1. Introduction

Spectroscopic analysis has advanced scientific knowledge of the planets by identifying chemical elements. In particular, the gamma-ray spectrometer [1] and alpha-proton X-ray spectrometer [2] on board Mars probes have accomplished various space missions in the past. Some known shortcomings of these techniques range from an extended time required for sample analysis and low sensitivity. Laser-Induced Breakdown Spectroscopy (LIBS) is a technique that has been proven possible for overcoming these known difficulties.

The LIBS analysis can be performed in real time without sample preparation. Also stand-off detection capability is particularly ideal for the mobile chemical sensor for space exploration of all sorts of material components. Of course there are disadvantages of LIBS such as matrix effect and difficulty of molecular detection. However, molecular detection is possible since molecular signal appears as a band like the CN band. Despite such disadvantages, real time analysis is still quite promising for reducing time and power consumption of identifying chemical elements in space missions [3,4].

LIBS is a type of atomic emission spectroscopy that uses a laser beam at high irradiance beyond 10⁹ W/cm² as an excitation source. At this high irradiance, the target surface temperature rises substantially above the vaporization temperature. The surface undergoes an explosive phase transition to a state where the gaseous molecules coexist with the electrons and positive ions. As a result, high pressure plasma and shockwave are generated. The emitted light from the plasma provides characteristic spectra for each element and by identifying the spectra, the chemical composition of each sample can be rapidly determined.

The LIBS plasma characteristics are strongly dependent on the ambient pressure [5]. In general, lower pressure causes a rapid expansion of the plasma volume, leading to a faster decay of the excited species’ number density and shorter plasma lifetime. However it is known that the signal to noise ratio increases at low pressure due to lower continuum and background signal. Figure 3 from Ref [6]. compares LIBS spectra taken at atmospheric condition and at vacuum conditions ~10⁻⁶ torr. Though the intensity of the LIBS spectrum taken at vacuum is less intense than the LIBS spectrum at atmospheric condition, it is clear that the LIBS spectrum at vacuum has higher resolution. In this study, we analyzed elemental lifetime of specific targets (Graphite, Ni, Cu, Sn, Al, Zn) staged within a vacuum chamber for analysis of the effect of surrounding pressure. The LIBS plasma peculiarity in the vicinity of 1 torr is identified and the possible cause of such interesting phenomena is given.

2. Experimental setup

A LIBS system typically consists of short pulse laser to generate plasma, a spectrometer for detection and analysis of the plasma light, optical components such as mirror and lens, and a computer. A vacuum chamber is added to analyze plasma characteristics in the low pressure. A Q-switched Nd:YAG laser (Surelite I, Continuum) which operates at 1064 nm with a pulse duration of 5-7 ns and output pulse energy of 45 mJ at 10 Hz repetition rate is focused onto the target surface inside of a vacuum chamber by a BK7 lens of 300 mm focal length. Single shot was applied to the experiment and it was repeated 5 times in each case.

The emitted light from the plasma was collimated on the entrance slit of the echelle grating spectrometer (Andor Mechelle), coupled to the gated ICCD (Andor, iStar), which allows the simultaneous spectral recording in the range 200 to 900 nm with a spectral resolution (λ/Δλ) of 5000. To collect the plasma, 50 μm fiber optic cable (ME-OPT-8004) was used.

Timing between laser pulse and spectrometer triggering (gate delay) was controlled by a delay generator (BNC 565 Digital pulse generator). The gate width was 50 μs and gate delay was varied from 100 ns to determine dynamic changes in the elemental lifetime of the target. The gate width is the acquisition time of the signal, and the gate delay is the time between laser irradiation and turning on the spectrometer. These temporal elements are important in enhancing signal to noise ratio and proper detection of spectra. During the first few microseconds after plasma initiation, the plasma energy is dominated by a strong continuum. Thus we set the optimum delay time for turning on the spectrometer to be one or two microseconds after the plasma initiation at atmospheric pressure. The electric signal output from the Q-switch of the laser and spectrometer was fed to a sampling storage scope (Tektronix TDS1000B) to check the delay time. The gain of ICCD was 150.

The targets used for the experiments are carbon (graphite), nickel, copper, tin, aluminum, and zinc. Although the targets are not standard materials, the purity of each target is above 95%. The targets were mounted on a XYZ stage inside of a vacuum chamber of 760 to 10⁻⁵ torr where a set of turbo and rotary pump is used to evacuate the chamber.

3. Results and discussion

Figure 1 shows the elemental lifetime of neutral atoms according to the ambient pressure. Several representative emission lines of each element based on the NIST database [7] were analyzed. The elemental lifetime was measured within the range of detectable limit with more than 5 on the signal to noise ratio. As shown in the figure, the lifetime of carbon decreased for the entire range of pressure reduction. The remaining metal targets all showed peak lifetime at 1 torr and started to decrease with pressure reduction until all the elemental lifetimes reached constant value below 10⁻² torr. This significant increase of elemental lifetime at 1 torr except for the non-metallic (carbon) receives a closer look. From the literature one finds that Yalcin and associates suggested a spectral enhancement of Al at 4 torr compared to atmospheric conditions at all gate delays [8]. The enhancement seen at 4 torr compared to atmosphere (760 torr) is likely a result of reduced plasma cooling at low pressures. Plasma cooling occurs when the ambient gas acts as cooling material for the hot vapor plasma [9]. Consequently, the lifetime of LIBS plasma created at 1 to 7 torr is much greater than that of a LIBS plasma created at an atmospheric condition. Another note from the literature is by Dreyer et al. where they observed a maximum spectral intensity of Ca, Mg, and Fe near 7 to 5 torr while decreasing intensity seen near 5 to 1 torr, suggesting a rapid decrease in electron density [10]. Also Cowpe et al. calculated electron number density from the Stark broadening of the 288.16 nm Si (I) emission line [11]. Figure 4 from Ref [11]. reveals that the step change in electron number density is evident, dropping by approximately an order of magnitude between ambient pressures of 7 and 0.7 torr. The electron number density is proportional to line broadening. Hence it is possible to get LIBS spectrum with higher resolution at low pressure condition compared with atmospheric pressure.

 

Fig. 1 Elemental lifetimes of neutral atoms in low pressure conditions (1 bar = 760 torr).

Download Full Size | PDF

These literature findings indirectly suggest a ‘soft spot’ in the low pressure zone near 1 torr. The present work attempts to clarify the noted feature in the vicinity of 1 torr where the maxima of elemental lifetimes exist for metallic samples by considering the full range of ‘vacuum’ pressure.

3.1 Characteristics of lifetimes with pressure change

3.1.1 High vacuum, 10⁻⁵<P<10⁻² torr

The elemental lifetimes of LIBS plasma are not affected by the pressure change in this region as the extremely rapid plasma expansion results in a very short elemental lifetime of less than a microsecond. It is too short a time to be influenced by the cooling effect and to be reduced by the plasma shielding, i.e., the laser energy coupled to the surface decreases due to the absorption of the irradiated energy by the plasma [12]. Typically the optimum delay time for turning on the spectrometer is one or two microseconds after the plasma initiation at atmospheric pressure because of the strong continuum. Below 10⁻² torr, however, spectra have to be detected at earlier times because LIBS plasma disappears before 1 μs. However the intensity of continuum and background signal also drops under this condition, making such LIBS detection still possible.

3.1.2 Medium vacuum, 10⁻²< P<1 torr

While the pressure is further lowered in the medium vacuum range, all of the elemental lifetimes declined for the pressure ranging from 1 to 10⁻² torr. In general, the effects of plasma cooling and plasma shielding gradually decrease with the decreasing pressure because there is not enough material to take up the heat. Besides, if the mean free path is too large, the small chance of collision makes generation of excitation and ionization difficult. For this reason, there are not enough excited atoms, ions, and electrons for collision with each other, and this causes the elemental lifetime to decrease, despite reducing the plasma cooling effect under this pressure condition.

3.1.3 Low vacuum, 1<P<760 torr

From 1 to 760 torr, most of the elemental lifetimes increase while carbon showed a decrease. As mentioned, the low pressure causes reduction in the plasma cooling and shielding effect. During a slow decrease in the plasma temperature, the atoms maintain their excited state for a prolonged period of time. Furthermore the small mean free path around 760 torr results in insufficient collision velocity for generation of strong atomic excitation. At a particular pressure state namely at 1 torr, optimal mean free path for atomic collision is allowed. Below 1 torr however, the mean free path is excessively large, and such the strong atomic excitation is very unlikely. These are just a few known complexities around low vacuum state, leading to the LIBS plasma peculiarity in the vicinity of 1 torr; namely, the maxima in elemental lifetimes of the samples (exception of a carbon) are reached. There is a strong evidence of proper pressure range that exists for optimal LIBS plasma spectroscopy in the low pressure conditions.

3.2 Analysis of elemental boiling point and electronegativity

In the case of carbon, the plasma lifetime is reduced with decreasing pressure due to a rapid plasma expansion as shown Fig. 2 . The difference of elemental lifetime characteristics results from the species’ physical and chemical properties such as boiling point and electronegativity.

 

Fig. 2 Elemental lifetimes of neutral carbon atom and C-N molecule.

Download Full Size | PDF

In the laser-material interaction at high irradiance, the target surface temperature rises substantially above the vaporization temperature. The species are excited and plasma is generated. The plasma generation is associated with the boiling point and heat of vaporization of target material. Under low pressure conditions, the plasma cooling effect is weakened, and thus most of elemental lifetimes increased except for carbon. As shown in Table 1 , carbon atom has high boiling point and heat of vaporization in atmospheric condition. This suggests that carbon is not affected by reducing the cooling effect because the carbon can condense at higher temperature. Zinc, however, has longest lifetime at 1 torr due to its low boiling point. Equation (1) is the Clausius-Clapeyron equation [13] which gives the pressure and temperature relationship during vaporization, where R ( = 8.3145 J mol⁻¹ K⁻¹) and Δh are the gas constant and heat of vaporization. For a given state of vaporization at state 1 (P₁, T₁), a new state (P₂, T₂) along the curve is known. The boiling temperature of each element can be given from

Table 1. Physicochemical properties of test samples [14]

View Table

Figure 3 shows plotted boiling temperatures of each element for various vapor pressures. One may infer a certain lower temperature than carbon curve where the effect of plasma cooling is insignificant. Below 3000 K at 1 torr for instance, the lifetime will be affected by reducing the plasma cooling as in all the metal samples. Then along this pressure dependent boiling curve, we can infer whether the specific element can be affected by reducing the cooling effect or not at full range of pressures.

 

Fig. 3 Estimated boiling point of each element.

Download Full Size | PDF

The emission line of carbon quickly disappeared with pressure decrease. At 760 torr, the neutral atom of carbon is detected until 5 μs. Shown in Fig. 2, above 1 torr, CN band is generated due to combination of carbon atom and nitrogen atom that exists in ambient atmosphere. The strongest CN band is appeared near 760 torr due to an abundance of nitrogen, and LIBS signal of a molecular band appeared because of the rotation and vibration of elements (Fig. 4 ). Table 1 shows carbon having the upper (or lower) limit of physicochemical properties amongst the tested samples. Both ionization energy and electronegativity increase with decreasing atomic radius. Electronegativity is a measure of the ability of an atom or molecule to attract pairs of electrons in the context of a chemical bond. Therefore despite the initial pressure of 760 torr, carbon atom line quickly disappears with pressure reduction because it combines easily with nitrogen atom; its large electronegativity being the main reason.

 

Fig. 4 LIBS spectra of CN band and neutral carbon atom from graphite target according to gate delay at 760 torr.

Download Full Size | PDF

There remains a possibility that additional factors may affect the lifetime variation since laser-material interaction is a complex phenomenon that often complicates itself with the plasma shielding effect, signal intensity, wavelength range of emission line, and generation of molecules by reactivity of elements. Nevertheless, it is possible to infer elemental lifetimes from the physicochemical properties of test samples.

4. Conclusion

In this study, elemental lifetimes of six different samples were investigated in a low-pressure chamber which has the pressure range of 760 to 10⁻⁵ torr. The peculiarity at 1 torr is observed for all targets except for carbon whose lifetime showed uniform decrease with lowering of the pressure due to a rapid plasma expansion. High boiling point and electronegativity are analyzed for explanation of the present observation of carbon. For the samples other than carbon, lifetimes increased during the prescribed pressure decline from 760 to 1 torr, and then decreased with continued pressure lowering from 1 to 10⁻² torr. Below 10⁻² torr, the lifetime of LIBS plasma was no longer affected by the pressure change. Possible reasons why maxima in the LIBS plasma lifetime for each sample occurred at 1 torr are believed to come from the plasma cooling and shielding effect, the plasma expansion, and the variation in the mean free paths for the surrounding pressure. The reported 1 torr peculiarity requires further investigation in order to better understand the elemental detection of the outer planet such as the Mars near 1 torr and the Moon at below 10⁻² torr.

Acknowledgments