Laser-Induced Breakdown Spectroscopy for Microanalysis

Transcription

1 Laser-Induced Breakdown Spectroscopy for Microanalysis Robert Fedosejevs, Y. Godwal, M.T. Taschuk, S. L. Lui, Y.Y. Tsui Department of Electrical and Computer Engineering University of Alberta, Edmonton, Alberta Presented at the 3rd INTERNATIONAL CONFERENCE ON THE FRONTIERS OF PLASMA PHYSICS AND TECHNOLOGY Bangkok, March 5, 2007 Research Funded by: MPBT/NSERC/UofA Senior Industrial Research Chair Natural Sciences and Engineering Research Council of Canada U of A - R. Fedosejevs p.1

2 Outline Introduction to LIBS Scaling of LIBS to µj Energies µlibs Applications 2D Surface Microanalysis Fingerprint Detection & Imaging Two Pulse Technique to Improve Limit Of Detection Measurement of Elemental Contaminants in Water µlibs in Microfluidic Systems for Lab on a Chip Analysis Conclusions U of A - R. Fedosejevs p.2

3 Overview of LIBS Process 1 Laser Strikes Target Material 2 Plasma Plume Expands 3 Spectra is obtained. 1.6 Focussing Lens Spectrometer Target Material Characteristic Radiation Wavelength (nm) U of A - R. Fedosejevs p.3

4 Overview of LIBS Plasma Expansion Laser pulse 20 ns wavelength (nm) Target Material Target Material Target Material Laser initiates breakdown Plasma forms Portion of sample taken into plasma Plasma expands rapidly Continuum radiation dominates Shockwave launched Atomic emission dominates Continuum decreases U of A - R. Fedosejevs p.4

5 Typical µlibs Experimental Set-up Alignment Camera CCD Spectrometer/OMA Rieger et al., Appl. Spect. 56, 689 (2002) Laser Pulse: 10 ns, E = µj I = GW/cm 2 R = 99% 50 ps, E = µj I = TW/cm fs, E = µj I = PW/cm 2 ~ 5 µm focal spot Dichroic Mirror Laser Pulse Microscope Objective Plasma PD PM Counts OMA Sample Filter Wavelength (nm) U of A - R. Fedosejevs p.5

6 Time Evolution of an Aluminum Alloy Plasma 1e+5 Al Al+ Al Al 3003, 10 µm slit 248 nm, 10 ns, T G = 300 ns E av = 200 µj Al+ Intensity 1e+4 Mn+ Mn Fe Mn 1e+3 0 Delay (ns) Si W avelength (n m ) AlO-bands U of A - R. Fedosejevs p.6

7 Overview of LIBS :Typical Spectra SPS steel, 0.6% Chromium SPS Chromium Nominal Composition: 0.6% PH steel, 16.5% Chromium 17-4 PH Chromium Nominal Composition: 16.5% Cr I Triplet nm, nm, nm Observing elements at less than a single percent concentration is straightforward Choose spectral window according to the material being observed to maximize information gathered Wavelength (nm) Wavelength (nm) U of A - R. Fedosejevs p.7

8 Advantages of LIBS Laser-Induced Breakdown Spectroscopy: offers rapid analysis requires no sample preparation sensitive to all elements scalable in sample size requires no contact with the sample work in hostile environments Laser beam being directed through the lead glass shield window to measure radioactive materials U of A - R. Fedosejevs p.8

9 LIBS Inspection of Gas Cooled Reactor Using a fiber optically coupled LIBS system for finding low ductility joints in superheated steam tubes by anomalously high copper content Applied Photonics U of A - R. Fedosejevs p.9

10 Scaling of LIBS to µj Energies Lower laser pulse energies 100 µj: Smaller spot sizes reduces damage to sample Allows micron scale resolution Higher repetition rate laser systems can be used Possibility of portable LIBS systems LODs achieved are comparable to mj LIBS New Subfield of µlibs Applications 3D surface Microanalysis with µm lateral and sub-µm depth resolution On line pollution monitoring of industrial effluents Monitoring of drinking water standards Microfluidic point of care medical diagnostic systems U of A - R. Fedosejevs p.10

11 Definition of Limit of Detection Noise is evaluated from the pixel to pixel variation on either side of the signal Signal LOD (limit of detection) is the point where signal within the full linewidth of the emission line is 3σ above the average noise scaled to the integration width Noise U of A - R. Fedosejevs p.11

12 Single Shot Surface Probe Capability U of A - R. Fedosejevs p.12

13 Single Shot µlibs Aluminum Precipitates Correlation of elements within precipitates - identification Aluminum 2024 Alloy Cravetchi et al., Spectrochimica Acta. 59, (2004) U of A - R. Fedosejevs p.13

14 Surface Mapping at sub-µj Energies with Femtosecond Pulses 0.5 µj Spectra Accumulation of µj pulses yields useful spectra High rep-rate fiber or microchip lasers have potential for LIBS Remaining Issues: Scaling to sub micron resolution with different materials Integration of high rep-rate laser source with ICCD Al2024, 100 shot average, 0.5 µj, 266 nm, 130 fs 600 l/mm, 100 µm slit, Gate Delay 2.5 ns, Gate Width 100 ns, Pixel Time 16 µs, Gain 275 counts/photoelectron, nm, nm U of A - R. Fedosejevs p.14

15 Surface Mapping at sub-µj Energies Al 2024 Alloy, 0.85 µj, 266 nm, 120fs Grey background matrix, Black Al 2 CuMg, White - Al 6 (Cu,Fe,Mn) Can build up a map of aluminum alloy surfaces with many single shots 2D map of aluminum alloy possible with sub microjoule energies However, a limited number of photons are available at these energies U of A - R. Fedosejevs p.15

16 LIBS Fingerprint Detection and Imaging Experimental Setup for µlibs measurement of fingerprints U of A - R. Fedosejevs p.16

17 Fingerprint Detection Characteristic Spectra Na nm Sample spectra from a fingerprint ridge and gap between fingerprint ridges on silicon wafer Si nm (2 nd order) 2D mapping technique can be applied to latent fingerprints 80 µj, 130 fs fs pulses at 400 nm N shots = 1, E laser = 80 µj, 400 nm 130 fs pulse T delay = 5 ns, T gate = 1 µs, Slit = 100 µm, Readout Time = 16 µs 1200 lines/mm grating U of A - R. Fedosejevs p.17

18 Fingerprint Detection - Line Scan Na signal Si signal suppressed at locations with a fingerprint ridge Si signal Femtosecond probe pulses only sensitive to surface layer N shots = 1, E laser = 80 µj, 400 nm 130 fs pulse T delay = 5 ns, T gate = 1 µs, Slit = 100 µm, Readout Time = 16 µs 1200 lines/mm grating U of A - R. Fedosejevs p.18

19 Fingerprint Detection 2D Scans Na signal 2D LIBS scan of a 1 mm by 5 mm area of a latent fingerprint from right thumb Ridge detail is clearly visible in the Na image (upper) and Si image (lower) Si signal M. Taschuk et al., Applied Spectroscopy 60, pp (2006) N shots = 1, E laser = 80 µj, 400 nm 130 fs pulse T delay = 5 ns, T gate = 1 µs, Slit = 100 µm, Readout Time = 16 µs 1200 lines/mm grating U of A - R. Fedosejevs p.19

20 Durability of Na Signature Sodium signals: Original After 2 cleaning wipes with alcohol soaked lens tissue After 4 cleaning wipes with alcohol soaked lens tissue N shots = 1, E laser = 80 µj, 400 nm 130 fs pulse T delay = 5 ns, T gate = 1 µs, Slit = 100 µm, Readout Time = 16 µs 1200 lines/mm grating U of A - R. Fedosejevs p.20

21 Shift to UV Excitation and Lower Energies Thus far used 80 µj 400nm femtosecond pulses Sample is mostly destroyed using a 50 µm sampling grid Try with 5 µj 266nm femtosecond pulses stronger UV absorption allows lower pulse energy threshold Much smaller 10 µm craters Large surface area preserved for future analysis if necessary Better suited to lower energy, higher repetition rate laser U of A - R. Fedosejevs p.21

22 Fingerprint Detection 5 µj 120 fs 266nm Probe Pulses Reflective laser focusing and achromatic plasma imaging system to collect broadband spectrum Schwarzchild Objective U of A - R. Fedosejevs p.22

23 Craters from Scanning with 5 µj 266 nm Pulses Large amount of the surface area remains undamaged by the craters U of A - R. Fedosejevs p.23

24 SNR scaling with Pulse Energy SNR scaling for 3 fingerprints using 266 nm pulses SNR approaching limit for single shot acquisitions at ~ 3 uj N shots = 100 shot average, 266 nm, 130 fs pulses T delay = 1-5 ns, T gate = 1 µs, Slit = 100 µm, Readout Time = 16 µs 600 lines/mm grating U of A - R. Fedosejevs p.24

25 Fingerprint Imaging 5 µj 266 nm Pulses Na 2D LIBS scan of a 2 mm by 5 mm area of a latent fingerprint Ridge detail is clearly visible in the Na image (upper) and Si image (lower) Si Energy requirements reduced to levels easily compatible with fiber or microchip lasers Portable system at khz acquisition rate may be possible N shots = 1, E laser = 5 µj, 266 nm 130 fs pulse T delay = 5 ns, T gate = 1 µs, Slit = 500 µm, Readout Time = 16 µs 600 lines/mm grating U of A - R. Fedosejevs p.25

26 Two Pulse LIBS: Laser Ablation - Laser Induced Fluorescence Utilizes two pulse technique One pulse to ablate the sample and create a plume Second pulse resonantly excites the atomic species of interest Improvement of detection limit to ppb from ppm level Must optimize the parameters for the two laser pulses Pulse energies Inter-pulse temporal separation Detector efficiency U of A - R. Fedosejevs p.26

27 Typical LA-LIF Experiment Layout Waterjet diameter: 1 mm Breakdown Pulse Probe Pulse 50 cm lens 2 ω Dye laser wavelength set to 257 nm for Al and 283 for Pb U of A - R. Fedosejevs p.27

28 Single Pulse µlibs of 500 ppm Pb At shorter gate delays plasma exhibits significant continuum background As the plasma cools continuum decreases rapidly Optimization of LOD requires optimum gate time N shots =100, E laser =260 µj, 500 ppm Pb in water Gate width =100ns, Slit = 300µm, Detector gain = 255 U of A - R. Fedosejevs p.28

29 LA-LIF of Pb 6p 1/2 7s 1/2 J=1 J= P P 0 Excitation wavelength nm 5.8 x 10 7 s nm 8.9 x 10 7 s nm 3.4 x 10 7 s -1 6p 1/2 6p 3/2 J=2 J=1 Fluorescence wavelength nm 1.5 x 10 8 s P P 1 6p 1/2 6p 1/2 Ground J=0 3 P 0 U of A - R. Fedosejevs p.29

30 LA-LIF spectrum of Pb Millions nm intensity (count) nm 368nm wavelength (nm) N shots =1000, E laser = 170 µj, E 2pulse = 10 µj, T=300ns, Slit width = 300 µm, Grating 1200l/mm, [Pb]: 50ppm U of A - R. Fedosejevs p.30

31 Time Resolved LA-LIF Signal for Pb Signal only appears with the probe pulse Enhancement is short-lived, on nanosecond time scale Signal (counts) single pulse LA-LIF signal Fluorescence Signal Time (ns) N shots =100, E laser =260 µj, E 2nd pulse = 45nJ, 500ppm of Pb in water, T= 700ns, Gate delay =700ns, Slit = 300µm, Detector gain = 255 U of A - R. Fedosejevs p.31

32 Selective enhancement of LA-LIF Conc (ppm) Upper level (cm -1 ) Pb Al * The two spectra have been offset vertically Al Pb Fluorescence Signal intensity (count) 2 pulses 1 pulse wavelength (nm) N shots =100, E laser = 170 µj, E 2pulse = 10 µj, Gate width =100ns, T=300ns, Slit = 300 µm, Detector gain = 255 U of A - R. Fedosejevs p.32

33 Optimization Pulse Separation 7e+6 6e+6 5e+6 Signal (count) 4e+6 3e+6 2e+6 1e Pulse Separation (ns) N shots =1000,E laser = 170 µj E 2pulse = 8 µj, Gate width =100ns, Slit = 300 µm, Detector gain = 255, [Pb]= 50 ppm in water U of A - R. Fedosejevs p.33

34 Scaling with Probe Pulse Energy 1.00E E+00 Signal (count) 6.00E E+00 First Pulse Energy 2.00E E Second Pulse Energy (µj) N shots =1000, T= 300ns, Slit = 300 µm, Detector gain = 255, [Pb]= 50 ppm in water U of A - R. Fedosejevs p.34

35 LA-LIF spectrum of 100 ppb Pb sample: 1000 shot SNR~5 Intensity (count) Wavelength (nm) N shots =1000, E laser = 170 µj E 2pulse = 10 µj, T= 300ns, Slit = 300 µm, Detector gain = 255, Gate width = 300 ns U of A - R. Fedosejevs p.35

36 Pb spectrum of 100 ppb Pb sample: shots SNR~12 Intensity (count) Wavelength (nm) N shots =10000, E laser = 170 µj E 2pulse = 10 µj, T= 300ns, Slit = 300 µm, Detector gain = 255, Gate width = 300 s U of A - R. Fedosejevs p.36

37 100 shot LOD for Pb in Water 1.E+07 Ablation pulse (µj) 2 nd Pulse (µj) T (ns) Gate width (ns) LoD (3σ) ppb 1.E+06 normalized signal 3σ noise floor 1.E+05 Counts 1.E ppb 1.E+03 1.E Pb Concentration (ppm) U of A - R. Fedosejevs p.37

38 1000 shot LOD for Pb in Water Ablation pulse (µj) 2 nd Pulse (µj) T (ns) Gate width (ns) LoD (3σ) ppb 1.E+08 1.E+07 normalized signal 3σ noise floor count 1.E+06 1.E ppb 1.E+04 1.E Pb Concentration (ppm) U of A - R. Fedosejevs p.38

39 Comparison with Previous Pb LOD Reports All values scaled to 100 shot equivalents µlibs [1] R. Kopp, et. al. Fresenius J. Anal. Chem. 355, 16 (1996). [2] G. Arca, et. al. IGARSS 96, Vol 1, May 1996, [3] M. Taschuk, et. al. EMSLIBS 2003, Heraklion, Crete October 1st, 2003 [4] K. M. Lo, et. al. Appl. Spectrocopy, vol. 56, Number 6, 2002 [5] X.Y.Pu, Appl. Spectroscopy 57,5, method ICP-MS ICP-AES Detection limit of Pb using other techniques Limit of detection ppb 1.5ppb source icppage.htm tern.edu/ [6] Le Bihan et. al. Annal. Bioanal. Chem 2003 ETA-LEAF 0.3ppt (20 shots) Le Bihan et. al.2003 U of A - R. Fedosejevs p.39

40 µlibs in Microfluidic Systems Apply LIBS in microfluidic system Detection of single cell contents Lab-on-a-chip application micro Total Analytic Systems (µtas) x Counts (Bg Corrected) µm droplet Wavelength nm LIBS probe orifice, ~ few µm microchannel ~50µm Drop-on-demand actuator (thermal or piezoelectric) U of A - R. Fedosejevs p.40

41 Microdroplet Generation Rapid thermal heater µs-pulse Piezoelectric pulser U of A - R. Fedosejevs p.41

42 Prototype Thermal Droplet Ejector microheater orifice microchannel microheater Micro- Heater Element The orifice, the channel, and the reservoir are all machined by laser-micromachining U of A - R. Fedosejevs p.42

43 Conclusions Development of µlibs High resolution and small probe spot size demonstrated LODs in ppm range demonstrated Initial µlibs Applications: Surface mapping of alloys for quality control of metal manufacturing Micron scale size resolution Fingerprint detection both by Na and substrate lines Overcomes fluorescence masking for some materials µj energies with fs uv pulses leaves large surface area for further investigation or as evidence Can increase acquisition speed to multi-khz repetition rates High resolution 3D scans possible µm lateral and sub-µm depth U of A - R. Fedosejevs p.43

44 Conclusions LA-LIF µj energies sufficient for excitation and resonant probing Increase sensitivity to ppb levels Initial LA-LIF Applications: Monitoring of water quality 25 ppb detection of Pb in water with shots High repetition rate lasers ( khz) would allow 2.5 ppb sensitivity in second measurement times i.e. real time water quality monitoring Can be scaled to portable systems using upconverted fiber lasers with fiber Bragg gratings to generate the exact probe wavelengths Future applications in lab on a chip for medical diagnostics in the doctors office U of A - R. Fedosejevs p.44

45 The End U of A - R. Fedosejevs p.45