inhomogeneity more than 10 microns will be clearly observed from each laser pulse measurement. The intensity of element-specific spectra provides a simple qualitative analysis. Their method was sufficient to characterize the nature of the defect by a simple estimation of the elemental composition between the basic material and the defect.

#### d. Capturing emission light

The LIBS signal is instantaneous and decays quickly. Temporal control of the detecting device is very important. In spite of the fact, overall emission can be captured by opening entire time of the spectrometer, most of LIBS measurement is controlled by time gate operation. Time control improves the signal-to-noise ratio by eliminating the continuum emission. A typical emission profile shown in the Figure 7, recorded at the different heights from the sand/ soil mixture sample[29]. As soon as the laser fires with the duration of a few nsec of pulse width, the plasma intensity is propagating outward from the sample surface. At about 0.5 µsec, plasma is observed at 0.3 mm away from the surface. At the propagating distance is 3 mm at 12 µsec, then plasma cool down with decreasing intensity until 20- 30 µsec. The plasma size will be much smaller and life is shorter when a weak laser power is used. The experiment uses aluminum[30] with a diode-pumped Nd:YAG laser, which can run at a faster repetition rate (kHz) with a laser energy of 80 µJ, was setup under the microscope excitation and detection optics. Like other flash lamp pumped lasers, the temporal profile of continuum emission is shown in Figure. 8 for aluminum atom (Al 396.1 nm) and aluminum ion (Al II at 358.6 nm) emission lines. The laser pulse was fired at the zero time of the x- axis. This profile indicates, the broad band continuum emission, which comes from high temperature heated plasma and regardless of the species in the plasma, has a lifetime of about 13 nsec. The ionic line from Al ion has shorter lifetime about 24 nsec. The neutral lines stay much longer, up to 80 nsec.

136 Advanced Aspects of Spectroscopy

material and the defect.

d. Capturing emission light

severe matrix effects at the real sample. It will mitigate the biggest advantage, i.e., nosample process. In other words, if the sample is measured as it is, the species in same concentration do not make a consistent signal, the analytical result will be severely diverted.

The fluctuation will be more serious because LIBS takes only a small amount of the sample, usually a micron sized spot. Two possible sample types are depicted in the Figure. 6. A solid sample is the most convenient sample type. Metal and ceramic samples include elements with strong atomic and ionic emission. Their emission spectra are measured at the range from UV to visible light, which is feasible by the most spectrometers. The spectra from many elements from the tool steel are shown by the Nd:YAG laser excitation[28]. In this research, the microscopic view of the ablated holes made on tool steel is about 10 microns in diameter. This resolution indicates that any inhomogeneity more than 10 microns will be clearly observed from each laser pulse measurement. The intensity of element-specific spectra provides a simple qualitative analysis. Their method was sufficient to characterize the nature of the defect by a simple estimation of the elemental composition between the basic

The LIBS signal is instantaneous and decays quickly. Temporal control of the detecting device is very important. In spite of the fact, overall emission can be captured by opening entire time of the spectrometer, most of LIBS measurement is controlled by time gate operation. Time control improves the signal-to-noise ratio by eliminating the continuum emission. A typical emission profile shown in the Figure 7, recorded at the different heights from the sand/ soil mixture sample[29]. As soon as the laser fires with the duration of a few nsec of pulse width, the plasma intensity is propagating outward from the sample surface. At about 0.5 µsec, plasma is observed at 0.3 mm away from the surface. At the propagating distance is 3 mm at 12 µsec, then plasma cool down with decreasing intensity until 20- 30

**Figure 6.** Solid sample and liquid sample under LIBS measurement

**Figure 7.** Spectra as a function of decay time measured at three observation distance from the sample surface. The original figure is rearranged to indicate observation height more clearly.

The lifetimes of laser-induced plasma are easily compared at various excitation energies from the silicon sample[25]. The time profile shows the plasma emission signal depends on the excitation pulse energy. Absolute intensity of the signal will increase by increasing laser pulse energy. The decay plot in the Figure 9 is normalized to a maximum intensity for comparison. This research explains the decay time dependence by the excitation energy that the probability of excitation to higher energy level is increased and more populated, leading to a longer decay time. Also, the upper state of the monitored transition receives population from this higher state at later times and lengthening of the rise time of the signal

**Figure 8.** Temporal profile of continuum emission and aluminum (atom and ion) emission.

will result. As a result, the lifetime point of LIBS will be changed by the system setup, especially using laser power. Capturing time of emission signal should be determined empirically by looking at the profile, usually at peak intensity point.

**Figure 9.** Time resolved signal-to-background ratio of the silicon line at 251 nm at various excitation energies.

#### e. Spectrometer and detector

Spectrometer completes the detecting part with a photo sensor and a manipulating computer. The spectrometer must have proper resolution and sensitivity. Also, in many cases the plasma emission needs to be separated from the continuum background signal, the detector has to be operated by timing control or gating operation. Various types of spectrometer with CCD array detector are available in the market. The wavelength range needed for LIBS is UV to visible range to have detection of most elements. If the dispersion of the spectrometer is 0.3 nm to measure 1 nm peak with three pixel, the pixel to pixel dispersion should be 0.3 nm. Total of 1024 pixel CCD array can have coverage 1024 x 0.3 = 307 nm, which can assign the range as 250 nm to 557 nm span. In many cases, the sample will have mixed elements and the emission lines will be overlapped and difficult to distinguish with 0.3 nm resolution. A conventional CCD array detector may not provide sufficient resolution and coverage to measure LIBS.

**Figure 10.** Echelle spectrometer dispersion image (a) Hg lamp, (b) LIBS spectrum of Sn metal

A correction of the array detector resolution is accomplished using multiple stacked spectrometers. For example, 5 spectrometers with 1000 array CCD stacks will cover a 500 nm span, in which each spectrometer covers a 100 nm range with 0.1 nm resolution. Echelle spectrometer uses very high orders of dispersion. One or two prisms are used to separate each diffraction order. As a result, the spectra are dispersed in two dimensional surfaces as shown Figure 10. The CCD detector in the Echelle spectrometer should be a two dimensional, the same as in the image camera. The continuum emission from the spark also engages in the Echelle spectrometer, so the detector must have gated operation. To satisfy those requirements, such as two dimensional, sensitive and gated operation, the cost of CCD detectors for the Echelle spectrometer is still significantly high.
