study laser-inducedbreakdown spectroscopy from micro ...€¦ · ieeesensors2006, exco, daegu,...

IEEE SENSORS 2006, EXCO, Daegu, Korea / October 22-25, 2006

Study of laser-induced breakdown spectroscopy from

micro-droplet ofNaCl Solution

Satoshi Ikezawa, Muneaki Wakamatsu and Toshitsugu UedaThe Graduate School of IPS, Waseda University

2-7 Hibikino, Wakamatsu-ku, City of Kitakyushu, Fukuoka-ken/Japan 808-0135Email: ikezawagfuji.wasedajp

Abstract- The laser-induced breakdown spectroscopy (LIBS)using micro-droplet sodium chloride solution is presented.Micronizing the sample made possible that whole volume of asample was confined in the laser beam spot area and wasseparated from its surrounding condition. If the sample'sphysical state was liquid, the density of solution could becontrolled as needed. Originally designed ink-jet system forsampling procedure was presented. According to the newmethod, improved calibration curves for the LIBS quantitativemeasurement were obtained.

I. INTRODUCTION

In the laser-induced breakdown spectroscopy (LIBS)technique, a high energy laser pulse is focused on a sampleto create plasma. Emission from the atoms and ions in theplasma is collected using lenses, guided to a spectrograph, astreak camera or some other gated detector and analyzed by acomputer.

LIBS is investigated extensively to establish properchemical analysis of specimens. LIBS is a useful method fordetermining the elemental composition of various materialsregardless of their physical state, (solid, liquid and gas)without any preprocessing. While qualitative analysis resultscan be obtained with such spectrum using a wavelengthcalibration reference, quantitative analysis can be obtainedon the basis of the time integrated amount of intensities fromthe emission.

In the quantitative analysis, optimizing experimental keyparameters of proper spark alignment, time gate delay, andwavelength gate width allowed determining experimentallythe limit of detection (LOD). Qualitative results can beobtained with such spectrum using a wavelength calibrationreference. There are several methods for LIBS analysis notonly to obtain spectral wavelength but also the amount ofintensities from the emission. The intensities can becompared with standard atomic line references. It is calledoptical emission spectroscopy (OES) analysis. Anothermethod is chemometric analysis based on comparing tosamples of known composition. Those methods are intensity

calibration references. Calibration curves can be obtainedfrom this intensity calibration method.

We were researching and developing the LIBS techniqueconcerning the quantitative analysis that could be used foreach LIBS system for various fields. To establish a practicalproper calibration method for quantitative analysis in variousstates, namely gas, liquid, and solid, we proposed the LIBStechnique using a micro-droplet ejection system. The micro-droplet ejection technique of the piezo type ink-jet systemwas used to delimit the accuracy of solution's sample.

First measurements were performed within a bulk liquidcell. Then liquid was ejected by ink-jet system. In this paper,an originally designed LIBS set-up and experimental data ofthe sodium chloride solution ejected by an ink-jet headpresented here contributes to improvement of calibrationcurves quality.

II. EXPERIMENTAL SET UP

A. LIBS system using a bulk liquid cellThe LIBS system using a bulk liquid cell is shown

schematically in Fig. 1. Nd:YAG laser (Big Sky Laser modelultra) was controlled by a delay pulse generator (StanfordResearch System, Inc., Model DG-535).

This laser operated at the 1064nm fundamentalwavelength used to generate 5OmJ/pulse of 8ns width(FWVHM). When used for the ablation, the laser pulse wasfocused with a 50mm focal length lens, yielding a powerdensity of a 1012 W/cm2. Emission from the laser-producedplasma was collected using lenses and guided to aspectrograph (CHROMEX 250is). Then the emitted lightwas dispersed by a 1200-lines/mm diffraction grating and anelectrical signal was recorded using a streak camera(Hamamatsu Photonics) with 10-ps time resolution or higher.Finally the signal was processed and stored in a computer.

1-4244-0376-6/06/$20.00 }2006 IEEE 291Authorized licensed use limited to: Donghua University. Downloaded on June 12,2020 at 05:35:54 UTC from IEEE Xplore. Restrictions apply.

IEEE SENSORS 2006, EXCO, Daegu, Korea / October 22-25, 2006

B. LIBS system using a micro-droplet ejection systemThe LIBS system using a micro-droplet ejection system

is shown in Fig. 3. The micro-droplet was released usingoriginally designed set up with ink-jet head. The samplesolution was injected into the piezo head from the syringethrough the filter. Electric signal of certain shape wasgenerated by a wavy synthesizer (NF Corporation, modelWF1945B), and it was transmitted through the piezo driveunit. The ink-jet head was covered with the quartz beaker.Function of the wind breaking, the liquid recovering, and theadditional beam condensing had been provided.

breakdown cell

Fig. 1. Schematic diagram ofLIBS experimentusing bulk liquid cell

In Fig. 2, the comparison of emission focused (a) insideof the sodium chloride solution and (b) on the solutionsurface is shown. We compared the LIBS data obtained fromthose measurements.

Fig. 3. Schematic diagram ofLIBS experimentusing micro-droplet ejection system

Fig. 2. Comparison ofbreakdown emission made (a)in the solution and (b) at the solution surface

The surface emission had greater intensity than that ofinner emission. The life time of plasma in a bulk liquid wasshorter than that generated at the liquid surface because mostof the energy produced by laser was utilized to heat thesample. Also, the laser beam was affected by refractionthrough the sample liquid and then plasma region wasnarrowed.

We used the method (b): focusing on the sample surfaceand recorded the emission intensity of sodium atomic speciesat the spectral region near 589nm repeatedly.

In Fig. 4, a piezo head that called "shear mode type" usedin this experiment is shown. Some channels and actuatorswere formed of piezoelectric zirconate titanate (PZT)substrate of a piezo-electric element. Wiring that added thevoltage was bonded to the actuator respectively. The actuatorwas bonded to the upper surface of the channel with thecover plate and fixed. The nozzle plate was bonded in frontof the channel.

solution

Fig. 4. Structure ofshear modepiezometric head

1-4244-0376-6/06/$20.00 }2006 IEEE 292Authorized licensed use limited to: Donghua University. Downloaded on June 12,2020 at 05:35:54 UTC from IEEE Xplore. Restrictions apply.

IEEE SENSORS 2006, EXCO, Daegu, Korea / October 22-25, 2006

Fig. 5 shows the micro-photograph of the nozzle platesurface. The nozzle diameter and the channel pitch were52.4ptm and 0.46mm, respectively. Adding the electric fieldin an orthogonal direction where the piezoelectric elementwas polarized leaded to the actuator winding transformation,and to pressurizing of solution in the channel. The pressuregenerated in the channel spread, the pressure wave reflectedbetween the nozzle and the solution supply room, and thedamped oscillation was generated while resonating. That wasa mechanism to eject micro-droplet by receiving the pressureofresonance with time.

III. RESULTS AND DISCUSSION

We obtained the results about the breakdown emissionintensities from 0.1M to 1.OM density of sodium chloridesolution with a bulk liquid cell. Fig. 6 shows the error rangesof this result. Analysis based on data taken from 100 shotslaser pulse, the central gate point was 18.1 ts from trigger-inand the gate width was 3.9[ts. This analytical condition wasoptimized at the sample of 0. IM sodium chloride solution. Itindicates that when one draws a calibration curve, there are afew data points from the optimum data point. There weresome problems associated with sample solution splashingwhen measurements were made at the solution surface. Forcalibration purpose, the fluctuation in droplet sizedistribution at the time of laser spark created problems.

DI Peak Calibration Curve

400

; ttO.1 0.2 0.3 0.4 0.5 0.6 0.7

NaCI solution Emol/l]

I4

Fig. 5. Photo ofa nozzle plate: Nozzle diameter 52.4 ,m

At first, the channel room capacity was shrunk by thefirst pulse created by a piezo-electric element voltage withrising edge and the solution was pushed out. Next, thechannel room capacity returned with falling edge of thepulse, and the meniscus on the surface of the solution at thenozzle was drawn in.

In the pull-push drive, meniscus was ejected from thenozzle when the velocity of the droplet became themaximum, and decelerated monotonously until the ejectionof the liquid reversed. Therefore, the ejection velocity ofdroplet at the front edge was faster than the rear end, thenejected in the shape of spindle. The micro-droplet volumewas calculated using the equation (1).

(1)V - _r2vd 2f

Vd - volume of micro-droplet,

r - nozzle radius,

v - micro-droplet ejection velocity,

f- resonance frequency,

Fig. 6. The error ranges ofeach concentrationfrom 0. JM to 1. OM at 100 laser shots

On the other hand, in the case of LIBS using a micro-droplet ejection system, we obtained different data. Thecomparison between the experimental data of micro-dropletejection system and the one of previous bulk-liquid type isshown as follows.

bulk-liquid 0.1 mol/

6550

6050

5550

E 5050

4550

4050587 587.5 588 588.5 589 589.5 590

Wavelength Enm]590.5 591

Fig. 7. Intensity ofNa 589 nm emission lines, bulk-liquid type, 0. JM density ofsolution

1-4244-0376-6/06/$20.00 }2006 IEEE

300

250

200

150

100

50

0.8 0.9 1.0

293Authorized licensed use limited to: Donghua University. Downloaded on June 12,2020 at 05:35:54 UTC from IEEE Xplore. Restrictions apply.

IEEE SENSORS 2006, EXCO, Daegu, Korea / October 22-25, 2006

ink-jet 0.1 mol/I

6550

6050

5550

E 5050

4550

4050 - - 1-587 587.5 588 588.5 589 589.5 590

Wavelength Enm]590.5 591

Fig. 8. Intensity ofNa 589 nm emission lines,micro-droplet type, 0. JM density ofsolution

bulk-liquid 0.7mol/I

6550

-6050

> 5550

E 5050

4550

4050 7587 587.5 588 588.5 589 589.5 590

Wavelength Enm]

The experimental data for 0. IM and 0.7M density sampleas an example of comparing the experiments on bulk-liquidtype and micro-droplet type are shown in Figs. 7 - 10. In Fig.7 and Fig. 8, data show that ink-jet micro-droplet's intensitywas a little higher than in the case of bulk-liquid type. On theother hand, it was clearly shown that huge difference ofintensity exists between the solution of 0.7M density in Fig.9 and Fig. 10. It was proved that in fixed quantitymeasurement of LIBS, calibration curve calculated withmicro-droplet type cannot be used with the bulk-liquidquantizing method.

IV. CONCLUSIONS

0.7 M density solution showed low intensity in the caseof sample forn of bulk-liquid type. It turned out that adifferent data were obtained and different calibration curveswould be obtained even with the same solution, if differenttypes of LIBS experimental set up were used. It could becaused by the increased emission within the solution ofdecentralized and unifornly distributed NaCl particles. Thiswas the first time, to our knowledge, that micro-dropletejection system in LIBS experiment was employed.

590.5 591

Fig. 9. Intensity ofNa 589 nm emission lines, bulk-liquidt type, 0. 7M density ofsolution

ink-jet 0.7mol/I

6550

- 6050

5550n

E 5050

4550

4050587 587.5 588 588.5 589 589.5 590

Wavelength Enm]590.5 591

REFERENCES

[1] Satoshi Ikezawa, Muneaki Wakamatsu, and Toshitsugu Ueda, "BasicResearch of LIBS Calibration Method using NaCl Solution,", ThePapers of Technical Meeting on Physical Sensor, IEE Japan, pp.9-13 ,2006.

[2] Muneaki Wakamatsu, Toshitusugu Ueda, Hisanori Hayashi, "ParticleElements and Size Measurement Using LIBS, ", InternationalConference on Electrical Engineering 2004 Vol.3-2, p823-826.

[3] Yoshio Takeuchi and Shinichi Nishi, "Features of Shear Mode PiezoInkjet Head,", Journal of the Imaging Society of Japan Vol.43, No.6,pp.509-514 ,2004.

Fig. 10. Intensity ofNa 589 nm emission lines,micro-droplet type, 0. 7M density ofsolution

1-4244-0376-6/06/$20.00 }2006 IEEE

v

294Authorized licensed use limited to: Donghua University. Downloaded on June 12,2020 at 05:35:54 UTC from IEEE Xplore. Restrictions apply.