Open Access
Add to BibSonomy
Abstract
In this study, two-dimensional (2D) correlation analysis was utilized for achieving a significant improvement in the signal-to-noise (S/N) ratio of laser-induced breakdown spectroscopy (LIBS) data. Time-resolved LIBS spectra of metallic, bimetallic targets and the normal LIBS spectra of bimetallic targets with varying compositions were used for the detailed analysis. The diagonal of the matrix in the synchronous spectra was used to demonstrate the improvement in the signal-to-noise ratio (SNR). An improvement in the peak intensities by few orders of magnitude accompanied by suppression in the noise was observed. The correlations between LIBS peaks were also visualized using the 2-D plots. The correlation strengths of atomic transitions were visualized in aluminium (Al), copper (Cu), and brass whereas correlation strengths of atomic and ionic transitions were visualized in Au-Ag bimetallic targets with different compositions (Au30Ag70, Au50Ag50, Au80Ag20). The improved spectra were subsequently used in the principal component analysis for classification studies of four compositions of bimetallic targets (Au20Ag80, Au30Ag70, Au50Ag50, and Au80Ag20). The variance of the first three principal components was found to be improved from the analysis. The accumulated percentage of explained variance of ∼95 was achieved with the first three components from improved spectra whereas only ∼80 was achieved with the regular LIBS spectra from PCA studies. Furthermore, using this correlation analysis we demonstrate a significant improvement in the SNR of CN and C2 peaks in the femtosecond LIBS spectra of two polymer samples obtained from a standoff distance of 6 m, which may prove substantial for improved classifications studies.
© 2021 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
1. Introduction
Laser induced breakdown spectroscopy (LIBS) analysis employs the spectral emissions from the plasma formed when intense laser pulses are focused on to the sample [1]. Typical LIBS spectra consist of ionic, atomic, and molecular transitions [2]. LIBS technique has the ability to identify individual elements in any material. Furthermore, the information in the varying intensities with different compositions of elements coupled with machine learning techniques can be used for classification and identification of samples. Because of its versatile nature, being quick and minimal sample preparation requirements, LIBS found fantastic applications in diverse fields such as identification and classification of bacteria [3–5], in planetary, mineral exploration [6], analyzing the archelogy samples [7,8], explosive detection [9–11], trace element detection [12,13], in the study of historic paintings [14], and in the study of fundamental plasma properties [15]. This technique found diverse applications because of its capability to combine with Raman spectroscopy [16,17], to be able to operate easily in double-pulse and/or stand-off mode [18–21]. The advantage of signal enhancement using nanoparticles [22], and adopting the new trends in machine learning for the LIBS data analysis made it a powerful tool [23].
Two-dimensional correlation spectroscopy (2DCOS) was initially developed by Noda et al. [24]. It was primarily used to analyze the data from nuclear magnetic resonance (NMR), near-infrared (NIR), and Raman spectroscopy [24–27]. The resolution of the spectrum was shown to be improved with the 2D spectroscopy [28]. The simultaneous variations or coupling between the two spectral emissions with a small perturbation in the system can be studied. Better interpretation capabilities can be achieved when compared to regular spectra from the shapes and strengths in the 2D correlation plot [29,30]. In the 2D correlation plots the strength of the points on diagonal of the two-dimensional plot represents the self-correlation (or the auto correlation) of the peaks. It gives the information on the time correlation of the particular peak intensity with itself. The strength of the off-diagonal points represents the time correlation of the peak intensity with other peaks. Similarly, the 2D analysis on the spectra with varying concentration gives information on the auto correlation and cross correlation with composition. Apart from visualization this analysis can be used in improving the resolution and S/N ratio of the spectrum. We have demonstrated the 2D correlation analysis with 2D contour fill plots on time-resolved and composition varying LIBS spectra. Our analysis confirms that at least 3-4 orders of magnitude improvement in the signal-to-noise (SNR) of the LIBS data can be achieved. It is pertinent to note that the observed improvements were (a) not limited to few peaks or (b) was not random and it was observed on the complete LIBS spectrum. Here we have used it to visualize the correlations in the atomic transitions of Aluminium, Copper, Brass, and a bimetallic target of Au-Ag. We have also demonstrated that improved spectra can further improve the classification capabilities (here we used the PCA analysis). This can find impending applications in stand-off LIBS and other techniques where the SNR is generally very poor. This has significant ramifications in the standoff detection of energetic molecules wherein the signals are too weak for identification and classification studies [11]. Though this technique has been well established in the improvement/analysis of NMR and IR data we have demonstrated this for the first time, to the best of our knowledge, using nanosecond LIBS data. We have also discussed few prospective applications in the conclusions section. The LIBS spectra of organic materials and few metals contain molecular emissions (AlO, CN, C2) along with atomic and ionic emissions [31–33]. These molecular emissions can be used to understand the structure-property correlation using LIBS data [34,35]. For example, in the above-mentioned works the detonation parameters of the explosive materials such as oxygen balance, velocity of detonation, and detonation pressure were correlated with the LIBS spectra of different functional and structural isomers. We believe that the 2D correlation analysis could be extended further to the molecular emissions in LIBS.
2. Experimental procedures
Figure 1(a) depicts the experimental schematic of the nanosecond LIBS setup. A nanosecond Nd:YAG laser of 7 ns duration at a wavelength of 532 nm and with an average pulse energy of 42 mJ was used for the breakdown of the samples. The LIBS signal collected was fed to an ICCD and spectrometer combination. Since the plasma emissions from metals as well as bimetals were strong the gate delay and gate width were kept long enough and were optimized for metals and bimetals separately. For the collection of the time-resolved LIBS spectra we have used a 0.5 µs of initial gate delay and 0.5 µs of gate width for metal samples and 1 µs of initial gate delay and 1 µs of gate width for bimetal samples. In the case of polyvinyl chloride (PVC) and hydroxy propyl cellulose (HPC) an initial gate delay of 120 ns was used. The acquisition parameters are tabulated in the Table 1. Each spectrum was an accumulation (average) of 10 acquisitions taken from a fresh spot each time. The LIBS spectra of Au-Ag bimetallic targets were acquired for four different compositions of 20-80, 30-70, 50-50, and 80-20, respectively. The alloy samples were prepared by purchasing the Ag and Au locally (>99.9% purity) and mixing them carefully. A 600 µm optical fiber was used to couple the collected light to the spectrometer and ICCD (Mechelle ME5000 Echelle spectrograph with a wide wavelength range (200-850) nm, ANDOR iStar, which was triggered by the delay generator output. Here 20 time-resolved spectra were collected with the above-mentioned gate delays and gate widths and were averaged at each time delay from the time resolved spectra and the final mean spectra were used for the 2D correlation studies as shown in the Fig. 1(b). The averaging was done to minimize the matrix effect, if any. The nonlinear variations in the intensity of the peaks with increasing the compositions are the primary motivation for this work to study the correlation between the transitions in the LIBS spectra. The coupling between the Au and Ag peaks for different compositions are illustrated at different ranges. The 45°-collection geometry was used since it was convenient for alignment. No additional optimization was performed since the LIBS signal was strong.
Fig. 1. (a) Schematic of the nanosecond LIBS experimental setup used for metals, alloys, and bimetallic targets (b) the steps followed in data acquisition and data preparation for the analysis.
Table 1. Summary of the acquisition parameters used in the ns LIBS experiments.
3. 2D correlation plots
A set of ‘m’ LIBS spectra of the system under systematic disturbance inducing change in the spectral intensities are represented by$\; A({{\lambda_j},{t_i}} )$. Where the discrete variable ${\lambda _j}$ for $j = 1,2,3..n\; \; \; $related to the single spectrum represents the wavelength sampled over ‘n’ values. The second variable ${t_i}\; $for $i = 1,2,3\ldots m$ represents the effect of applied perturbation on the system sampled over m values.
The dynamic spectrum$\; \tilde{A}({{\lambda_j},{t_i}} )\; $in the time interval$\; {t_1}\; $to$\; {t_m}\; $is defined in the below equation as
where$\; \bar{A}({{\lambda_j}} )\; \; $is the reference spectra used in the calculations of the dynamic spectra. The reference spectra can be either first spectrum or last spectrum of the time resolved spectra and it can also be set to zero as well [29].The average spectra can be calculated as shown in the Eq. (2). The reference spectrum is set to zero in our studies.
The synchronous $\Phi ({{\lambda_1},{\lambda_2}} )$ and asynchronous $\Psi ({{\lambda_1},{\lambda_2}} )$ correlation spectra were calculated from
where $\widetilde {{y_i}}$ is the spectral intensity at a point of variable ${t_i}$s
where ${N_{ik}}$ Hilbert-Noda transformation matrix
We have used the matrix multiplications method for the calculations and better methods like 2D projection methods were well reported [30,36] . The synchronous and asynchronous spectra as represented by the symbols Ф, Ψ with A as the spectral data matrix, where each row is the LIBS spectrum taken at particular time, as shown in Eq. (6) and (7).
N is the Hilbert Noda transformation matrix form the Eqs. 5.
In techniques studied earlier (other than the LIBS technique), the spectra were measured with the perturbation on the system of interest such as magnetic field, electrical field, thermal, chemical, optical, mechanical, or the system varying with time were used for the analysis. The analysis divulges the similarities and dissimilarities in the spectra. Correlation analysis was applied on the time varying metal, bimetallic targets, and composition varying bimetallic targets LIBS spectra. The correlations with systematic variations would be easy to interpret after the two-dimensional (2D) spectroscopy. The resulting similarity spectra is called synchronous and dissimilarity spectra as asynchronous 2D spectra, which are complementary to each other. The asynchronous spectra in this case of study were not considered here. Figure 2 illustrates the time-resolved LIBS spectra of Aluminium [Fig. 2(a)], Copper, [Fig. 2(b)], and Brass [Fig. 2(c)] at five different times with regular interval of 1 µs and the same data was used for the 2D correlation studies of Aluminium [Fig. 2(d)], Copper, [Fig. 2(e)], and Brass [Fig. 2(f)], respectively.
Fig. 2. The time-resolved LIBS spectra of (a) Aluminium (b) Copper (c) Brass at five different gate delays with regular interval of 1 µs. These were used for the 2D correlation studies for (d) Aluminium (e) Copper (f) Brass targets, respectively. “arb. u” stands for “arbitrary units”
4. Results and discussion
Four different reference spectra were tried for the analysis as shown in Fig. 3(a) with the first spectrum (from the time series spectra). Figure 3(b) depicts the last spectrum (from the time series spectra) whereas Fig. 3(c) is the reference taken as zero (no spectrum) and Fig. 3(d) is the average spectrum (obtained from the time series spectra) used as reference. The variation in the amplitude and the width of the peaks were observed for different reference spectra. The peaks were broad and high when the first spectra in Fig. 3(a) was taken as reference and the width, height were reduced for the cases of last spectrum as reference [Fig. 3(b)] and the zero as reference spectrum [Fig. 3(c)]. The intensities were too low when the reference spectrum used was an average of all the spectra [Fig. 3(d)]. The amplitude of peaks in the synchronous correlation plot is a function of both the LIBS signal intensity of the correlated peaks and the strength of correlation from the changes between them. The contributions from the noise, which is random, is constant on to the diagonal of 2D synchronous plots because of lack of autocorrelation. Hence, 2D synchronous plots can be used to recover weak signal from a noisy background. As shown in Fig. 4, the linear spectra along the diagonal of matrix in the 2D analysis were considered to demonstrate the improvement in the signal. Figure 4(a) depicts the regular LIBS spectra of Aluminium in the spectral range of 303-312 nm and the corresponding 2D correlation spectra is shown in Fig. 4(b) while Fig. 4(c) illustrates the plot of diagonal of the 2D correlation analysis of the time-resolved LIBS spectra.
Fig. 3. 2D correlation analysis performed with four different reference spectra (a) with the first collected spectrum of the time-resolved spectra (b) last collected spectrum (c) with reference taken as zero and (d) average spectrum of the time-resolved spectra as reference. “arb. u” stands for “arbitrary units”
Fig. 4. The regular LIBS spectra of (a) Aluminium target in the spectral range of 303-312 nm and the corresponding (b) 2D correlation spectra and (c) diagonal of the 2D correlation analysis of the LIBS spectra is plotted with reference taken as zero. “arb. u” stands for “arbitrary units”
4.1 Correlation studies
The 2D correlation spectra of the LIBS have wavelength as both X-axis and Y-axis. The peaks on the diagonal represent the autocorrelation of the peak called auto-peaks and the off-diagonal peaks correspond to the correlation with the other peaks at the particular wavelength and are called cross-peaks. Python was used for the analysis and in the calculation of the synchronous and asynchronous spectra from the dynamic spectra with the average spectra taken as zero. The noise in the 2D spectra was suppressed as the randomly varying noise does not have correlations with varying time but the signal with correlation is enhanced thus improving the signal to noise ratio in the 2D plots. Figure 2(d) shows that the atomic peaks at 308.21 nm and 309.28 nm of Aluminium are correlated positively and the peak are tabulated in the Table 2. In Fig. 2(e) the Copper peaks at 324.78 nm and 329.05 nm exhibited a good correlation and the identified peaks are tabulated Table 3. The intensity of the cross-peaks shows how strongly the peaks are correlated. In Fig. 2(f) for Brass shows that the atomic peak of Copper at 324.75 nm is correlated to the atomic peak of Zinc at 334.5 nm this correlation was found to be stronger than the self-correlation of the Zinc peak. The LIBS peaks for Brass are summarized in Table 4. The correlation spectra of the Brass and Copper were similar in the range of 220-250 nm but were highly distinguishable in the range 320-340 as expected. The auto peak 327.39 nm was found to be weaker than the peak at 324.7nm and the cross peak suggesting a coupling between the transition at 327.39 nm and 324.7 nm.
Table 2. Identified LIBS peaks from the Aluminium target
Table 3. Identified LIBS peaks from the Copper target.
Table 4. Identified LIBS peaks from the Brass target.
The LIBS signal intensity of the atomic peaks was improved by 5 orders of magnitude in the case of Aluminium and the data is shown in Fig. 5 [5(a) representing Aluminium, 5(b) representing Copper, 5(c) representing Brass] while 4 orders of magnitude improvement was observed for bimetallic targets as shown in Fig. 5 data [(d) illustrating Au30Ag70, (e) illustrating Au50Ag50, and (f) illustrating Au80Ag20 targets data, respectively]. It is worth mentioning that the observed improvements in the SNR is for these particular sets of data and the experimental conditions mentioned earlier. We believe that similar improvements can be obtained for other LIBS data as well. Specifically, this analysis will be an excellent choice where the SNR is very poor. For example, standoff LIBS spectra of explosive molecules and that too in the single shot mode. However, further detailed studies are essential to confirm this hypothesis. Recently, Quaroni et al. reported the signal improved by 3 orders of magnitude for time resolved infrared spectra [37] . It was observed that the correlated peaks were found to be improved and the non-correlated noise was suppressed. Furthermore, the widths of the peaks were reduced suggesting an improvement in resolution of the spectrum. This clearly demonstrates that multiple spectra under various conditions can be used to improve the quality of the spectra. Further studies are also necessary for exploring the application of this method to overcome the matrix effects in LIBS. The observed off-diagonal peaks were brighter than the diagonal peaks suggesting good correlations. The 2D correlations analysis from the time resolved LIBS spectra of bimetal targets shown in Fig. 6(a) for Au30Ag70, Fig. 6(b) for Au50Ag50, and Fig. 6(c) for Au80Ag20. Five different gate delays with regular interval of 1 µs was used for the 2D correlation studies shown in the Fig. 6(d) for Au30Ag70, Fig. 6(e) for Au50Ag50, and Fig. 6(f) for Au80Ag20 targets, respectively. The target Au50Ag50 exhibited superior correlations than the other two. It is also observed that the first ionized peak of Ag at 241.32 nm does not show much variation with the change in the Ag percentage but the intensity of the off-diagonal peaks varies. The cross peak between the first ionized peak of Ag at 241.32 nm and the atomic peak of Au at 242.7 nm was stronger.
Fig. 5. The diagonal of the 2D correlation analysis on LIBS spectra is plotted. The signal to noise ratio is improved enormously for all the samples (a) Aluminium (b) Copper (c) Brass (d) Au30Ag70, (e) Au50Ag50 (f) Au80Ag20 targets, respectively. “arb. u” stands for “arbitrary units”
Fig. 6. Time-resolved LIBS spectra of (a) Au30Ag70, (b) Au50Ag50, (c) Au80Ag20 bimetallic targets at five different gate delays with regular interval of 1microsecond is used for the 2D correlation studies on the time-resolved spectra of (d) Au30Ag70, (e) Au50Ag50, (f) Au80Ag20 targets, respectively.
The varying composition of Ag/Au in Ag-Au alloy target was also used as a perturbing parameter and the obtained results are depicted in Fig. 7. The 2D analysis for three different compositions was compared between data obtained for 1 µs and 2 µs gate delays. The correlations for three different spectral regions (a) 240-250 nm, (b) 260-280 nm, and (c) 520-550 for the first µs and (d) 240-250 nm, (e) 260-280 nm, (f) 520-550 nm for the second µs after the pulse was incident, respectively, were compared. The correlations improved in the 240-250 nm range but they diminished in the other two ranges 260-280 nm and 520-550 nm. LIBS spectra of Au-Ag bimetal targets at three different compositions were used for the 2D correlation studies at two different gate delays, in the range 240-250 nm [Fig. 7(a)], 260-280 nm [Fig. 7(b)], 520-550 nm [Fig. 7(c)] at 1 µs and 240-250 nm [Fig. 7(d)], 260-280 nm [Fig. 7(e)], 520-550 nm [Fig. 7(f)], at 2 µs. The correlations between different compositions grew as the time increased in the range 240-250 nm and diminished in other two ranges. The identified LIBS peaks of both the Au and Ag are summarized in the Table 5.
Fig. 7. LIBS spectra of Au-Ag bimetal at three different compositions were used for the 2D correlation studies at two different gate delays in the range (a) 240-250 nm (b) 260-280 nm (c) 520-550 nm at 1 µs and (d) 240-250 nm (e) 260-280 nm (f) 520-550 nm at 2 µs. “arb. u” stands for “arbitrary units”
Table 5. Identified LIBS peaks from the Au-Ag bimetallic target.
4.2 Classification studies
Principal component analysis on the diagonal of the matrix in the 2D synchronous spectra and the regular LIBS spectra were performed and the results are depicted in Fig. 8. PCA maximizes the variance of the principal components [38] and thus dimensional reduction is achieved and with fewer dimensional data we could be able to explain the system. The larger the variance, the greater is the amount of information the principal component contains. The cumulative explained variance provides the information of how many principal components should be included to describe the data accurately. The PCA score plot is shown in Fig. 8(a) and the variance plot is depicted in Fig. 8(b). The cumulative explained plot is illustrated in 8(c) for the regular LIBS spectra. The score plot [Fig. 8(d)], variance plot [Fig. 8(e)], and the cumulative variance plot [Fig. 8(f)] of the improved spectra were also compared. Here in this case removing the noise and improving the peaks intensity resulted in the improvement of the explained variance of each first three principal components and we could achieve 95% variance only with the first 3 components whereas in the case of regular spectra with noise it took more than 5 components to get 80%, which is much lower compared to earlier number.
Fig. 8. PCA on the regular spectra of bimetal targets and the improved diagonal of the 2D correlation analysis (a) PCA score plot (b) variance plot (c) cumulative explained variance on regular spectra and (d) PCA score plot (e) variance plot (f) cumulative explained variance on the signal to noise improved spectra. Solid line is a guide to the eye.
The SNR of the spectra improved enormously in the case of the intense peaks in regular spectra and the highly correlated peaks with them. The correlation between the atomic peaks in Aluminium (303–310 nm range), Copper (320-340 nm range) and Brass (320-340 nm range) were visualized in Fig. 2. The coupling between the Zinc and Copper lines was strong in the case of Brass in the 320-340 nm range. The distinguishable coupling for different compositions of Au and Ag were visualized and the cross-peaks were observed to be strong for the Au50- Ag50 sample when compared to the remaining compositions in the range 240-250 nm. The Au-Ag correlations peaks were plotted from the time-resolved LIBS spectra for varied compositions and the observed cross-peaks were found to be strong in the 260-280 nm range, as depicted in the Fig. 6. These 2D plots could also help distinguish the samples of different compositions by filtering out the features which do not change. Even with the diagonal from the 2DCOS data we can improve the classification. Additional studies with different perturbation methods will be an enormously useful tool for a better understanding of the coupling between different transitions and also to improve the classification. This could be an interesting tool also to explore the molecular emissions in nanosecond and femtosecond laser-induced plasma [39–41] and in exploring the data for machine learning applications (such as PCA, artificial neural networks, convolution neural networks).
4.3 2D correlation analysis of femtosecond stand-off LIBS data
The 2D correlation analysis was also applied on the spatially resolved femtosecond stand-off LIBS spectra of two polymer samples [polyvinyl chloride (PVC) and hydroxy propyl cellulose (HPC)]. The femtosecond (∼50 fs) pulses at 800 nm, 1 kHz repetition rate and 1.8 W (average power) were focused on to the target at a distance of 6 meters. To further strengthen our argument on 2D correlation analysis it is demonstrated here that the analysis is very useful in improving the signal-to-noise ratio of the stand-off LIBS data. Due to the pandemic new fs LIBS data could not be obtained.
Figures 9 and 10 clearly illustrate the improvement in the SNR (by few orders of magnitude) for PVC and HPC fs LIBS data reported in our previous work [18]. The CN (Δν = 0) and C2 (Δν = 0) peaks were considered for the analysis. The average of five spectra were taken for comparing with the 2D correlation analysis on five spatially resolved spectra. In femtosecond LIBS data the molecular peaks are very important for discrimination and, therefore, we had focused on these two peaks only for further analysis. It is advisable to select a proper wavelength range when working with the LIBS data on traces. This is significant since in the standoff geometry (more so when one wishes to collect single shot spectra) the signals are very poor (hence poor SNR) and classifications studies will not provide satisfactory results. Any improvement in the SNR will be significant for classification studies, especially in the case of explosives detection. Additionally, one could use advanced techniques such as convolution neural networks on the whole image form 2D correlation analysis which could further improve the classification capability and accuracy. However, this requires additional computational resources since the size of each image data is huge.
Fig. 9. The average of five fs stand-off LIBS spectra of PVC (a)-(c) were compared with the improved spectra. (d)-(f) after 2D correlation analysis the CN and C2 peaks in both cases were compared.
Fig. 10. The average of five fs stand-off LIBS spectra of HPC (a)-(c) were compared with the improved spectra. (d)-(f) after 2D correlation analysis the CN and C2 peaks in both cases were compared.
5. Conclusions
Primarily the SNR is enhanced in 2D spectroscopy as the randomly varying noise does not hold any correlations with time. The linear spectroscopic data cannot provide information regarding the correlations in the transitions. It was observed from the detailed analyses that while the intensity of noise was diminished the intensity of signal was significantly improved. Additionally, 2DCOS appears to be a powerful tool in resolving the complex spectra as well. 2D correlation analysis simplifies the interpretation of correlations between the transitions in LIBS spectra and is advantageous in visualising the time-dependent decay of different LIBS peaks and wavelength shift, if any. We summarize here the important results obtained from this work and suggestions for future works:
- • These studies will provide a better understanding of the molecular peaks and their coupling, especially in the femtosecond LIBS spectra. For example, the molecular emissions such as CN, C2, AlO, TiO in the ultrashort-pulse LIBS spectra could add more insights to the LIBS analysis.
- • 2DCOS studies combined with time-resolved LIBS spectra and/or other perturbation methods would result in better resolution with the same spectrometer, due to the spreading of data over a second dimension, improved resolution, and SNR from which overlapped peaks can be distinguished, which are not possible otherwise.
- • Similar studies can be performed with other parameters such as changing ICCD gain, gate delays, input laser energy, the distance at which the emissions are collected, compositions, etc. to get a better understanding of the transitions.
- • In the case of the stand-off LIBS the signal to noise ratio deteriorates as compared to nearfield case and this method can be used for improving it, especially in the case of explosives detection [11,42]. This will also be very useful in the case of single pulse LIBS wherein the SNR is expected to be poor.
Funding
Defence Research and Development Organisation [#ERIP/ER/1501138/M/01/319/D (R&D)].
Acknowledgements
Authors acknowledge DRDO, India for financial support [#ERIP/ER/1501138/M/01/319/D (R&D)] and Director, ACRHEM for his support and encouragement. Authors also thank Dr. Chandu Byram for his help in the experiments.
Disclosures
The authors declare no conflicts of interest.
Data availability
Data underlying the results presented in this paper are not publicly available at this time but may be obtained from the authors upon reasonable request.
References
1. S. N. Thakur and J. P. Singh, “Fundamentals of laser induced breakdown spectroscopy,” in Laser-Induced Breakdown Spectroscopy (Elsevier, 2007), pp. 3–21.
2. S. S. Harilal, J. Yeak, B. E. Brumfield, J. D. Suter, and M. C. Phillips, “Dynamics of molecular emission features from nanosecond, femtosecond laser and filament ablation plasmas,” J. Anal. At. Spectrom. 31(6), 1192–1197 (2016). [CrossRef]
3. D. Prochazka, M. Mazura, O. Samek, K. Rebrošová, P. Pořízka, J. Klus, P. Prochazková, J. Novotný, K. Novotný, and J. Kaiser, “Combination of laser-induced breakdown spectroscopy and Raman spectroscopy for multivariate classification of bacteria,” Spectrochim. Acta, Part B 139, 6–12 (2018). [CrossRef]
4. S. J. Rehse, “A review of the use of laser-induced breakdown spectroscopy for bacterial classification, quantification, and identification,” Spectrochim. Acta, Part B 154, 50–69 (2019). [CrossRef]
5. Q. Wang, G. Teng, X. Qiao, Y. Zhao, J. Kong, L. Dong, and X. Cui, “Importance evaluation of spectral lines in Laser-induced breakdown spectroscopy for classification of pathogenic bacteria,” Biomed. Opt. Express 9(11), 5837 (2018). [CrossRef]
6. R. S. Harmon, J. Remus, N. J. McMillan, C. McManus, L. Collins, J. L. Gottfried, F. C. DeLucia, and A. W. Miziolek, “LIBS analysis of geomaterials: geochemical fingerprinting for the rapid analysis and discrimination of minerals,” Appl. Geochem. 24(6), 1125–1141 (2009). [CrossRef]
7. V. Lazic, F. Colao, R. Fantoni, and V. Spizzicchino, “Recognition of archeological materials underwater by laser induced breakdown spectroscopy,” Spectrochim. Acta, Part B 60(7-8), 1014–1024 (2005). [CrossRef]
8. F. Ruan, T. Zhang, and H. Li, “Laser-induced breakdown spectroscopy in archeological science: a review of its application and future perspectives,” Appl. Spectrosc. Rev. 54(7), 573–601 (2019). [CrossRef]
9. A. Kumar Myakalwar, N. Spegazzini, C. Zhang, S. Kumar Anubham, R. R. Dasari, I. Barman, and M. Kumar Gundawar, “Less is more: Avoiding the LIBS dimensionality curse through judicious feature selection for explosive detection,” Sci. Rep. 5(1), 13169 (2015). [CrossRef]
10. F. C. De Lucia, J. L. Gottfried, and A. W. Miziolek, “Evaluation of femtosecond laser-induced breakdown spectroscopy for explosive residue detection,” Opt. Express 17(2), 419 (2009). [CrossRef]
11. A. K. Shaik and V. R. Soma, “Standoff discrimination and trace detection of explosive molecules using femtosecond filament induced breakdown spectroscopy combined with silver nanoparticles,” OSA Continuum 2(3), 554 (2019). [CrossRef]
12. R. González, P. Lucena, L. M. Tobaria, and J. J. Laserna, “Standoff LIBS detection of explosive residues behind a barrier,” J. Anal. At. Spectrom. 24(8), 1123 (2009). [CrossRef]
13. J. Moros, F. J. Fortes, J. M. Vadillo, and J. J. Laserna, “LIBS detection of explosives in traces,” in Springer Series in Optical Sciences (Springer, 2014), pp. 349–376.
14. L. Burgio, K. Melessanaki, M. Doulgeridis, R. J. H. Clark, and D. Anglos, “Pigment identification in paintings employing laser induced breakdown spectroscopy and Raman microscopy,” Spectrochim. Acta, Part B 56(6), 905–913 (2001). [CrossRef]
15. D. W. Hahn and N. Omenetto, “Laser-induced breakdown spectroscopy (LIBS), part I: review of basic diagnostics and plasma–particle interactions: still-challenging issues within the analytical plasma community,” Appl. Spectrosc. 64(12), 335A–336A (2010). [CrossRef]
16. S. K. Sharma, A. K. Misra, P. G. Lucey, and R. C. F. Lentz, “A combined remote Raman and LIBS instrument for characterizing minerals with 532 nm laser excitation,” Spectrochim. Acta, Part A 73(3), 468–476 (2009). [CrossRef]
17. A. Giakoumaki, I. Osticioli, and D. Anglos, “Spectroscopic analysis using a hybrid LIBS-Raman system,” Appl. Phys. A 83(4), 537–541 (2006). [CrossRef]
18. N. L. Murthy, S. Abdul Salam, and S. V. Rao, “Stand-off femtosecond laser induced breakdown spectroscopy of metals, soil, plastics and classification studies,” 2019 Workshop on Recent Advances in Photonics (WRAP), 2019/12 2019. DOI:10.1109/WRAP47485.2019.9013674
19. J. K. Antony, N. J. Vasa, V. S. Raja, and A. Laxmiprasad, “Single laser based dual-wavelength ablation technique for emission enhancement during LIBS,” J. Phys. D: Appl. Phys. 45(36), 365401 (2012). [CrossRef]
20. V. S. Kumar, N. J. Vasa, and R. Sarathi, “Remote surface pollutant measurement by adopting a variable stand-off distance based laser induced spectroscopy technique,” J. Phys. D: Appl. Phys. 48(43), 435504 (2015). [CrossRef]
21. S. Abdul Kalam, S. V. Balaji Manasa Rao, M. Jayananda, and S. Venugopal Rao, “Standoff femtosecond filament-induced breakdown spectroscopy for classification of geological materials,” J. Anal. At. Spectrom. 35(12), 3007–3020 (2020). [CrossRef]
22. A. De Giacomo, M. Dell’Aglio, R. Gaudiuso, C. Koral, and G. Valenza, “Perspective on the use of nanoparticles to improve LIBS analytical performance: nanoparticle enhanced laser induced breakdown spectroscopy (NELIBS),” J. Anal. At. Spectrom. 31(8), 1566–1573 (2016). [CrossRef]
23. H. Rajavelu, N. J. Vasa, and S. Seshadri, “Laser-Induced Breakdown Spectroscopy Combined with Temporal Plasma Analysis of C2 Molecular Emission for Carbon Analysis in Coal,” Appl. Spectrosc. 75(7), 893–900 (2021). [CrossRef]
24. I. Noda, A. E. Dowrey, and C. Marcott, “Recent developments in two-dimensional infrared (2D IR) correlation spectroscopy,” Appl. Spectrosc. 47(9), 1317–1323 (1993). [CrossRef]
25. I. Noda, “Generalized two-dimensional correlation method applicable to infrared, Raman, and other types of spectroscopy,” Appl. Spectrosc. 47(9), 1329–1336 (1993). [CrossRef]
26. I. Noda, A. E. Dowrey, and C. Marcott, “Two-dimensional infrared (2d IR) spectroscopy,” in Modern Polymer Sectroscopy (Wiley-VCH Verlag GmbH, 2008), pp. 1–32.
27. I. Noda, Y. Liu, and Y. Ozaki, “Two-dimensional correlation spectroscopy study of temperature-dependent spectral variations of N-Methylacetamide in the pure liquid state. 2. two-dimensional Raman and Infrared− Raman heterospectral analysis,” J. Phys. Chem. 100(21), 8674–8680 (1996). [CrossRef]
28. L. Ma, V. Sikirzhytski, Z. Hong, I. K. Lednev, and S. A. Asher, “Insight into resolution enhancement in generalized two-dimensional correlation spectroscopy,” Appl. Spectrosc. 67(3), 283–290 (2013). [CrossRef]
29. M. A. Czarnecki, “Interpretation of two-dimensional correlation spectra: science or art?” Appl. Spectrosc. 52(12), 1583–1590 (1998). [CrossRef]
30. Y. Park, I. Noda, and Y. M. Jung, “Two-dimensional correlation spectroscopy in polymer study,” Front. Chem. 3, 14 (2015). [CrossRef]
31. C. G. Parigger, “Atomic and molecular emissions in laser-induced breakdown spectroscopy,” Spectrochim. Acta, Part B 79-80, 4–16 (2013). [CrossRef]
32. C. G. Parigger and J. O. Hornkohl, “Computation of AlO B2Σ+→ X2Σ+ emission spectra,” Spectrochim. Acta, Part A 81(1), 404–411 (2011). [CrossRef]
33. S. J. Mousavi, M. Hemati Farsani, S. M. R. Darbani, A. Mousaviazar, M. Soltanolkotabi, and A. Eslami Majd, “CN and C2 vibrational spectra analysis in molecular LIBS of organic materials,” Appl. Phys. B 122(5), 106 (2016). [CrossRef]
34. S. A. Kalam, N. L. Murthy, P. Mathi, N. Kommu, A. K. Singh, and S. V. Rao, “Correlation of molecular, atomic emissions with detonation parameters in femtosecond and nanosecond LIBS plasma of high energy materials,” J. Anal. At. Spectrom. 32(8), 1535–1546 (2017). [CrossRef]
35. E. N. Rao, P. Mathi, S. A. Kalam, S. Sreedhar, A. K. Singh, B. Jagatap, and S. V. Rao, “Femtosecond and nanosecond LIBS studies of nitroimidazoles: correlation between molecular structure and LIBS data,” J. Anal. At. Spectrom. 31(3), 737–750 (2016). [CrossRef]
36. I. Noda, “Projection two-dimensional correlation analysis,” J. Mol. Struct. 974(1-3), 116–126 (2010). [CrossRef]
37. L. Quaroni and E. Normand, “Two-dimensional correlation spectroscopy analysis for the recovery of weak bands from time-resolved infrared spectra of single cells,” AIP Conf. Proc. 1214, 66–68 (2010). [CrossRef]
38. P. Pořízka, J. Klus, E. Képeš, D. Prochazka, D. W. Hahn, and J. Kaiser, “On the utilization of principal component analysis in laser-induced breakdown spectroscopy data analysis, a review,” Spectrochim. Acta, Part B 148, 65–82 (2018). [CrossRef]
39. E. N. Rao, S. Sunku, and S. V. Rao, “Femtosecond laser-induced breakdown spectroscopy studies of nitropyrazoles: the effect of varying nitro groups,” Appl. Spectrosc. 69(11), 1342–1354 (2015). [CrossRef]
40. E. J. Kautz, P. J. Skrodzki, M. Burger, B. E. Bernacki, I. Jovanovic, M. C. Phillips, and S. S. Harilal, “Time-resolved imaging of atoms and molecules in laser-produced uranium plasmas,” J. Anal. At. Spectrom. 34(11), 2236–2243 (2019). [CrossRef]
41. H. Rajavelu, N. J. Vasa, and S. Seshadri, “Effect of ambiance on the coal characterization using laser-induced breakdown spectroscopy (LIBS),” Appl. Phys. A 126(6), 395 (2020). [CrossRef]
42. A. K. Shaik, N. R. Epuru, H. Syed, C. Byram, and V. R. Soma, “Femtosecond laser induced breakdown spectroscopy based standoff detection of explosives and discrimination using principal component analysis,” Opt. Express 26(7), 8069 (2018). [CrossRef]
