Elemental Analysis of Heated Soil Samples Using Laser-Induced Breakdown Spectroscopy Assisted with High-Voltage Discharges

Abstract

In this study, a high-voltage (HV) oscillatory square pulse was used to re-excite the laser-induced breakdown spectroscopy (LIBS) signal produced by a nanosecond laser on different loamy soil samples at two different temperatures: ambient and 400 °C. The optimal delay was found for each experimental scheme to maximize the emission signal-to-noise ratio. The detection limits of various contaminants in the soil were studied for different experimental configurations. It was found that the lowest detection limits were achieved by combining HV discharges with LIBS on heated samples, resulting in improvements of up to a factor of 7 compared to LIBS on room temperature samples. Plasma characterization shows that the increased detection sensitivity is due to the rise in plasma temperature and electron density with HV re-excitation, while an increase in removed matter contributes to the emission intensification observed when samples are heated.

1. Introduction

The assessment and management of soil quality, including nutrient and toxic elements, is critical for the sustainability of agricultural production and food safety. This, in turn, has a potential impact on human health [1]. For this, it is necessary to know and evaluate the chemical, physical, and biological properties of the soil, such as nutrient content, pH, and organic matter content, among others [2].

Several techniques are necessary to evaluate these properties, such as atomic absorption spectroscopy (AAS) or inductively coupled plasma optical emission spectroscopy (ICP-OES). Most of these techniques require elaborate sample preparation meaning it takes several hours, or even days, to see the results [3]. For this reason, these techniques are not suitable for real-time measurements. In this context, faster analysis techniques represent an advantage, such as laser-induced breakdown spectroscopy (LIBS). The development of portable LIBS-based measurement systems that meet the demands of in-line and in situ analytical applications is feasible [4].

The LIBS technique is a spectrochemical analysis tool for a wide variety of materials that has several advantages such as competitive cost of analysis, no additional sample preparation required, it is mostly non-destructive, and can analyze a large number of materials without the need for vacuum systems or a controlled atmosphere, etc. Studies can be performed either in situ or remotely, allowing work in hostile environments [5,6]. LIBS analysis is a promising method for developing rapid analysis methodologies [7].

The application of LIBS to analyze soil samples is discussed in the review paper of Villas-Boas et al. [8,9]. The difficulties generated by the intricate matrices of soils, due to their heterogeneous chemical composition and physical structure, are mentioned there. For this reason, specific calibration models have been developed for each type of soil with similar composition to minimize matrix effects. LIBS has also been used to infer the soil’s physical and chemical properties [10]. In the work of Santos et al. [11], several problems are mentioned for quantitative analysis of complex soil and sediment matrices due to their inhomogeneous character. Toxic elements are naturally present in soil, generally in low concentrations. However, these elements can accumulate due to anthropogenic activities such as mining, fertilizer application, pesticide use, wastewater irrigation, and petrochemical spills, among others. Some of the toxic elements that contaminate the soil are lead, chromium, arsenic, zinc, cadmium, copper, mercury, and nickel [12].

Due to the aforementioned advantages, LIBS is a suitable method for soil contamination assessment. It has great potential for application in environmental sustainability studies, especially for providing fast and reliable results without generating chemical residues [13,14]. In addition, the LIBS technique is an alternative for obtaining geochemical data, analyzing the natural distribution of elements in the earth’s crust, and conducting environmental studies [15]. In spite of all the advantages mentioned above, the technique is limited by its low sensitivity and poor accuracy in the detection of trace elements [16].

In order to increase the trace detection sensitivity of LIBS, several methods have been implemented. Some of them employ spatial confinement of the plasma, others use a second laser, re-excitation with microwaves or a high voltage pulse, etc. [17]. In particular, the use of a spark discharge has the advantage of being inexpensive and relatively easy to implement. This configuration allows elemental analysis of both conductors and dielectrics [18]. For example, it has been used to detect phosphorus in fertilizer samples [19] and to detect copper in onion leaves [20]. In soil analysis, Wang et al. [21] used LIBS assisted with a low power unipolar arc discharge to improve sensitivity detection of cadmium. Srungaram et al. [22] found a significant improvement in the detection of mercury in soil samples using a high voltage discharge. Likewise, Kexue et al. [23] reported the analysis of lead and arsenic in soil samples with LIBS assisted by a fast electric discharge of 11 kV. In this instance, an increase in the signal-to-noise ratio of a factor of three was obtained. Using similar conditions (LIBS combining 11 kV fast electric discharge), Xiafen et al. [24] analyzed lead in soil samples and reduced the detection limit.

Sample heating has seldom been used for the improvement of detection sensitivity in LIBS. Increasing the sensitivity of trace detection by heating samples has mainly focused on the study of metallic samples [25,26]. Recently, a study was published showing its potential applicability for soil analysis [27]. However, in this work, it was restricted to a single element and a single acquisition delay.

The aim of this work is to investigate the increment in the detection sensitivity of trace contaminants present in soils when a high voltage discharge that re-excites the LIBS signal is used in combination with sample heating. The proposed configuration has the advantage of being affordable while maintaining the portability of the system. The investigation is focused on the study of several contaminant elements present in reference loam soil samples. The results were compared using four different experimental schemes: LIBS alone, high-voltage-assisted LIBS (HV-LIBS) on samples at room temperature and heated to 400°C temperature. Finally, the improvement of the LIBS signal was investigated by measuring the matter removed through laser ablation, characterizing the plasma temperature and its electron density.

2. Materials and Methods

The schematic diagram of the experimental setup is shown in Figure 1. The ablation laser employed was Nd:YAG laser (Surelite III from Continuum). It was operated at its fundamental wavelength of 1064 nm, with a repetition rate of 0.5 Hz and an energy of 50 mJ. The laser output has a beam diameter of 9 mm and a beam propagation ratio M² = 4 for the employed wavelength. The laser beam was focused on the sample with a 10 cm focal length plano-convex lens, producing a spot diameter of approximately 160 µm corresponding to an irradiance of about 50 GW cm⁻². The samples were placed on a x-y-z translation stage to adjust the focal distance and to shift it to an adjacent position every 10 laser shots to analyze different spatial regions. Experiments were performed under atmospheric pressure, at a temperature of 22 ± 1 °C and a relative humidity of 30 ± 5%.

The spatially integrated light emitted by the plasma was collected with a 5 mm diameter quartz collimator located at about 45° from the sample and directed to an echelle spectrograph (Aryelle 200 from LTB Lasertechnik) by means of a 400 μm diameter optical fiber. The spectrograph spans a range of 210–820 nm and has a spectral resolving power of 9000. The spectra were detected with an intensified charge-coupled device (ICCD) camera (334-18F-03 from Andor). The laser and the ICCD were synchronized by an eight-channel delay generator (575 from Berkeley Nucleonics). In most experiments, the light collection was time-integrated as a gate width of 100 µs was used for the ICCD, while its delay varied within the 10–70 μs range. For time-resolved plasma temperature and electron density determination, a gate width of 100 ns was employed.

In this work, we investigated the effect of heating the soil samples to improve the sensitivity of LIBS to detect contaminant traces in soils. For this purpose, the sample was placed on a steel plate that was heated by means of a 500 W halogen lamp. The temperature was fixed and maintained at 400 ± 20 °C and was monitored through a thermocouple.

In half of the experiments reported here, the laser-induced plasma was re-excited by means of a high voltage discharge [28]. Usually, this technique involves the discharge of a single (lumped) capacitor over the sample. In such a case, the current applied oscillates at a high frequency (dependent on the parasitic inductance) and, consequently, the power is applied only during short periods of time with several interruptions in the power delivered to the plasma each time the current crosses the zero value. In the present work, a 50 m long, RG 58 coaxial cable with a total capacitance of 5 nF played the role of the capacitor and was charged through a 10 kΩ resistor by a high-voltage power supply (Bertan 205B). It is well-known [29] that when a coaxial cable is discharged through a short circuit the current that flows has a bipolar square pulse shape. In this scheme, the current remains at a constant peak amplitude during 500 ns in any semi-cycle before it swings back to the opposite polarity where it again remains constant for that amount of time, and so on.

The electrodes employed were hemispherical capped rods 6 mm in diameter made from a commercial brass alloy containing ~63% Cu, 33% Zn, 3.7% Pb, and other elements. These elements were not included in the analysis of this experiment because part of the electrode evaporated during the discharge. The current was monitored by means of a Rogowski coil on the low voltage side of the circuit, connected to a digital oscilloscope (DPO 4104B from Tektronix). Finally, the crater analysis for the different experimental condition was performed with a profilometer (Dektak IIA from Veeco).

In this work, the targets used were loam soil certified reference materials from High-Purity Standards, LLC. The major elements are Si, Al, Ca, and Fe, while K, Mg, Mn, and Ti are minority elements.

Table 1. Concentrations of investigated traces present in loam soil samples of certified reference material (CRM).

Element (Symbol)	Loam Soil B (µg g−1)	Loam Soil C (µg g−1)	Loam Soil D (µg g−1)
Cd	84.3 ± 5.9	1.2 ± 0.2	0.520 ± 0.017
Cr	52.5 ± 3.7	42 ± 4	269 ± 10
Pb	107 ± 8	64 ± 7	129 ± 6
V	80.6 ± 5.8	57 ± 8	127 ± 5
Zn	231 ± 16	164 ± 20	104 ± 7
As	49.5 ± 5.9	52.6 ± 2.4	29 ± 1.4
Be	9.27 ± 0.74	1.3 ± 0.2	2.35 ± 0.06
Ni	55.4 ± 3.7	141 ± 20	51.6 ± 3.7

shows the certified composition of the trace elements investigated. To perform the tests, the samples were molded into a 1.25 cm diameter and 3.7 mm thick disk shape using approximately one gram of material that was pressed to 2000 kg/cm² for 1.5 min.

3. Results

3.1. Electric Circuit Characteristics

As mentioned previously, half of the experiments use an electrical discharge produced by the discharge of an RG58 cable charged up to 12 kV. When a plasma is induced on the surface of the sample it closes the circuit, and a pulse of current is sent to ground through the second electrode. The height and distance between the electrodes were adjusted to create a “V”-shaped arc. Different distances were used in order to control the discharge delay and investigate the effect on the LIBS signal. Figure 2 shows typical measurements of the discharge used to re-excite the laser-produced plasma. In this case, the time delay with respect to the ablation onset was 1.5 µs and the electrode spacing used was approximately 5 mm at a height of 2 mm above the target. As can be seen, the current applied was bipolar, reaching a maximum of 190 A during the first 500 ns and it quenches in the span of 10 µs. Figure 2a also shows that the voltage is maintained during discharge at a relatively low value due to the low resistance of the plasma, which can be estimated to be of the order of a few ohms.

From the product of current and voltage, the power as a function of time was obtained. Subsequently integrating its absolute value, the total energy deposited in the laser-produced plasma was calculated to be 220 mJ (see Figure 2b). It is evident from the graph that 80% of the energy is deposited in the first 3 µs after the onset of the HV pulse. The energy dissipated by the HV discharge is greater than that produced by the ablation plasma, but this is deposited in a much longer time. Thus, the re-excitation process occurs during the first few microseconds and thereafter, the plasma expands adiabatically.

3.2. Spectroscopy Results

Before the tests, an optimization process of the ICCD acquisition time delay was performed to maximize the signal-to-noise ratio (SNR) of the emission. The process was carried out using major elements because these have a higher SNR and have a smaller variation in their concentration in the sample than a trace. The analysis considered both ionic and neutral lines since the maximum may occur at different times due to the variation in plasma temperature. Moreover, the analysis was performed for the different experimental conditions used in the work: single-pulse ablation (LIBS), high voltage re-excitation of the LIBS signal (HV-LIBS) with the sample at room temperature, and then repeated with it heated to 400 °C.

Over 100 individual spectra were acquired at 10 different spots and averaged for different acquisition time delays. Lines were adjusted with a Lorentzian profile to obtain the peak-to-base intensity. Subsequently, the SNR was calculated from the ratio between the line height and the standard deviation of the noise near the transition. Figure 3 shows the results obtained for one ionic and one neutral silicon transitions obtained for the LIBS and HV-LIBS experiments, both performed at room temperature. The error bars were obtained from the standard deviation of the error-propagated quotient of the emission intensity and noise values. As can be observed for the laser ablation scheme, the best SNR is obtained in the interval between 500–1000 ns (see Figure 3a). However, as expected, neutrals reach the maximum at a later time. In the same way, when heating the sample, the delay that maximizes the SNR is 700 ns for the ionic lines and 1.5 µs for the neutral lines.

The experiment was then performed applying high voltage for inter-pulse delays between 1–6 µs. Keeping the voltage fixed at 12 kV, the delay was adjusted by varying the separation of the electrodes. For each condition, the ICCD acquisition onset time was varied, with time t = 0 being the laser pulse application. Figure 3b shows the results obtained for the case where the inter-pulse delay is 2 µs. As can be observed, the ionic lines show a maximum at around 3–4 µs after laser onset, while for the neutral lines, the SNR maintains a stable value for several microseconds. Thus, for each experimental condition used, the condition that maximizes the SNR was found. For the HV-LIBS configuration, the optimal delay between pulses is about 2–3 µs and the acquisition started about 1 µs after application of the HV pulse for ionic species and 2–3 µs for neutral species. Note that the optimal delays also depend on the element analyzed and the specific transition. For example, the increase in SNR for neutral transitions is, in general, lower than for ionic transitions.

Figure 4 shows the results for selected major, minority, and trace elements. The results shown in this figure correspond to a comparison of the results obtained for each experimental setup under optimal conditions for maximization of the SNR. The improvement in the signal intensity for HV-LIBS with heated samples is up to seven times better than that of LIBS at room temperature. It is also shown that under these conditions, it is possible to observe the arsenic present in the loam D sample, with a concentration of 29 µg g⁻¹; this element is not observed in the other experimental configurations used in the present work. Furthermore, it can be observed that no noticeable enhancement of the background signal is obtained. Therefore, the increase in line intensity leads to a straight enhancement of the detection limit.

Subsequently, 100 individual spectra were acquired at ten spots under the different conditions used. The results were averaged to obtain the SNR and the detection limits. For each element, lines that had no spectral interference and maximized the sensitivity of trace detection were chosen.

Table 2. Limit of detection (LOD) of LIBS compared to the values obtained by intensification techniques.

Species; Transition (nm)	LIBS (20 °C) (µg g−1)	LIBS (400 °C) (µg g−1)	HV-LIBS (20 °C) (µg g−1)	HV-LIBS (400 °C) (µg g−1)	LOD Enhancement	Lower Guideline Value (µg g−1) a,b	Reference Values (µg g−1)
Cd II; 214.44	10	7.6	3.2	1.5	6.7	37; 25	1.3 [11]; 57 [21]
Cr I; 520.60	8	2.3	15	2.2	3.6	280; 42	17 [7]; 8 [32]
Pb I; 405.78	35	17	-	-	2.1	400; 220	20 [7]; 61 [33]
V II; 311.07	24	10	19	6.4	3.7	78; 26	260 [13]; 65 [34]
Zn I; 481.05	62	27	-	-	2.3	-; 270	43 [13]; 31 [13]
As I; 228.81	LD	LD	LD	15	-	22; 20	4 [23]; 30 [35]
Be II; 313.04	0.57	0.25	0.28	0.14	4.1	150; 25	10 [36]; 0.07 [37]
Ni II; 230.30	14	4.8	8	4.7	3.0	1 600; 100	8.87 [13]; 7.86 [13]

^a total reference concentrations for agricultural land use according to Mexican standard [30]; ^b priority contaminants of ecological concern (Dept. Ecology, Washington State, USA) [31]. LD: Below detection limit.

shows the limits of detection (LOD) obtained using single-pulse laser ablation and applying intensification techniques. For reference, the Mexican regulation for maximum concentrations of contaminants in soil for agricultural, residential, and/or commercial use are shown alongside the Washington State Department of Ecology’s [30,31] ones. These are reference concentrations that, if exceeded, mean the soil is considered contaminated and represents an ecological or health risk. The table also shows some reference values obtained with LIBS and using LIBS assisted by other intensification enhancement techniques [7,11,13,21,23,32,33,34,35,36,37]. However, the comparison cannot be made directly since all these works have been carried out with different re-excitation methods, irradiances, with different ablation wavelengths, and using intensified and non-intensified detectors. In addition, it must also be taken into account that, in some works, the soil samples have been artificially contaminated. This does not guarantee reliable results because it is not clear whether the signal obtained by LIBS for a given contaminant does not vary with time. On the other hand, it is worth mentioning that lead and zinc are not reported using HV-LIBS because these elements are present in the electrodes. This was obtained by performing a qualitative analysis of the electrodes through LIBS.

The detection limits achieved by LIBS are below the reference values shown for most of the contaminants except for arsenic, which is not observed. In addition, for Cd and V, the limits of quantification LQ (3.3 higher than LOD) are higher than the shown reference values. It is observable for all cases that the combination of the techniques shows an increase in the sensitivity of trace detection and the values obtained are below recommended LOD concentrations. In addition, it is shown that the decrease in LOD is generally greater for ionic species. This is expected, as re-excitation processes such as HV discharges generally increase the plasma temperature and, consequently, ionic species are more favored than neutral species [28,38,39,40,41]. As a consequence, no improvement in the detection limit is observed in the case of chromium. Possibly for this element the optimum delay for acquisition or re-excitation might be different.

For most of the investigated elements, the increase in LOD when using the HV-LIBS configuration ranges between two and four, except for cadmium, for which an improvement of a factor of nearly seven is obtained, reaching a LOD of 1.5 µg g⁻¹. This value is similar to that obtained in the work of Santos et al. [11], where a much higher ablation energy of 365 mJ was employed combined with cryopreservation of soil samples. Arsenic could not be measured with LIBS because it is below the detection limit under the experimental employed conditions. However, by applying the signal enhancement techniques, it could be detected, reaching a LOD of 15 µg g⁻¹.

Decreasing the detection limits is useful for this type of sample, because the complex matrices of the soils make the quantification of trace elements challenging. Thus, the improvement in detection sensitivity obtained by combining intensification techniques is a viable and cost-effective approach.

3.3. Plasma Characterization

To better understand the physical mechanisms behind the observed increase in signal intensity, the temporal evolution of the plasma temperature and electron density were obtained for all the experimental schemes employed. Assuming local thermal equilibrium [42], the electron temperature was determined through the Saha–Boltzmann method from neutral and ionic magnesium emission including: 277.98, 285.21, 383.83, and 518.36 nm belonging to Mg I and 279.08, 279.55, 279.80, 280.27, and 292.86 nm to Mg II. The method is a modification of the normal Boltzmann plot method to obtain the plasma temperature using two consecutive ionization degrees of an element and was originally proposed in ref. [43]. Thus, the emissivity from level j to i of a ionic state z of a transition ${\epsilon }_{ji}^z$ is written in terms of the plasma temperature and its electronic density as:

$$ln\left(\frac{{\epsilon }_{ji}^z {\lambda }_{ji}}{A_{ji}^z g_j^z}\right)-zln\left(2 {\left[\frac{2 \pi m k}{ h^2}\right]}^{\raisebox{1ex}{$3$}\!\left/ \!\raisebox{-1ex}{$2$}\right.} \frac{T^{\raisebox{1ex}{$3$}\!\left/ \!\raisebox{-1ex}{$2$}\right.}}{n_e}\right)=-\frac{1}{k T} \left(E_j^z+E_{\infty }^0+\mathsf{\Delta }{E}_{\infty }^0\right)+ln\left(\frac{h c N}{Q_{}^0\left(T\right)}\right),$$

(1)

where ${\lambda }_{ji}$ is the transition wavelength; $A_{ji}^z$ is the transition probability; $g_j^z$ the level degeneracy; m is the electron mass; k and h are the Boltzmann and Planck constant; z = 0,1 correspond to neutrals and ionic species, respectively; T the plasma temperature; n_e the electron density; $E_j^z$ is the energy of the upper level; $E_{\infty }^0$ is the ionization energy of neutrals; and $\mathsf{\Delta }{E}_{\infty }^0$ is a correction to ionization potential due to interaction in the plasma. Finally, c is the speed of light, N is the number density of neutrals, and $Q_{}^0\left(T\right)$ is the partition function. To obtain the plasma temperature, a guess value for the temperature has to be assigned first to the left hand side of the equation. Afterwards, an iterative process is performed to obtain a linear plot with a slope 1/kT and from it the plasma temperature is obtained as a function of time. It is necessary to previously calculate the electron density, which is a function that slowly depends on the temperature. This value is obtained from the broadening of the lines, which are dominated by the Stark effect. Therefore, the electron density was obtained from the broadening of the lines through [44]:

$$n_e=\mathsf{\Delta }{\lambda }_{Stark}\frac{n_{ref}}{ w_{Stark}}$$

(2)

where $\mathsf{\Delta }{\lambda }_{Stark}$ is the full width at half maximum of the lines after subtracting the instrumental contribution, $w_{Stark}$ is the Stark broadening parameter, and $n_{ref}$ is the electron density employed as reference. In this work, the following transitions were used to obtain $n_e:$ 279.08 nm (Mg II), 281.62 m (Al II), and 285.21 nm (Mg I). In addition, the width of the H alpha line (656.28 nm) was used to further corroborate the results obtained, while the line it was observable (up to 5–7 µs) [45].

Figure 5a shows the spatially integrated time evolution of the electron density (n_e) for the different configurations. The error bars in each case correspond to the standard deviations of the n_e values obtained from different transitions. As can be observed, the n_e for the laser ablation on samples at ambient temperature and high temperature exhibit a similar pattern. In all cases, initially, the densities decay exponentially in the first few hundred nanoseconds. It is evident that the lifetime (up to 70 μs) of the plasma produced with the hot sample is longer than that of the other experimental conditions. In fact, the 285.21 (Mg I) transition was detected up to 70 µs, which contributes to the observed increase in emission. Also shown in the figure is the fact that after the arc strikes (t > 1.5 μs), the electron density sharply increases by a factor of two. However, after most of the electric power has been deposited, the electron density rapidly drops and reaches values similar to those measured for the LIBS experiment at 10 µs.

Figure 5b shows the calculated evolution of plasma temperature. The error bars stem from the linear fit error in the Saha–Boltzmann plots. It is observable that the plasma temperature in laser ablation decays exponentially over time, as expected. It is also noted that the heating of the sample alone does not result in an increase in plasma temperature or electron density. However, in experiments where a high-voltage discharge is applied, the plasma temperature is found to be somewhat higher compared to the laser ablation scheme’s results. This suggests that there are more ions that can potentially emit, leading to an increase in signal intensity. Thus, it can be assumed that the high-voltage discharge mainly contributes to the re-excitation of the laser-produced plasma.

3.4. Crater Analysis

The observed increase in emission could also be attributed to an enlarged number of emitters as more material is removed from the target. To estimate the ablated mass for the different configurations, a profilometer was used. Figure 6 displays the crater profiles formed by three laser shots for each experimental condition. By assuming a parabolic geometry, the average ablated volume per shot was determined for each experiment. The results show that the amount of removed material for the LIBS and HV-LIBS experiments for samples at room temperature is similar, indicating that the high-voltage discharge does not remove additional material from the target and that the observed improvement in emission is mainly due to plasma re-excitation. Conversely, in the experiments where the sample is heated, with or without HV, the ablated volume increases by about six or seven times compared to those performed with the target at room temperature. The fact that more matter is removed implies that there are more emitters present in the plasma.

In the experiments performed with the sample at high temperature, the air density near the target is estimated to be 44% of its normal value. Assuming that the plasma behaves as an ideal gas, the plasma volume is expected to be about twice as large as the volume obtained with the sample at room temperature. Since the matter removed is at least six times larger, the electron density in the heated case would be about three times larger. However, this was not observed experimentally, suggesting that not all the removed matter was actually atomized. On the other hand, the electron density results when the sample is heated show no significant change (see Figure 5a). Thus, due to the enlarged plasma volume, it is expected to see an increase of a factor of two in the number of emitters. This corresponds with the experimental results when heating the samples, since the average intensity and detection sensitivity increase approximately two-fold. Therefore, the increase achieved by heating the sample is mainly due to an increase in the plasma volume and the number of emitters.

4. Conclusions

In this work, we investigated the simultaneous use of high voltage discharges in combination with sample heating to improve the sensitivity of LIBS to analyze soil samples. The analysis of detection limits was performed on eight elements considered as contaminants naturally present in certified loamy soil samples. After performing a time delay optimization process, an improvement of close to a factor of two was obtained by heating the samples, and this was greater than four when high voltage was combined with heating.

The results from electron density and plasma temperature measurements indicate that the application of high voltage enhances the signal intensity by re-exciting the laser-generated plasma. In addition, analysis of the produced crater profiles shows that the matter removed in the heating experiments increases considerably, which contributes to the improvement in emission sensitivity.

The proposed method has the potential to measure various elements at concentrations below the levels established by international standards. This, it makes it possible to increase the sensitivity of LIBS while maintaining its portability.