<%BANNER%>

Microsecond Pulsed Grimm Glow Discharge Time-of-Flight Mass Spectrometry Study of Aerosols Generated by Nebulization and...

Permanent Link: http://ufdc.ufl.edu/UFE0024978/00001

Material Information

Title: Microsecond Pulsed Grimm Glow Discharge Time-of-Flight Mass Spectrometry Study of Aerosols Generated by Nebulization and Laser Ablation
Physical Description: 1 online resource (182 p.)
Language: english
Creator: Fani-Pakdel, Farzad
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2009

Subjects

Subjects / Keywords: cathode, cds, dark, discharge, flight, gd, gdms, geometry, glow, la, lagdms, laser, mass, modification, ms, nd, pulsed, space, spectrometry, time, tof, yag
Chemistry -- Dissertations, Academic -- UF
Genre: Chemistry thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: This dissertation discusses the experimental results for interaction of particles with the glow discharge plasma. The objective is to investigate the potential ability of a glow discharge (GD) for ionization of aerosols. In this work, a Grimm-type microsecond pulsed-GD source coupled with a time-of-flight mass spectrometer (TOFMS) were used for producing and detecting the ions formed after the introduction of the aerosol stream into the plasma. MICROSECOND PULSED GRIMM GLOW DISCHARGE TIME-OF-FLIGHT MASS SPECTROMETRY STUDY OF AERSOLS GENERATED BY NEBULIZATION AND LASER ABLATION At first, as a source for aerosols, an ultrasonic nebulizer-dehydrator unit was used to generate simple salt particles such as cesium iodide. The geometrical design of choice was found to be a modified Grimm-type source in which particles enter the plasma through an orifice in the middle of the cathode and instead of a conventional cylindrical anode, an anode with four off-centered holes was used. The effect of argon flow, source pressure, pulse potential and repeller delay on signal intensity was studied for NaI and CsI particles. The optimum condition was used for quantitative studies. The signal had a linear correlation with the concentration of the nebulized salt solution. Heating the cathode was found to improve the signal reproducibility and also sensitivity (approximately 1.3 times). At the optimum condition and heating the cathode to 75 degrees Celsius, a sensitivity of approximately 1.8 mV/mM for cesium was achieved. Another interesting observation was that the higher-mass ions such as cesium and iodide resulted in larger signals compared to lower-mass ions such as sodium (approximately 10:1 ratio). It was concluded that this was mainly due to the fact that more massive ions experience less scattering at the skimmer orifice. This discrimination between higher-mass and lower-mass ions was reduced by using a larger skimmer orifice (Cs+: Na+ = 2:1). Finally, an aerosol stream was generated by laser ablation of solids in an external cell. The ablation lasers were a power-chip Nd:YAG laser and a Q-switched, flashlamp-pumped Nd:YAG laser. The samples investigated were aluminum, stainless steel, brass, bismuth alloy, ceramic and pressed pellets of inorganic salts. The effects of laser pulse frequency, energy and discharge potential on particle ionization were studied. Due to better signal stability and intensities, the laser ablation was found to be a more suitable method for aerosol generation when glow discharge is the ion source.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Farzad Fani-Pakdel.
Thesis: Thesis (Ph.D.)--University of Florida, 2009.
Local: Adviser: Omenetto, Nicolo.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2009
System ID: UFE0024978:00001

Permanent Link: http://ufdc.ufl.edu/UFE0024978/00001

Material Information

Title: Microsecond Pulsed Grimm Glow Discharge Time-of-Flight Mass Spectrometry Study of Aerosols Generated by Nebulization and Laser Ablation
Physical Description: 1 online resource (182 p.)
Language: english
Creator: Fani-Pakdel, Farzad
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2009

Subjects

Subjects / Keywords: cathode, cds, dark, discharge, flight, gd, gdms, geometry, glow, la, lagdms, laser, mass, modification, ms, nd, pulsed, space, spectrometry, time, tof, yag
Chemistry -- Dissertations, Academic -- UF
Genre: Chemistry thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: This dissertation discusses the experimental results for interaction of particles with the glow discharge plasma. The objective is to investigate the potential ability of a glow discharge (GD) for ionization of aerosols. In this work, a Grimm-type microsecond pulsed-GD source coupled with a time-of-flight mass spectrometer (TOFMS) were used for producing and detecting the ions formed after the introduction of the aerosol stream into the plasma. MICROSECOND PULSED GRIMM GLOW DISCHARGE TIME-OF-FLIGHT MASS SPECTROMETRY STUDY OF AERSOLS GENERATED BY NEBULIZATION AND LASER ABLATION At first, as a source for aerosols, an ultrasonic nebulizer-dehydrator unit was used to generate simple salt particles such as cesium iodide. The geometrical design of choice was found to be a modified Grimm-type source in which particles enter the plasma through an orifice in the middle of the cathode and instead of a conventional cylindrical anode, an anode with four off-centered holes was used. The effect of argon flow, source pressure, pulse potential and repeller delay on signal intensity was studied for NaI and CsI particles. The optimum condition was used for quantitative studies. The signal had a linear correlation with the concentration of the nebulized salt solution. Heating the cathode was found to improve the signal reproducibility and also sensitivity (approximately 1.3 times). At the optimum condition and heating the cathode to 75 degrees Celsius, a sensitivity of approximately 1.8 mV/mM for cesium was achieved. Another interesting observation was that the higher-mass ions such as cesium and iodide resulted in larger signals compared to lower-mass ions such as sodium (approximately 10:1 ratio). It was concluded that this was mainly due to the fact that more massive ions experience less scattering at the skimmer orifice. This discrimination between higher-mass and lower-mass ions was reduced by using a larger skimmer orifice (Cs+: Na+ = 2:1). Finally, an aerosol stream was generated by laser ablation of solids in an external cell. The ablation lasers were a power-chip Nd:YAG laser and a Q-switched, flashlamp-pumped Nd:YAG laser. The samples investigated were aluminum, stainless steel, brass, bismuth alloy, ceramic and pressed pellets of inorganic salts. The effects of laser pulse frequency, energy and discharge potential on particle ionization were studied. Due to better signal stability and intensities, the laser ablation was found to be a more suitable method for aerosol generation when glow discharge is the ion source.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Farzad Fani-Pakdel.
Thesis: Thesis (Ph.D.)--University of Florida, 2009.
Local: Adviser: Omenetto, Nicolo.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2009
System ID: UFE0024978:00001


This item has the following downloads:


Full Text

PAGE 1

1 MICROSECOND PULSED GRIMM GLOW DISCHARGE TIME OF -FLIGHT MASS SPECTROMETRY STUDY OF AERSOLS GENERATED BY NEBULIZATION AND LASER ABLATION By FARZAD FANI -PAKDEL A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSI TY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2009

PAGE 2

2 2009 Farzad Fani -Pakdel

PAGE 3

3 To my mom, Farideh Arjmand, and My dad, Manouchehr Fani Pakdel To my brothers, Kiumars and Kiarash To my sister, Azar

PAGE 4

4 ACKNOWLEDGMENTS I thank my advisor, Professor Nicolo Omenetto, for giving me the opportunity to work in his lab and for supporting my research over the course of the last five years. He provided me w ith insightful advice throughout my education, and always gave encouragement when I faced difficulties in my research. He analyzed my results with extreme thoughtfulness and suggested better approaches and new ideas. I also learned from him to have the cou rage to address my results scientifically. I will always appreciate his patience, wisdom and kindness; lo ringrazio dalle profondit del mio cuore. I also want to mention my previous advisor, Dr. Jason Telford at the University of Iowa, where I earned my M aster of Science in chemistry. From him, I learned lots of important things about approaching problems. He taught me that it is important not to be afraid of problems and always think of things as if they are simple and manageable. I will never forget his wisdom and kindness and will always appreciate being his student. I would also like to thank Professor James Winefordner for his scientific and moral support during my years of study. He was always a source of energy and encouragement. Dr. Ben Smith was al ways a reliable source of advice in science and academia. I will always remember his kindness and hospitality upon first applying to join the chemistry department at the University of Florida. I also thank Dr. Igor Gornushkin for his contributions to my re search and for his friendship. I am grateful for the additional guidance by Dr. Harrison in the field of glow discharge and his kindness in lending me beneficial books and also allowing me to use some of the images from his group for this dissertation. I t hank Dr. Kevin Turney for helping me at the beginning with the basics of the glow discharge mass spectrometer.

PAGE 5

5 I express my appreciation to past and current students in the Omenetto -Smith Winefordner group for their friendship and support during these yea rs: Jonathan Merten, Ronald Whiddon, Dr. Mariela Rodriguez, Dan Shelby, Anderson Warren, Heh young moon, Dr. Benoit Lauly, Dr. Nick Taylor, Dr. Kate Herrera, Dr. Lydia Edwards, Akua Oppong-Anane, Dr. Joy Guingab and Lydia Edwards Jonathan was a great frie nd, and excelled in aiding me with any and all technical issues, as well as fielding my questions about optics. I always enjoyed discussing scientific and social matters with him. Through his electronics expertise, Ron helped me a lot with my research and the time -of -flight instrument. Mariela and Kate were good friends who imparted a lot of useful information to me. Mariela was always a source of energy and encouragement as well. Benoit and Nick always helped me with my optical questions. Dan Shelby was a great friend with whom I enjoyed discussing experimental matter s. Andrew was a great friend who was always there to help when asked Heh young was a kind friend and I specially thank her for filming my final defense. I also thank Lydia and D an for helping me with my tie before the oral exam and final defense, respectively. I also thank my PhD committee members, Dr. Yost, Dr. Powell, Dr. E y ler, Dr. Hahn and the late Dr. Zaman for their scientific contributions to my dissertation. Dr. Yost is a great teacher, and my interest in mass spectrometry was really shaped after taking his class. He was always available to answer my random questions. Dr. Powell is a very friendly scientist that always had his door open to helping me solve problems with my instruments. H e was often the first person I would seek advice from. I also thank Dr. Eyler for supporting me when I needed help. He also helped me with editing my dissertation with accuracy and extreme care. I thank him for reading my work carefully and suggesting usef ul changes. I thank Dr. Hahn for sharing interesting ideas during my final defense and also during the laser ablation symposium in New

PAGE 6

6 Orleans. Dr. Zaman was a very kind scientist at the Department of Material S cience and I always enjoyed talking to him. I will always remember him and wish the best for his wife and family. I also thank Dr. Martin for what I learned about nano -science while in his group, especially making conical nanopores. I thank the Department of Chemistry at the University of Florida f or supporting me during the last 5 years with teaching assistantship. I thank all the kind faculty members whom I was a teaching assistant for: Dr. Young, Dr. Horvath, Dr. Keaffaber, Dr. Mitchell, Dr. Christou, Dr. Fanucci, Dr. Angerhofer, Dr. Ucak and fin ally Dr. Williams. I thank Dr. Kathryn Williams for being a very kind and considerate supervisor. During my time as her TA, I learned a lot of useful techniques and concepts in analytical chemistry while teaching for her Without her support I would have a very difficult time. Dr. Ucak was also a good friend who always gave me her support. I would like to acknowledge Joe Shalosky, Brian Smith and Todd Prox in the machine shop, Steve n Miles and Larry Harley in the electronic shop and Stephen Pritchard and J oseph Carvson in IT shop for providing me with technical support during my research. Without Joes help, many designed parts that helped me during my research would not be built. I would also like to mention the kind and effective help by Jeanne Karably an d Antoinette Knight, the analytical chemistry division secretaries. Without their help things would not be as easy. It is important to remember the kind consideration of LECO Corporation in providing technical support in maintenance and repair of the Renai ssance time -of -flight mass spectrometer. I especially thank Dr. Eric Oxley for his kind attention during my communications with LECO. I also like to thank Dr. Fred Smith from CETAC Corporation for providing me with technical information on the nebulizer -de hydrator. I express my appreciation to the kind members of Dr.

PAGE 7

7 Hieftje s group at Indiana University, Dr. William Wetzel and Dr. Steven Ray, who always helped me with troubleshooting the time of -flight mass spectrometer. I would like to acknowledge my chem istry high school teacher Mr. Soleymani who was one of the main reasons for my interest in chemistry. Also, I thank my undergraduate advisors during Iranian National Chemistry Olympiad and undergraduate studies, Dr. Abedini, Dr. Seyyedi, Dr. Nejat, Dr. Agh aei, Dr. Boghaei, and Dr. Maghari in the University of Tehran and the Sharif University of Technology in Iran. Without their guidance it would not have been possible for me to do my best. Finally, I thank my family and friends for their support during the last six years. I truly appreciate the friendship of my dear Iranian friends in Gainesville: Dr Masoumeh Rajabi ( also known as Masi), Ashkan Behnam, Babak Mahmoodi, Saeed Moghaddam and Ayyoub Mehdizadeh. Ashkan helped me with modeling the electric field and Ayyoub was nice enough to run a modeling of flows in fluids. Masi was a great help in editing my dissertation and pushing me to get things done. She is a very strong and giving person. Babak helped me with processing my mass spectra on computer using M ATLAB. Saeed was always a great help with my computer related issues. I also like to thank my former students and friends K atherine Winter, Samantha McCall, Alexandria Berry, Kasia Cudzilo, Chenan Zhang and Bernadette Boac for their friendship. Katie Winte r and Sam helped me with parts of this dissertation in editing the English. My friend Heera Sharma was also a great help in editing Chapter 5 of my dissertation. I also thank Valori Chapman for doing a good job in helping me with typing and editing. My goo d friend Julie heaweon Park was also a great friend who always encouraged me not to give up and try hard.

PAGE 8

8 Last, but not the least, I express gratitude to my parents, Manouchehr and Farideh, my brothers Kiumars and Kiarash, my sister Azar and my brother in law Mohamad -Reza and my sisters -in -law Shokofeh and Atusa for their encouragement during all stages of my life. My beautiful nephews and nieces Alvand, Sahand, Armita and Raha, were always inspiration when I looked at their photos. I cannot wait to see and hug them in person. I thank my dad for being a kind father and always teaching me interesting things with his wisdom. I especially thank my mom for always caring for me and doing all she can to keep me healthy and strong. She was always there to support m e during all the years of my education. For bringing and sending me food from thousands of miles away and for cooking for me when she visits, I am particularly grateful.

PAGE 9

9 TABLE OF CONTENTS page ACKNOWLEDGMENTS .................................................................................................................... 4 LIST OF TABLES .............................................................................................................................. 12 LIST OF FIGURES ............................................................................................................................ 13 ABSTRACT ........................................................................................................................................ 19 CH A P T E R 1 GLOW DISCHARGE ................................................................................................................. 21 Introduction ................................................................................................................................. 21 Different Regions of Glow Discharge ....................................................................................... 21 Cathode Dark Space ............................................................................................................ 22 Negative Glow ..................................................................................................................... 23 Faraday Dark Space ............................................................................................................. 24 Glow Discharge Reactions and Processes ................................................................................. 25 Cathode Sputtering .............................................................................................................. 25 Collisional Re actions in Plasma ......................................................................................... 28 Electron behavior ......................................................................................................... 29 Sputtered species excitation and ionization ................................................................ 29 Excitation ...................................................................................................................... 30 Ionization ...................................................................................................................... 30 Modes of Operation..................................................................................................................... 31 Direct Current ...................................................................................................................... 31 Pulsed Discharge .................................................................................................................. 32 Radio Frequency .................................................................................................................. 32 Source Configurations ................................................................................................................ 32 Hollow Cathode Geometry ................................................................................................. 33 Diode (Coaxial) Geometry .................................................................................................. 34 Grimm Geometry ................................................................................................................. 34 2 EXPERIMENTAL DETAILS AND BACKGROUND ........................................................... 46 Introduction ................................................................................................................................. 46 Aerosols ....................................................................................................................................... 46 Aerosol Generation .............................................................................................................. 46 Nebulizer -Dehydrator Efficiency ....................................................................................... 47 Particle Size Distribution Measurements ........................................................................... 49 Microscopy method...................................................................................................... 49 Particle impaction method ........................................................................................... 50 Light interaction method .............................................................................................. 50 Electrical property method .......................................................................................... 50

PAGE 10

10 Sedimen tation method ................................................................................................. 51 Limitations of Particle Sizing Methods .............................................................................. 51 Presentation of Particle Size Measurement Results .......................................................... 52 Glow Discharge Source .............................................................................................................. 54 Mass Spectrometer ...................................................................................................................... 54 Modifications ....................................................................................................................... 55 Repeller Modulator Timing ................................................................................................. 56 TOF -MS Detector and Collecting Mass Spectrum ............................................................ 58 3 A NO VEL PULSED GRIMM TYPE GLOW DISCHARGE SOURCE FOR IONIZATION OF SALT AEROSOLS ..................................................................................... 69 Introduction ................................................................................................................................. 69 Experimental ................................................................................................................................ 71 Aerosol Generation and Particle Size Distribution ............................................................ 71 Mass Spectrometer ............................................................................................................... 72 Grimm T ype GD Ion Source .............................................................................................. 73 Cathode modification ................................................................................................... 73 New anode design ........................................................................................................ 73 Pressure and Flow Measurements ...................................................................................... 74 Results and Discussion ............................................................................................................... 74 Anode A vs. B ...................................................................................................................... 75 Cathode Dark Space (CDS) ................................................................................................ 76 Effect of the Flow and Source Pressure ............................................................................. 77 Effect of the First Stage Pressure ........................................................................................ 79 Effect of the Cathode Potential Pulse Width ..................................................................... 79 Effect of the Cathode Potential ........................................................................................... 80 CsI vs. NaI at Different Repeller Delay Times .................................................................. 80 Ionization Mechanism and Sodium vs. Cesium Signal Comparison ............................... 81 4 PUL SED GLOW DISCHARGE MASS SPECTROMETRY OF PARTICLES: FUNDAMENTAL STUDIES AND DIAGNOSTICS ............................................................. 96 Introduction ................................................................................................................................. 96 Experimental ................................................................................................................................ 97 Results and Discussion ............................................................................................................... 97 Effect of Solution Concentration ........................................................................................ 97 Relative Sensitivity Factors ................................................................................................. 98 Effect of Skimmer Orifice ................................................................................................... 98 Effect of Cathode Temperature ........................................................................................... 99 Particle Vaporization Efficiency ......................................................................................... 99 Effect of Background Salt ................................................................................................. 103 Effect of Particle Ionization on Dischar ge Species ......................................................... 103 Effect of Cathode Potential Particle Ionization Mechanism .......................................... 104

PAGE 11

11 5 PULSED GRIMM GLOW DISCHARGE TIME OF FLIGHT MASS SPE CTROMETRY STUDY OF AEROSOLS GENERATED BY LASER ABLATION OF SOLID SAMPLES .............................................................................................................. 119 Introduction ............................................................................................................................... 119 Glow Discharge and Lasers .............................................................................................. 120 LA -ICP -MS ........................................................................................................................ 121 Ideal LA -ICP -MS System ................................................................................................. 121 Fractionation ...................................................................................................................... 121 Particle Size Effect And Use Of Femto -Second Lasers .................................................. 122 Carrier Gas of Choice ........................................................................................................ 122 Laser Pulse Duration ......................................................................................................... 123 Experimental .............................................................................................................................. 123 Ablation Cell ...................................................................................................................... 124 Lasers .................................................................................................................................. 124 Materials ............................................................................................................................. 125 Results and Discussion ............................................................................................................. 125 Relativ e Signal Intensities for Different Matrices ........................................................... 128 The Effect of Cathode Potential on Particle Ionization ................................................... 129 Crater Shape in Bismu th Sample ...................................................................................... 130 Effect of Laser Pulse Energy ............................................................................................ 130 Investigation of Particle Ionization Mechanism .............................................................. 130 Laser Back Ablation of Thin Films .................................................................................. 131 Effect of Laser Ablation and Particle Introduction on Other Discharge Species .......... 133 6 CONCLUSION AND FUTURE WORK ................................................................................ 160 APPENDIX: COMPLEMENTARY FIGURES ............................................................................. 164 LIST OF REFERENCES ................................................................................................................. 176 BIOGRAPHICAL SKETCH ........................................................................................................... 182

PAGE 12

12 LIST OF TABLES Table page 1 1 Ionization processes in a glow discharge plasma. Ioni zation is characterized by primary and secondary ionization processes. ....................................................................... 36 2 1 Results of efficiency measurements for ultrasonic nebulizer dehydrator. ......................... 60 2 2 Results of nebulizer dehydrator efficiency using condu ctivity measurements .................. 60 2 3 Typical operating conditions for time -of -flight mass spectrometer. ................................... 61 3 1 Summary of the experimental parameters. ........................................................................... 84 3 2 Thermodynamic properties of CsI and NaI.81 ...................................................................... 85 4 1 Dry particle mean volume diameters (DV) for different concentrations of C sI and NaI. ....................................................................................................................................... 105 5 1 The chemical composition of sample material used in laser ablation in m ass%. ............ 135 5 2 Relative intensities of particle ionization from ablation of different solid samples. ....... 135 5 3 Particle mean volume diameters (DV) for aerosol generated by laser ablati on of different samples .................................................................................................................. 136

PAGE 13

13 LIST OF FIGURES Figure page 1 1 Diagram of voltage vs. current re gi mes for various discharge ............................................ 37 1 2 Vari ous regions of a glow discharge ..................................................................................... 38 1 3 The three main regions in a simple glow discharge vs voltage distribution as a function of distance to cathode. ............................................................................................. 39 1 4 Major events in the glow discharge: Cat hode sputtering and ionizationexc itation in the negative glow. .................................................................................................................. 40 1 5 Cathode sputtering d ue to fast ion/fast atom impact ............................................................ 41 1 6 Sputter yield for various elements in the periodic table ...................................................... 41 1 7 The most common types of interaction in the plasma. ........................................................ 42 1 8 Direct and pulsed glow discharge modes are compared ...................................................... 43 1 9 Hollow cathode glow discharge source for mass spectrometry .......................................... 44 1 10 Diode geometry with a direct insertion probe (DIP) ............................................................ 44 1 11 Grimm glow discharge geometry for mass spectrometry. ................................................... 45 2 1 The schematics of the ultrasonic nebulizer dehydrator. ...................................................... 62 2 2 The schema tics of the particle impactor ............................................................................... 62 2 3 The differential mobility analyzer measures the particle size based on their mobility vs. their attraction w ithin elec tric field ................................................................................. 63 2 4 The condensation particle counter (CPC) counts the number of particles after t hey pass through the DMA unit ................................................................................................... 63 2 5 Comparing the results of diffe rent part icle sizing methods ................................................. 64 2 6 The diagram shows continuous ion source (red) vs. the modulator repeller gating pulses, for the original ICP -MS design ................................................................................. 64 2 7 The diagram shows the pulsed GD ion source (red) vs. the modulator repeller gating pulses ....................................................................................................................................... 65 2 8 The schematics of repeller and modulator with respect to the pulsed ion source .............. 65 2 9 The timing diagram for the time of -flight trigger vs. the repeller and modulator ............. 66

PAGE 14

14 2 10 The effect of changing different instrument variables on timing diagram ......................... 67 2 11 A sample mass spectrum from the instrument software ...................................................... 68 2 12 A sample mass spe ctrum from the oscilloscope ................................................................... 68 3 1 The overall schematic of the instrumentation for pulsed GDMS of particles .................... 86 3 2 Schematic of the cathode design for particle introduction .................................................. 86 3 3 Anode A vs. B ........................................................................................................................ 87 3 4 Deta iled view of the i onization source with anode B .......................................................... 87 3 5 The sharp edge of the cathode orifice, where the particles enter the discharge, was etched due to sputtering ......................................................................................................... 88 3 6 The Cs+ ion signal vs. repeller delay time from CsI aerosols for the two anode configurations ......................................................................................................................... 88 3 7 The electric field of anode A vs. B ....................................................................................... 89 3 8 Ar and Cs signal vs. repeller delay at three different aerosol flow rates into th e source ...................................................................................................................................... 90 3 9 Effect of the first stage pressure o n signal profile and intensity ......................................... 91 3 10 Pulse width, where the two pulse durations of 30 s and 150 s are compared for Cs signal ....................................................................................................................................... 92 3 11 Effect of the cathode potential on particle ionization for CsI and NaI ............................... 93 3 12 Signal vs. delay time for different elements ......................................................................... 94 3 13 Particles enter from the cathode (left). After vaporization, atomization and ionization, if atoms make it out of the CDS, they can ionize i n the negative glow ........... 94 3 14 The heavier Cs ion travels 6 times farther than Na before it goes back towa rds the cathode .................................................................................................................................... 95 4 1 Qu antitative study of cesium iodide with di fferent solution concentrations .................... 106 4 2 Quantitative study of sodium iodide with different solution concentrations ................... 106 4 3 Modeling the flow for the glow discharge source .............................................................. 107 4 4 Effect of skimmer orifice size and heating the cathode on sensitivity for Cs, I and Na .. 108

PAGE 15

15 4 5 Effect of cathode potential for CsI 4 mM at cathode temperature = 75C and skimmer or ifice diameter = 0.56 mm .................................................................................. 109 4 6 Effect of cath ode potential for NaI 4 mM. Cathode temperature = 75C and skimmer orifice diameter = 0.56 mm. ................................................................................................ 110 4 7 The change in signal vs. cathode potential for iodine ion using a 4.0 mM cesium iodide s olution. ..................................................................................................................... 111 4 8 Comparing the ion signal vs. salt concentration for the experimental results vs. partial vaporization model. .................................................................................................. 112 4 9 Mean volume diameters f or different CsI concentrations ................................................ 113 4 10 Cs+ ion signal from 2 mM CsI in different NaI backgrounds. .......................................... 114 4 11 Signal vs. concentration for Na+ ion at different NaI concentrations at the presence of 2 mM CsI background compared with no background. ..................................................... 115 4 12 Iodine ion signal resulted from dif ferent NaI solutions at the presence of 2 mM CsI. .... 116 4 13 The effect of particle introduction on argon and copper ion signals ............................... 117 4 14 Comparing the effect of cathode potential on different ion signals. ................................. 118 5 1 Overall experimental set up for LA GD -MS. Only the skimmer of timeof -flight is shown. ................................................................................................................................... 137 5 2 The geometrical dimensions of the ablation cell. All dimensions are in millimeters. ..... 137 5 3 Iron signal resulted from ablation of stainless steel using power -chip Nd:YAG laser .... 138 5 4 Mass spectrum resulted from ablation of stainless steel using the Q -switched, flashlamp -pumped Nd:YAG laser a t its maximum pulse power ...................................... 138 5 5 Monitoring the 56Fe ion signal as a fun ction of time for 300 seconds .............................. 139 5 6 Monitoring the 56Fe ion signal as a function of time for ablation of stainless steel u sing 11 single laser shots ................................................................................................... 140 5 7 Monitoring the 209Bi ion signal as a function of time for laser a blation of bismuth alloy ....................................................................................................................................... 141 5 8 Monitoring the 56Fe and 63Cu ion signals as a function of time for approximately 13 minutes o f stainless steel ablation ....................................................................................... 142 5 9 Iron, copper and zinc ion signals at 4 different time domains ........................................ 142

PAGE 16

16 5 10 The signal ratio for 56Fe to 63Cu, almost remains constant for the 13 minutes duration of laser ablation. ................................................................................................................... 143 5 11 The cathode signal from sputtering the brass cathode increases du e to ablation of a brass sample .......................................................................................................................... 143 5 12 Mass spectrum from l aser ablation of a Macor sample ..................................................... 144 5 13 Mass spectrum from laser abl ation of a bismuth alloy sample ......................................... 144 5 14 Cesium signal vs. mole % of cesium at different salt pellets. ........................................... 145 5 15 Particle size distributions for aerosols generated from different samples. ....................... 145 5 16 The effect of cathode pot ential on ionization of stainless steel particles (56Fe signal) is compared with its effect on cathodic sputtering of brass (63Cu signal). ....................... 146 5 17 The effect of cathode potential on ionizatio n of Macor particles. .................................... 147 5 18 The effect of cathode potential on ionization of salt particles. ......................................... 147 5 19 Crater image of bismuth sample after ablation by 219 laser shots at 10 Hz and maximum e nergy .................................................................................................................. 148 5 20 The crater profile shape and dimension measured by an optical profilometer. After ablation by 219 laser shots at 10 Hz and maximum laser energy. .................................... 149 5 21 The effect of laser pulse energy on LA -GD MS Bi ion signal of bismuth alloy particles. ................................................................................................................................ 150 5 22 The effect of laser pulse energy on 209Bi to 208Pb signal ratio. ......................................... 150 5 23 Memory effect investigation: The initiation of glow discharge after ablation of bismuth sample (305 shots) ended resulted in a spike in Bi signal. .................................. 151 5 24 Particle deposition and laser back ablation of bismuth alloy on quartz window. ............ 152 5 25 Ablation cell set up for laser back ablation of sample from the surface of the quartz window. ................................................................................................................................. 152 5 26 Ablation cell set up for laser back ablation of a thin stainless steel film locate d above a sample of aluminum. ......................................................................................................... 153 5 27 Signal vs. time for back ablation of stainless steel followed by regular abla tion of aluminum sample ................................................................................................................. 153 5 28 Crater diameters for stainless steel and aluminum. ............................................................ 154

PAGE 17

17 5 29 Spike in copper signal due to entrance of stainless steel particles into the discharge, by a single laser shot. ........................................................................................................... 154 5 30 Increase in copper signal due to entrance of a 20 second ablated package of stainless steel particles into the discharge. ......................................................................................... 155 5 31 The effect of ablation of stainless steel on Ar+ and ArH+ ion signals. The right hand Y axis is for the argon ion only. .......................................................................................... 156 5 32 The effect of bismuth alloy ablation on Ar+ ion si gnal. ..................................................... 157 5 33 The effect of 4 single shots of laser ablation of bismuth alloy on 63Cu+ ion signal. No significant change i n copper ion signal is observed ........................................................... 158 5 34 The effect of ablation of Macor on 40Ar+ and ArH+ ion signals. ....................................... 159 5 35 The effect of ablation of Macor on 63Cu+signal ................................................................. 159 A 1 Calibration of the digital flow meter by the soap bulb method. The real value of flow in standard cubic centimeter per min is shown. ................................................................. 164 A 2 Calibration of the argon flow into the glow discharge with respect to the first stage pressure ................................................................................................................................. 165 A 3 Signal intensities for Cs and I at three different concentrations of CsI solution are shown. ................................................................................................................................... 166 A 4 Repeller and modulator ringing fro m the oscilloscope is shown ...................................... 167 A 5 The sources of particle and ion loss before the skimmer of tim e -of -flight are shown. ... 168 A 6 Calculation of the number of ions expected to arrive at the detector assuming 100% ionization and ion transfer efficiency. ................................................................................ 168 A 7 The overall efficiency of the particle vaporization atomization ionization and ion transfer to detector is calculated. ......................................................................................... 169 A 8 Calibration graph for conductivity vs. concentration of CsI ............................................ 170 A 9 Result of the particle size distribution measurement for ablation of Macor. Relative abundances of dN/dD is compared to N ............................................................................. 171 A 10 The particle size distribution for Macor aerosol is shown as dV/dD ................................ 171 A 11 The effect of ablating stainless steel by the power -chip Nd: YAG laser on c oppe r signal ..................................................................................................................................... 172 A 12 Geometrical modification of cathode resulted in ion signal as high as 22 mV for Cs+ upon nebulization of a 4 mM CsI solution. ........................................................................ 173

PAGE 18

18 A 13 The Cs+ signal vs. delay time is sh own as well as the error bars ...................................... 174 A 14 The signal for Cs+ ion vs. CsI concentration show a linear dynamic range from 0.50 mM to at least 10.0 mM ....................................................................................................... 175

PAGE 19

19 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy MICROSECOND PU LSED GRIMM GLOW DISCHARGE TIME OF FLIGHT MASS SPECTROMETRY STUDY OF AERSOLS GENERATED BY NEBULIZATION AND LASER ABLATION By Farzad Fani -Pakdel August 2009 Chair: Nicolo Omenetto Major: Chemistry This dissertation discusses the experimental results for interaction of particles with the glow discharge plasma. The objective is to investigate the potential ability of a glow discharge (GD) for ionization of aerosols In this work, a Grimm type microsecond pulsed GD source coupled with a time -of -flight mass spectrometer ( TOFMS) w ere used for producing and detecting the ions fo rmed after the introduction of the aerosol stream into the plasma. At first as a source for aerosols, an ultrasonic nebulizer -dehydrator unit was used to generate simple salt particles such as cesium iodide The geometrical design of choice was found to be a modified Grimm type source in which particles enter the plasma through an orifice in the middle of the cathode and instead of a conventional cylindrical anode an anode with four of fcentered holes was used. The effect of argon flow, source pressure, pulse potential and repeller delay on signal intensity was studied for NaI and CsI particles. The optimum condition was used for quantitative studies. The signal ha d a linear correlatio n with the concentration of the nebulized salt solution Heating the cathode was found to improve the signal reproducibility and also sensitivity ( approximately 1. 3 times) At the

PAGE 20

20 optimum condition and heating the cathode to 75 degree s Celsius, a sensitivity of approximately 1.8 m V /mM for cesium was achieved. Another interesting observation was that the higher -mass ions such as c esium and i odide resulted in larger signals compared to lower -mass ions such as sodium (approximately 10:1 ratio). It was concluded that this was mainly due to the fact that more massive ions experience less scattering at the skimmer orifice. This discrimination between higher -mass and lower -mass ions was reduced by using a larger skimmer orifice ( Cs+: Na+ = 2:1). Finally an aeros ol stream was generated by laser ablation of solids in an external cell. The ablation lasers were a power -chip Nd:YAG laser an d a Q -switched, flashlamp pumped Nd:YAG laser. The samples investigated were aluminum, stainless steel, brass, bismuth alloy, cera mic and pressed pellets of inorganic salts. The effects of laser pulse frequency, energy and discharge potential on particle ionization were studied. Due to better signal stability and intensities, the laser ablation was found to be a more suitable method for aerosol generation when glow discharge is the ion source.

PAGE 21

21 CHAPTER 1 GLOW DISCHARGE Introduction When a cathode and an anode are immersed in an inert gas such as argon and sufficient electrical potential is applied to that gas, an electric breakdown occurs. The gas breakdown results in the formation of electrons and ions, which allows current to flow through the gaseous medium. This breakdown of the gaseous medium and the resulting current is called a discharge or plasma.13 Gas pressure, gas type, cathode material, temperature and electrode distance are all factors that play a role in determining the gas breakdown potential.4, 5 Electrons, positive ions and negative ions are all formed during breakdown and maintain the discharge. The cathode surface is bombarded by the positive ions and a sputtering event is initiated. Analyte atoms are released from the bulk sample due t o the sputtering and can then be analyzed. The discharge becomes self -sustaining due to charged species being continually produced by collisional events within the plasma .36 Voltage and current characteristics de fine the operating regimes of discharges and differentiate them from each other. The operating regimes of discharges, with features close to a glow discharge, are shown in Fig ure 1 1 For both the optical emission and mass spectrometry detection methods, t he discharge is maintained in the abnormal mode with the purpose of obtaining the greatest amount of analytical information. Throughout this document, an abnormal glow discha rge will be referred to as the glow discharge Different Regions of Glow Discharge The glow discharge is made up of multiple regions as shown in Figure 1 2 The formation of these regions depends on the configuration of the glow discharge cell, in particular the distance separating the electrodes. A highly luminescent positive column re gion is most

PAGE 22

22 important when the inter electrode distance is large enough; however, to sustain a glow discharge it is not required. As the distance between electrodes decreases, this region shortens and eventually disintegrates. Decreasing the electrode se paration also causes other regions of the discharge to be diminished, with the exception of the cathode dark space. The cathode dark space must exist to sustain a glow discharge. Most analytical glow discharge devices utilize short electrode distances, so they generally exhibit only the cathode dark space and negative glow regions. When the electrode separation is only a few times larger than the thickness of the cathode dark space, an obstructed glow occurs. This is a discharge in which mos t of the regions have collapsed .7, 8 A distorted dark space region and an unstable, or extinguished, plasma will occur with a further decreased electrode separation. Three prominent regions of the glow discharge include: the cathod e dark space, negative glow and Faraday dark space. Cathode Dark Space The cathode dark space is a region with low luminosity next to the cathode surface as shown in Figure 1 2 The negative cathode potential causes electrons to be repelled away from this region, which creates a positive space where most of the potential difference between the two electrodes is dropped.4 For this reason, the cathode dark space is often called the cathode fall region. Figure 1 3 shows a profile of potentia l distribution as a function of electrode distance for the three major regions of the glow discharge. The figure shows the large potential fall of the cathode dark space region. Electrons are accelerated away from the cathode dark space region and therefor e have too much energy for excitation reactions. An excitation cross -section shows a dip at these high energies ,5, 6 accounting for the low luminosity of the cathode dark space. Since electrons and ions gain extra energy by electric field, the energy distribution of charged species

PAGE 23

23 in this region does not follow a Maxwell distribution. As a result, the cathode fall region is far from a hydrodynamic equilibrium.7 A glow discharge cannot exist without the cathode dark space region, even though there is a lack of excitation and ionization processes in th is region. Electrons that are accelerated aw ay from the cathode dark space region are responsible for ionizing the gas (Ar) in other regions of discharge. These collisions create electrons, which may cause further ionization and the resulting argon ions are attracted to the cathode. These complex pr ocesses are a continuous cycle which leads to a self -sustaining discharge, originating from the acceleration of electrons in the cathode dark space region. Negative Glow The negative glow region is a large, bright region adjacent to the cathode dark space and is analytically the most important region of the glow discharge.2, 4 Both fast, highly energetic electrons and slow, thermal electrons enter the negative glow. Fast secondary electrons are the electrons that are created at the cathode and have passed through the cathode dark space without losing significant energy by collisions. Due to their inherent high energies, these electrons (group I electrons) are only capable of ionizing collisions. It has been shown that group I electrons attain e nergie s of ~20 25 eV and number densities of around 106 cm3 in the negative glow region.8 Another name for t hese electrons is MC (Mon te Carlo) electrons (page 161, reference 2).3 L ow -energy or thermal electrons are separated into two groups. Group II electrons are secondary electrons of gas phase ionization collisions that have electron temperatures of ~2 10 eV and number densities of about 107 108 cm3.8 Electrons from group I or group II that have been through several elastic and inelastic collisions within the plasma are termed group III electrons. These electrons have electron temp eratures of 0.05 0.6 eV and number densities in

PAGE 24

24 the range of 109 1011 cm3.8 Group II and group III electrons may either excite atomic species or ionize excite d state species.9 The number of electrons in the negative glow is typically matched by a similar number of positive ions4 and the combination of these free ions and electrons results in an essentially field free region. This region is represented in the potential distribution of a linear segment just above ground potential (Vo) as seen in Figure 1 3 It has been shown that plasma potentials within the negative glow region va ry less than 1 V for discharge voltages between 800 and 1000 V.8 In this field -free region, electrons ar e not significantly accelerated and are capable of causing excitation collisions resulting in a bright glow. The discharge gas and, to a lesser degree, the cathode material are responsible for the color of this glow.3 Both the discharge pressure and the distribution of electron energies affect the size of the negative glow region. The number of collisi ons within the discharge is increased with high operating pressures; this will result in shortening of the negative glow region. A wide range of electron energies, however, can extend the length of this region. Electron multiplication by ionization process es results in the largest population of electrons existing at the interface between the cathode dark space and the negative glow. The greatest emission intensity is also seen at this edge of the negative glow, which fades as a function of cathode distance. Mass spectrometers are able to analyze the high population of ions that exist at this negative glow interface.2, 4 Faraday Dark Space T he Faraday dark space region is the region just before the anode end of the negative glow .24 Electrons that make up the Faraday dark space are thermal electrons, having lost most of their energy due to excit ation and ionization collisions.4 The separation between the cathod e and anode is often so small that this region of low luminosity is not observed.

PAGE 25

25 Glow Discharge Reactions and Processes The glow discharge is a simplified analytical device full of chaotic processes involving multiple species of varying energy. Summarizin g these basic processes will be beneficial to understanding the experimental results in the following chapters. This section will present an overview of the mechanisms that occur within the spatial regions depicted in the previous section. Figure 1 4 illus trates the fundamental processes that will be described in this section, including sputtering, excitation and ionization. Cathode Sputtering Sputtering is the phenomenon that occurs when an energetic particle strikes a solid surface and results in the ini tiation of many processes. Sputtering is a substantial contributor to the analytical utility of the glow discharge. Ion bombardment of the sample surface causes atomization of the sample and provides an atomic population for later excit ation and ionization processes.4 An overview of the sputtering process, as well as quantitative principles for characterizing sputtering, are given in this section. Sputtering process : Figure 1 5 illustrates the particle impact, which can lead to extensive sample damage through atom rearrangement or particle implantation. The figure simplifies the sputtering process, since the incident argon ion typically undergoes several charge exchange reactions with neutral gas atoms before impacting the cathode surface .10 Argon atoms can therefore also contribute to a certain degree to the sputter yield. Upon impact, the kinetic energy of the incident particle (in this case, Ar+) is transferred to the sampl e (M) through collisional events, and the potential energy of the particle results in electron ejection. The incident argon ion can either be backscattered from the sample surface or will penetrate the sample and transfer its energy to the surface. If the velocity of the impinging ion is greater than 30 eV, material may be sputtered (i.e., ejected) from the sample surface in the form of atoms or molecules in neutral

PAGE 26

26 (Mo) or ionized (M+ or M-) states, or electrons may be ej ected.4 Sputtere d particles usually have low energies (few eV) and can be ejected at many possible angles. Positively charged ions (M+) make up about one percent of the total sputtered particle flux and are redeposited back onto the sample surface, depending on the negati ve cathode potential .11 Negative ions (M-) are accelerated away from the cathode surface. The ejected material is made up mainly of neutral sample atoms, which generally possess kinetic energies of approximately 5 10 eV .12 When these sample atoms are ejected, they lose their momentum to elastic collisions with discharge species due to the small mean free paths at the cathode surface (~0.1 mm for pressures of 0.1 10 torr). Harrison and Bruhn13, 14 have shown that up to 95% of these atoms may be deposited back onto the cathode surface. Some atom s are not re -de posited, and these can either diffu se into the negative glow region, which is essential for atomic emission, or undergo excitation or ionization processes, which is critical for mass spectrometry measurements. Sputter rate : The amount of sam ple ejected for a given analysis time is d efined as the sputter rate ( ng/s) and is given by the E quation 1 1 :2, 4, 5 t W q (1 1) W here g and t (s) is the total time of sputtering. Th e above equation is generally known as the net sputtering rate for glow discharge because it recognizes that the glow discharge inherently involves re -deposition of sputtered material .9, 15 Multipl e parameters affect sputtering rates, including voltage, current and pressure of the discharge. It is possible that the type of fill gas affects s puttering, since the mass of the sputtering agent depends on the fill gas. Sputtering rate can also be altered by the type of sample. Multi component alloys of the same matrix demonstrate sputtering rates dependent on the

PAGE 27

27 concentration of other minor const ituents contained within the sample. Preferably, all elements (and allo y s made of the same matrix) would exhibit the same sputtering rates. This is not the case ; however, and sputtering rates vary between elements, with the potential to adversely affect ca libration curves. The calibration curves can be normalized based on established sputtering rates .15 Sputter yield : The sputtering efficiency of a sample can a lso be characterized by calculating sputter yield (atoms/ions), which represents the number of sputtered atoms per incident sputtering atom/ion .2, 4 The target atom receives the energy from the incident ion through a transfer in the lattice (sputtered atom). The following is the expression for sputter yield (S ): 0 2 2 1 2 1 24 4 3 U E m m m m S (1 2 ) Where is the function of the ratio m2/m1 and the angle of collision by the incident ion with respect to the sputtering surfac e .16, 17 Energy of the incident ion and the surface binding energy are denoted by E and U0, respectively The sputter yield depends on the relationship between the ion mass (m1) and the target atom mass (m2). As ( m1/m2) approaches unity, the mass transfer term maximizes, and thus sputter yield increases An atoms electron concentration in the d -shell orbital is related to sputter yield. In theory, a greater d -shell filling is indicative of a greater sputter yield.2, 4, 19 In addition, the penetration depth of primary ions is partially controlled by the electronic structure of the target atoms. An incident ion can penetrate further int o a cathode that has a more open electronic structure, which results in less e fficient transmission of energy back to the cathode surface and, thus, a reduction in the sputtering process.

PAGE 28

28 Fig ure 1 6 is a collection of sputter yields for various elements at conditions similar to a glo w discharge, 400 eV argon ions.18 where S is the sputtering coefficient in atoms/ion s overall agreement with t he electron density of d-shell electrons. Note that the res ults in Fig ure 1 6 are for ion beam, vacuum sputtering and not GD.18 Sputter yield is affected not only by sputter rate but by the masses of incident ion and sputtered atom, incident ion energy, angle of incidence and the surface binding energy. More neutral atoms are available for ionization if there is a higher sputter yield. Collisional Reactions in P lasma M any of the processes within the glow discharge can be accounted for by collisions, and collisions also contribute to the self -sustaining nature of the plasma. There are two broad types of collisions that occur within the glow discharge: elastic and inelast ic.14 The simpler of these two types of collisions are elastic collisions, since kinetic energy is conserved. Energy is not transferred; therefore excitation and ionization processes within the plasma are not usual ly attributed to these types of collisions.2, 9 In order for excitation ( ionization ) to occur, an impinging electron would have to have more energy than that needed to excite ( remove ) an electron from the atom ground state ( excitation /ionization energy). I f the electron energy is insufficient, the electron will be deflected from the atom.2, 9 The significantly large differen ce between masses of electrons and atoms leads to a negligible amount of energy transferred during an electronatom elastic collision, as suggested by the energy transfer function. Even though elastic collisions are unable to cause ionization, they are able to redistribute kinetic energy and help to thermalize the plasma Inelastic collisions are the other type of collision within the glo w discharge plasma and involve energy transfer

PAGE 29

29 between plasma species.6, 7, 17, 18 Inelastic collisions are responsible for forming species of analytical interest and for maintaining the self -s ustaining nature of the plasma.4, 5 Electron behavior Within a glow discharge, electron s are accelerated away from the cathode, possibly colliding with gas atoms and causing ionization. However, many of the electrons formed from collisional processes are lost due to recombination .19 It is also possible for electrons to convey their energy to an atom, raising the internal energy of the atom to an excited s tate. When this newly excited atom relaxes to a ground state, a photon is released which contributes to the characteristic glow of the discharge. These photons can also be useful for investigating optical characteristics of the discharge. Electrons are gen erally accelerated with up to 2000 eV at 2 kV of kinetic energy, which is enough to ionize atoms within the discharge. For example, argon is primarily ionized by collisions with electrons and has ionization energy of 15.76 eV. This process is called electr on ionization and is illustrated in F igure 1 4 Mass spectrometry detects doubly charge d argon (Ar+2) ions, which have an ionization energy of 27.63 eV This show s that electron ionization is a k ey process within the discharge since only electrons have e nou gh energy to form these ions.9, 20 Sputtered species excitation and ionization Sputtering the surface of a sample with argon ions (and atoms) results in the atomization of a sample in a glow discharge. Atoms that are ejected from the sample surface can diffuse into the negative gl ow region of the discharge where they may go through a series of collisions with electrons, meta -stable atoms and other ions. It is these collisions that are responsible for excitation and ionization of the sample material, but for these to be possible in the first place, there must be an inelastic collision between an atom and a particle with kinetic or potential energy. There are multiple excitation and ionization mechanisms that can occur in the negative

PAGE 30

30 glow region; F igure 1 7 lists the three most comm on of such processes which can t ake on different roles depending on the operating conditions of the discharge.2, 4, 5 Excitation In general, thermal electrons in the virtually field -free negative glow region posses s low kinetic energy, mainly because of loss of energy due to multiple collisions. The electron energy distribution function (EEDF) can show the abundance of electrons at different energies. It has been shown that thermal electrons will likely result in ex citation mechanisms that preferentially involve energy transfer to levels less than or equal to 34 eV above ground state .8 The atomic transitions that occur in the ultraviolet -visible region are primarily caused by these ground state transitions, and they are wholly responsible for the emission from the negative glow. Glow discharge atomic emission spectrometry (GD -AES) can measure the photons that are released. E lectron impact causing electronic excitations can also be cited as a cause for the population of argon meta -stable atoms. Argon has two long -lived (approximately one millisecond) excited states with energies of 11.55 and 11.72 eV, respectively. A Penning c ollision20 describess the possibility that collisions between the meta -stable gas atoms and sample atoms could lead to energy transfer and resulting excitation of the sample atom. Figure 1 7 illustrates this process. The negative glow region has a third possible excitation mechanism, known as asymmetric charge exchange .21, 22 However, this mechanism has been found by most researchers to be too selective to contribute significantly to excitation processes .20 Ionization Ionization mechanisms within the negative glow region occur due to the collision mechanisms described previously. Multiple mass spectrom etric techniques can detect the ions that are formed through these mechanisms. The goal in atomic mass spectrometry is to produce and detect the maximum possible number of elementally uniform ions .4

PAGE 31

31 Electron ionization comprises of a sma ll portion of ionization within the negative glow. The mechanism a s shown in F igure 1 7 is essentially unselective since any atom can be ionized by an electron with sufficient energy. The cross -section curve for electron ionization is similar in shape an d magnitude for all elements of the periodic table.23, 24 However, as mentioned there are only a few electrons that will have sufficient energy to ionize metals with ionization energies between 5 10 eV. It is fo r this reason that electron ionization likely only makes up a small portion of ionization within a glow discharge. The dominant ionization mechanism in the glow discharge is known as Penning ionization .11, 20, 2528 Based on its energies, the argon meta stable atom should be able to sufficiently ionize most elements (with ionization energy smaller than 11.5 eV) Penning ionization is similar to electron ionization in that it is unselective for elements with lower io nization energies than the meta -stable atom energy. It has been shown through elemental quantitati ve studies that Penning ionization could account for about 50 to 95 percent of sample ionization in the discharge .29 Table 1 1 lists the other possible mechanisms for ionization in addition to electron and penning ionizations. Figure 1 7 shows that asymmetric charge exchange or charge transfer2 could ionize sample at oms in the negative glow region; however the process is highly selective and controversial and is not considered to be a significant ionization mechanism. The process of asymmetric charge exchange occurs when an electron is transferred from a sputtered atom to an argon ion but is likely only if the difference in ener gy between the species is small.17 M odes of O peration Direct Current By applying a constant negative potent ial to the cathode, a direct current glow discharge 19 is operate d. Fig ure 1 8 is a schematic of a direct current (dc ) discharge showing the area of

PAGE 32

32 sputtered species in comparison to the glow region. Typical operating conditions of a dc discharge are 1 3 kV and less than 30 mA current.9 Owing to the simplicity, reproducibility, cost -effectiveness and steady -state supply of ions produced, this powering mode is the most conventional. There are some disadvantages to the dc discharge, and these include the low operating power (<10 W) and the inability to analyze non conducting samples.3, 9 Pulsed Discharge Fig ure 1 8 shows a comparison of ion generation in a dc and pulsed discharge. Sputtering ions are created from the electrical breakdown of argon atoms after a repetitive high voltage pulse is applied to the cathode in pulsed mode and at pulsed initiation. The ions that are created participate in analyte ejection from the cathode surface. Individual pulse s create atom packet s, which is subjected to ionization These packets expand and diffuse away from the cathode with each pulse. T emporal profiles are created as a result of the variation in the elemental response in the induction period. The temporal profiles provide discrimination between gas and analyte ions.4, 5, 33 39 Benefits of the pulsed discharge include short term high power operation, enhanced sputter yield, greater excitation and ionization, less sample consumption and temporal resolution.9, 3033 Radio F r equency Non -conducting samples, such as glass and ceramics, can be analyzed using the radio frequency powering mode. In this mode, an alternating current applied at high pulse frequencies (1 MHz) induces a self biasing direct current (dc) potential.40, 41 Source Configurations The glow discharge has a wide array of applications, which has led to the development of multiple source configurations. Each configuration is able to efficiently analyze different sample shapes, collectively making them w ell -suited for a range of applications. Included in this section

PAGE 33

33 are descriptions of the hollow cathode, diode and Grimm design glow discharge configurations. E ach configuration has its own strengths and weaknesses. There is no one configuration that has p roved optimum for efficient analysis of all sample types. Hollow Cathode Geometry The oldest glow discharge configuration, dating back to the 1930s,34 is the hollow cathode discharge. Hollow cathode sources26, 28, 35 are reasonably successful f or atomic emission measureme nts but are rarely implemented for mass spectrometric applications .36 The plasma in a hollow cathode is restricted within a hollow cavity within the sample, forcing sputtering to occur on the sides of the cavity wa lls. Highly energetic atoms and electrons are trapped within the hollow area, and extended residence times lead to enhanced excitation and ionization of the sample atoms. Due to efficient ionization capabilities, hollow cathode configurations demonstrate d etection limits down to the sub -nano -gram range.37 Typically operating conditions for a hollow cathode configuration are voltages between 200500 V, currents of 10 100 mA and argon pressure s between 0.1 10 torr .38 Figure 1 9 shows a n early design of a hollow cathode configuration designed exclusi vely for mass spectrometry, be cause of the high population of ions that can be formed within the cavity. The figure also illustrates the confined nature of a hollow cathode plasma, which provides an improved population of sample atoms and ions which must then be extracted to from the hollow well and transported to the sampling region, additional steps that can prove complex. Eve n with the potentially complex and numerous s teps involved the potential advantages for hollow cathode mass spectrometry studies, which includ e enhanced detection limits provide justification for developing methods to avoid these concerns.

PAGE 34

34 Diode (Coaxial) Geometry Diode geometry is the most popular glow discharge configuration for mass spectrometry. It is not typically used for atom ic emission, however .39, 40 This design is usually called a pin -type source and is used to analyze sample types that can be mounted on the end of a direct insertion probe (DIP ). Figure 1 10 illustrates h ow the pr obe and sample assembly are introduced into the glow discharge chamber through a vacuum interlock. The discharge then forms on the end of the sample and ions are extracted from the negative glow region of the plasma. Diode geometries generally employ condi tions of 500 1000 V, 1 5 mA and 0.4 2.0 torr Ar pressures.38 In this configuration, the sample serves as the cathode, and the surrounding discharge chamber becomes the grounded anode. There are some limitations to this geo metry including problems with sample place ment, thermal effects and re -deposition on source components. Glow discharge applications that require layer analysis, such as depth profiling, are difficult using this configuration. Grimm Geometry The Grimm -type configuration is widely used for atomic e mission measurements and is gaining popularity for mass spectrometry applications. The Grimm -type configuration makes use of flat cathode geometry and was introduced by Grimm as an atomic em ission source in the late 1960s .35 In this configuration, the sample is pressed against an O ring on the flat cathode plate, ensuring adequate sealing and proper vacuum conditions, shown i n Figure 1 11. Util izing a constricted plasma, the discharge is laterally restricted to the sample surface and confined to the size of the surrounding anode. Maintaining a distance between the anode tip and sample that is smaller than the mean free path of the electrons allows the restriction of the plasma to the sample surface.41 The plasma is also confined in circumference by the cylindrical anode, allowing planar sputtering of the sample. Typical operating conditions of the Grimm type source are approximately 5001500V, 3 20 mA and 2 6 torr a r gon pressure .38

PAGE 35

35 The Grimm -type source provides a significant advantage over other source configurations due to the easy sample interchange .4245 There can be quick sample turnaround time and precise sample placement because samples are mounted externally. There are also no sample positioning concerns .35, 46 The Grimm configuration also has the ability to obtain surface and in depth analysis by properly controlling the discharge parameters to obtain planar sputtering15, 42, 47 and can also analyze thin sample layers with pulsed operation .15, 42, 44, 48, 49 The concerns for using the Grimm type configuration include the ana lysi s of flat samples, since a flat smooth sample surface is required to guarantee adequate sealing and any other sample shapes with smaller diameters than the anode diameter are less adaptable for analysis. For mass spectrometric applications, ion transport is also a limitation of the Grimm configuration.

PAGE 36

36 Table 1 1 Ionization process es in a glow discharge plasma. Ionization is characterized by primary and secondary ionization processes. I. Primary Ionization Processes 1. M o + e M + + 2e 2. M o + Ar M + + Ar 0 + e II. Secondary Ionization Processes 1. M o + Ar + M + + Ar 0 MX + Ar + M + + X + Ar 0 2. Ar + M o ArM + + e 3. M + h M + + e 4. M o + e M + e M + + 2e

PAGE 37

37 10910 107105103101 Current (A) VbVd VnVoltage (V) Townsend Discharge Transition Range Normal Glow Discharge Abnormal Glow Discharge Arc Discharge 10910 10710510310110910 107105103101 Current (A) Current (A) VbVd VnVoltage (V) Townsend Discharge Transition Range Normal Glow Discharge Abnormal Glow Discharge Arc Discharge Figure 1 1 Diagram of v oltage v s. current regimes for various discharges. [Reprinted with permission from E. P. Hastings, Reactive gases in glow discharge ion sources: sputtering and ionization considerations. Ph.D. dissertation, Universit y of Florida, Gainesville, Florida, 2004.]

PAGE 38

38 Cathode Layer Negative Glow Positive Column Anode Glow ( ) (+) Aston Dark Space Cathode Dark Space Faraday Dark Space Anode Dark Space Cathode Layer Negative Glow Positive Column Anode Glow ( ) (+) Aston Dark Space Cathode Dark Space Faraday Dark Space Anode Dark Space Figure 1 2 Various regions of a glow discharge. The three main regions are the cathode dark space (CDS), negative glow (NG) and Faraday dark space. [Reprinted with permission from E. P Hastings, Reactive gases in glow discharge ion sources: sputtering and ionization considerations. Ph.D. dissertation, University of Florida, Gainesville, Florida 2004. ]

PAGE 39

39 Anode dark space Negative glow Cathode dark space Vo VbCathode Anode (Ground)(-)Distance to cathodeVoltage Figure 1 3 T he three main regions in a simple glow discharge vs. voltage distribution as a function of distance to cathode [Adapted from E. S. Oxley, The microsecond pulsed glow discharge: developments in time -of -flight mass spectrometry and atomic emission spectrometry. Ph.D. dissertation, University of Florida, Gainesville, Florida 2002]

PAGE 40

40 Cuo Ar+ eCuo Cuo Cuo Cuo Cu+ Ar+ Ar+ Cathode Dark SpaceCathode Sample (Cu) (-) Aro Ar* ehv Cuo Ar* Aro Cu+ eCu+ eCuo eeNegative Glow (+) eFigure 1 4 Major events in the glow discharge : Cathode sputtering and ionization excitation in the negative glow [Reprinted with permission from E. S. Oxley, The microsecond pulsed glow discharge: developments in time of -flight mass spectrome try and atomic emission spectrometry. Ph.D. dissertation, University of Florida, Gainesville, Florida, 2002]

PAGE 41

41 +Incident Ion Sputtered Atom Surface Reflected Ion or Neutral Secondary Electron + +Incident Ion Sputtered Atom Surface Reflected Ion or Neutral Secondary Electron Cathode Dark Space Negative Glow Figure 1 5 Cathode sputtering due to fast ion/fast atom impact. [Adapted from E. P. Hastings, Reactive gases in glow discharge ion sources: spu ttering and ionization considerations. Ph.D. dissertation, University of Florida, Gainesville, Florida, 2004.] Figure 1 6 Sputter yield for various elements in the periodic table The elements are bombarded with 400 eV argon ions. (F rom N. Laegreid et al .18)

PAGE 42

42 Arm* Au0 Au+ Ar0 eArm* Au0 Au+ Ar0 eeeAu0 Au+ eeeAu0 Au+ eeeeElectron impact ionization (EI) Ar+ Penning ionization Charge exchange Figure 1 7 The most common types of interaction in the plasma. [ Adapted from E. S. Oxley, The microsecond pulsed glow discharge: developments in time -of -flight mass spectrometry and atomic emission spectrometry. Ph.D. dissertation, University of Florida, Gainesville, Florida, 2002]

PAGE 43

43 Figure 1 8 Direct and pulsed glow discharge modes are compared. [Reprinted with permission from E. S. Oxley, The microseco nd pulsed glow discharge: developments in time of flight mass spectrometry and atomic emission spectrometry. Ph.D. dissertation, University of Florida, Gainesville, Florida, 2002]

PAGE 44

44 Anode Anode Cathode Cathode Collimating Slit Collimating Slit Ions Ions GD Plasma GD Plasma Early Ion Source ~ 1920 s Figure 1 9 Hollow cathode glow discharge source for mass spectrometry. [ Courtesy of Dr. W. W. Harrison, University of Florida, Gainesville, Florida )] HV HV Connector Connector Handle Handle Ball Valve Ball Valve Gas Inlet Gas Inlet Disc Sample Disc SampleDirect Insertion Probe (DIP) Figure 1 10. Diode geometry with a direct insertion probe (DIP). [Courtesy of Dr. W. W. Harrison, University of Florida, Gainesville, Florida)]

PAGE 45

45 Cathode O -ring Figure 1 1 1 Grimm glow d ischarge geometry for mass spectrometry. [Reprinted with permission from E. S. Oxley, The microsecond pulsed glow discharge: developments in time -of -flight mass spectrometry and atomic emission spectrometry. Ph.D. dissertation, University of Florida, Gaine sville, Florida, 2002]

PAGE 46

46 CHAPTER 2 EXPERIMENTAL DETAILS AND BACKGROUND Introduction The experimental pa rt of this research is consist of three major components: Aerosols: g eneration and sizing Pulsed glow discharge ion source Time -of -flight mass spectro meter In all future chapters, there will be a publication -style experimental section, which briefly discusses the set up and material used for this work. In this chapter, more details and back ground related to these experiments is provided. Aerosols Two methods for aerosol generation have been utilized in our research. The first method is simply based on nebulization of salt solution by an ultrasonic nebulizer. In the work discussed in Chapters 3 and 4 the salt particles generated by this method were use d as a source of aerosol. The other method of particle generation is laser ablation of solid samples. Laser ablation of solid samples is a well known technique used in tandem with ICP mass spectrometry. In this chapter particle generation by the ultrasonic nebulizer will be discussed The information about laser ablation is discussed in Chapter 5. Aerosol Generation An ultrasonic nebulizer -dehydrator57, 58 was used to generate a flow of dry particles from their corresponding salt solutions Figure 2 1 shows the schematics of the nebulizer dehydrator A p eri sta l tic pump directs the flow of aqueous salt solution onto the surface of a piezoelectric i n the nebulizer unit, vibrating at a frequency of about 1 MHz.50 About 10% of the solution is converted to a wet aerosol, and a mist is formed. The resulting wet droplets are carried by the argon flow into glass tubing heated to 150C. Part of the moisture from these wet particles is

PAGE 47

47 evaporated, and a stream of relatively dry particles and water vapor is formed. This stream then goes through a cooling chamber (4C) to condense the water vapor. The aerosol then goes through a porous tube (made of a special polymer) in the dehydrator unit. A stream of argon heated to 150C moving in the opposite direction of the aerosol outside the porous tube will completely remove any trace of water vapor from the aerosol. The out put of the dehydrator unit is a stream of dry salt particles in argon. In order to verify the necessity of the dehydrator, the nebulizer was directly connected to the glow discharge; it was observed that no plasma was formed due to presence of water vapor. In contrast, when both nebuli zer and dehydrator units are used, the discharge plasma can be ignited and remains stable. Nebulizer -Dehydrator Efficiency The easiest way to estimate the efficiency of our nebulizer -dehydrator unit is to collect and measure the weight of solutions exitin g the waste tubes of the nebulizer. There are three solution collection tubes in the nebulizer unit. The first one (W1) is located at the nebulization chamber which collects that part of the solution that is not nebulized at all. The second tube (W2) coll ects the solution formed at the beginning of the heating chamber. This solution is formed by water vapor that is condensed at the end of the heating chamber. The last tube (W3) collects the condensed vapor from the cooling chamber of the nebulizer. Three s amples of 4 m M CsI in water with a total solution mass of ~25 30 g were nebulized. The resulting solutions were carefully collected and weighed at the end of all three waste exit ports. Table 2 1 shows the mass and weight percentage of these experiments. I t is important to note that the summation of water weight collected at the three waste tubes does not add up to 100%. This is due to loss of water left in the nebulizer and tubing itself. The conductivity of collected solutions was measured to be 94.3 S.c m1, 2.76 S.cm1, and 1.66 S.cm1 for W1, W2, and W3, respectively (S stands for

PAGE 48

48 S iemens ). This shows that W1 is mainly un -nebulized solution, while W2 and W3 mostly contain the condensed water resulting from drying the wet particles. T his experiment shows that ~10% (W2 + W3) of solution is nebulized. This is the maximum possible yield, as some of the water exiting W2 and W3 can simply be the result of vaporization of water from the nebulization chamber and not from the aerosols. One should also consider the potential for loss of aerosol during the transfer process through different dehydrator and external tubings to the glow discharge source. The overall efficiency of the nebulizer dehydrator could even be less than the 10% measured by the above procedure In order to obtain a more accurate aerosol transfer efficiency to the discharge, another experiment was designed: the outlet of the dehydrator was transferred by Teflon tubing (same length used for aerosol transfer into glow discharge) under 250 .0 mL of water. After nebulizing a 4 mM solution of CsI for ~5 minutes, the conductivity of water was measured. After comparing this conductivity with the calibration graph ( Appendix Figure A 8 ) that was prepared separately, one can estimate the concentration of C sI in the aerosol collection solution. From this concentration, the mass of salt transferred as aerosol and efficiency of the nebulizer can be calculated. Table 2 2 shows the results of three such experiments, and the efficiency of the nebulizer -dehydrator unit was found to be ~ 4%. The expected efficiency for such a nebulizer is about 10%.51 Efficiencies as high as 25% for certain experimental conditions, with optimized nebulizing power and liquid flow rate ha ve been repo rted.52 Our measured efficiency is much lower than the reported values. This might be due to the fact that we a re nebulizing solutions of much higher concentration compared to the normal range of concentrations used for ICP experiments in these nebulizers.

PAGE 49

49 Particle Size Distribution Measurements In order to study the particle interaction with the glow discharge pl asma, one must consider the size of particles present in the aerosol. In order to measure the size distribution of particles in our aerosol, at first an attempt was made using a differential mobility analyzer (DMA). However, due to high concentration of sa lt in our nebulizer intake and also insufficient pumping, the size measurement experiment did not result in any meaningful results. Another problem was split ting the particle flow in order to protect the instrument from contamination by a large amount of p articles. Finally, a n aerodynamic particle sizer (Chapter 3) was used to measure the particle size distribution. This instrument measures the particle sizes based on their time of flight. This method is an example of light interaction methods of particle s ize measurements. In this instrument, particles pass through a path crossed by two laser beams. The scattering of laser light is detected and used to measure the velocity of particles, which can be related to particle diameter. Here we d iscuss the general categories of particle sizing techniques used in the literatu re .6165 Microscopy method In this method, the particles are collected on the proper microscope stage. The resulting collection of particles is then placed under the microscope (or microscopic probe/tip) for observation. The resulting image will show the shape and diameter of each individual particle collected on the collection surface. This method is accurate, but is time consuming and requires the collection of particles from the aerosol state to the surface. Optical microscopy, transmission electron microscopy (TEM), scanning electron microscopy (SEM), and atomic force microscopy (AFM) are examples within this category.53

PAGE 50

50 Particle impaction method In this classic method, a series of impaction plates are placed at the path of particles. The aerosols are accelerated through jet n o zzles towards the plates (Figure 2 2).62, 63 Each i mpactor will trap a group of pa rticles within a specific size range. At the end, weighing or chemical methods c an be used to estimate the over all volume / mass of particles in that size range.64, 65, 68 Light i nteraction m ethod In these methods, the interaction of light (usually in the form of a laser beam ) with particles is used to measure particle size in the gas phase.53, 54 Laser diffraction, single particle ligh t scattering,54 multi angle light scattering, time of -flight and laser doppler velocimetry are examples of light interaction methods of particle size measurement .55 For example, the light scattering technique uses the Mie Theory.53 According to Mie Theory,53 the light scattered by an individual particle is a function of the scattering pattern produced by spherical particles of a specific size. The scattered light intensity is a function of the angle, wavelength, particle size and optical properties of the system, such as refraction index of the particles. Therefore, collecting and analyzing the pattern of scat tered light at different angles along with complicated calculations can result in particle size d istribution of the aerosol sample. Electrical p roperty method In this method, usually particles are charged and then the charged particles will have different physical properties due to their size.56 The d ifferential mobility analyzer61, 64, 65 (DMA) is an instrument that works based on such properties. In this method particles are charged before they go through a chamber under the electric field of a central rod (Figure 2 -3). The intensity of attraction that particles have to this rod depends on their mobility, which also depend s on particle size and shape. This is assuming the fact that all particles have the same charge (more likely single charges). By scanning the electric field only particles of a size within a certain range can

PAGE 51

51 pass through the chamber. A condensation parti cles counter56 (CPC) unit is used in series after the output of DM A (Figure 2 4). The particles are first exposed to vapor of a liquid such as butyl alcohol to enlarge the size of particles by condensation. Then a light scattering system is used for counting the particles in CPC.56 Sedimentation method This category includes measurement techniques such as p hoto, centrifugal and X ray sedimentation.53, 57 Fluorescence Activated Cell Sorting (FACS) and Field F low Fractionation (FFF) are a lso within this category.55 As the meaning of the word sediment expresses, this method is based on the rate in which the particles precipitate (usually in a liquid). For example, in X -ray sedimentation, a horizontally collimated beam of X ray s passes through a liquid. Then the precipitation of particles will affect the beam intensity and as a result the number of particles is measured. The larger particles will cross the b eam first and the smaller particles will precipitate later. Limitations of Particle Sizing Methods It is important to understand that a specific particle sizing device can not measure the particle size directly, but rather measures a physical property that is related to particle size / diameter.58 For example the result of a DMA CPC particle sizing experiment is indeed the mobility diameter of particl es (dm), or a particle impactor works based on aerodynamic properties and the result is aerodynamic diameter (da). Another important point is that real particles are not always perfectly shaped spheres ; in some cases, they are irregularly shaped or agglome rated.59 Agglomeration can result in false data about particle size distributions. This means that the measu red diameter does not represent the diameter of a particle that has an equal volume to that of the real particle (dve measured diameter).59 For example, irregularities from spherical shape

PAGE 52

52 will result in reduction of aerodyn amic diameter. As a result a 200 nm agglomerate can have the same trajectory that a 40 nm spherical particle has. In contrast, in a DMA experiment an agglomerated or irregular particle will have a lower mobility. In other words the irregular ly shaped pa rticle will behave like a larger particle in a DMA system and will exit the analyzer in conditions similar to that of larger particle s .59 Consequently the impaction and DMA will result in falsely low and falsely high results respectively (Figure 2 5). Presentation of Particle Size Measurement Results The result of most particle size measurements is a graph in which the population of particles (absolute or relative) is plotted vs. particle size. However the population is usually found for a certain range of diameters. For example, N% (relative abundance) or N (absolute abundance) will have a constant value in a size range and then another value in the next diameter range.60, 61 Unfortunately, since the diameter bins are not equally distributed, this met hod can be misleading.60 For example, let us assume 20% of particles are within the range of 300500 nm, while only 10% are within the range of 10 30 nm. If we assume an equa l distribution within each range, one might think that a particle in the first range is more abundant than a particle in the second one. This would be correct if both ranges had equal width (20 nm vs. 200 nm). A closer look will show that the 20% of the pa rticles must be distributed among many sizes (from 300 to 500 nm) while the second range only covers 20 nm. In order to avoid such biases, we can normalize the abundance by the related range. Dividing the abundance by the size range width (20% / 200 nm for example) will result in more realistic values of abundance. This way we can see that a 20 nm particle is likely to be 5 times more abundant than a 400 nm particle. (0.5 vs. 0.1 as shown in equations 2 1 and 2 2) nm nm / 5 0 20 % 10 (2 1)

PAGE 53

53 nm nm / 1 0 200 % 20 (2 2) This way of presentation of particle size is normally referred to as dN/dD where N is the number of particles per unit volume and D is the diameter range.60, 61 If particle size is presented in micrometers and number density of particles is in number of particle s per cubic centimeter then the Y axis is labeled with dN/dD and units of cm3. m1. Although particle abundance in each size range provides some information, it does not necessarily provide the useful physical information such as their contribution in ion signal after they enter the plasma. For example, here if the number of 20 nm particles is 5 times of that of 400 nm part icles, one might misguidedly think that the 20 nm particles will contribute to the signal of our glow discharge 5 times more than that of 400 nm particles. This is correct if we ignore the obvious fact that the diameter of 400 nm particles is 20 times of t he diameter of 20 nm particles. Considering this and the cubic relation of volume and mass with diameter, the signal provided by a single 400 nm particle is 8000 times larger than a single 20 nm particle. Considering the factor of 5 for abundance of 20 nm particles, the 400 nm particles will have an advantage by factor of 1600. The importance of the role of particle volume leads to another method of presentation of particle size. The dV/dD represents the total volume of particles per unit volume of air (or gas) that a re within a specific size range and are normalize d by that size range.60, 61 The unit of this value is m3.cm3. m1. If the absolute abundances are not available and relative volume intensity in each range is available, then it will be shown as (dV%/dD) and the unit will simply be m1.

PAGE 54

54 If the surface of the particle is the physical property of interest ( i.e. surface chemistry and catalysis applications ) then one can present the dS/dD which deals with the surface instead of the volume.60, 61 Finally, when analyzing different aerosols, it is not always easy to reach a quantitative conclusion by just looking at the size distribution plots. They can look very similar while indeed they represent different aerosols. The volume mean diameter (DV) is a qu antity that is defined as the diameter of a particle, whose volume is equal to the average volume of all particles present in the aerosol.60, 61 O n e can define the surface area mean diameter in a similar way Glow D ischarge S ource The glow discharge source used in this experiment was modified from the original design used by Harrison group for depth profiling. The details of this geometrical modifica ti on are well explained in C hapter 3. Mass Spectrometer The mass spectrometer is a time -of -flight instrument (Renaissance, LECO Corp., St. Joseph, MI) that was modified for our application. Table 2 3 shows the typical instrument parameters used during expe rimentation. These parameters are very similar to those previously used by Harrison Group9 with a Grimm glow discharge source. The primary difference was the use of ion deflection at 1.346 s to deflect the excess of argon ions, which was required since in the new source geometry, the argon flow was much higher (800 ml/min compared to 150 ml/min). Because the argon signal was saturating the detector, deflection of argon ions was used to diminish i t. It was also observed that increasing repeller delay time from 1.2 s to 2.0 s increased the particle ionization signal. An important point of note is that values of less than 16 s for repeller delay time are experimentally meaningless. This was discovered by carefully

PAGE 55

55 investigating the repeller ringing on the oscillosc ope; when delay times shorter than 16 s are selected on the software, the real delay time remains at 16 s and does not go below this value. More details about repeller -modulator delay times follow in this chapter. Modifications Original ICP TOF-MS: The t ime -of -flight mass spec trometer was originally equipped with an ICP ion source. The red li ne in Figure 2 6 shows a constant supply of ions provided by the ICP plasma Since the TOF detector requires a gated ion source the modulator gate s this continuous i on source. The m odulator (yellow) is a cylindrical part that normally has a positive potential (100V) to prevent ions from entering the flight region. However for a short time it is pulsed to a n egative potential to let ions pass in a gated way. Its dela y time is 45.11 s and its negative pulse duration is 4.89 s It allow s for a pa cket of ions of ~ 1.5cm in length to pass. Modulation delay is the timing between opening the gate and the end of the last modulator pulse. Addition of the modulator delay to i ts width results in 50 s timing between the start of two consecutive pulses. As a result the factory -made ICP MS works at 20, 000 Hz ( As reported in t he instrument manual). The r epeller is adjacent to the modulator and it provides a h i gh positive voltage (1000V) to shoot the ions into the flight tube. The repeller was set to have a pulse width of 1 0 s (in the ICP mode) and the instruments manual does not clarify the delay time between modulator and repeller. It is predicted that there will be a very sh ort delay time right after the modulator and that they will work in tandem with each other to gate ions One thing that is clear, however, is that the role of the repeller here is not gating but to shoot a pa cket of ions ( partially or completely ) into th e flight tube, right after the m odulator

PAGE 56

56 Modifi ed GD -MS source : The LECO instrument has been modified for use with a pulsed Grimm -type glow discharge source for ionization. The supply of ions is no longer continuous but pulsed. According to Eric Oxleys dissertation the modification includes some structural s pecifications for replacement of the old ICP with a homemade GD source as well as some software modification s Without these changes, the software may not have accepted the microsecond pulses from the GD source. A repeller delay time adjustment also needed to be made to span a range of 10 s to 300 s allowing ion extraction time to be vari able so that temporal resolution measurements could be achieved.4, 33 A complementary software (Alternate timing) we added to allow the switch between ICP and GD modes. The ICP mode of detection (20,000 Hz) for the instrument is still available when the corresponding box is not selected on the software. In all experiments discussed, the only variable delay time is referred to as repeller delay time. Figure 2 7 s how s the timing diagram for the HV ionization p ulse in the source with respect to the modulator and repeller pulses The schematics of ion extraction are also shown in Figure 2 -8. The exact meaning of the repeller delay used previously was not discussed. In other words, it was not clear that this is the delay from the start of the HV pulse or the end of it, and is it the delay to the repeller or modulator. It was neither understandable what the exact timing between modulator and repeller is. Repeller Modulator Timing Since t he timing between the external trigger (which initiates the glow discharge pulse) and the repeller -modulator pulse s was not clear this was carefully studied by looking at the ringing of the repeller and modulator on an oscilloscope Figure 2 9 shows the timing diagram for different pulses of glow discharge and mass spectrometer. The time of -flight mass

PAGE 57

57 spectrometer sends an external trigger (green) to the pulsed high voltage generator (t = 0), and a high voltage (HV) pulse (black) is then app lied to the glow discharge source with a pulse width determined by the HV pulse generator. The repeller of the time -of -flight will always have its high voltage pulse (red) start according to the repeller delay time value from the software. T he modulator i s synchronized with the repeller in a unique way and works as follow s : t he trailing edge of the modulator is always 1 s after the leading edge of the repeller. This means any change in the width of the modulator pulse will change the location of its leading edge with respect to t = 0, but not its trailing edge. The arrows on the pulse widths in Figure 2 9 indicate which edge of the pulse will move with respect to t = 0 as a result of the change in pulse width. For example, Figure 2. 10, A shows a 10 s high v oltage pulse applied to the source. The repeller delay time is 60 s, and the modulator and repeller pulse widths are 5 s and 1.2 s, respectively. In 2 10 B, the modulator and repeller pulse widths are increased to 20 s and 10 s, respectively. In C, th e repeller delay time is reduced from 60 s to 40 s. It is important to note that the modulator is physically located a few centimeters behind the repeller, otherwise such timing would not be logical. E shows the source producing a 30 s package of ions w hich will move through the skimmer and ion lenses toward the modulator -repeller. It will take about 20 s for this package to arrive at the modulator. Since the delay time is set at 40 s, the modulator gate will allow a 5 s package of ions to pass (from t = 36 s to t = 41 s). However, considering a 2 s flight time from the modulator to repeller, this package will arrive at the repeller from t = 38 s to t = 43 s. The repeller will then choose a smaller portion of this package (from 40 s to 41.2 s) a nd accelerate it towards the flight tube. If the repeller width is increased to 2 s a larger package of ions and a larger signal is expected, which is in agreement with experimental results.

PAGE 58

58 Increasing the repeller pulse width beyond 2 s does not increas e the signal any more, as expected. Another int eresting observation is that repeller delay times of less than 16 s on the instrument software are all shown to equal 16 s. This must be due to some electronic oversight in design when modifications were done for pulsed timing. In order to reach shorter delay times, one can place an external delay generator before the HV pulse (Figure 2 10 F). Appendix Figure A 4 shows the repeller -modul ator ringing, along with argon signals, collected on the oscilloscope. TOF -MS Detector and Collecting Mass Spectrum The detector of the mass spectrometer is a discrete dynode active film multiplier that produces electrons when ions strike it. To acquire and generate a full mass spectrum, the mass spectrometer divide s the whole mass range into 10 sections. Then it goes through and integrate s for a certain period at low masses (from 0 30 amu), moving to the next window (30 60 amu) and so on. Considering the older technology of the 1990s, taking a full spectrum can take 30 40 sec onds (depending on integration time). A sample peak from a typical mass spectrum collected by the software is depicted in Figure 2 11. An SMA to BNC converter was used to connect an oscilloscope to the detectors signal output on the instrument control mod ule (ICM) of the mass spectrometer. The modulator repeller ringing was used to trigger the oscilloscope In order to obtain a stable signal, it is very important that the trigger main level be at 77 mV. Connecting the detector signal directly to an externa l oscilloscope results in a mass spectrum with much less noise (Figure 2 1 2 ). T he oscilloscope also makes possible the collection of the whole spectrum, and saving it as a spreadsheet (file.csv) on a computer, in just a few seconds.

PAGE 59

59 All mass spectrometry signals in this dissertation refer to positive ion signals. If a specific isotope is not specified, the signal is for the most abundant isotope of that element. For example, copper, argon, cesium and iron signals refer to 63Cu+, 40Ar+, 133Cs+ and 56Fe+ ion signals, respectively.

PAGE 60

60 Table 2 1. Results of efficiency measurements for ultrasonic nebulizer dehydrator. An average efficiency of 10.9 2.0 resulted (95% confidence interval). Trial 1 Trial 2 Trial 3 Mass / g Mass % Mass / g Mass % Mass / g Mass % Total feed 26.5717 100 29.2761 100 28.2120 100 W1 22.6236 85.14171 25.5053 87.12 23.7122 84.05 W2 2.0286 7.634438 2.1430 7.32 2.4347 8.63 W3 0.7996 3.009217 0.8549 2.92 0.8859 3.14 Total 95.78537 97.36 95.82 W2 + W3 (ffii ) 10.64365 10.24 11.77 Table 2 2. Res ults of nebulizer -dehydrator efficiency using conductivity measurements. The average efficiency was found to be 4.5 0.6 (95% confidence interval). A 4 mM CsI solution was used for both nebulization and conductivity calibration gra ph. The conductivity values were compared to the calibration graph (Appendix A, Figure 8) and the volume s of CsI solution found from the graph w ere multiplied by 2.5 (250 mL volume of collector vs. 100 mL for calibration graph). The result is the volume 4 mM of CsI that would lead to the same conductivity as the particles transferred to the collecting solution. Comparing this equivalent volume with the original volume of nebulized solution will result in the efficiency of the nebulizer -dehydrator system. T rial 1 Trial 2 Trial 3 Volume of CsI nebulized / mL 5.97 5.74 7.08 Conductivity of 250 mL collector / S.cm 1 1.02 0.93 1.12 Equivalent volume of CsI transferred / L 282.5 246.4 322.7 Efficiency % 4.73 4.29 4.56

PAGE 61

61 Table 2 3. Typical operating conditi ons for time -of -flight mass spectrometer. Ion Lens 1 500 V Ion Lens 2 550 V Flight Tube 1490 V Detector 1600 V Reflectron Low 215 V Reflectron High 1536 V Noise Reduction 500 V Deflection Time 1 (start/width) 0.986 s / 0.05 s Deflection Time 2 (start/width) 1.346 s / 0.150 s Deflection Pulse 400 V X Steering 1473 V Y Steering 1630 V Einzel Lens 1 1370 V Einzel Lens 2 799 V Modulation Positive 120 V Modulation Negative 133 V Modulation Delay 40 s Modulation Pulse Width 5 s R epeller Bias 4 V Repeller Pulse 1000 V Repeller Pulse Width 1.2 2.0 s Repeller Delay Time 15 300 s Alternate Frequency 400 Hz Third Stage Pressure 0.4 1.2 torr Integration Time 500 ms 2 s

PAGE 62

62 Drying argon Condensed water Heat 150 oC Cold 4 oC pump Salt solution In water Heat 160 oC Argon vibrating surfaceTowards GD source Figure 2 1. The schematics of the ultrasonic nebulizer -dehydrator. Figure 2 2. The schematics of the particle i mpactor. [Adopted from Kuhn et. al.59]

PAGE 63

63 Figure 2 3. The differential mobility analyzer measures the particle size based on their mobility vs. their attraction within electric field. [ht tp://www.cac.yorku.ca/mozurke/Analyzer.htm] Figure 2 4. The condensation particle counter (CPC) counts the number of particles after they pass through the DMA unit. [Adopted from Clifford et. al.56]

PAGE 64

64 A B Cda< dve Ddm> dvedveAgglomerate Figure 2 5. Comparing the results of different particle sizing methods. The agglomerated particles (A) will result in false particle size measurements compared to the ir real equivalent volume diameter (B). The impaction wi ll result in a falsely low aerodynamic diameter (C), while a DMA measures a falsely high mobility diameter (D). [Adopted from Kuhn et. al.59] 45.11 s 4.89 s 1 s Figure 2 6. The diagram shows continuous ion source (red) vs. the modulator repeller gating pulses, for the original ICP -MS design.

PAGE 65

65 Ion Gating system Pulse interval time 2500 s (400Hz) Gate delay time 10300 s Pulse width 20 s Pulsed ion source Figure 2 7. The diagra m shows the pulsed GD ion source (red) vs. the modulator repeller gating pulses. High voltage Pulse ( -2kV) Modulator repeller t = 0 t = 0 -300 s time Ion lenses skimmer High voltage Pulse ( -2kV) Modulator repeller t = 0 t = 0 -300 s time Ion lenses skimmer Figure 2 8. The schematics of repeller and modulator with respect to the pulsed ion source.

PAGE 66

66 Modulator width t = 0 HV pulse width t1t2 Repeller width Trigger from TOF t2 = t1+ 1 s t1: repeller delay ( 16 300 s) Figure 2 9. The timing diagram for the time -of -flight trigger vs. the repeller and modulator.

PAGE 67

67 Increasing modulator and repeller width Reduce the repeller delay from 60 s to 40 s Reduce modulator and repeller width Generate 20 s delay before the HV pulse Increase the HV pulse width 0 20 40 60 80 Time / sA. B. C. D. E. F. Figure 2 10. The effect of changing different instrument variables on timing diagram

PAGE 68

68 Figure 2 11. A sample mass spectrum from the instrument software. The 68Zn peak is shown for comparison purposes with the same peak in figure 2 12. 67.5 67.6 67.7 67.8 67.9 68 68.1 68.2 68.3 68.4 68.5 0 0.5 1 1.5 2 2.5 m/zSignal (mV) Figure 2 12. A sample mass spectrum from the oscilloscope The 68Zn peak is shown for comparison purposes with the same peak in f igure 2 11.

PAGE 69

69 CHAPTER 3 A NOVEL PULSED GRIMM TYPE GLOW DISCHARGE SOURCE FOR IONIZATIO N OF SALT AEROSOLS Introduction Glow discharge is a well known plasma source in analytical chemistry for atomization and ionization. The resulting atoms and ions can be ide ntified using optical emission spectroscopy (GDOES) or mass spectrometry (GDMS). Although glow discharge i s mainly used to study solid samples, the increas ing interest in the study of aerosols has led analytical chemists to find functionality in glow discharge as a source for atomization and ionization of particles. The need for an online detector for liquid chromatography was the primary motivation for development of particle ionization techniques. As for most other low pressure ion sources, the challenge was to send a stream of dry particles into the ionization source. Since a small amount of solvent can severely suppress the plasma, the complete removal of solvent from the aerosol is very important. Marcus et al. ha ve extensively used a particle beam samp le introduction device to generate a dry flow of particles into the glow discharge source .62 67 The particle beam strikes a hot target (or the heated cathode itself) and evaporates. The resulting vapor will then enter the glow discharge plasma for further atomization and ionization. Both optical emission spectroscopy64, 65 and mass spectrometry62 64, 66, 67 were used for quantitative and qualitative analysis of the particles. Grimm -type glow discharges were originally used for depth profiling of solid samples. Pulsed glow discharge time -of -flight mass spectrometry was later introduced.31, 6972 Pulsing has the advantages of enhanced sample sputtering, excitation and ionizati on and reduced thermal effects .32, 33, 68 Due to the timed nature of the pulse, spatial resolution of the sputtered ato ms from the plasma ions can also be achieved if the pulse is synchronized with the repeller of the time of flight mass spectrometer. It has been shown that the pulsed GDMS can be used to obtain

PAGE 70

70 elemental fractional and molecular information from gaseous analytes.69 This is due to different ionization energies at the initiation, plateau and after termination of the pulse. In recent years the interest in the study of aerosols has prompted scientists to investigate the possibility of the development of aerosol mass spectrometers .70 72 Most of the aforementioned techniques utilize a laser beam for ionization of particles.70 Non -laser methods usually include extreme heating of the particles. For example, flash vaporization of part icles in a closed -ended tube that was heated up to 600 C followed by electron ionization (EI) generated ions from particles.71 To our knowledge, no one has used a pulsed Grimm type GD as a source of ion s for mass spectrometry of particle s The previous particle beam GDMS uses a DC hollow cathode GD source with heating serving as the main source of particle vaporization.63, 64, 67, 7678 It has actually been indicated that vaporization from the cathode walls must be the major means of particle vaporization. This is due to the fact that plasma ionization energies are less than 25 eV and the kinetic temperature of the plasma is not capable of dissociating particles in the gas phase. However, it is wel l known18, 26, 76, 77 that the high energy atoms and ions in th e cathode dark space of a Grimm type GD are capable of sputtering the metal atoms from the cathode surface. This brings about the potential for vaporization of particles in this high energy region. This chapter investigates the interaction of particles with high energy argon ions and atoms in the cathode dark space (CDS) of the Grimm glow discharge (without heating the cathode). Most plasma sources such as ICP and glow discharge only provide atomi c ions for mass spectrometry. Molecular and fragment ed ions have been reported in a few cases. For example, microwave plasma GD of volatile organic compounds73 and variable frequency, switched, direct current GD of halogenated hydrocarbons74 resulted in atomic and molecular ions. Another

PAGE 71

71 purpose of this work is to investigate the possibility of observing atomic versus molecular information at different pulse zones. This application of pulsed-GD has been shown by s ending molecular vapors of simple organic compounds into a pulsed GDMS source.7578 Since the focus of this research is on fundamental study of interaction of aerosols with Grimm pulsed GD TOF -MS, a commercial u ltrasonic nebulizer dehydrator unit (efficiency of less than 5%) w as used to generate simple aerosols from inorganic salt solutions. For quantitative purposes, this is a disadvantage when compared with a particle beam device w h ich has an efficiency of up to 97%.67 Experimental The overall schematic of the instrum entation is shown in Figure 3 1. A flow of sal t particles in argon is generated and transferred through a needle valve into the low pres sure GD source for ionization. A high voltage s -pulsed power supply (M3k20, Instrument Research Company, MD) that is triggered by the instrument control module (ICM) of the mass spectrometer provides a negative potential to the cathode of t he GD source. The ICM also triggers the repeller of the mass spectrometer with a variable delay time (0 300 s), controlled by the software. The temporal pulse voltage and current can be monitored on a 500 MHz digital oscilloscope (TDS 724D, Tektronix Inc., Beaverton, OR) A 5k 10W resistor was placed in the circuit and the average current is digitally presented on the pulsed power generator. Table 3 1 provides a list of major operating conditions. Additional parameters not included can be found in previously reported literature .32 Aerosol G eneration a nd P article S ize D istribution An ultrasonic nebulizer -dehydrator (U 6000, CETAC Inc., Omaha, NE) was used to produce a stream of dry particles from the corresponding salt solution. The solution i s transferred

PAGE 72

72 on to the sur face of a piezo electric vibrating stage by a pe ristaltic pump (M312, Gilson, France) whe re argon flow generates a mist. The resulting aerosol goes through a heating-cooling stage and then enters the dehydrator unit, where an opposite directional stream o f hot argon completely dries the particle flow. Industrial argon (Grade 2, Air Gas Inc Kennesaw, GA) was used as an aero sol carrier into the GD source. Salt solutions were prepared by dissolving the corresponding solid salts (Fisher Scientific, Pittsburgh PA) in deionized, distilled water. NaI, CsI and acetate salts of some transition metals were used. In an independent experiment, a particle size analyzer (Aerosizer LD, API Inc., TSI Inc, Shoreview, MN) was used for particle size distribution measurement s. Mass Spectrometer The mass spectrometer is an axial commercial ICP TOF instrument (Renaissance, LECO Corp., St. Joseph, MI) that was modified by removing the ICP source to ac cept an in -house designed Grimm type GD source.32 In addition, a variable repeller delay was added to the instrument and its software (LECO Renaissance version 1.16, LECO Corp.). The repeller delay indicates the time difference fro m the start of the high voltage pulse at the GD source t o the start of the high voltage pulse on the repeller. There is a modulator that always functions in synchrony with the repeller in a way that the trailing edge of the modulator pulse always ends 1 s after the leading edge of the repeller pulse. The pulse width s used for the repeller and the modulator were 2 s and 5 s respectively. The transient mode of the software was used to record the m/z values for selecte d peaks as a function of time. Integ ration time was set t o 2.0 s. For collecting full mass spectra, the detector was connected to the oscilloscope and the resulting mass spectrum was saved on the computer through a GPIB -PCI interface.

PAGE 73

73 Grimm -Type GD Ion Source The GD source geometry was the s ame as the Grimm source us ed by Harrison et al previously.32 In order to send particles into the discharge, the particles were introduced through the gas -directing sleeve entrance from the space between the a node and the sampler. However, sending particles through the gas directing sleeve entrance resulted in a l oss of particles and clogging. As a result, no atomic signal related to the particles was observed. Cathode m odification In order to overcome this p roblem, a small orifice (0.5 mm) was placed at the center of the cathode so that the particles c ould enter the discharge from the back of the cathode. In orde r to center the cathode orifice with respect to the anode, the c athode base was also threaded. Fig ure 3 2 shows a schematic of this cathode design. Since the back sleeve was no longer required, the sampler was also replaced with a flat sampling device. New a node d esign After cathode modification, the particle ionization signal was observed using a typ ic al cylindrical Grimm type anode (anode A ). However, implement ing a different anode (anode B ) that is a concealed cylinder with 4 off axis orifices instead of a hollow cylinder resulted in improved signal intensity and discharge stability (Figure 3 3 ). On e of the advantages of this new anode design was preventing large particles fro m mov ing directly towards the skimmer by blocking them This reduce d the risk of particles entering into the second stage of TOF wh ich can contaminate ion lenses. Figure 3 4 sh ows the detailed GD source geometry with anode B. The cathode is machined from br ass and the anode from copper. The effe ctive cathode diameter is 5 mm. The cathode anode distance is 0.50 mm. The four orifices of the anode have diameter s of 1.5 mm each and are 2 mm away from the c enter. The particles move along the argon flow and enter the area between the cathode and the anode, where a high voltage pulse causes

PAGE 74

74 ionization The ions pass through the 4 holes of the anode into the first stage of TOF where som e of them reach the skimmer. Since the skimmer and sampler are both grounded, the majority of the flow (including the ions) is pumped out at this stage. An adjustable diaphragm valve (Model: Sp35K, BOC Edwards, Wilmington, MA) controls the amount of pumpin g and the pressure of the first stage. The ions that pass through the skimmer orifice, after passing through the second stage ion lenses and accelerat ion by the repeller, will be detected in the TOF analyzer. Pressure and Flow M easurements The pressure of all three stages of the instrument is displayed on the mass spectrometer software. For the measurement of the source pressure (the anode -cathode gap) a dual sensor vacuum gauge (Model: 2002, Teledyne Hastings Instruments, Hampton, VA) was connected to the source through th e back sleeve gas entrance. Since particles can harm the digital flow meters, the flow of argon gas without the presence of particles was measured using 2 parallel digital mass flow meter s /controllers (Model: M100B, MKS Instruments, Andover, MA). A soap bubble meter was used to calibrate the digital flow meters The argon flow was then calibrated with respect to the first stage and the source pressure (P1 and Ps). These pressures were used to calculate the flow during aerosol introduction. Results and Discussion At first, a 10.0 mM solution of NaCl was used for generating salt particles and sodium signal was monitored. Since the back sleeve of the source had clogging problems and no sodium signal was detected, the cathode was modified as m entioned in the experimental section. The n ext course of action was choosing the best salt one that produced the highest signal and could be used for fundamental studies. The modified cathode along with anode A resulted in ionization of a variety of sal t particles generated from different salt solutions (10.0 mM). The positive ion signal for both anion and cation of such salts was observed at the correspondent m/z

PAGE 75

75 value: NaI (Na+ = 23, I+ = 127), NaCl (Na+ = 23, Cl+ = 35, 37), NaF (Na+ = 23) RbI (Rb+ = 8 5, 87, I+ = 127), CsI (Cs+ = 133, I+ = 127), Mg(NO3)2 (Mg+ = 24), MnAc2 (Mn+ = 55, CO2 + = 44, H CO2 + =45), CoAc2 (Co+ = 59, CO2 + = 44), NiAc2 (CO2 + = 127). Ac represents the acetate ion. The signal intensity for I+ was higher than other non-metals due to t he lower ionization energy of iodine. The i onization potentials (in e V ) for I, Cl, F and Ar are 10.45, 12.97, 17.42 and 15.76, respectively. Since Ar+ is the major source of ionization in the plasma, Fl uorine can not be ionized. The signal for chlorine was also much lower than that of iodine. It is important to consider that the major source of ionization in the negative glow is the P enning ionization which is due to meta -stable argon atoms (Ar* =11.55 eV and 11.72 e V ). The sodium in iodide salt showed a higher signal than that of chloride. This can be explained by lower energy of vaporization for NaI, which has a melting point of 670 C compared to NaCl = 801 C. Among the metals, the estimated relative signal intensities for Cs, Rb, Na, Co, Mn and Mg were 15, 13, 6, 5, 5 and 1, respectively. It appears that singly charged cations produce a larger signal (for example Na vs. Mg) and th a t as the mass of the atom increases the signal also increases. This will be discussed in more detail later in this chapter The signal for Ni had a very low intensity; this can not be all explained by ionization energy (IE) or atomic mass as the IE of Ni is lower than Co and they almost have the same mass. However, both Co and Mn have only one naturally occurring isotope, whi le Ni has multiple isotopes (58Ni = 63% the most abundant). This could cause the signals for Ni to be lower. CsI and NaI were found to be good choices for further studies, as both salts showed a high degree of ionization and both offer cations with only on e naturally occurring isotope. Anode A vs. B Although the reconfiguration of the cathode allowed particle introduction into the discharge and resulted in its ionization, it was faced with a few problems. When the argon flow

PAGE 76

76 in this configuration exceeded 300 ml/min, no ionization signal (even from argon) was detected. It seems that at higher flow rates the atoms do not spend enough time at the electric field of the cathode for ionization. This limits the number of particles transferred to the discharge and potentially cause s clog ging problem s at the needle valve. Another problem was lack of stability in the discharge. The discharge would slowly extinguish after ~30 min and clos e inspection showed contamination on the sharp edge of the cylindrical anode. Fi nally, after a period of about one month, no particle ionization was observed. The careful inspection of the cathode under a microscope showed that the sharp edge of the cathode orifice, where the particles enter the discharge, was etched (Figure 3 5). The sharp edges are responsible for an intense electric field that can enhance ionization. The use of anode B with a higher surface area was a solution for such aforementioned problems. In this configuration, the flow does not pass through the discharge quick ly, but ra t her has to travel in a diagonal way towards the four off axis anode orifices. As a result the atoms and the particles will spend a higher time in the discharge area, allow ing the use of higher flow rates. Figure 3 6 shows the Cs signal vs. repe ller delay time from CsI aerosols for the two anode configurations. The larger surface area of this new anode also provide s higher signal stability. In addition to discussed advantages of anode B it also provides a more intense electric field (Figure 3 7) at the point that particles enter the discharge; this compensates for the lack of sharp edges and allows for stabl e particle ionization. Cathode D ark S pace (CDS) Understanding the cathode dark space 79 (or fall region) is es sential in explaining the results of particle ionization. Cathode dark space is the thin layer in front of the cathode which is necessary to maintain the discharge. The electric field decreases linearly across this region from a large value at the cathode to eventually zero at the end of this region. When the anode -cathode distance decreases, the positive column and later the F araday dark space and finally the negative

PAGE 77

77 glow will disappear in the anode. While the fall region also shrinks, it remains as a com ponent to sustain the discharge. Anything that facilitates the function of the discharge will shorten the cathode dark space. Increasing temperature, pressure and voltage, as well as utilizing a gas with lower ionization energy, or a cathode material that can readily emit electrons will all reduce the length of the cathode dark space.7 In glow discharge mass spectrometry, if an atom is ionized in the cathode fall region, the electric field of the cathode will attract the ion towards the cathode : as a consequence, the ion will not be detected. Atomic emission does not face such a problem. As a result, a shorter cathode fall region and a longer mean free path of the atoms are desired for mass spectrometry .9 The mean free path needs to be long to prevent atoms from undergoing collision ionization with argon ions and fast argon atoms ( the main elements for sputtering) until they leave the catho de fall region. Ideally the mean free path of the atoms must be longer than the cathode fall region. Effect of the F low and S ource P ressure To study the effect of the particle flow, 3 different particle flows of 440, 630 and 875 ml/ min were used. Figure 3 8 shows that the increase in flow did not change the shape of the Cs signal profile significantly ; t he signal maximum was only shifted a few microseconds to earlier delay times. This was to be expected as the ions move faster in the higher flow. When the aerosol flow doubles from 440 to 875 ml/min, the Ar and Cs signals increase. However, closer examination shows that although the amount of analyte transferred to the discharge is only doubled, the signal has increased by the factors of 4 and 9 for Ar and Cs respect ive ly This findin g indicates that a factor other than flow itself must play a role here. It is important to note that anode B provides a smaller effective surface area for the flow to proceed towards the first stage compared to anode A (5.4 mm2 vs. 20 mm2). As a result the flow

PAGE 78

78 is at the sonic regime and the pressure of the source is much higher than the first stage pressure. At this regime, the flow is a direct function of the effective surface area and the upstream pressure.80 F = constant. Pup stream. As (3 1 ) In this equation, F is the flow from the source into the first s tage and As is the total surface area of the holes on anode B As mentioned previous ly, an increase in pressure reduces the length of the cathode dark space which allows more ions to escape the fall region. It has been shown that a pressure increase fro m 50 to 100 Pa in a 1400V Grimm -type argon discharge resulted in a decrease in length o f the cathode fall region (CDS) from 3.5 mm to less than 1 mm.7 Another effect of the increase in pressure is an increase in mea n energy of the argon ions and fast argon atoms. These are the main species responsible for sputtering, and also vaporization atomization and ionization of the particles at the cathode dark region. For instance, as pressure is increased from 50 to 100 Pa, mean energy of the argon ions at the maximum of their profile increases from 170 eV to 220 e V For fast a rgon atoms, this mean energy increase is from 44 to 47 eV .7 Finally, higher pressure means higher population density of ions and atoms. This density increases the number of collisions and assists in vaporization, atomization, and ionization. Additionally, it is important to note that the increase in pressure reduces the mean free path of the atoms and increases the possibility of ionization. This increases the ion signal due to increase in ionization in the negative glow. However, the ions that are formed in the CDS will not be abl e to leave this region; hence the cathodes electric field will attract them back. Experimental results illustrate that other factors overcome this problem and the overall signals increase as flow increases.

PAGE 79

79 Effect o f the Firs t Stage Pressure The pressure effect was studied using the optimum flow of 875 ml/min. Figure 3 9 shows that the signal profile and intensity changes with an increas e in the pressure from 1.36 torr to 3.1 torr. The 1.45 torr offers the best signal intensity and also shows 3 maxim a The signal profile change due to the first stage pressure (P1) does not follow a very clear pattern. It seems that the increase in pressure at first increases the signal while further change in pressure reduces the signal. It is important to note that the pr essure of the first stage is increased by reducing the pumping at this stage. Such a change does not affect the pressure of the source (Ps), as the source pressure is a function of the flow. However the increase in P1 increases the flow of ions into the s econd stage and consequently increases the signal. The number of collisions (with other species and the walls) is also a function of the pressure which has a negative effect on ion transfer yield. Effect of the Cathode Potential P ulse Width The pulse wid th was changed from 30 s to 150 s for a 4 mM CsI solution. The increase in pulse width does not change the intensity or the pattern (2 maximums) of the profile. The higher pulse duration just stretches the signal along the time domain. This can be seen i n Figure 3 10, in which where the two pulse durations are compared for Cs. Since the power generator is limited to maximum of 8 mA, one can not just increase the pulse at the same pulse potential. In order to keep the average current within instrument rang e, the 400 Hz frequency was reduced to 40 Hz in the 150 s pulse. This was done using a homemade frequency divider with the reduction in frequency by a factor of 10. Using an external frequency (home built) divider the TOF will send 10 triggers to the pulse generator of which only one provides a high voltage to the source. However the detector collects data for all 10 triggers and average s these collections As a result not only the signal and

PAGE 80

80 background noise increases due to less sampling per integration time interval (40 vs. 400 per second), but also the signal intensity drops significantly, as it is the average of one signal and 9 backgrounds. The signal intensity in Figure 3 10 was modified to compensate for this signal decrease. Effect of the C athode P otential The effect of the cathode potential on pa rticle ionization for both CsI and NaI particles is depicted on Figure 3 11 at a fixed repeller delay of 62 s. When the pulse potential increases, the energy of species in the discharge goes up and, as a result, more particle vaporization atomization -ioni zation will occur. This will result in a signal increase. Furthermore, the increase in pulse potential reduces the CDS length, and consequently increases the signal. It seems the plot of signal vs. potential reaches a plateau. This effect is caused by the fact that when the CDS becomes too short, then the particles spend an extremely brief time in this high -energy region, therefore jeopardizing the vaporization process. More importantly, an increase in cathode potential will increase the electric field of t he cathode, allowing fewer ions to escape the cathode fall region. In the following section, the effect of the electric field will be examined more in depth. CsI vs. NaI at Different Repeller D elay T imes A solution of CsI (2 mM ) and NaI (5 mM ) was used a nd the signal for Na and Cs, as well as i odi n e was studied (Figure 3 12). Due to lower sensitivity for Na, its signal is much lower than Cs even if the concentration of Na is higher. Iodine has almost the same sensitivity as Cs and its signal is bigger due to higher concentration (7 mM vs. 2 mM ). The argon and copper profiles for the same experiment are shown for comparison. The two argon maxima are located in between the corresponding m axima in the Na and Cs profiles. This observation matches the atomic weight of the three. That means the smaller Na atoms reach the repeller faster than Ar Cs

PAGE 81

81 and I. Cesium and i odi n e have almost the same mass ( m/z = 133 and 127) and their profiles are very similar to each other. Ionization M echanism and Sodium vs. Cesium S i gnal Comparison As Figure 3 11 demonstrate s the Cs signal from CsI particles is much higher than the Na signal from the NaI particles, whereas both aerosols are generated from salt solutions of equal concentrations (4.0 mM ). The signal for iodine, however is the same for both CsI and NaI. This difference in signal sensitivity could be due to the advantage of Cs and I signals over Na signal at a certain repeller delay time, however F igure 3 12 shows such discrimination exists at all repeller delays (Na vs Cs and I). O ne might attribute this finding to the higher ionization energy of sodium (5.14 eV) compared to Cs (3.89 eV). This cannot be true for a variety of reasons: f irst, t he energy of discharge species in the negative glow is so high that these spec ies are not able to discriminate between Na and Cs. The energ ies for the meta -stable argon atoms are 11.55 eV and 11.72 eV while Ar+ is 15.76 e V Secondly, when NaI particles are used, the iodine signal is much higher than Na, and iodine has a higher ioni zation energy of 10.46 eV. Another possibl e explanation of the low Na signal could be the higher melting point or heat of evaporation of NaI. Table 3. 2 shows the thermodynamic information for these two compounds.81 It is clear that the process es of vaporization for CsI and NaI do not differ energetically. In addition, when a solution containing both NaI and CsI was nebulized, the Na signal is still much lower than Cs. In this case both compounds share the same crystal so vaporization energy will be the same for them. The last process to consider as the reason for the difference in Na vs. Cs signal is atomization of NaI (g) and CsI (g). This possibility is also rejected, as iodine has a much higher signal than Na (Figure 3 11). If the atomization was the reason, the signal for I must also be much lower for the NaI particles compared to that of CsI. However, the plots of Figure 3 11 indicate that both NaI and CsI with the same concentration

PAGE 82

82 result in almost the same signal for iodine. Any significant difference in vaporization or atomization would not allow this to happen. Another way to compare the energies involved is through the use of lattice energy combined w ith electron affinity ( 297 kJ/mol) and ionization energy of iodine. For example, for CsI we have:81 CsI (s) > Cs +(g) + I -(g ) H = 604 k J/mol I -(g) > I (g) H = 297 k J/mol I (g) > I +(g) H = 1009 k J/mol A dding the above reactions will result in the overall reaction: CsI (s) Cs +(g) + I +(g) H = 1910 k J/mol In other words, the total energy for vaporization atomization and ionization of both atoms is 1910 kJ /mol for CsI. Similar calculations will result in 2010 kJ /mol for NaI a mere 5% higher than CsI. Suppression of the Na signal can be explained if we assume that most of the vaporization and atomization of the particle occurs in the cathode dark space of the disch arge (Figure 3 13). Since the cathode is not heated to high temperatures and none of the negative glow species have enough energy for particle vaporization the assumption of vaporization in CDS region must be valid. On e other possibl e mechanism for partic le ionization is the aerosol deposition on the cathode surface and its sputtering from the surface. This phenomenon must be very unlikely, as the geometry used in this experiment does not result in particles impacting the cathode. On the other hand the pa rticles enter from the back of the cathode. Another experiment was also run in which particles where allowed to enter discharge for about 3 mins while the discharge was off. Later the discharge was turned on and no signal was observed. In another investig ation, a droplet of 10 mM CsI solution was placed on the surface of the cathode and

PAGE 83

83 after vaporization of the water it was placed into the source. Although a small ion signal was observed, it was much lower than the case of dynamic introduction of partic les. These all indicate that the major source of ion signal is due to particle ionization in the CDS by high energy argon ions atoms. If the resulting ions reach the space beyond the cathode dark space, they will be detected (Figure 3 13). However, it seem s that most of the atoms are ionized within the cathode dark space and will not make it to the detector. This is due to the short mean free path of the atoms ( 10 m for Cs, I and 25 m for Na)82 at the discharge condition compared to the CDS length (le ss than 0.7 mm = 700 m). The reason for detection of some ions, despite this large loss of ions in CDS, is the momentum of these ions. If an ion is formed in the CDS, even if it is attracted towards the cathode it is still possible for this ion to escape the CDS region due to its momentum. Assume a Cs and Na atom are moving away from the cathode with the velocity of V0 at the same distance from the cathode. If they both get ionized at this point (Figure 3 14) the cathode will attract both of them with an equal force (F = E. q ) ; h owever for the same force the smaller ion will have a higher acceleration towards the cathode (a = F/m) In other words, Na atoms will accelerate 6 times faster than Cs towards the cathode. The time that it takes for Cs to stop before it moves back will be 6 times longer. Thus Cs will travel a distance 6 times larger than Na before it is attracted towards the cathode. As a result Cs has a bett er chance of leaving the CDS and mak ing it to the detector. Since the reaction cross section of these ions with other species is not well known, it is hard to evaluate the effect of the cross section of ionization on the ionization rate.

PAGE 84

84 Table 3 1. Summary of the experimental parameters. Category Parameter Fixed V alue or Range Nebulizer Dehydrator Argon T an k P ressure 90 psi Nebulizer A rgon F low 1.0 L /min Dehydrator A rgon F low 2.0 L / m in Solution F low 1.0 ml/min Nebulizer T emperature (high) 150 C Nebulizer T emperature (Low) 4 C Dehydrator T emperature 150 C High V oltage P ulse Pulse P ote ntial (V) 1.5 2.5 kV Pulse W idth 20 150 s Pulse Frequency 100 1000 Hz Average Current ( I ) 1 8 mA Gas F low and P ressure In the S ource Flow 4001000 ml/min Vacuum V alve P osition Turn to A djust P1 Stage 1 Pressure (P1) 1.3 3.0 torr Back S leeve P ressure (Ps) 2 30 torr

PAGE 85

85 T able 3 2. Thermodynamic properties of CsI and NaI.81 Property CsI NaI Melting point / C 632 661 Heat of melting / kJ /mol 23.9 23.6 Heat of evaporation / kJ /mol 150.2 less than 160* Lattice energy / kJ /mol 604 704 *The info rmation was not available for NaI, but the value for NaBr is 160.

PAGE 86

86 trigger exhaust GD TOFMS delay Power supply ICM Scope Dehydrator Nebulizer Ar pump Salt in water I ave I V R Figure 3 1 The overall schematic of the instrumentation for pulsed GDMS of particles. Anode Cathode base Insulator particles Figure 3 2. Schematic of the cathode design for particle introduction.

PAGE 87

87 A A C C B B A B Skimmer Skimmer Figure 3 3. Anode A v s. B. cathode Anode Insulator TOF detectorpump particles 2ndstage Gauge (P1) Gauge (PS)1ststage ions Vacuum valve Figure 3 4. Detailed view of the ionization source with anode B. Only the discharge interface of the cathode is shown for simplicity.

PAGE 88

88 Corrosion (due to sputtering) Cathode Anode Figure 3 5. The sharp edge of the cathode orifice, where the particles enter the discharge, was etched due to sputtering. -0.5 0 0.5 1 1.5 2 2.5 3 3.5 0 10 20 30 40 50 60 70 80 90 100 Repeller delay time ( s)Cs Signal (mV)Cs signal for 4mM CsI Figure 3 6. The Cs+ ion s ignal vs. repeller delay time from CsI aerosols for the two anode configurations. Refer to figure A 13 for a sample with standard deviation of the mean.

PAGE 89

89 Iso potential lines are shown for both configurations2000V ~0V Figure 3 7. The electric field of anode A vs. B. [Courtes y of Ashkan Behnam, Department of Electrical Engineering, University of Florida, for modeling the electric field using Medici 4.0. Durham, NC : Synopsys, Inc. (2004)]

PAGE 90

90 Signal ( a.u ) 0.9 1.4 1.9 2.4 2.9 3.4 3.9 4.4 4.9 0 10 20 30 40 50 60 70 80 90 100 440 630 875 Signal ( a.u )440 630 875 Ar -0.5 0 0.5 1 1.5 2 2.5 3 3.5 0 10 20 30 40 50 60 70 80 90 100 875 ml/min 630mL/min 440 mL/min Repeller delay time ( s)Signal (mV)Cs Figure 3 8. Ar and Cs signal vs. repeller delay at three different aerosol flow rates into the source. Refer to figure A 13 for a sample with standard deviation of the mean.

PAGE 91

91 Repeller delay time ( s)Signal (mV) -0.5 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 20 30 40 50 60 70 80 90 100 3.10 torr 2 torr 1.45 torr 1.36 torr RSD ~10% Figure 3 9. Effect of the first stage pressure on signal profile and intensity. Refer to figure A 13 for a sample with standard deviation of the mean.

PAGE 92

92 Delay time ( s) Signal ( a.u ) 0.9 1.4 1.9 2.4 2.9 3.4 0 20 40 60 80 100 120 140 160 180 200 220 30 s (400Hz) 150 s (40Hz) 60 180 Delay time ( s) Signal ( a.u ) 0.9 1.4 1.9 2.4 2.9 3.4 0 20 40 60 80 100 120 140 160 180 200 220 Delay time ( s) Signal ( a.u ) 0.9 1.4 1.9 2.4 2.9 3.4 0 20 40 60 80 100 120 140 160 180 200 220 30 s (400Hz) 150 s (40Hz) 60 180 signal ( a.u .) Repeller delay time ( s) 0 20 40 60 80 100 120 140 160 180 200 220 0 0.5 1.0 1.5 2.0 2.5 Figure 3 10. Pulse width, where the two pulse durations of 30 s and 150 s are compared for Cs signal. Refer to figure A 13 for a sample with standard deviation of the mean.

PAGE 93

93 Cathode potential ( kV)Signal (mV) 0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50 1 1.2 1.4 1.6 1.8 2 2.2 2.4 2.6 Cs I 0.00 0.50 1.00 1.50 2.00 2.50 3.00 1 1.2 1.4 1.6 1.8 2 2.2 2.4 2.6 2.8 Cathode potential ( kV)Signal (mV) Na+ I+ Figure 3 11. Effect of the cathode potential on particle ionization for C s I and NaI. Standard deviation of the mean is shown for an average of n ~ 15. The error bars for iodine in CsI are not shown for demonstration purposes as the RSD for I and Cs are similar.

PAGE 94

94 0 20 40 60 80 100 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 0 10 20 30 40 50 I Cs Na Cu Ar Signal (mV)Ar signal (mV)Repeller delay time ( s) Figure 3 12. Signal vs. delay time for different elements. particle CDS Ar(fast)and Ar+ NaI(g) particle Na I Na + I + Towards The skimmer particle CDS Ar(fast)and Ar+ NaI(g) particle Na I Na + I + Towards The skimmer Figure 3 13. Particles enter from the cathode (left). After vaporization, atomization and ionization, if atoms make it out of the CDS, they c an ionize in the negative glow Then pass through the anode and reach the skimmer ( s izes are not to the scale). The negative glow region is not shown here due to small anode -cathode distance and for demonstration purposes.

PAGE 95

95 CDS Na + Cs + a = E.e / mNaa = E.e / mCs vovo Cs signal Na signal CDS Na + Cs + a = E.e / mNaa = E.e / mCs vovo Cs signal Na signal Figure 3 14. The heavier Cs ion travels 6 times farther than Na before it goes back towards the cathode. The Cs and Na ions formed close to the end of the CDS region can still escape the electric field due to their momentum. The Cs and Na signal arrows show the relative distance these at oms have to the CDS border

PAGE 96

96 CHAPTER 4 PULSED GLOW DISCHARGE MASS SPECTROMETRY OF PART ICLES: FUNDAMENTAL STUDIES AND DIAGNOSTICS Intro d uction In the previous chapter, the task of introducing particles into the glow discharge was discussed. Of particular i nterest was the effect of different variables such as aeroso l flow rate, discharge pressure and discharge potential on particle ionization. The vaporization atomization of particles is believed to mostly happen in the cathode dark space of the discharge, where high-energy argon ions and fast argon atoms have energies as high as 200 eV. Subsequently, the atoms will transfer into the negative glow, and their ionization occurs in this region. In the cathode dark space (CDS) some ions will be formed as well, a lthough most of these ions ultimately return to the cathode due to the large electric field in this region. Nonetheless those ions formed in part of CDS that is adjacent to the negative glow can still escape the electric field of the cathode due to their momentum. T he question remains as to what deg ree particle vaporization occurs ; i n other words, does a particle completely evaporate or does it go through partial vaporization? The use of different particle sizes might bring light to such questions. In orde r to produce salt particles of variable size s in this investigation different concentrations of salt will be used for nebulization. In this chapter the results of varying salt concentration and its effect on particle size and ionization signal will be stu died. Another interesting observation was higher signal intensities for Cs and I compared to ion intensity of Na. In the previous chapter, the transfer of ions across the CDS region was recognized to be the main reason for this phenomenon. However, another place in which the possibility for discrimina tion between the ions may occur is the GD interface with the TOF, where ions are transferred to the second stage of the mass spectrometer In this chapter, the effect

PAGE 97

97 of skimmer orifice on different ion signals and t he effect of cathode temperature on signal intensity will be discussed. Experimental Most of t he experimental setup used in this chapter is identical to that of C hapter 3. A DC power supply along with two heating cartridges was used to heat the catho de base to the desired temperature The cathode temperature was measured by a thermocouple thermometer (Digi Sence, Oakton Instruments Vernon Hills, IL ). Two inch holes and one 1/16 inch hole were machined on the cathode based in order to place the heat ing cartridges and the thermometer respectively. The skimmers with variable orifice diameters were made by LECO Corp. ( St. Joseph, MI ). For particle size distribution measurements a particle size analyzer (Aerosizer LD, API Inc., TSI Inc, Shoreview, MN) was used. For modeling the flow, Fluent 6.3.1 (ANSYS Inc. Canonsburg, PA) was used. Results and D iscussion Effect of Solution Concentration Solutions of CsI and NaI with different concentrations were used to generate aerosol. Signal vs. concentration was a linear function for these salts. Sensitivities for Cs, I, and Na were 0.61, 0.55, and 0.07 (mV/mM), respectively. The ion signal vs. concentration for cesium iodide and sodium iodide is shown in Figures 4 1 and 4 2. The iodine signal is independent of the type of salt used; this is because the difference between heats of vaporization of these two salts is small and in both cases the entire particle evaporates. The linearity of signal vs. concentration also indicates that the discharge is completely evapo rating the particles of both salts. The particle v aporization efficiency will later be discussed in more detail.

PAGE 98

98 Relative Sensitivity Factors In the previous chapter, it was observed that the relative sensitivity for Cs and I was much higher than Na. In ge neral, the ion sensitivity in GDMS is a functi on of : (1) events occurring in the GD source, (2) the ion transfer efficiency through the ion lenses and mass analyzer, and (3) the detector efficiency .29 The effect of atomic mass on ion transfer in the GD source along with the role of the cathode dark space electric field was discussed in Chapter 3 The extraction of ions from the first stage to the second stage through the skimmer orifice is another crucial step in the transfer of ions. In our system, the skimmer and sampler are both grounded and the pressure difference ( 1.2 torr vs. 0.001 torr) between the two stages is the only driving force for ion extraction. The first ion extraction lens is loca ted about 2 0 mm behind the skimmer orifice. A large portion of the flow will be scattered and pumped out just past the skimmer orifice. The ions with smaller drift velocities (heavy ions), such as Cs and I, will face less scattering and have a higher chanc e of reaching the ion extraction lenses. On the other hand, the smaller ions such as Na will drift and pump out; as a result, they carry a lower possibility for detection. This is potentially another explanation for higher Cs and I signal compared to Na. F igure 4 3 shows t he fluid dynamic modeling for Grimm source geometry similar to the employed source configuration. As it is demonstrated, the flow will deviate from the center after it passes through the skimmer orifice. If this explanation is correct, cha nging the skimmer orifice size should affect the relative sensit ivity for heavy and light atoms; thus the increase in skimmer orifice diameter should result in higher signa l, especially for lower -mass ions such as Na. Effect of Skimmer Orifice As the skimm er orifice diameter was increased from 0.43 mm to 0.56 mm, signal sensitivity increased for all elements as expected. As the orifice diameter/surface area increases, more flow is transferred into the second stage and the signal intensifies for all ions. Fi gure 4 4

PAGE 99

99 shows the relative sensitivity for Cs, I, and Na at two different skimmer orifice diameters. It is also observed that the Cs:Na signal ratio changes from ~10 to ~3 as skimmer orifice diameter increases. This is due to reduction of ion scattering ( which influences the smaller ions more) at the extraction orifice when a larger orifice is used. Effect of Cathode Temperature As the cathode temperature increased from 23 C to 75 C, the particle signal increased by a facto r of ~1.25 (Figure 4 4). Figur e 4 5 show s the particle ionization signal for various cathode potentials at 75 C and a skimmer orifice diameter of 0.56mm for CsI. Figure 4 6 shows the same for NaI. Figure 4 7 compares the iodine signal vs. cathode potential at 23 C and 75 C. At 75 C the signal reaches a plateau at ~1.1 kV, while at 23 C it reaches its plateau at ~1.8 kV. This demonstrates that the particle vaporization is facilitated at the higher temperature. Since the temperature increase from 23 C to 75 C is not an intense inc rease in the energy of the species, there might be other reasons for these observations. One explanation could be that the higher temperature of the discharge decreases the length of the cathode dark space, and consequently the intensity of the electric fi eld increases. In such a condition, argon ions and fast argon atoms of higher energy participate in the vaporization of particles. In addition, in a shorter cathode dark space more atoms are able to leave the electric field before they ionize through colli sions with other species. This way, the atoms will ionize in the field free glow region, where they will not be attracted back to the cathode surface. Thus, more ions will be transferred to the skimmer orifice. Particle Vaporization Efficiency The ion sig nal intensity at the detector (S) in a mass spectrometer is a function of vaporization, atomiza tion, ionization, ion transport and ion detection efficiencies: S = Const f (vap) f(atom ization ). f (ionization ). f (ion transport ) f (detection ) (4 1)

PAGE 100

100 A tomization efficiency is only a function of the salt type and doe s not depend on concentration. Ionization, ion transport and detection efficiencies are also constant for a specific ion and are concentration independent. If the vaporization efficiency is 1 00% then the signal will be proportional to particle mass which linearly corresponds to solution concentration. Thus, the linearity of the signal -concentration graph is indicative of complete particle vaporization. If the particle vaporization is not co mplete, this means that only a constant thickness of particles (x) is sp uttered by the glow discharge plasma As a result, the amount of mass evaporated from each particle will be a function of particle surface area: 2 2.. 4 r const x r x A V m (4 2) Where V is volume of the dry particle, is particle (salt) density, A is particle surface area and r is particle radius. If one assumes constant total number of particles and constant volume mean diameter (DV) of wet aerosol regardless of salt concentration, the mean volume of the dry particle will be proportional to the salt concentration (C). It is important to note that as it was discussed in Chapter 3, the particles that arrive in the glow discharge plasma are dry salt particles. Since radius is proportional to the cubic root o f volume in a dry particle in consider ation of E quation 4 2 the sputtered mass should be proportional to C2/3: 3 / 2. C const m (4 3) If the above assumption s are correct and particle vaporization is partial, then the signal would not show a linear r elationshi p with concentration. Therefore the linearity of signal vs. concentration is a further evidence for complete particle vaporization. Figure 4 8 depicts the comparison between a partial vaporization model and the experimental data. However, one might argue that the assumption of constant diameter for the wet particles is erroneous and wet particle diameter varies as a function of the salt concentration. Table 4 1

PAGE 101

101 shows the experimental values for volume mean diameter (DV) for particles generated from diffe rent salt concentrations. It is well known52 that the wet particle diameter produced by an ultrasonic nebulizer is a f unction of nebulizer frequency (f), solution surface tension ( and solution density ( ): 3 / 1 2) 8 ( f (4 4) 34 0 vD (5 4) Using a nebulizer frequency of 1 MHz and the density an d surface tension of pure water, the wet particle diameter for this nebulizer was calculated to be about 5 m. However, this theoretical value for average wet particle diameter does not agree with experimental and known51 parameters (10 15 m) of our commercial nebulizer ( due to reasons that are beyond the scope of this work ). Figure 4 9 shows the experimental DV values for dry CsI particles compared with calculated ones based on different wet particle diameters. Both experimental and calcul ated DVs for dry particles show similar trends. Based on experimental dry particle diameters it is expected that the mean volume diameter of wet particles generated by the nebulizer dehydrator unit must be about 8 10 m for CsI at salt concentrations of 1 4 mM. As salt concentration increases, the particle diameter increases as well. Although the experimental results for particle diameter are not exactly the same as calculated results, they would still not result in a linear signal vs. concentration graph. One interesting observation was the evidence of decrease in the diameter of the wet particle as the salt concentration increases. For example, the experimental DV value for dry particles generated from a 0.50 mM solution is similar to what is expected from a wet particle diameter of 12 m. However, for the 4.0 mM solution the expected wet particle diameter is

PAGE 102

102 closer to that of the 8 m curve. This behavior can not simply be explained by E quation s 4 4 and 4 5 The only factor in E quation 4 4 a ffected by salt concentration is surface tension which does not change significantly as a function of concentration. For instance, the surface tension of a solution of 0.58% NaCl (wt%) in water is 72.92 dynes / cm, compared to 74.93 dynes / cm when concentration of salt increases nearly 10 times (5.43%).81 This small increase in surface tension (1.028 tim es) when enter ed into E quation 4 4 will result in a minor increase (0.9 %) in wet particle diameter. It is not easy to find a study in whic h wet particle diameters are measured directly. Only one example was found in which DV of wet aerosol generated by 2 different ultrasonic nebulizers and an air -jet nebulizer were measured by laser diffraction analysis.83 The study was aimed to measure the particle size as an indicator of medical nebulizer performance in pulmonary drug delivery.83 The res ult of this study showed that the wet particle diameter is dependent on salt concentration. I nterestingly this is in agreement with our results ; as salt concentration increased the wet particle diameter decreased.83 A dditional evidence for complete particle vaporization is the result of signal vs. cathode potential. Figure 4 5 shows the signal for Cs and I increasing as cathode potential increases at first, and later the signal reaches a plateau. This shows that at a certain potential, the particles have reached complete vaporization and the potential increase does not improve the signal anymore. Finally, another experiment was devised to investigate the extent of particle vapor ization. In this experiment, four different solutions with a constant concentration of CsI = 2 .0 mM were prepared. The concentration of NaI in these solutions ranged from 0 to 6.0 mM. As Figure 4 10 shows, the Cs signal was almost constant regardless of Na I concentration. If the particle

PAGE 103

103 vaporization had not been absolute, Cs signal would be expected to decrease as the concentration of background NaI salt increases. Effect of Background Salt Four different solutions with a constant concentration of CsI = 2.0 mM were prepared. The concentration of NaI in these solutions was 0, 2.0 4 .0 and 6 .0 mM. Figure 4 10 shows the Cs signal vs. NaI concentrations for these four solutions. Interestingly, the increase in concentration of NaI (background salt) does not hav e a significant effect on Cs signal. Figure 4 11 shows the Na signal for these four solutions vs. NaI concentration and compares it with a similar plot in which no Cs background is present. The results show that no significant change in sensitivity of Na i s observed due to the presence of 2 .0 mM CsI. Figure 4 12 shows the iodine signal produced from these solutions. The concentration of iodine is equal to the total concentration of both NaI and CsI solutions. These results reconfirm the complete vaporizatio n of salt particles in the glow discharge plasma. Effect of Particle Ionization on Discharge Species When CsI particles enter the discharge the copper signal increases while the argon signal is subject to a minor depression. Figure 4 13 shows the ion signa l vs. time for Cs, Cu, and Ar for an experiment in which particles were generated by nebulization of a 4 mM CsI solution for 30 seconds. The spike in copper signal is indicative of increasing cathode sputtering efficiency. This can be explained by the temp oral increase in press ure due to particle vaporization and also the bombardment of the cathode by Cs and I ions. The decrease in argon signal is due to loss of argon ions in ion exchange reactio ns between argon ions and Cs, I and Cu atoms. Another reason for the decrease of argon signal could simply be the plasma cooling down due to introduction of particles. The temporal decline in argon signal upon particle introduction has al so been reported in ICP studies

PAGE 104

104 Effect of Cathode Potential Particle Ionization Mechanism Figure 4 14 compares the iodine signal from the particle ionization with the signal s from argon and copper. The increase in copper signal (due to particle introduction) is also shown as a function of cathode potential. The Cu signal increases li nearly with cathode potential due to an increase in cathode sputtering. Conversely, the iodine signal reaches a plateau and does not follow a linear pattern. The pattern of particle ionization signal (iodine) is similar to that of argon signal. This indica tes that particles ionize in the gas phase as opposed to the cathode surface. The change in Cu signal indicates the increase in Cu signal due to introduction of particles. This signal increase follows a pattern similar to that of particle ionization, which indicates the ions resulting from particle ionization are the source of this increase in sputtering.

PAGE 105

105 Table 4 1. Dry particle mean volume diameters (DV) for different concentrations of CsI and NaI. A particle size analyzer (Aerosizer LD, API Inc., TSI In c, Shoreview, MN) was used for particle size distribution measurements. Salt concentration /mM DV /nm for CsI DV /nm for NaI 0.1 290 10 NA 0.5 333 6 316 8 1 380 5 350 10 2 441 8 385 9 4 510 5 470 10 10 610 15 510 15 40 820 20 NA 100 1023 20 NA

PAGE 106

106 CsI concentration (mM)Signal (mV) R2 = 0.9983 0 0.4 0.8 1.2 1.6 0 0.5 1 1.5 2 Cs I I Cs Figure 4 1. Quantitative study of cesium iodide with different solution concentrations. The error bars for iodine are not shown for demonstration purposes as the RSD for I and Cs are similar. Ref er to figure A 14 in Appendix A for more information about linear dynamic range and detection limit. R2 = 0.9988 0 0.4 0.8 1.2 0 0.5 1 1.5 2 2.5 Na I NaI concentration (mM)Signal (mV) Na I Figure 4 2. Quantitative study of sodium iodide with different solution concentrations.

PAGE 107

107 1ststage, p = 1.2 torr Na+ I+ Ion Lens 1 I+ Na+ 2nd stage, p = 0.001 torr Figure 4 3. Modeling the flow for the glow discharge source. Courtesy of Ayyoub Mehdizadeh, Department of Mechanical engineering, University of Florida. The software used is Fluent 6.3.1 (ANSYS Inc. Canonsburg, PA). Divergence of smaller ions at skimmer o rifice can explain the low signal value for sodium in comparison with iodine and cesium.

PAGE 108

108 0 0.5 1 1.5 2 Cs I Na Small skimmer orifice (d= 0.43mm) Large skimmer orifice (d= 0.56mm) Large skimmer orifice (heated cathode) Sensitivity (mV/mM) Figure 4 4 Effect of skimmer orifice size and heating the cathode on sensitivity for Cs, I and Na.

PAGE 109

109 Cathode potential ( kV)Signal (mV) 0.00 2.00 4.00 6.00 8.00 0.5 1 1.5 2 2.5 Cs I Figure 4 5. Effect of cathode potential for CsI 4 mM at cathode temperature = 75C and skimmer orifice diameter = 0.56 mm. The error bars for iodine are not shown for demonstration purposes as the RSD for I and Cs are similar.

PAGE 110

110 Cathode potential ( kV)Signal (mV) 0.00 2.00 4.00 6.00 8.00 0.5 0.7 0.9 1.1 1.3 1.5 1.7 1.9 2.1 Na I Figure 4 6. Effect of cathode potential for NaI 4 mM C athode temperature = 75C and skimmer orifice diameter = 0.56 mm.

PAGE 111

111 Cathode potential ( kV)I+ ion signal (mV) 0.00 2.00 4.00 6.00 8.00 0.5 1 1.5 2 2.5 3 75oC 23oC Figure 4 7. The change in signal vs. cathode potential for iodine ion using a 4.0 mM cesium iodide solution.

PAGE 112

112 -0.4 0 0.4 0.8 1.2 1.6 2 0 0.5 1 1.5 2 Experimental Partial evaporation NaI concentration (mM)Signal (mV) Figure 4 8. Comparing the ion signal vs. salt concentration for the experimental results vs. partial vaporization model.

PAGE 113

113 0.0 100.0 200.0 300.0 400.0 500.0 600.0 700.0 800.0 0 1 2 3 4 Experimental caculated based on wet = 5 calculated based on wet = 8 calculated based on wet = 10 caculated based on wet = 12 CsI concentration (mM)Dry particle diameter (nm) 0.0 200.0 400.0 600.0 800.0 1000.0 1200.0 1400.0 1600.0 1800.0 0.1 1 10 100 Experimental caculated based on wet = 5 calculated based on wet = 8 calculated based on wet = 10 caculated based on wet = 12 CsI concentration (mM)Dry particle diameter (nm) Figure 4 9. Mean volume diameters for different CsI concentrations. Top: 0.1 mM to 4.0 mM, Bottom 0.1 to 100 mM (logarithmic scale) Experimental values are compared to c alculated values based on 4 different wet particle diameters

PAGE 114

114 0 1 2 3 4 5 0 2 4 6 8 Cs signal NaI concentration (mM) Cs Signal (mV) Figure 4 10. Cs+ ion signal from 2 mM CsI in different NaI backgrounds

PAGE 115

115 R2 = 0.9943 0 2 4 6 0 2 4 6 8 Na-in 2mM CsI Na from NaI only 0 2 5 2 6 4 2 4 3 2 2 2 2 0 1 CsI NaI # C / mM NaI concentration (mM)Na Signal (mV) Figure 4 1 1. Signal vs. concentration for Na+ ion at different NaI concentrations at the presence of 2 mM CsI background compared with no background.

PAGE 116

116 Total NaI + CsI concentration (mM)I Signal (mV) R2 = 0.9904 0 2 46 8 10 12 14 16 0 2 46 8 10 From NaI & CsI From CsI only From NaI only Figure 4 12. Iodine ion signal resulted from different NaI solutions at the presence of 2 mM CsI.

PAGE 117

117 40Ar63Cu Cssignal (mV)0 0 180Time (s)3.5 124.5 114.0 14.2 10.8 Figure 4 13. The effect of particle introduction on argon and copper ion signals: A package of particles is nebulized to the plasma will result in a decrease of argon ion, while it increases the copper signal. At t = 150 s, the di scharge is turned off.

PAGE 118

118 Cathode potential ( kV)Signal (mV)Ar signal (mV) -2.00 0.00 2.00 4.00 6.00 8.00 10.00 0.4 0.8 1.2 1.6 2 -25 0 25 50 75 100 125 change in Cu Cu I Ar Figure 4 14. Comparing the effect of cathode potential on different ion signals.

PAGE 119

119 CHAPTER 5 P ULSED GRIMM GLOW DIS CHARGE TIME OF FLIGHT MASS SPECTROM ETRY STUDY OF AEROSOLS GE NERATED BY LASER ABL ATION OF SOLID SAMPL ES Introduct ion The field of inorganic mass spectrometry deals with the elemental analysis of solid and liquid samples. I nductively coupled plasma mass spectrometry (ICP -MS) is the most common way for elemental analysis of inorganic solutions. However solid samples a re harder to analyze. These samples must first go through sample preparation and dissolve in proper solvent before being introduced into the ICP -MS. Sample preparation i s usually time consum ing and a method that allows for direct analysis of solids is desirable. In the past decades techniques such as spark source mass spectrometry (SSMS), glow discharge mass spectrometry (GDMS), secondary ion mass spectrometry (SIMS) and thermal ionization mass spectrometry (TIMS) were used as methods of ch oice for direct solid analysis.84 In recent years the high sensitivity of ICP MS along with the adva nces in laser technology have le d to development of Laser A blation In ductively Coupled P lasma M ass S pectrometry (LA ICP -MS) This technique h as become a very popular analytical technique for element and isotope -selective analyses of solid samples such as geological, archaeological, environmental and even biological matrices.85 This is due to the simplicity and versatility of this tech nique for solid samples, since t here is no need for time consuming sample preparations as needed for traditional ICP methods .86 Although GDMS is another common plasma source that has been employed for solid analysis, less work has been done that combines laser ablation with GDMS This is due to the fact that ICP is a much more powerful plasma that can reach higher sensitivity and lower limits of detection. In th e work, presented in this chapter the objective is to understand more about the

PAGE 120

120 fundamentals of particle interaction wit h our puls e d discharge plasma, and l aser ablation can therefore provide a potential pulsed source of particles for this purpose. Glow D ischarge and Lasers Laser Ablation coupled with Inductively Coupled Plasma Optical Emission Spectroscopy (LA ICP -OES) or Mass Spectrometry (LA ICP -MS) are well -established techniques of chemical analysis. M any of the articles which include both concepts of glow discharge and laser, are about the application of a laser beam for the purpose of plasma diagnostics .100 On the other hand, few reports are found describing the coupling of Laser Ablation with Glow Discharges (LA GD) .8791 Most of these articles use O ptical E mission S pectroscopy (OES ), as opposed to Mass Spectrometry for detection of the resulting ions/atoms .8890 In addition, in almost all the aforementioned GD laser studies, the laser interacts with the sample wi thin the glow discharge chamber .87 91 Indeed, in many of these cases, the sample itself is the cathode of the glow discharge. For example the polarity of a Grimm type glow discharge was reversed to have a hollow cathode formed by the anode of t he Grimm source, now acting as cathode.89 Then a Nd: YAG laser at 1064 nm and 1 kHz ( 0.76 mJ / pulse) was used for laser ablation of a sample of stainless st eel placed a s the cathode (anode considering the reverse phase) of the discharge The steel sample was directly ablated into the plasma and analyzed by OES. In another stu dy, King et al. us ed a pulsed glow discharge source as an auxiliary for excitation of ablated copper that was the cathode of the discharge as well.90 The ablated material was studied by means of OES. In a recent paper by Gunther et al LA -GD TOFMS was introduced and investigated to evaluate the analytical capabilities of this technique.87 Although in part of this work, the ablation sample was not the cathode itself, the experimental set up was different from that of a typical LA ICP. The sample and the pulsed (2 ms) hollow cathode discharge were both placed in a low pressure ch amber, from which ions are extracted towards the

PAGE 121

121 mass spectrometer. The closest analog to LA ICP MS is a recent work in which an atmospheric helium GD was used to detect laser ablated organic compounds such as caffeine and acetaminophen.92 To our knowledge, the study of the transfer of laser ablated metallic aerosols from an external ablation cell into a pulsed GD -MS has not yet been reported. LA -ICP -MS As mentioned above the field of LA GD -MS is not yet fully inv estigated by scientists. Due to this lack of literature resources for LA GDMS some of the principles of LA ICP -MS the closest field that has been studied extensively, will be discussed here It has been shown that experimental parameters such as the las er wavelength, energy and pulse width of the laser, the geometry of the ablation cell, the cell carrier gas, the particle transport efficiency and its atomization in the ICP plasma have an impact on the analytical performance of LA ICP -MS .93 Ideal L A -ICP -MS System An ideal laser ablation ICP -MS system must have three characteristic components :93 The ablated particles have the same overall composition as the sample itself The transport efficiency of aerosols is 100% The particles will go through complete atomization and ionization in the plasma source without affecting the plasma parameters.93 Fractionation However, the challenge of LA -ICP is performing fractionation-free analys is without the necessity that sample and standards matrices h ave to match .94 Therefore only matrices that have variety of standards to provide a matrix -matched calibration graph can be analyzed by this method. Consequently, LA ICP -MS can only be used to analyze samples like stainless steel, minerals and silicate glasses that have a range of available standards.95

PAGE 122

122 Particle Size Effect And Use Of Femto -S econd Lasers It has been shown that the use of femto -second lasers can reduce the effect of fractionation This is because the a erosols generated by this type of laser are stoichiometric and exactly reflect the composition of the sample itself.95, 96 The aerosols generated by laser ablation usually include particles that vary in size and c over a wide range of diameters from as small as 10nm to larger particles of a micron. The resulting particles have different chemical composition based on their size. In other words a source of fractionation could be due to the fact that the particles of different size have different composition and also different transfer efficiency. As a result the particles of a specific size that will transfer the most will influence the result of analysis the most. However these particles might have a chemical compos ition that does not necessary represent the bulk of the sample .97 Particle transport efficiencies of up to 80% have been reported using femto second lasers and helium as carrier gas.98 This was done by particle impacting an d measuring the weight loss in the sample. However, in order for this technique one must be able to neglect the re -deposition of particles on the sample surface, which might be a good approximation with helium but not with argon.86, 99, 100 Carrier Gas o f Choice Helium is usually the carrier gas of the choice and results in better accuracy and sensitivity when compared to argon.86 This is in part due to the higher ionization energy of he lium. The plasma breakdown and shielding is less intense for helium which reduces the propagation of material in the hori zontal mode (as opposed to vertical mushroom cloud mode) and less sample re -deposition on surface occurs in case of helium .93 However argon is easily available and cheaper. In addition, argon has lower ionization energy, and it requires a lower pot ential to form a discharge plasma. In our pulsed glow discharge system the introduction of helium resulted in

PAGE 123

123 total suppression of the discharge and the maximum possible potential delivered by our pulsed supply could not ignite a stable plasma discharge Laser P ulse D uration Short laser pulses interact with the sample directly, while longer pulses can result in heating the plasma.93 This can increase the electron density which will reduce laser sample interaction. As a result, the mass removal efficie ncy will decrease. This phenomenon is more pronounced at higher laser fluence. Another drawback of longer laser pulse s is the possibility of melting th at can result in fractionation .93 In order to reduce this fractionation it is recommended to focu s the laser below the surface of the sample to reduce laser fluence. Experimental The experimental se tup is very similar to that of Chapter 3 only the source of aerosol has changed; instead of an ultrasonic nebulizer -dehydrator for particle generation, l aser ablation of solid samples is the source of aerosol. Figure 5 1 shows the overall experimental setup for laser ablation glow discharge mas s spectrometry. The laser pulse ablate s the sample in the ablation cell from which the generated aerosol is transf erred by the flow of argon into the discharge plasma. The needle valve is located before the ablation cell to break the pressure from atmospheric to low pressure. The pressure of the ablation cell is about 20 torr, while the discharge pressure is 12 torr. The glow discharge geometry use d is the anode B geometry from Chapter 3 and mass spectrometer and discharge parameters are similar to those values used in the previous chapter unless otherwise noted. For a majority of the experiments, the optimum conditio n listed in Table 3 1 was used. The TOF detector was connected to a 500 MHz digital oscilloscope (TDS 724D, Tektronix Inc., Beaverton, OR) to obtain mass spectra with larger signal to noise ratio compared to the spectra collected by the instrument software

PAGE 124

124 Ablation Cell The ablation cell was machined from aluminum in three parts (Figure 5 1 ). Figure 5 2 also shows the geometric dimensions of the cell. The laser beam goes through a lens (focal length = 55 mm) and then enters the cell through a quartz window at the top of the cell. The sample is machined to the shape of a disc with a diameter of 12 mm and thickness of 10 mm, and i ts surface is located 2 0 mm below the quartz window. The surface of the sample is 3 mm above the center of the argon inlet which ha s a 5 mm diameter The center of the argon outlet is slightly higher than the argon inlet so that it is at the same height as the sample surface. The overall cell volume is 10.3 cm3. This volume include s the v olume of argon inlet and outlet but not that of the transfer tubing At 10 mm below the quartz window, the cell narrows downs to form an edge to provide support for additional thin samples or quartz windows. This can be used for the purpose of laser induced forward transfer of thin films into the ablat ion cell. Lasers Two different lasers we re employed for sample ablation. The first one was a passively Q switched diode -pumped Power -chip laser (JDS Uniphase, Santa Rosa, CA ). This laser is a solid state Nd : YAG laser that operates at 1064 nm with pulse r epetition frequency of 1 kHz. The pulse width of the laser is 500 ps with 50 J pulse energy. The other laser was a Q -switched, flashlamp -pumped Nd: YAG laser (Big Sky Laser Technologies, Bozeman, MT ). This laser has variable frequency of 1, 2, 5, 10 and 20 Hz, temporal pulse lengt h of 5 ns and maximum energy of ~ 75 mJ /pulse) The l aser probe was set on a stand perpendicularly to aim at the top window of the ablation cell. The stand was also set on an XYZ stage, allowing for minuscule movements of the laser probe in the XY direction. The lens (focal length = 55 mm) is supported on the same stand as the laser probe, and as a result, will move along with the laser beam. This way, the ablation spot c ould be moved on the sample's surface without moving the ablation cell, which is

PAGE 125

125 constrained by tubing that connects it to the discharge sou rce. The laser beam was focused 3 mm below the sample surface to minimize fractionation. Materials Commercially available samples of stainless steel, bismuth alloy, aluminum, brass and Macor were machined into small discs. Table 5 1 shows the chemical composition of these samples. For generating salt partic les, a mixture of desired salt ( for example CsI) and a buffer salt ( for example NaCl ) was ground as finely as possible and the resulting mixture was pressed into a disc and placed into the cell for ablati on Results and Discussion In order to compare the efficiency of the two lasers in laser ablation, both lasers were utilized at their maximum power on a stainless steel sample, and the iron signal at m/z = 56 was monitored. When the Power -chip laser was used, no detectable iron signal was observed using the instrument's software. However, when the detector of time of -flight mass spectrometer was directly connected to the oscilloscope, due to reduction of background noise, a very small iron peak appeared at m/z = 56. The resulting mass spectrum is depicted in Figure 5 3 .When the flash lamp -pumped Nd:YAG laser (20 Hz) was used, the iron signal was significantly stronger than that of the Powerchip laser (~15 times), and was also detectable with the instrument's software. Thus, the flash lamp -pumped Nd:Y A G was used throughout the rest of the experimentation. Figure 5 4 shows the mass spectrum resultant from ablation of stainless steel by the flash lamp -pumped Nd:YAG laser. As is shown in Figure 5 4 in addition to a large 5 6Fe signal ion s of 52Cr, 53Cr, 54Fe (over laps with 54Cr), 57Fe, 55Mn 58Ni and 60Ni were detected as well. To investigate the stability of aerosol generation by Nd:YAG laser, the stainless steel sample was ablated for ~2 4 0 seconds, and the 56Fe si gnal was monitored by time. The results shown in Figure 5 5 indicate acceptable signal stability with relative standard deviation of 1 0 %.

PAGE 126

126 Laser freq uency, glow discharge frequency and integration time for detector of the time of -flight were 10 Hz, 400 Hz, and 500 ms respectively. As was expected, the particles generated by the Nd:YAG laser provided a continuous source of aerosol for the mass spectrometer. This is due to broadening of the particle package generated by each laser pulse during aerosol transpo rt from the ablation cell to discharge. This broadening is clearly demonstrated in Figure 5 6 when laser frequency of 1 Hz and integration time of 10 ms was used. At this low laser frequency, individual signal pulses from each laser shot could be detected The width of each particle package at the discharge source was about 150 ms. This indicates that for a laser frequency of 10 Hz or higher, these packages of particle will start to overlap and provide a continuous flow of aerosol into the glow discharge. Similar behavior was also observed for b ismuth alloy. Figure 5 7 shows an experiment in which bismuth alloy was ablated for ~10 seconds. The laser frequency and glow discharge frequency were 5 Hz and 400 Hz, respectively. The integration time of detector was 5 ms in order to see the particle profil e with higher resolution. As can be seen, the individual pulses can still be resolved at the maximum of each pulse. However, the bases of each pulse (a package of particle) broaden and merge into each other. At h igher frequencies (10 and 20 Hz), no individual particle package (due to a single laser pulse) can be detected, and a continuous signal is observed. At 1 Hz, however, each laser pulse can be completely resolved. When the laser is employed at a frequency of 5 Hz, the particle packages are at the threshold of merging into each other. As a result, the signal vs. time for such frequency will present borderline characteristics between a continuous and pulsed source of aerosol (Figure 5 7 ). This shows that at la ser frequencies of 10 Hz or higher, there is no need to synchronize the

PAGE 127

127 laser and discharge pulses with respect to each other as the particle flow into the discharge is continuous and not pulsed. Figure 5 8 shows the result (Cu and Fe signals vs. time) for laser ablation of stainless steel for ~12 minutes. The ablation spot was not moved during this experiment, and interestingly, the sample was able to provide a relatively stable source of aerosol from the same spot during this experiment. At the very beginning of ablation, the signal had a slight reduction (in the first minute) and then remain ed stable for about 9 minutes (to the 10th minute). Figure 5 9 also shows the 56Fe 63Cu and 64Zn peaks for four different time zones in this experiment. Th e early red uction in both Cu and Fe signals could be due to the change in the nature of the cathode and plasma due to introduction of particles, which will reach a steady state after the first minute. After about 10 minutes, the iron signal started to decay along wit h the copper signal. However, after the laser abl ation ceases, the copper signal (resulting from the sputtering of the cathode) becomes stable. Figure 5 10 shows that the iron to copper signal ratio is very stable, and the decay of the copper and iron signals are related to each other. This means that the decay in both iron and copper signals is due to contamination of the cathode and plasma as a result of particle introduction that changes the nature of the discharge and the quality of the plasma. This is not a surprising phenomenon, as the glow discharge cathode requires regular cleaning, even without the introduction of particles. However, as a result of particle introduction, more frequent cleaning of the cathode is required. Figure 5 11 also shows the c opper signal vs. time where a sample of brass was ablated for a period of ~30 seconds, followed by a time lapse of ~20 seconds in which the particle introduction was halted, and then resumed for another ~ 10 seconds. The origin of the primary 63Cu ion signa l is the sputtering of the brass cathode. Every time the brass sample was ablated

PAGE 128

128 there was a boost in 63Cu signal, which represents the contribution of brass particles in providing copper ions. It is interesting to note that the standard deviation of the signal increases as a result of particle introduction. However, after the termination of particle flow, the copper signal returns to its previous state without any significant perturbation. Ablation of Different Materials The Nd:YAG laser was used to gene rate aerosol from different samples. Figure 5 12 shows the resulting spectrum for Macor Peaks for silicon, aluminum and magnesium were detected. However, no signal for fluorine or boron was observed. This is due to high ionization energy of fluorine and boron. A peak for potassium at m/z = 39 was expected ; however, for some unknown reason, no significant potassium peak was detected. Only at higher resolution and with the oscilloscope was a small 39K signal detected. Figure 5 13 shows the result of LA -GD MS for bismuth alloy. The peaks for bismuth ions (m/z = 209), lead ions (m/z = 206, 207, and 208), cadmium ions (m/z = 110, 111, 112, 113, 114, and 116) and tin ions (m/z = 112, 114, 116, 117, 118, 119, 120, 122 and 124) were observed. It is important to not e that m/z = 112, 114 and 116 are common to both cadmium and tin. Isotopes m/z = 106 and 108 for cadmium, and m/z = 115 for tin are not detected due to low isotopic abundance. Three samples of salt pellets with different CsI percentages ( 7.65 %, 16. 42% and 2 0.16%) in an NaCl matrix were prepared. The Cs signal vs. CsI molar percent resulted in a linear relationship, which is depicted in Figure 5 14. Relative S ignal I ntensities for Different Matrices A list of relative intensities for different elements in th ese samples can be found on Table 5 2 Bismuth alloy resulted in the highest sensitivity, due to the low heat of vaporization of lead and bismuth, while Macor resulted in a lowe r sensitivity due to high heat of vaporization of this material. The ionization energies of the elements in the bismuth alloy are also lower than those

PAGE 129

129 of Macor elements. In addition to heat of vaporization, higher atomic weights of bismuth and lead are also important factors in their higher sensitivities (Chapter 4 ). Figure 5 15 sho ws the particle size distribution for these materials. The lower sensitivities for aluminum, Macor and salt pellets c ould be explained by the higher DV values of these particles (Table 5 3) The Effect of Cathode Potential on Particle Ionization The effec t of cathode potential on the ionization of stainless steel particles is shown in Figure 5 16 and compared to the copper signal that originates from the brass cathode. As expected (Chapter 4 ), the copper signal is almost a linear function of cathode potent ial. However, interestingly, the iron signal reduces as cathode potential increased. This phenomenon can be explained by considering the small mean volume diameter of stainless steel particles (200 nm). The particles are so small that even a low cathode po tential is enough to ensure their complete vaporization. Any further increase in cathode potential will result in reducing the length of the cathode dark space. Consequently, particles will spend a shorter time in this high energy region, which can reduce the vaporization efficiency. A more important explanation for the decay of the iron signal is the increase in the electric field by increasing the cathode potential. It was previously mentioned in Chapter 3 that the electric field of the cathode dark space can significantly suppress the ion transfer out of this region. Figures 5 17 and 5 18 show the effect of cathode potential on Macor and salt particles. It can be seen that for these larger particles that are also harder to evaporate, the cathode potential behavior is similar to that of salt particles generated by an ultrasonic nebulizer -dehydrator. The particle signal increases by cathode potential at first; as the cathode potential increases further, the signal starts to decline. This difference from the stainless steel behavior can be explained by the larger diameter of Macor and salt particles (DV~ 700 nm ).

PAGE 130

130 Crater S hape in B ismuth S ample Figure 5 19 shows the crater image for the ablation of bismuth alloy obtained by optical microsco pe. The crater diamet er was measured to be 1.1mm, and the depth of the crater was estimated to be 500 m In another experiment 219 laser sho ts (at maximum pulse energy) were used to ablate another crater. The crater profile shape and dimension w ere measured by an optical pro filometer. Figure 5 20 shows the result for this experiment. The diameter of the crater was about 1000 m on top in agreement with what was measured by optical microscope (Micromaster, Fischer Scientific) However, t he diameter at the bottom of the crater was only 400 m. The crater depth was 50 m. Considering a cylindrical shape for the crater and using average of 700 m for its diameter, a crater volume of 1.92 .105 cm3 was calculated. Considering density of the bismuth alloy (9.37 g/ cm3) will result in mass removal of 0.82 g /p ulse for the bismuth alloy. Effect of Laser Pulse Energy As the laser pulse energy was increased from 1 to 10, both bismuth and lead signal increased as expected (Figure 5 21). However, the increase in signal for lead had a higher rate than that of bismuth. Figure 5 22 shows the Bi to Pb signal ratio at different laser pulse energies. This is an example of the effect of laser fluence on fractionation. Investigation of Particle Ionization M echanism In the previous chapters, we discu ssed some evidence that showed particles are ionized in the cathode dark space of the glow discharge plasma, as opposed to cathode sputtering of particles deposited on the cathode. In order to eliminate the possibility of the ionization of particles from t he cathode surface, an experiment was devised. In this experiment, the bismuth sample was ablated for ~30 s (305 shots) while the glow discharge source was off. Then the laser

PAGE 131

131 ablation was stopped and the glow discharge plasma was initiated. Figure 5 -23 shows the bismuth signal vs. time for this experiment. Only a small spike in the signal was observed at the beginning of plasma initiation. This signal is much lower than the signal observed during previous experiments in which the alloy was ablated into the plasma while plasma was on. In other words, only a small portion of the signal results from the particles that are deposited on the surface of the cathode. Its also important to note that no signal was observed when only the laser was operating without t he discharge plasma glowing. This means that no ions generated by laser plasma can reach a detector of TOF. In another experiment, a 200 m filter (porous metal disc with pore diameter of 200 m) was placed between the laser ablation cell and the plasma source while both laser and plasma were on. The presence of the filter resulted in no iron signal. This shows that no atomic vapor from the sample is transferred to the discharge, and the majority of the material is transferred in the form of particles. Laser Back Ablation of Thin Films After ablation of bismuth sample during few experiments, it was observed that a very thin film of material is f or med on the quartz window. As F igure 5 24 shows, this is a very minute amount of deposition that might not even be considered a film. A closer investigation of the window shows that parts of this deposited material are removed by the laser beam from t he center o f the window. This is shown in F igure 5 24 by zooming into the center of the quartz window where the laser beam passed through. This is an evidence for laser back induced forward transfer of material. As mentioned in the experimental section, t he ablation cell was designed in a versatile way so that it can accommodate another quartz window or a thin film for laser back ablation studies (Figure 5 25). The secondary quartz window was spiked with a few micro liters of CsI solution

PAGE 132

132 and placed on the ablation cell. The back ablation of CsI was investigated. Although there was physical evidence for the removal of some CsI from the window, no Cs signal was detected. In another experiment, a tape with a thin layer of copper (5 m) was placed on the surfa ce of the quartz window. Although the laser beam clearly removed the copper layer from the surface of the tape, no increase in copper signal was detected. It is then conceivable that the laser back ablation of aforementioned samples does not provide the ap propriate amount of aerosol for detection. In another experiment, a thin, circular stainless steel film was affixed in place of the quartz window (Figure 5 26). The laser was operated at maximum energy and frequency (20 mJ and 20 Hz), and the iron signal was monitored as a function of time. In order to distinguish between laser back ablation and regular ablation, a pure aluminum sample (more than 99 % Al) was placed below the stainless steel film in the usual sample position, and aluminum and iron signals were monitored. Figure 5 27 shows that for about the first thousand laser shots, no signal for either aluminum or iron was observed. The aluminum signal appeared after 1,002 shots, and continued until the laser beam was terminated. This shows that after 1, 002 laser shots, the stainless steel film has been penetrated and the beam reaches the aluminum. It was interesting to see that a small iron signal appeared a few shots before occurrence of aluminum signal. This iron signal started to decay immediately aft er the aluminum signal appeared. One explanation could be that right before complete penetration of the laser through the stainless steel, laser back ablation of the remaining thin layer of stainless steel has occurred. Another explanation might involve t he transfer of iron particles from the sides of the crater in the stainless steel through a small hole at the bottom of the crater. Another experiment is required to evaluate the validity of these explanations. In this experiment, the ablation must be term inated

PAGE 133

133 immediately after the appearance of iron signal, and then the stainless steel must be examined for existence of any hole at the bottom of the crater. Figure 5 28 shows the hole in the stainless steel film, and the crater on the aluminum sample. Effe ct of Laser Ablation and Particle Introduction on Other Discharge Species When the stainless steel was ablated and particles entered the glow discharge, it caused a minor decline in argon signal while copper and argon hydride signals increased. Figure 5 29 shows an increase in copper signal upon ablation of stainless steel by a single laser shot. Figure 5 30 also shows the increase in copper signal as a package of stainless steel particles produced by a laser frequency of 10 Hz enters the discharge for dura tion of ~20 seconds. The change in Ar and ArH signals is shown in Figures 5 31. The effect of particle introduction is negative and positive for Ar and ArH signals respectively. As discussed in previous chapters, the increase in copper signal is due to increase in discharge pressure due to introduction and vaporization of stainless steel particles. The bombardment of the cathode by resulting iron atoms is another reason for increase in the sputtering of the cathode, and copper signal. The decrease in argon signal can also be explained by charge transfer reactions between iron atoms and argon ions. An increase in concentration of iron atoms after introduction of the particles in the negative glow will result in the following reaction: Fe + Ar+ > Fe+ + Ar Due to higher ionization energy of argon, this reaction is thermodynamically favorable and will decrease the population of argon ions. Bismuth alloy ablation also reduces the argon signal for the same reason (Figure 5 32). However, as shown in figure 5 33, the copper signal does not change much in the case of the bismuth alloy. This can be explained by the fact that ionization energies of bismuth and lead (7.3 and 7.4 eV) are slightly lower than that of copper (7.7 eV). As a result introduction of this Bi -Pb

PAGE 134

134 alloy will reduce the population of copper ions due to ion exchange reactions. Since the sputtering increase is still expected, overall there is not much change in copper signal. The Macor particles decrease the argon ion population for a similar re ason as did the stainless steel particles (Figure 5 34). However, unlike stainless steel, when Macor particles enter the discharge, they result in a reduction of copper signal (Figure 5 35). There could be a few explanations for this different behavior. Fi rst, Macor is a non -conducting material that can disrupt the function of discharge upon its introduction by contaminating the cathode surface. In addition, Macor particles are larger than stainless steel particles (DV= 900nm compared to 200nm), and have a higher enthalpy of vaporization As a result, the vaporization of Macor particles is not complete, and less temporal increase in discharge pressure is caused by Macor when compared with stainless steel. Finally, the majority of ions generated by Macor part icles (Si, Al, and Mg) have lower masses compared to atoms generated by stainless steel (Fe, Cr, and Mn), and heavier ions accelerated in the electric field of the CDS can sputter the cathode more efficiently.

PAGE 135

135 Table 5 1. The chemical composition of sampl e material used in laser ablation in mass%. Macor Stainless steel Bismuth alloy Brass Aluminum SiO2 46% MgO 17% Al2O3 16% K2O 10% B2O3 7% F 4% Mn 2% Si 1% Cr 18 20% Ni 8 10.5% Fe 71 66.5% Bi 50% Pb 26.7% Sn 13.3% Cd 10% Cu 61.5% Zn 35.4% Pb 3.1% Al 99% Table 5 2 Relative intensities of particle ionization from ablation of different solid samples. Sample isotope mole % of isotope i l Signal ii Normalized ii Stainless steel 56 Fe 61.6 40.3 65.4 Bismuth alloy 209 Bi 42.1 72.1 171.3 Macor 28 Si 14.2 6.5 45.8 Aluminum 27 Al 99 25 25.3

PAGE 136

136 Table 5 3. Particle mean volume diameters (DV) for aerosol generated by laser ablation of different samples. A particle size analyzer (Aerosizer LD, API Inc., TSI Inc, Shor eview, MN) was used for particle size distribution measurements. Sample DV /nm Stainless steel 180 10 Brass 239 10 Bismuth alloy 332 12 Aluminum 908 22 Macor 934 35 Salt pellet 900 15

PAGE 137

137 Figure 5 1 Over all experimental set up for LA GD -MS. Only the skimmer of time -of -flight is shown. 63 sample 10 30 22 5 17 Figure 5 2 The geometrical dimensions of the ablation cell. All dimensions are in millimeters.

PAGE 138

138 0 1 2 3 4 5 52 54 56 58 60 62 64 signal ( a.u.)m / z Figure 5 3. Iron signal resulted from ablation of stainless steel using power chip Nd:YAG laser Cathode potential = 1.75 kV The peaks at m/z = 63 and 64 are the 63Cu and 64Zn from the brass cathode. 0.0 2.0 4.0 6.0 51 52 53 54 55 56 57 58 59 60 61 62 63 64 signal ( a.u.)m / z Figure 5 4 Mass spectrum resulted from ablation of stainless steel using the Q -swit ched, fl ashlamp -pumped Nd:YAG laser at its maximum pulse power Cathode potential = 1.75 kV The peaks at m/z = 63 and 64 are the 63Cu and 64Zn from the brass cathode.

PAGE 139

139 Laser on Laser off 56Fe signal (mV)Time (s)300 0 9 Figure 5 5. Monitoring the 56Fe ion signal as a function of tim e for 300 seconds. Laser ablation of stainless steel was performed for 240 seconds of the time. Cathode potential = 1.75 kV Laser frequency = 10 Hz, Integration time = 500 ms

PAGE 140

140 Laser on Laser on 56Fe signal (mV)Time (s)15 0 5.58 1.34 Figure 5 6 Monitoring the 56Fe ion signal as a function of time for ablation of stainless steel using 11 single laser shots. Cathode potential = 1.75 kV Laser frequency = 1 Hz, Integration time = 10 ms

PAGE 141

141 Bi signal (mV) Time (s) Figure 5 7 Monitoring the 209Bi ion signal as a function of tim e for laser ablation of bismuth alloy. Cathode potential = 1. 78 kV Laser frequency = 5 Hz, Integration time = 50 ms

PAGE 142

142 0 5 10 15 20 25 30 0.00 5.00 10.00 15.00 20.00 Fe56 Cu63 Time (min)Peak area ( a.u .) Laser on Laser off Figure 5 8 Monitoring the 56Fe and 63Cu ion signal s as a function of time for approximately 13 minutes o f stainless steel ablation Cathode potential = 1.75 kV Laser frequency = 10 Hz, Integration time = 500 ms Time ( min:sec) 0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 23.3 23.35 23.4 23.45 23.5 flight time / s ~ ~ Flight time ( s)56Fe Time (min)signal ( a.u.) 0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 24.45 24.5 24.55 24.6 24.65 flight time / s signal / a.u 0:15 9:30 12:20 14:40 ~ ~64Zn63Cu Figure 5 9 Iron, copper and zinc ion signals, at 4 different time domains. The laser is turned off at t = 13 min The x axis shows t he flight time instead of the m/ z.

PAGE 143

143 0 0.2 0.4 0.6 0.8 1 1.2 1.4 0.00 5.00 10.00 15.00 20.00 Fe/ Cu ion signal ratioTime (min) Figure 5 10. The signal ratio for 56Fe to 63Cu almost remains constant for the 13 minutes duration of laser ablation. 0 4 8 12 16 20 0 20 40 60 80 100 120 GD source on Laser on Laser off 56Fe/ 63Cu ion signal ratioTime (min) Figure 5 11. T he cathode signal from sputtering the brass cathode increases due to ablation of a brass sample

PAGE 144

144 24 24.5 25 25.5 26 26.5 27 27.5 28 28.5 0 1 2 3 4 5 6 7 8 Mg Mg Mg Si Alsignal ( a.u.)m/z Figure 5 12. Mass spectrum from laser ablation of a Macor sample. -10 0 10 20 30 40 50 202.0 204.0 206.0 208.0 210.0 212.0 0 10 105.0 107.0 109.0 111.0 113.0 115.0 117.0 119.0 121.0 123.0 125.0 Cd Sn signal ( a.u.)m/z Figure 5 13. Mass spectrum from laser ablation of a bismuth alloy sample.

PAGE 145

145 R2 = 0.9834 0 1 2 3 4 5 6 0 5 10 15 20 25 CsI mole % in NaCl pelletsignal ( a.u.) Figure 5 14. Cesium signal vs. mole % of cesium at different salt pellets. 0 4 8 12 16 20 0 0.5 1 1.5 2 Diameter ( m)% relative abundance of dV/dDBi SS, Brass, Macor, Al, Salt Figure 5 15. Particle size distributions for aerosols generated fr om different samples.

PAGE 146

146 0 5 10 15 20 25 30 35 0.5 1 1.5 2 2.5 3 Fe56 Cu63 Fe/Cu Cathode potential (kV)signal (mV) F igure 5 16. The effect of cathode potential on ionization of stainl ess steel particles (56Fe signal) is compared with its effect on cathodic sputtering of brass (63Cu signal).

PAGE 147

147 0 1 2 3 4 5 6 0 0.5 1 1.5 2 2.5 3 Cathode potential (kV)28Si signal (mV) Figure 5 17. The effect of cathode potential on ionization of Macor particles 0 1 2 3 4 5 6 0 0.5 1 1.5 2 2.5 3 Cathode potential (kV)Cs signal (mV) Figure 5 18. The effect of cathode potential on ionization of salt particles.

PAGE 148

148 1.1mm Figure 5 19. Crate r image of bismuth sample after ab lation by 219 laser shots at 10 Hz and maximum energy. C rater diameter = 1.1 mm

PAGE 149

149 20 depth parameters 50 200 um Figure 5 20. The crater profile shape and dimension measured by an optical profilometer. After ablation by 219 laser shots at 10 Hz and maximum laser energy.

PAGE 150

150 0 5 10 15 20 25 0 20 40 60 80 Laser pulse energy ( mJ)Bi signal ( a.u) Figure 5 21. The effect of laser pulse energy on LA -GD MS Bi ion signal of bismuth alloy particles. Laser pulse energy ( mJ)Bi / 208Pb signal 0 2 4 6 8 0 10 20 30 40 50 60 Real value Figure 5 22. The effect of laser pulse energy on 209Bi to 208Pb signal ratio.

PAGE 151

151 Figure 5 23. Memory effect investigation: The initiation of glow discharge after ablation of bismuth sample (305 shots) ended resulted in a spike in Bi signal.

PAGE 152

152 Figure 5 24. Particle deposition and laser back ablation of bismuth alloy on quartz window. Figure 5 25. Abl ation cell set up for laser back ablation of sample from the surface of the quartz window.

PAGE 153

153 Al Stainless steel thin film Figure 5 26. Ablation cell set up for laser back ablation of a thin stainless steel film located above a sample of aluminum. 80D = 0.63mm D = 0.58mm 2306 1002 Number of shots Figure 5 27. Signal vs. time for back ablation of stainless steel followed by regular ablation of aluminum sample. After about 1000 shots the stainless steel sample is completely etched and ablation of aluminum starts.

PAGE 154

154 D = 0.63mm D = 0.48mm Figure 5 28. Crater diameters for stainless steel and aluminum. Figure 5 29. Spike in copper signal due to entrance of stainless steel particles into the discharge, by a single laser shot.

PAGE 155

155 -5 0 5 10 15 20 25 30 35 0 20 40 60 80 100 -0.8 0 0.8 1.6 2.4 3.2 4 4.8 5.6 Fe56 (Analog) Cu63 (Analog) Signal (mV)Time (s) Figure 5 30. Increase in copper signal due to entrance of a 20 second ablated package of stainless steel particles into the discharge

PAGE 156

156 Signal (mV) -5 0 5 10 15 20 25 30 35 0 20 40 60 80 100 80 100 120 Fe56 (Analog) K41 (Analog) Ar (Analog) Time (s) ArH Figure 5 31. The effect of ablation of stainless steel on Ar+ and ArH+ ion signals. The right hand Y axis is for the argon ion only.

PAGE 157

157 Figure 5 32. The effect of bismuth alloy ablatio n on Ar+ ion signal.

PAGE 158

158 Figure 5 33. The effect of 4 single shots of laser ablation of bismuth alloy on 63Cu+ ion signal. No significant change in copper ion signal is observed.

PAGE 159

159 0 4 8 12 16 0 50 100 150 0 25 50 75 100 Al (Analog) K41 (Analog) Mg24 (Analog) Si28 (Analog) Ar (Analog) Time (s)Signal (mV) ArH Figure 5 34. The effect of ablation of Maco r on 40Ar+ and ArH+ ion signals. 0 2 4 0 50 100 150 Al (Analog) Cu63 (Analog) Mg24 (Analog) Si28 (Analog) Cathode potential (kV)Signal (mV) Figure 5 35. The effect of ablation of Macor on 63Cu+signal.

PAGE 160

160 CHAPTER 6 CONCLUSION AND FUTURE WORK This work showed that the low pressure ionization source of a pulsed glow discharge plas ma can provide enough energy for production of ions from particles. In other words, the Grimm source is capable of vaporization atomization and ionization of aerosols with volume mean diameters ranging from 100 nm up to 900 nm. It is believed that only the cathode dark space (CDS) of our glow discharge is capable of providing enough energy for this process. High energy species such as fast argon ions (population of about 4 x 1013),101 fast argon atoms and fast electrons in this region7 will sputter the particle while it passes through this region. The linearity of signal vs. concentration of salt even in the presence of a buffer background salt provides evidence for complete particle vaporization in the CDS. Th e fact that the signal vs. cathode potential reaches a plateau also proves this hypothesis as the particles are completely vaporized. The problem of sending a nebulized aerosol into the glow discharge is mostly due to the interface between the atmospheric pressure sample and a low pressure source (Chapters 3 and 4). The laser ablation generation of aerosols allowed for producing the particles at a low pressure similar to that of the discharge source (Chapter 5). Therefore, no restriction between the aerosol source and plasma was required. The LA GD -MS resulted in higher intensities compared to that of solution nebulization GD -MS. The introduction of particles into the discharge resulted in the change in plasma conditions. The plasma cooling always caused a decrease in 40Ar+ signal. Most particles boosted the 63Cu+ as the ions from the particles enhanced cathode sputtering. The cathode becomes contaminated if a large amount of aerosol is introduced and requires cleaning. Usually metal samples are less problem atic in comparison with non -conductive samples such as Macor and salt particles. However, if aerosols are introduced in short packages in terms of time, (1 20 seconds) the

PAGE 161

161 discharge will return to its original condition shortly after the particle introduct ion is halted. The argon signal will rise up to its previous intensity and the cathode will be cleaned by cathodic sputtering. It is important to note that our aim was the characterization of particle glow discharge interactions, and therefore not strictl y analytical. For analytical purposes, the low pressure plasma of glow discharge is no match for a powerful technique such as ICP -MS. The gas kinetic temperature of a pulsed glow discharge is about 800 K102, 103 and might reach as high as 1000 K.101 On the other hand, gas kinetic temperatures as high as 5700 K for argon ICP plasma has been reported. The limits of detection in aqueous solution were 13 ppm, 14 ppm and 5 ppm for cesium, iodine and sodium, respectively. Using the same TOF mass spectrometer, detection limit of 0.1 ppm for lead in a wine sample was reported when ICP was the source of ionization.104 For analysis of solid samples by LA GD -MS, the detection limit was 1% for lead in the bismuth alloy. Detection limits of 0.1 1 ppb for LA ICP -MS ha s been reported.84 One of the main reasons for such high limits of detection in our set up is the poor ion transfer and extraction. The majority of positive ions formed in the CDS will return to the cathode due to the high electric field in this region. Reversing the cathode anode polarity might offer a solution to this problem. Another reason for low ion signal is the poor extraction efficiency at the skimmer. The skimmer of the instrument is grounded and practically there is no sampler aperture. The only force to extract the ions from the first stage into the second stage is the pressure difference between the two stages. As a result, the majority of the ions are just pumped out with the flow of argon. The use of ion funnel s to transfer more ions into the second stage and geometrical modifications of the source could alleviate this problem and increase the ion transfer efficiency.

PAGE 162

162 Future work should continue in the following three areas. First, the geometry of the glow discharge should be modified to allow an easier access to the laser beam (in the case of laser ablation studies) and a better sampling efficiency of the ions produced Second, the study should be extended to other elements and other type of aerosols (for example, biological aerosols). Third, the electrical operating characteristics of the source should be modified, in order to allow the use of much longer voltage pulse s (of the order of milliseconds) and therefore attempt a more complete characterization of the particle composition (elemental as well as molecular). As this work showed, the geometry of the glow discharge ionization source can play an important role in pa rticle ionization efficiency. For example, in a preliminary experiment in which the cathode surface was made smaller, the signal intensity increased by a factor of ~4. As Figure A 12 in A ppendix A shows, in this configuration, the smaller cathode surface w ill intensify the ion and electron current per unit area of the cathode. Another future work consideration for modification of GD geometry would be to design a discharge configuration in which the laser beam can ablate the sample immediately next to the di scharge plasma. This will increase the aerosol transfer efficiency into the plasma. However, a shorter transfer pathway of aerosol might result in smaller packages of particles, which might require synchronization between laser and glow discharge pulses to observe particle -plasma interaction. Although our present LA GD -MS set up does not provide analysis of elements with subpercentage composition, it can still be used for sample identification. For example, some geological samples can be distinguished from each other by analysis of major elements, usually present at percentage levels in the sample. Moreover, the lower cost of a GD source is an advantage over ICP.

PAGE 163

163 Elemental analysis, even at limits of detection as high as 1%, can provide meaningful informati on in some biological applications. For instance, elemental analysis of fish scales105 for the purpose of water pollution studies and mapping of bio-mineral structures such as uroliths106 can be done by LA GD -MS. Another interesting application for LA GDMS would be in cases where sample ablation occurs at low pressures, while ICP is limited to atmospheric pressure. For example, LIBS analysis of hematite under a Mars like atmosphere has been i nvestigated.107 In order to investigate the same system by mass spectrometry one can use the low pr essure (~10 20 torr) ablation cell along with the pulsed GD source. One of the objectives in this research was to obtain molecular information as well as elemental analysis. However, the results of particle ionization did not show any evidence toward the e xistence of molecular ions, which could be due to the short duration of the ionization pulse. Using a millisecond ionization pulse (instead of microsecond) could increase the possibility for observation of molecular ions at the after pulse. It would be int eresting to study biological and organic aerosols, because they are easier to vaporize and the possibility of observing molecular ions increases.

PAGE 164

164 APPENDIX COMPLEMENTARY FIGURES y = 1.6102x + 1.0319 R2 = 1 y = 1.1458x + 0.3323 R2 = 1 0 50 100 150 200 250 300 350 0 50 100 150 200 250 Channel 2 Channel 1 Flow reading from digital flow meter ( a.u .)RSD ~ 0.3 %Calibrated flow ( scc min1) Figure A 1. Calibration of the digital flow m eter by the soap bulb method. The real value of flow in standard cubic centimeter per min is shown.

PAGE 165

165 y = 1326.9x3 3051.5x2 + 2841.1x 682.11 R2 = 0.9998 0.0 200.0 400.0 600.0 800.0 1000.0 1200.0 0.5 0.7 0.9 1.1 1.3 1.5 1ststage pressure (torr)Argon flow (mL.min1) Figure A 2. Calibration of the argon flow into the glow discharge with respect to the first stage pressure. The position of vacuum valve must remain constant (all open) for this calibration graph to be valid.

PAGE 166

166 Cs+ I+2.5 mM 5.0 mM 10.0 mM Intensity (relative) m/z Figure A 3. Signal intensities for Cs and I at three different concentrations of CsI solution are shown.

PAGE 167

167 0 1 2 3 4 5 6 7 8 -500 0 500 Modulator pulse width Repeller pulse widthRelative intensitytime ( s) -70 -60 -50 -40 -30 -20 -10 0 10 20 30 -400 -200 0 200 400 600 25 s HV pulse Repeller start Modulator start Ar ArH signal Relative intensitytime ( s) Figure A 4. Repeller and modulator ringing from the oscilloscope is shown. The HV pulse is also detected at a delay time of ~65 s before the repeller pulse (bottom). The repeller and modulator pulses are shown with highe r magnification as well (top).

PAGE 168

168 pump Nebulizer dehydrator 5% exhaust Figure A 5 The sources of particle and ion loss before the skimmer of time -of -flight are shown. 4 mM CsI 1.5 ml/min 1 6min 10 6 mol % 5 Nebulizer efficiency % 67 Flow Loss to exhaust s 60 min 1 mol Cs 1 10 02 623 1 8min 10 1 mol Cs atoms / s at the GD source 1510 2 % 5 1ststage Pumping losses 1410 1 ions / s assuming 100% ionization at the source Figure A 6. Calculation of the number of ions expected to a rrive at the detector assuming 100% ionization and ion transfer efficiency.

PAGE 169

169 20 mV signal RI V 610 4 R= 5K C/s 710 5 2 ions/s Detector gain Electron charge e ion C e 6 10 1 19 10 6 1 1 Ionization efficiency = % 10 5 2 100 10 1 10 5 25 14 7 Figure A 7. The overall efficiency of the particle vaporization atomization ionization and ion transfer to detector is calculated.

PAGE 170

170 Volume CsI 4mM ( L )Conductivity ( S.cm1) y = 0.0062x + 0.3164 R2 = 0.9999 0 0.5 1 1.5 2 2.5 0 100 200 300 400 Figure A 8. Calibration graph for conductivity vs. concentration of CsI. The volume of the 4 mM CsI solution added to a 100 ml solution is directly proportional to conductivity of the final solution.

PAGE 171

171 0 5 10 15 20 25 30 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 Relative abundanceDiameter ( m)dN/dD N Figure A 9. Result of the particle size distribution measurement for ablation of Macor Relative abundances of dN/dD is compared to N. At larger bin sizes the dN/dD is smaller than N and vice versa. 0 2 4 6 8 10 12 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 dV/dD %Diameter ( m) Figure A 10. The particle siz e distribution for Macor aerosol is shown as dV/dD. The relative intensity at larger particle diameters increases.

PAGE 172

172 Figure A 11. The effect of ablating stainless steel by the power -chip Nd: YAG laser on copper signal. Although, no iron signal is detected, the copper signal is influenced by laser ablation of the stainless steel sample.

PAGE 173

173 0 5 10 15 20 25 0 10 20 30 40 50 60 70 80 90 100 Cs signal (mV)Repeller delay time ( s) insulator cathode anode Figure A 12. Geometrical modification of cathode resulted in ion signal as high as 22 mV for Cs+ upon nebulization of a 4 mM CsI solution.

PAGE 174

174 -0.5 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 0 20 40 60 80 100 Repeller delay time ( s)Signal (mv) Figure A 13. The Cs+ signal vs. delay time is shown as well as the error bars. The error bars represent the standard deviation of the mean for 5 series of collection. The pressure of the first stag e is 1.45 torr. Refer to figure 3 9 for data related to other pressures.

PAGE 175

175 y = 1.7172x + 0.1632 R2 = 0.9963 0 2 4 6 8 10 12 14 16 18 20 0 2 4 6 8 10 CsI concentration (mM)Signal (mV) Figure A 14. The signal for Cs+ ion vs. CsI concentration show a linear dynamic range from 0.50 mM to at least 10.0 mM. The concentrations higher tha n 10.0 mM were not used to avoid sending an excess amount of salt into the system. Limit of detection of 0.1 mM was calculated based on standard deviation of background = 0.05 mV. Cathode was heated to 75C and cathode potential = 1.80 kV.

PAGE 176

176 L IST OF REF ERENCES 1. J. D. Cobine, Gaseous Conductors: Theory and Engineering Applications Dover Publications, Inc. 1958. 2. A. M. Howatson, An Introduction to Gas Discharges. 2nd Ed. Pergamon Press Ltd. 1976. 3. R. K. Marcus and J. A. C. Broek aert, Glow Discharge Plasmas in Analytical Spectroscopy John Wiley & Sons, Inc. 2003. 4. D. Fang and R. K. Marcus, Glow Discharge Spectrosc. 1993, 1766. 5. H. A. Hyman, Physical Review A: Atomic, Molecular, and Optical Physics 1979, 20, 855 859. 6. R. J. Carman, Journal of Physics D: Applied Physics 1989, 22, 55 66. 7. A. Bogaerts, Mathematical modeling of a direct current glow discharge in argon. Ph.D. dissertation, Universiteit Antwerpen, Antwerpen, Belgium, 1996. 8. D. Fang and R. K. Marcus, Spectr ochimica Acta, Part B: Atomic Spectroscopy 1990, 45B 10531074. 9. E. S. Oxley, The microsecond pulsed glow discharge: developments in time -of -flight mass spectrometry and atomic emission spectrometry. Ph.D. dissertation, University of Florida, Gainesvil le, Florida, 2002. 10. G. F. Weston, Cold Cathode Glow Discharge Tubes Iliffe. 1968. 11. J. M. Schroeer, T. N. Rhodin and R. C. Bradley, Surface Science 1972, 34, 571580. 12. R. V. Stuart and G. K. Wehner, Journal of Applied Physics 1964, 35, 18191824. 13. C. G. Bruhn, B. L. Bentz and W. W. Harrison, Analytical Chemistry 1978, 50, 373375. 14. C. G. Bruhn and W. W. Harrison, Analytical Chemistry 1978, 51, 16 21. 15. E. Oxley, C. Yang and W. W. Harrison, Journal of Analytical Atomic Spectrometry 2000, 15, 12411245. 16. P. Sigmund, Physical Review 1969, 184, 383416. 17. P. W. J. M. Boumans, Analytical Chemistry 1972, 44, 12191228. 18. N. Laegreid and G. K. Wehner, Journal of Applied Physics 1961, 32, 365369. 19. K. Wagatsuma, Glow Discharge Opti cal Emission Spectrometry 1997, 167175. 20. R. L. Smith, D. Serxner and K. R. Hess, Analytical Chemistry 1989, 61, 11031108.

PAGE 177

177 21. E. B. M. Steers and R. J. Fielding, Journal of Analytical Atomic Spectrometry 1987, 2 239244. 22. E. B. M. Steers and F. Leis, Journal of Analytical Atomic Spectrometry 1989, 4 199204. 23. A. Bogaerts and R. Gijbels, Spectrochimica Acta, Part B: Atomic Spectroscopy 1998, 53B 1 42. 24. A. Bogaerts, L. Wilken, V. Hoffmann, R. Gijbels and K. Wetzig, Spectrochimica Acta, P art B: Atomic Spectroscopy 2001, 56B 551564. 25. E. W. Eckstein, J. W. Coburn and E. Kay, International Journal of Mass Spectrometry and Ion Physics 1975, 17, 129138. 26. B. L. Bentz, C. G. Bruhn and W. W. Harrison, International Journal of Mass Spect rometry and Ion Physics 1978, 28, 409425. 27. K. R. Hess and W. W. Harrison, Analytical Chemistry 1988, 60, 691696. 28. K. C. Smyth, B. L. Bentz, C. G. Bruhn and W. W. Harrison, Journal of the American Chemical Society 1979, 101 797799. 29. W. Vieth and J. C. Huneke, Spectrochimica Acta, Part B: Atomic Spectroscopy 1991, 46B 137153. 30. W. W. Harrison and W. Hang, Journal of Analytical Atomic Spectrometry 1996, 11, 835840. 31. W. W. Harrison, W. Hang, X. Yan, K. Ingeneri and C. Schilling, Journal of Analytical Atomic Spectrometry 1997, 12, 891 896. 32. E. Oxley, C. Yang, J. Liu and W. W. Harrison, Analytical Chemistry 2003, 75, 64786484. 33. C. Yang, M. Mohill and W. W. Harrison, Journal of Analytical Atomic Spectrometry 2000, 15, 12551260. 34. H. Schuler and H. Gollnow, Zeitschrift fuer Physik 1935, 93, 611619. 35. W. Grimm, Spectrochimica Acta, Part B: Atomic Spectroscopy 1968, 23, 443 454. 36. C. M. Barshick, D. C. Duckworth, D. H. Smith and Editors, Inorganic Mass Spectrometry: Fundame ntals and Applications. [In: Pract. Spectrosc., 2000; 23] 2000. 37. S. L. Mandelstam and V. V. Nedler, Spectrochimica Acta 1961, 17, 885 894. 38. W. W. Harrison, C. Yang and E. Oxley, Glow Discharge Plasmas in Analytical Spectroscopy 2003, 7196. 39. W. A. Mattson, B. L. Bentz and W. W. Harrison, Analytical Chemistry 1976, 48, 489491.

PAGE 178

178 40. W. W. Harrison, C. M. Barshick, J. A. Klingler, P. H. Ratliff and Y. Mei, Analytical Chemistry 1990, 62, 943A 949A. 41. J. A. C. Broekaert, Glow Discharge Spectrosc. 1993, 113174. 42. A. Bengtson and L. Danielsson, Thin Solid Films 1985, 124, 231236. 43. K. Wagatsuma and S. Suzuki, Fresenius' Journal of Analytical Chemistry 1997, 358, 581 584. 44. C. Yang, K. Ingeneri, M. Mohill and W. W. Harrison, Analytical Che mistry 1999, 71, 53285334. 45. A. Bengtson, C. Yang and W. W. Harrison, Journal of Analytical Atomic Spectrometry 2000, 15, 12791284. 46. R. Payling, D. G. Jones, A. Bengtson and Editors, Glow Discharge Optical Emission Spectrometry 1997. 47. K. Wetzi g, S. Baunack, V. Hoffmann, S. Oswald and F. Prassler, Fresenius' Journal of Analytical Chemistry 1997, 358, 25 31. 48. C. Yang, K. Ingeneri and W. W. Harrison, Journal of Analytical Atomic Spectrometry 1999, 14, 693 698. 49. C. Yang, K. Ingeneri, M. Moh ill and W. W. Harrison, Journal of Analytical Atomic Spectrometry 2000, 15, 73 78. 50. A. Montaser, Inductively Coupled Plasma Mass Spectrometry Wiley -VCH, Inc. 1998. 51. Personal communications with Fred Smith, CETAC Corp, 2009. 52. M. A. Tarr, G. Zhu a nd R. F. Browner, Applied Spectroscopy 1991, 45, 1424 1432. 53. K. Willeke and P. A. Baron, Aerosol Measurement: Principles Techniques and Applications Van Nostrand Reinhold. 1993. 54. C. J. Tsai, H. M. Chein, S. T. Chang and J. Y. Kuo, Journal of Aerosol Science 1998, 29, 839853. 55. http://frontpage.okstate.edu/nanotech/Assignments/Readings/DSL/ primer_particle_sizing_static.pdf 56. R. H. Clifford, H. Tan, H. Liu and A. Montaser, Spectrochim. Acta, Part B 1993, 48B 12211235. 57. http://www.ceramic industry.com/Articles/Cover_Story/ f8997184a6ac7010VgnVCM100000f932a8c0.

PAGE 179

179 58. K. Willeke and P. A. Baron, Aerosol Measurement: Principles Techniques and Applications Van Nostrand Reinhold. 2001. 59. H. -R. Kuhn, J. Koch, R. Hergenroeder, K. Niemax, M. Kalbe rer and D. Guenther, Journal of Analytical Atomic Spectrometry 2005, 20, 894900. 60. J. H. Seinfeld and S. N. Pandis, Atmospheric Chemistry and Physics: From Air Pollution to Climate Change John Wiley & Sons, Inc. 1996. 61. P. F. Decarlo, J. G. Slowic, D. R. Worsnop, P. Davidovits and J. L. Jimenez, Aerosol Science and Technology 2004, 38, 11851205. 62. M. V. B. Krishna and R. K. Marcus, Spectrochimica Acta, Part B: Atomic Spectroscopy 2008, 63B 673684. 63. J. Castro, M. V. B. Krishna and R. K. Marc us, American Laboratory (Shelton, CT, United States) 2008, 40, 8 10. 64. T. M. Brewer, J. Castro and R. K. Marcus, Spectrochimica Acta, Part B: Atomic Spectroscopy 2006, 61B 134149. 65. T. M. Brewer, B. Fernandez and R. K. Marcus, Journal of Analytical Atomic Spectrometry 2005, 20, 924 931. 66. T. E. Gibeau and R. K. Marcus, Analytical Chemistry 2000, 72, 38333840. 67. C. M. Strange and R. K. Marcus, Spectrochimica Acta, Part B: Atomic Spectroscopy 1991, 46B 517526. 68. W. Hang, C. Baker, B. W. Sm ith, J. D. Winefordner and W. W. Harrison, Journal of Analytical Atomic Spectrometry 1997, 12, 143149. 69. A. Sola -Vazquez, A. Martin, J. M. Costa -Fernandez, R. Pereiro and A. Sanz -Medel, Analytical Chemistry (Washington, DC, United States) 2009, 81, 25 912599. 70. S. Wang and M. V. Johnston, International Journal of Mass Spectrometry 2006, 258, 50 57. 71. J. T. Jayne, D. C. Leard, X. Zhang, P. Davidovits, K. A. Smith, C. E. Kolb and D. R. Worsnop, Aerosol Science and Technology 2000, 33, 49 70. 72. A. C. Aiken, P. F. DeCarlo, J. H. Kroll, D. R. Worsnop, J. A. Huffman, K. S. Docherty, I. M. Ulbrich, C. Mohr, J. R. Kimmel, D. Sueper, Y. Sun, Q. Zhang, A. Trimborn, M. Northway, P. J. Ziemann, M. R. Canagaratna, T. B. Onasch, M. R. Alfarra, A. S. H. Prevot J. Dommen, J. Duplissy, A. Metzger, U. Baltensperger and J. L. Jimenez, Environmental Science & Technology 2008, 42, 44784485. 73. S. J. Ray and G. M. Hieftje, Analytica Chimica Acta 2001, 445, 35 45.

PAGE 180

180 74. J. P. Guzowski, Jr. and G. M. Hieftje, Journal of Analytical Atomic Spectrometry 2000, 15, 27 36. 75. D. Fliegel, R. Waddell, V. Majidi, D. Guenther and C. L. Lewis, Analytical Chemistry 2005, 77, 18471852. 76. C. L. Lewis, M. A. Moser, D. E. Dale, Jr., W. Hang, C. Hassell, F. L. King and V. Majidi Analytical Chemistry 2003, 75, 19831996. 77. C. L. Lewis, M. A. Moser, W. Hang, D. E. Dale, Jr., D. C. Hassell and V. Majidi, Journal of Analytical Atomic Spectrometry 2003, 18, 629636. 78. V. Majidi, M. Moser, C. Lewis, W. Hang and F. L. King, Journal of Analytical Atomic Spectrometry 2000, 15, 19 25. 79. H. A. Wilson, Physical Review 1916, 8 227237. 80. G. O. Nelson, Controlled Test Atmospheres Ann Arbor Science, Ann Arbor, Michigan. 1980. 81. R. C. Weast, CRC Handbook of chemistry and physics Boca Raton, Fl. 1986. 82. I. N. Levine, Physical Chemistry. 2nd Ed, McGraw Hill. 1983. 83. H. Steckel, F. Eskandar and K. Witthohn, Journal of Aerosol Medicine 2003, 16, 417432. 84. J. S. Becker and H. J. Dietze, International Journal of Mass Spectromet ry 2003, 228, 127 150. 85. S. F. Durrant and N. I. Ward, Journal of Analytical Atomic Spectrometry 2005, 20, 821829. 86. C. C. Garcia, M. Waelle, H. Lindner, J. Koch, K. Niemax and D. Guenther, Spectrochimica Acta, Part B: Atomic Spectroscopy 2008, 63B 271276. 87. M. Tariq and D. Gunther, Spectrochimica Acta, Part B: Atomic Spectroscopy 2009, In Press, Accepted Manuscript. 88. T. M. Naeem, H. Matsuta and K. Wagatsuma, Analytical and Bioanalytical Chemistry 2004, 379 115119. 89. T. M. Naeem, H. Mat suta and K. Wagatsuma, Analytical Sciences 2004, 20 17171720. 90. C. Lewis, S. K. Doorn, D. M. Wayne, F. L. King and V. Majidi, Applied Spectroscopy 2000, 54, 12361244. 91. K. R. Hess and W. W. Harrison, Oxford Series on Optical Sciences 1990, 1 205222.

PAGE 181

181 92. J. T. Shelley, S. J. Ray and G. M. Hieftje, Analytical Chemistry (Washington, DC, United States) 2008, 80, 83088313. 93. C. C. Garcia, H. Lindner and K. Niemax, Journal of Analytical Atomic Spectrometry 2009, 24, 14 26. 94. D. Guenther and B. Hattendorf, TrAC, Trends in Analytical Chemistry 2005, 24, 255265. 95. J. Koch and D. Guenther, Analytical and Bioanalytical Chemistry 2007, 387, 149153. 96. B. Fernandez, F. Claverie, C. Pecheyran and O. F. X. Donard, TrAC, Trends in Analytical Chemis try 2007, 26, 951966. 97. K. Niemax, Fresenius J Anal Chem 2001, 370, 332340. 98. C. C. Garcia, H. Lindner and K. Niemax, Spectrochimica Acta, Part B: Atomic Spectroscopy 2007, 62B 13 19. 99. I. Horn and D. Gunther, Applied Surface Science 2003, 207, 144 157. 100. S. M. Eggins, L. P. J. Kinsley and J. M. G. Shelley, Applied Surface Science 1998, 127129, 278286. 101. A. Bogaerts and R. Gijbels, Journal of Analytical Atomic Spectrometry 2000, 15, 895905. 102. A. Bogaerts, A. Okhrimovskyy and R. Gi jbels, Journal of Analytical Atomic Spectrometry 2002, 17, 10761082. 103. G. Gamez, A. Bogaerts and G. M. Hieftje, Journal of Analytical Atomic Spectrometry 2006, 21, 350 359. 104. M. Balcerzak, Analytical Sciences 2003, 19, 979 981. 105. M. Hola, K. N ovotny, J. Kaiser, D.W. Hahn, P.K. Diwakar, J. Kalvoda, Hana Novakova and V. Kanicky, Elemental distribution in fish scales using LIBS and LA ICP MS methods, Poster presentation, North America Laser Induced Breakdown Spectroscopy Symposium, New Orleans, LA United States, July 2009. 106. J. Kaiser, M. Hola, K. Novotny, D. W. Hahn, P. K. Diwakar, R. Malina, M. Galiova, V. Kanicky and M. Liska, LIBS and LA ICP -MS complementary study of uroliths, Poster presentation, North America Laser Induced Breakdown Spect roscopy Symposium, New Orleans, LA, United States, July 2009. 107. A. Bauer, J. Black, M. Schmidt and A. J. R. Bauer, LIBS Analysis of Hematite Under a Mars -like Atmosphere, Poster presentation, North America Laser Induced Breakdown Spectroscopy Symposium, New Orleans, LA, United States, July 2009.

PAGE 182

182 BIOGRAPHICAL SKETCH Farzad was born in the city of Mashhad, in the province of Khorasan, Iran. He finished high school in 1996, after winning the National gold medal in C hemistry Olympiad in 1995. He got h is B .Sc. in chemistry from the University of Tehran -Iran in 1999, and went on to mentor the Iranian team for the 32nd International C hemistry Olympiad in Denmark in July 2000. It was there that he received a visitor visa from the American embassy in Copenhagen. In August 2000, Farzad came to the United States and spent six months in Berkeley, California, where he participated in English classes in Berkeley adult school and took the GRE and TOEFL exams. From January 2001 to August 2003, Farzad worked toward his m asters degree in Professor Jason Telfords research group at the University of Iowa, where his research focused on organic synthesis of calixarene and cyclodextrine derivatives. Farzad began his Ph.D. studies at the University of Florida in 2003. He sta rted his research as a graduate student under the supervision of Professor Nicolo Omenetto in the fall of 2004. His research focuses on fundamental aspects of pulsed glow discharge mass spectrometry and investigation of particle ionization in the pulsed di scharge, using time -of -flight mass spectrometry. He also worked for a short period of time in 2003, in Dr. Martins laboratory, on making single conical nano pores on polycarbonate membranes. During his time at the University of Florida, Farzad was a teach ing assistant for multiple courses and labs, including general, inorganic, organic, physical and analytical chemistry. Farzad defended his Ph.D. dissertation on July 24, 2009 and obtained his Ph.D. in analytical chemistry in August 2009.