List of Publications

Publications  

  Page 200  

List of Publications

Thesis related

1. “Optimized LIBS setup with echelle spectrograph for multi elemental

analysis”, V. K. Unnikrishnan, Kamlesh Alti, Rajesh Nayak, Rodney

Bernard, Niyati Khetarpal, V. B. Kartha, C. Santhosh, G. P. Gupta and B. M.

Suri, Journal of Instrumentation, 5, p4005, 2010.

2. “Measurements of plasma temperature and electron density in laser-induced

copper plasma by time-resolved spectroscopy of neutral atom and ion

emissions”, V K Unnikrishnan, Kamlesh Alti, V B Kartha, C Santhosh, G

P Gupta and B M Suri, Pramana Journal of Physics, Vol.74, No 6, pp 983-

993, 2010.

3. “Spectroscopy of Laser-Produced Plasmas: Setting up of High Performance

Laser-Induced Breakdown Spectroscopy (LIBS) System”, V. K.

Unnikrishnan, Kamlesh Alti, Rajesh Nayak, Rodney Bernard, V. B. Kartha,

C. Santhosh1, G. P. Gupta and B. M. Suri, Pramana Journal of Physics, Vol.

75, pp. 1145-1150, 2010.

4. “Quantitative elemental analysis of nickel alloys using calibration-based

laser-induced breakdown spectroscopy”, G. P. Gupta, B. M. Suri, A. Verma,

M. Sundararaman, V. K. Unnikrishnan, K. Alti, V. B. Kartha and C.

Santhosh, Journal of Alloys and Compounds, 509, 3740–3745, 2011.

5. “Trace Element Analysis using Laser Induced Breakdown Spectroscopy

(LIBS) Technique”, V.K. Unnikrishnan, K. Mridul, Rajesh Nayak, Kamlesh

Alti, V.B. Kartha, C. Santhosh, G.P. Gupta and B.M. Suri, Proceedings of the

55th DAE Solid State Physics Symposium 2010, AIP Conf. Proc. 1349, 475-

476, 2011

6. “Analysis of Trace Elements in Complex Matrices by Laser Induced

Breakdown Spectroscopy (LIBS)”, V K Unnikrishnan, K Mridul, Rajesh

Nayak, Kiran Aithal, Kamlesh Alti, V B Kartha, C Santhosh, G P Gupta and B

M Suri (Under Review)

Publications  

  Page 201  

7. “Calibration-free laser-induced breakdown spectroscopy for quantitative

elemental analysis of materials”, V K Unnikrishnan, K Mridul, R Nayak, K

Alti, V B Kartha, C Santhosh, G P Gupta and B M Suri (Accepted in Pramana

Journal of Physics)

8. “Quantitative Analysis of Manganese Concentration in Manganese-Doped

Glasses by Laser-Induced Breakdown Spectroscopy Using a Nanosecond

Ultraviolet Nd: YAG Laser”, V. K. Unnikrishnan,  Rajesh Nayak, V. B.

Kartha, C. Santhosh, M. S. Sonavane, R. G. Yeotikar, G. P. Gupta and B. M.

Suri (Under Review) 

Other publications

1. “Evaluation of high-performance liquid chromatography laser-induced

fluorescence for serum protein profiling for early diagnosis of oral cancer”,

Ajeetkumar Patil, Vijendra Prabhu, K.S. Choudhari, V.K. Unnikrishnan,

Sajan D. George, Ravikiran Ongole, Keerthilatha M. Pai, Jayarama K. Shetty,

Sujatha Bhat, Vasudevan Bhaskaran Kartha, Santhosh Chidangil, Journal of

Biomedical Optics, 15(6), 067007, 2010.

2. “Parameter optimization of a Laser-Induced Fluorescence (LIF) system for

in-vivo screening of oral cancer”, V.K. Unnikrishnan , Rajesh Nayak,

Rodney Bernard, K. Jeena Priya, Ajeetkumar Patil, J. Ebenezer, Keerthilatha

M. Pai, Sajan D. George, V.B. Kartha and C. Santhosh, Journal of Laser

Applications, Vol 23, pp. 1-7, 2011

3. “Surface Enhanced Fluorescence of Tryptophan by Silver-Nano-particles”,

Ajeetkumar Patil, Unnikrishnan.V.K. and Santhosh Chidangil, Proceedings

of the 55th DAE Solid State Physics Symposium 2010, AIP Conf. Proc. 1349,

216-217, 2011

4. “Laser Induced Fluorescence Spectroscopy of Soft Tissues of the Oral

Cavity”, Ajeetkumar Patil, Unnikrishnan.V.K., Rodney Bernard,

Keerthilatha M.Pai, Ravikiran Ongole, V.B.Kartha and Santhosh Chidangil,

Proceedings of the 55th DAE Solid State Physics Symposium 2010, AIP Conf.

Proc. 1349, 224-225, 2011

Publications  

  Page 202  

5. “Early Detection of Oral Pre-malignancy and Malignancy by Salivary Protein

Profile analysis: A non-invasive technique for screening of oral cancer”,

K.S.Choudhari,Ajeetkumar Patil,Vijendra Prabhu, Nandita Shenoy, Ravikiran

Ongole, Keerthilatha M.Pai, V.K. Unnikrishnan, V.B.Kartha and C.

Santhosh (Under Review)

6. “Highly Sensitive High Performance Liquid Chromatography- Laser Induced

Fluorescence (HPLC-LIF) for Proteomics Applications”, Ajeetkumar Patil,

Choudhari K.S, Vijendra Prabhu, Unnikrishnan V.K., Sujatha, Keerthilatha

M Pai, Kartha V.B. and Santhosh C (Under Review)

Conference Proceedings

International

1. “Setting up of a Sensitive Laser-Induced Breakdown Spectroscopy (LIBS)

System with Echelle Spectrograph - A Method for Surface Analysis of

Materials”, V. K. Unnikrishnan, Kamlesh Alti, Rajesh Nayak, Rodney

Bernard, V. B. Kartha, C. Santhosh, G. P. Gupta and B. M. Suri, International

Conference on Recent Trends in Material and Characterization RETMAC,

NITK, Surathkal, 2010.

2. “Laser-Induced Breakdown Spectroscopy using echelle spectrograph-ICCD

system: An effective tool for ultra trace elemental analysis”, V. K.

Unnikrishnan, K. Mridul, R. Nayak, K. Alti, V. B. Kartha, C. Santhosh, G. P.

Gupta and B. M. Suri, Meghnad Saha Memorial International Symposium-

cum-workshop on “Laser Induced Breakdown Spectroscopy”, MMISLIBS,

Allahabad University, Allahabad, 2010.

3. “Quantitative Analysis of Brass by Calibration-Free Laser-Induced

Breakdown”, G. P. Gupta, B. M. Suri, V. K. Unnikrishnan, K. Alti, V. B.

Kartha and C. Santhosh, Meghnad Saha Memorial International Symposium-

cum-workshop on “Laser Induced Breakdown Spectroscopy”, MMISLIBS,

Allahabad University, Allahabad, 2010.

Publications  

  Page 203  

4. “Quantitative Analysis of Chromium Concentration in Nickel-Based-Alloys by

Laser-Induced Breakdown Spectroscopy at Atmospheric Pressure Using a

Nanosecond Ultraviolet Nd:YAG Lase”r, G. P. Gupta, B. M. Suri, A. Verma,

M. Sundararaman, V. K. Unnikrishnan, K. Alti, V. B. Kartha and C.

Santhosh, International congress on Analytical Science – iCAS, University of

Science and Technology(CUSAT), Kochi, 2010.

5. “Biomedical Applications of Laser-Induced Breakdown Spectroscopy (LIBS):

a preliminary study”, V. K. Unnikrishnan, Rajesh Nayak, Praveen

Devangad, M. K. Dinoop, V. B. Kartha, B. M. Suri and C. Santhosh,

International Conference on Biomedical Engineering, Manipal, 2011.

National

1. “Development of a Laser-Induced Breakdown Spectroscopy (LIBS)

system for remote sensing applications”, V.K. Unnikrishnan, Ajeetkumar

Patil, Kamlesh Alti, V. B. Kartha and C. Santhosh, G.P. Gupta and B.M.

Suri, Eighth DAE-BRNS National Laser Symposium (NLS-08), Laser

Science and Technology Centre (LASTEC), Delhi, 2008.

2. “Spectroscopy of Laser-Produced Plasmas: Setting up of High Performance

Laser-Induced Breakdown Spectroscopy (LIBS) System”, V. K.

Unnikrishnan, Kamlesh Alti, Rajesh Nayak, Rodney Bernard, V. B. Kartha,

C. Santhosh1, G. P. Gupta and B. M. Suri, Ninth DAE-BRNS National Laser

Symposium (NLS-09), BARC, Mumbai, 2009.

3. “Characterization of laser-induced copper plasma by time resolved

spectroscopy of neutral atom and ion emissions”, G. P. Gupta, B. M. Suri, V.

K. Unnikrishnan, Ajeetkumar Patil, Kamlesh Alti, V. B. Kartha, and C.

Santhosh, 24th National Symposium on Plasma Science & Technology, NIT

Hamirpur , 2009.

4. “Trace Element Analysis using Laser Induced Breakdown Spectroscopy

(LIBS) Technique”, V.K. Unnikrishnan, K. Mridul, Rajesh Nayak, Kamlesh

Alti, V.B. Kartha, C. Santhosh, G.P. Gupta and B.M. Suri, DAE Solid State

Physics Symposium (SSPS-55), Manipal University, Manipal, 2010.

Publications  

  Page 204  

5. “Effect of magnetic field on Laser-Induced plasmas using an optimized LIBS

system with echelle spectrograph”, V.K. Unnikrishnan, K. Mridul, Rajesh

Nayak, Kamlesh Alti, V.B. Kartha, C. Santhosh, B.M. Suri and G.P. Gupta,

DAE-BRNS National Laser Symposium (NLS-19), RRCAT, Indore, 2010.

6. “Laser-Induced Breakdown Spectroscopy with echelle spectrograph: A pilot

study for remote analysis”, K. Mridul, V.K. Unnikrishnan, Rajesh Nayak,

Kamlesh Alti, V.B. Kartha, C. Santhosh and B.M. Suri, DAE-BRNS National

Laser Symposium (NLS-19), RRCAT, Indore, 2010.

7. “Quantitative Analysis of Manganese Concentration in Manganese-Doped

Glasses by Laser-Induced Breakdown Spectroscopy Using a Nanosecond

Ultraviolet Nd:YAG Laser”, V. K. Unnikrishnan, Rajesh Nayak, V. B.

Kartha, C. Santhosh, M. S. Sonavane, R. G. Yeotikar, G. P. Gupta and B. M.

Suri, DAE-BRNS National Laser Symposium (NLS-20), Anna University,

2011.

8. “Time Evolution and Dynamics of Laser-Produced Copper Plasma in LIBS”,

V. K. Unnikrishnan, Rajesh Nayak, M. K. Dinoop, V.B. Kartha, C.

Santhosh, B.M. Suri and G.P. Gupta, DAE-BRNS National Laser Symposium

(NLS-20), Anna University, 2011.

2010 JINST 5 P04005

PUBLISHED BY IOP PUBLISHING FOR SISSA

RECEIVED: January 11, 2010REVISED: February 19, 2010

ACCEPTED: February 27, 2010PUBLISHED: April 15, 2010

Optimized LIBS setup with echellespectrograph-ICCD system for multi-elementalanalysis

V.K. Unnikrishnan,a K. Alti,a R. Nayak,a R. Bernard,a N. Khetarpal,b V.B. Kartha,a

C. Santhosh,a,1 G.P. Guptac and B.M. Suric

aCentre for Atomic and Molecular Physics, Manipal University,Manipal, India

bDepartment of Biotechnology, Manipal University,Manipal, India

cLaser and Plasma Technology Division, Bhabha Atomic Research Centre,Mumbai, India

E-mail: santhosh.cls@manipal.edu

ABSTRACT: Laser-Induced Breakdown Spectroscopy (LIBS) is well recognized as a promisingtool for in situ/remote elemental analysis of environmental, archeological, clinical, and hazardoussamples. With the aim of quantifying trace elements in such samples, using LIBS technique, anechelle spectrograph-ICCD system with high sensitivity and good resolution has been assembled.Various important parameters of this system were studied and optimized. Conditions for gettinggood quality LIBS spectra and signal for multielemental analysis have been achieved, and these arediscussed and illustrated in this paper.

KEYWORDS: Plasma generation (laser-produced, RF, x ray-produced); Pulsed power; Plasma di-agnostics - interferometry, spectroscopy and imaging

c© 2010 IOP Publishing Ltd and SISSA doi:10.1088/1748-0221/5/04/P04005

2010 JINST 5 P04005

Contents

1 Introduction 1

2 Experimental details 2

3 Results and discussion 3

4 Conclusion 13

1 Introduction

In recent years, Laser Induced Breakdown Spectroscopy (LIBS), also known as Laser InducedPlasma Spectroscopy (LIPS), has been shown to be a versatile elemental analysis tool attractingincreased attention because of the broad range of applications. LIBS can be conveniently usedin the laboratory as well as for in situ/remote analysis of environmental samples (soil, water, airetc), clinical samples (tissue and body fluids), hazardous materials, and planetary surfaces [1]–[11]. With the development of industrial activities, the living systems as well as environment arebeing contaminated by numerous pollutants, with their increasing levels causing concern. Thereis an enormous increase in the output of pollutants from various industries (thermal power plants,semiconductor, leather, textile, soft drinks, fertilizers, oil and gas industries), which is a cause forgreat concern. Considerable attention is now being paid to find the relationship between variousforms of cancer and other diseases with trace metal pollutants in the human body as well as in theenvironment [12–15]. Many trace elements are nutrients for all living systems and play importantroles in many biochemical processes. However, these elements must be kept below certain levels;otherwise they become toxic to human/living beings [12, 13]. Thus both their deficiency and excesspresence are harmful for the proper functioning of cellular processes. The versatility of LIBStechnique for multi-element analysis, and its applicability to different sample types (solid, liquidand gas), makes it attractive in detecting and quantifying pollutants in water, soil, clinical samples,food products etc [8, 16–19]. Especially relevant in this respect is the trace element analysis ofsamples of living systems, environments in the neighbourhood of industrial centers, power plants,material storage locations and high density population areas. Earlier studies have already shownthe relation between metal contents in serum and liver cancer [12]. Similarly a link has been shownbetween zinc and copper in serum with lung cancers [13]. LIBS technique can be used to determineall these trace elements in soil, water, plants and clinical samples [14, 15].

The present work aims at exploring the potential of LIBS combined with advanced instru-mentation, on the simultaneous detection and quantification of several elements in environmen-tal/clinical samples. In general, samples like soil, plants, tissue etc. contain various metallic ele-ments such as Al, Ca, Zn, Cu, Fe, Mg, Na and K in large amounts, which often give rise to extensivebackground spectra, interfering with analytical lines of trace elements present in the sample, when

– 1 –

2010 JINST 5 P04005

Figure 1. Schematic diagram of LIBS setup.

conventional spectrometers with low or moderate resolution are used. It may also happen thatrequired analytical lines, free from spectral interference, may fall over a wide range (200 nm to1000 nm).With the use of high dynamic range and high resolution echelle type spectrographs, it ispossible to cover the wide range 200–950 nm, and record simultaneously interference free spectrallines of each element of the sample.

A portable LIBS instrument has been reported previously for the detection of metal contami-nants on surfaces [20]. Efforts have been made to build a LIBS system for simultaneous detectionof several elements using multiple spectrographs and synchronous charged coupled detectors [21].Czerny-turner type spectrometers have been utilized in most of the LIBS applications for a longperiod [8, 14, 22–26]. Recently, combination of high resolution and high dynamic range echellespectrograph coupled with sensitive intensified charge coupled device (ICCD) has been used forLIBS experiments [7, 27–32]. The present paper discusses assembling and standardization of asensitive LIBS system employing an echelle spectrograph coupled with an ICCD for in situ/remotetrace element analysis of environmental and clinical samples.

2 Experimental details

Figure 1 shows a schematic diagram of the LIBS set-up assembled in our laboratory. It uses the3rd harmonic laser light (355 nm, ∼400 mJ, 10 Hz repetition rate and 6 ns pulse width) of anNd:YAG laser (Spectra Physics PRO 230-10) for plasma generation. The laser is focussed on tothe specimen surface with appropriate optics.

A 355 nm specific high power mirror is used to reflect only 355 nm laser light (transmitting532 nm and 1064 nm laser lines) from the laser head towards a Pellin-Broca prism. The smallamounts of 532 nm and 1064 nm remaining in the beam are dispersed from the Pellin-Broca prismand are blocked using a beam dump. A beam splitter, which transmits 80% of incident light andreflects remaining 20% at 45◦, has been used for the material ablation. Neutral density filters ofdifferent optical densities are used to adjust the laser energy of 20% beam incident on the sample.

– 2 –

2010 JINST 5 P04005

Figure 2. Echelle spectrum of Hg Ar lamp covering a broad range from 250–800 nm.

This is then focused on to the sample using a high-energy bi-convex lens of focal length 20 cm inorder to achieve appropriate breakdown threshold irradiance for different samples. When hazardousmaterials or remote samples are to be excited this arrangement can be replaced with a telescopicoptical set-up [10;11].

The plasma emission is collected by a light collection system which consists of UV gradequartz lenses/mirrors, and is imaged on to the spectrograph slit by using an optical fiber couplingsystem. A high resolution, cross dispersion echelle spectrograph-ICCD system (Andor MechelleME5000-DH734-18U-03PS150) records the spectrum. It is a 195mm focal length system (F/7) andcovers 200–975 nm spectral range in one setting with 0.05nm wavelength resolution. With 10µmslit, spectral dispersion (λ/∆λ ) will be very high with an echelle grating (52 lines/mm blazed at575 nm). A thermoelectric cooled ICCD which is sensitive in the whole UV-VIS-NIR region isused to collect the dispersed signal from spectrograph. The detector is gated in synchronizationwith the Nd-YAG laser pulse to get maximum signal to noise ratio.

A Hg-Ar lamp, which provides sharp lines from 200–1000 nm, has been used for wavelengthcalibration of the system. Intensity calibration of the echelle spectrograph-ICCD system was doneusing NIST certified Deuterium-Quartz-Tungsten- Halogen lamp [Ocean Optics, USA].

3 Results and discussion

Wavelength and intensity calibrated spectrum of Hg-Ar lamp are shown in figures 2 and 3 respec-tively. Figure 2 shows sharp-well resolved lines (the 577–579 nm doublet, for example) of mercuryand argon spread over a selected wavelength range of 250–800 nm, which clearly indicates theresolution and range of the system.

However, it can be seen from the spectrum that the 254 nm mercury line, which should be verystrong (in fact the strongest mercury line) appears only slightly more intense than the 436 nm line.This happens because, echelle gratings produce spectra in several orders, and the efficiency variesacross diffraction orders. If one wants quantitative intensity information, (which may be useful

– 3 –

2010 JINST 5 P04005

Figure 3. Intensity calibrated spectrum of Hg-Ar lamp.

and necessary for remote analysis), the spectrograph has to be calibrated with standard sources forintensity. This can be done with intensity calibration lamps. The result of such calibration is shownin figure 3.

As seen from figure 3 the intensity of the 253.7 nm mercury line (which is the strongest lineof all) is now very high compared to all other lines in the spectrum.

In many samples, several elements may be present in much larger amounts compared to theelement to be quantified, and spectral lines of these elements can overlap with usable, intenseanalytical lines of the elements of interest. Unique advantage of using the echelle spectrographcan be visualized by comparing its performance with a conventional czerny turner spectrograph(Spectra-pro 150). Figure 4(a) shows Mercury lines recorded using Spectra-pro 150 spectrograph.The 435.8 nm line of Mercury is shown in expanded scale in figure 4(b). FWHM of this peak is ameasure of the resolution of this system and it is 0.74 nm.

The Hg-Ar spectrum recorded with the echelle spectrograph is shown in figure 5(a) and (b).The FWHM of the 435.8 nm line is only 0.058 nm, better by a factor of 12 compared to the 150mmspectrograph. (This means that we have to have a ∼ 2meter czerny-turner spectrograph to achievethe same dispersion, but such a system will cover only an extremely small range in one setting witha 25mm CCD).

As already mentioned, this LIBS system is built for trace analysis in mainly two classes ofsamples; clinical samples (saliva, serum, and tissue) and environmental samples (soil, water andagricultural/food products). The samples for analysis will thus range from those which give verysimple spectra, easy to handle and analyze in the lab (saliva, water) to those which will haveextremely complex spectra, difficult to handle and may require remote operation (soil).

Each of the different sample types requires detailed investigations for developing quantitativeanalysis technique. It is therefore very essential to understand the advantages and disadvantages ofthe echelle- LIBS system compared to conventional emission spectrographic analysis systems aswell as the influence of various parameters on the emission lines in a matrix. We illustrate belowhow information on these aspects can be obtained with plasma from a copper plate.

– 4 –

2010 JINST 5 P04005

Figure 4. (a) Hg-Ar spectrum recorded using Spectra-Pro 150 (b) Resolution on an expanded scale as inset.

Figure 5. (a) Hg-Ar spectrum recorded using echelle spectrograph (b) Resolution on an expanded scale asinset.

For comparison of echelle and conventional systems, the plasma was studied by simultane-ously recording the LIBS spectra on both systems. A 200mm focal length lens was used to focusthe laser beam on to the target. Copper plate was mounted on a motorized X and Y translationstage in order to ensure that each laser pulse is falling on a fresh copper surface. The LIBS spectrausing echelle and diffraction (Spectra-Pro 150) spectrographs are shown in figure 6(a) and (b) forcomparison.

A comparison of figure 6(a) and (b) shows the relative merits and demerits of the two sys-

– 5 –

2010 JINST 5 P04005

Figure 6. LIBS spectra of copper recorded using (a) Echelle (Mechelle 5000) with inset showing the fullrange of echelle copper spectrum (200–950 nm) and (b) diffraction (Spectra Pro 150) spectrographs.

tems. The range covered, and the number of lines observed are much more in the echelle system,indicating the advantages of the high resolution (0.05 nm) and broad range (200–950 nm, 750 nmat a stretch) provided by the system. This also has the advantage that any overlapping lines ofinterfering elements will be separated out and analytical results will be more reliable. Secondlythe larger number of lines offers a wider choice of analytical lines. On the other hand, diffractionspectrographs can only cover 250 nm at a stretch (1/3rd of the range that echelle systems can cover)as shown in figure 6(b) with low spectral resolution.

On the other hand, the intensities of the lines are much lower (8000 Vs 900,000 for the 521 nmline) in the echelle system. This has happened because of two reasons; first, because of the higheraperture (F/4) of the Spectra-Pro compared to that of the echelle (F/7); and second, the very lowresolution of the Spectra-pro which results in the intensities of the weak lines around 521 nm gettingadded to its intensity. It should be noted that for complex samples this may create large interferencefrom major components in the low-resolution system.

The above result shows that to get best intensities we have to increase the radiation collectionefficiency of the echelle system. This can be done by using an F/1 collection lens and using f-matching optics to input this into the spectrograph and figure 7 shows the improvement with this.

In continuation of the above preliminary studies we have further optimized the echellespectrograph-ICCD system to achieve good signal with better signal-to-noise ratio. In quantitativetrace analysis, source stability and spectrum reproducibility are very crucial. Though variations inthese parameters can be compensated to a large extent by the use of internal standards, in manyapplications with the LIBS system (remote/hazardous material analysis) addition of internal stan-dards may not be possible. In order to check the source stability we have done short (30 minutes)and long (∼ 5 hours) term measurements using Nd-YAG laser. The root mean square value of thepulse energy variation was found to be 400±0.4 mJ and 400±1.0 mJ for short term and long term

– 6 –

2010 JINST 5 P04005

Figure 7. LIBS spectrum of copper with improved collection efficiency.

Figure 8. Reproducibility check of the LIBS system using Hg-Ar lamp.

measurements respectively. Similarly we have monitored the detector reproducibility by recordingseveral spectra using Hg-Ar lamp. The variation in relative intensity of the two Mercury lines i.e.546.07 nm and 435.8 nm is shown in figure 8. The relative standard deviation (RSD) of these valueswas found to be negligible (i.e. 0.11%).

We have also checked the reproducibility of the LIBS (4.46 x 108 W/cm2) spectra of copper

– 7 –

2010 JINST 5 P04005

Table 1. Spectrum reproducibility check for laser induced copper plasma.

Intensity ratio of Cu lines RSD (%)I521.820/I330.795 5I521.820/I465.112 19I521.820/I515.324 8I521.820/I529.252 12I521.820/I578.213 18

Figure 9. LIBS signal variation of copper plasma for different laser pulse intensities.

using relative intensities of 521.82 nm line with respect to other copper emission lines at delay700ns as shown in table I.

It is observed that the RSD of relative intensities varies as 5%, 19%, 8%, 12% and 18%for 521.820/330.795, 521.820/465.112, 521.820/515.324, 521.820/529.252 and 521.820/578.213respectively. Variations of these magnitudes are acceptable for trace analysis.

Influence of laser intensities (irradiance) on the plasma generation was studied using copperfoil as target at different incident laser energies. This is achieved by the use of neutral density filterskept before the focusing lens. Gate delay and gate width were kept constant at 750ns and the signalwas accumulated for 120 laser pulses. The results are shown in figure 9.

As expected, the signals keep increasing in direct proportion with increase in the laser irradi-ance. This is very crucial since it is essential that day-to-day or operator variations in laser powershould not lead to any errors in the results. In figure 9, the intensities of 521.82 nm and 515.32 nmlines is the average of three experiments under same conditions. An interesting result from theabove studies is the observation that the slope of the intensity variation with power is different for

– 8 –

2010 JINST 5 P04005

Figure 10. Spatial distribution of laser ablated copper plasma.

the two lines. This can give information on the dynamics of the plasma, since the levels corre-sponding to these two lines may be getting populated through different channels.

Usually, collecting radiation from the whole plasma may be better for increasing intensity andimproving sensitivity. However, there are situations in which it may be necessary to select suitableregions of the plasma. This is especially necessary in remote LIBS (for example planetary surfaces,industrial waste processing units), where we have no control over the atmosphere. Also, when weuse only single line intensity (see later) the region of plasma where this line will be most intense willhelp in recording a suitable spectrum. We have studied the spatial distribution of copper plasma forgetting this information. Here we have formed a 1:1 image of the copper plasma plume (∼4mm)using appropriate lens (5 cm focal length lens kept at 10 cm away from the copper plasma). Avariable position slit having width of 350µm was used to scan the plasma image. By moving thislens we could image the required plasma regions on to the slit. We have kept our signal collectorprobe 20 cm away from this slit to satisfy f-matching conditions. For same delay and width (i.e.1000 ns) we have recorded spectra (average of three trials) for different positions of the slit andcorrelated these positions with plasma regions as shown in figure 10.

Study of temporal evolution of plasma is of great importance as it helps to determine the timeslot when signal of our interest is maximum. Hence, we have recorded time- resolved LIBS copperspectra (figure 11) and studied their characteristics.

It can be seen from the figure 11 that the spectrum changes very much as plasma evolves tem-porally. Initial delays are dominated by continuum and ionic spectra, but as the plasma decays, thecontinuum intensity decreases and we could see good quality LIBS spectrum suitable for qualita-tive and quantitative elemental analysis of the sample. The narrowing of the lines on the time scaleis an indication of the rate of cooling of the plasma.

The results also show that LIBS, unlike D.C. arc or ICP, requires detailed studies to establishsuitable conditions for different types of samples. As mentioned earlier, our set- up is to be used

– 9 –

2010 JINST 5 P04005

Figure 11. Gated copper spectrum of LIBS plasma as a function of time after plasma initiation.

for relatively simple systems (water,saliva, serum, tissue) having little background, as well as forsamples which may give very denser spectra (soil, minerals) and hence high background, andsamples where remote operation is required (hazardous materials). We discuss below some ofour results for highly complex systems which illustrate the advantages and disadvantages of theLIBS-echelle system.

Usually for spectrochemical or other methods of analysis of trace elements in complex ma-trices like soil, minerals, and ores, the element of interest is pre-concentrated by dissolution andseparation by complex formation or ion exchange, to avoid the possible high interference from thelarge amounts of other elements in the matrix. It should be emphasized here that in the presentmethod this is not necessary because of the high resolution. Also, any small interference can beremoved by running a synthetic- simulated- blank of the matrix materials without the element ofinterest, and subtracting out the blank spectrum from the sample spectrum. This technique can beespecially useful for samples like soil, which can be sand, dolomite, granite or of volcanic originsince the simulated matrix can compensate to a large extent the variations in sample type, eliminat-ing any problems in the usability of the method for any type of soil. In the present study the soilused was of sandy loam (paddy soil) type.

For samples like soil and minerals, the LIBS technique has to be used in the laboratory orremote conditions. Figure 12 shows a comparison of the typical LIBS spectrum of a soil sample (1x 1012W/cm2, 300ns delay and 6000ns gate width) with that of a copper sample. Because of thepresence of many elements like alkali and alkaline earth metals, transition metals, silica, aluminaetc., the soil spectrum is extremely complex. The other main elements present in the sand loam soilused for this study apart from copper are zinc, magnesium and calcium. It is very difficult, if notimpossible, to isolate suitable spectral lines for quantitative analysis, of elements present at tracelevels with low resolution instruments. In conventional methods, this problem is usually solvedby extracting the element of interest and analyzing it in a suitable matrix [33]. Obviously this is

– 10 –

2010 JINST 5 P04005

Figure 12. LIBS spectra recorded from copper and soil samples.

not possible for in situ or remote analysis. We have developed the LIBS approach for direct traceanalysis of such samples and this is illustrated here for analysis of copper in soil.

By comparing LIBS spectrum from a pure copper sample (figure 12(a)) with that from a soil infigure 12(b), we see that it is extremely difficult to identify lines of copper for quantification in soil.The strong lines of copper atoms fall in the regions around 270, 325, 380, 425, 520, 580, 780 and820 nm. It is possible to computer-match the soil and copper spectra and identify interference-freeregions.

Figure 13 shows this for the strong Cu I lines where the regions of interest are plotted inexpanded scale for copper and soil. Combining the observed half width (0.06 nm) and the averagerun-to-run error (0.03 nm), any line within 0.18 nm can cause interference. It is seen that the lines521.82 nm, 515.324 nm and 510.554 nm are relatively free from interference from lines of soilspectra. The high resolution and wide range of the echelle system thus enables one to chooseinterference-free lines for analysis even with highly complex samples.

Figure 14 shows a spectrum of 400 ppm copper in soil. Sections around copper lines are shownin the expanded scale. If we look at the lines discussed above, it can be seen that even at 400 ppmlevel quantitative analysis is not easy in this sample, presumably because of the slight overlap oflines with weaker lines of other elements at chosen delay.

It thus looks like that trace analysis below these levels in soil samples will be difficult with theLIBS technique compared to D.C. arc/ICP, because the total number of available atoms of a givenemitter is distributed in a much larger number of levels. Two methods are available to overcomethis problem. One is to increase exposure time, which will be practically difficult beyond a certainrange. The second method is to make use of the fact that the total trace element excited in theplasma is represented by the total lines emitted by that element. That is, make use of as muchemission (as many atomic and ionic lines) as possible.

– 11 –

2010 JINST 5 P04005

Figure 13. Comparison of LIBS copper and soil spectra at region of our interest.

Figure 14. LIBS spectrum of 400ppm copper in soil. Inset shows zoomed 521.82 nm line of copper.

To carry out the above method, pre-processing of the data is needed. Pre-processing involvesthe following steps: Background correction (to remove detector background/plasma continuum),normalization and calibration. Background correction is achieved by fitting a polynomial to back-ground levels at regular intervals, generating the background curve, and subtraction of this curvefrom the observed spectrum. Normalization is done for more accurate evaluation of difference be-tween a sample (here soil with 400 ppm copper) and a matrix (blank soil). For normalization asuitable line of a matrix is chosen in the region of interest and the sample spectrum is multiplied by

– 12 –

2010 JINST 5 P04005

Figure 15. LIBS Soil sample analysis — figure (a) spectrum obtained subtracting soil blank from soil with400ppm copper (521.82 nm line of copper is shown inset), (b) spectrum of soil with 400ppm copper and (c)blank soil spectrum.

the ratio of the intensity of this line to intensity of the same line in the sample, making intensitiesof all matrix lines more or less equal in both spectra. Before calculating the difference betweensample and blank, we have to ensure that errors from small changes in calibration are reduced asmuch as possible. For this, sharp lines of the matrix distributed across the spectrum are chosen inboth sample and blank. A polynomial function of pixel number against wavelength is generatedusing these lines. All the lines in both spectra are then recalibrated using this polynomial, so thatlines generated from elements in the blank superpose in both spectra. Subtraction of the blankfrom the sample spectrum will now contain only lines from the element being analyzed. The wholetechnique is illustrated in figure 15.

Figure 15(c) shows the spectrum of soil blank and figure 15(b) is the spectrum of a soil samplewith 400 ppm copper given earlier, after background correction, normalization, and recalibration.Figure 15(a) shows the “difference spectrum”. It is seen that by this method we can reduce con-siderably the interference from matrix spectrum and get high sensitivity for trace element analysis.This is possible in the echelle-ICCD system because it can cover a very wide range enabling theemission from several lines to be observed and data processing of the ICCD output can be done toremove matrix interference. Hence the “difference spectrum” discussed above, in principle, willcontain only spectral lines of elements which are not present in the blank.

4 Conclusion

As seen from above results and discussion, the special features that make LIBS an attractive toolare: (a) rapid elemental analysis (one measurement per laser pulse), (b) minimum or no priorsample preparation, (c) detection of many elements (high and low atomic number) simultaneously,(d) low detection limits (ppm), (e) point detection capability, enabling spatial distribution in a

– 13 –

2010 JINST 5 P04005

sample at micron level separation, as well as surface species (f) possibility of in-situ investigationof environmental samples, and (g) remote operation enabling probing of hostile or not so easilyaccessible surfaces.

In view of the above, a sensitive, high resolution and broad range laser-induced breakdownspectroscopy (LIBS) system having echelle type spectrometer coupled with an ICCD has been as-sembled and tested. Using this system copper plasma was studied at several laser irradiances anddelays. Broad spectral coverage and high resolution of the echelle system give great advantage tothe LIBS technique as spectral lines of elements can be easily selected in any part of the spectrum.Extensive studies done on copper using echelle system shows that this system can be successfullyused to perform LIBS for ultra-trace elemental analysis in the laboratory, and in situ/remote condi-tions. It can also be inferred that using “difference spectrum” method we can reduce considerablythe interference from matrix and get high sensitivity for trace element analysis of complex samples.

Acknowledgments

This work was carried out under the project titled “Trace Element Analysis for Environmentaland Biomedical Applications – Development of Laser Induced Breakdown Spectroscopy (LIBS)Technique”. Project No. 2007/34/14-BRNS/87, Board of Research in Nuclear Sciences (BRNS),Department of Atomic Energy (DAE), Govt. of India.

References

[1] P. Yaroshchyk, R.J.S. Morrison, D. Body and B.L. Chadwick, Quantitative determination of wearmetals in engine oils using laser-induced breakdown spectroscopy: A comparison between liquid jetsand static liquids, Spectrochimica Acta. Part. B 60 (2005) 986.

[2] A.I. Whitehourse, Laser-induced breakdown spectroscopy and its application to the remotecharacterisation of hazardous materials, Spectroscopy Asia 2 (2006) 13.

[3] M. Corsi et al., Double Pulse Calibration-Free Laser-Induced Breakdown Spectroscopy: a newtechnique for in situ standard-less analysis of polluted soils, Appl. Geochem. 21 (2006) 748.

[4] V. Hohreiter and D.W. Hahn, Dual-pulse laser induced breakdown spectroscopy: time-resolvedtransmission and spectral measurements, Spectrochimica Acta. Part. B 60 (2005) 968.

[5] M.A. Gondal, T. Hussai, Z.H. Yamani and M.A. Baig, Detection of heavy metals in Arabian crude oilresidue using laser induced breakdown spectroscopy, Talanta 69 (2006) 1072.

[6] S. Acquaviva, E.D. Anna, M.L. De Girogi and F. Moro, Laser-induced breakdown spectroscopy forcompositional analysis of multielemental thin films, Spectrochimica Acta. Part. B 61 (2006) 810.

[7] K. Song, Y. III Lee and J. Sneddon, Recent Developments in Instrumentation for Laser InducedBreakdown Spectroscopy, Appl. Spectroscopy Rev. 37 (2002) 89.

[8] R.S. Harmon, F.C. DeLucia, C.E. McManus, N.J. McMillan, T.F. Jenkins and M.E.Walsh,Laser-induced breakdown spectroscopy. An emerging chemical sensor technology for real-timefield-portable, geochemical, mineralogical, and environmental applications, Appl. Geochem. 21(2006) 730.

[9] A. Giakoumaki, K. Melessanaki and D. Anglos, Laser-induced breakdown spectroscopy (LIBS) inarchaeological science. Applications and prospects, Anal. Bioanal. Chem. 387 (2007) 749.

– 14 –

2010 JINST 5 P04005

[10] S.M. Clegg, R.C. Wiens, S.K. Sharma, P. Lucey, A. Misra and J. Barefield, LIBS–Ramanspectroscopy of minerals using remote surface modification techniques, Lunar and Planetary ScienceConf. XXXVII (2006).

[11] S.K. Sharma, A.K. Misra, P.G. Lucey, R.C. Wiens and S.M. Clegg, Combined remote LIBS andRaman spectroscopy of minerals coated with hematite and covered with basaltic dust At 8.6 m,GEORAMAN (2006).

[12] O. Miatto, M. Casaril, G.B. Gabriell, N. Nicoli, G. Bellisola and R. Corrocher, Diagnostic andprognostic value of serum copper and plasma fibrinogen in hepatic carcinoma, Cancer 55 (1985)784.

[13] M. Diez, F.J. Cerdan, M. Arroyo and J.L. Balibrea, Use of the copper/zinc ratio in the diagnosis oflung cancer, Cancer 63 (1989) 726.

[14] F. Capitelli, F. Colao, M.R. Provenzano, R. Fantoni, G. Brunetti and N. Senesi, Determination ofheavy metals in soils by Laser Induced Breakdown Spectroscopy, Geoderma 106 (2001) 45.

[15] M. Hanafi, M.M. Omar and Y.E.E-D. Gamal, Study of laser-induced breakdown spectroscopy ofgases, Radiat. Phys. Chem. 57 (1999) 11.

[16] A. Kumar, F.Y. Yueh, J.P. Singh and S. Burgess, Characterization of Malignant Tissue Cells byLaser-Induced Breakdown Spectroscopy, Appl. Optics 43 (2004) 5399.

[17] R.S. Harmon, F.C. DeLucia, A.W. Miziolek, K.L. McNebsby, R.A. Walters and P.D. French,Laser-induced breakdown spectroscopy (LIBS) an emerging field-portable sensor technology forreal-time, in-situ geochemical and environmental analysis, Geochem.-Explor. Env. A. 5 (2005) 21.

[18] S. Bouudjemai, T. Grasmi, R. Boushaki, R. Kasbadji and F. Medajahed, Laser induced breakdownspectroscopy in water, J. Appl. Sci. Env. Mgt 8 (2004) 13.

[19] M.H. Ebinzer et al., Extending the Applicability of Laser-Induced Breakdown Spectroscopy for TotalSoil Carbon Measurement, Soil Sci. Soc. Am. J. 27 (2003) 1616.

[20] K.Y. Yamamoto, D.A. Cremers, M.J. Ferris and L.E. Foster, Detection of metals in the environmentusing a portable laser-induced breakdown spectroscopy instrument, Appl. Spectrosc. 50 (1996) 222.

[21] D. Body and B.L. Chadwick, Simultaneous elemental analysis system using laser induced breakdownspectroscopy, Rev. Sci. Instrum. 72 (2000) 1625.

[22] S. Morel, N. Leone, P. Adam and J.Amouroux, Detection of Bacteria by Time-ResolvedLaser-Induced Breakdown Spectroscopy, Appl. Optics 42 (2003) 6184.

[23] P.L. Garcia, J.M. Vadillo and J.J. Laserna, Real-Time Monitoring of High-Temperature Corrosion inStainless Steels by Open-Path Laser-Induced Plasma Spectrometry, Appl. Spectrosc. 58 (2004) 1347.

[24] N.M. Shaikh, B. Rashid, S. Hafeez, Y. Jamil and M.A. Baig, Measurement of electron density andtemperature of a laser-induced zinc plasma, J. Phys. D 39 (2006) 1384.

[25] M.A. Hafez, M.A. Khedr, F.F. Elaksher and Y.E. Gamal, Characteristics of Cu plasma produced by alaser interaction with a solid target, Plasma Sources Sci. Technol. 12 (2003) 185.

[26] R.D. Harris, D.A. Cremers, M.H. Ebinger and B.K. Bluhm, Determination of nitrogen in Sand usinglaser-induced breakdown spectroscopy, Appl. Spectrosc. 58 (2004) 770.

[27] V. Detalle, R. Heon, M. Sabsabi and L. St-Onge, An evaluation of a commercial Echelle spectrometerwith intensified charge-coupled device detector for materials analysis by laser-induced plasmaspectroscopy, Spectrochim. Acta Part B 56 (2001) 1011.

– 15 –

2010 JINST 5 P04005

[28] A. Uhl, K. Loebe and L. Kreuchwig, Fast analysis of wood preservers using laser induced breakdownspectroscopy, Spectrochim. Acta Part B 56 (2001) 795.

[29] S. Hamilton, R. Al-Wazzan, A. Hanvey, A. Varagnat and S. Devlin, Fully integrated wide wavelengthrange LIBS system with high UV efficiency and resolution, J. Anal. Atom. Spectrom. 19 (2004) 479.

[30] J. Yun, R. Klenze and J. Kim, Laser-Induced Breakdown Spectroscopy for the On-Line MultielementAnalysis of Highly Radioactive Glass Melt Simulants. Part II: Analyses of Molten Glass Samples,Appl. Spectrosc. 56 (2002) 852.

[31] B. Salle, D.A. Cremers, S. Maurice and R.C. Wiens, Laser-induced breakdown spectroscopy forspace exploration applications: Influence of the ambient pressure on the calibration curves preparedfrom soil and clay samples, Spectrochim. Acta Part B 60 (2005) 479.

[32] M. Tripathi, Echelle Spectrographs: a flexible tool for spectroscopy: Raman and LIBS Spectroscopy,Andor (2006).

[33] V.B. Kartha, T.S. Krishnan, N.D. Patel and S. Gopal, BARC 284, Bhabha Atomic Research CenterReport (1967).

– 16 –

PRAMANA c© Indian Academy of Sciences Vol. 75, No. 6— journal of December 2010

physics pp. 1145–1150

Spectroscopy of laser-produced plasmas: Setting upof high-performance laser-induced breakdownspectroscopy system

V K UNNIKRISHNAN1, KAMLESH ALTI1, RAJESH NAYAK1,RODNEY BERNARD1, V B KARTHA1, C SANTHOSH1,∗,G P GUPTA2 and B M SURI21Centre for Atomic and Molecular Physics, Manipal University, Manipal 576 104, India2Laser and Plasma Technology Division, Bhabha Atomic Research Centre,Mumbai 400 085, India∗Corresponding author. E-mail: santhosh.cls@manipal.edu

Abstract. It is a well-known fact that laser-induced breakdown spectroscopy (LIBS)has emerged as one of the best analytical techniques for multi-elemental compositionalanalysis of samples. We report assembling and optimization of LIBS set up using highresolution and broad-range echelle spectrograph coupled to an intensified charge coupleddevice (ICCD) to detect and quantify trace elements in environmental and clinical sam-ples. Effects of variations of experimental parameters on spectroscopy signals of copperand brass are reported. Preliminary results of some plasma diagnostic calculations usingrecorded time-resolved optical emission signals are also reported for brass samples.

Keywords. Spectroscopy; laser-induced plasma; plasma temperature.

PACS Nos 42.62.Fi; 52.50.Jm; 52.38.Mf; 52.70.-m

1. Introduction

Laser-induced breakdown spectroscopy (LIBS), also known as laser-induced plasmaspectroscopy (LIPS), is basically an emission spectroscopy technique which uses in-tense, short pulses of laser radiation to ablate the sample surface [1–3]. Ablationof sample results in plasma generation. Spectral lines of atoms and ions of thisradiant plasma are used to develop quantitative and qualitative analytical informa-tion about the sample. Recent applications of LIBS technique for multi-elementalanalysis include environmental samples, biological samples, radioactive waste mate-rials etc. [4–6]. The versatility of LIBS technique for multi-element analysis and itsapplicability to different types of samples (solid, liquid and gas) make it attractivein detecting and quantifying hazardous pollutants using in-situ remote excitation.

In this paper, we report the assembling and standardization of an LIBS systemusing high-resolution broadband echelle spectrograph coupled with a sensitive ICCD

1145

V K Unnikrishnan et al

Figure 1. LIBS experimental set-up for multi-elemental analysis.

to detect and quantify trace elements in environmental, clinical and radioactivewaste samples. Various important experimental conditions of this system werestudied and optimized to increase the signal strength and detection efficiency of thisLIBS set-up for plasma spectroscopy of copper and brass samples. Time-resolvedplasma temperature for the brass plasma is also studied for finding out the localthermodynamic equilibrium (LTE) conditions required for LIBS elemental analysis.

2. Experimental methods

The schematic diagram of the LIBS set-up used for this study is shown infigure 1. The third harmonic of Nd:YAG laser (Spectra Physics PRO 230-10) witha pulse duration of 6 ns, pulse repetition frequency of 10 Hz and pulse energy of400 mJ was used for ablation of samples to form the plasma. This laser was focussedonto the sample using a bi-convex lens of focal length 20 cm to achieve appropri-ate breakdown threshold irradiance for different samples. A collecting/collimatinglens/mirror system was used for collecting the emission light from the generatedplasma for the best performance of the broadband echelle spectrograph (AndorMechelle ME5000-DH734-18U-03PS150) of the LIBS system. It was optimized toensure that all the wavelengths in the range 200–975 nm were collected using fibre-optic cable to the entrance slit of the spectrograph. The detector was kept inproper synchronization with the laser using delay generator to get time-resolvedinformation of plasma evolution.

3. Results and discussion

Wavelength and intensity calibration of echelle spectrograph–ICCD system wasdone using NIST certified lamps (mercury–argon, deuterium–quartz–tungsten–halogen) as shown in figure 2.

Figure 2b shows that after the intensity calibration, the intensity of 253.7 nmmercury line (which is the strongest line of all) is highest compared to all otherlines in the spectrum.

A comparison of the spectral resolution of the present LIBS set-up with Czerny-turner-based system was done by recording the mercury (Hg) spectrum and the

1146 Pramana – J. Phys., Vol. 75, No. 6, December 2010

Spectroscopy of laser-produced plasmas

Figure 2. (a) Wavelength and (b) intensity calibration of the LIBS systemusing NIST lamps.

Figure 3. Spectral resolution of (a) echelle spectrograph and (b) Czerny-turner spectrograph.

results are shown in figure 3. The estimated full-width at half-maximum (FWHM)of the 546.1 nm Hg line in the spectra recorded by both systems is found to be 0.1and 1 nm respectively. It is evident from figure 3 that the spectral resolution ofechelle spectrograph is 10 times higher than that of Czerny-turner spectrograph.LIBS spectrum of a complex sample containing several elements can have overlap-ping spectral lines of different elements. The echelle system has very high resolution

Pramana – J. Phys., Vol. 75, No. 6, December 2010 1147

V K Unnikrishnan et al

Figure 4. Variation in LIBS signal with (a) gate width, (b) number of laserpulses and (c) gain.

as shown in figure 3a and so it can very well separate out individual lines from dif-ferent elements.

We have also optimized different parameters of the LIBS system like laser irra-diance, collector probe distance and angle from the sample, detector parameters,etc. to achieve good signal with better signal-to-noise ratio. Some of the resultsare discussed here.

Detector parameters, namely, gate width, gain and accumulation (number oflaser pulses) are optimized using copper plasma signals generated by irradiatingthe sample with a laser intensity of 4.46 × 108 W/cm2 at a gate delay of 1000 ns.Figure 4 describes how the LIBS signal changes with these parameters. Each pointrepresents an average of three measurements. Well-defined dependence of LIBSsignal on gate width, number of laser pulses and gain is clearly seen from the figure.Hence, we have set an optimum possible value of these parameters to achieve good-quality LIBS signal.

After optimizing the experimental conditions, we have carried out time-resolvedLIBS study of copper, zinc and brass samples to locate suitable time window ofinterest for a particular sample under investigation. Temporal history of copper,zinc and brass samples also helps us to monitor the plasma characteristics. Figure5 shows LIBS spectra recorded from these samples at a gate delay of 1000 ns anda gate width of 1000 ns using a laser intensity of 4.88× 108 W/cm2.

1148 Pramana – J. Phys., Vol. 75, No. 6, December 2010

Spectroscopy of laser-produced plasmas

Figure 5. LIBS spectra of (a) copper, (b) zinc and (c) brass samples.

Table 1. Brass plasma temperature measurements.

Gate delay Plasma temperature(ns) (eV)

300 0.788500 0.843700 0.806

1000 0.7752000 0.774

From figure 5c, one can see that the LIBS spectrum of brass contains all the majorlines of copper and zinc. Hence we have recorded brass plasma at different delaysand calculated the plasma temperature using the Boltzmann plot. Five Cu I linesfrom brass plasma (465.112 nm, 510.554 nm, 515.324 nm, 521.82 nm and 578.213nm) were used for these calculations at 300 ns, 500 ns, 700 ns, 1000 ns and 2000ns delays. The estimated plasma temperatures are given in table 1. It is observedthat after 500 ns the plasma cools down exponentially. Quantitative measurementof this information is also possible by identifying suitable characteristic lines andemploying proper calibration methods.

Pramana – J. Phys., Vol. 75, No. 6, December 2010 1149

V K Unnikrishnan et al

4. Conclusions

A high-performance laser-induced breakdown spectroscopy system has been setup and optimized for studying spectroscopy of laser-produced plasmas of differentmaterials. Broad spectral coverage with high resolution of echelle system providesgreat advantage to carry out useful, precise, qualitative and quantitative elementalanalysis.We have also done time-resolved LIBS study of copper, zinc and brasssamples and estimated the brass plasma temperature using this present set-up atdifferent gate delays.

Acknowledgement

The authors are thankful to BRNS, DAE, Government of India for the financial(Project No. 2007/34/14-BRNS) support.

References

[1] D A Cremers and L J Radziemski, Handbook of laser-induced breakdown spectroscopy(Wiley, New York, 2006)

[2] A W Misiolek, V Palleschi and Schechter, Laser-induced breakdown spectroscopy (Cam-bridge University Press, Cambridge, 2006)

[3] J P Singh and S N Thakur, Laser-induced breakdown spectroscopy (Elsevier Science,Amsterdam, 2006)

[4] K Y Yamamoto, D A Cramers, M J Ferris and L E Foster, Appl. Spectrosc. 50, 222(1996)

[5] O Samek, D C S Beddows, H H Telle, J Kaiser, M Liska, J O Caceres and A G Urena,Spec. Chem. Acta Part B56, 865 (2001)

[6] S Y Oh, F Y Yueh, J P Singh, C C Herman and K Zeigler, Spec. Chem. Acta PartB64, 113 (2009)

1150 Pramana – J. Phys., Vol. 75, No. 6, December 2010

Author's personal copy

Journal of Alloys and Compounds 509 (2011) 3740–3745

Contents lists available at ScienceDirect

Journal of Alloys and Compounds

journa l homepage: www.e lsev ier .com/ locate / ja l l com

Quantitative elemental analysis of nickel alloys using calibration-basedlaser-induced breakdown spectroscopy

G.P. Guptaa,∗, B.M. Suria, A. Vermab, M. Sundararamanb, V.K. Unnikrishnanc, K. Alti c,V.B. Karthac, C. Santhoshc

a Laser & Plasma Technology Division, Bhabha Atomic Research Centre, Trombay, Mumbai 400085, Maharashtra, Indiab Structural Metallurgy Section, Materials Group, Bhabha Atomic Research Centre, Mumbai, Indiac Centre for Atomic and Molecular Physics, Manipal University, Manipal 576104, India

a r t i c l e i n f o

Article history:Received 25 August 2010Accepted 28 December 2010Available online 5 January 2011

Keywords:LIBSNickel alloysInternal standardizationCalibration curve

a b s t r a c t

This work reports on the quantitative elemental analysis of nickel alloys using laser-induced breakdownspectroscopy (LIBS) in air at atmospheric pressure. The LIBS plasma is generated using a Q-switchedultraviolet Nd:YAG laser, which evolves with time. The LIBS spectra of three samples with known com-position are recorded at five detector gate delays. Employing the internal standardization method, thecalibration curves for Cr present in the samples are produced. The Cr concentration in the samples isdetermined using the generated linear calibration curves having varying slopes and regression coeffi-cients at different delays. The effect of slopes and regression coefficients of the linear calibration curveson the analytical predictive capability of the LIBS system is studied through the correlation of the LIBSdetermined concentration of Cr with its known value. The analytical predictive capability of the LIBSsystem is noted to be the best when the calibration-based analysis is performed at an appropriate delay(2000 ns in the present experiment) where the linear calibration curve has both the regression coefficientand the slope close to the ideal value.

© 2011 Elsevier B.V. All rights reserved.

1. Introduction

Laser-induced breakdown spectroscopy (LIBS) [1,2] has beenwell recognized as a simple, fast and direct analytical techniqueof elemental analysis of multi-element materials by a number ofresearch groups all over the world. It is based on the focusing of ahigh-power pulsed laser beam with a power density >100 MW/cm2

onto a sample surface followed by optical emission spectroscopyof the plasma produced over the surface. The LIBS technique haswell-known attractive features over other widely used analyticaltechniques of atomic emission spectroscopy. In particular, vapor-ization and excitation of samples are possible in a single stepand no restriction has to be placed on the sample size or spe-cific requirement of sample preparation. Moreover, simultaneousmulti-element analysis can be performed irrespective of atomicnumber of elements and this technique can be employed for onlineanalysis remotely in hostile environments. During the last twodecades, LIBS has attracted a lot of attention, leading to an ever-increasing list of applications, both in laboratory and in industry.One of the major applications of this technique is in the compositionanalysis of metallic materials [3].

∗ Corresponding author. Tel.: +91 22 25595058; fax: +91 22 25505151.E-mail address: gpgupta@barc.gov.in (G.P. Gupta).

For quantitative elemental analysis of unknown samples, thecalibration curve, which is a plot of intensity of an analyte emissionline versus its concentration, is commonly made with standards ofknown elemental concentrations in a matrix close to that of theunknown samples. Although the calibration curve method is theconvenient approach for quantitative analysis, it is most suitedfor matrix-matched samples. For achieving accurate quantitativeresults from the linear calibration curves, the experimental con-ditions affecting the analytical performance need to be carefullystudied [4–6]. The slope of the calibration curve that represents thesensitivity and the regression coefficient of the calibration curvethat represents the precision are the two important parametersaffecting the analytical predictive capability of the LIBS system tounknown samples. Several LIBS research papers on metallic solidsamples in air at atmospheric pressure have been published in theliterature where the calibration curves are obtained with varyingslopes and regression coefficients [6–15]. But, to the best of ourknowledge, their effect on the correlation of the LIBS determinedconcentration with that determined by any other analytical methodis not reported in the literature.

In the present work, we have produced calibration curves for Cr,at five values of the detector gate delay relative to the laser pulse,from the LIBS spectra of three samples of nickel alloys with knowncomposition in air at atmospheric pressure using a nanosecond,ultraviolet Nd:YAG laser, employing the internal standardization

0925-8388/$ – see front matter © 2011 Elsevier B.V. All rights reserved.doi:10.1016/j.jallcom.2010.12.189

Author's personal copy

G.P. Gupta et al. / Journal of Alloys and Compounds 509 (2011) 3740–3745 3741

Fig. 1. Schematic diagram of the experimental set-up for LIBS studies.

(IS) method [10–12]. We have used the ultraviolet laser becauseit generates plasmas of samples in atmospheric air with improvedemission characteristics [16]. We have investigated the effect ofslopes and regression coefficients of the linear calibration curvesthat vary with the detector gate delay, on the quantitative ele-mental analysis considering three certified samples as unknownsamples. We have discussed the influence of these two parametersof the linear calibration curve on the analytical predictive capabilityof the LIBS system.

2. Experimental

A schematic diagram of the experimental set-up for LIBS studies is shown inFig. 1. A Q-switched Nd:YAG laser (Spectra Physics PRO 230-10) working at thethird harmonic wavelength of 355 nm with a 6 ns pulse duration and a 10 Hz rep-etition rate was focused at right angles on the sample surface, placed in air atatmospheric pressure using a bi-convex lens of focal length 20 cm. For an estimatedvalue of the diameter of the focussed beam of ∼120 �m, the laser pulse irradianceon the sample surface was about 1.5 × 1010 W/cm2. The sample was placed on aX–Y translation stage having speed 6 mm/s so that each laser pulse was incidenton a fresh surface, thus maintaining the sample in the same conditions for differ-ent measurements. The spatially integrated plasma light emission was collectedand imaged on to the spectrograph slit using an optical-fibre-based collection sys-tem. This collection system was positioned at a distance of about 20 cm from theplasma, making an angle 45◦ to the laser beam. An Echelle spectrograph-ICCD system(Andor Mechelle ME5000-DH734-18U-03PS150) was used to record the emissionspectrum. The spectrograph, with an Echelle grating, covers 200–975 nm spectralrange in a single shot with a good wavelength resolution (0.05 nm). The dispersedlight from the spectrograph was collected by a thermoelectrically cooled ICCD cam-era which is sensitive in the whole UV–VIS–NIR spectral region converting thespectral signal into digital signal. A Hg–Ar lamp, which provides sharp lines from200 to 1000 nm, was used for wavelength calibration of this system. Intensity cal-ibration of the Echelle spectrograph-ICCD system was done using NIST certifieddeuterium–quartz–tungsten–halogen lamp (Ocean Optics, USA).

The detector was gated in synchronization with the laser pulse to get maximumsignal-to-background (S/B) ratio. The detector gate width of 6 �s, which was found tobe most advantageous in terms of the S/B ratio, was kept constant. The detector gatedelay was varied in the time span 300–2000 ns for recording the plasma emissionsignals, discriminating the continuum radiation which is intense at initial delaytimes less than 300 ns and decreases at later times. We used a large detector gatewidth as used by Davies et al. [11] rather than a narrow gate width commonly usedby various research groups, in order to enhance the S/B ratio. All LIBS spectra arederived from an integration of 120 laser pulses for obtaining enough reproducibilityof the spectra.

For the present LIBS experiments, we have employed three samples of Ni2 (Cr,Mo) alloys whose elemental compositions are listed in Table 1. These nickel alloysare prepared by melting Ni (99.99% purity), Cr (99.99% purity) and Mo (99.99%purity) in appropriate ratio in a non-consumable arc furnace with tungsten elec-trode and a water-cooled copper hearth under purified argon atmosphere. Thesecompositions were chosen so that the alloys are in the single phase region in theternary phase diagram. The alloys were homogenized at 1200 ◦C for 24 h in flowingargon atmosphere followed by furnace cooling and then hot rolled to about 0.5 mm

thickness. The specimens from these alloys of 5 mm × 5 mm dimension were solu-tion treated at 1150 ◦C for 2 h. X-ray diffraction as well as transmission microscopystudies confirmed that all the specimens are in single phase. Electron probe micro-analysis also confirmed that the alloys have the same compositions as concentrationof elements used for alloy preparation.

3. Results and discussion

3.1. LIBS spectra

Typical LIBS spectra of one of the certified Ni-alloy samples(sample 1) recorded at 2000 ns delay relative to the laser pulse aredisplayed in Fig. 2(a) and (b). The spectral lines of Cr I and Ni Iof interest are shown in the inset of the figure. Three sets of suchspectra are recorded for all the samples at each of the five delaysof interest. A mean value calculated from the three spectra is usedfor the emission intensity of atomic lines, needed for obtaining thecalibration curves.

3.2. Selection of spectral lines

The selection of the spectral lines of the analyte element as wellas the internal standard (reference) element is required to be care-fully made for the reproducibility of the analytical results using theIS method [10–12]. As stated in [10–12], the spectral lines of boththe analyte element and the reference element selected for the cal-ibration curves should fulfill the four conditions: (i) they shouldbe reasonably strong and isolated to avoid interference with otherspectral lines, (ii) they should be non-resonant lines with lowerlevels much above the ground level particularly at higher concen-trations of the elements of interest to avoid saturation of emissionintensities due to self-absorption of the emission lines, (iii) theyshould have the upper levels with energies close to each othersuch that their energy level difference is near zero (<2000 cm−1)to minimize the plasma temperature effect on the reproducibilityof the line intensity ratio and (iv) they should be simultaneouslydetected in a single laser shot to avoid complexities in the line

Table 1Elemental concentrations (wt%) in three certified samples of Ni2 (Cr, Mo) alloys.

Sample Cr Mo Ni

1 16.67 16.67 66.662 21.01 12.33 66.663 25.01 8.33 66.66

Author's personal copy

3742 G.P. Gupta et al. / Journal of Alloys and Compounds 509 (2011) 3740–3745

a

400395390385380375370365360355350

0

75000

150000

225000

300000

375000

450000

525000

600000

675000

750000

825000

900000

975000

1050000

1125000

1200000

353.00352.75352.50352.25352.00351.75

0

10000

20000

30000

40000

50000

60000

70000

80000

Inte

nsit

y (

co

un

ts)

Wavelength (nm)

Ni I

Inte

nsit

y (

co

un

ts)

Wavelength (nm)

Ni I

b

550545540535530525520515510505500

0

2500

5000

7500

10000

12500

15000

17500

20000

22500

25000

27500

30000

32500

35000

37500

40000

42500

45000

47500

50000

522.0521.5521.0520.5520.0

0

5000

10000

15000

20000

25000

30000

35000

40000

45000

50000

Inte

nsit

y (

co

un

ts)

Wavelength (nm)

Cr I

Inte

nsit

y (

co

un

ts)

Wavelength (nm)

Cr I

Fig. 2. Typical LIBS spectra of a Ni alloy (certified sample 1) at 2000 ns delay with inset showing Ni I line at 352.45 nm of interest (a) and with inset showing Cr I line at520.84 nm of interest (b).

intensity calibration. The foremost requirement of similar upperlevel energies of the analyte and reference lines is not fulfilled inmany LIBS works based on calibration curves [6,10–12], affectingslope and regression coefficient of the calibration curves. For thecalibration curves of Cr using the IS method, we have chosen theanalyte Cr and the reference Ni lines as shown in Table 2. The relatedspectroscopic parameters of the spectral lines, taken from the NISTatomic database [17], are also shown in the table. The energy level

Table 2Atomic emission lines used for quantitative analysis, along with the related spec-troscopic parameters.

Element Species Wavelength (nm) Transition

Lower levelenergy (cm−1)

Upper levelenergy (cm−1)

Cr Cr I 520.84 7593.2 26,787.5Ni Ni I 352.45 204.8 28,569.2

difference of the upper levels of the chosen line pairs is about1780 cm−1, thus satisfying the first requirement. These lines arenon-resonant as well as interference-free, thus satisfying the sec-ond and third requirements. The fourth condition is automaticallysatisfied because the spectral range detected in a single laser shotusing the present LIBS system is very large, covering the chosen linepairs.

3.3. Quantitative analysis using the IS method

The quantitative analysis of the spectral emission from LIBSplasma involves relating the emission intensity of an atomic lineof any element in the plasma to the concentration of that elementin the sample. It relies on the assumption that the plasma is opti-cally thin and in local thermodynamic equilibrium (LTE). Under thisassumption, the intensity ratio of an atomic spectral line emittedby an analyte element (denoted by subscript a) to that emitted bythe internal standard (reference) element (denoted by subscript r),

Author's personal copy

G.P. Gupta et al. / Journal of Alloys and Compounds 509 (2011) 3740–3745 3743

-1.2-1.3-1.4 -0.9-1.0-1.1-1.4

-1.2

-1.0

-0.8

-0.6

-0.4

-0.2a b

c d

e

y = 1.620 + 2.109 x

R2 = 0.945

ln (concentration ratio Cr / Ni)

-0.9-1.0-1.1-1.2-1.3-1.4-1.2

-1.1

-1.0

-0.9

-0.8

y = - 2.408E-4 + 0.850 x

R2 = 0.978

ln (concentration ratio Cr/Ni)

-0.9-1.0-1.1-1.2-1.3-1.4-1.1

-1.0

-0.9

-0.8

-0.7

-0.6

y = 0.053 + 0.769 x

R2 = 0.996

ln (concentration ratio Cr/Ni)

-0.9-1.0-1.1-1.2-1.3-1.4-0.7

-0.6

-0.5

y = - 0.271 + 0.286 x

R2 = 0.966

ln (concentration ratio Cr/Ni)

-0.9-1.0-1.1-1.2-1.3-1.4-0.8

-0.6

-0.4

-0.2

0.0

y = 1.390 + 1.483 x

R2 = 0.998

ln (

inte

nsit

y r

ati

o I

Cr / I

Ni)

ln (

inte

nsit

y r

ati

o I

Cr / I

Ni)

ln (

inte

nsit

y r

ati

o I

Cr / I

Ni)

ln (

inte

nsit

y r

ati

o I

Cr / I

Ni)

ln (

inte

nsit

y r

ati

o I

Cr / I

Ni)

ln (concentration ratio Cr/Ni)

Fig. 3. Calibration curves for Cr at delay times 300 ns (a), 500 ns (b), 700 ns (c), 1000 ns (d) and 2000 ns (e). Linear regression equations and their coefficients are also given.

considered in the IS method, is given as [10–12,18]

IaIr

= Na

Nr

ga

gr

Aa

Ar

Zr

Za

�r

�aexp

[− (Ea − Er)

kT

](1)

Here I stands for the intensity of the spectral lines, N is the totalatom number density, Z is the partition function, E and g are theenergies and the degeneracies of the upper levels respectively, Aand � are the Einstein coefficient and the wavelength respectivelyfor the observed line transitions, k is the Boltzmann constant and Tis the plasma temperature.Taking the natural logarithms in Eq. (1),

one obtains:

lnIaIr

= lnNa

Nr+ ln

gaAaZr�r

grArZa�a− Ea − Er

kT(2)

Thus, if the logarithm of the intensity ratio versus the logarithmof the concentration ratio is plotted, the calibration curve as astraight line with a slope of unity is obtained. Using the regressionequation of the linear calibration curve one may obtain the ana-lyte concentration from the measured analyte spectral intensityin unknown samples. Since the LIBS data are normally not per-

Author's personal copy

3744 G.P. Gupta et al. / Journal of Alloys and Compounds 509 (2011) 3740–3745

Table 3Slope (m) and linear regression coefficient (R2) of the calibration curves of Cr at fivevalues of the detector gate delay using the three certified samples of Ni alloys.

Delay time (ns) m R2

300 2.109 0.945500 0.850 0.978700 0.769 0.996

1000 0.286 0.9662000 1.483 0.998

fectly reproducible, they do not fit perfectly on a straight line withregression coefficient equal to one. The slope values of the linearcalibration curves are commonly observed to differ from the idealslope of unity, owing to the inhomogeneity and matrix effect of thesamples [10,18]. Both the regression coefficient and the slope of thelinear calibration curve affect the analytical predictive capability ofthe LIBS system. When the energy difference of the upper energylevels is much smaller as compared to kT, the effect of the last termin Eq. (2) that contributes only to the intercept of this straight line,affecting the reproducibility of the spectral line intensity ratio, isnegligible. Under this condition the intercept is determined by thesecond term on the right hand side of Eq. (2). Here all the compo-nents except Z remain constant for the selected lines, whereas Z isdependent on the plasma temperature.

3.4. Calibration curves for Cr

Using the analyte Cr and reference Ni lines given in Table 2 andEq. (2), we have produced the calibration curves for Cr at five delaytimes and depicted these in Fig. 3. The logarithm of the relativeintensity ratio of the analyte (Cr) element and the reference (Ni)element lines are plotted against the logarithm of the given rela-tive concentrations of the three Ni alloy samples for the calibrationcurves of Cr. The linear regression equation and the regression coef-ficient for each of the calibration curves are also given in the figure.As seen from the figures, the calibration curves in our experimentare well characterized by a straight line without any saturationeffect. The slopes and the regression coefficients of these linearcalibrations curves are tabulated in Table 3. As evident from thistable, the slope values differ from the ideal slope of unity, whereasthe regression coefficient values of the linear fit are near unity withthe best value equal to 0.998 at 2000 ns delay. From the regressionequations, it is seen that the intercept varies widely over the delaytimes of interest. The radiative lifetimes of electronic energy levelsare in ns, and the observation time of 6 �s more or less ensures that,for any delay, we observe the emissions from all the atoms reach-ing the levels involved. This intercept is varying from a positivevalue to zero to slightly negative, and then to positive. This changeis ascribed to the dependence of Z on the plasma temperature thatvaries significantly with the delay time.

3.5. Analytical results

For the demonstration of analytical predictive capability of thepresent LIBS system, we have used the linear regression equations

0.40.30.2

0.2

0.3

0.4

y = 0.002 + 0.993 x

R2 = 0.998

LIB

S d

ete

rmin

ed

co

ncen

trati

on

rati

o

Certified concentration ratio

Fig. 4. Correlation of the LIBS determined concentration ratio Cr/Ni and certifiedconcentration ratio Cr/Ni using the calibration curve at a delay time of 2000 ns.

of the produced calibration curves and determined the relativeconcentration of Cr in three nickel-alloy samples with known com-position at 5 detector gate delays. We have also evaluated theaccuracy of the elemental determinations using the LIBS methodby their relative deviation from the certified values of the samples.Table 4 shows the correlation of the LIBS determined concen-tration ratio Cr/Ni with its certified value and the correspondinguncertainty of the three Ni-alloy samples using the producedcalibration curves at 5 detector gate delays. It is observed thatthe high regression coefficient with its value very close to unity(0.998) along with the slope (1.483) of the linear calibration curveclose to the ideal value of unity, obtained at 2000 ns gate delayin the present experiment, yields the best LIBS analytical resultsin all the three samples. Thus, the analytical predictive capabil-ity of the LIBS system is strongly dependent on both the slopeand the regression coefficient of the calibration curve, the bestcapability occurring at a particular gate delay where both theslope and the regression coefficient of the linear calibration curveare close to the ideal value of unity. In Fig. 4, we have depictedthe correlation of the LIBS determined concentration ratio Cr/Niand certified concentration ratio Cr/Ni for the three Ni-alloy sam-ples using the calibration curve at a delay time of 2000 ns wherethe correlation coefficient and the slope of the linear calibrationcurve are very close to the ideal value. As seen from the fig-ure, the LIBS analytical results are fitted satisfactorily by a linearequation with the regression coefficient of 0.998 and the slopeof 0.993. These results confirm high analytical predictive capa-bility of the LIBS system provided the linear calibration curve isproduced at an appropriate detector gate delay, having both the

Table 4Correlation of the LIBS determined concentration ratio Cr/Ni with its certified value and the corresponding uncertainty of the three samples of Ni alloys using calibrationcurves at five values of the detector gate delay.

Sample Certified concentration ratio Cr/Ni LIBS determined concentration ratio Cr/Ni (correlation uncertainty, %)

Delay time (ns)

300 500 700 1000 2000

1 0.250 0.256 (2.4) 0.246 (1.6) 0.252 (0.8) 0.245 (2.0) 0.249 (0.4)2 0.315 0.298 (5.4) 0.326 (3.5) 0.311 (1.3) 0.329 (4.4) 0.318 (1.0)3 0.375 0.387 (3.2) 0.368 (1.9) 0.378 (0.8) 0.365 (2.7) 0.373 (0.5)

Author's personal copy

G.P. Gupta et al. / Journal of Alloys and Compounds 509 (2011) 3740–3745 3745

regression coefficient and the slope close to the ideal value ofunity.

4. Conclusions

We have recorded the LIBS spectra of three Ni-alloy samplesof known composition in air at atmospheric pressure using a Q-switched Nd:YAG ultraviolet laser at five detector gate delays witha view to investigate the effect of the slopes and the regressioncoefficients of linear calibration curves on the analytical predic-tive capability of the LIBS system. We have chosen the appropriatespectral lines of both the analyte and the reference elements, com-plying with the requirement of the IS method for the calibrationcurves. We have produced the calibration curves for Cr from theLIBS spectra, recorded at five detector gate delays, of three samplesof nickel alloys and obtained the linear regression equations withtheir slopes and regression coefficients. Using these equations, wehave determined the Cr concentration in three certified Ni-alloysamples from the measurement of its LIBS spectral intensity. Wehave studied the effect of slopes and regression coefficients of thelinear calibration curves on the analytical predictive capability ofthe LIBS system through the correlation of the LIBS determinedconcentration of Cr with its certified value at five detector gatedelays. The analytical predictive capability of the LIBS system isfound to be the best when the linear calibration curve with regres-sion coefficient and slope close to the ideal value corresponding toan appropriate detector gate delay, which occurred in the presentexperiment at 2000 ns gate delay, is utilized for the quantitativeelemental analysis.

Acknowledgments

This work was carried out under the project titled “Trace Ele-ment Analysis for Environmental and Biomedical Applications –Development of Laser Induced Breakdown Spectroscopy (LIBS)Technique.” Project No. 2007/34/14-BRNS/87, Board of Researchin Nuclear Sciences (BRNS), Department of Atomic Energy (DAE),Govt. of India.

References

[1] D.A. Cremers, L.J. Radziemski, Handbook of Laser-Induced Breakdown Spec-troscopy, John Wiley & Sons Ltd., England, 2006.

[2] A.W. Miziolek, V. Pallesschi, I. Schechter (Eds.), Laser-Induced Breakdown Spec-troscopy (LIBS), Cambridge University Press, New York, 2006.

[3] W.B. Lee, J.Y. Wu, Y.I. Lee, J. Sneddon, Appl. Spectrosc. 39 (2004) 27.[4] X. Mao, W.-T. Chan, M. Caetano, M.A. Shannon, R.E. Russo, Appl. Surf. Sci. 96–98

(1996) 126.[5] J.A. Aguilera, C. Aragon, F. Penalba, Appl. Surf. Sci. 127–129 (1998) 309.[6] C. Aragon, J.A. Aguilera, F. Penalba, Appl. Spectrosc. 53 (1999) 1259.[7] N. Andre, C. Geertsen, J.-L. Lacour, P. Mauchien, S. Sjostrom, Spectrochim. Acta,

Part B 49 (1994) 1363.[8] M. Sabsabi, P. Cielo, Appl. Spectrosc. 49 (1995) 499.[9] A. Gonzalez, M. Ortiz, J. Campos, Appl. Spectrosc. 49 (1995) 1632.

[10] B. Bescos, J. Castano, A.G. Urena, Laser Chem. 16 (1995) 75.[11] C.M. Davies, H.H. Telle, D.J. Montgomery, R.E. Corbett, Spectrochim. Acta, Part

B 50 (1995) 1059.[12] I. Bassiotis, A. Diamantopoulou, A. Giannoudakos, F.R. Kalantzopoulou, M. Kom-

pitsas, Spectrochim. Acta, Part B 56 (2001) 671.[13] L. Fornarini, F. Colao, R. Fantoni, V. Lazic, V. Spizzicchino, Spectrochim. Acta,

Part B 60 (2005) 1186.[14] A. De Giacombo, M. Dell’Aglio, O. De Pascale, R. Gaudiuso, R. Teghil, A. Santagata,

G.P. Parisi, Appl. Surf. Sci. 253 (2007) 7677.[15] B.C. Windom, D.W. Hahn, J. Anal. Atom. Spectrom. 24 (2009) 1665.[16] K. Kagawa, K. Kawai, M. Tani, T. Kobayashi, Appl. Spectrosc. 48 (1994) 198.[17] NIST Atomic Spectra Database, http://physics.nist.gov.[18] R.S. Adrain, J. Watson, J. Phys. D: Appl. Phys. 17 (1984) 1915.

PRAMANA c© Indian Academy of Sciences Vol. 74, No. 6— journal of June 2010

physics pp. 983–993

Measurements of plasma temperatureand electron density in laser-inducedcopper plasma by time-resolved spectroscopyof neutral atom and ion emissions

V K UNNIKRISHNAN1, KAMLESH ALTI1, V B KARTHA1, C SANTHOSH1,∗,G P GUPTA2 and B M SURI21Centre for Atomic and Molecular Physics, Manipal University, Manipal 576 104, India2Laser and Plasma Technology Division, Bhabha Atomic Research Centre,Mumbai 400 085, India∗Corresponding author. E-mail: santhosh.cls@manipal.edu

MS received 7 January 2010; revised 3 February 2010; accepted 4 February 2010

Abstract. Plasma produced by a 355 nm pulsed Nd:YAG laser with a pulse durationof 6 ns focussed onto a copper solid sample in air at atmospheric pressure is studiedspectroscopically. The temperature and electron density characterizing the plasma aremeasured by time-resolved spectroscopy of neutral atom and ion line emissions in the timewindow of 300–2000 ns. An echelle spectrograph coupled with a gated intensified chargecoupled detector is used to record the plasma emissions. The temperature is obtainedusing the Boltzmann plot method and the electron density is determined using the Saha–Boltzmann equation method. Both parameters are studied as a function of delay timewith respect to the onset of the laser pulse. The results are discussed. The time windowwhere the plasma is optically thin and is also in local thermodynamic equilibrium (LTE),necessary for the laser-induced breakdown spectroscopy (LIBS) analysis of samples, isdeduced from the temporal evolution of the intensity ratio of two Cu I lines. It is foundto be 700–1000 ns.

Keywords. Laser-induced plasma; spectroscopy; plasma temperature; electron density.

PACS Nos 52.50.Jm; 52.70.Kz; 52.25.Os

1. Introduction

Pulsed laser-induced plasmas (LIPs) of metals and alloys formed at laser pulse ir-radiances near the plasma ignition threshold are of great interest since they haveseveral important applications, e.g. material processing, thin film deposition andmetal analysis in solid samples [1]. Optical emission spectra of an LIP consist ofatomic and ionic lines, superimposed on a continuum of radiation. Elemental analy-sis of the sample based on the optical emission spectra from an LIP is known as

983

V K Unnikrishnan et al

laser-induced plasma spectroscopy (LIPS), also called as laser-induced breakdownspectroscopy (LIBS). The LIBS technique, utilizing a pulsed LIP formed near theplasma ignition threshold as a spectroscopic source is a well-known analytical tech-nique to provide remote, in-situ, rapid and multi-elemental analysis of bulk andtrace sample in any phase (solid, liquid and gas) with no or minimal sample prepa-ration [2–4].

The characterization of LIPs by determining their temperature and electron den-sity is essential and has gained considerable interest in recent years for the un-derstanding and exploitation of these complex and versatile spectroscopic sources.The plasma characteristics are dependent on laser irradiance, wavelength, pulseduration, target material, atmospheric conditions, space and time. For elementalanalysis using LIBS, it is important that LIP should be optically thin and in LTE

Under LTE condition in the plasma, the excitation temperature governing thedistribution of energy level excitation through the Boltzmann equation and the ion-ization temperature governing the ionization equilibrium through the Saha equationare equal to the electronic temperature describing the Maxwellian distribution ofelectron velocities [5]. Thus, one describes the plasma in LTE by a common tem-perature T , called the plasma temperature. Optical emission spectroscopy has re-cently attracted a lot of attention for characterizing an LIP. The most widely usedspectroscopic method for the determination of T is the Boltzmann plot method[5] which employs the ratio of integrated line intensities for two or more atomiclines. Among several diagnostic methods for measuring the plasma electron den-sity ne, plasma spectroscopy based on either Stark broadening of spectral lines orthe Saha–Boltzmann equation is considered as the simplest method [5].

For the spectroscopic investigation of solid targets, several workers studied LIPfrom solid copper. Lee et al [6] carried out time-integrated, space-resolved studies oflaser-ablated plasma emission with Cu target in air at atmospheric pressure using a193 nm pulsed excimer laser and determined T using the emission spectra. Mao et al[7] characterized an LIP from a solid Cu target in air using a 248 nm pulsed excimerlaser. They have employed time-integrated emission spectroscopy in the plasmacharacterization. Pietch and his co-workers [8] studied the expansion of Cu plasmaand its distribution, formed by a 308 nm pulsed excimer laser in air at reducedpressure (20 mTorr), using a gated intensified charge-coupled device (ICCD) camerafor spectroscopic applications. Wu et al [9] investigated the dynamics of Cu plasmagenerated by a 308 nm pulsed excimer laser in air at reduced pressure (<1 mTorr)by optically examining the plasma plume. Hafez et al [10] studied the characteristicsof Cu plasma produced by a 355 nm pulsed Nd:YAG laser interaction with a solidtarget in vacuum and argon buffer gas using the plasma spectroscopy and Langmuirprobe methods. They determined T using the Boltzmann plot and ne using theStark line broadening.

In this paper, we report the measurements of plasma temperature T and elec-tron density ne in Cu plasma formed by irradiation of a solid Cu target in airat atmospheric pressure with a 355 nm pulsed Nd:YAG laser, using time-resolvedspectroscopy of atom and ion emissions with a gated ICCD camera coupled withan echelle spectrograph. The laser irradiance of 4.5× 108 W/cm2 employed in thiswork was near the plasma ignition threshold. The aim of this investigation was toidentify and optimize laser ablation parameters suitable for elemental composition

984 Pramana – J. Phys., Vol. 74, No. 6, June 2010

Measurements in laser-induced copper plasma

analysis of samples using the LIBS technique. We have used the Boltzmann plotmethod for determining T and the Saha–Boltzmann equation method for deter-mining ne, instead of using Stark line broadening. From these measurements wehave found the time window where the plasma is optically thin and in LTE, whichis a necessary requirement for the applicability of the equilibrium equations andemission signals to elemental analysis using the LIBS technique.

2. Experimental details

The schematic diagram of the experimental set-up for the LIBS study is presentedin figure 1. The Q-switched Nd:YAG laser (Spectra Physics PRO 230-10) wasoperated at the third harmonic wavelength of 355 nm, pulse width of 6 ns andrepetition rate of 10 Hz. The laser was focussed on a solid copper target placedin air at atmospheric pressure using a bi-convex lens of focal length 20 cm. Thisprovides a laser pulse irradiance of 4.5 × 108 W/cm2, which is near to the plasmaignition threshold, forming the plasma over the target surface. The target wasplaced on an X–Y translation stage having a speed of 6 mm/s so that every laserpulse was incident on a fresh location of the target. The spatially integrated plasmalight emission was collected and imaged onto the spectrograph slit using an opticalfibre-based collection system. This collection system was positioned at a distance ofabout 20 cm from the plasma, making an angle 45◦ to the laser beam. An echellespectrograph–ICCD system (Andor Mechelle ME5000-DH734-18U-03PS150) wasused to record the emission spectrum. The spectrograph with an echelle gratingcovers 200–975 nm spectral range in one setting with a good wavelength resolution(0.05 nm). The spectrally dispersed light from the spectrograph was collected bya thermoelectrically cooled ICCD camera which is sensitive in the whole UV–VIS–NIR region, converting the spectral signal into digital signal. The detector was

Figure 1. Experimental lay-out of LIBS system.

Pramana – J. Phys., Vol. 74, No. 6, June 2010 985

V K Unnikrishnan et al

Figure 2. Mercury doublet spectrum recorded using echelle spectrograph(dotted lines) and Czerny-turner spectrograph (continuous lines).

gated in synchronization with the laser pulse to get maximum signal-to-noise ratio.The detector gate width was kept constant at 750 ns whereas its delay time was var-ied in the time span 300–2000 ns for recording the plasma emission signals. A Hg–Arlamp, which provides sharp lines from 200 to 1000 nm, was used for wavelengthcalibration of this system. Intensity calibration of the echelle spectrograph–ICCDsystem was done using NIST certified deuterium–quartz–tungsten–halogen lamp(Ocean Optics, USA).

The unique feature of the LIBS system used for our study is the usage of anechelle spectrograph which provides broad spectral coverage with a good resolu-tion. The echelle spectrograph disperses the spectral components of the collectedplasma light in both X and Y directions to fill a 2D CCD. Echelle spectrographsdiffer from Czerny-turner spectrographs in the fact that the first one has got twodispersive elements (a dual order prism and a grating) whereas the latter has onlyone grating. These elements disperse light at 90◦ to one another. Hence the spread-out light on the CCD gives both X and Y spectral information. This provides widespectral band pass, as given by a low-dispersion Czerny-turner, together with high-est resolution of a high-dispersion Czerny-turner. It is observed that the resolutionof the spectrum recorded with the echelle spectrograph is better by a factor of∼10 compared to Czerny-turner spectrograph as shown in figure 2. Apart fromthis, in the present echelle spectrograph there are no moving parts which make thesystem more convenient and reliable for calibration and fast analysis. This echellespectrograph–ICCD system will be useful for multi-elemental composition analysisof samples using the LIBS technique planned in future.

3. Theoretical description

For the interpretation of spectroscopic data, one requires a plasma ionization modelto describe the ionization state and atom/ion energy level populations in terms

986 Pramana – J. Phys., Vol. 74, No. 6, June 2010

Measurements in laser-induced copper plasma

of plasma temperature and electron density. We present below the methods fordetermining the LIP parameters T and ne for optically thin and LTE plasmas.

3.1 Boltzmann plot method for T

For plasma in LTE, the energy level populations of the species are given by theBoltzmann distribution law [5],

nk,Z

nZ=

gk,Z

PZexp

(−Ek,Z

kBT

). (1)

Here, the index Z refers to the ionization stage of the species (Z = 0 and 1 corre-sponding to the neutral and singly ionized atoms respectively), kB is the Boltzmannconstant, T is the plasma temperature, nk,Z , Ek,Z and gk,Z are the population, en-ergy and degeneracy of the upper energy level k respectively, nZ is the numberdensity and PZ is the partition function of the species in ionization stage Z. Theintegrated intensity IZ of a spectral line occurring between the upper energy levelk and the lower energy level i of the species in ionization stage Z in optically thinplasma, i.e. plasma in which only very little radiation is absorbed, is given as

IZ =hc

4πλki,ZAki,Znk,ZL, (2)

where h is the Planck constant, c is the speed of light, L is the characteristic lengthof the plasma, Aki,Z is the transition probability and λki,Z is the transition linewavelength. Using eq. (1), eq. (2) can be rewritten as

IZ =hc

4πλki,ZAki,ZL

nZ

PZgk,Z exp

(−Ek,Z

kBT

). (3)

By taking the natural logarithm, eq. (3) can be rewritten as

ln(

IZλki,Z

gk,ZAki,Z

)= − 1

kBTEk,Z + ln

(hcLnZ

4πPZ

). (4)

This yields a linear plot (the so-called Boltzmann plot) if one represents the mag-nitude on the left-hand side for several transitions against the energy of the upperlevel of the species in ionization stage Z. The value of T is deduced from the slopeof the Boltzmann plot. As eq. (4) is obtained under the assumption of plasmabeing optically thin as well as in LTE, the applicability of this equation is limitedto LTE and optically thin plasmas.

3.2 Saha–Boltzmann equation method for ne

The electron density using atom and ion spectral lines emitted from the plasma isdetermined from the Saha–Boltzmann equation as [5,11]

Pramana – J. Phys., Vol. 74, No. 6, June 2010 987

V K Unnikrishnan et al

ne =I∗Z

I∗Z+1

6.04× 1021(T )3/2

× exp[(−Ek,Z+1 + Ek,Z − χZ)/kBT ] cm−3, (5)

where I∗Z = IZλki,Z/gk,ZAki,Z and χZ is the ionization energy of the species in theionization stage Z. The lowering of the ionization energy due to the interactions inthe plasma is negligibly small which has been omitted in eq. (5).

3.3 Optically thin plasma

The elemental composition analysis from the line intensities in a LIBS experimentbecomes simple if the plasma is optically thin and is also in LTE. It is thus necessaryto know the time window for time-evolving plasma like LIPs where the plasma isoptically thin as well as in LTE. Using eq. (3), the intensity ratio of two lines ofthe same species of ionization stage Z is expressed as

I1

I2=

(λnm,Z

λki,Z

) (Aki,Z

Anm,Z

)(gk,Z

gn,Z

)exp

(−Ek,Z − En,Z

kBT

), (6)

where I1 is the line intensity from the k–i transition and I2 is that from the n–mtransition. If we consider two emission lines having the same upper level or as close

Figure 3. LIBS spectra of copper recorded using ICCD-based echelle spec-trograph with a gate delay time of 700 ns at a laser irradiance of 4.5×108 W/cm2, showing (A) Cu I atomic lines and (B) Cu II ionic lines usedfor the characterization of the laser-induced Cu plasma.

988 Pramana – J. Phys., Vol. 74, No. 6, June 2010

Measurements in laser-induced copper plasma

Table 1. Wavelength, lower and upper energy levels, upper level degeneracyand transition probability for the Cu I and Cu II emission lines used in thiswork.

Wavelength Upper level Lower level Upper level TransitionAtom/ion (nm) energy (eV) energy (eV) degeneracy probability (s−1)

Cu I 465.11 7.740 5.072 8 3.8×107

Cu I 510.55 3.817 1.389 4 2.0×106

Cu I 515.32 6.191 3.786 4 6.0×107

Cu I 521.82 6.192 3.817 6 7.5×107

Cu I 578.21 3.786 1.642 2 1.65×107

Cu II 268.93 13.392 8.783 7 4.1×107

Cu II 271.35 13.432 8.864 5 6.8×107

as possible, the temperature effect of the Boltzmann factor on the reproducibilityof the line intensity ratio is minimized and at the same time the consideration ofthe efficiency factor of the collecting system is avoided. Neglecting the exponentialfactor in that condition, one can find out the theoretical value of the intensity ratioof the two lines by using the atomic parameters of the transitions. By matchingthis ratio with the measured values at different delay times, one finds out the timewindow where the plasma is optically thin.

4. Results and discussion

Using the neutral atom and ion emission spectra recorded at different delay times inthe time span 300–2000 ns, we have characterized the LIP in terms of its transientT and ne. Figure 3 shows a typical spectrum recorded using ICCD-based echellespectrograph with a gate delay time of 700 ns, depicting Cu I and Cu II emissionlines from the LIP. Five Cu I and two Cu II emission lines which are well resolvedand free from spectral interference are chosen in the present work. These lines alongwith their spectroscopic parameters, taken from the NIST atomic database [12], areshown in table 1. The value of T is obtained from the Boltzmann plot made fromthe analysis of the five recorded Cu I lines at a given delay time. Figure 4 showsone such Boltzmann plot from the intensities of these Cu I lines at a delay time of700 ns, the slope of which gives T = 0.79 eV. The estimated values of T at severaldelay times are presented in table 2. These are plotted in figure 5. It is observedthat after 500 ns delay time the plasma cools down exponentially.

The value of ne is obtained from eq. (5) using the measured intensity ratio ofCu I and Cu II lines at a given delay time. We have considered three intensityratios, 515.32 nm Cu I and 268.93 nm Cu II, 515.32 nm Cu I and 271.35 nm Cu IIand 521.82 nm Cu I and 268.93 nm Cu II and obtained the values of ne. As seenfrom figure 3, Cu II lines at 277 and 279 nm are more intense than the chosen CuII lines for the analysis. However, we could not use these more intense Cu II linesas the Cu II line at 277 nm overlaps with the three Cu I lines at 276.637, 276.639and 276.888 nm and the transition probability of the Cu II line at 279 nm is not

Pramana – J. Phys., Vol. 74, No. 6, June 2010 989

V K Unnikrishnan et al

Figure 4. Boltzmann plot made from the analysis of five Cu I lines, 465.11,510.55, 515.32, 521.82 and 578.21 nm, considering the intensities at a delaytime of 700 ns. The continuous line represents the result of a linear bestfit. I and λ are the intensity and the wavelength of a transition from upperlevel k of energy Ek and statistical weight gk to lower level i with Aki asthe corresponding transition probability. The slope gives the temperatureas 0.79 eV.

Figure 5. Variation of plasma temperature with delay time.

990 Pramana – J. Phys., Vol. 74, No. 6, June 2010

Measurements in laser-induced copper plasma

Figure 6. Variation of electron density with delay time.

Table 2. Plasma temperature and electron density as a function of delay timeof the detector relative to the onset of the laser pulse on the sample.

Delay time Plasma temperature Electron density(ns) (eV) (cm−3)

300 0.78 2.0×1014

500 0.84 1.1×1015

700 0.79 5.7×1014

1000 0.75 2.9×1014

2000 0.69 4.5×1013

given in the NIST atomic database. The arithmetic mean of the three values of ne

is represented as the average value of ne. We have presented these average valuesas the electron density at several delay times in table 2 and graphically in figure 6.It is observed that after 500 ns delay the electron density decreases exponentially.

The time window where the plasma is optically thin and is also in LTE is inferredfrom the temporal evolution of the intensity ratio of two Cu I lines, 515.32 and521.82 nm, which have upper levels having very close energy as shown in table1. Figure 7 shows the temporal evolution of the intensities of these lines and theintensity ratio between them. We have calculated the intensity ratio for this couple

Pramana – J. Phys., Vol. 74, No. 6, June 2010 991

V K Unnikrishnan et al

Figure 7. Temporal evolution of intensities of two Cu I lines 515.32 and521.82 nm and their intensity ratio. The straight line indicates the theoreticalintensity ratio = 1.85 for this couple of lines, which is the condition of anoptically thin and LTE plasma. The time window for the thin and LTE plasmais 700–1000 ns.

of lines using eq. (6) which is equal to 1.85 and shown this theoretical value as astraight line in the same figure. Comparing the experimental data of the intensityratio with the theoretical one, we have inferred the time window 700–1000 ns wherethe LIP produced is thin as well as in LTE.

5. Conclusions

We have determined time-resolved values of the plasma temperature from the Boltz-mann plots made from the analysis of five observed Cu I spectral lines. We have also

992 Pramana – J. Phys., Vol. 74, No. 6, June 2010

Measurements in laser-induced copper plasma

determined time-resolved values of the electron density from the Saha–Boltzmannequation which relates the electron density with the intensity ratio of atomic andionic emission lines. The average value of the electron density at a given delay timeis calculated by considering the measured intensity ratios of Cu I and Cu II linesat various wavelengths. From the temporal evolution of the intensity ratio of twoCu I lines and matching it with the known value, we have inferred the time windowwhere the plasma is optically thin and is also in LTE. This time window is foundto be 700–1000 ns.

Acknowledgement

The authors are thankful to BRNS, DAE, Govt. of India for the financial supportprovided through a LIBS project (Project No. 2007/34/14-BRNS).

References

[1] D B Chrisey and G K Hubler, Pulsed laser deposition of thin films (Wiley, New York,1994)

[2] L J Radziemski and D A Cremers, Laser-induced plasmas and applications (MarcelDekker Inc., New York, 1989)

[3] A W Miziolek, V Pallesschi and I Schecchter, Laser-induced breakdown spectroscopy(Cambridge University Press, Cambridge, 2006)

[4] D A Cremers and L J Radziemski, Handbook of laser-induced breakdown spectroscopy(John Wiley & Sons Ltd, West Sussex, 2006)

[5] H R Griem, Principles of plasma spectroscopy (Cambridge University Press, Cam-bridge, 1997)

[6] Y I Lee, S P Sawan, T L Thiem, Y Y Teng and J Sneddon, Appl. Spectrosc. 46, 436(1992)

[7] X L Mao, M A Shannon, A J Fernandez and R E Russo, Appl. Spectrosc. 49, 1054(1995)

[8] W Pietch, B Dubreuil and A Briand, Appl. Phys. B61, 267 (1995)[9] J D Wu, Q Pan and S C Chen, Appl. Spectrosc. 51, 883 (1997)

[10] M A Hafez, M A Khedr, F F Elaksher and Y E Gamal, Plasma Source Sci. Technol.12, 185 (2003)

[11] J M Gomba, C D’ Angelo, D Bertuccelli and G Bertuccelli, Spectrochimica Acta B56,695 (2001)

[12] NIST Atomic Spectra Database, http://physics.nist.gov

Pramana – J. Phys., Vol. 74, No. 6, June 2010 993

Trace Element Analysis Using Laser Induced Breakdown Spectroscopy (LIBS) Technique

V.K. Unnikrishnan1, K. Mridul1, Rajesh Nayak1, Kamlesh Alti1, V.B. Kartha1, C. Santhosh1* and B.M. Suri2

1Centre for Atomic and Molecular Physics, Manipal University, Manipal-576104, INDIA 2Laser & Plasma Technology Division, Bhabha Atomic Research Centre, Mumbai - 400085, INDIA.

E.mail : santhosh.cls@manipal.edu.

Abstract. Laser-Induced Breakdown Spectroscopy (LIBS) is well recognized as a promising tool for in situ/remote elemental analysis of environmental, archeological, clinical, and hazardous samples. With the aim of quantifying trace elements in such samples, using LIBS technique, an echelle spectrograph-ICCD system with broad range and good resolution has been assembled. Various important parameters of this system were studied and optimized. Conditions for getting good quality LIBS spectra and signal for multielemental analysis have been achieved, and these are discussed and illustrated in this paper.

Keywords: Laser ablation; Echelle spectrograph; Time resolved laser spectroscopy.PACS: 52.38.Mf; 07.60.Rd; 42.65.-k

INTRODUCTION

The Laser Induced Breakdown Spectroscopy (LIBS) has developed as an analytical technique over the past two decades [1-3]. The technique employs a pulsed laser and a focusing lens which vaporizes exposed region of sample generating plasma. Recombining plasma emit sample specific radiations which is then collected by a spectrograph and detector. The versatility of LIBS technique for multi-element analysis and its applicability to different sample types (solid, liquid and gas) makes it attractive in detecting and quantifying various elements [4-6].

EXPERIMENTAL

The schematic of the LIBS set-up used for this study is shown in Figure 1. The 3rd harmonic of Nd-YAG laser (Spectra Physics PRO 230-10) with pulse duration 6ns @10Hz and energy 400mJ was used for ablation of materials to form the plasma. This is then focused on to the sample material using a bi-convex lens in order to achieve appropriate breakdown threshold irradiance for different samples. A collecting/collimating lens/mirror system is used for collecting the generated plasma for the best performance of the broad band Echelle spectrograph (Andor Mechelle ME5000-DH734-18U-03PS150) of

the LIBS system. It is optimized to ensure that all the wavelengths in the range 200-950nm are collected evenly into fiber coupled with the entrance slit of the spectrograph. The detector kept in proper synchronization with the laser using delay generator to get time resolved information of plasma evolution.

FIGURE 1.LIBS experimental setup for multielemental analysis

RESULTS AND DISCUSSIONS

A comparison of the spectral resolution of the present LIBS setup with Czerny Turner based system was done by recording the Mercury spectrum. The estimated FWHM’s of the 546.1nm line in the spectra recorded by both systems are found to be 0.1nm and 1nm respectively (Figure 2). It is evident from the figure that the spectral resolution of Echelle is 10 times higher than that of Czerny Turner. A LIBS spectrum of

Solid State Physics, Proceedings of the 55th DAE Solid State Physics Symposium 2010AIP Conf. Proc. 1349, 475-476 (2011); doi: 10.1063/1.3605940

© 2011 American Institute of Physics 978-0-7354-0905-7/$30.00

475Downloaded 22 Aug 2011 to 117.211.113.66. Redistribution subject to AIP license or copyright; see http://proceedings.aip.org/about/rights_permissions

sample is constituted by several elements and spectral lines of different elements can overlap. Since the Echelle system has got very high resolution as shown in Figure 2 (a), it can very well separate out individual lines from different elements.

FIGURE 2. Spectral resolution of (a) Echelle spectrograph and (b) Czerny Turner spectrograph

Study of temporal distribution of plasma is of great interest as it helps to discriminate the regions i.e. time window where a signal of our interest is predominant. Hence, with an ultimate aim to study environmental samples we have recorded time resolved LIBS soil spectra (Figure 3) and studied its characteristics.

FIGURE 3. Gated soil spectrum of LIBS plasma as a function of time after plasma initiation

In order to find the limit of detection (LOD) of LIBS technique for soil analysis of copper, we have recorded the soil spectrum with different known concentrations of copper. Spectra generated for 400ppm, 200ppm, 100ppm and 60ppm of copper in soil and a calibration graph is plotted by taking the corresponding intensities for 521.82nm line of copper as shown in Figure 4. LOD was calculated from this plot (3σ/slope, where σ is the standard deviation at lowest concentration) and it was found to be 13ppm.

FIGURE 4. Calibration curve for copper in soil to calculate the LOD

ACKNOWLEDGMENTS

This work was carried out under the project titled “Trace Element Analysis for Environmental and Biomedical Applications – Development of Laser Induced Breakdown Spectroscopy (LIBS) Technique.” Project No. 2007/34/14-BRNS/87, Board of Research in Nuclear Sciences (BRNS), Department of Atomic Energy (DAE), Govt. of India.

REFERENCES

1. P.Yaroshchyk, D.Body, R.J.S.Morrison, and B.L.Chadwick, Spec. Acta Part:B, 61, 200-209 (2006).

2. A.K.Misra, P.G.Lucey, R.C.Wiens, and S.M.Clegg,, Lunar and Planetary Science XXXVIII (2007).

3. David A. Cremers and Leon J. Radziemski, Handbook of Laser-Induced Breakdown Spectroscopy, 134 (2006).

4. N.M.Shaikh, B.Rashid, S.Hafeez, Y.Jamil, M.A.Baig, Journ. of Physics D: Appl. Physics, 39, 1384-91 (2006).

5. M.A.Hafez, M.A.Khedr, F.F.Elaksher, Y.E.Gamal, Plasma Sources Science and Technology, 12, 185-98 (2003).

6. R.S. Harmon, F.C. De Lucia Jr, A.W. Miziolek, K.L. McNesby, R.A. Walters, and P.D. French,Geochemistry-Exploration Environment Analysis, 5(1), 21-28 (2005).

476Downloaded 22 Aug 2011 to 117.211.113.66. Redistribution subject to AIP license or copyright; see http://proceedings.aip.org/about/rights_permissions