Determination of thorium and uranium in solutionby laser-induced breakdown spectrometry

Arnab Sarkar,* Devanathan Alamelu, and Suresh Kumar AggarwalFuel Chemistry Division, Bhabha Atomic Research Centre, Trombay, Mumbai-400 085, India

*Corresponding author: asarkar@ymail.com

Received 3 April 2008; revised 17 July 2008; accepted 18 July 2008;posted 18 July 2008 (Doc. ID 94648); published 4 September 2008

Laser-induced breakdown spectrometry (LIBS) has been successfully developed to quantify thorium (Th)and uranium (U) in solution using membrane-based filter paper as a sample support. The filter paper iseasy to handle and the entire analysis process takes only a few minutes for each sample. The developedmethod provides a practical approach for fast and sensitive determination of actinides in aqueous solu-tions using LIBS. We obtained calibration curves for U and Th individually as well as as a mixture. Weobserved that the quantification of trace levels of U in a bulk amount of Th was not possible because ofstrong spectral interference from Th at the most intense lines of U, but traces of Th in a bulk amount of Ucould be determined. The concentrations of U and Th in unknown solutions were determined by use ofLIBS, and these results agreed to within 4% and 2%, respectively, with the expected values. The limits ofdetection for Th and U were calculated from the experimental data and were in the range of a few partsper million by weight to a few tens of parts per million by weight. © 2008 Optical Society of America

OCIS codes: 120.6200, 140.3440, 300.2140, 300.6365.

1. Introduction

Laser-induced breakdown spectrometry (LIBS) is apromising and attractive technique for applicationto nuclear materials as the technique is nondestruc-tive and involves a minimum number of steps of sam-ple preparation. The latter feature helps to reduce theradiation exposure of the analysts involved in theanalysis of nuclear materials. Also, since the techni-que can be applied directly to analysis of solid sam-ples, it eliminates the problems associated with therecovery of actinide elements from solutions. LIBShasbeenused extensively for awide range of scientificand industrial applications as a spectrochemical tool;its principles have been well-described elsewhere[1–6]. Many reports of LIBS describe its wide applic-ability for qualitative detection and quantitative de-termination of trace elements in gas, solid, and liquidsamples [7–14]. LIBS ismostly applied for analysis ofsolid samples that provide high sensitivity and repro-ducibility for most of the elements in the periodic

table. The application of LIBS for liquid sampleshas some inherent drawbacks such as splashing, sur-face ripples, quenching of emitted intensity, and ashorter plasma lifetime [15–17]. Different methodol-ogies have been developed for analysis of liquids.These include (a) freezing samples [18], (b) laminarflows and liquid jets [19,20], (c) the double-pulse tech-nique [15], (d) the liquid-to-solid conversion technique[21] on droplets [22,23], and (e) solid surface analysis[24,25]. LIBS has been used for the determination ofseveral elements (Mg, Ca, Cr,Ni)[16] and some chlori-nated hydrocarbons (CHCs)[26] in aqueous solutions.LIBS has also been applied for bulk analysis of aqu-eous solutions that contain alkali metals [15,27], al-kaline earth metals [15,27,28], and other metals[15]. Analytical experiments ondroplets havebeen re-ported for the determination of Li, Na, Mg, Ca, Mn,and Al [22,23] as well as for K [29,30], Ba [29], Rb,Li, and Cd[31] in a liquid jet. Species in organic sol-vents have also been determined [15].

The solid surface analysis technique exploits thestrengths of LIBS without the problems inherentto liquid analysis [24,25]. We describe an experimen-tal setup that can be used to analyze thorium (Th)

0003-6935/08/310G58-07$15.00/0© 2008 Optical Society of America

G58 APPLIED OPTICS / Vol. 47, No. 31 / 1 November 2008

and uranium (U) separately and also in the presenceof each other in aqueous solutions. While maintain-ing the main advantages of the LIBS, i.e., fewer sam-ple preparation steps and less in situ measurement,our current research demonstrates the capability ofLIBS for determination of Th and U.

2. Experimental Setup

A. Apparatus

We used the Spectrolaser 1000M from Laser Analy-sis Technologies (now known as XRF scientific), Vic-toria, Australia. The instrument is an integratedanalysis system comprising an excitation laser, opti-cal spectrograph, and charge-coupled-device (CCD)array camera and is fully software controlled. TheSpectrolaser 1000M instrument is equipped with ahigh-power Q-switched Nd : YAG laser that yields200mJ of pulse energy at the fundamental IR wave-length (1064nm) with a 7ns pulse width, at a repeti-tion rate of 10Hz. The laser is focused onto thesample by a plano–convex lens with a focal lengthof 5 cm (Fig. 1). The sample holder is fixed to a rotaryholder so that every laser shot hits on a fresh site.The emitted plasma radiation is then collected ata 45° angle with respect to the laser beam direction.The emission output is focused through four fiber op-tic cables onto the entrance slit of spectrographs(Czerny–Turner configuration) with CCDs as detec-tors. The system has a wide spectral coverage from180 to 850nm while maintaining the resolution of0:6nm at 300nm that is required for observationof elemental emissions with minimal interferences.The output from each of the four CCDs is digitizedby analog-to-digital converter (ADC) circuitry. Thespectral output over the entire spectral range is thusrecorded simultaneously for each laser pulse. Thespectral acquisition can be delayed by a software con-trol delay generator to achieve the required delay be-tween plasma ignition and recording of the emissionspectrum. For this study we averaged the data of 50laser shots for each replicate measurement. In thisaveraging, we excluded the spectra that had a great-

er than 10% variation from the mean value. Threesuch independent replicate analyses were carriedout, and the mean of these replications was usedin the construction of the calibration curve. The errorbars for all mean calibration values are �1σ (stan-dard deviation) of the three independent analyses.We then analyzed the spectral data by using non-linear and linear equations.

B. Sample Preparation and Recording Spectra

Stock solutions of Th and U were prepared by dissol-ving known amounts of U3O8 and ThðNO3Þ4. 5H2O insuprapure nitric acid (Merck). Concentration of U inthe stock solution was determined using biampero-metry, employing Ti(III) as reductant. For Th stocksolution, we used complexometric titration withethylenediaminetetraacetic acid (EDTA) with xyle-nol orange as the indicator. Suitable dilutions weremade from these stock solutions to prepare sampleswith Th and U concentrations in the 0–5000 ppmwrange to established calibration plots. Synthetic mix-tures of Th and U were prepared by mixing indivi-dual solutions of the two elements on the basis ofweight All the mixtures were made by keeping theamount of U nearly the same (�2%), and the Thamount was varied so that the Th to U ratio rangedfrom 0.005 to 0.5.

A volumetric pipette (Eppendorf, Germany) wasused to transfer 40 μl of solution dropwise to theWhatman-42 membrane filter paper. The solutionspread circularly over the filter paper with a radiusof approximately 1 cm. The solution was evaporatedto dryness with the help of a hot air blower. Subse-quently, the paper was allowed to cool to room tem-perature and was then mounted in the sampleholder. The spatial extent of the sample was markedon the filter paper so that the analysis could be per-formed over a region of 1 cm diameter where both Uand Th are expected to be uniformly distributed.

All the experiments were carried out under iden-tical conditions of laser fluence, repetition frequency,and acquisition delay. Triplicate analyses were per-formed on all the samples including the calibrationsamples as well as the unknowns. The emission linesfor Th and U determination were selected from theregions in which minimal spectral interferences wereobserved and their intensities were ratioed with re-spect to C (I) 193:029nm. Among the two most in-tense C(I) lines, 193:029nm and 247:856nm, thelatter was relatively intense but the emission peakprofile was much cleaner for the former and hencewas the one used. Fe(I) at 249:064nm, which is pre-sent in the filter paper, contributes to the backgroundat the C(I) 247:856nm line as shown in Fig. 2.

3. Results and Discussion

A. Selection of Emission Line for Analysis

The spectral region (180–850nm) used to record theemission lines contained most of the strong emissionlines of both U and Th. The spectra obtained using

Fig. 1. Schematic of a laser-induced breakdown spectrometrysystem.

1 November 2008 / Vol. 47, No. 31 / APPLIED OPTICS G59

pure Th and pure U solutions under identical condi-tions of analyses (laser fluence and acquisition delay)are shown in Fig. 3. It is obvious that there are anumber of emission lines that can be employed fordetermination. The selection of emission lines forqualitative and quantitative determination of U andTh was based on the following criteria:

a. The spectral lines that are usually employedin other emission spectroscopic experiments suchas inductively coupled plasma (ICP) atomic emissionspectroscopy (AES).

b. No strong spectrally interfering emission fromthe other actinide within �0:3nm, so as to allow de-termination of U and Th in a mixture.

c. An emission line with Elower ¼ 0 or as low aspossible was preferred. Although an emission linewith its ground state near to or equal to zero hasthe possibility of self-absorption, it has the highestsensitivity, which is especially useful when the con-centration of the species of interest is low.

d. The emission line should have a high transi-tion coefficient (A).

e. A high signal-to-noise ratio (SNR) should existunder the conditions of analysis by LIBS.

By careful comparison of the emission spectra, twolines for U and three lines for Th were selected. Thespectral lines considered for analysis and their avail-able spectroscopic data are listed in Table 1.

B. Effect of Substrate

Our approach consisted of evaporation of the analytesolution upon a commercially available filter paper(Whatman-42) substrate followed by LIBS analysisof the substrate surface. The data reported for theconcentrations of different trace elements in theWhatman-42 membrane filter paper are listed inTable 2. Since both Th and U are not present inthe filter paper and no observable spectral linesfor these two elements were observed during theLIBS analysis of blank filter paper, this filter paperwas selected as the substrate to minimize the blankcontribution. It should also be mentioned that caremust be taken to dry the liquid completely on the fil-ter paper to obtain reliable results. A small amountof moisture, if present, drastically reduces the emis-sion intensity of the lines of interest and leads toerratic results. At fluence corresponding to laser en-ergies higher than 100mJ, slight charring of the fil-ter paper was observed and hence the laser energywas fixed at around 100mJ for the current work.

C. Time Resolution

As reported in the application of LIBS on solids[3,32,33], time-resolved spectroscopy is essential forimproving the sensitivity in any LIBS experiment. Ingeneral, the optimal acquisition delay time dependson the laser fluence, the kind of sample being

Fig. 2. Carbon emission lines from Whatman-42 membranebased filter paper.

Fig. 3. Emission spectra of U and Th obtained by LIBS underidentical conditions of analysis.

Table 1. Characteristics of Spectral lines Used for LIBS Analysis ofUranium and Thorium [37]a

Elementλij

ðnmÞAij

ð108 s−1ÞEj

ðcm−1ÞEj

ðtermÞEi

ðcm−1ÞEi

ðtermÞU(II) 367.007 0.26 914.765 6K° 28154.45 6LU(II) 409.013 0.19 1749.123 6L° 26191.309 6MTh(II) 283.729 - 6213.490 4F 41447.959 4G°

Th(II) 318.019 0.88 1521.896 4F 32957.429 4G°

Th(II) 401.913 0.41 0 - 24873.983 -aλij is the transition wavelength, Aij is the transition probability,

Ei and Ej are the energies of the upper and lower level, respec-tively, and gi and gj are statistical weights of upper and lower level,respectively.

G60 APPLIED OPTICS / Vol. 47, No. 31 / 1 November 2008

studied, and the surrounding atmosphere. On theother hand, its dependence on different species inthe sample is not significant at high fluence and, gen-erally, a unique acquisition delay time can be chosenfor all the elements [34]. In the laser-induced plasma,at time intervals in the range of a few 100 ns, thespectral emission lines are significantly broadenedbecause of the high electronic density and are alsosuperimposed on the intense continuum, which de-creases with time. At a later stage, as the plasma ex-pands and cools, lines from the excited elementalspecies dominate the plasma emission. An adequateselection of acquisition delay time involves a compro-mise between signal-to-background ratio and line in-tensities. In this study, an acquisition delay of 600nsand a gate width of 10 μs were selected to record spec-tral lines corresponding to 100mJ of laser energy toobtain the optimum SNR.

D. Reproducibility

The reproducibility was evaluated by observing theintensity of the Th(II) 401:913nm emission line bycarrying out replicate analysis from the same filterpaper as well as by using a number of filter papers.Replicate analysis by using the same filter paperyields a relative standard deviation (RSD) of approxi-mately 20%. The signal intensity was also found todecrease with successive analyses. This was attribu-ted to the loss of a significant amount of absorbedanalyte or could be due to damage of the filter paper.A RSD of approximately 2% was obtained for the lineintensity for ten analyses using ten independentfilter papers, which indicated good reliability ofthe LIBS analysis. Figure 4 shows the results ofthese measurements. Reproducibility for the U(II)367:007nm emission line was also determined andwas found to be nearly the same as that for the Th(II) 401:913nm line.

4. Calibration Curves

A. Individual Element Calibration and Limit of Detection

To obtain the calibration curve for Th and U, we usedeight standard solutions of known concentrations.The experimental data between the concentrationsof Th or U and the normalized intensities of the spec-tral lines were fitted to polynomial equations of de-grees one and two, and the results are listed in

Table 3. Figure 5 shows the calibration curves forTh and U. Saturation effects were observed for Thand U above 1000ppmw and 800ppmw, respectively.The calibration curves deviated from linearity due toself-absorption. This effect was much more pro-nounced for Th and hence the data correspondingto the higher concentration of Th are not shown inFig. 5. For higher concentrations of U, a polynomialequation of degree two was used to fit the data. In thecase of Th, even in the low concentration range of 0to 1000 ppmw, we obtained two distinct concentra-tion regions, namely, 0 to 180ppmw and 180 to1000 ppmw with different calibration curves.

The limit of detection (LOD) for both Th and U wasdetermined according to the definition 3σB=s, whereσB is the standard deviation of the blank and s is thesensitivity determined by the slope of the calibrationcurve. We determined σB from ten measurements ofblank signals under the same experimental condi-tions, where the sample was a filter paper with onlydeionized water. Table 4 lists the LODs determinedfrom different spectral lines of Th and U. To the bestof our knowledge, this is the first report of Th quan-tification in a solution by LIBS; the LOD for U ismuch lower than the previously reported LOD of0:1 g=L [35]. Th quantification in dry aerosol by LIBShas been reported with a LOD of 0:45mg=ml, i.e.,450ppmw [36]. To verify the calibration of our inves-tigations, two standard solutions of U and Th wereprepared and analyzed. In the case of U, the367:007nm and 409:013nm lines yielded accuracyvalues of 2.2% and 4.8%, respectively. The relativeprecision in both cases was less than 5%. Amongthe three spectral lines of Th, the 283:729nm and318:019nm lines gave an accuracy of approximately10% with more than 10% precision. The calibration

Table 2. Trace Elements in the Whatman-42 Membrane Filter Papera

Element ppmw Element ppmw Element ppmw

Al 2 Cr 0.3 N 12Sb <0:02 Cu 0.3 K 1.6As <0:02 F 0.2 Si <2Ba <1 Fe 6 Na 33B 1 Pb 0.2 S <5Br 1 Mg 1.8 Zn 0.6Ca 13 Mn 0.05Cl 80 Hg <0:005

aRef. [38].

Fig. 4. Reproducibility of the Th(II) 401:913nm emission linefrom ten measurements carried out using independent filter pa-pers under identical experimental conditions. The horizontalstraight line represents the mean value and the inset shows thepeak region used for integration.

1 November 2008 / Vol. 47, No. 31 / APPLIED OPTICS G61

curve corresponding to the 401:913nm emission linegave an accuracy of 4% with 7% precision.

B. Thorium–Uranium Mixture Calibration and Limit ofDetection

LIBS is always associated with and affected by a ma-trix effect to a high degree. Therefore, the calibrationcurves obtained by use of pure elements cannot beused for a solution that contains a mixture of twoor more elements. In the case of U and Th, this pro-blem is further aggravated by the complex emissionspectra of both. From pure Th spectra [Fig. 3(b)], one

can observe a band of spectral lines in the 405nm to425nm regions, making the U(II) 409:013nm lineunsuitable for any analytical purposes. The U(II)367:007nm emission line also had significant spec-tral interference from the Th(I) 366:998nm line. Thismakes it difficult to quantify trace amounts of U in abulk amount of Th. However, determination of traceamounts of Th in the presence of a bulk amount of Uis possible, since the Th(II) 401:913nm emission lineis free from any spectral interference from U emis-sion lines. The other two lines of Th, namely, Th(II) 283:729nm and Th(II) 318:019nm, are affectedeither completely or partially by spectral interfer-ence from U emission lines. The calibration curve ob-tained by plotting the intensity ratio of Th(II)401:913nm with respect to U(II) 367:007nm versusthe Th/U amount ratio is shown in Fig. 6. The re-sponse of the Th/U signal was found to be linearup to the Th/U ratio of 0.035, beyond which the sa-turation effect was observed.

To verify the calibration, two fuel pellets of Thand U were used as unknowns. The amount of the

Table 3. Fitted Parameters Obtained from Calibration Curves for Uranium and Thorium Using Polynomial Equation y ¼ aþ bxþ cx2

Element Analytical line (nm) Concentration range (ppmw) a b c R a

U 367.007 0–5000 0.0078 5:59E–05 −4:71E–09 0.999367.007 0–800 0.0081 5:30E–05 0 0.999409.013 0–5000 0.0097 4:56E–05 −3:56E–09 0.999409.013 0–800 0.0106 4:12E–05 0 0.997

Th 283.729 0–180 0.0124 4:56E–04 0 0.995283.729 180–1000 0.0673 1:59E–04 0 0.998318.019 0–180 0.0166 4:03E–04 0 0.931318.019 180–1000 0.0700 1:92E–04 0 0.999401.913 0–180 0.0315 1:32E–03 0 0.997401.913 180–1000 0.1840 4:92E–04 0 0.998

Th (In the case ofmixture calibration)

401.913 0–0:035b 0.6475 15.122 0 0.999401.913 0–0:5b 0.6901 11.281 0 0.999

aR denotes regression coefficient.bDenotes Th/U amount ratio.

Fig. 5. Calibration curve of (a) U and (b) Th in aqueous solutionfitted with a polynomial equation. Error bars represent the stan-dard deviation frommultiple measurements at each concentrationvalue.

Fig. 6. Calibration curve of the U-Th mixture in aqueous solu-tion. Error bars show standard deviation from multiple measure-ments at each value.

G62 APPLIED OPTICS / Vol. 47, No. 31 / 1 November 2008

Th/U ratio in these pellets was determined by anelectrochemical process for U and by complexometrictitration for Th after quantitative dissolution of thepellets. The results obtained from triplicate analysesby LIBS from the solutions agreed to within 3% withthe expected amount of the Th/U ratios with arelative standard deviation of approximately 3%(Table 5). The LOD was determined to be 0.0007for the Th/U ratio.

5. Conclusions

LIBS has been developed for the determination of Thand U in aqueous solutions. From the above observa-tions we can conclude that LIBS is useful for the de-termination of trace amounts of Th in the presence ofa bulk amount of U, while the same is not possible fora trace amount of U in the presence of a bulk amountof Th because of the significant interference from Themission lines. Thus LIBS is a promising techniquefor the determination of trace amounts of Th in a Umatrix without separation of the matrix during theThorex process. This feature of LIBS is highly usefulsince the radioactive analytical waste generated byother analytical methods can be reduced signifi-cantly. LIBS can also be employed for the determina-tion of Th and U in mixed oxide pellets of Th and U,but it might be necessary to dissolve the pelletsquantitatively before analysis. LIBS also providesan independent approach based on different physico-chemical principles for the determination of differenttrace constituents in a Th and U matrix. This wouldbe useful for the development of trace element certi-fied reference materials for these matrices of interestin a Th-based fuel cycle.

The authors are grateful to V. Venugopal, Director,Radiochemistry and Isotope Group, Bhabha AtomicResearch Centre, for his constant support and inter-est in LIBS research.

References

1. D. A. Rusak, B. C. Castle, B. W. Smith, and J. D. Winefordner,“Fundamentals and applications of laser-induced breakdownspectroscopy,” Crit. Rev. Anal. Chem. 27, 257–290 (1997).

2. I. Ahmed and B. J. Goddard, “An overview of laser inducedbreakdown spectroscopy,” J. Fiz. Malays. 14, 43–54 (1993).

3. L. J. Radziemski, “Review of selected analytical applicationsof laser plasmas and laser ablation,” Michrochem. J. 50, 218–234 (1994).

4. K. Song, Y-I. Lee and J. Sneddon, “Applications of laser in-duced breakdown spectrometry,” Appl. Spectrosc. Rev. 32,183–235 (1997).

5. L. J. Radziemski, “From laser to LIBS, the path of technologydevelopment,” Spectrochim. Acta Part B 57, 1109–1113(2002).

6. E. Tognoni, V. Palleschi, M. Corsi, and G. Cristoforetti, “Quan-titative microanalysis by laser-induced breakdown spectro-scopy: a review of the experimental approaches,”Spectrochim. Acta Part B 57, 1115–1130 (2002).

7. M. Hanafi, M. M. Omar, and Y. E. E-D. Gamal, “Study of laser-induced breakdown spectroscopy of gases,” Radiat. Phys.Chem. 57, 11–20 (2000).

8. C. D'Angelo, J. M. Gomba, D. Iriarte, and G. Bertuccelli,“Trace element analysis in water by LIBS technique,” III Iber-oamerican Conference on Optics 98, 534–541 (SPIE, 1999).

9. B. Charfi and M. A. Harith, “Panoramic laser-induced break-down spectrometry of water,” Spectrochim. Acta Part B 57,1141–1153 (2002).

10. M. Sabsabi and P. Cielo, “Quantitative analysis of aluminumalloys by laser-induced breakdown spectroscopy and plasmacharacterization,” Appl. Spectrosc. 49, 499–507 (1995).

11. C. Aragón, J. A. Aguilera, and F. Peñalba, “Improvements inquantitative analysis of steel composition by laser-inducedbreakdown spectroscopy at atmospheric pressure using an in-frared Nd : YAG laser,”Appl. Spectrosc. 53, 1259–1267 (1999).

12. J. M. Gomba, C. D'Angelo, D. Bertuccelli, and G. Bertuccelli,“Spectroscopic characterization of laser induced breakdownin aluminum—lithiumalloy samples for quantitative determi-nation of traces,” Spectrochim. Acta Part B 56, 695–705 (2001).

13. M. Kuzuya and H. Aranami, “Analysis of a high-concentrationcopper in metal alloys by emission spectroscopy of a laser-produced plasma in air at atmospheric pressure,” Spectro-chim. Acta Part B 55, 1423–1430 (2000).

14. M. Kuzuya, M. Murakami, and N. Maruyama, “Quantitativeanalysis of ceramics by laser-induced breakdown spectro-scopy,” Spectrochim. Acta Part B 58, 957–965 (2003).

15. D. A. Cremers, L. J. Radziemski, and T. T. Loree, “Spectroche-mical analysis of liquids using the laser spark,” Appl. Spec-trosc. 38 721–729 (1984).

16. G. Arca, A. Ciucci, V. Palleschi, S. Rastelli, and E. Tognoni,“Trace element analysis in water by laser-induced breakdownspectroscopy technique,” Appl. Spectrosc. 51, 1102–1105(1997).

17. L. St-Onge, E. Kwong, M. Sabsabi, and E. B. Vadas, “Rapidanalysis of liquid formulations containing sodium chlorideusing laser-induced breakdown spectroscopy,” J. Pharm.Biomed. Anal. 36, 277–284 (2004).

18. J. O. Cáceres, J. Tornero López, H. H. Telle, and A. GonzálezUreña, “Quantitative analysis of trace metal ions in ice usinglaser-induced breakdown spectroscopy,” Spectrochim. ActaPart B 56, 831–838 (2001).

19. C. W. Ng, W. F. Ho, and N. H. Cheung, “Spectrochemical ana-lysis of liquids using laser-induced plasma emissions: effectsof laser wavelength on plasma properties,” Appl. Spectrosc.51, 976–983 (1997).

20. B. T. Fisher, H. A. Johnsen, S. G. Buckley, and D. W. Hahn,“Temporal gating for the optimization of laser-induced

Table 4. LODs for Uranium and Thorium

ElementAnalyticalLine (nm) LOD (ppmw)

Uranium (pure) 367.007 18.5409.013 24.6

Thorium (pure) 283.729 1.6318.019 2.25401.913 0.72

Thorium (in the presenceof bulk uranium)

401.913 0.0007a

aDenotes Th=U amount ratio.

Table 5. Comparison of Analytical Results of LIBSwith Expected Values

Sample No.

Th/U Amount Ratio

ðX=YÞLIBS Value(X) Expected Value (Y)

1 0:2394� 0:0016a 0:2373� 0:0003 1.0092 0:0839� 0:0021 0:0859� 0:0003 0.977

aDenotes standard deviation of three replicate measurements.

1 November 2008 / Vol. 47, No. 31 / APPLIED OPTICS G63

breakdown spectroscopy detection and analysis of toxic me-tals,” Appl. Spectrosc. 55, 1312–1319 (2001).

21. D. M. Díaz Pace, C. A. D'Angelo, D. Bertuccelli, andG. Bertuccelli, “Analysis of heavy metals in liquids using laserinduced breakdown spectroscopy by liquid-to-solid matrixconversion,” Spectrochim. Acta Part B 61, 929–933 (2006).

22. D. E. Poulain and D. R. Alexander, “Influences on concentra-tion measurements of liquid aerosols by laser-induced break-down spectroscopy,” Appl. Spectrosc. 49, 569–579 (1995).

23. H. A. Archontaki and S. R. Crouch, “Evaluation of an isolateddroplet sample introduction system for laser-induced break-down spectroscopy,” Appl. Spectrosc. 42, 741–746 (1988).

24. R. L. Vander Wal, T. M. Ticich, J. R. West, and P. A. House-holder, “Trace metal detection by laser-induced breakdownspectroscopy,” Appl. Spectrosc. 53, 1226–1236 (1999).

25. P. Yaroshchyk, R. J. S. Morrison, D. Body, and B. L. Chadwick,“Quantitative determination of wear metals in engine oilsusing LIBS: the use of paper substrates and a comparison be-tween single- and double-pulse LIBS,” Spectrochimica ActaPart B 60, 1482–1485 (2005).

26. L. M. Berman and P. J. Wolf, “Laser-induced breakdown spec-troscopy of liquids: aqueous solutions of nickel and chlorinatedhydrocarbons,” Appl. Spectrosc. 52, 438–443 (1998).

27. R. Knopp, F. J. Scherbaum, and J. I. Kim, “Laser inducedbreakdown spectroscopy (LIBS) as an analytical tool for thedetection of metal ions in aqueous solutions,” Fresenius J.Anal. Chem. 355, 16–20 (1996).

28. T. Kitamori, T. Matsui, M. Sakagami, and T. Sawada, “Laserbreakdown spectrochemical analysis of microparticles in li-quids,” Chem. Lett. 12, 2205–2208 (1989).

29. W. F. Ho, C. W. Ng, and N. H. Cheung, “Spectrochemicalanalysis of liquids using laser-induced plasma emissions:effects of laser wavelength,” Appl. Spectrosc. 51, 87–91(1997).

30. N. H. Cheung and E. S. Yeung, “Single-shot elemental analysisof liquids based on laser vaporization at fluences below break-down,” Appl. Spectrosc. 47 882–886 (1993).

31. N. H. Cheung and E. S. Yeung, “Distribution of sodium and po-tassiumwithin individual human erythrocytes by pulsed-laservaporization in a sheath flow,”Anal. Chem.66, 929–936 (1994).

32. V. Majidi and M. R. Joseph, “Spectroscopic applications oflaser-induced plasmas,” Crit. Rev. Anal. Chem. 23, 143–162(1992).

33. P. Fichet, P. Mauchien, and C. Moulin, “Determination of im-purities in uranium and plutonium dioxides by laser-inducedbreakdown spectroscopy,” Appl. Spectrosc. 53, 1111–1117(1999).

34. R. Wisbrun, I. Schechter, R. Niessner, H. Schroder, andK. L. Kompa, “Detector for trace elemental analysis of solidenvironmental samples by laser plasma spectroscopy,” Anal.Chem. 66, 2964–2975 (1994).

35. J. R. Wachter and D. A. Cremers, “Determination of uraniumin solution using laser-induced breakdown spectroscopy,”Appl. Spectrosc. 41, 1042–1048 (1987).

36. http://www.arl.army.mil/www/default.cfm?Action=247&page=250

37. NIST, Basic atomic spectroscopic data, http://physics.nist.gov/PhysRefData/Handbook/periodictable.htm

38. Cellulose Filters, Whatman, www.whatman.com/CelluloseFilters.aspx

G64 APPLIED OPTICS / Vol. 47, No. 31 / 1 November 2008