xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader nam 2200421Ka 4500
controlfield tag 001 002047321
007 cr mnu|||uuuuu
008 100113s2008 flu s 000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0002764
Laser remote sensing of trace chemical species using 10.6 m CO laser enhanced breakdown spectroscopy and differential absorption lidar
h [electronic resource] /
by Avishekh Pal.
[Tampa, Fla] :
b University of South Florida,
Title from PDF of title page.
Document formatted into pages; contains 251 pages.
Dissertation (Ph.D.)--University of South Florida, 2008.
Includes bibliographical references.
Text (Electronic dissertation) in PDF format.
Mode of access: World Wide Web.
System requirements: World Wide Web browser and PDF reader.
Non-Latin script record
Advisor: Dennis Killinger, Ph.D.
ABSTRACT: Several different laser remote sensing techniques related to the detection of trace chemical species were studied. In particular, a Differential-Absorption lidar (DIAL), a Laser-Induced-Breakdown Spectroscopy (LIBS) lidar, and a Raman lidar were studied. Several of the laser spectroscopic techniques that were used were common throughout these different studies. More precisely, 10.6 m CO laser related spectroscopy was common for the DIAL and LIBS studies, and 266 nm Nd:YAG laser related spectroscopy was used for the LIBS and Raman studies. In the first system studied a tunable CO DIAL system was developed for the first time to our knowledge for the potential detection of the explosive Triacetone Triperoxide (TATP) gas clouds. The system has been used to measure gas samples of SF, and has shown initial absorption measurements of samples of TATP contained within an enclosed optical absorption cell.The LIBS plasma emission covering the range of 240 800 nm was enhanced by use of a nearly simultaneous 10.6 m CO laser that increased the LIBS plasma emission by several orders of magnitude. The emission spectrum was used to detect and identify the species of interest. Plasma temperatures on various solid substrates were measured. An increase in the plasma temperature of about 5000 K was measured and analyzed, for the first to our knowledge, due to the addition of the CO laser pulse to the LIBS plasma generated by the Nd:YAG laser. An optimum temporal overlap of the two laser pulses was found to be important for the enhancement. Finally, in a third related lidar system, initial 266 nm Raman lidar studies were conducted at detection ranges of 15 m. However, significant spectroscopic background interferences were observed at these wavelengths and additional optical filtering is required.DIAL/Lidar returns from a remote retro-reflector target array were used for the DIAL measurements after passage through a laboratory cell containing the TATP gas. DIAL measured concentrations agreed well with those obtained using a calibrated Ion Mobility Spectrometer. DIAL detection sensitivity of the TATP gas concentration in the cell was about 0.5 ng/l for a 0.3 m path-length. However, the concentration of TATP was found to be unstable over long periods of time possibly due to re-absorption and crystallization of the TATP vapors on the absorption cell windows. A heated cell partially mitigated these effects. In the second set of studies, a Deep UV LIBS system was developed and studied for the remote detection of solid targets, and potentially chemical, biological, and explosive substances. A 4th harmonic Q-Switched Nd:YAG laser operating at 266 nm was used for excitation of the LIBS plasma at standoff ranges up to 50 m .
Laser breakdown spectroscopy
Boltzmann temperature measurements
t USF Electronic Theses and Dissertations.
Laser Remote Sensing of Trace Chemical Species Using 10.6 m CO2 Laser Enhanced Breakdown Spectroscopy and Differential Absorption Lidar. by Avishekh Pal A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy Department of Physics College of Arts and Sciences University of South Florida Major Professor: Dennis Killinger, Ph.D. Sarath Witanachchi,Ph.D. Myung K.Kim,Ph.D. Chun Min Lo, Ph.D Wei Chen,Ph.D. Date of Approval: November 2008 Keywords: laser breakdown spectroscopy, DIAL, lidar,Raman spectroscopy, Boltzmann temperature measurements Copyright 2008, Avishekh Pal
ACKNOWLEDGEMENTS I would like to pay a heartfelt dept of gr atitude to Dr. Killinger for his support and guidance during the course of this project. Your patience, understanding and mentoring during the course of my rese arch is highly appreciated. I would like to thank the Department of Physics at USF for the opportunity. All the favors by the department staff and LAMSAT Lab at USF for loaning equipment necessary for my research was appreciated. Thanks a lot to Phil Bergeron and Robert Hyde for their help. I would like to express my thanks to my dissertation committee members Dr. S.W. Witananchchi, Dr. M. K. Kim, Dr. W. Chen and Dr C. M. Lo for serving on my committee and Dr. Gregory McColm for chairing my defense examination. I would like to thank Dr. M. Sigman and Douglas Clark for preparing and providing me with the required TATP sample for my research. I am very appreciative to have worked closely with Ed Dottery, Rob Waterbury, Jim Bernier and Chis Stefano at Alakai and want to thank them for their assi stance during my research at Alakai in Largo. The LIB/TEPS and Raman remote sensing pa rt of the research reported in this dissertation was supported in part by a grant from Alakai Consulting & Engineering, Inc. through a contract W911QX-07-C-0044 with the U.S. Army Research Laboratory, and a Florida Matching Grant FHT 08-11. The DIAL re mote sensing part of the research was partially supported by AFOSR Grant FA96550-06-1-0363.
TABLE OF CONTENTS LIST OF TABLES v LIST OF FIGURES vi ABSTRACT xv CHAPTER 1. INTRODUCTION A ND MOTIVATION 1 CHAPTER 2. BACKGROUND INFORM ATION ON REMOTE SENSING BY DIAL LIDAR, LIBS AND RAMAN SPECTROSCOPY 7 2.1 Description of Lidar 9 2.1.1 Hard Target Lidar Equation 11 2.2 Differential Absorption Lidar (DIAL) 12 2.3 Lidar S/N and Detector Noise 16 2.4 Review of Absorption Spectroscopy Theory 17 2.5 Introduction to Laser Induced Breakdown Spectroscopy (LIBS) 20 2.6 LIBS by Different Kinds of Laser 23 2.6.1 Femtosecond Laser Ablation 23 2.6.2 Picosecond Laser Ablation 24 2.6.3 Nanosecond Laser Ablation 24 2.7 Time evolution of the LIBS signal 28 2.8 Raman Spectroscopy 33 2.9 Different Raman Techniques 36 2.9.1 Resonance Raman 36 2.9.2 Surface Enhanced Scattering (SERS) 40 2.9.3 Ultraviolet Resonance Raman Spectroscopy (UVRRS) 41 2.9.4 Time-Resolved and Pulsed Raman 42 i
2.9.5 Confocal Raman Microscopy 43 2.10 Some Raman Spectroscopy Applications 44 2.11 Laser Wavelength Selection and Eye Safety 45 CHAPTER 3. REMOTE SENSING OF EXPLOSIVE RELATED GASES USING CO2 DIAL LIDAR 50 3.1 Motivation 50 3.2 Introduction to TATP 51 3.3 Preliminary TATP Data and LIDAR Signal Return Calculations 52 3.4 Experimental Setup 67 3.5 DIAL Calibration with SF6 69 3.6 DIAL Lidar Detection of TATP Gas 81 3.7 Conclusion for CO2 DIAL detection of TATP 92 CHAPTER 4. ENHANCED LIBS PL ASMA EMISSION FOR STANDOFF DETECTION USING A 266 nm AND 10 m CO2 LASER SYSTEM 93 4.1 Single Pulse LIBS Experiments 93 4.2 Deep UV LIBS and 10 m laser TEPS Experimental Apparatus 97 4.3 CO2 Laser LIBS Measurement 104 4.4 Enhanced 266 nm Nd:YAG LIBS Emission Using Simultaneous 10.6 m Laser Pulse 104 4.5 Timing Overlap of Two Lasers and Its Effect on LIBS Plasma 109 4.6 Geometrical Overlap of Lase r Intensities on Substrates 111 4.7 Orthogonal Geometry for the LIBS/TEPS Setup 119 CHAPTER 5. STUDY OF CRATERS, RANGE DEPENDENCE AND EFFECTS DUE TO PRESENCE OF ARGON GAS ENVIRONMENT ON LIBS/TEPS SIGNAL 125 ii
5.1 LIBS/TEPS Induced Crater Study on Al Substrates 125 5.2 Crater Depth Measurements for LIBS/TEPS as a Function of Standoff Range 127 5.3 Standoff LIBS/TEPS Measurement as a Function of Range 139 5.4 LIBS/TEPS Experiments in Argon Environment 141 CHAPTER 6. ELECTRON TEMPERATU RE MEASUREMENT OF THE LIBS/TEPS PLASMA BY BOLTZMANN PLOT METHOD 147 6.1 Basic Theory 147 6.2 Experimental Setup 151 6.3 Electron Plasma Temperature Measurements of LIBS/TEPS by 10.6 m and 1.06 m Nd:YAG lasers 161 6.4 Additional Electron Plasma Temperature Measurements of LIBS/TEPS of 10.6 m and 266 nm Nd:YAG laser on various Substrates and emission lines 166 6.5 Electron Plasma Temperature Measurements of LIBS/TEPS on Substrates in Argon Environment 174 6.6 Background Theory for Electron Density Measurement Using Stark Width 179 6.7 Experimental Stark Width Calculation 182 CHAPTER 7. LIBS/TEPS ON VARI OUS SUBSTRATES AND STUDY OF SIGNAL ENHANCEMENT DUE TO DUAL PULSE 266 nm Nd:YAG AND 10.6 m CO2 LASER 186 7.1 Multiple Pulse LIBS 186 7.2 Dual Pulse Nd:YAG LIBS 188 7.3 266 nm Nd:YAG Dual Pulse LIBS with the Addition of 10.6 m CO2 laser 191 7.4 Survey of LIBS/TEPS Signal on Various Substrates 197 iii
7.5 Peak Ratio Analysis for Impur ity Detection on Substrates 205 CHAPTER 8. DETECTION OF SPECIE S BY 266 nm Nd:YAG LASER STANDOFF RAMAN LIDAR SYSTEM 214 8.1 Experimental Setup for St andoff Detection Using Raman Spectroscopy 214 8.2 Conclusion for Raman Lidar Detection 219 CHAPTER 9. POSSIBLE LIBS/TE PS ENHANCEMENT MECHANISMS AND PLASMA INTERACTIONS 224 9.1 Basics of Laser-Matter Interaction 224 9.2 Plasma Process 228 9.3 Theoretical Models for Plasma 230 9.3.1 Corona Model 230 9.3.2 Local Thermodynamic Equilibrium Model 231 9.4 Possible TEPS/LIBS Mechanisms 233 CHAPTER 10. CONCLUSIONS AND FUTURE WORK 237 REFRENCES 240 APPENDICES 248 APPENDIX A. OPERATIONAL AMPLIFIER FOR THE MCT DETECTOR. 248 APPENDIX B. LABVIEW INTERF ACE PROGRAM USED TO ACQUIRE AND STORE DATA FROM THE DIAL SYSTEM. 249 ABOUT THE AUTHOR End Page iv
LIST OF TABLES Table 3.1 Specific parameters for TATP 53 Table 4.1 Enhancement of a few id entified lines comparing the 266 nm Nd:YAG LIBS spectrum and the 266 nm/10.6 m LIBS/TEPS signals. 108 Table 5.1 Technical Specifications for 3030 Dektak profilometer 133 Table 5.2 Data for the mass ablated with the use of 266 nm Nd:YAG laser al one LIBS and both Nd:YAG and CO2 lasers together (LIBS/TEPS). Target range is 30 m. 137 Table 5.3 Measured averaged values for the LIBS/TEPS craters on Aluminum substrate as a function of range 137 Table 6.1 Spectroscopic constants of the neutral Fe(I) lines used in Boltzmann pl ot temperature determination. (from NIST Database55) 156 Table 6.2 Spectroscopic constants for Al (I) lines chosen in Boltzmann plot temperature determination. (from NIST database55) 167 Table 6.3 Spectroscopic constants for Al (I) lines chosen in Boltzmann plot temperature determination (from NIST database55) 170 Table 6.4 Spectroscopic constants for Pb (I) lines chosen in Boltzmann plot temperature determination (from NIST database55) 175 Table 6.5 Spectroscopic constants for Ar(I) lines chosen in Boltzmann plot temperature determination (from NIST database55) 177 v
LIST OF FIGURES Figure 2.1 Typical setup for passive remote sensing using absorption spectroscopy. 8 Figure 2.2 Typical setup for active remote sensing. 10 Figure 2.3 Typical setup of DIAL remote sensing equipment. 13 Figure 2.4 DIAL detection of atmospheric CO2 compared to in-situe gas analyzer measurements. DIAL measurements are the blue doted data and in-situe are the green (grey) solid line. (from G. Koch3; reprinted with permission from Applied Optics). 15 Figure 2.5 Typical LIBS setup. 21 Figure 2.6 Typical LIBS spectrum fro m metallic Pb showing neutral Pb emission Pb(I). Lines due to sing le ionized Pb, Pb (II), are also marked ( from V. Babushok15). 26 Figure 2.7 Typical LIBS spectra of Al at 25 m range using (a) Femtosecond laser, (b) Picosecond laser, (c) Nanosecond laser (from P.Rohwetter16). 27 Figure 2.8 Typical timing delay after th e plasma formation for the emission of different species in a LI BS spectra. (from A.Miziolek11). 29 Figure 2.9 Typical LIBS emission spectra as observed at different timing delays.(from A.Miziolek11). 30 Figure 2.10 Life cycle diagram of LI BS, showing the progression of events (adapted from A. Miziolek11). 32 Figure 2.11 Energy level diagram show ing the states involved in Raman spectroscopy. 35 Figure 2.12 Typical Continuous Wa ve (CW) Raman layout. 37 Figure 2.13 Typical Raman spectra for Napthalene. (adapted from http://www.chemistry.ohiostate.edu/~ rmccreer/freqcorr/images/nap.ht ml , Richard L. McCreery, The Ohio State University20) 38 vi
Figure 2.14 Remote Raman spectra of et hyl benzene and tolune over the spectral range 180-1800 cm-1 at 10 m range (from S.K Sharma21). 46 Figure 2.15 Chart of Maximum Permi ssible Energy versus exposure time for pulsed lasers at different wavelengths (derived from ANSI Z136.6 Standard25). 48 Figure 2.16 Chart of Maximum Permissi ble Energy versus wavelength for pulsed lasers (derived from ANSI Z136.6 Standard25). 49 Figure 3.1 (a) Picture of Acetone pe roxide powder. (b) Ball-and-stick model of the acetone peroxide trimer (TATP). 54 Figure 3.2 The acid (H+) used in the synthesis is typically concentrated HCl or H2SO4. The alkyl groups (R and R') in the starting ketone, acetone, are CH3 (methyl); however, these groups are not required to be identical. 55 Figure 3.3 Vapor phase FTIR spectrum of TATP. 56 Figure 3.4 Solid phase FTIR-ATR spectrum of TATP. 58 Figure 3.5 Absorbance spectrum for vapor phase TATP (cell path length of 5 cm; TATP concentratio n/partial pressure 8.28 Pa; 28oC ). 59 Figure 3.6 Predicted absorption spectr a for 1m cloud of TATP (Path 1 m ; saturated vapor of 7 Pa). 60 Figure 3.7 Composite spectrum of 100 m atmosphere and 1m TATP plume. 61 Figure 3.8 Composite spectrum of 100 m atmosphere and 1m TATP plume and overlay of CO2 laser lines. 63 Figure 3.9 Lidar equation works file showing parameters for calculating S/N and lidar return signal using USF-LIDAR-PC software. 64 Figure 3.10 Predicted lidar return signal as a function of range for lidar returns from a hard target. 65 Figure 3.11 Predicted lidar return signal as a function of range for lidar returns from a retro-reflector. 66 Figure 3.12 Schematic of laborator y DIAL/Lidar setup. 68 Figure 3.13 Photogr aph of laboratory CO2 DIAL/Lidar setup for TATP detection. 70 vii
Figure 3.14 Picture of the single retro-reflector and retr o reflector array used for the experiment. 71 Figure 3.15 Qualitative transmission spectra of gaseous SF6 as a function of wavelength near the R(24) and P(24) CO2 laser lines. 72 Figure 3.16 175 cm long PVC plastic ce ll with Mylar windows with injection ports for TATP. 74 Figure 3.17 DIAL returns from a single retro-reflector at a distance of 100 m for the CO2 offline R (24) and online P(24) for SF6. 75 Figure 3.18 DIAL returns from the retroreflector array at a distance of 100m for some of the different CO2 laser lines. 77 Figure 3.19 Picture of th e vacuum sealed cell for SF6. 78 Figure 3.20 DIAL transmission measur ements as a function of time for 0.2 Torr of SF6 (5 cm cell) and lidar target range of 100 m. Predicted absorption for the on-line P(24) line is shown as a dotted line. 79 Figure 3.21 DIAL transmission measurem ents as a function of time for 0.5 Torr of SF6 (5 cm cell) and lidar target range of 100 m. Predicted absorption for the online P(24) line is shown as a dotted line. 80 Figure 3.22 Expected transmission spectra of TATP for a 175 cm path length and concentration of 4.3 Pa near the R(24) and P(24) CO2 laser lines. 82 Figure 3.23 Transmission plot as func tion of time as TATP was injected into the 1.75 m te st cell. 84 Figure 3.24 30 cm long heated glass test cell with Mylar wi ndows for TATP measurements. 85 Figure 3.25 Online DIAL transmission as a function of time showing injection of TATP into a heated 30 cm long absorption cell. 87 Figure 3.26 DIAL measured TATP concen tration inside a heated 30 cm long absorption cell as function of time. 88 Figure 3.27 Online and offline DIAL tr ansmission of TATP as a function of time in a heated 30 cm long absorption cell after a period of 4 hours. 90 viii
Figure 3.28 Superimposed transmission spectra of TATP and acetone. 91 Figure 4.1 Schematic of LIBS setup us ed for standoff measurements. 94 Figure 4.2 LIBS signal from a pure aluminum target (range 25 m) using a 266nm Nd:YAG laser initiated plasma. 96 Figure 4.3 Schematic of Deep UV 266 nm LIBS and CO2 laser enhancement LIBS standoff system. 98 Figure 4.4 (a) Photograph of the LIBS/TEPS setup (b) Photograph of the CO2 laser used in LIBS/TEPS setup. 99 Figure 4.5 CO2 beam profile as measured by Spiricon beam profiler. 101 Figure 4.6 Beam profile of 266 nm Nd :YAG laser as measured by Spricon beam profiler. 102 Figure 4.7 Burn patterns on paper due to CO2 laser: (a) 18 mm diameter spot, (b) 20 mm diameter spot, (c) 7.5 mm diameter spot, (d)10 mm diameter spot. 103 Figure 4.8 LIBS signal on a pure ceramic substrate using a CO2 pulsed laser initiated plas ma as a function of CO2 energy density on the target. 105 Figure 4.9 Enhanced LIBS emission signal on Al target due to 10.6 m CO2 laser for 266 nm of Nd:YAG initiated plasma. 107 Figure 4.10 Pulse shape and relative timing of CO2 and 266 nm Nd:YAG laser pulses ( No plasma). 110 Figure 4.11 CO2 and 266 nm Nd:YAG laser pulse observed directly by use of silicon and pyro electri c detector with the plasma. 112 Figure 4.12 LIBS spectrum as a function of CO2 laser pulse delay compared to 266nm Nd:YAG laser pulse, (standoff range of 35 m). 113 Figure 4.13 Various elemental LIBS emission lines as a function CO2 laser delay, (stand-off range of 35 m). 114 Figure 4.14 Beam geometry of CO2 laser Nd:YAG laser pulse on the substrate. 115 Figure 4.15 LIBS/TEPS signal as a f unction of energy density of the CO2 laser pulse. 117 ix
Figure 4.16 LIBS spectrum as a function of CO2 laser spot size on the target Substrate. 118 Figure 4.17 Schematic of orthogonal beam geometry for TEPS/LIBS setup. 120 Figure 4.18 LIBS spectrum with and without addiction of CO2 laser to 266 nm Nd:YAG laser with the ort hogonal geometry for both beams. 121 Figure 4.19 CO2 and 266 nm Nd:YAG laser pulse observed directly by use of silicon and pyro electric detector with the plasma for the collinear geometry for both beams. 123 Figure 4.20 Dependence of CO2 and Nd:YAG laser beam overlap on orthogonal geometry. 124 Figure 5.1 (a) SEM images of smooth Al surface. (b) 266 nm Nd:YAG induced LIBS crater from 30 m distance on Al substrate. (c)&(d) Increased magnification within the LIBS crater. 126 Figure 5.2 SEM images of CO2 laser induced craters on Al substrate from 30 m. (a) 100 m resolution, (b) 500 m resolution and (c) 50 m resolution. 128 Figure 5.3 SEM images of CO2 and Nd:YAG induced craters on Al at 30m range (a) 100 m resolution, (b) 500 m resolution and (c) 50 m resolution. 129 Figure 5.4 SEM images of 266 nm Nd:YAG induced craters at 1m range on Al substarte at (a) 100 m resolution, (b) 500 m resolution and (c) 50 m resolution. 130 Figure 5.5 SEM image of CO2 and 266 nm Nd:YAG induced. 131 Figure 5.6 Photograph of the 3030 Dektak profilometer. 132 Figure 5.7 (a) Photograph of the crater formed at 25 m range on a aluminum substrate (b) Screen shot of the Pr ofilometer Display for the crater. 135 Figure 5.8 Diagram for the typical crat er formed due to LIBS along with the parameters for the depth and radius of the average cones formed. 136 Figure 5.9 (a) Volume of the craters formed on the substrate as function of range. (b) Volume of the metal lost on substrate as function of range. 138 x
Figure 5.10 Dual pulse LIBS/TEPS si gnal as function of range. 140 Figure 5.11 Schematic of Deep UV 266 nm LIBS and CO2 laser enhancement LIBS/TEPS standoff system in Argon gas environment. 142 Figure 5.12 (a) LIBS/TEPS spectra of Cu substrate. (b) LIBS/TEPS spectra of Cu when the substrate is placed in Ar gas bath. 143 Figure 5.13 (a) LIBS/TEPS spectra of Al substrate. (b) LIBS/TEPS spectra of Al when the substrate is placed in Ar gas bath. 145 Figure 5.14 (a) LIBS/TEPS spectra on bare ceramic substrate. (b) LIBS/TEPS spectra on ceramic when the substrate is placed in Ar gas bath. 146 Figure 6.1 Schematic of modified set up for LIBS/TEPS plasma readings on iron substrate. 152 Figure 6.2 LIBS spectrum on Iron substrate when only 266nm Nd:YAG laser was only used to generate the plasma. 153 Figure 6.3 LIBS spectrum on Iron s ubstrate when both 266nm Nd:YAG and 10.6 m CO2 lasers ware used to generate the plasma. 154 Figure 6.4 Boltzmann plot of Fe (I) lines of the laser-induced plasma for 266nm Nd:YAG LIBS and enhanc ed LIBS/TEPS plasma using both Nd:YAG and CO2 laser pulses. 157 Figure 6.5 Plot of electron plasma temp erature as a function of interpulse delays between two lasers. 159 Figure 6.6 Plot of electron plasma temp erature as a function of interpulse delays between two lasers superimposed with the enhancement lines of Fe(I) and O-777.42, N-746.83. 160 Figure 6.7 Schematic of two laser LIBS system for CO2 laser enhancement of Nd:YAG induced LIBS plasma emission (from D.K Killinger 42). 162 Figure 6.8 LIBS signal from ceramic alumina target for Nd:YAG laser initiated plasma; emission lines are tentatively identified and listed in nm. (from D.K Killinger 42). 164 xi
Figure 6.9 LIBS signal from ceramic alumina target for Nd:YAG and CO2 laser initiated plasma emission li nes are tentatively identified and listed in nm. (from D.K Killinger 42). 165 Figure 6.10 Boltzmann plot of Al (I) lines of the laser-induced plasma for 1064nm Nd:YAG LIBS and enha nced LIBS/TEPS plasma using both Nd:YAG and CO2 laser. 168 Figure 6.11 LIBS spectrum on Al substrate when both 266 nm Nd:YAG and 10.6 m CO2 lasers ware used to generate the plasma. 169 Figure 6.12 Boltzmann plot of Al (I) lines of the laser-induced plasma for 1064 nm Nd:YAG LIBS and CO2 LIBS/TEPS plasma. 171 Figure 6.13 LIBS spectrum on Pb subs trate when both 266 nm Nd:YAG and 10.6 m CO2 lasers ware used to generate the plasma. 172 Figure 6.14 Boltzmann plot of Pb (I) lines of the laser-induced plasma for 1064 nm Nd:YAG LIBS and CO2 LIBS/TEPS plasma. 173 Figure 6.15 LIBS spectrum on Cu substrat e with a constant Ar flow when both 266 nm Nd:YAG and 10.6 m CO2 lasers ware used to generate the plasma. 176 Figure 6.16 Boltzmann plot of Ar (I) lines of the laserinduced plasma generated by 1064 nm Nd:YAG LIBS and CO2 Lasers. 178 Figure 6.17 Superimposed image of the Stark broadening of Fe line along with its Lorenzian fit when both YAG and CO2 laser is used for plasma generation. 183 Figure 6.18 Superimposed image of the Stark broadening of Fe line along with its Lorenzian fit when only YAG laser is used for plasma generation. 184 Figure 7.1 Experimental setup for dual pulse 266 nm Nd:YAG and CO2 laser. 187 Figure 7.2 LIBS signal as a functio n of varying delay between two 266 nm Nd:YAG lasers on a pure ceramic substrate for some selected lines. 189 Figure 7.3 LIBS signal on a pure cera mic substrate using a single and double 266 nm ND:YAG laser initiated plasma 190 xii
Figure 7.4 LIBS signal on a pure cera mic substrate using a single and double 266 nm ND:YAG laser initiated plasma by switching the order. 192 Figure 7.5 LIBS signal on selected lines when CO2 laser is added to dual 266 nm Nd:YAG laser on a pure ceramic substrate and sweeping its delay with the YAG-YAG delay of 10 s. 194 Figure 7.6 Dual pulse LIBS signal comp ared to LIBS/TEPS for Al (I) 396 nm line for varying CO2 to 266 nm Nd:YAG interpulse delays. 195 Figure 7.7 LIBS signal on selected lines when CO2 laser is added to dual 266 nm Nd:YAG laser on a ceramic substrate and sweeping its delay with the YAG-YAG delay of 4 s. 196 Figure 7.8 LIBS signal on ceramic when CO2 laser is added to single 266 nm Nd:YAG. 198 Figure 7.9 LIBS signal on ceramic when CO2 laser is added to dual 266 nm Nd:YAG. 199 Figure 7.10 LIBS/TEPS signal on bare ce ramic substrate at 20 m range. 200 Figure 7.11 LIBS/TEPS signal on bare C opper substrate at 20 m range. 201 Figure 7.12 LIBS/TEPS signal on Iron substrate at 20 m range. 202 Figure 7.13 LIBS/TEPS signal on Lead substrate at 20 m range. 203 Figure 7.14 LIBS/TEPS signal on Plas tic (cd) at 20 m range. 204 Figure 7.15 LIBS/TEPS spectra of Acet onitrile on pure Al substrate. 207 Figure 7.16 LIBS/TEPS spectra of Alcohol on pure Al substrate. 207 Figure 7.17 LIBS/TEPS spectra of bl each on pure Al substrate. 208 Figure 7.18 LIBS/TEPS spectra of hydroge n peroxide on pure Al substrate. 208 Figure 7.19 LIBS/TEPS spectra of thin la yer of paint on pure Al substrate. 209 Figure 7.20 LIBS/TEPS spectra of ga soline on pure Al substrate. 209 Figure 7.21 LIBS/TEPS spectra of WD -40 oil on pure Al substrate. 209 xiii
Figure 7.22 LIBS/TEPS spectra of th in diesel fuel film on pure Al substrate. 210 Figure 7.23 LIBS/TEPS spectra of thin la yer of dirt on pure Al substrate. 211 Figure 7.24 Peak ratios for elementa l lines of oxygen, hydrogen, Al and CN bands for various substan ces. (a) Plastic, (b) Wd-40 oil, (c) Thin layer of paint, (d) Hydrogen peroxide 212 Figure 7.25 Peak ratios for elementa l lines of oxygen, hydrogen, Al and CN bands for various substan ces. (e) Bleach, (f) Acetonitrile, (g) Alcohol, (h) Gasoline. 213 Figure 8.1 Schematic of the labor atory lidar Raman setup. 215 Figure 8.2 UV Raman signal due to 266 nm on Teflon with 1s integration and 70 mJ/ pulse of laser with the 266 nm beam blocked in the spectrometer shown by the blue line. 217 Figure 8.3 UV Raman signal due to 266 nm on Teflon with 1s integration and 70 mJ/ pulse of laser. 218 Figure 8.4 UV Raman signal due to 266 nm Nd:YAG laser on thin film of paint with 1s integration and 70mJ/ pulse of laser. 220 Figure 8.5 UV Raman signal due to 266 nm Nd:YAG laser on Acetonitrile dissolved in water with 1s integr ation and 70 mJ/ pulse of laser. 221 Figure 8.6 UV Raman signal due to 266 nm Nd:YAG laser on sugar crystals with 1s integration and 70 mJ/ pulse of laser. 222 Figure 9.1 Schematic diagram of ex panding laser produced plasma in ambient gas. Plasma plume is divided into many zones having high-density hot and low-density cold plasma. The farthest zone from the target has minimum plasma density and temperature. Laser is absorbed in lowdensity corona. (from J.P Singh13). 227 Figure 9.2 Plasma expansion diagram. 235 xiv
LASER REMOTE SENSING OF TRAC E CHEMICAL SPECIES USING 10.6 m CO2 LASER ENHANCED BREAKDOWN SPECTROSCOPY AND DIFFERENTIAL ABSORPTION LIDAR Avishekh Pal ABSTRACT Several different laser remote sensing techniques related to the detection of trace chemical species were studied. In particular, a Differential-Absorption lidar (DIAL), a Laser-Induced-Breakdown Spectro scopy (LIBS) lidar, and a Raman lidar were studied. Several of the laser spectroscopic techniques that were used were common throughout these different studies. More precisely, 10.6 m CO2 laser related spectroscopy was common for the DIAL a nd LIBS studies, and 266 nm Nd:YAG laser related spectroscopy was used for the LIBS and Raman studies. In the first system studied a tunable CO2 DIAL system was developed for the first time to our knowledge for the potent ial detection of the explosive Triacetone Triperoxide (TATP) gas clouds. The system has been used to measure gas samples of SF6, and has shown initial absorption measur ements of samples of TATP contained within an enclosed optical absorption cell. DIAL/Lidar returns from a remote retroreflector target array were used for the DIAL measurements after passage through a laboratory cell containing the TATP gas. DIAL measured concentrations agreed well with those obtained using a calibrated I on Mobility Spectrometer. DIAL detection sensitivity of the TATP gas concen tration in the cell was about 0.5 ng/ l for a 0.3 m xv
xvi path-length. However, the concentration of TATP was found to be unstable over long periods of time possibly due to re-absorpti on and crystallization of the TATP vapors on the absorption cell windows. A heated cell partially mitigated these effects. In the second set of studies, a D eep UV LIBS system was developed and studied for the remote detection of solid ta rgets, and potentially chemical, biological, and explosive substances. A 4th harmonic Q-Switched Nd:YAG laser operating at 266 nm was used for excitation of the LIBS pl asma at standoff ranges up to 50 m The LIBS plasma emission covering the range of 240 800 nm was enhanced by use of a nearly simultaneous 10.6 m CO2 laser that increased the LIBS plasma emission by several orders of magnitude. The emission spectrum was used to detect and identify the species of interest. Plasma temperatures on various solid substrates were measured. An increase in the plasma temperature of about 5000 K was measured and analyzed, for the first to our knowledge, due to the addition of the CO2 laser pulse to the LIBS plasma generated by the Nd:YAG laser. An optimum temporal overlap of the two laser pulses was found to be important for the enhancement. Finally, in a third related lidar system, initial 266 nm Raman lidar studies were conducted at detection ranges of 15 m. However, significant spectroscopic background interferences were observed at these wavelengths and additional optical filtering is required.
1 CHAPTER 1. INTRODUCTION AND MOTIVATION Lidar (light detection and ranging) ha s been widely used since the 1960s to study atmospheric particles and clouds.1-10 A lidar is an instrument that uses short pulses of laser light to detect particles or gases in the atmosphere much like radar bounces radio waves off rain in clouds. A te lescope collects and m easures the reflected laser radiation, leading to a map of the atmo sphere's structure. Researchers can then determine the location, distribution, and nature of atmospheric particles and clouds and, under special circumstances, molecular specie s. There are many different types of spectroscopic interactions that can be used with a lidar system that allow one to measure different atmospheric properties. For example, different atmospheric molecules absorb light at certain wavelengths so that a tunable laser can be used to detect a particular atmospheric gas usi ng the difference in the absorption of the backscattered lidar returns. In this case, the DIAL (Differential Absorption Lidar) technique can be extremely sens itive and is able to detect gas concentrations as low as a few hundred parts per billion (ppb). This makes it possible to measure trace pollutants in the ambient atmosphere and monitor stack emissions in the parts per million ranges. It offers excellent oppor tunity for remote sensing of known species, especially ones which are hazardous and are best detected from afar. Laser induced breakdown spectroscopy (LIBS) is another emerging technique which has gained popularity in recent y ears for determining elemental composition. With the ability to analyze solids, liquids and gases with little or no sample preparation, it is more versatile than c onventional wet chemistry methods and is ideal for on-site
2 analysis. The sample size for LIBS is minimal, with typically 0.1 g to 0.1 mg of material ablated from the solid sample. The sample size required minimizes the destructiveness and invasi veness of the technique. Laser induced breakdown spectroscopy (LIBS) is basically an emission spectroscopy technique where atoms and ions are primarily formed in their excited states as a result of an inte raction between a tightly focused laser beam and the material sample.11-13 The interaction between matter and high-density photons generates a plasma plume, which evolves with time and may eventually acquire thermodynamic equilibrium. Samples in the form of solids, liquids, gels, gases, plasmas and biological materials (like teeth, leaf, or blood) can be studied with almost equal ease. LIBS has rapidly developed into a major analytical technology with the capability of detecting many chemical elements in a sample, with realtime response and the potential for stand-off analysis of the target sample. The work presented in this dissertation can be broadly divided into two main parts. In the first part a tunable CO2 DIAL system has been developed, for the first time to our knowledge, for the potential dete ction of explosive related Triacetone Triperoxide (TATP) gas clouds. The system has been used to measure gas samples of SF6 (an absorption calibration ga s), and has shown initial absorption measurements of samples of TATP contained w ithin an enclosed optical ab sorption cell. DIAL/Lidar returns from a remote retro-reflector target array were used for the DIAL measurements after passage through th e laboratory cell containing the TA TP gas. DIAL measured concentrations agreed well with those obtained using a calibrated Ion Mobility
3 Spectrometer. DIAL detection sensitivity of the TATP gas concentration contained in the cell was about 0.5 ng/ l. For the second part of the dissertation, a stand-of f Townsend Effect Plasma Spectroscopy (TEPS) and Laser Induced Breakdown Spectroscopy (LIBS) system, developed with Alakai Consulting an d Engineering, was used where a 10.6 m CO2 laser was used, for the first time to our knowledge, to enhance a remote LIBS plasma formed by a 266 nm Nd:YAG laser focused onto a distant target at a range of about 35 m or more using eye safe wavelengths. La ser pulse energies on target were about 0.05 J/mm2 for both lasers, with LIBS emission depending upon both the Nd:YAG and the CO2 laser pulse intensities. An enhancemen t in the LIBS emission on the order of 10x to 100x was observed. In addition, an incr ease in the LIBS plasma temperature of about 3,000-5,000 K was measured for the firs t time to our knowledge. The temporal overlap of the two lasers wa s found to be optimized for a 0.5 s to 1.5 s delay. Related studies were conducted to analyze th e size and volume of the LIBS crater, and to measure the LIBS/TEPS emission from se veral other target substrates including copper, lead, and in an argon atmosphere. Finally, initial Raman lidar measurements were conducted in order to compare this re mote sensing technique with that of the LIBS/TEPS lidar. However, the Raman lidar measurements were tentative and further work is required to better quantify this data. The organization of this thesis is as follows. Chapter 2 discusses the basic theory and background information of the va rious laser detection and lidar techniques used in this dissertation for the remote se nsing of chemical species. Chapter 3 discusses the setup of our laboratory CW tunable CO2 Laser Differential Absorption
4 lidar (DIAL) system for the remote sensi ng of Triacetone Trip eroxide (TATP) gas vapors. The DIAL laser beam was transmitted through an enclosed absorption cell containing TATP or SF6, and backscattered returns we re measured from a retroreflector array target at ranges of 5 m to 100 m. Concentrat ion of known amounts of SF6 and TATP were measured and our results are discussed. Chapter 4 describes the stand-off TEPS LIBS system in detail. Single pulse LIBS by a 266 nm Nd:YAG laser or a 10.6 m CO2 laser were studied and compared with the LIBS spectra obt ained by simultaneous use of the CO2 laser with the 266 nm Nd:YAG laser. The eff ects of optimum delay between the two laser pulses along with the two different overla p geometries were studied. Chapter 5 describes the LIBS induced crat er studies on the target surf aces as a function of range along with LIBS/TEPS studies in the presen ce of Argon gas environment. Of major importance is our study of the electron plasma temperature using the Boltzmann method, which is presented in Chapter 6. Significant temperature increases in the plasma due to the TEPS process were measured for the first time on various substrates in air and in an Argon atmosphere. El ectron density measurements were also calculated using the Stark broadening met hod. Chapter 7 describes the experimental setup for multiple pulse LIBS/TEPS measur ements where dual pulse 266 nm Nd:YAG laser LIBS signals were enhanced by the addition of a 10.6 m CO2 laser. Tentative elemental identifications ware made from the LIBS/TEPS spectrum on various substrates. A peak ratio analysis method wa s used in an attempt to identify impurities and unknown species present in the known s ubstrates. Chapter 8 describes the experimental setup for a 266 nm Nd:YAG laser standoff Raman lidar system which is
5 being studied as an additional remote sensi ng technique for trace species. Chapter 9 discusses some of the possible mechanisms underlying the plasma interactions which are responsible for the enhancement of the emission signal due to the addition of the 10.6 m CO2 laser pulse on to the LIBS plasma formed by the 266 nm Nd:YAG laser. Conclusions and future research are discussed in Chapter 10. Finally it may be noted that the measur ements and results presented in this dissertation are a result of th e authors research, but also involved collaboration with other groups in a interdiscipl inary research program involvi ng several research groups. To be more specific of the authors contri bution, it may be noted that all the DIAL TATP research presented was conducted by the author at USF, but noting Prof. Michael Sigman and Dr. Douglas Clark who synthesized the TATP samples in our USF DIAL laboratory and performed the FT-IR absorption and Ion Mobility Spectroscopy Measurements. The TEPS/LIBS lidar experiments invol ved a collaborative research program and ARL grant with Alakai Inc and the Florid a High Tech Matching grant. As part of this, the author spent an extended graduate course Industrial Practicum and conducted subsequent research at the Al akai laser laboratory in La rgo, Florida. The author worked on the development and construction of the 266 nm LIBS/CO2 TEPS system, and performed experiments on remote targ ets with Rob Waterbury and Ed Dottery (Alakai Inc). In particul ar, the author collected the LI BS/TEPS data and analyzed the enhancement performed the LIBS crater measurements, performed all of the temperature measurement studies outlined in this dissertation, and worked on the
6 development of the initial Raman lidar stud ies (at Alakai). Another research group from Florida A&M University (Dr. Lewi s Johnson) supplied the second Nd:YAG laser used in the dual laser LIBS measurements. It may be added that additional studies have also been conducted by these collaborative groups including on-going LIBS/TEPS remote stand-off detection of explosives (Alakai Inc.), on-going femtosecond laser LIBS with CO2 TEPS enhancement (Florida A & M University), and CO2 plasma propagation studies (Arkansas State University). The results of these studi es will be published by these groups later, and the author will be a co-author on these papers. However these particular results are not pr esented in this dissertation.
7 CHAPTER 2. BACKGROUND INFORM ATION ON REMOTE SENSING BY DIAL, LIDAR, LIBS AND RAMAN SPECTROSCOPY In the broadest sense, remote sens ing is the measurement or acquisition of information of an object or phenomenon, by a recording device that is not in physical or direct contact with the obj ect. In practice, remote sens ing is the utilization at a distance (as from aircraft, sp acecraft, satellite, or ship) of any device for gathering information about the environment. Thus an aircraft taking photographs, weather satellites, and space probes are all examples of remote sensing. The main criteria for remote sensing are presence of an unique ab sorption or emission spectrum of a species compared to that of an interfering ba ckground, a transmission windows of atmosphere, and the availability of detectors, spectral de vices and lasers at a wavelength equal to the spectral signatures of the species. There are essentially two kinds of re mote sensing: Active remote sensing and Passive remote sensing. Both of these are covered in the following. Passive sensors detect natural energy (ra diation) that is emitted or reflected by the object or scene being observed. Reflected sunlight is the mo st common source of radiation measured by passive sensors. The detection of a gas is often made by measuring the absorption/emission spectrum of the gas. Figure 2.1 shows a typical experimental passive remote sensing setup. As can be seen in the figure a telescope is directed towards the target of interest and th e light collected from the telescope is then
Figure 2.1 Typical setup for passiv e remote sensing using absorption spectroscopy. 8
9 analyzed by a spectrometer to determine the presence and concentration of the elements and molecules present in the target. An active sensor emits radiation in the direction of the target to be investigated. The sensor then detects and m easures the radiation that is reflected or backscattered from the target. Figure 2.2 shows a typical experimental setup for active remote sensing. Microwave radar and laser radar (LIDAR) are widely known forms of active remote sensing. There are several different forms of active Lidar systems that have been used for remote se nsing, and some of thes e are covered in the following sections. 2.1 Description of Lidar A typical Lidar system uses a pulsed or cw laser to propagate a beam through the atmosphere. The laser beam is then re flected or scattered fr om hard targets or aerosol particles in the atmosphere. A frac tion of the light scatte red by the target is then collected by a telescope and focused onto a detector. The signal from the detector is then analyzed with the ai d of high speed electronics to give information about either the target itself or the atmosphere betw een the target and the receiver. The backscattered light from a hard target occurs delayed in time due to the speed of light in the atmosphere (about 1ns for 1ft of distan ce). The backscattered light from aerosols occurs due to the spatial di stribution of aerosols in the atmosphere, and produces a range resolved backscattered return that is determined by the pulse length of the transmitted laser pulse. In general,
Figure 2.2 Typical setup fo r active laser remote sensing. 10
the back scattered Lidar return s from aerosols in the atmosphe re is about six orders of magnitude less than that from a hard target. 2.1.1 Hard Target Lidar Equation This section discusses the lidar equa tion used to theoretically calculate the laser return signal for a hard target lidar system The lidar signal or power returned or backscattered from a hard target, Pr, may be described by the Lidar Equation as1,4 2 )(2t r0 ''e KA F(r) P P r r Tdrr (2.1) where Pt is the laser power transmitted, F(r) is the optical overlap factor of the transmitted beam/telescope field of view, A is the telescope area (m2), K is the system optical efficiency, is the hard target reflectivity, r is the target range (m), and T(r) is the total absorption coefficient (km-1) of the atmosphere at range r. As can be seen in the above equation th e return power is essentially the fraction of reflected light b ackscattered into a hemi sphere of area 2 r2 and collected by the telescope area, A. The return power is proportional to the transmitted power and the telescope collecting area, and decreases with range as 1/r2 and the Beer Lambert Law. 11
12 In the above equation the Beer Lambert attenuation, exp(r ), is integrated over the entire two way optical path. 2.2 Differential Absorption Lidar (DIAL) Differential Absorption lid ar or DIAL systems are typically used to measure the concentration of a gas in the atmosphere.1-4 DIAL works by taking the ratio of lidar signals taken at two or more different wavelengths and comparing the difference with the expected differential absorption valu es. Wavelengths are specifically chosen based on the species to be detected such that at one wavelength, called the onresonance wavelength, is tuned to where the species absorbs strong ly, and for the other wavelength, called the off-resonance wavele ngth, the species ha s low absorption. The absorption cross section for the particul ar absorption line bei ng studied is normally measured by using the lidar / laser beam and passing it thro ugh a gas cell containing a known amount of the gas being measured; prev iously measured absorption coefficients may also be used. Figure 2.3 shows a typical Differential Absorption Lidar setup. As can be seen, a line tunable lase r is used to send out beams at two DIAL wavelengths to the open atmosphere containing a cloud having th e chemicals which need to be sensed. Part of the laser beam is also sent to an absorption test cell in the lab with a known gas concentration present. The backscattered return signal is collected by a telescope onto a detector and analyzer.
Figure 2.3 Typical schematic of DIAL remote sensing equipment. 13
14 The molecular absorption can further be defined in terms of the molecular absorption cross section, mol (cm2), and gas density, N (molecules/cm3) from the Beer Lambert equation. The species concentration can be given by1,2,4 Na= ( ln( Pon/Poff ) ) / (2 aR) (2.4) where Na is the concentration of the species (molecules/cm3), Pon and Poff are the returned signal from the online and offline wavelengths, R is the range, and a is the absorption cross section of the species to be measured. With the DIAL technique, one can in vestigate pollutants in both the free atmosphere and in polluted areas, such as cities near industrial plan ts. The differential absorption technique can be extremely sensitive and is able to detect gas concentrations as low as a few hundred parts per billion ( ppb). This makes it possible to measure trace pollutants in the ambient atmosphere a nd monitor stack emissi ons in the parts per million ranges. Range-resolved DIAL systems are sensitive enough to measure the ambient air concentrations a nd distribution of most of the major polluting gases, including SO2, NO2, NO, and ozone. This technique makes it possible to obtain vertical profiles of the atmospheric gas c oncentrations from ground, airborne, or space platforms.1-6 As an example of DIAL measurements of atmospheric gases, Figure 2.4 shows a comparison of the concentration of CO2 measured in the atmosphere by a DIAL lidar at a range of 4.2 km and by an in-situ probe. As can be seen, the DIAL measurement is very close to that of the in-situ probe.3
Figure 2.4 DIAL detection of atmospheric CO2 compared to in-situe gas analyzer measurements. DIAL measurements are the blue doted data and in-situe are the green (grey) so lid line. (from G. Koch3; reprinted with permission from Applied Optics). 15
2.3 Lidar S/N and Detector Noise Lidar systems operating at wavelengths in the near to far IR (1 to 10 m) require the use of various photodiodes and phot oconductors as optical detectors. The noise of such detectors often determines the noise of the Lidar syst em. In general, the S/N in a detection system may be given as7 2 2/n Si i NS (2.5) where is the mean square signal current and is the mean square noise current. 2 Si 2 ni The system noise may arise from seve ral sources includi ng dark current, idc, Johnson (thermal) noise, ij and amplifier noise, ia. In general, the noise terms add independently and the total mean square system noise, , is given by (2.6) ...........2 2 2 2a dc j niiii In most IR Lidar systems, the major sour ce of noise is due to IR radiation from the 300 K background falling on the detector For such systems a good approximation of the Noise Equivalent Power (NEP) may be calculated from the Detectivity, D*, area, and 16
electronic bandwidth of the detector/ pr e-amp. The NEP of a background limited photoconductor is given by7 *D BA NEPd (2.7) where Ad is the detector area (cm2), B is the bandwidth (Hz), and D* is the detector sensitivity (cm Hz1/2 Watt-1). In this case the S/N is usually given as the ratio of the returned lidar signal, Pr and NEP, so that S/N = Pr / NEP (2.8) 2.4 Review of Absorption Spectroscopy Theory This section reviews the equations used to translate the observed transmittance spectra into spectroscopic parameters. Light transmitted through a medium is absorbed when the frequency of the photon is resonant with one or a combination of the medias energy levels. The amount of ab sorption that occurs is proportional to the quantity of the material present, the populat ion distribution within the energy states involved with the transmission, and how closel y the photons frequency is to that of the optical transition. The equation that describes the transmittance of monochromatic light through a medium may be expressed by the Beer Lambert Law in different forms, such as,8 I(v,L) = Ioe(v) P z = Ioe(v) N z (2.8) where Io is the incident intensity, I (v, L) the intensity of light after passing through the medium, (v) is the absorption coefficient (cm-1), P is the partial pressure of the 17
gas (atm), z is the optical path length (cm), (v) is the absorption cross section (cm2), and N is the concentration of the molecule (molecules / cm3) The transmission, T, of the beam is given by I / I0, the absorption, A, is 1 T, and the absorbance, Ab, is log 10 ( T ). The absorption coefficient can be related to the molecular line intensity parameters as (v) = S g(v-vo) N (2.12) where S is the line intensity ( cm-1/( molecule .cm-2)), g(v) is the line shape of the molecular transition (cm) and N is the total number densit y of absorbing molecules per unit volume per atmosphere. The line intensity, S of the transmission is given by 1 2 1] 296 )[ ( T N NNB c h SL (2.13) where, B is the Einstein coefficient describing the rate of transitions between two energy levels, N1 N2, is the population difference between the two energy levels, T is the temperature of the medium (K), v is the transition frequency (rad/s), h is Plancks constant, c is the speed of light, and NL is Loschmidts number, the molecular density of the atmosphere at standard temperature and pressure (NL = 2.479 x 1019 molecules cm-3atm-1). The first term of the equation is a te mperature scaling term that results from Boltzmann population changes in the energy states as a function of temperature, and the second term is due to changes in the gas density as a result of temperature changes. The absorption line shape term in Eq. (2.12) is usually described by one of the three line-shape functions, either Gaussian, Lorent zian, or a Voigt profile. The choice of 18
functions is determined by the total pressure of the medium. The Gaussian lineshape function, gD, also known as the Doppler distribution, is used in low pressure atmospheres in which the linew idths are Doppler broadened. At higher atmospheric pressures where pressure broadening dominate s, the transition lineshape is described by a Lorentzian line shape. At intermediate pressures, the line shape is described by a Voigt profile and is an integral over both prof iles. At atmospheric pressure (1atm or 760 Torr), the Lorentian pressure broadened lineshape profile is usually given by10 ])[( )( )(22 0 P o P Pg (2.14) where p is the pressure broadene d line width (HWHM), and vo is the resonance frequency for the transition. In this case, p can also be given as t n PP T g ) 296 ( (2.15) where g is the pressure broa dened linewidth parameter, Pt is the total background gas pressure, T is the temperature, and n is the temperature scalin g exponent.. In atmospheric gases in which constituent concentrations are relatively small, the value used for g is usually the air broadened linewidth parameter. This value is primarily influenced by molecular collisions with at mospheric nitrogen molecules. The final term, N, of Eq. (2.12) is the total number of molecules per cm3 per atmosphere. This term is the temperature adjusted Loschmidts number given by ) 296 ( T NNL (2.16) 19
20 where T is the temperature in degrees Kelvin. 2.5 Introduction to Laser Induced Breakdown Spectroscopy (LIBS) Laser-induced breakdown spectroscopy (L IBS) is a type of atomic emission spectroscopy which utilizes a highly energetic laser pulse as the excitation source. LIBS can analyze most matter regardless of its physical state, be it solid, liquid or gas. Even slurries, aerosols, gels, and more can be readily investigated. Because all elements emit light when excited to suffici ently high temperatures, LIBS can detect many elements, limited usually by the power of the laser as well as the sensitivity and wavelength range of the spectrograph and de tector. The LIBS method makes use of optical emission spectrometry and is very similar to arc/spark emission spectroscopy.11,12 LIBS operates by focusing the laser onto a small area at the surface of a specimen. A schematic of a common LIBS set up is shown in Fig. 2.5. When a highpower laser pulse is focused onto a target surface, it can ab late a very small amount of material, in the range of nanograms to pic ograms,due to intense heating by the laser pulse to produce a plasma plume w ith temperatures of about 10,000,000 K.13 At these temperatures, the ablated material can dissociate into excited ionic and atomic species. During this time, the plasma emits a continuum of radiation due to the blackbody radiation of the heated plasma. Within a short timeframe the plasma expands at supersonic velocities and cools. At this point the characteristic atomic emission lines of the elements can be observed.
Figure 2.5 Typical LIBS setup 21
22 Because such a small amount of material is consumed during the LIBS process the technique is considered essentially non-destructive or minimally-destructive, and with an average power density of less than one watt radiated onto the specimen there is almost no specimen heating surrounding the ablation site. One of the major advantages of the LIBS technique is its ability to depth profile a specimen by repeatedly discharging the laser in the same position, effectively going deeper into the specimen with each shot. This can also be applied to the removal of surface contamination, where the laser is discharged a number of times prior to the analyzing shot. LIBS is also a very rapid technique giving results wi thin seconds, making it particularly useful for high volume analyses or on-line industrial monitoring. The use of specialized optics or a mechanically positioned specimen stage can be used to raster the laser over the surface of the specimen allowing spatially resolved chemical analysis and the creation of 'eleme ntal maps'. This is very significant as chemical imaging is becoming more impor tant in all branches of science and technology. One of the disadvantages, for the LIBS, how ever, is variability in the laser spark and resultant plasma which often limits reproducibility. The accuracy of LIBS measurements is typically better than 10% a nd precision is often better than 5%. The detection limits for LIBS vary from one element to the next, depending on the specimen type and the experimental apparatus used. Even so, detection limits of 1 to 30 ppm by mass are common.
23 2.6 LIBS by Different kinds of Lasers The use of different types of lasers has been studied for LIBS detection. In particular, the plasma physics can be different for different laser pulse lengths and energy levels.11-14 2.6.1 Femtosecond Laser Ablation When a femtosecond laser pulse inter acts with a solid sample, different electronic mechanisms are excited, depending on the sample material. For conducting samples, free-electrons inside the solid can directly absorb laser energy and form a hot electron-hole plasma. For semiconductor a nd wide bandgap dielectr ics, the electronhole plasma is created thr ough nonlinear processes such as multi-photon absorption and ionization, tunneling, and av alanche ionization. At hi gh energy, the electron-hole plasma created on the surface of the solid will induce emission of x-rays. Hot electrons and photoemission can produce highly charged ions through a phenomenon called Coulomb explosion or non-thermal melting. Once electron-hole plasma is formed inside the solid, the carriers can absorb additional laser pho tons, sequentially moving to higher energy states. When a femtosecond laser pulse is focused in air, optical emission of nitrogen molecular lines will ex ist in the air plasma several picoseconds after the laser pulse; usually, the molecular structure is preserved for this time period. Overall, the femtosecond pulsed laser induced plasma has a shorter overall lifetime compared to those plasmas initiated using longer laser pulses.
24 2.6.2 Picosecond Laser Ablation For picosecond laser ablation, the lattice can be melted through thermal and/or non-thermal processes, depending on the laser irradiance. Electrons are ejected from the target surface during the laser pulse. The free electr ons can interact with the air and absorb laser energy to initiate air plasma during the ps lase r pulse duration. The plasma forms long before the plume forms. The air plasma above the sample can absorb a part of the incoming laser beam radiation. Unlike ns laser ablation, plasma shielding is not caused by absorption from the vapor plume; on the picosecond time scale, plasma shielding is caused by the air plasma. 2.6.3 Nanosecond Ablation When the pulse duration is on the or der of a few nanoseconds, and the laser irradiance is on the order of 107-1011 W/cm2, the mechanisms involved in ablation are melting, fusion, sublimation, vaporization and ionization. If the irradiance is high enough, non-thermal ablation is also important and can co-exist with these thermal mechanisms. For the ns pulsed laser induced plas ma, continuum emission appears during the laser pulse and lasts for several hundred nanoseconds. Ion emission also dominates on the ns timescale. Atomic and molecular lin e emission occurs after about 1 microsecond. Molecular line emission measured at later time s is from the recombination of species in the plasma. As an example of the typi cal line emission from a LIBS plasma,
Figure 2.6 shows a typical LIBS spectra obtained from Pb.15 As can be seen, strong emission lines associated with optical transi tions within the neutral Pb atom, usually notated as Pb (I), and from the single-ionized Pb ion, usually notated as Pb (II), are easily measured. When the laser irradiance is less than 108 W/cm2, thermal processes are dominant. The temperature at the target surface will rise during the laser pulse and eventually the target may melt and vaporize. The temperature distribution in the target can be calculated with heat conduction equation 12 ),( ] ),( )[( ),( txI Cx txT Cxt txTsp sp (2.17) where T represents the temperature inside the ta rget, x is the position from the surface, k, Cp, s and denote the thermal conductivity, heat capacity, mass density and absorption coefficient of the solid target material respectively. Finally, as an illustration of the difference in the different pulse-width, Fig. 2.7 shows a LIBS emission spectra from an alumin ium substrate using a 75 fs laser pulse, a 200 ps laser pulse and 5ns laser pulse for Fig. 2.7 (a), Fig. 2.7 (b), and Fig. 2.7 (c), respectively; here the range to the sample target was 25 m.16 As can be seen, the LIBS spectra is different in some aspects acco rding to the different physics processes occurring. This is explained in more detail in the next section. 25
Figure 2.6 Typical LIBS spectrum from me tallic Pb showing neutral Pb emission Pb(I). Lines due to single ionized Pb, Pb (II), are also marked (fromV. Babushok15; reprinted with permission from Applied Optics). 26
Figure 2.7 Typical LIBS sp ectra of Al at 25 m range with delay of 345 ns and gatewidth of 10 s on ICCD camera using a (a) Femtosecond laser, (b) Picosecond laser, (c) Nanos econd laser (from P.Rohwetter16; reprinted with permission from J.Anal. At Spectrom). 27
28 2.7 Time evolution of the LIBS signal The LIBS emission spectra depend on th e laser-induced plasma properties. Figure. 2.8 shows a typical timeline for th e emission of different species in LIBS emission spectra. At the earliest time, the plasma light is dominated by a background continuum that has little intensity va riation as a function of wavelength.11 This is caused by Bremsstrahlung and recombinati on radiation from the plasma as free electrons and ions recombine in the cooling plasma. If this plasma background is integrated over the entire emission time of the plasma, this continuum background can seriously interfere with the detection of weaker emissions from minor and trace elements in the plasma. For this reason, LIBS measurements are sometimes carried out using time-resolved detection. In this way the strong white light at early times can be removed from the measurements by turning the detector on after this white light has significantly subsided in intensity but when the atomic emissions are still present. As can be seen in Fig. 2.8 there is a strong c ontinuum emission for the first 100 ns after the laser pulse followed by ions, neut rals and molecular emission. As an experimental example, Fig.2.9 shows the evolution of the LIBS spectrum from a soil sample.11 Several important features shoul d be noted. First the significant decrease in line widths as th e delay time changes from 0 to 7 s. This is particularly evident in the two strongest lines (single-ioni zed Ca) on the left side of the figure. Second, as the line widths decreases, it becomes evident at td = 0.5 s that two additional lines (neutral Al between the Ca lines) appear that were masked by the strong Ca lines.
Figure 2.8 Typical timing delay after th e plasma formation for the emission of differentspecies in LIBS spectra. (adapted from A.Miziolek11) 29
Figure 2.9 Typical LIBS emission spectra as observed at different timing delays. (adapted from A.Miziolek11) 30
31 Comparison of the relative intensities of the Ca and Al lines shows that these change as the plasma cools with the si ngle-ionized Ca lines decreasing more in comparison with the neutral Al lines with in creased delay time. These same features are evident in the expanded portions of the spectrum (400 nm 410 nm) displayed on the right side of the figure showing Fe and Sr lines. A life cycle schematic for laser-i nduced plasma on a surface is shown schematically in Figure 2.10. Briefly, th ere are two steps leadi ng to breakdown due to optical excitation.11 The first involves having or ge nerating a few free electrons that serve as initial receptors of energy through free body co llisions with photons and neutrals. The second is avalanche ionizati on in the focal region Classically, free electrons are accelerated by the electric fields associated w ith the optical pulse in the period between collisions, which act to therma lize the electron energy distribution. As the electron energies grow, collisions produce ionization, other electrons, more energy absorption, and an avalanche occurs. In the photon picture, absorption occurs because of inverse Bremsstrahlung. The breakdown threshold is usually specified as the minimum irradiance needed to generate visible plasma. Following breakdown, the plasma expands outward in all directions from the focal volume. However, the rate of expansion is greatest towards the focusing le ns, because the optical energy enters the plasma from that direction. A cylindrical shaped appearance results from this non isotropic expansion. The initial rate of plasma expansion is on the order of 105 m/s. The loud sound that one hears during a LIBS experiment is caused by the shockwave coming from the focal volume.11
Figure 2.10 Life cycle diagram of LIBS, showing the progression of events (adapted from A. Miziolek11) 32
332.8 Raman Spectroscopy The Raman optical effect occurs wh en light impinges upon a molecule and interacts with the electron cloud of the bonds of that molecule. The incident photon excites one of the electrons in to a virtual state. For the spontaneous Raman effect, the molecule will be excited from the ground state to a virtual energy stat e, and relax into a vibrational excited state, wh ich generates Stokes Raman s cattering. If the molecule was already in an elevated vibrational energy state, the Raman scattering is then called anti-Stokes Raman scattering. A molecu lar polarizability change or amount of deformation of the electron cloud, with respec t to the vibrational c oordinate is required for the molecule to exhibit the Raman effect The amount of the polarizability change will determine the Raman scattering intensity, whereas the Raman shift is equal to the vibrational level that is involved.17,18 Although inelastic s cattering of light was predicted by Smekal in 1923, it was not until 1928 that it was observed in practice. The Raman effect was named after one of its discoverers, the Indian scientist Sir C. V. Raman who observed the effect by means of sunlight in 1928, together with K. S. Krishnan, and independently by Grigory La ndsberg and Leonid Mandelstam. Raman won the Nobel Prize in Physics in 1930 for this discovery accomplished using sunlight, a narrow band photographic filter to create monochromatic light and a "crossed" filter to block this monochromatic light. He f ound that light of changed frequency passed through the "crossed" filter. Subsequently the mercury arc became the principal light source for Raman experiments, first with photographic detection and then with spectrophotometric detection. Currently, lasers are usually used as the Raman excitation light source.
34 The scattering of monochromatic radiation that is incident on a sample can tell the Raman spectroscopist something of the sa mple's molecular structure. If the frequency (wavelength) of the scattered radi ation is analyzed, not only is the incident radiation wavelength observed via elastic s cattering (Rayleigh scattering), but also a relatively small amount of radiation is scattered inelastically at different wavelengths corresponding to the the specifi c energy levels of the molecule. Approximately only 1 x 10-7 of the scattered light is inelastically Raman scattered. The scattered radiation occurs in all directions and may also have observable chan ges in its polarization along with its wavelength. It is the energy shift in wavelength of the inelastically scattered radiation that provides the chemi cal and structural information.17-19 Figure 2.11 shows a simple diagram to il lustrate the energy level diagrams for the Raman signal.17 Stokes radiation occurs at lo wer energy (longer wavelength) than the Rayleigh radiation, and anti-Stokes radiation has greater energy. The energy increase or decrease is related to the vibr ational energy levels in the ground electronic state of the molecule, and as such, the obs erved Raman shift of the Stokes and antiStokes features are a direct m easure of the vibrational energies of the molecule. The energy of the scattered radiation is less than the incident radiation for the Stokes line and the energy of the scattered radiation is mo re than the incident radiation for the antiStokes line. The energy increase or decrea se from the excitation is related to the vibrational energy spacing in the ground electronic state of the molecule and therefore the wavenumber of the Stokes and anti-Stok es lines are a direct measure of the vibrational energies of the molecule.
Figure 2.11 Energy level diagram showing th e states involved in Raman spectroscopy. 35
36 In Fig. 2.11, it can be noted that St okes and anti-Stokes lines are equally displaced from the Rayleigh line. This occu rs because in either case one vibrational quantum of energy is gained or lost. It can be noted also that anti-Stokes line is much less intense than the Stokes line. This occurs because only molecules that are vibrationally excited prior to irradiation can give rise to the anti-Stokes line and the Boltzmann population of the upper energy levels is less than that of the ground state. Hence, in Raman spectroscopy, only the more intense Stokes line is normally measured. Raman scattering is a relatively w eak process. Figure 2.12 shows a typical experimental Raman spectroscopy setup.19 Figure 2.13 shows a typical Raman spectrum obtained from a naphthalene sample which is often used as a calibration sample because it has a large Raman cross section.20 2.9 Different Raman Techniques There are several different, yet related, Raman laser sp ectroscopic techniques. Some of these different techniques are de scribed here for background information. 2.9.1 Resonance Raman If the wavelength of the ex citing laser coincides with an electronic absorption of a molecule, the intensity of Raman-active vi brations associated with the absorbing species are enhanced. This resonance enha ncement or resonance Raman effect can be extremely useful, not just in significantly lowering the detecti on limits, but also in introducing electronic selectivity.
Figure 2.12 Typical Continuous Wa ve (CW) Raman spectroscopy layout. 37
Figure 2.13 Typical Raman spectra for Napthalene,(adapted from http://www.chemistry.ohiostate.edu/~ rmccreer/freqcorr/images/nap.html Richard L. McCreery, The Ohio State University20) 38
39 Thus the resonance Raman technique is used for providing both structural and electronic insight into a species of intere st. Resonance selectivity has a further practical use, in that the spectrum of the chemical of biological chromophoric majority is resonance enhanced and that of the surr ounding environment is not. For biological chromophores, this means that absorbing active centers can be specifically probed by visible excitation wavelengths, and not the surrounding protein matrix (which would require UV lasers to bring into resonance). Resonance Raman spectroscopy is also an important probe of the chemistry of metal centered complexes, fullerenes, polydiacetylenes, and other "e xotic" molecules which strongl y absorb in the visible. Although many more molecules absorb in the ul traviolet, the high cost of lasers and optics for this spectral region have lim ited ultraviolet (UV) resonance Raman spectroscopy to a small numb er of specialist groups. Vibrations which are resonantly enhanced fall into two or three general classes.1719 The most common case is Franck-Condon e nhancement, in which a component of the normal coordinate of the vibration occu rs in a direction in which the molecule expands during an electronic excitation. The more the molecule expands along this axis when it absorbs light, the larger the enhancement factor. The easily visualized ring breathing (in-plane expansion) modes of porphyrins fall into this class. Vibrations which couple two electronic ex cited states are also res onantly enhanced, through a mechanism called vibronic enhancement. In both cases, enhancement factors roughly follow the intensities of the absorption spectrum. Resonance enhancement does not begin at a sharply defined wavelength. In fact, enhancement of 5x to 10x is observed if
40 the exciting laser is within even a few 100 wavenumbers below the electronic transition of a molecule. This "pre-resonance" e nhancement can be experimentally useful. 2.9.2 Surface Enhanced Raman Scattering (SERS) Raman scattering from a compound (or ion) adsorbed on or even within a few Angstroms of a structured metal surface can be 100x greater than that in solution. This surface-enhanced Raman scattering (SERS) is st rongest on silver, but is observable on gold and copper as well. SERS arises from two mechanisms: The first is an enhanced electromagnetic field produced at the surface of the metal. When the wavelength of the incident light is close to the plasma wave length of the metal, conduction electrons in the metal surface are excited in to an extended surface electr onic excited state called a surface plasmon resonance. Molecules adsorbed or in close proximity to the surface experience an exceptionally large electroma gnetic field, vibrational modes normal to the surface are most strongly enhanced. Th e second mode of enhancement is by the formation of a charge-transfer complex betw een the surface and analyte molecule. The electronic transitions of many charge transfer complexes are in the visible, so that resonance enhancement occurs. Molecules w ith lone pair electrons or pi clouds show the strongest SERS. The effect was first discovered with pyridine, other aromatic nitrogen or oxygen containing compounds, such as aromatic amines or phenols, are strongly SERS active. The effect can also been seen with other electron-rich functionalities such as carboxylic acids. The intensity of the surface plasmon resonance is dependent on many factors incl uding the wavelength of the incident light and the morphology of the metal surface. The wavelength should match the plasma
41 wavelength of the metal. This is about 382 nm for a 5 m silver particle, but can be as high as 600 nm for larger ellipsoidal silver pa rticles. The plasma wavelength is to the red of 650 nm for copper and gold, the ot her two metals which show SERS at wavelengths in the 350-1000 nm region. The best morphology for surface plasmon resonance excitation is a small (<100 nm) pa rticle or an atomically rough surface. SERS is commonly employed to study monolay ers of materials adsorbed on metals, including electrodes. Other popular surfaces include colloids, meta l films on dielectric substrates and, recently, arrays of metal pa rticles bound to metal or dielectric colloids through short linkages. 2.9.3 Ultraviolet Resonance Raman Spectroscopy (UVRRS) UVRRS is a powerful tool in the molecular analysis of complex biological systems. Most biological systems absorb UV radiation and hence have the ability to offer resonance with UV Raman excitation. This results in the highly selective resonance Raman effect enabling enhancement of important biological targets such as protein or DNA. For example, excitation around 200 nm enhances the Raman peaks from vibrations of amide groups; excita tion around 220 nm enhances peaks from certain aromatic residues. The Raman scatter from water is weak, allowing for analysis of very weak aqueous systems. Due to the selective nature of UVRRS, a tunable laser is typically required as the excitation source. Since truly tunable co ntinuous-wave lasers are not yet available, a Nd:YAG-pumped dye laser with frequencydoubled output is one suitable UVRRS system. Depending on the dyes used, this la ser setup can give almost any required UV
42 wavelength. Intensified CCDs (ICCDs) w ith UV photocathodes, back-illuminated CCDs or CCDs can be used as detector s for UVRRS. These detectors are used on account of their high detection efficien cy and multichannel capabilities. 2.9.4 Time-Resolved and Pulsed Raman Vibrational Raman spectroscopy is no w widely recognized as a useful technique for chemical analysis of stable species, since the technology which underpins Raman measurements has matured. Similarly, time-resolved Raman spectroscopy has also become established as an excellent method for the ch aracterization of transient chemical species. Many of the technical advances which have reduced the cost and increased the reliability of conventional Rama n systems can also be exploited in studies of transient species. Pulsed lasers are typically utilized in the study of short-lived species. A laser pulse can be supplied to a molecular syst em with enough energy to redistribute the electrons in a molecule causing the formation of an excited state as illustrated on the right. The Raman spectrum of this excited st ate molecule can be studied either using the same laser pulse or a different pulse fr om a second laser (single color and two-color pulsed Raman). Excited states of interest can have lifetimes, from picoseconds to milliseconds, but the majority can be studied us ing gating in the order of 5 ns. As the majority of excited states are generated using UV and visible lasers, a photocathode with high UV and visible quantum e fficiencies are typically used. In Time Resolved Resonance Raman sp ectroscopy, pairs of laser pulses of different wavelength are used to optically pump and then to Raman probe the transient
43 species of interest. The spectral window of the spectrograph/detector is chosen so that it corresponds to the frequenc y range of the Raman scattering from the probe laser. The time evolution of the transient signal is monitored by recording a series of spectra at different delays after the phot olysis event, i.e. at a seri es of time delays between the excitation and probe pulses. The ICCD camera (Andor model DH720-18F-03 for instance) or either of the lasers can supply the trigger. A delay generator is used to controls the delays. 2.9.5 Confocal Raman Microscopy In Raman microscopy, a research grade optical microscope is coupled to the excitation laser and the spect rometer, thus producing a platform capable of obtaining both conventional images and in addition ge nerating Raman Spectra from sample areas approaching the diffraction limit (~1 micron). Imaging and spectroscopy can be combined to generate "Raman cubes", 3dimensional data sets, yielding spectral information at every pixel of the 2D image. A motorized xyz microscope stage can be used to automatically record spectral file s, which will constitute the basis of Raman images, Raman maps, or a set of Raman spect ra recorded from pre-selected points. Specific software routines will allow the quick and easy reconstruction of these maps. The possibility of generating two-dimensi onal and three-dimensional images of a sample, using various special features, is an evident advantage over either traditional spectroscopy or microscopy.
44 2.10 Some Raman Spectroscopy Applications Raman spectroscopy is commonly used in chemistry, since vibrational information is very specific for the chemical bonds in molecules. It therefore provides a fingerprint by which the molecule can be id entified. The fingerprint region of organic molecules is in the range 500-2000 cm-1. 22 Another way that the technique is used is to study changes in chemical bonding, e.g., wh en a substrate is added to an enzyme. Raman gas analyzers have many practical applic ations. For instance, they are used in medicine for real-time monitoring of anesth etic and respiratory gas mixtures during surgery. In solid state physics, spon taneous Raman spectroscopy is used to, among other things, characterize materials, measur e temperature, and find the crystallographic orientation of a sample. The spontaneous Raman signal gives information on the population of a given phonon mode in the ra tio between the Stokes (downshifted) intensity and anti-Stokes (upshifted) intensity. Raman active fibers, such as aramid and carbon, have vibrational modes that show a shift in Raman frequency with applie d stress. Polypropylene fibers also exhibit similar shifts. Spatially Offset Raman Spect roscopy (SORS), which is less sensitive to surface layers than conventional Raman, can be used to discover counterfeit drugs without opening their internal packaging, a nd for non-invasive monitoring of biological tissue.21-24 Raman spectroscopy can be used to investigate the chemical composition of historical documents and contribute to knowledge of the social and economic conditions at the time the documents were produced.23 This is especially helpful
45 because Raman spectroscopy offers a non-invasive way to determine the best course of preservation or conservation treatment for su ch materials. In nanotechnology, a Raman Microscope can be used to analyze nanow ires to better understand the composition of the structures. Finally Raman spectroscopy can be join ed with remote sensing of lidar techniques for Raman detection of remote trace species. Figure 2.14 shows such Raman Lidar data obtained for the remote de tection of a sample of ethyl benzene and a sample of toluene at a range of 10 m away.21 2.11 Laser Wavelength Selection and Eye Safety In laser remote sensing it is of ten important to choose a laser wavelength that is considered more eye safe usually using wavelengths in the UV (say below 300 nm) or in the mid-IR (say beyond 1.4 microns). The maximum permissible exposure (MPE) is the highest power or energy density (in J/cm2 or W/cm2) of a laser source that is considered safe and has a negligible probability for creating damage. The MPE is measured at the cornea of the human eye or at the skin, for a given wavelength and exposure time. A calculation of the MPE for ocular exposure takes into account the various ways light can act upon the eye. Co llimated laser beams of visible and nearinfrared light are especially dangerous at relatively low powers because the lens focuses the light onto a tiny spot on the reti na. Although the MPE is specified as power or energy per unit surface, it is based on the power or energy that can pass through a fully open pupil (0.39 cm2) for visible and near-infrared wavelengths.
Figure 2.14 Remote raman spectra of ethyl benzene and tolune over the spectral range 180-1800 cm-1 at 10 m range (from S.K Sharma21; reprinted with permission from Spectrochimica Acta). 46
47 This is relevant for laser beams that have a cross-section smaller than 0.39 cm2. 25 The IEC-60825-1 and ANSI Z136.1 standards in clude methods of calculating MPEs. Figure 2.15 shows the maximum permissible energy of various pulsed lasers at different wavelengths as a function of e xposure times from pico seconds to one second.25 Figure 2.16 shows maximum permissible energies as a function of different wavelength pulsed lasers.25 Most of the LIBS work reported in this dissertation has been performed using either a 266 nm quadrupled Q-switched Nd:YAG laser (fundamental wavelength of 1064 nm) or a pulsed 10.6 micron CO2 laser. As can be seen from Fig. 2.16 the main benefit of th is is that the Maximum Permissible Energy (MPE) at 266 nm is about 600 times the MPE at 1064 nm.42 While the laser output energy is reduced from 890 mJ at 1064 nm to 90 mJ at 266 nm, the TEPS technique (CO2 enhancement mechanism) provides a 100x signal enhancement which more than compensates for the reduction in power. Additionally the CO2 enhancement laser operates at a wavelength that has a slightly higher MPE compared to the 266 nm Nd:YAG. Thus using a 266 nm wavelength over 1064 nm for the production of LI BS plasma seems to have eye safety advantage without losing much of the signal
Figure 2.15 Chart of Maximum Permi ssible Energy versus exposure time for pulsed lasers at different wavelengths (derived from ANSI Z136.6 Standard25). 48
Figure 2.16 Chart of Maximum Permi ssible Energy versus wavelength for pulsed lasers (derived from ANSI Z136.6 Standard25). 49
50 CHAPTER 3. REMOTE SENSING OF EXPLOSIVE TATP VAPORS USING CO2 DIAL LIDAR A CW tunable CO2 Laser Differential Absorption lidar (DIAL) system has been developed, for the first time to our knowledge, for the remote sensing of explosive Triacetone Triperoxide (TATP) gas vapors. The DIAL laser beam was transmitted through an enclosed absorpti on cell containing TATP or SF6, and backscattered returns were measured from a retro-reflector array target at ranges of 5 m to 100 m. DIAL sensitivity for the detection of TATP was about 0.5 ng/ l for a 0.3 m path. 3.1 Motivation With the advent of peroxide base d explosives like Triacetone triperoxide (TATP) the danger of more sophisticated improvised explosive devices (IEDs) have increased. The ingredients to make TATP are easily available. The simplicity of synthesis of TATP and absence of high-sensitivity detection of TATP and its precursors has resulted in present international air travel security inspections of all small amounts of carry-on liquids. Thus there is need to develop instruments which can detect these explosives. Toward this end, we have developed a t unable Differential Absorption Lidar (DIAL) system for the detection of gas plumes of TATP. It may be noted that TATP has a very high vapor pressure unlik e the very low vapor pressure of most nitrogen based
51 explosives. The vapor pressure of TATP is about 7 Pa which is about 14,000 times that of TNT.27 3.2 Introduction to TATP Acetone peroxide (triacetone triperoxide, peroxyacetone, TATP, TCAP) is an organic peroxide and a primary high explosive. It takes the form of a white crystalline powder with a distinctive acrid smell. It is highly susceptible to heat, friction, and shock. Acetone peroxide was discover ed in 1895 by Richard Wolffenstein.28,29 He was the first chemist who used inorganic acids as a catalyst. He was also the first researcher who received a patent for using the peroxide as an e xplosive compound. It detonates with about 80 % to 85 % of the power of TNT. Also known as "peroxyacetone", acetone peroxide most commonly refers to the cyclic trimer TCAP (tri-cyclic acetone peroxide, or tri-cyclo), also called Triacetone Triperoxide (TATP), and is produced by a reaction between hydrogen peroxide and acetone. The cyclic dimer (C6H12O4) and open monomer and dimer are also formed, but under proper conditions the cyclic trimer is the primary product. A tetrameric form has also been observed. In mildly acidic or neutral conditions, the reaction is much slower and produces more monomeric organic peroxide than the r eaction with a strong acid catalyst. Due to significant strain of the chemical bonds in the dimer and especially the monomer, they are even more unstable than the trimer. At room temperature, the trimeric form slowly sublimes reforming as larger crystals of the same peroxide.30,31
52 Table 3.1 lists most of the known chemical parameters for TATP including its density, molecular mass and boiling point. Figure 3.1 shows a photo of a sample of powder TATP, and a chemical ball and stick model of TATP. Acetone peroxide is notable as a high explosive not containing nitrogen. TCAP generally burns when ignited, unconfin ed, in quantities less than about 2 grams. More than 2 grams will usually detonate when ignited; smaller quantities might detonate when even slightly confined.31 Completely dry TCAP is much more prone to detonation than the fresh product still wetted with water or acetone. TATP is typically synthesized by us ing a highly concentric acid like hydrochloric acid (HCl) or Sulphuric acid (H2SO4) with hydrogen peroxide as the starting chemicals. The typical path for this reaction along with the intermediate product Diacetone Diperoxide (DADP) is shown in Fig. 3.2. 3.3 Preliminary TATP data and LIDAR Signal Return Calculations The absorption spectrum of solid state and vapor phase TATP was measured by M. Sigman (UCF) using FTIR spectrometers. For the solid phase absorption measurements, a Fourier Transform Infrar ed-Attenuated Total Reflectance (FTIRATR) microscope was used. Figure 3.3 s hows the measured absorbance spectrum of vapor phase TATP. For the vapor phase measurements, a 5 cm long optical absorption cell was used at a temperature of 28oC. The concentration of TATP in the vapor phase was estimated to be 8.28 Pa (0.062 Torr) under these conditions. As can be seen, similar absorption spectral features near 3000 cm-1, 1350 cm-1, and 950 cm-1 are present.
Acetone peroxide 3,3,6,6-tetramethyl-1,2,4,5-tetraoxane (dimer) 3,3,6,6,9,9-hexamethyl-1,2,4, 5,7,8-hexaoxacyclononane (trimer) C6H12O4 (dimer) Chemical formula C9H18O6 (trimer) 148.157 g/mol (dimer) Molar mass 222.24 g/mol (trimer) Shock and Friction sensitivity Very high / moderate when wet Vapor Density 7 Pa (0.05 torr) Density 1.18 g/cm Explosive velocity 5300 m/s 17,384 fps 3.29 Miles per second RE factor (relative explosivity) 0.83 Melting point 91 C Auto ignition temperature 97-160 degrees Celsius Appearance white crystalline solid CAS number 17088-37-8 Table 3.1 Specific parameters for TATP.(from http://webbook.nist.gov/chemistry/ ). 53
Acetone peroxide (a) (b) Figure 3.1 (a) Picture of Acetone per oxide powder and (b) Ball-and-stick model of the acetone peroxide trimer (TATP) (from http://webbook.nist.gov/chemistry/ ) 54
Figure 3.2 The acid (H+) used in the s ynthesis is typically concentrated HCl or H2SO4. The alkyl groups (R and R') in the starting ketone, acetone, are CH3 (methyl); however, these groups ar e not required to be identical. 55
Figure 3.3 Vapor phase FTIR absorb ance spectrum of TATP ( path 5 cm; concentration 8.28 Pa). 56
57 Figure 3.4 shows the measured absorbance sp ectrum of a solid pha se TATP obtained using an FTIR-ATR system.The vapor phase measured FTIR absorption spectrum of TATP was baseline corrected and the abso rbance was re-plotted as a function of wavelength ( m) which is the reciprocal of the wavenumber (cm-1). The wavelength of the strong absorption lines of the TATP spectrum are identified in Fig. 3.5. As can be seen in Fig. 3.5, the absorption lines of TATP are fairly str ong with lines occurring near 8.36 m, 7.29 m, 3.38 m, and near 10.59 m. For example, the absorbance, A, of TATP in Fig. 2 at 10.59 m is about 0.0024. Relating the absorbance, A, to the transmission, T, as A = -log10(T), and noting the Beers-Lambert relation of T = exp(Pg L), one calculates that the attenuation coefficient, is 0.014 / Pa m (or 1.75 / Torr-m), where Pg is the partial pressure of the abso rbing gas in Pa (Pascals), and L is the path length in meters. It is informative to predict the absorp tion due to a remote 1-m diameter cloud of TATP. Figure 3.6 shows the normalized one-way transmission spectra of 1 m long TATP vapor cloud having a saturated vapor phase concentration of about 7 Pa at 25oC. The absorption for a 2-way DIAL path would be 31% for the 3.32 m line, 73% for the 8.36 m line, and 17% for the 10.59 m line. As such because of these large absorption values, a cloud of TATP could poten tially be detected using a DIAL system at wavelengths near 7.29 m, 8.4 m, 3.3 m, 10.6 m or 11.2 m. To see this further, a composite transmission spectrum fo r a 100 m path of the atmosphere and a 1 m TATP plume is plotted in Fig. 3.7 using the HITRAN database.9
Figure 3.4 Solid phase FTIR-A TR absorbance spectrum of TATP. 58
Figure 3.5 Absorbance spectrum for vapor phase TATP (cell path length of 5 cm; TATP concentrat ion/partial pressure 8.28 Pa; 28oC ). 59
Figure 3.6 Predicted absorption spectra fo r 1m cloud of TATP (Path 1 m; Saturated vapor of 7 Pa). 60
Figure 3.7 Composite spectrum of 1 00 m atmosphere and 1m TATP plume. 61
62 It can be seen from this figure that th ere is a fairy clean optical window with no interfering absorbing species pr esent at wavelengths near 10 m. Figure 3.8 shows the same plot of the composite spectrum of 100 m atmosphere and 1 m TATP vapor plume, but also including an overlay of the different CO2 laser lines at which a CO2 laser can be tuned to operate. We chose to use the lines near 10.6 m since they coincide with the lase r lines of a tunable CO2 which was available in our lab. Also, this wavelength region has less spectral inte rferences from other TATP mixing solvents such as acetone as re ported by other groups.26-28 Expected lidar return signals were calc ulated using the lidar equation, Eq. (2.1) for the parameters used in our DIAL system. Figure 3.9 shows the Lidar equation work file with the parameters for calculating S/N and Lidar return signal using the LIDARPC software. The predicted lidar return signal as a function of range was calculated and is shown in Fig. 3.10. As can be seen the return signal is fairy weak. Note: the typical Noise-Equivalent-Power (NEP) of the lidar detector is on the order of about 10-7 W. By adding a retro-reflector target the return signal improved substantially by a few orders of magnitude as can be seen in Fig. 3.11; the retro-reflect or array was modeled bychanging the target reflectivity to a high valu e of 100 instead of a value of 0.5, which essentially negates the R-2 term in the standard lidar equa tion. Because of this greatly enhanced signal, in our experiments a retroreflector array was usually used as a lidar target.
Figure 3.8 Composite spectrum of 100 m atmosphere and 1m TATP plume and overlay of CO2 laser lines. 63
Figure 3.9 Lidar equation work file show ing parameters for calculating S/N and lidar return signal using LIDAR-PC software. 64
Figure 3.10 Predicted lidar return signal as a function of range for lidar returns from a hard target. 65
Figure 3.11 Predicted lidar return signal as a function of range for lidar returns from a retro-reflector 66
67 3.4 Experimental Setup A laboratory DIAL LIDAR sy stem was constructed usi ng a grating tuned CW CO2 laser. A schematic of the DIAL LIDAR system is given in Fig. 3.12. A grating tuned CW CO2 laser (Edinburg Inst. Model WL-86T) was used for producing the line tunable emission near 10.6 m. The laser had a CW power level of about 1W and could be tuned over about 40 different lines from 9.7 m to about 11.2 m. The output from the laser was se nt through an optical choppe r (SRS Model # SR540), and directed via beam-splitters toward a CO2 laser line spectrum analyzer (Optical Engineering Model # LSA 16-A), a pyro-electric detector (Eltec Model # 420-0-1491) for power monitoring, and through eith er an absorption cell containing SF6 or through a Test Absorption Cell containi ng the TATP gas sample; the SF6 cell was 5 cm long with ZnSe windows, while the Test Abso rption Cell was 1.75 m long with Mylar windows and was constructed using PVC pipe The laser beam was sampled and detected after passage thr ough the cells, but the majo r portion of the beams was directed via mirrors to a large beam steer ing mirror toward targets outside our lab window or towards a retro-reflector array targ et. The retrore flector array target consisted of a grouping of thirty 3 inch diameter gold coated retro-reflectors.
68 Figure 3.12 Schematic of laboratory CO2 DIAL/Lidar setup.
69 The backscattered lidar returns were co llected by a 16 inch diameter telescope (Meade Model # DS 16) and detected by a liquid nitrogen cooled Mercury Cadmium Telluride (MCT) detector (Electro Optical Systems Inc, Model # MCT10-040). The chopped signal was detected using a lock-i n amplifier (SRS, Model # SR810 DSP) and interfaced to a computer with a Labview software program; a chopper frequency of 330 Hz was usually used in our experiments. A photograph of our DIAL setup is show n in Fig. 3.13. DIAL experiments were conducted by tuning the CO2 laser wavelength to an off-line wavelength and then to an on-line wavelengt h, and deducing the concentration of the target gas from the differential absorption or different intensities of the on and offline lidar returned signals. A photograph of the single retro-reflector target and the gold coated retroreflector array target is given in Fig. 3.14. 3.5 DIAL Calibration with SF6 SF6 was used as a calibration gas for our DIAL system because of its strong absorption lines near 10.6 m and its past use by various DIAL groups for this purpose.33-36 Figure 3.15 shows the qualitative transmission spectra of SF6 gas as a function of wavelength between 10 m and 11m obtained from the NIST data base.37 Taking into account the absorption stre ngth of TATP at the different CO2 wavelengths and the expected CO2 laser output power at these various wavelengt hs, we selected the P (24) line at 10.632 m to be used as the on-resonance wavelength, and the R (24)
Figure 3.13 Photogr aph of laboratory CO2 DIAL/Lidar setup for TATP detection. 70
Figure 3.14 Picture of the single retro-reflector and retro reflector array used for the experiment. 71
Figure 3.15 Qualitative transmission spectra of gaseous SF6 as a function of wavelength near the R(24) and P(24) CO2 laser lines. 72
73 line at 10.220 m to be used as the off-resonance wavelength for the DIAL measurements. A home made 175 cm long PVC plastic cel l with Mylar windows was initially used in our DIAL/LIDAR setup and is shown in Fig. 3.16. Mylar transmission at 10.6 m is about 90%. The DIAL laser beam was directed toward targets outside our laboratory window. The transmission of the DIAL beam through the open atmosphere was measured for returns from a single retro-reflector located at a ra nge of 100 m outside the laboratory window in a pa rking lot. Figure 3.17 show s the measured DIAL returns as a function of time for a 20 minute peri od showing both the on-resonance and offresonance lines. In this case an unknown amount of SF6 was filled inside the test cell and thus it can be observed that for the on-resonance wavelengths at P(24) that the optical beam is absorbed heavily. In a ddition there is some variability for the offresonance R(24) wavelength due to atmospheric turbulance and target specular variation in the retro-reflector returns. It was found that less variability was measured by replacing the single retro-reflector with th e gold coated retro-re flector array as the target.
Figure 3.16 175 cm long PVC plastic cell w ith Mylar windows with injection ports for TATP. 74
Figure 3.17 DIAL returns from a single re tro-reflector at a di stance of 100 m for the CO2 offline R (24) and online P (24) for SF6. 75
76 For example, Figure 3.18 shows the DI AL signal measured for four different CO2 lines using the retro-reflector arra y at a distance of 100 m with three onresonance/absorbed lines and one off-resonance line. A 5 cm long aluminum cell with ZnSe windows was used, and filled with 0.2 or 0.5 Torr of SF6 to conduct DIAL/LIDAR measurem ents with a more controlled concentration of SF6. The picture of the vacuum sealed cell is shown in Fig. 3.19. A vacuum pump (Sergent Welch, Model # 140 2) and gas manifold was used to control the pressure of SF6 gas inside the cell. The DIAL beam was directed through the SF6 cell and then toward the retroreflector array target placed at a range of 100 m outside ou r lab window. Lidar returns as a function of time for each of the on/off resonance CO2 wavelengths were recorded and are shown in Fig. 3.20 and Fig. 3.21 for two different SF6 concentrations; here, an average over 10s of data is shown. As can be seen in Fig. 3.20, there was about 80% transmission (i.e. 20% absorption) for the case of SF6 gas at 0.2 Torr and a 5 cm pathlength. Similarly, Fig. 3.21 shows about 50% transmission (i.e. 50% absorption) for the case of 0.5 Torr of SF6 gas and a 5 cm path-length. The theoretical or expected absorption values are shown as dotted lines in the figures using an attenuation coefficient value of 26 / Torr-m for the 10 P(24) line at 10.632 m obtained from other studies.14
Figure 3.18 DIAL returns from the retroreflector array at a distance of 100 m for some of the different CO2 laser lines. 77
78 Figure 3.19 Picture of the vacuum sealed cell for SF6.
Figure 3.20 DIAL transmission measurements as a function of time for 0.2 Torr of SF6 (5 cm cell) and lidar target ra nge of 100 m. Predicte d absorption for the on-line P(24) line is shown as a dotted line. 79
Figure 3.21 DIAL transmission measuremen ts as a function of time for 0.5 Torr of SF6 (5 cm cell) and lidar target range of 100 m. Predicted absorption for the on-line P(24) line is shown as a dotted line. 80
81 As can be seen, our DIAL measurements are consistent with these previously measured values. Using the Beer-Lambert equation, Eq. (2.8), the attenuation coefficient value, was found to be 27.7/Torr-m and 22.31/Torr-m for the case of 0.5 Torr and 0.2 Torr concentrations of SF6, respectively. These values are consistent with previous measurments.14 3.6 DIAL Detection of TATP Gas The CO2 DIAL system was used with TATP ga s in the laboratory cell. In this case, the test absorption cell in Fig. 3.16 was a 175 cm long PVC plastic pipe cell with Mylar windows to transmit the CO2 wavelengths. The cell had injection ports on the side for the delivery of TATP into the cell. The expected TATP transmission spectrum for a 1.75 m path of TATP vapor at a concen tration or partial pressure of 4.3 Pa was calculated from the FTIR spectra of TATP of Fi g. 3.5 and is shown in Fig. 3.22. As can be seen from Fig. 3.22 the absorption between the offline and the online resonance DIAL wavelength should be expected to be about 10% if the assumed TATP gas concentration for saturated vapor is present throughout the cell Initial DIAL experiments were conducted by passing the CO2 laser beam through the 1.75 m Test Absorption Cell, towa rd the retro-reflector array target, and detecting the backscatter using the telescop e and liquid nitrogen c ooled MCT detector. In this case, the target wa s at a range of 5 m to increase the S/N due to loses experienced directing the beam through the ce ll using the beam splitters shown in Fig. 3.12.
Figure 3.22 Expected transmission spect ra of TATP for a 175cm path length and concentration of 4.3 Pa near the R(24) and P(24) CO2 laser lines. 82
83 A small sample of TATP (about 1 mg) was prepared by M. Sigman and D. Clark, and left in a chloroform (CHCl3) solvent. The sample was then injected via a syringe into the absorption cell and allowed to mix for se veral minutes. During this time, the CO2 laser was operating on the on-resonance P ( 24) line so that absorption due to TATP could be observed. However, in this case negligible absorption was observed as can be seen in Fig. 3.23. It was determined that the TATP was difficult to disperse evenly throughout the entire cell, and that possibly stratification could have occurred even though large (>200cc) syringe pumps were used to circulate and mix the gas inside the cell. Calculations indicated that 1 mg of TATP within th e cell (volume of 17,500 cm3) would produce a concentr ation of about 0.05 ng/ l, and thus produce only about 2% absorption of the single-pass online beam. Our measurements are consistent with these limitations. It was determined that use of a larger sample size ( 1 mg) was not prudent. It should be ad ded that the conversion be tween partial pressure Pg and concentration is 1 Pa = 0.088 ng/ l at 25oC and 1 Pa = 0.077 ng/ l at 70oC using the ideal gas law. To increase the TATP vapor pressure and the optical path length inside the absorption cell, a second (different) glass ab sorption cell was used that could be heated so that the vapor pressure of TATP was increased; the cell was 30 cm long and had Mylar windows. The photograph of this new te st cell is shown in Fig. 3.24. The cell was heated to a temperature near 70 oC by using heating tape wrapped around the 20 cm central portion of the cell. A 200 l sample of TATP solution (concentration of 1 g of TATP in 1 l of CHCl3 solution) was injected into the cell, and the onresonance DIAL laser beam transmi ssion through the cell was recorded.
Figure 3.23 Transmission plot as function of time as TATP was injected into the 1.75 m test cell. 84
Figure 3.24 Heated glass test cell, 30 cm long with Mylar windows for TATP measurements. 85
86 Figure 3.25 shows our measured transmis sion signal as a function of time over the period when the TATP was injected. As can be seen, there was about 10 % absorption when the gas was injected. Using the attenuation coefficient of TATP near 10.63 m and our DIAL data in Fig. 3.25, Figure 3.26 shows the deduced concentration of TATP as the TATP was injected and later for a period of about 2 hours. The measurements indicated a TATP concentration of about 1.8 ng/ l (or 26.4 Pa) when the sample was first injected, with sensitivity (noise level uncertainty) of about 0.5 ng/ l. In order to better quantify the concentration inside the cell, a small samp le of the gas inside the cell was obtained using a precession volumetric gas-tight syring e and the sample was injected into a calibrated Ion Mobility Spectrometer (IMS, Smith Detection 400B) operating in the positive ion mode. Two different IMS read ings were obtained over a period of about 20 minutes and yielded values of 0.72 and 0.99 ng/ l. Our DIAL measured value is consistent with that measured by the I on Mobility spectrometer taking into account some variability after the time of introduction of the sample into the cell. These values can be compared to the theoretical c oncentrations within the cell using 200 g injected into the cell volume of 135 cm3 which yields a maximum concentration of about 1.5 ng/ l. Since our DIAL related measurement is slightly higher than this value, it is possible that some stratificat ion or layering of the TATP vapor could have occurred within the cell, thus leading to an incr eased concentration be ing measured by the relatively small laser beam as it transverses the cell.
Figure 3.25 Online DIAL transmission as a function of time showing injection of TATP into a heated 30 cm long absorption cell. 87
Figure 3.26 DIAL measured TATP con centration inside a heated 30 cm long absorption cell as function of time. 88
89 It was noticed that there were often long term (few minutes) temporal changes in the DIAL signal. A second set of e xperiments were performed 4 hours after the TATP sample was injected inside the ce ll, and the cell was reheated to about 70oC with the heating tape. As can be seen in Fig. 3.27 there was substa ntial variation in the transmission values for the on-resonance and off-resonance wavelengths of P(24) and R(24) CO2 laser lines. The observed variation in transmission may be due to chemical induced changes in TATP with temperature and time as reported by other groups.27,29,30,31 Upon further examination of the cell, it was found that TATP crystals were being formed upon the unheated Mylar windows. Attempts to heat the windows slightly were not successful. To better control the TATP concentrati on in our absorption cell, and to reduce the statistical error in our DIAL meas urements, we are now working on a new absorption cell with ZnSe windows that can be heated uniformly including the windows. We anticipate that such a cell will be able to be tter stabilize the concentration of TATP within the absorpti on cell. Finally, field tests are being planned to test our CO2 DIAL system for detecti on of a potential TATP plume surrounding a large sample of TATP. Howe ver, TATP measurements in different solvent media may also need to be studied. For example, Fig. 3.28 shows the superimposed transmission spectra of TATP and acetone which can be used as the solvent carrier as opposed to chloroform (CHCl3) as in our DIAL measurements. As can be seen the acetone spectra overlaps with many of the strong TATP absorption lines, but has a clean window near 10.6 m.
Figure 3.27 Online and offline DIAL transm ission of TATP as a function of time in a heated 30 cm long absorption cell after a period of 4 hours. 90
Figure 3.28 Superimposed transm ission spectra of TATP and acetone. 91
92 3.7 Conclusion for CO2 DIAL detection of TATP A tunable CO2 DIAL system has been devel oped for the first time to our knowledge for the potential dete ction of TATP gas clouds. The system has been used to measure gas samples of SF6, and has shown initial ab sorption measurements of samples of TATP contained w ithin an enclosed optical ab sorption cell. DIAL/Lidar returns from a remote retro-reflector target array were used for the DIAL measurements after passage through th e laboratory cell containing the TA TP gas. DIAL measured concentrations agreed well with those obtained using a calibrated Ion Mobility Spectrometer. DIAL detection sensitivity of the TATP gas concentration in the cell was about 0.5 ng/ l. However, the concentration of TATP was found to be unstable over long periods of time possibly due to re-a bsorption and crystallization of the TATP vapors on the absorption cell windows. A heated cell partially mitigated these effects, but further detailed studies to control the TATP chemistry are required to better quantify our results. We plan to extend these preliminary one-way DIAL measurements to that of a two-way DIAL measurement by placing the absorption cell outside the laboratory window, if the TATP c oncentration within an external cell can be controlled. In addition, a more optimized high power pulsed CO2 laser DIAL system could be used for greater detection ranges, and that pulsed DIAL systems near 3.3 m, 7.3 m and 8.4 m could also be used for TATP detection.
93CHAPTER 4. ENHANCED LIBS PLASMA EMISSION FOR STANDOFF DETECTION USING a 266 nm AND 10 m CO2 LASER SYSTEM A Deep UV LIBS system has been studi ed for the remote detection of solid targets, and potentially chemical, biological, and explosiv e substances. A 4th harmonic Q-Switched Nd:YAG laser operating at 266 nm was used for excitation of the LIBS plasma at standoff ranges up to 50 m. The LI BS plasma emission covering the range of 240 800 nm was enhanced by use of a nearly simultaneous 10.6 m CO2 laser that increased the LIBS plasma emission by several orders of magnitude. 4.1 Single pulse 266 nm LIBS Experiments Laser-Induced-Breakdown Spectroscopy (LIBS) is a recognized laser detection technique for sensing the chemical compositi on of a wide range of materials including minerals, chemical substances, and trace species.11-13 Recently, LIBS is being studied for the remote detection of a wide variety of substances such as surface contaminants and other trace materials.38-41 There is a need to increas e the detection range of such standoff LIBS systems, and, as a result, th e need to increase the strength of the LIBS signal. Towards this end we develope d a single pulse 266 nm LIBS system. A schematic of the 266 nm Nd:YAG single pulse LIBS setup is given in Fig. 4.1. Our 266 nm LIBS system used a 4th harmonic nanosecond Nd:YAG laser to produce a LIBS plasma. As seen in Fig. 4.1, a Q-switched, 4th harmonic Nd:YAG laser (frequency quadrupled Quantel Brilliant B; 90 mJ/pulse, 10 Hz, 6 ns pulse length, M2 of 5) was focused onto the ta rget with a beam expander
266nm Nd:YAG LaserSpectrometer Computer Substrate Negative lens on a moving mount Focussing lens Pulse Delay Generator Optical Fibers Cassegrain Telescope Dichrioc Beam Splitter 20m-50m Si Detector 94Figure 4.1 Schematic of LIBS se tup used for standoff measurements.
95 and focusing optics using a 2.5 cm plano con cave lens (focal length of 100 mm) and 10 cm double convex lens ( focal length of 760 mm). The focused spot size was measured by burn patterns to be about 1 mm in diameter at ranges close to 40 m. The plasma produced on the target emitted LIBS emission into a 2 steradian solid-angle cone, which was collected using a 35 cm diameter telescope (Meade LX200-14) and sent to a dichroc beam splitter and then focused onto two separate 600 micron core fiber optic fibers. The signal was tran smitted to a 2-channel spectrometer (Ocean Optics: Model HR2000+; channel 1 for 300-570 nm, channel 2 for 640-850 nm) where it was detected by a linear CCD array with 2048 pixels and an optical resolution of 0.2 nm/ pixel. An optical orde r-sorting filter in each spectro meter eliminated higher order effects off the grating. The LIBS spectrum from the spectrometer was then transferred to a notebook computer using the Ocean Optics spectrometer software. The Nd:YAG laser beam was at near normal incidence to the sample and created a plasma spark on the remote sample surface. For most solids the necessary irradiance to produce a plasma is on the order of 107-108 W/cm2. 41 As such, the transmitted Nd:YAG laser beam was expanded and then focused down to a point at the remote target as shown in Fig .4.1 using a negative and positive lens. The focus distance to the sample was adjusted by alte ring the distance between the diverging lens and the primary mirror with the help of a motorized moving platform (Thorlabs Model# TST 001/ZST25). LIBS spectra using only the Nd:YAG laser at 266 nm were obtained. As an example, Fig. 4.2 shows the measured spectra obtained for an aluminum target at a distance of about 25 m. As can be seen in Fig. 4.2, there are several atomic type
Figure 4.2 LIBS signal from a pure alum inum target (range 25 m) using a 266nm Nd:YAG laser initiated plasma. 96
97 emission lines superimposed upon a broad continuum. For example the strong elemental line at 777.16 nm is due to O(I) and at 746.86 nm is due to N(I). Other lines in Fig. 4.2 are also identified. Ho wever some of the lines are hidden in a broad continuum due to the black body radiation emitted by the plasma. The C-C Swan bands near 460 nm and 520 nm as re ported by other groups can be clearly seen.38 The absolute error in the spectrometer wave length reading was about 0.2 nm. The statistical variability in the LIBS signal was approximately 15% for 100 laser pulses. 4.2 Deep UV LIBS and 10 m Laser TEPS Experimental Apparatus Our TEPS(Townsend Effect Plasma spect roscopy) 266 nm/10 m LIBS system used a 4th harmonic nanosecond Nd:YAG laser to produce a LIBS plasma, and a simultaneous CO2 laser pulse to overlap spatially and temporally the LIBS spark. A schematic of the simultaneous dual-laser standoff LIBS system is shown in Fig. 4.3. As seen in Fig. 4.3, this dual-laser (TEPS) system is the same as the single Nd:YAG laser LIBS system presented in Fig. 4. 1, but has the addition of a high power CO2 laser to enhance the LIBS plasma. A high-power, pulsed CO2 Transverse Electrode Atmospheric (TEA) laser (Lumonics Model 960; 1.4 J/pulse, 10 Hz) was used to produce 10.6 m laser pulses that were routed using mirrors onto the same LIBS emission target area. Due to the moderately high mode pattern and beam divergence of the CO2 laser (0.01 radians), it was not made co-linear with the main Nd :YAG laser and its beam was focused by a separate lens (5 cm ZnSe lens, focal le ngth of 25 cm) as shown in Fig. 4.3.
98Figure 4.3 Schematic of Deep UV 266 nm LIBS and CO2 laser enhancement LIBS standoff system.
99 The pulse length of the CO2 laser pulse had an initial TEA laser spike of about 100 ns length followed by a nitrogen-fed tail about 5 s long. The CO2 laser output beam size was controlled using a 5 cm focal length ZnSe lens to have a diameter on target between 3 mm to 15 mm. For a 6.5 mm diamet er beam, the energy density was about 40 mJ/mm2. The timing of the laser pulses was controlled with a digital time-delay generator (SRS Model DG535), and the laser pulses were detected using a fast Si photodiode (ThorLabs: Model 10A) for the Nd: YAG laser and a pyro-electric detector (Eltec:Model 420-0-1491) for the CO2 laser. The timing uncertainty (jitter) of the lasers was about 20 ns for the Q-switched Nd:YAG laser and about 500 ns for the TEA CO2 laser. The angular separation of the Nd:YAG/LIBS coll ection beam and CO2 beam was about 20 degrees. The Nd:YAG lase r beam was at near normal incidence to the sample. Figure 4.4 shows photogr aphs of the LIBS setup and CO2 laser resectively. The laser beam profiles were studied with a Spiricon beam profiler to understand the formation of the spark on the substrates. Figure 4.5 and Fig. 4.6 shows the measured beam profile of 10.6 m CO2 laser and the 266 nm Nd:YAG laser, and the 3-D representation of the beams respectively. As can be seen the CO2 beam is much wider compared to the 266 nm beam. The CO2 beam spot size on the target was studied and its relative size compared to the beam spot size of the 266 nm Nd:YAG laser beam. The beam size on the substrat e was varied using the ZnSe lens placed in front of the CO2 beam (see Fig. 4.4). Figure 4.7 shows the CO2 laser burn patterns from blackened paper placed at the substrate. As can be seen in the figure, the CO2 beam was approximately circular, but there were several hot spots (multi-mode) in the pattern.
Figure 4.4(a) Photograph of the LIBS/TEPS setup. Figure 4.4(b) Photograph of the CO2 laser used in LIBS/TEPS setup. 100
Figure 4.5 CO2 beam profile as measured by Spiricon beam profiler. 101
Figure 4.6 Beam profile of 266 nm Nd:YAG laser as measured by Spricon beam profiler. 102
(a) 18 mm diameter spot (b) 20 mm diameter spot (c) 7.5 mm diameter spot (d) 10 mm diameter spot Figure 4.7 Burn patterns on paper due to CO2 laser. 103
1044.3 CO2 Laser LIBS Measurements The influence of the CO2 laser by itself as a source of the LIBS emission was investigated. The energy density of the CO2 laser on the target substrate was varied by changing the iris in front of the CO2 laser beam and changing the position of the ZnSe focusing lens. The CO2 laser induced LIBS emission signal were measured as a function of the CO2 laser energy density. Our results are shown in the Fig. 4.8, showing the LIBS signal as a function of the CO2 laser energy density on a ceramic target. In this case no Nd: YAG laser was used and only the CO2 laser was used to generate the LIBS plasma. As can be seen in Fig. 4.8, the LIBS emission or plasma appears at a CO2 laser energy density of about 120 mJ/mm2. A lower density of about 100 mJ/mm2 produced a black body continuum while lower energy levels of 70 mJ/mm2 produced no LIBS signals. In subsequent sectio ns of this paper where the CO2 laser is used to enhance the 266 nm Nd:YAG laser produced LIBS signal, the energy density for the CO2 laser was about 30 mJ/mm2. As can be seen in Fig. 4.8, this energy density value does not cause breakdown. It is interes ting to note that the 100 mJ/mm2 energy density for a pulse length of 100 ns is equivalent to 108 W/cm2. 4.4 Enhanced 266 nm Nd:YAG LIBS Emission Using Simultaneous 10.6 m Laser Pulse The 266 nm Nd:YAG laser was used to ge nerate the plasma on an aluminum substrate at a distance 25 m away, and the 10.6 m CO2 laser was positioned about
Figure 4.8 LIBS signal on a pur e ceramic substrate using a CO2 pulsed laser initiated plasma as a function of CO2 energy density on the target. 105
106 5 m away from the substrate for our experiments. The CO2 laser beam was focused at the LIBS plasma on the substrate and had a pul se energy of around 1.5 J, a spot size of around 7 mm, and an energy density of 30 mJ/mm2 on the substrate. This value ensured that the LIBS emission was not due to the CO2 laser alone. The LIBS/TEPS spectrum from the remote aluminum target was measured and is shown in Fig. 4.9, where many of the lines have been tentatively identified. As can be seen in a comparison of Fig. 4.2 and Fi g. 4.9, the LIBS signal has been greatly enhanced by the addition of the CO2 laser beam; note that the amplitude scale in Fig. 4.9 is about 100x greater than that shown in Fig. 4.2, where no CO2 laser was used. As can be seen, the signal strength in some cas es is almost 100 times greater. This is consistent with our previous nearIR published work using a 1 m Nd:YAG laser for plasma excitation and a CO2 laser for enhancement.42 In addition to the previously observed lines of oxygen and nitrogen in Fig. 4.3, more lines are seen in Fig. 4.9 indicating that a change in the plasma condition has occurred by addition of the CO2 laser. Table 4.1 lists some of the nitr ogen, oxygen and aluminum lines of the LIBS spectrum from an Al substrate and the e nhancement ratios obtained at a standoff distance of 25 m from the data in Fig. 4.2 and Fig. 4.9. It can be seen from Table 4.1 that the enhancements are on the order of 40-240. To better understand this enhancement, additional studies involving the timing and plasma temperature were conducted, and ar e presented in the following sections. It should also be noted that our two wavelength laser e nhanced LIBS technique is similar to, but distinct from, previous work conducted using dual-pulse LIBS. For
107Figure 4.9 Enhanced LIBS/TEPS emi ssion signal on Al target due to 10.6 m CO2 laser for 266 nm of Nd:YAG initiated plasma
Wavelength (nm) Elemental Id Nd:YAG LIBS (AU) LIBS/TEPS (AU) Enhancement ratio 656.66 O(II) 14.48 3479.11 240 704.22 Al(II) 26.24 1061.04 40 742.52 N(I) 9.44 377.11 40 744.38 N(I) 12.38 750.23 61 746.95 N(I) 13.82 1209.75 88 777.47 O(I) 19.27 2286.12 119 Table 4.1 Enhancement of a few identified lines comparing the 266 nm Nd:YAG LIBS spectrum and the 266 nm/10.6 m LIBS/TEPS signals. 108
109 example, there has been extensive dual-pulse studies using both preand post-ablation secondary pulses which have shown enhancement using similar wavelengths for the dual pulses.58-61 The use of a same wavelength d ouble-pulse laser has been used for ablation of materials and enhancements in the plasma LIBS signal.65 UV excimer and CO2 lasers have been studied and used to enha nce ablation of selected materials for thin film deposition, various dual laser waveleng ths investigated for LIBS enhancement, and a second resonant laser pulse has been used for resonance photoionization of the initial plasma discharge. Our use of two di fferent laser wavelengths at 266 nm and at 10.6 m for LIBS is new, as far as we know and builds off from our previous TEPS studies that were conducted us ing a classical single pulse nanosecond 1064 nm laser for the LIBS excitation followed by the addition of a nearly simultaneous CO2 laser pulse which resulted in signal enhancements on the order of 25 300.42 4.5 Timing Overlap of Two Lasers and its Effect on LIBS/TEPS Plasma The enhancement of the LIBS/TEPS signal was found to be highly dependent on the timing overlap of the two lasers. The overlap was measured using the silicon photo detector which detected the 266 nm and the LIBS plasma emission, and the pyro electric photo detector that measured the CO2 laser pulses reflected from the target. As an example, Fig. 4.10 shows a dual oscilloscope trace of the two laser pulses. In this case the 266 nm beam was reduced in inte nsity so that no LIBS plasma emission occurred. As can be seen, the 266 nm laser pulse width was 6 ns and the CO2 laser pulse width was 0.1 s with a 5 s tail.
Figure 4.10 Pulse shap e and relative timing of CO2 and 266 nm Nd:YAG laser pulses (No plasma). 110
111 With the same experimental setup the 266 nm beam intensity was increased to induce a LIBS plasma. Figure. 4.11 shows the dual oscilloscope trace in this case. Here, the LIBS plasma emission can be observed about 0.5 s after the CO2 pulse. It can also be seen that the transmitted CO2 laser intensity was reduced which means the plasma was absorbing the CO2 laser intensity. A delay generator was used to change the timing between the two lasers. Figure 4.12 shows the measured LIBS signal ac quired at a range of 35 m as a function of the CO2 laser pulse delay with respect to the Nd:YAG laser pulse. As can be seen, a delay of around 0.5 s produced the most enhanced si gnal, however some elemental lines were more pronounced at 2.0 s and 2.8 s. Several of the lines were measured as a function of the CO2 laser pulse delay, and our results are shown in Fig. 4.13. As can be seen, the maximum enhancement was found to occur between 0.5 s and 1.5 s. 4.6 Geometrical Overlap of Laser Intensities on Substrate For optimal LIBS signal enhancement, it wa s critical to make sure the two laser beams on the substrate had good spatial overlap Figure 4.14 shows the basic geometry of the Nd:YAG and CO2 laser beam in two dimensions. The beam radius for the CO2 beam was about 1cm after focusing down with the ZnSe lens, and the 266 nm Nd:YAG had a beam radius of about 0.2 cm on the substrate after focusing down with the negative lens in the setup. To study this more, the 266 nm Nd:YAG laser beam was focused and directed almost perpendi cular to the substrate, while the CO2 laser beam
Figure 4.11 CO2 and 266 nm Nd:YAG laser pulse obser ved directly by use of silicon and pyro-electric detector with the plasma. 112
CO2 Delay (s) Figure 4.12 LIBS spect rum as a function of CO2 laser pulse delay compared to 266nm Nd:YAG laser pulse; standoff range (35 m). 113
114 Figure 4.13 Various elemental LI BS emission lines as a function CO2 laser delay (stand-off range of 35 m).
Figure 4.14 Beam geometry of CO2 laser and Nd:YAG laser on the substrate 115
116 was focused onto the same spot on the substr ate but at an angle of approximately 30 degrees. The energy density on the substrate was va ried by using an aperture placed in front of the CO2 laser beam and by using the ZnSe lens to focus onto the substrate. The energy density on the aluminum substrat e at a distance of 35 m was measured by using a power meter and by measuring the di ameter of the beam. Figure 4.15 shows the LIBS signal as the energy density was varied from 25 mJ/mm2 to about 49 mJ/mm2. As can be seen, the enhancement due to the CO2 laser is negligible at low energy density (< 25 mJ/mm2), increases up to an energy density near 40 mJ/mm2, and seems to saturate at values higher than that. A set of experiments were conducted to find the optimum CO2 laser spot size on the substrate. Attenuators and ZnSe lens were used to vary the spot size of the CO2 laser beam while keeping the energy density constant at around 18 mJ/mm2. The resultant LIBS spectrum as a function of the spot size of the CO2 laser beam on the substrate is shown in Fig. 4.16. As can be seen, maximum enhancement of the LIBS signals seemed to occur at a spot size near 7 mm for the CO2 beam. It should be noted that the spot size used for the 266 nm Nd:YAG laser beam was about 0.5 mm on the substrate. These results indicate that the CO2 laser beam should be larger than that of the 266 nm laser beam, possibly due to pulse-to-pulse variations in the CO2 laser intensity, multi-mode considerations, or direct ional stability. More refined experiments are required to better qua ntify these measurements.
117 Figure 4.15 LIBS/TEPS signal as a function of energy density of the CO2 laser pulse (Al target).
Figure 4.16 LIBS spect rum as a function of CO2 laser spot size on the Al target substrate. 118
119 4.7 Orthogonal Geomet ry for the LIBS/TEPS Setup The previous deep UV 266nm and CO2 laser enhancement LIBS standoff system was designed with almost collinear geometry for the two laser beams as can be seen in Fig. 4.4. In order to better study the CO2 laser target interaction, a set of measurements was made where the CO2 laser beam was directed parallel to the target surface and only interacted with the expa nding plasma plume. The beam overlap geometry of the laser beams was made orthogona l as shown in Fig. 4.17. All the other experimental setup was kept unchanged as in the previous LIBS/ TEPS experiments. The target used for the experiment was a ceramic substrate which was placed on a translation stage which could be moved perpendicular to the 266 nm Nd:YAG laser beam. The silicon photo detector and a pyro el ectric detector were placed in front of the substrate to record the laser beam and plasma intensity. Figure 4.18 shows the measured LIBS spectra from the ceramic target due to 266 nm Nd:YAG laser without CO2 laser pulse and the LIBS spectra on addition of the orthogonal CO2 laser beam to the 266 nm Nd:YAG la ser beam. As can be seen the LIBS/TEPS results had similar enhancemen t of the LIBS signal on the order of 25300x compared to LIBS due to just the 266 nm Nd:YAG laser.
Figure 4.17 Schematic of orthogonal b eam geometry for TEPS/LIBS setup. 120
Figure 4.18 LIBS spectrum with and without addiction of CO2 laser to 266 nm Nd:YAG laser with the orthogonal geometry for both beams. 121
122 As done in Section 4.5 similar studie s were undertaken to understand the LIBS/TEPS plasma. Figure. 4.19 shows a dual oscilloscope trace of the two laser signals when the beams were sampled for the orthogonal beam geometry of the laser beams. As can be seen, the 6 ns 266 nm laser pulse are evident. Figure 4.19 shows the CO2 laser pulse in the pres ence and absence of the 266 nm Nd:YAG laser induced plasma plume. For reference purposes, the average CO2 pulse energy transmitted is shown when there is no plasma and when there is LIBS plasma generated by 266 nm Nd:YAG laser pulse. From this it can be clearly seen from Fig. 4.19 that the CO2 pulse energy is strongly quenched or absorbed by the LIBS/TEPS plasma. LIBS/TEPS spectra of ceramic was measur ed as a function of beam overlap by moving the translation post of the substr ate as shown in Fig. 4.17, and these measurements are plotted in Fig. 4.20. As can be seen in Fig. 4.20 the LIBS/TEPS spectra remains unchanged for a distance of about 6 mm which is within the beam diameter of the CO2 beam (7 mm). Thus it can be concluded that the enhancement due to the orthogonal beam geometry of the laser beams are similar to the collinear beam geometry and the signal is more dependent of direct overlap of the two beams rather than the orientation angles of the two lasers used.
Figure 4.19 CO2 and 266 nm Nd:YAG laser pulse and plasma emission observed using silicon and pyro electric detector for the collinear geometry. 123
Figure 4.20 Dependence of CO2 and Nd:YAG laser beam overlap on orthogonal geometry. 124
125CHAPTER 5. STUDY OF LI BS CRATER FORMATION, RANGE DEPENDENCE AND EFFECTS DUE TO PRESENCE OF ARGON GAS ENVIRONMENT ON LIBS/TEPS SIGNAL The interaction of a high intensity lase r light with a solid target initially increases the surface temperature of the sample such that the material vaporizes. As a result of material vaporizati on and plasma formation, target erosion appears in the form of craters on the sample surface. In this section the crater geometry was studied. Using an aluminum substrate at various standoff ranges, LIBS/TEPS emission spectra were also studied in an Argon gas environm ent and the results compared to spectra obtained in air. 5.1 LIBS/TEPS Induced Crater Study on Al Substrates In order to study the geometry and composition of the formed LIBS/TEPS crater, several LIBS/TEPS shots were taken on the same spot of an Al substrate usually for an average of 50 laser pulse. SEM (scanning electron microscope) images were taken at USF of the craters that were form ed. Figure 5.1 (a) shows the SEM image of the untouched Aluminum surface which can be s een to be relatively smooth. Figure 5.1 (b) shows the SEM image of the crater formed by the LIBS plasma induced by 266 nm Nd:YAG laser alone at a distance of 30 m. Figure 5.1 (c) and Fig. 5.1 (d) shows SEM images at higher magnification of the burn patterns formed inside the crater due to the hot LIBS plasma.
(a) (b) (c) (d) Figure 5.1 (a) SEM images of smooth Al surface, (b) 266 nm Nd:YAG induced LIBS crater from 30m distance on Al substrate, (c)&(d) Increased magnification within the LIBS crater. 126
127 In a similar experiment using the CO2 laser, Figure 5.2 (a) s hows the SEM image of the LIBS crater induced by the 10.6 m CO2 laser alone at a distance of 30 m. Figure 5.2 (b) and Fig. 5.2 (c) shows SEM im ages at higher magnification of the burn patterns formed inside the crater due to the hot LIBS plasma. Figure 5.3 (a) shows the SEM image of the crater formed by the LIBS/TEPS plasma induced by the addition of the 10.6 m CO2 laser to the plasma created by the 266 nm Nd:YAG laser pulse at a distance of 30 m. Figure 5.3 (b) and Fig. 5.3 (c) shows SEM images at higher magnification of the superheated burn patterns formed inside the crater due to the hot LIBS plasma. Finally, Fig. 5.4 shows the SEM images of the LIBS crater formed at a close range of 1 m by the 266 nm Nd:YAG laser alone. Also of interest is Fig. 5.5 which shows the SEM image of the crater formed by LIBS/TEPS plasma. In this case, some dark spots inside the crater were observed. They might be residues of the atmospheric gaseous molecules present during the plasma formation process 5.2 Crater Depth Measurements for LIBS /TEPS as a Function of Standoff Range The size of the crater formed by the LIBS process was studied to better understand the intera ction of the CO2 laser. We measured the depth of the crater that was formed on the substrate as a function of range and as a function of the addition of the10.6 m CO2 laser pulse to the plasma formed by the 266 nm Nd:YAG laser. A Dektek 3030 profilometer was used to measure the size of crater. Figure 5.6 shows a photo of the profilometer. The technical sp ecifications of the profilometer are given in Table 5.1.
(a) (b) (c) Figure 5.2 SEM images of CO2 laser induced LIBS crater s on Al substrate from 30 m (a) at 500 m resolution,(b)50 m resolution and (c) 5 m resolution 128
(a) (b) (c) Figure 5.3 SEM images of CO2 and Nd:YAG induced LIBS/TEPS craters on Al at 30 m Range (a) 100 m resolution,(b) 500 m resolution and (c) 50 m resolution 129
(a) (b) (c) Figure 5.4 SEM images of 266 nm Nd:YAG induced craters at 1m range on Al substarte at (a) 100 m resolution,(b)500 m resolution and (c) 50 m resolution 130
Figure 5.5 SEM image of CO2 and 266 nm Nd:YAG induced crater. 131
Figure 5.6 Photograph of the 3030 Dektak profilometer. 132
133 Stylus: 12.5 m radius (Diamond) Scan Range x: 50 m 50 mm z Resolution: 1 /65 K ; 10 /655 K ; 20 /1310 K Force: 1-40 mg Zoom: 35x200x Scan Speed Ranges Low,Medium, High Leveling Manual, two-point programmable or cursor leveling Stylus Tracking Force Programmable, 1-40 mg (0.01 0.4 milli Newtons Maximum Sample Thickness 45 mm (1.75 inches) Maximum Sample Weight 0.5 Kg (1 lb) Sample Stage Diameter 165 mm (6.5 inches) Sample Stage Translation X Axis, 76 mm ( 3 inches) ; Y Axis, 76mm (3.0 inches) Sample Stage Rotation 3600 Power Requirements 100/115/220Vac 10%, 50-60Hz, 200V Warm up Time 15 Minutes Operating Temperature 21 C 3 C ( 70 F 5 F) Table 5.1 Technical Specifications for 3030 Profilometer.(from Dektak Manual).
134 Photos and profilometer measurements we re made of the LIBS/TEPS craters to help determine the size and depth. Figure 5.7 (a) shows a magnified image of the crater formed by the LIBS plasma, while Fig. 5.7 (b) shows the meas ured profile of the crater formed. As can be seen; material is displaced from the center and deposited on the sides of the crater. Figure 5.8 shows a schematic diagram of the typical crater profile and our designated geometrical parame ters that were measured. From the values for the depth and the diameter of the crater formed as measured with the profilometer, the approximate volume of th e crater was calculated by approximating a conical volume for the crater and the residues deposited on the side of the crater. The amount of metal lost was also calculated by calculating the vol ume deposited on the sides of the crater and su btracting it from the total volume of crater (cone). Initial LIBS and LIBS/TEPS crater meas urements were calculated for the case of the aluminum target located at a range of 30 m. Table 5.2 lists the average value of the volume of the crater formed due to the plasma and the volume of the residues formed on the sides of the craters which enable us to find the volume of the mass ablated. The amount of mass ablated by the addition of the CO2 laser beam is about 10 times more compared to the ablation due to the 266 nm Nd:YAG laser alone. Following this technique, crater parameters were measured at different ranges for the LIBS/TEPS configuration. Our results are shown in Table 5.3. Figure 5.9 shows a plot of the measured crater volume and the volume ablated as a function of range. As can be seen in Fig. 5.9 (a), the volume of the crater decreased as
(a) (b) Figure 5.7 (a) Photograph of the crater formed at 25 m range on a aluminum substrate. (b) Screen shot of the Profilometer Display for the crater. 135
Figure 5.8 Diagram for the typical crat er formed due to LIBS along with the parameters for the depth and radius of the average cones formed. 136
Laser source at distance (30 m) Volume of the crater(m3) Volume of residues outside the crater(m3) Volume lost (m3) LIBS : 266nm Nd:YAG 9.76 X 10-13 5.71 X10-13 4.04 X10-13 LIBS/TEPS : Nd:YAG & CO2 9.6 X 10-12 7.75 X10-12 1.84 X10-12 Table 5.2 Data for the mass ablated with the use of 266 nm Nd:Y AG laser alone LIBS and both Nd:YAG and CO2 lasers together (LIBS/TEPS). Target range of 30 m. Distance(mm) d1 (mm) d2 (mm) P3 (mm) P2 (mm) P1 (mm) Volume of crater(mm3) Volume ablated(mm3) 15000 1.038 1.412 0.07045 0.0578 0.0463 0.00813 0.00423 25000 0. 816 1.002 0.01730 0.0092 0.00785 0.0018 0.00164 35000 1.238 1.491 0.07378 0.0022 0.00238 0.00187 0.0018 Range Measured Crater Parameters Calculated Volumes Table 5.3 Measured averaged values for the LIBS/TEPS craters on Aluminum substrate as a function of range. 137
(a) (b) Figure 5.9 (a) Volume of the craters form ed on the substrate as function of range. (b) Volume of the metal lost on substrate as function of range. 138
the range increased. Our results are consiste nt with the results obtained by other groups showing an ablation rate which is proportiona l to the inverse third power of the range.41 However it was also noticed that the volum e of the crater reached a minimum at around 25 m and did not change dr astically at longer ranges Figure 5.9 (b) shows that the volume ablated or lost seems to follow th is trend also. This is consistent with the interplay of the minimum beam diameter reached at the longer ranges and the increased confocal beam parameter, zo, as a function of range, where the confocal beam parameter is given by10 : 2 0 02W z (5.1) The confocal beam parameter zo for the Nd:YAG laser was cal culated to be about 6 m by using Eq. (5.1) and using the beam width W0 as 0.5 mm.10 5.3 Standoff LIBS/TEPS Measurements as a Function of Range In order to study the range dependence of the LIBS/TEPS signal, a series of LIBS/TEPS experiments were c onducted at varying standoff ranges. In this case, due to CO2 laser transmitter limitations, the CO2 laser was kept at a constant distance of 5 m away from the target. Figure 5.10 shows a plot of LIBS/TEPS spectrum measured as a function of range for a target of pure Al. As can be seen, the LIBS signal strength decreased as the range became progressively longer. This is consistent with the effects 139
300 400 500 600 700 800 0 10 20 30 40 50 60 0 1000 2000 3000 4000 Wavelength (nm) Range (m) Intensity (counts) Figure 5.10 LIBS/TEPS signa l as function of range. 140
141 of atmosphere attenuation and the 1/R2 factor as found in the LIBS lidar equation due to the reduced fraction of LIBS emission cap tured by the telescope area as the range is increased. As can be seen, the intensity of the LIBS signal seems to be decreasing as the range is approaching 60 m primarily due to the 1/R2 dependence in the lidar equation. 5.4 LIBS/TEPS Experiments in Argon Environment In order to better understand the role th at atmosphere gases may play in the measured LIBS emission, our setup was slightly modified to make the plasma formation on the substrate in a predominantly Argon environment. The modified setup is shown in Fig. 5.11. Here, the substr ate was placed inside a box with a small open window to let the laser pulses in. An Ar gas tank was used to flood the box with argon gas. This configuration ensured that the LIBS plasma on the substrate was formed in an excessively Argon gas environment but di d not eliminate the atmospheric gases due to the open window in front of the box. The LIBS/TEPS emission spectrum was measured for several target substrates with and without the argon gas flush. Figure 5.12 shows the LIBS/TEPS spectra for a Cu target for an air atmosphere and for an argon gas flush. As expected, several Ar emission lines appear. Interestingly, th e oxygen and nitrogen lines at 777 nm and 746 nm do not show any change. It should be noted that the background continuum seems to increase in the presence of the argon gas. Similar expe riments were conducted in the presence and absence of argon gas for Al target s and for ceramic targets, and the results
142 Figure 5.11 Schematic of Deep UV 266 nm LIBS and CO2 laser enhancement LIBS/TEPS standoff system in Argon gas environment.
(a) (b) Figure 5.12 (a) LIBS/TEPS spectra of Cu substrate. (b) LIBS/TEPS spectra of Cu when the substrate is placed in Ar gas bath. 143
144 are shown in Fig. 5.13 and Fig. 5.14. As expected a number of argon emission lines were observed in the presence of th e argon gas and the background continuum increased as it in the case of the Cu substrate. However a different mechanism seems to occur here as the intensity of the oxyge n and the nitrogen emission lines at 777 nm and 742, 744, 746 nm respectively actually increased in the presence of the argon gas. This may indicate that the argon flush was not complete and other atmospheric gases were still present. More controlled studies need to be c onducted in an absolute vacuum and with different gas baths to understand the various mechanisms that occur in the presence and absence of certain gases. It may be noted that the diffusion velocity of the nitrogen and oxygen is very high (500 m/s), and thus a pure argon gas atmosphere has to be used in order to exclude oxygen and nitrogen completely.39 It is interesting to note that our argon flushing LIBS experiments do not appear to have excluded ambient atmospheric gas emissions. Further discussion with Prof. M. Sigman and his controlled LIBS experiments indicated the difficulty in elim inating the atmospheric Nitrogen using an argon flush. In particular his group found that they had to use a sealed quartz cell containing a controlled argon environment. No CN emission was found on a pure carbon target in an argon atmosphere, if the atmospheric nitrogen had been properly purged by the argon.46
(a) (b) Figure 5.13 (a) LIBS/TEPS spectra of Al substrate. (b) LIBS/TEPS spectra of Al when the substrate is placed in Ar gas bath. 145
(a) (b) Figure 5.14 (a) LIBS/TEPS spect ra on bare ceramic substrate. (b) LIBS/TEPS spectra on ceramic when the substrate is placed in Ar gas bath. 146
147 CHAPTER 6. ELECTRON TEMPERATURE MEASUREMENT OF THE LIBS/TEPS PLASMA BY BOLTZMANN PLOT METHOD Laser induced breakdown spectroscopy l eads to the formation of a plasma where the molecules of the substrate tend to di ssociate into ions and electrons. Some of the molecules, atoms and ions are excited to higher energy states and the kinetic energy of all these particles an d of the free electrons increases. The spectroscopic determination of the electron temperature of a source of radiation is based on the assumption that local equilibrium conditions must exist in each small volume that contributes to emission. In this chapter th e plasma temperature and the electron density are determined using the intensity of the emitted spectral lines. 6.1 Background Theory of the Boltzmann Temperature Measurements The Boltzmann plot method is a simple and widely used method for spectroscopic measurement, especially for measuring the electron temperature of a plasma from using the relative intensity of two or more line spectra having a relatively large energy difference. However, in orde r to practically apply the Boltzmann plot method for the measurement of electron temper ature, the excitation level needs to be reached under a local thermodynamic equilibr ium (LTE) condition. The basic principle of the Boltzmann plot method is descri bed below and mostly comes from the reference.46,47 Plasma descriptions start by trying to characterize properties of the assembly of atoms, molecules, electrons and i ons rather than the individual species. If thermodynamic
148 equilibrium exists, then plasma properties, such as the relative populations of energy levels, and the distribution of the speed of the particles, can be described through the concept of temperature. In fact, thermodyna mic equilibrium is rarely complete, so physicists have settled for a useful appr oximation, local thermodynamic equilibrium (LTE). All one demands is that equilibration occurs in small regions of space, although it may be somewhat different from region to region. A useful approximation usually exists after a sufficient number of collisions have occurred to thermalize the plasma, which means to spread the energy in the plasma across volume and species. Even then, not all species may be in thermodynamic equ ilibrium. It is common for heavy species (atoms and ions) and light species (electrons ) to equilibrate separately more quickly, and later in time with each other. The fundamental physical reason is that energy between collisions partners is shared mo re equally the closer the masses of the colliding particles. Note that there may be more than one temperature. A variety of tests have been devel oped to ascertain whether thermodynamic equilibrium exists. Probably the simplest is that the relative intensities of atomic spectral lines from closely upper spaced le vels in the same multiplet agree with predictions from basic theory.46,47 Self absorption and interferences from nearby lines can limit the usefulness of this approach. Another test has to do with the electron density being high enough for collisions to dominate the populati on of levels. Griem discusses various criteria for the hydr ogen atom, which depend on the energy difference between the levels involved.48 The larger that difference, the more difficult it is to establish equi librium. The worst case in neut ral atoms (except for the ground to first excited states of monatomic gases) is the 10.2 eV difference between the ground
and first excited state of hydrogen. Griem's an alysis suggests that for a temperature of 1eV (~11 000 K) and atmospheric pressure, an electron density of 1017/cm3 would ensure LTE. The approximations imply that LIBS plasmas generated by an irradiance of > 108 W/cm2, at atmospheric pressure, would be sufficiently thermalized several hundred nanoseconds after initiation. The situation could be different for plasmas formed with femtosecond pulses, or at low pressure. If experiment determines that LTE exis ts, the distribution of several quantities, including electron speeds and populations of energy levels or ion stages are dependent on a single quantity, temperature. The Ma xwellian velocity distribution function fM is: (6.1) )2/exp()2/(2 2/3kTmv kTmfM where m is the electron mass and v the el ectron speed. Relative populations of energy levels, whether atomic or molecular in or igin, are given by the Boltzmann distribution with respect to the ground state )/exp()/(/0kTEZgNNj j j (6.2) or for relative population as ] /)(exp[)/(/ kTEEggNNij ijij (6.3) where i and j refer to the two levels, No is the total species population, Ni, and N,j are the populations of levels Ei. and E.j gi. is the statistical weights of the levels (2Ji, +1), 149
Ji is the total angular momentum quantum number of the term, a nd Z is the partition function usually taken as the statistical wei ght of the ground state. The spectral line radiant intensity is given by: 150 ) /exp()4/(4/0kTEZgAhcN gANhI (6.4) I is in units of W/sr, is the line frequency, and A is the transition probability (Einstein A coefficient). N may be the absolute number density. If the latter is the case, then Equation (6.4) gives the radian t intensity per unit volume of source. The ratio of the intensities of two lines is: ]/)'(exp[)'/''(/' kTEEgAAgII (6.5) Choosing lines for which the g, A and E va lues and the wavelengths are known, and measuring the relative intensities, enables one to calculate T by th e two-line method. If the lines have significantly di fferent line widths, then integrated intensities are the measurement of choice. Relative intensities are not easy to measure precisely. A way to improve temperature values is to use many lines simultaneously and perform a graphical analysis by rearranging the Equation (6.4) onto the form: kTEgAIhcNZ /)/ln()/4ln(0 (6.6) or ) /4ln(/)/ln(0hcNZkTEgAI (6.7)
151 This is the equation of a straight line with slope of -1/ kT. Hence if one plots the quantity on the left against E (of the upper state for emission), and if there is a Boltzmann distribution, a strai ght line is obtained. Some of the crucial factors in obtaining a good Boltzmann plot are accurate line intensities, accurate transition probabilities, and well spaced upper levels. Th e further apart the extremes of the upper level values, the easier it will be to define the slope of the line. Temperatures achieved in LIBS plasmas are of course dependent on the energy deposition, hence the fluence and irradiance. For irradiance values of ~1010 W/cm2, the temperature is typical ly 8,000-12,000 K at 1-2 s into the plasma lifetime.12 6.2 Experimental Setup The experimental setup of Fig. 4.3 was modified so that the collection optics bypassed the telescope and used input optical fi bers that were placed directly next to the LIBS target substrate in order to increase the signal strength and enhance the spectral resolution of the observed lines. As such, the optical LIBS signal was fed to the Ocean Optics spectrometer through an opti cal fiber. A schematic of our setup is given in Fig. 6.1. We conducted ini tial LIBS/TEPS temperature experiments using iron as our target because iron has been shown to work well in past LIBS temperature measurements by other groups. Figure 6.2 and Fig. 6.3 show our LIBS and LIBS/TEPS measurements for an iron target at a standoff range of 25 m. Figure 6.2 shows the LIBS
Figure 6.1 Schematic of modified setup for LIBS/TEPS plasma readings on iron substrate. 152
Figure 6.2 LIBS spectrum on Iron substrate when only the 266 nm Nd:YAG laser was used to generate the LIBS plasma. 153
Figure 6.3 LIBS spectrum on Iron subs trate when both 266 nm Nd:YAG and 10.6 m CO2 lasers ware used to generate the plasma. 154
155 spectrum from the iron substrate due to the 266 nm Nd:YAG laser on ly, while Fig. 6.3 shows the emission spectrum from the iron substrate generated by the 266 nm Nd:YAG and 10.6 m CO2 lasers. As can be seen in Fig. 6.3, the background noise and the LIBS intensity increased with the addition of the CO2 laser. Note that the y-axis scale for Fig. 6.3 is about 30 times greater than th e scale in Fig. 6.2. The assignments of these lines as the levels belonging to the lo wer and upper state conf igurations are well known and are tabulated in the NIST database.55 There are seve ral lines between 380 nm and 450 nm belonging to the neutral ir on Fe (I), and the assignments of these observed lines are tabulated in Table 6.1, along with their trans ition probabilities, A, and uncertainty in the A value, A. Assignments were conducted using the wavelengths recorded in the spectral data file from the spectrometer, and were more accurate than that implied by the arrows designa ting the lines in Fig. 6.2, and Fig. 6.3. The electron temperature was calculated using the ratio of the relative line to background intensities of about seven spectral lines using Equation (6 .6) and the data in Table 6.1. In order to increase the accuracy of this technique, lines were chosen which had a large range of upper energy levels, espe cially over the range of 4.37 eV to 6.58 eV. Figure 6.4 is the resultant Boltzmann plot of the Iron (I) lines measured from the 266 nm laser LIBS and that with the addition of the CO2 laser; here there was a 0.5 s delay between the two laser pulses. The slope of the linear regression yields the Boltzmann temperature. As can be seen, the temperature measured for the LIBS plasma
156 (nm) Configurations A(s-1) A(%) g E(eV) 404.58 3p63d7(4F)4s 3p63d7(4F) 7.50E+07 25 9 4.55 411.85 3p63d7(2H)4s 3p63d7(2H) 5.80E+07 25 13 6.58 426.05 3p63d6( 5D)4s4p(3P0) 3p63d6( 5D) 3.70E+07 50 11 5.31 430.79 3p63d7( 4F)4s 3p63d7( 4F) 3.50E+07 25 9 4.44 432.58 3p63d7( 4F)4s 3p63d7( 4F) 5.10E+07 25 7 4.47 438.35 3p63d7( 4F)4s 3p63d7( 4F) 4.60E+07 25 11 4.31 440.48 3p63d7( 4F)4s 3p63d7( 4F) 2.50E+07 25 9 4.37 Table 6.1 Spectroscopic constants of the neutral Fe(I) lines used in Boltzman plot temperature determination (from NIST database55).
Figure 6.4 Boltzmann plot of Fe (I) lin es of the laser-induced plasma for 266 nm Nd:YAG LIBS and enhanced LIBS/ TEPS plasma using both Nd:YAG and CO2 laser pulses. 157
158 was about 8,500 K, and about 12,000 K for the LIBS/TEPS CO2 enhanced plasma. This suggests that the LIBS/TEPS plasma is hotter due to absorption of the CO2 laser radiation by the thermal electron in the LIBS plasma. Of importance is that the correlation value, R2, shown in Fig. 6.4 is high with a value of 0.77 to 0.87. This tends to indicate that Local Thermodynamic Equ ilibrium (LTE) conditions were present within the plasma because there was consistency of the slope between different energy spacings. It may be added that the line n ear 404.58 nm in Fig. 6.2 is somewhat wider and may be due to a combination of lines. However, this data point at 4.55 eV in Fig. 6.3 does not significantly change the derived temperature. The electron temperature of the LIBS /TEPS plasma was also measured as a function of the CO2 laser pulse delay. Figure 6.5 shows the measured Boltzman temperature as the delay between th e 266 nm Nd:YAG laser pulse and 10.6 m CO2 laser pulse was varied. As can be seen, the temperature reached a peak on the order of 25,000 K at a delay of about 1.5 s. In addition, we also measured the intensity of several prominent lines as a function of the pulse delay. This data is shown in Fig. 6.6 along with the temperature data from Fig. 6.5. As can be seen, both the emission and temperature seems to peak around a pulse delay of about 0.5 s to 3 s. It is interesting to note that the peak N and O emission seems to occur at a delay of 0.5 s, while the peak temperature and p eak Fe emission is at around 1.5 s. This may be due to the fact that Fe is much heavier than N and O and thus gets enhanced later than the lighter elements.
Figure 6.5 Plot of measured electron plasma temperature as a f unction of interpulse delays between two lasers (Fe target). 159
Figure 6.6 Plot of electron plasma temper ature as a function of interpulse delays between two lasers superimposed with the enhancement lines of Fe(I) and O-777.42,N-746.83. 160
161 6.3 Electron Plasma Temperature Measurements of LIBS/TEPS Using 10.6 m and 1.064 m Nd:YAG Lasers A previous LIBS/TEPS study was done by our group using a simultaneous CO2 laser pulse to enhance the initial 1.064 m Nd:YAG induced LIBS plasma emission. Enhancements on the order of 25 to 300 coul d be obtained for the LIBS emission from ceramic (alumina) targets. A schematic of our simultaneous 1.06 m/ 10.6 m duallaser enhanced LIBS system is shown in Fig. 6.7. The author (AP) revisited the past data and calculated the plasma temperature. These new temperature results are as follows. For background, the 1.06 m / 10.6 m LIBS/TEPS experiment used a Qswitched Nd:YAG laser (Big Sky Laser Mode l CRF200; 50 mJ/pulse, 5 Hz, 5 ns pulse length, M2 = 5) that was focused by a 10 cm focal length lens onto the target. The focused spot size was measured (by burn patterns) to be about 1 mm in diameter. The plasma produced on the target emitted LIBS emission into a 2 steradian solid-angle cone, which was collected using a 5 cm diam eter lens (focal length of 10 cm) focused onto a fiber optic cable. The 300 micron core fiber optic cable transmitted the light to a spectrometer (Ocean Optics:Model HR2000; 200 nm 1100 nm) where it was detected by a linear photodiode array. An order-sorting fi lter in the spectrometer reduced second order features by a factor of 100 fold. Th e LIBS spectrum from the spectrometer was then transferred to a notebook comp uter. A high-power Q-switched CO2 Transverse Electrode Atmospheric (TEA) laser (Lumonics Model 920; 1.3 J/pulse, 5 Hz) was used to produce 10.6 m laser pulses that were then ro uted using mirrors onto the same LIBS
Figure 6.7 Schematic of two laser LIBS system for CO2 laser enhancement of Nd:YAG induced LIBS plasma emission (from D.K Killinger 42 ;reprinted with permission from Optics Express) 162
163 emission target area. The pulse length of the CO2 laser pulse had an initial TEA laser spike of 100 ns length followed by a nitrogen-fed tail about 5 s long. The CO2 laser output beam size was controlled using a lens to have a diameter on target between 3 mm to 15 mm. For a 6.5 mm diameter b eam, the energy density was about 40 mJ/mm2. The timing of the laser pulses was controlled with digital time-delay generators (SRS Model DG535), and the laser pulses were de tected using fast Si photodiodes (Nd:YAG laser) and photon-dr ag detectors (CO2 laser). The timing uncertainty (jitter) of the lasers was about 20 ns for the Q-switched YAG laser and about 500 ns for the TEA CO2 laser. Timing errors of the gate electron ics were negligible, on the order of a ns. It should be added that the pulse-to-pulse stability of the Nd:YAG laser pulse energy was about 2% and about 5% for the CO2 laser. The spatial pr ofile of the Nd:YAG laser was approximately a lower order top-hat mode and appeared stable in mode pattern. However, the Lumonics laser may have had a less stable spatial mode pattern due to small changes in the gain characteristics wi thin the TEA discharge on a pulse-to-pulse basis. The angular separati on of the Nd:YAG beam and CO2 beam was about 30 degrees, with the LIBS collection optics aligned at an angle of about 20 degrees from that of the Nd:YAG beam. The Nd:YAG laser beam was at near normal incidence to the sample. Similar to Section 6.2, the LIBS spectra was obtained on a ceramic alumina substrate. Figure 6.8 shows the 1.06 m Nd:YAG laser LIBS spectra, and Fig. 6.9 shows the LIBS/TEPS spectra using a simultaneous CO2 laser pulse to enhance the initial 1.064 m Nd:YAG induced LIBS plasma emission.
Figure 6.8 LIBS signal from ceramic al umina target for Nd:YAG laser initiated plasma; emission lines are tentatively identified and listed in nm.(from D.K Killinger42 ;reprinted with permission from Optics Express) 164
Figure 6.9 LIBS signal from ceramic alumina target for Nd:YAG and CO2 laser initiated plasma emission lines are te ntatively identified and listed in nm. (from D.K Killinger42 ;reprinted with permission from Optics Express) 165
166 It should be noted that the yaxis scale for Fig.6.9 is a bout 25 times greater than the scale in Fig. 6.8. The assignments of thes e lines using the NIST database is shown in Table 6.2. 15 Several neutral Al(I) lines were id entified and tabulated in Table 6.2, along with their transition probabilities, A, and uncertainty in the A value, A. The electron temperature was calculated using the ratio of the relative line to background intensities of about seven spectral lines using Equation (6 .6) and the data in Table 6.2 and Table 6.3. In order to incr ease the accuracy of this technique the best possible lines were chosen which had a la rge range of upper energy levels over the range of 3.14 eV to 5.47 eV and were reason ably isolated from any adjacent lines. Figure 6.10 is the resultant Boltzmann plot of the Al (I) lines measured from the 266 nm laser LIBS and that with the addition of the CO2 laser; delay between the two laser pulses was 0.5 s. As can be seen, the temperature measured for the LIBS plasma was about 10,600 K, and about 16,400 K for the LIBS/TEPS CO2 enhanced plasma. This suggests that the LIBS/TEPS plasma is hotter due to absorption of the CO2 laser radiation by the thermal electron in the LIBS plasma. Like in Section 6.2, the correlation value, R2, shown in Fig. 6.10 is high with a value of 0.79 to 0.82, and tends to indicate that Local Thermodynamic Equilibrium (LTE) conditions were present within the plasma. 6.4 Additional Electron Plasma Te mperature Measurements of LIBS/TEPS of 10.6 m and 266 nm Nd:YAG Lasers on Va rious Substrates and Emission Lines Using the LIBS/TEPS spectrum from the setup described, a number of substrates were used. Figures 6.11Fig. 6.14 show the LIBS/TEPS spectra obtained from a pure
167 (nm) Configurations A (100000000s1) A(delta)% g Ag E(eV) 226.91 3s23p 3s25d 5.20E+07 10 8 5.20E+07 5.476 257.51 3s23p4(1D)4p 3s23p4(1D)5s 3.04E+07 10 6 9.20E+07 4.82719 308.22 3s23p 3s23d 84800000 25 4 2.50E+08 4.0214 309.27 3s23p 3s23d 57200000 25 2 4.40E+08 4.0216 394.4 3s23p 3s24s 3.04E+07 10 8 9.86E+07 3.143 396.15 3s23p 3s24s 3.04E+07 10 6 2.00E+08 3.142 Table 6.2 Spectroscopic constants for Al (I) lines chosen in Boltzman plot temperature determination (from NIST database55).
Figure 6.10 Boltzmann plot of Al (I) lines of the laser-induced plasma for 1064 nm Nd:YAG LIBS and enhanced LI BS/TEPS plasma using both Nd:YAG and CO2 laser. 168
Figure 6.11 LIBS spectrum on Al substrate when both 266 nm Nd:YAG and 10.6 m CO2 lasers ware used to generate the plasma. 169
170 (nm) Configurations Ag E(eV) 308.22 3s23p 3s23d 2.5E+08 4.0214 309.27 3s23p 3s23d 4.4E+08 4.0216 394.4 3s23p 3s24s 98600000 3.143 396.15 3s23p 3s24s 2E+08 3.142 Table 6.3 Spectroscopic constants for Al (I) lines chosen in Boltzmann plot temperature determination (from NIST database55).
171 Figure 6.12 Boltzmann plot of Al (I) lin es of the laser-induced plasma for 266 nm Nd:YAG and CO2 enhanced LIBS/TEPS plasma.
Figure 6.13 LIBS spectru m on Pb substrate when both 266 nm Nd:YAG and 10.6 m CO2 lasers ware used to generate the plasma. 172
Figure 6.14 Boltzmann plot of Pb (I) lines of the laser-induced plasma for 266 nm Nd:YAG and CO2 enhanced LIBS/TEPS plasma. 173
174 Aluminum substrate and pure Lead substrate, and the temperature obtained using a few isolated neutral Al (I) lines and Pb (I) lines, respectively.50 Table 6.3 and Table 6.4 lists the upper energy level values, configurations along with product of their transition probabilities, A and the statistical weight, g fo r the selected neutral Al (I) lines and the selected neutral Pb(I) lines respectively. Figure 6.12 and Fig. 6.14, shows the resultant Boltzmann plot of the Al (I) lines and Pb (I) lines respectively. As can be seen from Fig. 6.12 and Fig. 6.14 the sl ope of the linear regression yields the Boltzmann temperature 14,000 K and 10,500 K, re spectively, for the Al (I) and Pb (I) lines on the individual substrates. Like in Section 6.2, the high correlation value R2 of ~0.9, for both Al and Pb substrates tends to indicate that Local Thermodynamic Equilibrium (LTE) conditions were present within the plasma. 6.5 Electron Plasma Temperature Measu rements of LIBS/TEPS on Substrates in Argon Environment The LIBS/TEPS system was slightly modi fied to make the plasma formation on the substrate in a predominantly Argon e nvironment, and the modified setup was shown in Fig. 5.11. The LIBS/TEPS spectrum for a Cu substrate was measured and is shown in Fig. 6.15. The spectroscopic c onstants for the Argon Ar(I) lines are given in Table 6.5 The Boltzmann plot is shown in Fig. 6.16. The temperature was found to be around 28,000 K with a high regression coefficient, R2 of 0.9 implying the presence of LTE condition.
175 (nm) A (100000000s-1) A(delta)% g Ag E(eV) 424.49 115000000 10 8 9.2E+08 11.47 438.64 155700000 10 6 9.34E+08 11.47 560.89 84800000 25 4 3.39E+08 9.58 666 57200000 25 2 1.14E+08 9.33 Table 6.4 Spectroscopic constants for Pb (I) lines chosen in Boltzmann plot temperature determination (from NIST database55).
Figure 6.15 LIBS spectrum on Cu substrate with a cons tant Ar flow when both 266 nm Nd:YAG and 10.6 m CO2 lasers ware used to generate the plasma. 176
177 (nm) Configuration A (108s-1) A(delta)% g E(eV) 444.88 3s23p4(1D)4p 3s23p4(1D)5s 3.04E+07 10 6 24.284 463.70 3s23p4(1D)4s 3s23p4(1D)4p 3.04E+07 10 8 21.127 476.48 3s23p4(3P)4s 3s23p4(3P)4p 3.04E+07 10 8 19.867 480.60 3s23p4(3P)4s 3s23p4(3P)4p 3.04E+07 10 6 19.222 484.78 3s23p4(3P)4s 3s23p4(3P)4p 84800000 25 4 19.3 488.90 3s23p4(3P)4s 3s23p4(3P)4p 57200000 25 2 19.8 496.50 3s23p4(3P)4s 3s23p4(3P)4p 3.04E+07 10 8 19.76 Table 6.5 Spectroscopic constants for Ar(I) lines chosen in Boltzmann plot temperature determination (from NIST database55).
28,300 K Figure 6.16 Boltzmann plot of Ar (I) lines of the laser-induced plasma generated by 266 nm Nd:YAG LIBS and CO2 Lasers. 178
6.6 Background Theory for Electron Density Measurement Using Stark Width Collisions between radiating atoms a nd surrounding particles can lead to a broadening of the lines. Such collision broa dening can be classified into three groups: (a) Stark broadening (b) Doppler broa dening (c) Resonance broadening. Stark broadening is the statistical shift of the spectral lines by micro-fields produced by electrons and ions surrounding th e radiating atoms, which is utilized for plasma density measurement 11,12,13,49. In this section, the electron density is evaluated using Stark broadening of the Fe(I) (538.38 nm) line. Resonance broadening occurs between neutrals of the same species and is confin ed to transitions with the upper and lower states connected by electric dipole transition (r esonance line) to the ground state. Thus, resonance broadening is not expected to play a role for the line broadening of the Fe(I) (538.38 nm) line in our experiments. Doppler broadening is the broadening of spectral lines due to the Doppler effect in which the thermal movement of atoms or molecules shifts the apparent frequency or wavelength of each emitter. The many different velocities of the emitting gas results in many small shifts, the cumulative effect of which is to broaden the line. The broadening is dependent onl y on the lines wavelength, the mass of the emitting particle, and the temperature, and can theref ore be a very useful method for measuring the temperatureof an emitting gas. A simple equation based on the Maxwellian distribution law can be used to estimat e the half width FWHM for the Doppler broadening 48,49,50: 1/2 22ln2 2 kT mc (6.7) 179
180 In this equation, is the wavelength (nm), k is the Boltzmann constant (JK1), T (K) is the absolute temperature, m (kg) is th e atomic mass for iron (55 amu or 9.1 x 10-26 kg) and c is the speed of light (ms1). At a temperature T of 10,000 K, the corresponding width is estimated to be 0.003 nm for a transition near 538.38 nm. However, an exact measurement of St ark broadening is not possible unless the broadenings of the spectral line due to other effects such as the Doppler effect are considered. The instrumental width of a sp ectral line, arising from the resolution of a spectrometer, depends on the instrumental slit function. This instrumental width is generally expressed by the Gaussian function. 49-52 In our present LIBS/TEPS study, the instrumental width of the spectrom eter over a wavelength range of 300 nm was found to be about 0.2 nm, which cannot be neglected in the present case, and must be accounted for to determine the Stark broadening. The natural broadening of a spectral line is the width arising from the spontaneous (natural) emission. This natural line shift, arising from uncertainty in the energy of the states involved in the transi tion, occurs when an excited atom is deexcited to a lower level by spontaneous em ission. The FWHM (FullWidth at Half Maximum) natural line width, is determined by the transition probability A as 51 =( 2A)/(2 c) (6.8) A typical natural line width corresponding to a tran sition probability of 108 s1 and a wavelength 538.38 nm calculated using Eq. (6.8) is found to be about 1.53 105 nm,
which is very small compared with the inst rumental width (0.2 nm), and can therefore be neglected in the present study. The instrumental width and the Doppler width can be expressed by a Gaussian lineshape function, while the Stark width is expressed by a Lorentzian function. 48,52 Therefore, in order to evaluate the Stark width, it is necessary to remove the Gaussian contribution from it. Since both the Doppler wi dth and the instrumental width are Gaussian profiles, these two widths can be expressed by a single width by overlapping them with each other, which can be expressed by 52 2 Gauss = 2 Instrument + 2 Doppler (6.9) where Instrument and Doppler are the instrumental wi dth and Doppler width, respectively. In our LIBS/TEPS study, th e instrumental width and the Doppler width were found to be about 0.2 nm and 0.003 nm respectively. Therefore, the Gaussian width calculated using Eq. (6.9) is found to be approximately 0.2 nm. The electron number density related to the FWHM of Stark broadening lines is given by the expression 49,50,51,52 5/4 1/3 1/2 16 163 ()2()3.5() 10104ee DNN nm A N (6.10) In this equation, (nm) is the electron impact width parameter also known as the Stark coefficient or half-width, A (nm) is the ion broadening parameter, Ne (cm3) 181
is the electron number density, and ND is the number of particles in the Debye sphere. The first term in Eq. (6.10) refers to broadening due to the electron contribution whereas the second term is at tributed to the ion broadeni ng. Since the contribution of the ionic broadening is normally very small, it can be neglected and Eq. (6.10) reduces to a much simpler form of 1/2 162(). 10eN (6.11) In this equation, the value of the Stark coefficient ( ) corresponding to different plasma temperatures is obtained from the reference data 53. 6.7. Experimental Stark Width Calculations To calculate the stark line width, we chose the 538.33 nm line for different reasons: first, its Stark coefficient is reported in the literature with an uncertainty better than l0%, and second, this line is well isol ated in the spectrum, making its treatment easier. Figure 6.17 and Fig. 6.18 shows the plot of the stark broadening of line Fe(I) 538.33 nm and the Lorenzian fit on the plot which gives the value of the linewidth (FWHM) of 0.216 nm for both the cases where the plasma was enhanced by using both 266 nm YAG laser 10.6 m CO2 laser and the case where we used only 266 nm YAG laser to enhance the plasma. Using the value for the linewidth after it has been corrected by subtracting th e instrumental width and us ing Eq. (6.11) the electron density was found to be 0.568 x 1016 cm3. 53 182
Figure 6.17 Superimposed image of the Stark broadening of Fe line along with its Lorentzian fit when both YAG and CO2 laser is used for plasma generation. 183
Figure 6.18 Superimposed image of the St ark broadening of Fe line along with its Lorentzian fit when only YAG laser is used for plasma generation. 184
185 The McWhirter criterion 54 given in Equation (6.12) is used to check the condition for the validity of the LTE. Ne KT 1/2( E)3 (6.12) where Ne is the electron density in (cm-3), K is a collisional coefficient ( K~ 1.6 x 1012 eV3 K-1/2 cm-3) which is equal to K= 5/8 1/2( /a0EH)3(k/EH)1/2 with the fine structure constant, a0 the Bohr radius in (cm), k the Boltzmann constant in (eVK-1), EH the ionization potential of hydrogen in (eV), Te the electron temperature in (K) and E the transition energy from in (eV). For the case of Fe(I) line the value of Ne 0.181 x 1014 cm3 which is true from the density of state we calculated earlier, thus the LTE condition is satisfied. It must be noted that the values fo r the electron density calculated by the above method is an approximation as the instrument width which was 0.2 nm was very close to the observed stark width of 0.216 nm. Higher resolution spectral measurments are being planned to better quantify these initial measurements.
186CHAPTER 7. LIBS/TEPS ON VARIOUS SUBSTRATES AND STUDY OF SIGNAL ENHANCEMENT DUE TO DUAL PULSE 266 nm ND:YAG LIBS AND 10.6 m CO2 LASER LIBS/TEPS Enhancement of LIBS/TEPS emission signal by addition of 10.6 m CO2 laser pulse onto the plasma formed by the dual pul se 266 nm Nd:YAG laser was studied and compared to the signal obtained by single pul se LIBS and single pulse LIBS with the addition of CO2 laser pulse. Our LIBS/TEPS setup has been applied for determination of contaminants using a peak ratio analys is of the commonly observed elemental and molecular emission lines present in th e spectra and for studying the LIBS/TEPS emission spectra on various substrates to help identify the substrates. 7.1 Multiple Pulse LIBS Our experimental setup of LIBS si gnal enhancement by simultaneous use of 10.6 m CO2 laser on the plasma formed by the 266 nm Nd:YAG has been done for the first time to the best of our knowledge. The signal enhancement observed by this technique was found to be much higher th an that found previously by other groups using dual pulse LIBS where two laser pulses of the same wavelength were used.56-62 To expand on this idea, a multiple pulse LIBS setup was devised using two separate Nd:YAG lasers. The same experiment al setup shown in Fi g. 4.2 was modified by adding an additional 266 nm Q-switched, 4th harmonic Nd:YAG laser (frequency quadrupled Quantel Brilliant B; 90 mJ /pulse, 10 Hz, 6 ns pulse length, M2 of 5) about 4 m away from the substrate as shown in Fig. 7.1. The beam from the second Nd:YAG
187 Figure 7.1 Experimental setup fo r dual pulse 266 nm Nd:YAG and CO2 laser.
188 laser was also focused at the same spot on the substrate by the use of a mirror and focusing lens. The timing of the two Nd:YAG laser pulses was contro lled with a digital time-delay generator (SRS Model DG535), and the laser pulses were detected using a fast Si photodiode (ThorLabs: Model 10A) for the Nd:YAG laser and a pyro-electric detector (Eltec:Model 420-0-1491) for the CO2 laser. The timing uncertainty (jitter) of the lasers was about 20 ns for the Q-switched Nd:YAG lasers. 7.2 Dual Pulse Nd:YAG LIBS Experiments were conducted to study the enhancement of the LIBS signal by dual pulse LIBS without using the CO2 laser. Thus, in this case, from the experimental setup of Fig. 7.1, the CO2 beam was blocked and timing between the two 266 nm Nd:YAG lasers were altered by the use of the digital time delay generator. The interpulse delay between the two Nd:YAG laser pulses were varied from 0 to 16 s and the LIBS emission was observed. Figure 7. 2 shows the LIBS signal as a function of interpulse delay of the two Nd:YAG lasers for some of the selected emission lines of Al at 396 nm and 394 nm Oxygen at 777 nm Nitrogen at 746 nm, and Hydrogen at 656 nm. As can be seen in Fig. 7.2, the signal enhancement increased up to a pulse delay of 4 s and remained roughly unchanged till 12 s, and droped down afterwards. This result is consistent w ith similar dual pulse studies done by various other groups who reported signal enhancement for dual pulse LIBS stays for pulse delays of up to 10 s.56,57 For comparison purposes, Figure 7.3 show s the LIBS signal due to the single pulse 266 nm Nd:YAG laser and that due to the dual pulse 266 nm Nd:YAG lasers.
Figure 7.2 LIBS signal as a function of varying delay between two 266 nm Nd:YAG lasers on a pure ceramic substrate for some selected lines. 189
Figure 7.3 LIBS signal on a pure ceramic s ubstrate using a single and double 266 nm ND:YAG laser initiated plasma. 190
191 As can be seen in Fig. 7.3, we observed an enhancement of the order of 10 on most emission lines for the dual pulse compared to single pulse LIBS. As can be seen from the schematic in Fig. 7.2 one Nd:YAG lase r was placed much closer to the substrate compared to the other. To find out if changing the order in which the Nd:YAG lasers were fired made any difference more readings were taken. In Fig. 7.3 for the single pulse the 266 nm Nd:YAG placed closer wa s fired first and then the Nd:YAG laser placed at a distance of 25 m away from the substrate was fired at a pulse delay of 4 s. We switched the order and in Fig. 7.4 the 266 nm Nd:YAG laser placed at range of 25 m away was fired first for the single pulse LIBS emission and followed by the second Nd:YAG pulse placed closer at a pulse delay of 4 s for the dual pulse LIBS. As can be seen in Fig. 7.4 we observed similar enhancement of 10x for the dual pulse compared to the single pulse LIBS, however no signi ficant difference was observed by switching the order in which the lasers were fired. 7.3 266 nm Nd:YAG Dual Pulse LIBS With the Addition of 10.6 m CO2 laser LIBS/TEPS emission spectra were obtain ed using the experimental setup of Fig. 7.1. As it was found in Section 7.4 that the enhancement due to dual pulse LIBS stays constant for an inter pulse delay of up to 10 s, the timing between the two 266 nm Nd:YAG lasers was kept at 10 s and the timing of the CO2 laser pulse was swept from 1 s to 14 s with respect to the first Nd:YAG laser pulse.
Figure 7.4 LIBS signal on a pure ceramic substrate using a single and double 266 nm ND:YAG laser initiated pl asma by switching the order 192
193 Figure 7.5 shows the resultant LIBS signa l due to the dual pu lse Nd:YAG laser and CO2 laser as a function of the interpulse delay of CO2 and the first Nd:YAG laser for some of the selected emission lines of Aluminim at 396 nm and 394 nm, Oxygen at 777 nm, Hydrogen at 656 nm and Nitrogen at 746 nm. As seen in Fig. 7.5 it was observed that there was a small window for LIBS/TEPS enhancement of the signals. One mainly gets enhancement when the CO2 couples with the 266 nm Nd:YAG laser beams individually within a 2 s window, and then diminishes for delays beyond that. A comparison was made of the dual pulse LIBS signals from Fig. 7.2 and the dual pulse LIBS/TEPS signal enhanced by addition of CO2 laser pulse shown in Fig. 7.5. The results are given in Fig. 7.6 for the emission line of neutral Al(I) at 396 nm. As noted in previous chap ters, the addition of the CO2 laser pulse to the LIBS plasma by the single 266 nm Nd:YAG laser enhan ced the signal strength by a few 100 times. In this case it was observed that the signal enhancement due to the addition of the CO2 laser to the dual pulse Nd:YAG enhanced the signal strength by almost 200 times under the optimum timing de lay between the lasers. Similar to our data for a single pulse LIBS enhanced by the CO2 laser, the enhancement of the signals for the dual pul se LIBS along with the addition of the CO2 laser pulse was found to be very sensitive to the timing delay between the laser pulses. When the dual pulse delay between the tw o 266 nm Nd:YAG laser was kept at 4 s and the CO2 delay was swept from -2 s to 24 s with respect to the first Nd:YAG laser, some interesting features were observed. Figure 7.7 shows the LIBS signal due to the dual pulse Nd:YAG laser and CO2 laser as a function of the interpulse delay of the
Figure 7.5 LIBS signal fr om selected lines when CO2 laser is added to dual 266 nm Nd:YAG laser on a pure ceramic su bstrate and sweeping its delay (YAGYAG delay of 10 s). 194
LIBS/TEPS Signal Dual pulse LIBS Signal Figure 7.6 Dual pulse LIBS signal compared to Dual pulse and LIBS/TEPS for Al (I) 396 nm line for varying CO2 to 266 nm Nd:YAG interpulse delays. 195
Figure 7.7 LIBS signal on selected lines when the CO2 laser is added to dual 266 nm Nd:YAG laser on a ceramic substrate and sweeping its delay with the YAG-YAG delay of 4 s. 196
197 CO2 and the first Nd:YAG laser for some of the selected emission lines of Aluminim at 396 nm and 394 nm, Oxygen at 777 nm, Hydrogen at 656 nm and Nitrogen at 746 nm. The LIBS/TEPS enhancement of the signal seems to be coincident when the CO2 pulse couples with either of Nd:YAG laser pulses and diminishes later. This is similar to the data shown in Fig. 7.5. Also of interest in Fig. 7.7 is th e enhancement of the LIBS/TEPS emission signals that was observed at the pulse delay of 21 s for the CO2 laser with respect to the first Nd:YAG laser pulse. Finally, for comparison purposes, Fig. 7.8 is a plot of the LIBS signal induced by 266 nm Nd:YAG laser enhanc ed by addition of the CO2 laser pulse on a bare ceramic substrate. As observed in Section 4.4, the enhancement in the LIBS signal was about two orders of magnitude on most of the selected emission lines. Figure 7.9 shows the plot of the dual pulse LIBS signal enhanced by the CO2 laser pulse on bare ceramic. As can be seen from Fig. 7.9, similar enhancements for most emission lines were observed in this case compared to single pulse LIBS and CO2 laser. Some lines seem more intense, but overall sp ectral resolution seems similar. 7.4 Survery of LIBS/TEPS Signal on Various Substrates The original LIBS/TEPS setup as shown in Fig. 4.3 was used to record LIBS/TEPS spectra on various substrates. The 266 nm Nd:YAG laser and 10.6 m CO2 laser were used in our remote stand-off configuration to obtain enhanced LIBS/TEPS spectra for several different types of target. Figures 7.10 to 7.14 shows the measured LIBS signal on bare substrates of ceramic, Copper, Iron, Lead and plastic substrates at a standoff range of 20 m.
Figure 7.8 LIBS signal on ceramic when CO2 laser is added to single 266 nm Nd:YAG. 198
199 Figure 7.9 LIBS signal on ceramic when CO2 laser is added to dual 266 nm Nd:YAG.
200Figure 7.10 LIBS/TEPS signal on bare ceramic substrate at 20 m range.
201 Figure 7.11 LIBS/TEPS signal on bare Copper substrate at 20 m range.
202Figure 7.12 LIBS/TEPS signal on Iron substrate at 20 m range.
203Figure 7.13 LIBS/TEPS signal on Lead substrate at 20 m range.
204 Figure 7.14 LIBS/TEPS signal on Plastic (cd) at 20 m range.
205 As can be seen, a number of molecular a nd atomic lines can be observed. Also shown in the figure is a tentative identification of the lines using the NIST database. 55 As can be seen, ceramic has Al, O, N and Fe lines copper has Cu, Al, Fe, N,O lines, iron has Pb, Fe, Al, O, N lines and lead has Fe, Pb along with O, N lines. There are several lines identified for the plastic samp le, but several also overlap the C2-Swan bands near 500 nm. It is interesting to note that oxygen and nitrogen emission lines are present on all substrates, even though most substrates us ed are 99% pure. It is possible that the oxygen and nitrogen lines are due to the at mospheric air present around the LIBS/TEPS plasma and would not have been observed if the experiment was conducted in vacuum. However, further controlled ambient LI BS experiments are required to better understand this. 7.5 Peak Ratio Analysis for Dete ction of Impurities on Substrates One of the major advantages of the LIBS technique is its ability to depth profile a specimen by repeatedly discharging the lase r in the same position, effectively going deeper into the specimen with each shot. This can also be applied to the removal of surface contamination, where the laser is di scharged a number of times prior to the analyzing shot. LIBS is also a very rapi d technique giving resu lts within seconds, making it particularly useful for high volume analyses or on-line industrial monitoring Identification of organic compounds with LIBS has been reported in several papers.620 More often, oxygen, nitrogen a nd hydrogen emission lines and a number of molecular bands are used for this task. The analysis of molecular bands64,75 is focused on the detection of CN molecular violet bands at 386.17 nm, 387.14 nm, and
206 388.34 nm and C-C carbon Swan bands at 516. 52 nm. It is well known that the intensity of the Swan system is proportiona l to the concentration of the carbon dimmer in the excited state while the CN bands emission could be also due to the CN generation in the ambient air. Thus, only the measurement of the C-C bands is reliable for the analysis in the open atmosphere. Although the raw atomic emission has been used for the analysis of organics,67,6971 the peak ratio method has been found to yield better results.7274 The intensity ratio is related to the difference between the upper energy levels of the lines and also is propor tional to the ratio of Boltzmann factors, becoming independent of the ablated mass.75 In the present work, the peak ratio analysis in combination with analysis of molecular bands was used for identification. A number of organic and non organic samples in the form of solution thin films evaporated on solid surfaces were analyzed. A series of experiments were conducted to test the peak ra tio analysis method for LIBS detection. Towards this end various samples were prepared and a thin layer of the sample was deposited on pure a Alumin um substrate. The LIBS/TEPS emission spectra were collected at a standoff distance of 25 m. While taking the LIBS/TEPS spectra, care was taken to rotate the s ubstrate to ensure the acquired LIBS/TEPS spectra was of the sample deposited on Al uminum substrate and not of the pure Al itself. For the present dissertation, only non-e xplosive sample readings are shown. Figures 7.15 to Fig. 7.23 shows the LI BS/TEPS spectra for Acetonitrile, alcohol, bleach, hydrogen peroxide, thin layer of white paint, gasoline, diesel fuel, WD-40 oil and dirt.
Figure 7.15 LIBS/TEPS spectra of thin Acetonitrile film on pure Al substrate. Figure 7.16 LIBS/TEPS spectra of thin Alcohol film on pure Al substrate. 207
Figure 7.17 LIBS/TEPS spectra of thin bleach film on pure Al substrate. Figure 7.18 LIBS/TEPS spectra of thin hydrogen peroxide film on pure Al substrate. 208
Figure 7.19 LIBS/TEPS spectra of thin layer of paint on pure Al substrate. Figure 7.20 LIBS/TEPS spectra of thin gasoline film on pure Al substrate. 209
Figure 7.21 LIBS/TEPS spectra of thin WD-40 oil film on pure Al substrate. Figure 7.22 LIBS/TEPS spectra of thin di esel fuel film on pure Al substrate. 210
Figure 7.23 LIBS/TEPS spectra of thin la yer of dirt on pure Al substrate. Almost all the spectra show the pres ence of oxygen and nitrogen emission lines near 777 nm and 746 nm respectively. The C-C Swan band and CN bands near 500 nm are also distinctly present in most spectrums From the data in Fig. 7.15 to Fig. 7.23, the peak ratio for the oxygen, hydrogen, Al, and CN bands was calculated and some of the data is shown in Fig. 7.24 and Fig. 25. As can be seen,the peak ratios changes somewhat with different surfaces. However, these preliminary peak ratio analyses demonstrated only limited success in identifi cation of various samples and were not able to distinguish one organic compound fr om another. It is hoped that better processing analysis (PCA,ICA) coupled to di fferent laser detection schemes may help this. Such a system (Raman Lidar) is discussed in the following chapter. 211
All Samples0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50 4.00 4.50 5.00 O/NN/HO/HO/Al *0.1N/Al *0.1CN/C2 Peak RatioRatio Bare Al AN on Al All Samples0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50 4.00 4.50 5.00 O/NN/HO/HO/Al *0.1N/Al *0.1CN/C2 Peak RatioRatio Bare Al Wd-40 on Al (a) (b) All Samples0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50 4.00 4.50 5.00 O/NN/HO/HO/Al *0.1N/Al *0.1CN/C2 Peak RatioRatio Bare Al Paint on Al All Samples0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50 4.00 4.50 5.00 O/NN/HO/HO/Al *0.1N/Al *0.1CN/C2 Peak RatioRatio Bare Al H2O2 on Al (c) (d) Figure 7.24 Peak ratios for elemental lin es of oxygen, hydrogen, and Al and CN bands for various substances shown above 212
All Samples0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50 4.00 4.50 5.00 O/NN/HO/HO/Al *0.1N/Al *0.1CN/C2 Peak RatioRatio Bare Al Bleach on Al All Samples0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50 4.00 4.50 5.00 O/NN/HO/HO/Al *0.1N/Al *0.1CN/C2 Peak RatioRatio Bare Al Acetonitrile on Al (e) (f) All Samples0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50 4.00 4.50 5.00 O/NN/HO/HO/Al *0.1N/Al *0.1CN/C2 Peak RatioRatio Bare Al Alcohol on Al All Samples0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50 4.00 4.50 5.00 O/NN/HO/HO/Al *0.1N/Al *0.1CN/C2 Peak RatioRatio Bare Al Gasoline on Al (g) (h) Figure 7.25 Peak ratios for elemental lin es of oxygen, hydrogen, Al and CN bands for various substances shown above. 213
214CHAPTER 8. PRELIMINARY DETECTION OF SPECIES USING A 266 nm ND:YAG LASER STANDOFF RAMAN LIDAR SYSTEM Raman spectroscopy is a powerful technique for molecular analysis, suitable for detecting both organic and inor ganic materials in any state (s olid, liquid, gas). Raman spectroscopy provides direct informati on about the molecular structure and composition of the sample. The vibrational spectrum provided by Raman is a molecular fingerprint that can be used to differentiate the sample from a complex media or differentiate very simply struct ured molecules from one another and often provides unambiguous sample identification.18-24 An area of Raman spectroscopy that is seeing renewed interest is standoff Ra man spectroscopy. In standoff Raman, a pulsed laser is used to investigate the sample from a distance from which inelastically Raman scattered photons are colle cted with large optic and detected with dispersive spectrograph system. One caveat, however, is that the Raman emission process is often very weak, and is better suited fo r detection of high c oncentration species. 8.1 Experimental Setup for Standoff Detection Using Raman Spectroscopy For detection and identification of va rious species, a Raman lidar setup was constructed. Figure 8.1 shows a schematic of our standoff Raman lidar setup. Our Raman system used a 266 nm 4th harmonic nanosecond Nd:YAG laser to produce molecular excitation of the sample. This system was similar to our 266 nm LIBS lidar system. The spot size measured on the sample was about 2 cm in diameter at ranges close to 15 m. The scattered photons on the target emitted emission into a 2 steradian.
215 Figure 8.1 Schematic of the laboratory lidar Raman setup.
216 solid-angle cone, which was collected us ing a 35 cm diameter telescope (Meade LX200-14) and then focused onto two separate 600 micron core fiber optic fibers. The signal was transmitted to a 2-channel sp ectrometer (Ocean Optics: Model HR2000+; channel 1 for 300-570 nm, channel 2 for 640-850 nm) where it was detected by a linear CCD array with 2048 pixels and an optical re solution of 0.2 nm/ pixel. An optical order-sorting filter in each sp ectrometer eliminated high order effects off the grating. The spectrum from the spectrometer was then transferred to a not ebook computer using the Ocean Optics spectrometer software. A part of the signal was also sent to an spectrometer (Andor Model #SR303) by the optical fibers via appropriate laser rejection filters (Shamrock, Model # LP02266RU-25) to block the 266 nm laser beam and the signal was then sent to a note book computer for data analysis. The video camera in the setup was used only for ali gnment purposes. Except for the solid Teflon sample target, all other samples were placed inside a UV quartz cuvette (Starna cells,Model# 3-Q-40). The UV Raman signal due to the 70 mJ/ pulse 266 nm Nd:YAG laser excitation was obtained from a pure solid Teflon sample target at a range of 15 m. Figure 8.2 shows the measured Raman spectrum with th e blocked 266 nm shown as reference. Figure 8.3 shows the same Raman data but as a function of the Raman shift (in cm-1). Using one second integration of the sample some of the clear identif iable features were observed at 390 cm-1, 560 cm-1, 740 cm-1,1220 cm-1, 1305 cm-1and 1390 cm-1 Raman shifted wavelengths.
266 nm Nd:YAG Figure 8.2 UV Raman signal from Teflon ta rget as a function of wavelength.(266 nm excitation, 1s integration, Range 15 m). 217
Figure 8.3 UV Raman signal from Teflon targ et as a function of Raman shift.(266 nm excitation, 1s integration, Range 15 m). 218
219 More Raman spectra were obtained using the sa me setup for the sample of a thin film of paint on a metallic plate at a range of 15 m using the 266 nm Nd:YAG laser as the excitation source. As can be seen from Fig. 8.4, the Raman spectra has Raman peaks at 836 cm-1 and1228 cm-1, 1590 cm-1, 1860 cm-1. The peak at 1590 cm-1 can be possibly due to the atmospheric oxygen. Raman spectra of Acetonitrile (CH3CN) dissolved in wate r was obtained at a range of 15 m and plotted in Fig.8.5. A nu mber of distinct Raman peaks were observed at 405 cm-1, 922 cm-1, 1361 cm-1, 1530 cm-1 corresponding to possible CH3 and CN bonds and the atmospheric oxygen at 1530 cm-1. Under the same experimental conditions at a range of 15 m the Raman spectra of sugar crystals placed in a quartz cell were ta ken and is plotted in Fig. 8.6. As can be seen in Fig. 8.6, there are the Raman peaks at 855 cm-1, 1050 cm-1, 1135 cm-1, 2940 cm-1, 2990 cm-1. The peaks at 1550 cm-1and 2340 cm-1 corresponding to atmospheric O2 and N2, are also observed. 8.2 Conclusion for Raman lidar Detection Recent advances in instrumentation ha ve made it practical to develop fieldportable standoff Raman systems that are suit able for identifying hazardous compounds remotely at distances up to and beyond 50 m. This work also demonstrates that standoff Raman is a viable technique for standoff measurements of unknown species at standoff ranges of up to 15 m.
Figure 8.4 UV Raman signal due to 266 nm Nd :YAG laser on thin film of paint with 1s integration and 70 mJ/ pulse of laser (Range, 15m). 220
922 1530 1361 405 Figure 8.5 UV raman signal due to 266 nm Nd:YAG laser on Acetonitrile dissolved in water with 1s integration and 90 mJ/ pulse of laser ( Range 15 m ). 221
Figure 8.6 UV Raman signal due to 266 nm Nd:YAG laser on sugar crystals with 1s integration and 70 mJ/ pulse of laser.( Range ,15 m ) 222
223 Our wavelength selection of 266 nm as a Raman excitation source offers advantages over the conventionally used wavelength of 532 nm of the 2x Nd:YAG laser due to the increase in Raman intensity (i.e., scales as 4) which improves its potential for use in resonance Raman enhancements, a fl uorescence-free window, and the ability to operate in ambient light conditions without gated detection. Work presented here clearly demonstrates that it is possible to measure Raman spectra with characteristic spectral featur es of various organics, chemicals, and minerals using only a single laser pulse w ith a portable remote Raman system. However, significant spectroscopic background interferences were observed at these wavelengths, blocking many of the identifiable spectral features for the chemicals tested and additional optical filtering is required. Further work is planned to incorporate the Raman lidar system with the LIBS/TEPS system to improve the detection of remote trace species.
224 CHAPTER 9. POSSIBLE LIBS/TEP S ENHANCEMENT MECHANISMS AND PLASMA INTERACTIONS The interaction of high-power laser light with a target material has been an active topic of research not only in plasma physics but also in the field of material science, chemical physics and particularly in analytical chemistry. The high intensity laser beam impinging on a target (solid, liqui d or gas) may dissociate, excite, and or ionize. The constituent atomic species of the solid and produces plasma, which expands either in the vacuum or in the ambient gas depending on the e xperimental conditions. As a result of laser-matter interaction, vari ous processes may occur such as ablation of material target acceleration, high energy particle emission, generation of various parametric instabilities as well as emission of radiati on ranging from the visi ble to hard X-rays depending on the intensity of laser. These processes have many appl ications but we are mainly interested here in the study of optical emission from the plasma and the enhancement mechanisms. In this chapter, we indicate some initial potential plasma physics interaction of our LIBS/TEPS measurements, and present some possible explanations. 9.1 Basics of lase r-matter interaction When a high-power laser pulse is focused onto a material targ et (solid, liquid, gas and aerosols), the intensity in the fo cal spot produces rapid local heating and intense evaporation follo wed by plasma formation.81 The interaction between a laser
225 beam and a solid is a complicated process dependent on many characteristics of both the laser and the solid material. Various factors affect ablation of material, which includes the laser pulse width, its spatial and temporal fluctuations as well as its power fluctuations. The mechanical, physical and ch emical properties of the target material also play an important role in laser-induced ablation. The phenomena of laser-target interactions have been reviewed by several authors81, while the description of melting and evaporation at metal surfaces has been reported by Ready.81 It has been found that the thermal conductivity is an important para meter for vaporization of the solid. The plasma expands normal to the target surface at a supersonic speed in vacuum or in the ambient gas. The hot expanding plasma interacts with the surrounding gas and the expansion of high pressure plasma compre sses the surrounding gas and drives a shock wave, during this expansion, energy is tran sferred to ambient gas by the combination of thermal conduction, radiative transfer and he ating by shock wave. The evolution of the plasma depends on the intensity of the laser, its wavelength, size of focal spot, target vapor composition ambient gas composition, an d pressure. It has been found that the plasma parameters, such as radiative tran sfer, surface pressure, plasma velocity, and plasma temperature are str ongly influenced by the nature of the plasma. Since vaporization and ionization take place during th e initial fraction of laser pulse duration, the rest of the laser pulse energy is absorbed in the vapor and expanding plasma plume. This laser absorption in the expanding vapor /plasma generates three different types of waves as a result of different mechanisms of propagation of absorbing in front into the cool transparent gas atmosphere. These waves are (i) laser-supported combustion (LSC) waves, (ii) laser-supported detonati on (LSD) waves, and (iii) laser-supported
226 radiation (LSR) waves.81 Each wave is distinguishe d on the basis of its velocity, pressure, and on the effect of its radial expansion during the subsequent plasma evolution, which is strongly depende nt on the intensity of irradiation. At low irradiation, laser-supporte d combustion waves are produced, which comprise of a precursor shock that is separated from the absorption zone and the plasma. The shock wave results in an increase in the ambient gas density, pressure and temperature, where as the shock edges remain transparent to the laser radiation. The front edge of the expanding plasma and th e laser absorption zone propagate into the shocked gas and give rise to laser supporte d combustion wave as can be seen in Fig. 9.1. The major mechanism causing LSC wave propagation is radiative transfer from the hot plasma to the cool high pressure ga s in the shock wave. The plasma radiation lies primarily in the extreme ultraviolet and is generated by photo-recombination of electrons and ions into the ground-state atom. At intermediate irradiance, the precurs or shock is suffici ently strong and the shocked gas is hot enough to begin absorb ing the laser radia tion without requiring additional heating by energy transport from the plasma. The laser absorption zone follows directly behind the shock wave and moves at the same velocity. At high irradiance, the plasma is so hot that prior to the arrival of the shock wave, the ambient gas is heated to temperatur es at which laser absorption begins. In the ideal condition, laser absorption is init iated without any density change, and the pressure profile results mainly from the st rong local heating of th e gas rather than a propagating shock wave.
Figure 9.1 Schematic diagram of expandi ng laser produced plasma in ambient gas. Plasma plume is divided into many zones having high-de nsity hot and lowdensity cold plasma. The farthest zone from the target has minimum plasma density and temperature. Laser is absorbed in low-density corona.(adapted from J.P Singh13) 227
228 The laser supported radiative wave velo city increases much more rapidly with irradiance than those of the LSC and LSD waves. The temperature and pressure increase are quite low, which indicates that the LSR wave is effective in channeling the absorbed energy into heating a large amount of gas rather than increasing the local enthalpy. 9.2 Plasma Processes The initiation of plasma formation over a target surface begins in the hot target vapor. First of all, absorpti on of laser radiation takes plac e via electron-neutral inverse Bremsstrahlung process. But when sufficien t electrons are generated, the dominant laser absorption mechanism makes a transition to electron-ion inverse Bremsstrahlung process. Photo-ionization of excited states can also contribute in the case of interactions with short wavelength radiat ions. The same absorption processes are responsible for the absorption by the ambien t gas also. The laser-produced plasma expands into the vacuum or into the surrounding gas atmosphere, where the free electrons present in the plasma modify propagation of laser light.11,12,13,82 The plasma formed by a high intensity or small time duration laser has a very steep density and temperature gradient in comparison to the plasma formed by the low intensity or long time duration laser. The density gradient in the plasma plays a very important role in the mechanism of light absorption and in the partition of absorbed energy between thermal and non-thermal particle distribut ion. There are two basic mechanisms through, which intense laser light may interact with the plasma.83 The first mechanism is inverse Bremsstrahlung, where the electric field of the incident light collides with the
229 electrons, which then lose this energy in collision with ions. This mechanism is important with shallow density gradients in the plasma. The parametric processes also take part most efficiently, when the density gradient is shallow. There are three-wave parametric interaction processes in which intense laser field drives one or more longitudinal plasma waves out of the noise and also parametric decay processes where laser light decays into a high frequency elec tron acoustic wave and a low frequency ion acoustic wave conserving energy and moment um. Another important short pulse laser absorption mechanism is the resonance absorp tion. With a p-polarized light obliquely incident on plasma surface, the radial compone nt of electric field re sonates with plasma frequency and causes large tran sfer of energy to electrons near critical density (Ne) surface. Critical density for a given laser wavelength is49 Ne = 1021/ 2cm-3 (9.1) where is in micron. Energy absorbed at or belo w the critical density in plasma is then conducted towards the target surface by vari ous transport processes. The study of energy coupling to the target has many sub areas such as laser light absorption, nonlinear interaction, electron energy transport and ablation of material from the target surface. One of the important processes, in laser-plasma interaction, is emission of radiation from the plasma ranging from visible to hard X-rays and it is very relevant for the understanding of laser-induced breakdown sp ectroscopy. It has been found that Xrays are emitted from all parts of the absorption, interaction and transport regime. However, as the plasma expands away from th e target surface, its density as well as the
230 temperature decreases. As the plasma temperature decreases the wavelength of emission from the plasma increases, that is, the emission shifts from X-rays to visible region. 9.3 Theoretical Models for Plasma Emission Interpretation of the radiation em itted by the plasmas requires knowledge of both the charge state distribution and the ex cited level populations of different ions. This is possible by obtaining the solution of a complex system of rate equations, describing the population and depopulation of all the levels by the processes such as ionization, recombination, co llisional excitation and de-exc itation, radiative decay and absorption as well as the stimulated emission. Any given charge stat e is connected with its two neighboring states by the processes of ionization and recombination. Considering the difficulties associated in solving these equations, the approximation models used in order of in creasing electron density are: the corona model (CM), the local thermodynamic equilibrium (LTE) models.50,84 9.3.1 Corona Model In this approximation there is a ba lance between collis ional ionization (and excitation) and recombination (and spontane ous decay). This model, therefore, depends critically on the knowledge of atom ic cross sections. Assuming that free electrons have a Maxwellian velocity distribut ion, it is required that only a negligible number of ions be in excited levels (as compared to the ground level). Two neighboring ionization states, of charge Z and (Z + I), are then connected by
NzNeSz(Te)=Nz+1Ne z+1(Te,Ne) or )( ),(1 1 ez ee z z zTS NT N N (9.2) where Sz and z+1 are the ionization and recombinati on rate coefficients, respectively. Eg. (8.2) is, to a first approximation, independent of Ne. In this approximation a balance between the rate of collisional exc itation from the ground le vel and the rate of spontaneous radiative decay determines the popu lation densities of excited levels. This model requires that the electron density be sufficiently low for collisions not to interfere with radiative emission. 9.3.2 Local Thermodynamic Equilibrium Model In the LTE model it is assumed that mainly particle collision processes determine the distributio n of population densities of the el ectrons which are so fast that the latter take place with sufficient ra pidity and the distribution responds instantaneously to any change in the plas ma condition. In such circumstances each process is accompanied by its inverse and thes e pairs of processes occur at equal rate by the principle of detailed balance. Thus the distribution of populat ion densities of the energy levels of the electrons is the same as it would be in a system in complete equilibrium. The population distribution is determined by the statistical mechanical law of equipartition among the energy leve ls and does not require the knowledge of atomic cross sections. Actually the plasma temperature and density vary in space and 231
time, but the distribution of population dens ities at any instant and point in space depends entirely on the local values of temp erature, density and chemical composition of the plasma. The uncertainty in pr ediction of spectral lin e intensities from LTE model plasma depend mainly on the uncerta inty in the values of these plasma parameters and atomic transition probabiliti es. For analytical plasmas the condition of LTE is considered very much vital for getting any reliable quantitative information. In the case of thermal equilibrium all the processes in the plasma are collision dominated as discussed above and the plasma can be considered as having a single temperature Ti = Te. However in the expanding plasma this is possible only locally and for specific time segment during the evolution. The fo llowing criterion must be satisfied by the plasma to be in local thermodynamic equilibrium:83-84 (9.3) 2/1312106.1e eTExN where E(eV) is the largest observed transiti on energy for which the condition holds, and Te is the excitation temperature (K). It should be noted that the choice of the time delay is crucial for obtaining the best operati ng conditions in the LIBS plasma to ensure that LTE prevails during the measurements for obtaining the quantitative results. However, it has also been found that LTE is not an indispensable condition for qualitative analysis, once the measurement c onditions are kept constant and are exactly reproducible 232
2339.4 Possible TEPS/LIBS Mechanisms The phenomenology behind the microsecond long LIBS plasma begins with a deposition of energy into a surface or as a spar k in air. Materials are ejected from the surface beginning early in time (~1 ns) and ar e, partially ionized and partially excited in the resulting laser deposition over the remaining 6 or so ns. This excited and partially ionized mass and residual solids decay over a 20 s period of time. Decay of the plasma is also accompanied by a shock wave expanding in a spherical manner from the target. The shock wave is composed of a sheath of plasma. The wave has been monitored over tens of microsecond periods but never over a larg e distance from the target. Two widely accepted mechan isms or assumptions are discussed.13 In the first mechanism, using a double pulse of laser energy, the second pulse operates through the shock wave, the plasma being optically thin, and generates a second event in the vacuum region behind the shock wave. This being in a reduced N2 region and resulting in enhancement of th e LIBS signal by an order of magnitude.43-46 In the second mechanism, it is believed that the s econd pulse dissociates and ionizes the blow-off particles from the firs t laser event. It has been observed by all doing double laser pulse LIBS that enhancement occurs between 0.5 s and 10 s delta laser event, falling off after 10 s. The second ablation and plasma event is large enough to vaporize the droplets of material not only from the new ablation, but also left behind in the lower pressure re gion behind the first shock wave. 11,40,56,61,65 In both cases, the resultant plasma, emitting Bremsstrahlung radiation, in the first few microseconds follows the LTE mode l (Local Thermodynamic Equilibrium). The LIBS plasma generates the noted ionized particles and dissociated mass. There are
234 many free electrons and the plasma temperature is very high. The plasma is unstable as described above and lasts for about 2 s where Bremsstrahlung ra diation disappears. This is an indication that the plasma no longe r contains a large density of free electrons and now has transitioned from being coll ision dominated to being a radiation dominated plasma with recombination ra diation providing the dominant plasma emission. The plasma expands as a shock wave with a reduced pressure region behind. The loud audible noise one hears from the LIBS plasma is a supersonic pressure wave. With the introduction of an E field by 10.6 m laser radiation, the free electrons are available early in the plasma (< 2 us ) for acceleration further ionizing and exciting the gaseous elements. As can be seen in Fig.9.2, a the shock wave expands, the low pressure behind the shock wave contains the many fragments left over by the laser ablation and excitation. The mean free path of the electrons is quite large due to the low pressure (~10 m at 100 torr) so the E field con tinues to accelerate electrons and produce an ever-larger plasma. As the pre ssure wave decays, the low-pressure region collapses and the plasma excitation is quenche d. The plasma has been extended in time associated with the E field strength and the characteristics of the pressure wave. Laser radiation is absorbed primarily by inve rse Bremsstrahlung, which involves the absorption of a photon by free electrons during the collision with h eavy particles (ions and atoms). The inverse Bremsstrahlung absorption coefficient is given by :13
Figure 9.2 Plasma expansion diagram. 235
)] exp(1) 3 2 ( 3 4 [2 1 4 236 0 eB eBe e ie e IBTk hc x Tkm x mhc NZNe NQN (9.4) where Q is the cross section for photon absorpti on by an electron during the collision with atoms, c is the speed of light, h is the Plancks constant and Z is the charge on ions. Ne, N0 and Ni are the number density of electr on, atoms and ions, respectively, Te is the electron temperature, is the laser wavelength, me is the electron mass and kB is the Boltzmann constant. The CO2 lasers 10.6 m wavelength is much better coupled to the electrons in the plasma as absorp tion coefficient for inverse Bremsstrahlung is directly proportional to 3. It should be noted that pulsed CO2 enhancement occurs only during the very early plasma time when Bremsstrahlung radia tion and thus large el ectron densities are present. Double laser pulsed LIBS (typically 1.064 m Nd:YAG) enhancement (up to x10) occurs out to 10 s in plasma decay. This obviously indicates that double pulsed LIBS is a different phenomena based on ab lation and plasma initiation in a low pressure region and not E-fiel d coupled enhancement. For the case of TEPS/LIBS enhancement we assume that the second pulse comes and superheats the plasma generated by th e first laser pulse. This is different in mechanism compared to the dual pulse LIBS where the second pulse ionizes the fragments created by the first pulse. As it was found that the pulse timing between the two pulses in dual pulse la ser is in order of 1-10 s compared just about 1-2 s for the case of LIBS/TEPS mechanism to get ma ximum signal to noise enhancements. 236
237CHAPTER 10. CONCLUSIONS AND FUTURE WORK A tunable CO2 DIAL system has been devel oped for the first time to our knowledge for the potential dete ction of TATP gas clouds. The system has been used to measure gas samples of SF6, and has shown initial ab sorption measurements of samples of TATP contained within an encl osed optical absorption cell. DIAL/Lidar returns from a remote retro-reflector target array were used for the DIAL measurements after passage through th e laboratory cell containing the TA TP gas. DIAL measured concentrations agreed well with those obtained using a calibrated Ion Mobility Spectrometer. DIAL detection sensitivity of the TATP gas concentration in the cell was about 0.5 ng/ l. However, the concentration of TATP was found to be unstable over long periods of time, possibly due to re-a bsorption and crystallization of the TATP vapors on the absorption cell windows. A heated cell partially mitigated these effects, but further detailed studies to control the TATP chemistry are required to better quantify our results. We plan to extend these preliminary one-way DIAL measurements to that of a two-way DIAL measurement by placing the absorption cell outside the laboratory window, if the TATP c oncentration within an external cell can be controlled. In addition, a more optimized high power pulsed CO2 laser DIAL system could be used for greater detection ranges, and that pulsed DIAL systems near 3.3 m, 7.3 m and 8.4 m could also be used for TATP detection. A Deep UV LIBS system has been studied for the remote detection of solid targets, and potentially chemical, biological, and explosiv e substances. A 4th harmonic Q-Switched Nd:YAG laser operating at 266 nm was used for excitation of the LIBS
238 plasma at standoff ranges up to 50 m Th e LIBS plasma emission covering the range of 240 800 nm was enhanced by use of a nearly simultaneous 10.6 m CO2 laser that increased the LIBS plasma emission by several orders of magnitude. Laser pulse energies on target were about 0.05 J/mm2 for both lasers, with LIBS emission depending upon both the Nd:YAG and the CO2 laser pulse intensities. Enhancement in the LIBS emission on the order of 10x to 100x was observed. The temporal overlap of the two lasers was found to be optimized for a 0.5 s to 1.5 s delay. Electron temperature of the LIBS plasma was m easured by Boltzman method of using the relative intensity of two or more line spect ra having a relatively large energy difference for the levels which satisfy the condition of Local thermodynamic Equilibrium. An increase in the LIBS plasma temperature of about 3,000 K was measured, for the first time due to the addition of the CO2 laser pulse to the plasma generated by the Nd:YAG laser pulse. Electron plasma temperature and the electron density using the stark width were also measured on various substrates. LIBS/TEPS induced mass ablation of the substrates was studied by studying the crater s formed on substrates as a function of standoff ranges and use of different laser pulse s for plasma generation. Behavior of the LIBS emission spectra was observed on various substrates in an Argon gas environment. Peak ratio analysis me thod was employed for detection of impurities present on substrates by comparing the ratios of certain known LIBS emission lines. Experiments were conducted to study the effect of the addition of a CO2 laser pulse on dual pulse LIBS plasma and single pulse LIBS plasma generated by 266 nm Nd:YAG laser.
239 A 266 nm Nd:YAG laser standoff Raman lid ar system was built for detection of various species from a remote distance. Preliminary Raman signals were obtained from targets at range of about 15 m, and showed some detection capability for solid targets and thin films. It is anticipated that using the Raman lidar detection scheme along with the LIBS/TEPS will lead to enhanced detection and selectivity. The results obtained in this disserta tion have been presented in several conferences and publications.86-92 Our future plans call to better quantify the LIBS/TEPS plasma interaction in other ambient atmosphere.
240 REFRENCES 1. D. K. Killinger and N. Menyuk, Laser Remote Sensing of the Atmosphere, Science 235, 37 (1987). 2. N. Menyuk, D. K. Killinger, and W. E. DeFeo, Laser remote sensing of hydrazine, MMH, and UDMH using a differential-absorption CO2 lidar, Applied Optics 21, 2275 (1982). 3. Grady J. Koch, Jeffrey Y. Beyon, Fabi en Gibert, Bruce W. Barnes, Syed Ismail, Mulugeta Petros, Paul J. Petzar, Jirong Yu, Edward A. Modlin, Kenneth J. Davis, and Upendra N. Singh, S ide-line tunable la ser transmitter for differential absorption lidar measurements of CO2: design and application to atmospheric measurements, Applied Optics 47, 944 (2008). 4. E.R. Murray and J.E. Laan, Remote measurement of ethylene using a CO2 differential-absorption lidar, Applied Optics 17, 814 (1978). 5. R. M. Measures, Laser Remote Sensing Wiley & Sons, New York, 1984. 6. John R. Quagliano, Page O. Stoutland, R oger R. Petrin, Quantitative chemical identification of four gases in remote infrared (9-11 m) differential absorption lidar experiments, Applied Optics 36, 1915 (1997). 7. R. H Kingston, A. Jillian, Detection of Optical and Infrared Radiation Springer 1992. 8. P. F. Bernath, Spectra of Atoms and Molecules Oxford University Press, New York 1995. 9. L. S. Rothman, D. Jacquemarta, A. Barbeb, D. Chris Bennerc, M. Birkd, L.R. Browne, M.R. Carleerf, C. Chackerian Jr.g, K. Chancea, L.H. Couderth, V. Danai, V.M. Devic, J.-M. Fl audh, R.R. Gamachej, A. Goldmank, J. M. Hartmannh, K.W. Jucksl, A.G. Makim, J.Y. Mandini, S.T. Massien, J. Orphalh, A. Perrinh, C.P. Rins lando, M.A.H. Smitho, J. Tennysonp, R.N. Tolchenovp, R.A. Tothe, J. Vander Auweraf, P. Varanasiq, G. Wagnerd, The Hitran 2004 molecular spectrosc opic database, Journal of Quantitative Spectroscopy & Radiative Transfer 96,139 (2005); HITRAN database and HITRAN-PC/LIDAR-PC program obtainable at www.ontar.com. 10. Amnon Yariv, Quantum Electronics Wiley & Sons, New York, 1989. 11. A. Miziolek, V. Palleschi, and I. Schechter, Ed. Laser Induced Breakdown Spectroscopy, Cambridge University Press 2006.
241 12. A. Cremers, L.J. Radzeimki, Handbook of laser induced breakdown spectroscopy Wiley 2006. 13. J. P. Singh, S. N. Thakur, Laser-Induced Breakdown Spectroscopy Elsevier 2007. 14. J. Scaffidi, W. Pearman, J.C. Cart er and M. Angel, "Temporal dependence of the enhancement of material rem oval in femtosecond-nanosecond dual pulse laser-induced breakdown spectroscopy", Applied Optics 43, 6492 (2004). 15. Valeri I. Babushok, Frank C. DeLucia, Jr ., Paul J. Dagdigian, Michael J. Nusca, and Andrzej W. Miziolek, Kinetic m odeling of the laser-induced breakdown spectroscopy plume from metallic lead, Applied Optics 42, 5947(2003). 16. Ph. Rohwetter, J. Yu, G. Mejean, K. Stelmaszczyk, E. Salmon, J, Kasparian, J. Wolf, L. Woste, Remote LIBS with ultrashort pulses: characteristics in picosecond and femtosecond regimes, J. Anal. At. Spectrom, 19, 437, (2004). 17. J. R. Ferraro, K. Nakamoto, C. Brown, Introductory Raman Spectroscopy Academic Press 2002. 18. D. A. Long, The Raman Effect Wiley 2001. 19. J. Loader, Basic Laser Raman Spectroscopy Heyden 1970. 20. Richard L. McCreery, Ohio State University, http://www.chemistry.ohiostate.edu/~rmccreer/freqcorr/images/nap.html. 21. S. K. Sharma, A. K. Misra, B. Sh arma, Portable remote Raman system for monitoring hydrocarbon, gas hydrates and explosives in the environment, Spectrochimica Acta Part A 61, 2404 (2005). 22. M. Wu, M. Ray, K. Hang, M.W. Ruck man, D. Harder, A. J. Sedlacek III, Stand-off Detection of Chemical s by UV Raman Spectroscopy, Applied Spectroscopy 54,6 800 (2000). 23. N. Tarcea, M. Harz, P. Rosch, J. Popp, UV Raman spectroscopy-A technique for biological and mineralogical in situ planetary studies, Spectrochimica Acta Part A 68, 10029-1035 (2007). 24. A. Blanco, A. Quevedo, S. P. Rive ra, UV Raman Detection of 2,4-DNT in contact with Sand Particles, Proc. of SPIE Vol. 6217, 621737. 25. American National Standard for Safe use of Lasers Outdoors, Laser Institute of America, ANSI Z136.6 (2005).
242 26. C. Bauer, U. Willer, A. Sharma, W. Schade, A new Photonic Sensor Device for TATP Detection, OSA/ LACS EA Conference Proceedings (2008). 27. Anthony J Bellamy, "Triacetone Triper oxide: Its Chemical Destruction", Journal of Forensic Science 44, 603 (1999). 28. Ilya Dunayevskiy, Alexei Tsekoun, Manu Prasanna, Rowel Go, and C. Kumar N. Patel "High-sensitivity detecti on of triacetone tr iperoxide (TATP) and its precursor acetone", Applied Optics 46, 6397 (2007). 29. Robert Matyas, Jiri Pachman, Ho w-Ghee Ang, "Study of TATP: Spontaneous Transformation of TATP to DADP", Pr opellants, Explosives, Pyrotechnics 33,89 (2008). 30. David Armitt, Peter Zimmermann, and Simon Ellis-Steinborner, Gas chromatography/mass spectrometry analys is of triacetone triperoxide (TATP) degradation products, Rapid Co mmun. Mass Spectrum 22, 950 (2008). 31. Robert Matyas, Jiri Pachman, Therm al stability of Triacetone Triperoxide, Sci. Tech. Energetic Materials 68, 111 (2007). 32. Jimmie C. Oxyley, James L, Smith, Kajal Shinde, Jesse Morgan, "Determination of the Vapor Densit y of Triacetone Triperoxide (TATP) Using a Gas Chromatography Headspace T echnique", Propellants, Explosives, Pyrotechnics 30, 127 (2005). 33. C. K. N. Patel, R. E. Slusher, Selfinduced transparency in gases, Physical Review Letters19, 1019 (1967). 34. L. Lyman, G. Robert Anderson, A. Fi sher, B. J. Feldman, Absorption of pulsed CO2laser radiation by SF6 at 140K, Optics Letters 3, 238 (1978). 35. Hugh R. Carlon, Infrared ab sorption coefficients (3-15 m) for sulphur hexafluoride (SF6) and Freon (CCl2F2), Applied Optics 18, 1474 (1979). 36. Edward E. Uthe, "Airborne CO2 DIAL measurement of atmospheric tracer gas concentration distributions", Applied Optics 25, 2492 (1988). 37. NIST standard Reference database 79: Infrared database. 38. C. Lopez-Moreno, S. Palanco, J.J. Lase rna, F. DeLucia Jr., A. W. Miziolek, J. Rose, R. A.Walters, and A.I. Whitehouse, "Test of a stand-off laser-inducedbreakdown spectroscopy sensor for the detection of explosive residues on solid surfaces," J. of Analytical Atomic Spectrosc. 21, 55 (2006).
243 39. Mark Denny, Air and water: The Biology and Physics of Lifes Media Princeton University Press, New Jersey, 1995. 40. L. M. Cabalin, J.J. Laserna., Experi mental determination of laser induced breakdown thresholds of metals under nanosecond Q-switched laser operation," Spectrochimica Acta part B 53, 723 (1998). 41. S. Palanco, C Lopez-Moreno, J. J. Laserna, F. DeLucia Jr., Design construction and assessment of a fi eld deployable laser-induced breakdown spectrometer for remote elemental sensing," Spectrochimica Acta part B 61, 88 (2006). 42. D. K. Killinger, S. D. Allen, R. D. Waterbury, C. Stefano, and E. L. Dottery, "Enhancement of Nd:YAG LIBS emi ssion of a remote target using a simultaneous CO2 laser pulse," Opt. Express 15, 12905 (2007). 43. A. A Khalil, M. Richardson, C. Barnett, L. Johnson, "Double Pulse UV laser induced breakdown spectroscopy of stai nless steel," Applied spectroscopy 73,735 (2006). 44. K. J. Grant and G.L Paul, "Electron te mperature and density profiles of Excimer laser-induced plasmas," Applied Spectroscopy 44, 1349 (1990). 45. D. Lacroix and G. Jeandel, "Spectro scopic characterization of laser-induced plasma created during welding with a pulsed Nd:YAG laser," J. Applied Phys 81, 65599 (1997). 46. Michael Sigman, University of Central Florida, (Personal Communication). 47. H.E White, Introduction to Atomic Spectra, McGraw Hill, New York, 1934. 48. G. V. Marr, Plasma Spectroscopy, Amsterdam: Elsevier, 1968. 49. H.R. Griem, Spectral Line Broadening by Plasma Academic Press, New York, 1974. 50. H.R. Griem, Plasma Spectroscopy McGraw Hill, New York, 1964. 51. P.S. Moussounda and P. Ranson, Pressu re broadening of argon lines emitted by a high-pressure microwave discharge (Surfatron), J. Phys. B: At. Mol. Phys. 20, 949 (1987).
244 52. N. Konjevic, A. Lesage, Experimental Stark widths and Shifts for Spectral Lines of Neutral and Ionized Atoms, J. Phys. Chem. Ref. Data, Vol. 31, No. 3, (2002). 53. N. M. Shaikh, B Rashid, Measurement of electron density and temperature of a laser-induced zinc plasma, J.Phys.D:Appl.Phys.39 1384 (2006). 54. R W P McWhirter, R H Huddlestone S L Leonard, Plasma Diagnostic Techniques Academic Press, New York1965. 55. Ralchenko, Yu., Kramida, A.E., Reader, J., and NIST ASD Team (2008). NIST Atomic Spectra Database (v ersion 3.1.5), [Online]. Available: http://physics.nist .gov/asd3 [2008, August 29]. National Institute of Standards and Technology, Gaithersburg, MD. 56. W. Liu, H Xu, G. Mejean, Y. Kamali, J. Diagle, A. Azarm, P. Simard, P. Mathieu, G. Roy, S. Chin, Effici ent non-gated filament-induced breakdown spectroscopy of metallic sample, Spectrochimica Acta B 62, 76-81, (2007). 57. U. Panne, R. Neuhauser, C. Haisch, H. Fink, R. Niessner, Remote analysis of mineral melt by Laser-Induced plasma spectroscopy, Applied Spectroscopy, 56(3), 375, (2002). 58. D. N. Stratis, K. L. Eland, and S. M. Angel, "Enhancement of aluminum, titanium, and iron in glass using preablation spark dual-pulse LIBS," Appl. Spectrosc. 54, 1719 (2000). 59. J. Scaffidi, W. Pearman, J. C. Carter, and S. M. Angel, "Observations in collinear femtosecondnanosecond dual-pulse laser-induced breakdown spectroscopy," Appl. Spectrosc. 60, 65 (2006). 60. J. Scaffidi, J. Pender, W. Pearman, S. R. Goode, B. W. Colston, Jr., J. C. Carter, and S. M. Angel, "Dual pulse la ser-induced breakdown spectroscopy with combinations of femtosecond and nanosecond laser pulses," Appl. Opt. 42, 6099 (2003). 61. J. Gonzalez, C. Y. Liu, J. H. Yoo, X. L. Mao, and R. E. Russo, "Double-pulse laser ablation ICP-MS and LIBS," Spectrochim. Acta. 60, 27 (2005). 62. P. Garcia, J. Vadillo, J. Laserna, Real-time monitoring of high temperature corrosion in stainless steels by open-pa th laser-induced plasma spectrometry, Applied Spectroscopy, 58 (11), 1347 (2004). 63. S. Palanco, S. Conesa, J. Laserna, Analytical control of liquid steel in an induction melting furnace using a remote laser induced plasma spectrometer, J. Anal. At. Spectrom. 19, 462 (2004).
245 64. C. Lorez-Moreno, S. Palanco, J. Laserna, Remote laser-induced plasma spectrometry for elemental analysis of samples of environmental interest, J. Anal. At. Spectrom. 19, 1479 (2004). 65. V. I. Babushok, F. C. DeLucia Jr., J. L. Gottfried, C. A. Munson, and A. W. Miziolek, "Double pulse laser ablation and plasma: Laser induced breakdown spectroscopy signal enhancement," Spectrochimica Acta Part B-Atomic Spectrosc. 61, 999 (2006). 66. I. B. Gornushkin, B. W. Smith, H. Nasajpour and J. D. Winefordner, Identification of Solid Materials by Correlation Analysis Using a Microscopic Laser-Induced Plasma Spectrometer, Anal. Chem. 71, 5157 (1999). 67. L. St-Onge, R. Sing, S. Bechard and M. Sabsabi, Carbon emissions following 1.064 m laser ablation of graphite and orga nic samples in ambient air , Appl. Phys. A 69, 913 (1999). 68. A. Portnov, S. Rosenwaks and I. Bar, Emission Following Laser-Induced Breakdown Spectroscopy of Organic Co mpounds in Ambient Air, Appl. Opt. 42, 2835 (2003). 69. A. R. Boyain-Goitia, D. C. S. Beddows B. C. Griffiths and H. H.Telle, SinglePollen Analysis by Laser-Induced Breakdown Spectroscopy and Raman Microscopy, Appl. Opt. 42, 6119 (2003). 70. D. R. Anderson, C. W. McLeod and T. A. Smith, Rapid survey analysis of polymeric materials by laser-induced pl asma emission spectrometry, J. Anal. At.Spectrom 9, 67 (1994). 71. J. D. Hybl, G. A. Lithgow and S. G. Buckley, Laser-Induced Breakdown Spectroscopy Detection and Classifica tion of Biological Aerosols, Appl. Spectrosc 57, 1207 (2003). 72. R. Sattmann, I. Monnch, H. Krause, R. Noll, S. Couris, A. Hatziapostolou, A. Mavromanolakis, C. Fotakis, E. La rrauri and R. Miguel, Laser-Induced Breakdown Spectroscopy for Polymer Id entification, Appl. Spectrosc. 52, 456 (1998). 73. M. Tran, Q. Sun, B. W. Smith and J. D. Winefordner, Determination of C:H:O:N ratios in solid organi c compounds by laser-induced plasma spectroscopy, J. Anal. At.Spectrom.16, 628 (2001). 74. F. Ferioli, P. V. Puzinauskas and S. G. Buckley, Laser-Induced Breakdown Spectroscopy for On-Line Engine Equivalence Ratio Measurements, Appl. Spectrosc 57, 1189 (2003).
246 75. C. Aragon, J. A. Aguilera and J. Ca mpos, Determination of Carbon Content in Molten Steel Using Laser-Induced Brea kdown Spectroscopy, Appl. Spectrosc 47, 606 (1993). 76. R. Gronlund, M. Lundqvist, S. Svanberg, Remote Imaging Laser-induced Breakdown spectroscopy and Laser-Induced Fluorescence Spectroscopy Using Nanosecond Pulses from a Mobile Lida r System, Applied Spectroscopy, 60(8), 853 (2006). 77. Ph. Rohwetter, K. Stelmaszczyk, I. Woste, R. Ackermann, G. Mejean, E. Salmon, J. Kasparian, J, Yu, J. Wolf, Fillament-induced remote surface ablation for long range laser-indu ced breakdown spectro scopy operation, Spectrochimica Acta Part B, 60, 2005, 1025 (2005). 78. K. Stelmaszczyk, P. Rohwetter, Long distance remo te laser-induced breakdown spectroscopy using filamentati on in air, Applied Physics Letters, 85(18), 3977 (2004). 79. D. Leonard, Observation of Raman s cattering from the atmosphere using a pulsed nitrogen ultraviolet laser, Nature 216,142 (1967). 80. S. Sharma, New trends in telescopic remote Raman spectroscopy, Spectrochimica Acta Part A 2007, in press 81. J.F Ready, Effect of High Power Laser Radiation Academic Press, New York (1971). 82. G. Bekefi, Principle of laser plasma interaction John Wiley& Sons, New York 1976. 83. W.L. Kruer, The Physics of laser plasma interaction, Addison-Wesley, New York 1988. 84. R.H. Huddlestone, S. L Leonard, Plasma Diagnostics Techniques Academic Press, New York 1965. 85. A.P. Thorne, Spectrophysics, Chapman and Hall, London 1988. 86. R. Waterbury, A. Pal, D. K. Killinger, J. Rose, E. Dottery Enhancement of 266 nm laser induced LIBS emission using a simultaneous 10.6 m CO2 laser pulse, LThB5, OSA:LACSEA St Petersburg Fl (2008).
247 87. R. Waterbury, A. Pal, D. K. Killinger, J. Rose, E. Dottery, G. Ontai Standoff LIBS measurements of energetic materi als using a 266 nm ex citation laser, in Proceedings of Chemical, Biological, Radiological, Nuclear, and Explosives, A.W Fountain, P.J. Gardner, Proc. SPIE Vol. 6954 ,409 (2008). 88. E. Dottery, A. Pal, R. Waterbury, D. K. Killinger, J. Rose Standoff LIBS and Raman Measurements of Energetic Materials using a Single UV Excitation Laser, in Proceedings of Chemi cal & Biological Agent Detection and Identification, Lyle Malotky, Dan Kelley, IEEE 2008. 89. A. Pal, D. K. Killinger, M. Sigman, R emote sensing of explosive related gases using cw CO2 DIAL Lidar, LThB5, OSA:LACS EA St Petersburg Fl (2008). 90. A. Pal, D. K. Killinger, CW tunableCO2 differential absorption lidar for remote sensing of gaseous TATP, ISSSR 2008. 91. A. Pal, D. K. Killinger, R. Waterbury, E. Dottory, Enhanced LIBS plasma emission and temperature for sta ndoff detection using a 266 nm and 10 m CO2 laser system, To be submitted for publication. 92. A. Pal, C. D. Clark, M. Sigman, D. K. Killinger, CO2 DIAL lidar system for remote sensing of TATP related gases Accepted in Applied Optics. In press.
APPENDIX A. OPERATIONAL AMPLIFIER FOR THE MCT DETECTOR. An operational amplifier for the liquid nitrogen cooled Mercury Cadmium Telerium detector was made using low noise two JFET OP Amps with appropriate resistors to supply a BIAS current of around 50mA and amp lification of 1000x. The circuit diagram used is shown above in Fig 1.1 Figure 1.1 Circuit Diagram of OPAmp used for the MCT detector Wide Bandwidth Dual JFET Input Operational Amplifier provides very low input bias and offset currents. The connect ion diagram is given below. 248
249 Figure 1.2 Connection Pin Diagram for the OP Amp used. APPENDIX B. LABVIEW INTERFACE PROGRAM USED TO ACQUIRE AND STORE DATA FROM THE DIAL SYSTEM. The signal was collected at the Lecroy Waverunner (model No. LT344, 500 MHz) oscilloscope and it was interfaced to the computer with a GPIB card. A labview program was written by the author for extract ion of the signals and the calculation of transmission and absorption in realtime. The program was very user friendly and enables the user to monitor either or all th e siganals obtained from the three detectors used. Figure 2.1 and Fig. 2.2 shows the front panel and the bl ock diagram of this program.
250Figure 2.1 Front Panel
251 Figure 2.2 Block Diagram.
ABOUT THE AUTHOR Avishekh Pal grew up in India and receive d a Bachelor of Science degree with a major in Physics from St Xaviers Colle ge, Bombay, India in 1999 and obtained a Masters of Science degree in Physics in theore tical solid states from the Indian Institute of Technology Bombay, India in 2001 and obtained a second Masters of Science degree in Physics from University of South Florida, U.S.A. in 2005 under the guidance of Prof. Dennis K. Killinger. He completed an industrial practicum at Alakai Consulting and Engineering, Largo, Florida as part of the Applied Physics practical training. He has presented his results at several technical confer ences including SPIE, OSA and IEEE.