Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Development and characterization of an electrothermal vaporization parallel plate capacitively coupled… Rahman, Md. Mahbubur 2001

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_2001-611620.pdf [ 10.02MB ]
Metadata
JSON: 831-1.0059599.json
JSON-LD: 831-1.0059599-ld.json
RDF/XML (Pretty): 831-1.0059599-rdf.xml
RDF/JSON: 831-1.0059599-rdf.json
Turtle: 831-1.0059599-turtle.txt
N-Triples: 831-1.0059599-rdf-ntriples.txt
Original Record: 831-1.0059599-source.json
Full Text
831-1.0059599-fulltext.txt
Citation
831-1.0059599.ris

Full Text

DEVELOPMENT AND CHARACTERIZATION OF AN ELECTROTHERMAL VAPORIZATION PARALLEL PLATE CAPACITIVELY COUPLED PLASMA By MD. MAHBUBUR RAHMAN M. Sc . University of British Columbia, 1994 A THESIS SUBMITTED IN PARTIAL FULFILLMENT O F THE REQUIREMENTS FOR THE D E G R E E O F DOCTOR O F PHILOSOPHY in THE FACULTY OF G R A D U A T E STUDIES D E P A R T M E N T O F CHEMISTRY We accept this thesis as conforming to the required standard THE UNIVERSITY O F BRITISH COLUMBIA April, 2001 © Md. Mahbubur Rahman, 2001 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department The University of British Columbia Vancouver, Canada Date DE-6 (2/88) 11 Abstract The parallel plate capacitively coupled plasma (PP-CCP) is a fairly new and unique source for spectrochemical analysis. In this source, a radio frequency (rf) plasma is initiated and sustained in a quartz discharge tube placed between a pair of water-cooled conducting electrodes to which rf power is applied. The geometry of the discharge torch is flexible and allows the introduction of both liquid (using electrothermal vaporizer, ETV) and gaseous samples into the plasma. The primary objective of this work was to characterize the fundamental processes and properties of this C C P discharge using spectroscopic techniques. Spatially resolved emission profiles from plasma background species (e.g. He, N 2 + , and OH) reveal that the analyte emission or emission sensitive zone is, as a result of the transverse power coupling geometry, situated adjacent to the discharge wall. The spatial distributions of plasma species remain unchanged, but the signal intensity from all three species are altered with variations of plasma power and gas flow rate. The difference in measured electronic excitation (T e x c ) temperatures' suggests the absence of local thermodynamic equilibrium (LTE) conditions in P P -C C P . T e x c was calculated from the slope of the Boltzmann plot using Fe and He as the thermometric species and Pb excitation temperature was calculated using the two line method. Over a power range from 100-250 W, excitation temperatures are 3255-3900 K for He, 3540-4500 K for Pb, and 4300-4890 K for Fe. The rotational temperature (T^ ) was also measured using both O H and N * molecular spectra with I l l values occurring in the 828-911 K and 845-956 K range over a power range of 75-275 W, respectively. Results obtained from matrix interference studies show that the presence of either NaCI or N a N 0 3 > as a concomitant in silver analyte, causes an interference effect by both enhancing, and decreasing, the emission intensity dependent upon the amount of added easily ionizable element (EIE). The maximum signal interfe -rence, signal enhancement, was observed at 125 W plasma forward power. The degree of enhancement from an EIE decreases with increasing rf forward power switching to a signal depression at a power of 250 W. The degree of ionization for Mg and C d vaporized into this source has also been studied. Using electrothermal vaporization and a C C P operating at 200 W, the degree of ionization is 83% and 48% for Mg and Cd, respectively. An increase in applied power increases the degree of ionization. Changes in plasma gas flow rate, up to 1.0 L min 1 , also slightly changes the degree of ionization, likely due to mass transport effects and minor changes in the discharge conditions. Additions of EIE (Na as NaN0 3 ) up to 10 times (w/w) of analyte increases the degree of ionization somewhat. All these results indicate the potential of the P P - C C P as an ion source for mass spectrometry. The absolute detection limit for Pb and Ag, for a 4 cm long electrode, is found to be 0.33 ng and 24 pg. The signal - to - noise ratio intensifies with increasing rf power and gas flow rate, reaching its maximum value at 250 W rf power and 0.2 L m in 1 gas flow rate for Pb analyte. The optimum plasma operating conditions for ultimate signal - to - noise ratio is, however, analyte dependent. The precision, another useful tool in measuring the power of an analytical method, is approximately iv 4-10% at a concentration of 100 times the limit of detection. The effect of electrode length (hence, the plasma volume) on some fundamental and analytical characteristics has also been studied and the results show an improvement in detection limit with increasing electrode length, reaching as low as 0.96 pg for silver and 0.34 pg for magnesium for a 6 cm long electrode. Electrode length also influences the analyte ionization in this source. With changing electrode length, two opposing factors, namely "residence time" and "power density", become active. At a plasma power of 250 W and 1.0 L min1 gas flow rate, the residence time emerged as a dominant factor when the electrode measured up to 5 cm long; with a longer electrode, the power density outweighed the residence time factor. The degree of Mg ionization reaches up to its maximum of 87% for the 5 cm long electrode. V Table of Contents Abstract ii Table of Contents v List of Tables IX List of Figures IX List of Publications Arising from this Study XVII List of Conference Presentations Arising from this Study x v i i * List of Abbreviations XIX Acknowledgements XX Chapter 1 Introduction 1 1.1 Parallel-Plate Capacitively Coupled Plasma (PP-CCP) 2 1.2 Thesis Objective 3 1.3 Evolution of the P P - C C P 4 1.3.1 Plasma 4 1.3.2 Plasma as Optical Emission Sources 6 1.3.3 Plasma Formation 7 1.3.4 Electrothermal Vaporizer (ETV) 10 1.3.5 The Discharge in P P - C C P 12 1.4 A Brief Chronicle of C C P 13 1.5 Plasma Diagnostics 18 1.5.1 Spatial Distribution of Plasma Species 19 1.5.2 Plasma Temperature 21 1.6 Analyte Excitation and Ionization 24 1.7 Thesis Epitome 26 1.8 References 28 vi Chapter 2 Experimental 34 2.1 Instrumentation ,34 2.1.1 The Plasma Source 36 2.1.2 Electrothermal Vaporizer 38 2.1.3 Optical Configuration 39 2.1.4 Spectral Dispersion and Detection 41 2.1.5 Specially Resolved Intensity Measurements 42 2.1.6 ETV Temperature Measurements 43 2.2 Data Acquisition and Processing 43 2.3 Experimental Procedure 44 2.4 References 45 Chapter 3 Temperatures and Some Analytical Figures of Merit 46 3.1 Introduction 46 3.2 Experimental 47 3.2.1 Instrumentation 47 3.2.2 Procedure 48 3.2.2.1 Helium Excitation Temperature 48 3.2.2.2 Pb and Fe Excitation Temperatures 50 3.2.2.3 Rotational Temperature 53 3.3 Results and Discussion 56 3.3.1 Excitation Temperature 56 3.3.2 Rotational Temperature 60 3.3.3 Analytical Figures of Merit 61 3.4 Conclusions 67 3.5 References 68 vii Chapter 4 Spatial Distribution of Plasma Background Species 71 4.1 Introduction 71 4.2 Experimental 71 4.2.1 Instrumentation 71 4.2.2 Procedure 72 4.3 Results and Discussion 74 4.3.1 Spatial Distribution 74 4.3.2 Spatially Resolved Rotational Temperature 81 4.4 Conclusions 85 4.5 References 86 Chapter 5 Effects From Easily lonizable Elements on Silver Analyte 87 5.1 Introduction 87 5.2 Experimental 91 5.2.1 Instrumentation 91 5.2.2 Reagents 91 5.2.3 Procedure 92 5.3 Results and Discussion 92 5.4 Conclusions 112 5.5 References 114 Chapter 6 Analyte ionization in the PP-CCP Source 116 6.1 Introduction 116 6.2 Experimental 117 6.2.1 Background 117 6.2.2 Instrumentation 119 6.2.3 Reagents 120 6.2.4 Procedure 120 6.3 Results and Discussion 121 viii 6.3.1 Temporal Response 123 6.3.2 Effect of Plasma Power 129 6.3.3 Gas Flow Effect 136 6.3.4 Effect of Easily lonizable Element (EIE) 143 6.4 Conclusions 149 6.5 References 150 Chapter 7 The Effect of Electrode Length on Fundamental and Some Analytical Characteristics 153 7.1 Introduction 153 7.2 Experimental 154 7.2.1 Instrumentation 154 7.2.2 Reagents 155 7.2.3 Procedure 155 7.3 Results and Discussion 156 7.3.1 Different Electrode Lengths and Plasma Power 156 7.3.2 Different Electrode Lengths and Gas Flow 164 7.3.3 Analytical Figures of Merit 174 7.3.4 Rotational Temperature and Degree of Ionization 178 7.4 Conclusions 185 7.5 References 187 Chapter 8 Conclusions 188 8.1 Summary 188 8.2 Suggestions for Future Research 192 ix List of Tables Table Page 1.1 Comparison between MIP, ICP, and CCP 7 3.1 Spectral data for He I lines 50 3.2 Spectral data for Fe I lines 52 3.3 Spectral data for Pb I lines 59 6.1 Physical characteristics of atomic and ionic lines 122 List of Figures Figure Page 1.1 Schematic diagram of the plasma source 3 2.1 Block diagram of the experimental setup 35 2.2 Schematic diagram of the experimental system 37 2.3 Two different designs of the tantalum strip platforms 38 2.4 Schematic diagram of the electrothermal (made from graphite) vaporization system employed in PP-CCP source 40 3.1 Block diagram of experimental system 49 3.2 Fe emission spectrum in PP-CCP source at a 250 W plasma forward power and 0.1 L min1 gas flow rate 51 3.3 N 2 + emission spectrum in PP-CCP source at a 250 W rf power and 0.2 L min1 gas flow rate 54 X 3.4 OH spectrum in PP-CCP source for a 100 W plasma with a gas flow rate of 0.2 L min1 55 3.5 Boltzmann plot for (A) He I and (B) Fe I line for a 200 W plasma with a gas flow rate of 0.2 L min1. Correlation coefficient, r(A) = 0.95 and r(B) = 0.99 56 3.6 Influence of plasma forward power on excitation temperatures 57 3.7 Effect of the plasma forward power on both N 2 + and OH rotational temperatures with a fixed gas flow rate of 0.2 L min1 60 3.8 Effect of gas flow rate on (A) Pb analyte and (B) background signal for a 250 W plasma for a deposition of 1.0 ng of lead 62 3.9 Effect of gas flow rate on the (A) silver analyte and (B) background , signal for a 250 W plasma for a deposition of 1.0 ng of silver 63 3.10 Effect of rf power on atomic emission behavior of Pb analyte: (A) emission signal and (B) Signal-to-Background and Signal - to -Noise Ratios at 0.2 L min"1 gas flow rate for a deposition of 1.0 ng of Pb 65 3.11 Effect of rf forward power on atomic emission behavior of Ag analyte : (A) emission signal and (B) Signal-to-Background & Signal - to -Noise Ratios at 0.2 L min1 gas flow rate for a deposition of 1.0 ng of Ag 66 4.1 The spatial distribution of plasma species in PP-CCP source for a plasma of 150 W forward power and at 0.5 L min1 gas flow rate 75 4.2 The spatial distribution of (A) He I (388.87 nm), (B) OH (band head at 310 nm), and (C) N 2 + (band head at 391.4 nm) at 0.1 L min1 gas flow rate and at different applied rf power ranging from 75 to 200 W 77 xi 4.3 Effect of applied rf power on the emission intensity of He I, OH, and N 2 + molecule measured at 1.75 mm off the center of the discharge 78 4.4 Effect of different gas flow rates on the spatial distribution of (A) He atomic, (B) N 2 +, and (C) OH molecular emission in a150 W plasma 80 4.5 Spatially resolved (A) N 2 + and (B) OH rotational temperatures profile at different plasma forward powers and for a plasma gas flow rate of 0.1 L min1. Error bars represent ± 1 standard deviation (5 samples) 82 4.6 Spatially resolved (A) N 2 + and (B) OH rotational temperatures profile for a 150 W applied plasma power and a different plasma gas flow rates. Error bars represent ± 1 standard deviation (5 samples) 84 5.1 Effect of 150 ng of Na from NaCI on the temporal response of 1.0 ng of Ag at 150 W plasma forward power and 0.1 L min1 gas flow rate. The temperature profile on the electrothermal vaporizer is also shown 93 5.2 Effect of 150 ng of Na from NaN0 3 on the temporal response of 1.0 ng of Ag for a 150 W plasma forward power and 0.1 L min1 gas flow rate 96 5.3 Influence of added amounts of EIE (Na as NaCI, as an example) on He atomic emission signal for a plasma forward power of 150 W and gas flow rate of 0.1 L min1 in a PP-CCP source 98 5.4 Temporal response of Ag atomic emission from a deposition of 1.0 ng of Ag with different amounts of Na (as NaCI) for a 150 W plasma power and 0.1 L min1 gas flow rate 100 5.5 Temporal response of silver atomic emission for a deposition of 1.0 ng of Ag with different amounts of Na (as NaN03) for a 150 W plasma power and 0.1 L min1 gas flow rate 101 xii 5.6 Effect of varying amounts of EIE (Na from NaCI) on Ag emission signal (from 1.0 ng sample introduction) for a 150 W plasma forward power and 0.1 L min1 gas flow rate 103 5.7 Effect of varying amounts of EIE (Na from NaN03) on Ag emission signal (from 1.0 ng sample introduction) in a 150 W plasma and 0.1 L min1 gas flow rate 104 5.8 Effect of varying amounts of EIE (Na from NaCI) on Ag emission signal (from 1.0 ng sample introduction) at 150 W plasma forward power and 0.1 L min1 gas flow rate 105 5.9 Effect of varying amounts of EIE (Na from NaN03) on Ag emission signal (from 1.0 ng sample introduction) for a 150 W plasma forward power and 0.1 L min1 gas flow rate 106 5.10 Effect of different plasma applied powers on the temporal emission response of 1.0 ng silver: (A) with (150 ng) and (B) without the presence of matrix (Na as NaCI). The gas flow rate is 0.1 L min"1 108 5.11 Effect of different applied rf powers on the temporal emission response of 1.0 ng silver: (A) with (150 ng) and (B) without the presence of matrix (Na as NaN03). The gas flow rate is 0.1 L min1 109 5.12 Effect of different applied rf powers on analyte signal enhancement from a deposition of 1.0 ng silver with (150 ng) and without sodium matrix (Na as NaCI). The gas flow rate is 0.1 L min1 110 5.13 Effect of different applied rf powers on analyte signal enhancement from a deposition of 1.0 ng silver with (150 ng) and without sodium matrix (Na as NaN03). The gas flow rate is 0.1 L min1 111 xiii 6.1 Temporal response of Mg atomic and ionic signals, from a deposition of 5.0 ng of standard Mg solution, at 250 W plasma forward power and 1.0 L min 1 gas flow rate 124 6.2: Temporal response of Mgl and Mgll and calculated ratio of Mgll / Mgl as a function of time for a deposition of 5.0 ng of Mg in a 250 W plasma with a gas flow rate of 1.0 L min 1 125 6.3 Temporal response of Cd atomic and ionic signals, from a deposition of 5.0 ng of standard Cd solution, for a 250 W plasma forward power and 1.0 L min"1 gas flow rate 127 6.4 Degree of ionization (for Mg and Cd) as a function of time for a 250 W plasma forward power with a gas flow rate of 1.0 L min 1 128 6.5: Response of magnesium atomic and ionic emission signals (A: using peak height, B: using peak area), from a deposition of 5.0 ng of Mg, as a function of applied rf power and 1.0 L m in 1 gas flow rate 130 6.6: Effect of plasma forward power on magnesium ion : atom ratio and the degree of ionization measured from a deposition of 5.0 ng of Mg with a gas flow rate of 1.0 L m in 1 131 6.7: Integrated atomic and ionic emission signals from a deposition of 5.0 ng of Cd as a function of applied plasma forward power with a gas flow rate of 1.0 L m in 1 132 6.8: Effect of plasma forward power on magnesium ion : atom ratio and the degree of ionization measured from a deposition of 5.0 ng of Cd with a gas flow rate of 1.0 L min' 1 134 6.9 Influence of plasma (also the carrier) gas flow rate on C d (A) atomic and (B) ionic transient signals in a 250 W plasma. The signals were generated from the deposition of 5.0 ng of Cd standard sample 137 xiv 6.10 Influence of plasma (also the carrier) gas flow rate on Mg (A) atomic and (B) ionic transient signals in a 250 W plasma. The signals were generated from the deposition of 5.0 ng of Mg standard sample 139 6.11 Influence of gas flow rate on integrated Cd atomic and ionic signals, (A) using peak height and (B) using peak area, in a 250 W plasma discharge. The amount of deposited cadmium sample was 5.0 ng 140 6.12 Influence of plasma gas flow rate on Mg atomic and ionic signals, (A: using peak height and B: using peak area), in a 250 W plasma discharge. The amount of deposited magnesium sample was 5.0 ng 142 6.13 Influence of gas flow rate on cadmium and magnesium degree of ionization for a plasma operating at 250 W forward power 143 6.14 Mg II - Mg I ratio as a function of added Na mass for a 250 W plasma forward power and 0.5 L min 1 gas flow rate 145 6.15 Effect of Na amount on measured degree of ionization for magnesium analyte at 250 W plasma forward power and 0.5 L min 1 gas flow rate 146 6.16 Effects of Na amount on (A) Cd atomic and ionic emission and, hence on (B) the measured degree of ionization for a 250 W plasma and 0.5 L m in 1 gas flow rate 148 7.1 Influence of plasma forward power on (A) He I and (B) total N 2 + emission intensity for different electrode lengths at a constant gas flow rate of 1.0 L min' 1 157 7.2 The effect of rf forward power on (A) silver, (B) magnesium atomic and (C) magnesium ionic emission. The plasma gas flow rates are 0.2 and 1.0 L min 1 for a deposition of 3.0 ng Ag and 5.0 ng Mg, respectively 159 7.3 The effect of electrode length on (A) silver, (B) magnesium atomic and (C) magnesium ionic emission. The plasma gas flow rates are XV 0.2 and 1.0 L min1 for a deposition of 3.0 ng Ag and 5.0 ng Mg, respectively 162 7.4 The effect of plasma power on signal - to - noise ratio (SNR) from (A) silver and (B) magnesium analyte for different electrode lengths. The plasma gas flow rates are 0.2 and 1.0 L min1 for Ag and Mg, respectively 163 7.5 The effect of plasma power on signal - to - background ratio (SBR) from (A) Ag and (B) Mg analyte for different electrode lengths. The gas flow rates are 0.2 and 1.0 L min1 for Ag and Mg, respectively 165 7.6 The plasma gas flow rate and the emission intensity from silver analyte, (A) using peak height and (B) using peak area, for different electrode lengths. The plasma forward power is 250 W 167 7.7 The plasma gas flow rate and the emission intensity from magnesium analyte, using (A) peak height and (B) using peak area, for different electrode lengths. The plasma forward power is 250 W 168 7.8 The plasma gas flow rate and the emission intensity from He I for different electrode lengths, with a plasma forward power of 250 W 169 7.9 The effect of electrode length on (A) silver and (B) magnesium analyte for a deposition of 2.5 and 5.0 ng of Ag and Mg sample, respectively. The applied plasma power is 250 W with different gas flow rates 171 7.10 The effect of gas flow rate on signal-to-noise ratio (SNR) using (A) silver and (B) Mg as analyte for different electrode lengths. The applied plasma forward power is 250 W 173 7.11 The effect of electrode length on (A) signal-to-noise and (B) signal-to-background ratios for a250 W plasma with 0.2 and 0.75 L min1 (for Ag) and 1.0 L min'1 (for Mg) gas flow rates 176 xvi 7.12 The effect of electrode lengths on the analyte (for both silver and magnesium) detection limits for an optimized plasma operating conditions 177 7.13 The effect of (A) rf power (at a constant gas flow rate of 0.2 L min1) and (B) gas flow rate (at constant applied rf power of 250 W) on measured N 2 + rotational temperature for different electrode lengths 179 7.14 The effect of electrode length on calculated N 2 + rotational tempe -rature at a constant applied rf power of 250 W and two different gas flow rates of 0.2 and 1.0 L min1 180 7.15 The effect of electrode length on the degree of ionization (using Mg as a typical analyte) for different plasma forward powers and at a fixed gas flow rate of 1.0 L min1 182 7.16 The effect of plasma gas flow rate on magnesium ionization using different electrode lengths for a 250 W plasma 183 7.17 The effect of electrode length on the degree of ionization (using Mg as analyte) for a 250 W plasma at different gas flow rates 185 XVII List of Publications Arising from this Study 1. M. M. Rahman and M. W. Blades, Effect of electrode length on some fundamental properties of plasma and analytical figures of merit in a parallel plate capacitively coupled plasma (PP-CCP) , in preparation 2. M. M. Rahman and M. W. Blades, Ionization of Mg and Cd in an atmospheric pressure parallel plate capacitively coupled plasma, J . of Anal. Atom. Spectrom., 15, 2000, p.1313-1319 3. M. M. Rahman and M. W. Blades, Effect from easily ionized elements on Ag analyte in Atmospheric Pressure Parallel Plate Capacitively Coupled Plasma (PP-CCP) , Spectrochim. Acta, 55B, 2000, p.327-338 4. M. M. Rahman and M. W. Blades, Atmospheric Pressure, Radio Frequency, Parallel Plate Capacitively Coupled Plasma- Excitation Temperatures and Analytical Figures of Merit. Spectrochim. Acta, 52B, 1998, p.1983-1993 List of Conference Presentations Arising from this Study 1. Michael Blades, Adam Bass, M. M. Rahman, and Conard Chevalier, Approaches to forming micro-plasmas using capacitive coupling, will be presented in "2001 European Winter Conference on Plasma Spectrochemistry" at Lillehammer, Norway, Feb. 04-08, 2001 2. Michael Blades, Adam Bass, M. M. Rahman, and S. Lu, Miniature capacitively coupled plasmas, . Presented in " The 27 t h . Annual conference of the federation of analytical chemistry and spectroscopy societies (FACSS)" held at Nashville, TN, USA, Sep. 24-28, 2000 3. M. W. Blades and M. M. Rahman, Progress in the development of C C P for spectrochemical analysis. Presented in "2000 Winter Conference on Plasma Spectrochemistry" held at Florida, USA, January 10- 15, 2000 XVIU 4. M. M. Rahman and M. w. Blades, Degree of ionization in a parallel-plate CCP operating at atmospheric pressure. Presented in "2000 Winter Conference on Plasma Spectrochemistry" held at Florida, USA, January 10-15, 2000 5. M. M. Rahman and M. W. Blades, Effect of electrode length on fundamental and analytical characteristics of atmospheric pressure CCP. Presented in "The 261h Annual conference of the FACSS" held at Vancouver, Canada, Oct. 24-29, 1999 6. M. W. Blades, C. Chaqui, M. M. Rahman, and S. Lu, Temporal emission characteristics of CCPs using a Gated-IPDA.. Presented in" The 25th. Annual conference of the FACSS" held at Austin, Texas, USA, October 11-15,1998 7. M. M. Rahman, C. Chaqui, and M. W. Blades, Atomic Spectrometry using atmospheric pressure capacitively coupled plasma. Presented in "81st. Canadian society for chemistry conference and exhibition" held at Whistler, BC, Canada May 31 - June 04,1998 8. M. W. Blades and M. M. Rahman, Fundamental and Analytical Characteristics of an Atmospheric Pressure capacitively Coupled Plasma Discharge. Presented in "1998 Winter Conference on Plasma Spectrochemistry", held at Scottsdale, Arizona, USA, from Jan. 05 -10,1998 9. M. M. Rahman and M. W. Blades, Effect of Electrode Area on Fundamental and Analytical Characterestics of Atmospheric Pressure CCP. Presented in "The 24th Annual Conference of the FACSS" held at Providence, Rhode Island, USA, Oct. 26-30, 1997 10. M. M. Rahman and M. W. Blades, Characterization of a RF-PP-CCP Operating at Atm. Pressure for AES. Presented in "The 23rd Annual Conference of the FACSS" held at Kansas City, USA, Sep. 29-Oct.4,1996 11. M. W. Blades and M. M. Rahman Trace Element Analysis Using RF capacitively Coupled Discharges Couple with Electrothermal Sample Introduction Presented in "1996 Winter Conference on Plasma Spectrochemistry", held at, Florida, USA, from Jan. 08 -13, 1996 xix List of Abbreviations AAS Atomic Absorption Spectrometry ac alternating current ADC Analog - to - Digital Converter AES Atomic Emission Spectrometry CCP Capacitively Coupled Plasma CMP Capacitively Coupled Microwave Plasma CTE Complete Thermal Equilibrium dc direct current DCP Direct Current Plasma EIE Easily lonizable Element ETV Electrothermal Vaporizer eV Electron Volt FAPES Furnace Atomization Plasma Excitation Spectrometry GC Gas Chromatography GD Glow Discharge Hz Hertz ICP Inductively Coupled Plasma i. d. inner diameter IPDA Intensified Photodiode Array LTE Local Thermal Equilibrium MIP Microwave Induced Plasma ng nanogram o. d. outer diameter OES Optical Emission Spectrometry PP-CCP Parallel Plate Capacitively Coupled Plasma pg picogram PMT Photomultiplier Tube rf radio frequency RSD Relative Standard Deviation SBR Signal - to - Background Ratio SNR Signal - to- Noise ratio W Watt XX Acknowledgements This work is dedicated to my wife Mithu and my daughters Aurnee and Auddri, for their loving support, motivation, and patience during the completion of this work. I would like to express my sincere appreciation to my research supervisor Mike Blades for his invaluable and scholastic guidance, constructive criticism, and support throughout the course of this project. I would also like to thank my colleagues in the group for their encouragement and valuable discussions. The experimental apparatus used in this study was constructed with the assistance of the electrical engineering, mechanical engineering, and glassblowing services in the Department of Chemistry, University of British Columbia. Martin Carlisle of the electrical engineering, Des Lovrity of the mechanical engineering and Steve Rak of the glassblowing services deserve special recognition for their contributions. Acknowledgement is also made to the Canadian Commonwealth Scholarship and Fellowship Program and to the Natural Science and Engineering Research Council of Canada (NSERC) for financial support. 1 C h a p t e r 1 Introduction Plasma discharges are widely used in many industrial, research, and laboratory applications. The discharge has become the atomization, excitation and/or ionization source of choice for many applications. These have been reviewed in several publications over the past two decades [1-11]. One kind of plasma discharges is the Radio Frequency Capacitively Coupled Plasma (RF-CCP). An RF-CCP is usually created and sustained using a radiofrequency field in the frequency range of 20 kHz to 1 GHz. A striking feature of CCP's is the wide range of operating conditions that it allows. Depending on the plasma source, power levels can range from a few watts to several hundred watts; the discharge pressure can range from less than 20 torr to as high as atmospheric pressure, and different discharge gases (both noble and molecular) can be used, usually with a flow range of 0.02 to 10 L min 1. A result of this wide range of operational conditions is that discharges with different characteristics such as excitation temperature, gas temperature, electron density, degree of ionization, and chemical composition can be created. Moreover, the geometry of the plasmas can also be used to tailor the plasma characteristics. Because both practical applications and laboratory research usually call for plasmas with specific characteristics and geometry, CCPs may in many cases preferable to other plasmas. 2 1.1 Parallel-Plate Capacitively Coupled Plasma (PP-CCP) The parallel-plate capacitively coupled plasma (PP-CCP) operating at atmospheric pressure is a relatively new and fairly unique source [12-17] for spectrochemical analysis of trace and ultra-trace elemental concentrations. In this PP-CCP source, as shown in Fig. 1.1, pressure plasma is sustained by placing the quartz discharge tube (T) between a pair of water-cooled conducting electrode (E) to which radio-frequency power is applied. Capacitive coupling provides for very efficient energy transfer from the rf power supply to the plasma, thus enabling the discharge to be operated over a wide range of powers and plasma gas flow rates. The plasma torch and the electrodes are housed in an aluminum support structure. Macor™ (I) provides the electrical insulation between the electrodes and aluminum support structure. An electrically heated graphite platform serves as a liquid sample vaporizer. The plasma gas carries the vaporized sample into the plasma where the processes of dissociation, excitation, and ionization occur. Gaseous samples can also be introduced directly into the plasma. Due to the high transport efficiency, only a few microliters of analyte solution are needed for analysis. Although this type of discharge can be sustained using either argon or helium, the latter is more common as, at atmospheric pressure, the voltage required to breakdown and sustain a He-CCP is lower than that required for Ar-CCP. This is because, for the dimensions normally used for CCP's at atmospheric pressure, the ionization rate for helium is of a considerably higher magnitude than that for argon [18]. An added advantage of helium is, due to its high ionization energy (24.6 eV), it can efficiently excite and ionize analyte through resonant Penning and charge transfer processes [19-21]. Figure 1.1: Schematic diagram of the plasma source. E - copper electrodes, I -Macor® insulator, K- clamping knob, RF- radio frequency connector, T-quartz torch, S- plasma gas inlet along with the vaporized sample. 1.2 Thesis Objective As was previously stated, the atmospheric pressure parallel plate capacitively coupled plasma (PP-CCP) is a relatively new optical emission source. Since its introduction in 1988 [12], a number of studies have been completed on the application of this source as a detector for gas chromatography [13-15]. Despite these studies, the fundamental and analytical characteristics of this source remain 4 largely unknown. An investigation of the fundamental properties of plasma sources helps lead to an improved understanding, and aids in the development, optimization, and comparison of different plasmas as spectroscopic sources. Common fundamental properties that can be investigated include electron number density, excitation and rotational temperatures, specific heat, viscosity, and conductivity [22]. Analytical characteristics such as temporal and spatial characteristics, degree of ionization, matrix effects, and analytical figures of merit (sensitivity, precision, detection limits, etc.) determine the success of an analytical method. Gaining some information on these fundamental and analytical characteristics, as well as how these characteristics are affected by experimental parameters (e.g. applied plasma power, plasma gas flow rate, and plasma volume) will ultimately help to optimize the source. The studies reported in this thesis were designed to gain precisely such information. 1.3 Evolution of the PP-CCP 1.3.1 Plasma Although interest in electric-discharge phenomena may be traced back to the beginning of the eighteenth century, the term "plasma" was first introduced by Irving Langmuir in 1923 while he was investigating electric discharges. In a more generalized form, a plasma can be described as a partially ionized gas consisting of equal numbers of positive ions and negative electrons in a sea of neutral atoms and molecules. Since the number of positive ions and negative electrons are 5 approximately equal, plasma is virtually electrically neutral and highly conductive. Plasma is sometimes referred to as the fourth state of matter, distinct from the solid, liquid, and gaseous states. Occurring prominently in the sun, stars and interplanetary and interstellar space, nearly all the universe's visible matter exists in a plasma state. Auroras, lightning, and welding arcs are plasmas, and plasmas also exist in neon and fluorescent tubes, in the crystal structure of metallic solids, and in many other phenomena and objects. The Earth itself is immersed in a thin plasma called the solar wind and is surrounded by a dense plasma called the ionosphere. A plasma may be produced in the laboratory by heating a gas to an extremely high temperature, which causes such vigorous collisions between its atoms and molecules that electrons are ripped free, yielding the requisite electrons and ions. The properties of this gas may then be dominated by the effect of electromagnetic forces acting on these charged particles; such a gas is called plasma. A transfer of energy to the bound electron, by a less dramatic process than ionization, would enable the electron to jump to a higher energy level within the atom, with corresponding quantum absorption of energy that is known as excitation. These excited states are rather unstable and undergo a de-excitation process upon which the electron configuration returns to its original (ground) state through one or more transitions. Each transition is accompanied by the emission of a photon of very specific energy, equal to the energy difference between the relevant quantum levels. This photon emission is used as the basis of spectrometric analysis. 6 1.3.2. Plasma as Optical Emission Sources Atomic emission spectrometry (AES), sometimes denoted as "optical emission spectrometry" (OES), is one of the major branches of that part of analytical spectrometry that derives analytical information from atomic or elementary ionic spectra in the optical region of the electromagnetic spectrum. Optical emission sources have been used as analytical tools since Bunsen and Kirchhoff's laboratory flame experiments [23] in the mid 1800s. A few decades later, in 1882, Hartley [24] made the first spectroscopic quantitative analysis by measuring the amount of Beryllium in certain compounds using a "condensed spark" optical emission source. Apart from flames and laser-induced plasmas, current optical emission sources rely mainly on electrical discharges for creating plasmas. Electrical discharges that have found broad application as emission sources are the inductively coupled plasma (ICP), direct current plasma (DCP), and microwave induced plasma (MIP). Without a doubt, the ICP is the current workhorse of trace elemental analysis. Other electrical discharges such as dc (Arc) and ac (Spark) plasmas, capacitively coupled microwave plasma (CMP), and capacitively coupled plasma (CCP) are also used as emission sources for spectrochemical analysis. All these discharges differ both in their geometry and in the frequency of the applied electric fields. Their geometry varies from electrodes in direct contact with the plasma, as in the DCP, to induction coils separated from the plasma by quartz torches, as in the ICP. The frequencies of the applied field are "direct current" for DCP and GD (glow discharge), 1-100 MHz for rf plasmas with either capacitive or inductive coupling, and in excess of 1 GHz, usually 2.45 GHz, for microwave 7 plasmas. A brief comparison of frequency, operating parameters, and some fundamental properties for the MIP, ICP, and CCP is provided in Table 1.1. Table 1.1: Comparison between MIP, ICP, and CCP PARAMbIbH MIP ICP CCP Operating Frequency, MHz 2450 5-40 13.56/27.12 Operating Power, w <100 -1000 20 - 600 Operating Pressure, Atm. 0.01-1.0 1.0 1.0 Support Gas Ar/He Ar /He Ar /He Gas Consumption, L min'1 0.5-2.0 10-20 0.1-2.0 Excitation Temperature, K 3000- 5000 5500 - 7000 3000 - 5000 Gas Temperature, K 1100-2500 4000 - 5000 1100-1500 1.3.3 Plasma Formation Apart from solid-state plasmas, such as those in metallic crystals, plasmas do not usually occur naturally at the surface of the earth. For laboratory experiments and technological applications, plasmas therefore must be produced artificially. In most gases, before any significant degree of ionization is achieved, temperatures of approximately 10,000 K are required. As all substances melt at temperatures far below the required level, no container as yet can withstand an external application of 8 the heat necessary to form a plasma; therefore, any heating must be supplied internally. One technique is to apply an electric field to the gas to accelerate and scatter any free electrons, thereby heating the plasma. Electrons can be raised to much higher temperatures than other particles, because of their small energy loss in elastic collisions. For plasma formation to occur, a sufficiently high electric potential difference between two electrodes must be applied to breakdown the gases. This is known as breakdown potential. The exact value of this breakdown potential is dependent upon the separation, geometry, and composition of the electrodes, the composition and pressure of the gas between them, as well as the frequency. For the sake of simplicity, a direct current (dc) glow discharge is considered first, as Schwab et al. [25-27] have shown that, in many respects, high pressure rf discharges exhibit properties which are analogous to those of dc glow discharges. By applying a sufficient dc potential between two electrodes a dc glow discharge could be formed. The majority of the space between the electrodes is filled by a bright glow known as the negative glow and a comparatively thin dark region, adjacent to the cathode, known as the dark space or sheath. A cathode fall of potential develops across the sheath region, leaving the glow space nearly field free. In the plasma the electrons and ions become lost by ion-electron recombination. There is also a considerable energy loss from the discharge. Energetic particles impinge on the electrodes and walls of the system, resulting in heating; this energy loss is then conducted to the environment. As a result, in order to maintain a steady state discharge, there must be a balancing energy input to the discharge along with a good deal of ionization for current continuity. In a glow discharge, secondary 9 emission through positive ion bombardment on the cathode, is a major source of electrons, which then absorb energy from the applied electric field. After acquiring sufficient energy from the field, the electrons begin to ionize the discharge gas. The application of a low frequency ac voltage, as opposed to a dc voltage, will result in a series of short-lived discharges with the electrodes successively taking opposite polarities. The discharge will be extinguished when the cathode potential drops below a critical value because of the accumulation of the positive charges. Rapid polarity reversals of voltage pulses allows for rapid charge compensation and reapplication of the desired high voltage overcoming the inherent decay time constant. To achieve a pseudo-continuous discharge, pulse frequencies in the order of 1 MHz are required [28]. The key fundamental phenomenon occurring from the application of a high frequency ac potential is the "self-biasing" of the electrodes. This self-bias potential arises because of the differential mobility of electrons and positive ions in the discharge. Due to higher mobility, the electrons are easily collected on an electrode whenever it becomes positive with respect to the glow space. As the frequency of the applied potential increases, the minimum pressure at which the discharge can be operated is reduced [29]. This reduction indicates, as in the case of dc glow discharge, that there is an additional source of ionization other than the secondary electron emission from the electrode. This additional source of electrons arises when sufficiently highly energetic electrons, oscillating in the time dependent electric field, undergo inelastic collisions with the plasma gas atoms to cause ionization. It has been pointed out that this collision process is the primary requirement in sustaining the discharge. Therefore, the high potential electrode that 10 is necessary in a dc glow discharge for the secondary electron emission, is not required to sustain the rf discharge. 1.3.4 Electrothermal Vaporizer (ETV) The term "electrothermal vaporizer" or "ETV" can be used to encompass a wide range of physical platforms used for the resistive heating of a substrate upon which a (condensed phase) analyte sample has been placed so as to result in its (rapid) release into the gas phase. This process may result in the simple vaporization of the analyte with or without a matrix, or the partial or complete atomization of the analyte and/or matrix. Samples may be introduced in solid, slurry, or liquid form. In the field of atomic spectroscopy, a variant of the ETV technique, pertaining to its use as an atomizer has been well known since its first introduction by Lvov in 1961(30]. Although the ETV technique was first developed for atomic absorption spectrometry (AAS), it has been in use as a successful sample introduction source for plasma emission spectrometry since the mid 1970s. The requirements for successful application of the ETV as an atomization source for AAS are distinctly different from those set for its use as a vaporizer for sample introduction into plasma sources. Whereas complete atomization of the sample and confinement of the atomic vapor within the high temperature observation volume is sought for AAS, in plasma emission spectrometry it is sufficient to ensure complete vaporization of the analyte and its efficient transport from the ETV to the plasma. Formation of gaseous analyte molecular species or the adsorption / occlusion of molecules or atoms of the 11 analyte in a transportable aerosol is the principal objective [31]. The ETV may thus be considered one of the constituents of a tandem source [32] in which the best characteristics of the device can be separately optimized in order to take advantage of its unique sample treatment capabilities in combination with those offered by a variety of plasma sources. One of the key advantages of using the ETV technique is that, in addition to the high transport efficiency achieved, small discrete sample volumes can be conveniently processed. The ETV efficiently handles solutions containing dissolved solids, including solution with a 100% dissolved solid content (i.e. solids), and it can be directly used for the analysis of organic solvents. Enhanced detection power ensues [33, 34], often as a direct result of reduced solvent-related polyatomic spectral interferences (with lower oxide fractions) in addition to the possible sample pre-concentration designs which can be implemented with the ETV. The use of the ETV, however, is not without significant drawbacks. The discrete nature of sample introduction leads to transient responses, which are more difficult to quantitate precisely (5 - 10% relative standard deviation, RSD, typical) than a steady-state signal. As well, lower sample throughput, requiring several minutes per sampling, is an added disadvantage. Although rods, boats, filaments, strips, cups, and tubes have been used as ETV's, tubular graphite furnaces are the best characterized and provide the most controllable thermal environment [35]. When PP-CCP was first introduced as a spectrochemical source, a tantalum strip was used to introduce the liquid sample into the plasma [16]. Unfortunately the lifetime of the tantalum strip was limited to 15-12 20 heating cycles, limiting its capability as an ETV source. To address this problem, the tantalum strip was replaced by a specially designed pyrolytic graphite platform [17]. A detailed description of the aforementioned ETV is given in Chapter 2 of this thesis. 1.3.5 The Discharge in PP-CCP The PP-CCP is an atmospheric pressure discharge sustained by application of radio-frequency power to a pair of water-cooled electrodes that "sandwich" the quartz discharge tube forming a capacitive, transverse discharge. In the study reported herein, the plasma is operated using helium as the plasma gas at 13.56 MHz, although other gases and frequencies could also be used. A spot of "light-greenish purple" plasma becomes visible with the application of as low as 10 W of rf power. By increasing the applied power, the spot becomes bigger and brighter, eventually filling the entire volume of the torch. An advantageous feature of this plasma is that it is self-igniting, i. e. there is no need for external initiation. The source of electrons for breakdown and maintenance of the discharge in this plasma type is primarily drawn from support gas ionization rather than secondary electron emission [18]. According to Beneking [36], the structure of the plasma formed consists of two main regions, a positive column-like bulk plasma region and two sheath regions located between the discharge volume and the dielectric walls in front of the field-supplying electrodes. In this parallel plate geometry, the electrodes are isolated from the discharge by the dielectric quartz layer, causing no net flow of electric charge but only a 13 displacement charge. Equal ion and electron currents through the sheath region satisfy this requirement of no net charge flow. Most of the energy supplied to the discharge is dissipated in the sheath regions, resulting in two localized zones where most of the excitation and ionization take place. The discharge originates from equal electrode areas, causing identical electrical field strength upon the electrodes during each half-cycle. This makes the plasma source symmetric if we imagine a plane running longitudinally through the discharge torch. 1.4 A Brief Chronicle of CCP Capacitively coupled plasmas (CCPs) have been generated at various frequencies in both rf and microwave regions. CCPs were first described in 1941 by Critescu and Grigorovici [37, 38], who reported that a CCP discharge could be formed between two circular plates, separated by up to 15 cm. These plates formed a capacitor which, as part of the resonant circuit, determined the frequency of the rf generator, in the range of 60 - 90 MHz. Analytical applications of CCP were reported for the first time by Stolov [39] and Badarau et al. [40] in the mid 1950s. Several types of CCP continued to be explored and studied throughout the following two decades [25-27, 41-51]. Badarau et al. [40] used a capacitor comprised of a hollow cylinder and a coaxial electrode to generate a discharge, which was capacitively induced by a rectified rf source at 43 MHz. Another arrangement for capacitive coupling, consisting of an inductance coil of tubular copper with coaxial electrode, was used by both the Mavrodineanu et al. 14 and West ef al. groups [42, 46] in the mid 1960s. An unique CCP was described by Cristescu ef al. [48] in the late 1960s, where they superimposed a dc (direct current) component on the rf. The rf power was applied across two electrodes, and an adjustable dc voltage was applied between them and a third movable electrode. Adjusting the dc current from 0.1 to 1.0 ampere changed the apparent excitation temperature, as well as the appearance of the plasma. An atmospheric pressure capacitively coupled plasma known as FAPES (furnace atomization plasma excitation spectrometry) was developed by Liang and Blades [52] in the late 1980s, wherein a helium or argon CCP could be formed inside an otherwise normal graphite-furnace atomizer. The plasma is formed by applying high voltage rf power on a conductive central electrode located in the graphite furnace with a coaxial geometry. The reported excitation temperatures from this source were in the range of 2900 - 5000 K [53, 54]. In recent years, Cordos and coworkers reported an atmospheric pressure rf CCP torch with coaxial electrodes in tip-ring geometry [55-60] and central tube electrode geometry [61] for spectrochemical analysis. In the tip-ring geometry, the discharge is unipolar and is maintained between a water-cooled electrode with a sharp platinum [55] or tungsten [56] tip, which is connected to the rf high voltage. The electrode is placed inside an 18 mm i. d. quartz tube and a counter electrode, a metallic ring, is placed outside the tube (65 mm apart from the tip) and connected to the ground. The discharge is formed and sustained at atmospheric pressure in air, argon or a mixture of both by using a 27.12 MHz rf generator at low to medium power (85 - 275 W). The reported excitation temperatures are in the range of 2000 - 4700 K [57] for this source. For 15 this tip-ring type of torch, a part of the sample remains in the plasma mantle and does not reach the hot plasma core. To overcome this problem, Cordos et al. [61] came up with a new torch design with a central tube electrode and one or two ring counter electrodes. In this design, the torch's main body is made of brass and is water-cooled. A P T F E tube with an i. d. of 5 mm is placed in the central channel of the torch, into which a molybdenum tubular electrode with an i. d. of 2-4 mm is inserted and connected to the rf generator. A quartz tube is placed over the tubular electrode and the ring electrodes are placed outside the quartz tube and connected to the ground. The analytical performance is substantially improved by using this tubular electrode over the tip-ring electrode. For microwave capacitively coupled plasmas, the power is delivered to a single-electrode torch through a rectangular waveguide. These devices have the advantage of being able to tolerate the introduction of water vapor, are extremely stable at a wide range of powers, and can be operated using various support gases, including Ar, He, N 2, and air. They do, however, suffer from the problem of electrode erosion and contamination. These plasmas are commonly known as capacitively coupled microwave plasmas (CMPs). The configuration of CMP sources finally developed for analytical work are derived from the initial work of Cobine and Wilbur [62], with further developments contributed by Kessler et al. [63-66], and Murayama and Yamamoto [67-70]. Winefordner and his co-workers [71-80] have developed several new torch designs for the C M P employing tantalum, tungsten, and graphite electrodes through which the sample can be introduced. The single Pt-clad tungsten electrode atmospheric pressure microwave plasma described by Winefordner et al. 16 [71, 72] has considerably improved the performance of the CMP in comparison to the single Al electrode CMP system described by Murayama ef al. [67-69]. The tantalum tubular electrode (4 mm i. d., 3.8 mm o. d., 50 mm long), described by Patel et al. [74], allows the sample to be introduced directly into the plasma flame at the tip of the tubular electrode. The electrode was held in position inside an aluminum electrode holder which was screwed into a brass tube. A Teflon tube (4mm i. d., 7.7 mm o. d., 100 mm long) was inserted into the brass tube which formed a central channel running down the entire length of the quartz tube. This tantalum electrode was water-cooled. Hwang et al. [73] reported a graphite tubular electrode which operated similar to the tantalum torch. The graphite electrode does not require cooling since its melting point is much higher than that of aluminum. Ali et al. [77] subsequently adopted a microsampling technique for liquid samples in their CMP. They used a graphite electrode and a graphite cup from which micro-liter volumes of liquids [77] and solids [76] could be introduced into the plasma. Ali et al. [78, 81] also reported a tungsten filament electrode, where the filament (0.25 mm diameter) was coil-shaped in order to hold the sample primarily by adhesion. The extended vertical wires of the coil were slid into the CMP torch and the plasma was formed on the top of the electrode. The excitation temperatures for these torches were in the range of 4000 - 5000 K and the detection limits of several elements were reported in the picogram range. In all of the above mentioned plasmas, one or more electrode(s) are in direct contact with the plasma. This causes electrode erosion and contamination, both 17 problems for spectrochemical analysis. In 1947 Babat [82] reported the first CCP where no electrode was in direct contact (electrodeless) with the plasma. Babat produced the plasma using two external annular electrodes, separated from each other, and enclosing the discharge tube. Two decades later, Egrova [49] described a similar geometrical structure of a CCP source for spectrochemical analysis of solutions. Egrova produced the plasma in a water-cooled, cylinder around which two cylindrical electrodes, separated by 100 mm, were placed. At atmospheric pressure, the argon discharge was excited with a 36 MHz, 1.5 kW radio frequency generator. In 1975 Zvyagintsev and coworkers [50, 51] reported an electrodeless capacitive discharge similar to Egrova's device. The discharge torch was made from 10 to 70 mm i. d. quartz tubes and two annular electrodes separated by 5 - 25 cm placed around the torch. The discharge was operated using a 150 MHz rf generator at powers between 2 and 3 kW. The support gases utilized for the discharge were, air nitrogen, and helium. The temperature reported was 5500 K. The use of this annular electrode CCP as an element-specific detector for gas chromatography was first described by Platzer et al. [83].The discharge torch consisted of a 0.5 mm i. d. water-cooled fused silica tube surrounded by two annular electrodes separated by a short distance. The plasma support gas was helium. The plasma was sustained at atmospheric pressure using a 27.12 MHz rf generator and a forward power of 150 W. The use of an rf parallel plate capacitively coupled discharge for spectra -analytical determinations was first described by Winslow [84] in 1980. The discharge torch described by Winslow consisted of a 16 cm long, 8 cm i. d. quartz tube with 18 two rf electrodes located outside the torch in a parallel position. Winslow used oxygen as a plasma support gas. In this geometry, the plasma could be sustained at low pressure using a 13.56 MHz rf generator. An atmospheric pressure rf parallel plate CCP was first reported by Liang and Blades [12] in 1988 for spectrochemical analysis of small, discrete sample volumes of 1-10 \iL. The device consisted of two distinct parts, the CCP discharge, which was sustained in a cylindrical or square-bore quartz tube with two external parallel plate rf electrodes, and a tantalum strip, electrothermal vaporizer (ETV) sample introduction system. Both helium and argon could be used as the plasma support gas. The plasma could be sustained by using either 13.56 or 27.12 MHz with a forward power range of 20 - 400 W. Huang ef al. [13-15] demonstrated the use of this source as a detector for gas chromatography (GC). Recently Rahman and Blades [16, 17] replaced the tantalum strip atomizer with a graphite platform and reported some fundamental characteristics of this source. The analytical characteristics and some of the fundamental properties of this PP-CCP source are described in detail in the following chapters of this thesis. 1.5 Plasma Diagnostics An investigation of the fundamental properties of plasma sources helps lead to an improved understanding, and aids in the development, optimization, and comparison of different kinds of plasma as spectroscopic sources. Common fundamental properties which are investigated include, the spatial distribution of plasma species, electron number density, temperatures, specific heat, viscosity, and 19 conductivity [22]. Usually, spectroscopic diagnostic measurements are preferred for these investigations as they are non-invasive and the methodology is fairly developed. A brief discussion of plasma parameters are provided herein. For more complete information on plasma diagnostics please refer to the works cited in the bibliography [85-89]. 1.5.1 Spatial Distribution of Plasma Species Although a plasma seems to be a simple medium, many types of neutral and ionized species aside from electrons, can be present. These neutral and ionized species are either in excited or in ground states. Neutral atoms can be excited (He*, for a helium discharge), with the lowest levels being either resonant or metastable levels; molecular excited He (He2) and molecular He ion (He2+) may also be present. The introduced analyte (X) can be observed in several forms (X, X*, X+, X+*, X2* etc.), depending on the respective excitation and ionization energies. Some molecular species (i. e. OH, N2, N2+, NO etc.) may still exist in the plasma, at least at trace levels, even though the discharge has a sufficiently high temperature to dissociate them. Due to the geometry of the source and the mechanism by which the discharge is formed, the plasma species are not spatially homogeneous. Measurement of their spatial distributions yields valuable information from both a fundamental and analytical point of view. Knowledge of the atom and ion distribution within the discharge can be used to gain information on the ability of the source to atomize and 20 ionize analyte species. The population distribution, of atoms or ions, can be determined by measuring the emission intensity, (1^), which is proportional to the number density of the excited atomic or ionic state (Nq), by the following equation: •em = ^ A N q (1.1) where: h is the Planck's constant; v is the frequency of the emitted radiation, and A is the Einstein transition probability ( s'1) for spontaneous emission. The Boltzmann distribution allows us to define the ratio of the populations or number densities N p and N q of atoms in the state of energy E p and E q with q >p. •(Ea - ED) Nq g q M = TT e x P Np g P kBT, exc (1.2) where gq and gp are the statistical weights of the two levels, kB is the Boltzmann constant and T e x o is the excitation temperature, assuming the discharge is in p-LTE (partial local thermal equilibrium) condition. If we consider the total number density of the atom is N, we have g q exp — ^ _ Texc / 1 o\ N~ " Q(T) where Q (T) is the internal partition function. The Q (T) for an atom is defined as 21 Q CO = 2q 9q e x P (1.4) exc For "low" temperatures, the value of Q(T) is approximately equal to the statistical weight g0 of the ground state. For a temperature greater than 4000 K, the partition function may vary with the temperature and electron number density [90]. By substituting the value of Nq from equation 1.3 in equation 1.1, we obtain the expression for emission intensity as follows: where h is the Planck's constant. The spatial population distribution of atoms and ions can be determined by measuring the spatially resolved emission intensities of the interested species. Knowledge of these spatial distributions is vital for spectrochemical analysis, as this information helps to determine the analytical zone in a source. 1.5.1 Plasma Temperature Temperature is one of the most important characteristics for an optical emission source. The temperature can be used to compare different plasma sources with respect to their ability to atomize, excite, and ionize analyte species. A higher temperature makes the plasma "robust" and able to atomize and excite the analyte, resulting in an intense excitation signal with fairly low matrix interference effects. The (1.5) 22 plasma temperature appears in a number of thermodynamic functions [86, 87, 91] which can be used to describe the number density of analyte atoms and ions in a discharge. These distributions are: (a) Maxwell velocity distribution (gas kinetic temperature, TK and electron kinetic temperature, TE), (b) Boltzmann distribution of bound state populations (excitation temperature, T^), (c) Saha distribution of ionization products (ionization temperature, T^), and (d) Guldberg - Waage distribution of dissociation products (dissociation temperature, Td). Under the condition of complete thermodynamic equilibrium (CTE), each of the distributions is characterized by a unique temperature. For most laboratory plasmas the low optical density and presence of high concentration (of atoms, ions and electrons) and temperature gradients often prevents the establishment of CTE [92]. This deviation from equilibrium means that the plasma cannot be described using a unique temperature and leads to the use of several different temperatures (TK * T E * T e x c * T l o n * Td). However, when the source is in a state of local thermal equilibrium (LTE), a unique temperature can be achieved for each point in the source, allowing for the possibility of different temperatures at different points. Experimental techniques have been designed to measure each of these temperatures. In-depth descriptions are available in the cited sources, along with more comprehensive descriptions of plasma temperatures and their relevant distribution functions [86, 87, 90, 92, 93]. For the present study, the atomic and molecular excitation temperatures (commonly known as excitation temperatures T e x c and rotational temperatures T r o t , respectively) were measured. Due to their non-invasive characteristics, spectros -23 copic techniques were employed to measure these temperatures. There are various methods to determine these temperatures, which may be categorized depending on the use of (a) absolute line intensities, (b) relative line intensities, (c) linewidths, and (d) the line-to-continuum ratio [90]. Among these categories, "relative line intensities" are most commonly used, with variants of the Boltzmann plot technique and line pair technique. Both techniques assume that the populations of the species in the excited state follow the Boltzmann distribution: l n "^T = I T T 1 - + mC (1.6) gq A Kg I exc Therefore, the excitation temperature, Te x c, can be determined by measuring the relative intensity of each available spectral line and plotting In (I X / gq A) as a function of Eq. The slope of the line (-1/kBTexo) then provides the excitation temperature. The transition probability, A, may be replaced by oscillator strength, f, using the following relationship: g q A = ^ | I (1.7) and plotting In (I X31 gp f) as a function of E q can also be used to determine the excitation temperature from its slope (-1/kBTexc). This spectroscopic method of temperature determination is normally known as "the Boltzmann plot method". For the rotational temperature of molecules, Trot, the same equation is used with the appropriate statistical weights being replaced by gq. Detailed derivation from 24 equation 1.5 for rotational temperature calculations can be found sources cited [90, 94}. In the line pair method, Equation-1.6 is converted and simplified to: _ ^ - E 2 L S h A j X g , ,h 1 Texc = — r d-8) k B l n a , ' : ' 7 - l n r . I 92*2*1 l 2 j where supscripts 1 and 2 represent different parameters for line 1 and line 2 respectively. Using Equation 1.8, it is possible to calculate the excitation temperature by measuring the relative intensities of two lines at different energies, as a function of time during the atomization step. 1.6 Analyte Excitation and ionization Unlike the ICP, MIP and GD, for which much mechanistic information has been gathered, fundamental examination of analytical helium plasmas has not been extensive. However, in 1977 Beenakker reported that a number of possible reactions are responsible for the observed excited and ionized states [95] in an atmospheric pressure plasma. The most important of these are electron impact excitation and ionization, Penning ionization and excitation, charge transfer ionization (and excitation) involving He+ or He2+, radiative and three-body ion-electron recombi -nation, and radiative de-excitation. Electronic excitation due to electron impact is a major process in plasmas. In addition to the production of excited state analyte atoms, electron impact is also responsible for populating the metastable levels of discharge gas atoms (Hem\ He^' 25 for a helium plasma). Collisions between the metastable and neutral atoms may lead to an internal energy transfer from the metastable to the neutral, resulting in the excitation or ionization of the latter; this collision process is known as Penning excitation and ionization. This process has a resonance when the excited state atom or ion energy is in proximity to that of the metastable energy value of the discharge gas. Hess and coworkers [96, 97] have suggested that Penning-type collisions contribute significantly to the formation of excited state ionic species in a glow discharge. Charge transfer is another likely mechanism for excitation and ionization in helium discharge. Charge transfer is based on near-resonant energy transfer between ionized helium (in case of He discharge) and a ground state neutral atom or molecule. When comparing the helium discharge with other gas discharges, the former yields intense nonmetal ion emission signals [20]. Carnahan and coworker [20, 98, 99] suggested that the mechanism responsible for the intense nonmetal ion emission is charge transfer. Huang et al. [14] also suggested that charge transfer reaction plays an important role in the excitation processes in a helium capacitively coupled plasma. Recombination is the inverse of ionization. For radiative recombination, the product is an excited analyte atom and a continuum photon (fiv^. The photons produced in this reaction may have a continuous range of energies because the colliding electrons posses a range of energies. A third body, usually the gas atom, (dependent on the pressure of the discharge) may also take part in the recombination process, hence the name three-body recombination, to form the excited analyte atom. In radiative de-excitation, an excited atom (or ion) relaxes to a 26 lower (ground) state by emitting a quantum of radiation energy. This radiative de-excitation mechanism is, therefore, the most significant process, analytically, for emission spectrometry since it creates the characteristic line spectra used for chemical analysis. 1.7 Thesis Epitome This thesis represents an experimental study of the spectroscopic features of a radio frequency (rf) parallel plate capacitively coupled plasma (PP-CCP) operating at atmospheric pressure. Since the PP-CCP is a relatively new source for spectrochemical analysis, this thesis is focused on the characterization and better fundamental understanding of this CCP discharge. Various plasma properties and analytical characteristics, such as excitation temperature, rotational temperature, degree of analyte ionization, spatial distribution of plasma species, detection limit, influence of plasma applied power and gas flow rate on analyte emission, and the effect of electrode length on analytical performance, have been investigated using emission spectroscopy. Excepting the last, the following chapters in this thesis focus on specific aspects of the study. Chapters 3, 5, and 6 have been published as papers [16, 17,100] and Chapters 4 and 7 will be submitted in the near future. As a result, all chapters can be read separately and unavoidably contain some overlap material. An experimental system developed for the measurement of temporally and spatially resolved transient signals is described in Chapter 2. Chapter 3 describes the plasma temperatures and analytical figures of merit for the PP-CCP source using lead and silver as analytes. Atomic emission intensities from helium, lead, and iron, and molecular emission intensities from OH and N2 + were measured in an attempt to 27 characterize the excitation and rotational temperatures of this source. This chapter also demonstrates that some important fundamental and analytical characteristics of the parallel plate He plasma are similar to those of a microwave induced plasma. The spatial distribution of plasma background species is discussed in chapter 4 using helium atomic and OH and N2 + molecular emission. This chapter also discusses the influence of plasma applied power and the plasma gas flow rate on spatial distribution. In the study reported in Chapter 5, silver was used as an analyte to investigate the effects of easily ioniable elements (ElEs) on analyte emission. NaCI and NaN03 were used as ElEs. Both enhancement and suppression of analyte emission signals were observed depending upon the amount of matrix species. It is suggested that two opposing mechanisms for interference from EIE in the PP-CCP source are active, one causing an enhancement of analyte signal and the other causing a depression. In Chapter 6 the results of analyte ionization studies are shown, using magnesium and cadmium as the spectrometric species. It is shown that the applied forward power in the plasma has a significant effect on the degree of ionization for both magnesium and cadmium analytes. In this chapter, the influence of EIE on analyte ionization is also discussed. The effect of electrode length and hence plasma volume on analytical characteristics of this source, was also studied and is presented in Chapter 7. The results of this study have shown that the analyte detection limit improves with increasing electrode length, reaching a low of 0.96 pg for silver and 0.34 pg for magnesium for a 6 cm long electrode. The electrode length as reported in Chapter 7 also influences Analyte ionization. 28 1.8 References 1. R. Sing, Spectrochim. Acta, 54 Part B, (1999), 411 - 441. 2. D. Gunther, S. E. Jackson, and H. P. Longerich, Spectrochim. Acta, 54B, (1999), 381-409. 3. A. Bogaerts and R. Gijbels, Spectrochim. Acta, 53B, (1998), 1-42. 4. E. H. van Veen and M. T. C. de Loos-Vollebregt, Spectrochim. Acta, 53B, (1998), 639-669. 5. N. Jakubowski, L. Moens, and F. Vanhaecke, Spectrochim. Acta, 53B, (1998), 1739-1763. 6. H. Niu and R.S. Houk, Spectrochim. Acta, 51B, (1996), 779-815. 7. D. Beauchemin, J. C. Y. Le Blanc, R. Peters, and A. T. Persaud, Anal. Chem., 66, (1994), 462R-499R. 8. D. Beauchemin, J. C. Y. Le Blanc, R. Peters, and J. M. Craig, Anal. Chem., 64, (1992), 442R-467R. 9. M. W. Blades, P. Banks, C. Gill, D. Huang, C. Le Blanc, and D. Liang, IEEE Transactions on Plasma Science, 19, (1991), 1090-1113. 10. S. Caroli, J. Anal. Atom. Spectro., 2, (1987), 661-678. 11. J. P. Matousek, B. J. Orr, and M. Selby, Prog. Analyt. Atom. Spectrosc, 7, (1984), 275-314. 12. D. C. Liang and M. W. Blades, Anal. Chem., 60, (1988), 27-31. 13. D. Huang, D.C. Liang, & M. w. Blades, J. Anal. Atom. Spectro., 4,(1989), 789 14. D. Huang and M. W. Blades, Appl. Spectrosc, 45(9), (1991), 1468-1477. 15. D. Huang and M. W. Blades, J. Anal. Atom. Spectro., 6, (1991), 215. 29 16. M. M. Rahman and M. W. Blades, Spectrochim. Acta, 52B, (1997), 1983-93. 17. M. M. Rahman and M. W. Blades, Spectrochim. Acta, 55B, (2000), 327-338. 18. M. W. Blades, Spectrochim. Acta,, 49B, (1994), 47-57. 19. S. K. Chan and A. Montaser, Spectrochim. Acta, 40B, (1985), 1467. 20. P. G. Brandl and J. W. Camahan, Spectrochim. Acta, 49B, (1994), 105-115. 21. P. G. Brandl and J. W. Carnahan, Appl. Spectrosc, 49, (1995), 1781 - 1788. 22. W. Lochte-Holtgreven, Evaluation of Plasma Parameters, Plasma Diagnostics, W. Lochte-Holtgreven Ed. (John Wiley & Sons, New York, 1968) 23. G. Kirchhoff and R. Bunsen, Phil. Mag., 22, (1861), 329. 24. W. H. Hartley, J. Chem. Soc,, (1882), 210. 25. H. A. Schwab and C. K. Manka, J. Appl. Phys., 40, (1969), 696-706. 26. H. A. Schwab and R. F. Hotz, J. Appl. Phys., 41, (1970), 1503-1506. 27. H. A. Schwab, Proc. IEEE, 59, (1971), 613. 28. G. S. Anderson, W. Mayer, & G. K. Wehner, J. Appl. Phys., 33, (1962), 2991. 29. H. Norstrom, Vacuum, 29, (1979), 341-350. 30. B. V. LVov, Spectrochim. Acta, 17, (1961), 761 -776. 31. T. Kantor, Spectrochim. Acta, 43B, (1988), 1299-1320. 32. M. W. Borer and G. M. Hieftje, Spectrochim. Acta Rev., 14, (1991), 463. 33. J. M. Carey and J. A. Caruso, CRC Crit. Rev. Anal. Chem., 23, (1992), 397. 34. D. C. Gregoire, Can. J. Spectros., 21, (1997), 1. 35. R. E. Sturgeon, Fresenius'J. Anal. Chem., 355, (1996), 425-432. 36. C. Beneking, J. Appl. Phys., 68, (1990), 5435-5446. 37. G. Cristescu and R. Grigorovici, Naturwissenschaften, 29, (1941), 571-572. 30 38. G. D.Cristescu and R. Grigorovici, Bull. Soc. Roum. Phys., 42, (1941), 37. 39. A. L. Stolov, Uch. Zap. Kazan. Gos. Univ., 116, (1956), 118. 40. E. Badarau, M. Giurgea, G. H. Giurgea, and A. T. H. Trutia, Spectrochim. Acta, 11, (1957), 441-447. 41. G. D. Cristescu, Ann. Physik., 6, (1960), 153. 42. R. Mavrodineanu and R. C. Hughes, Spectrochim. Acta, 19, (1963), 1309-17. 43. V. Trunecek, Z. Chem., 4, (1964), 358. 44. H. Dunken, G. Pforr, and W. Mikkeleit, Z. Chem., 4, (1964), 237. 45. H. Dunken and G. Pforr, Z. Phys. Chem., 230, (1965), 48. 46. C. D. West and D. N. Hume, Anal. Chem., 36, (1964), 412-415. 47. G. Pforr and K Langner, Z. Chem., 5, (1965), 115. 48. G. D. Cristescu and M. Giurgea, Z. Chem., 7, (1967), 360. 49. K. A. Egorova, Zh. Prikl. Spektrosk, 6, (1967), 12-14. 50. A. V. Zvyagintsev, R. V. Mitin, and K. K. Pryadkin, Sov. Phys.-Tech. Phys., 20, (1975), 177-181. 51. N. I. Gonchar, A. V. Zvyagintsev, R. V. Mitin, and K. K. Pryadkin, Sov. Phys.-Tech. Phys., 20, (1975), 407-408. 52. D. C. Liang and M. W. Blades, Spectrochim. Acta, 44B, (1989), 1059-1063. 53. T. D. Hettipathirana and M. W. Blades, Spectrochim. Acta, 47B, (1992), 493. 54. R. E. Sturgeon, S. N. Willie, and V. T. Luong, Spectrochim. Acta 46B, (1991), 1021-1031. 55. E. A. Cordos, S. D. Anghel, T. Frentiu, A. Popescu, J. Anal. Atom. Spectro., 9, (1994), 635-641. 31 56. A. E. Cordos, T. Frentiu, A. M. Rusu, and G. Vatca, Analyst, 120, (1995), 725 57. A. E. Cordos, S. D. anghel, T. Frentiu, E. Darvasi, A. M. Rusu, and A. Simon, FresniusJ. Anal. Chem., 355, (1996), 250-251. 58. A. E. Cordos, S. D. Anghel, T. Frentiu, A. M. Rusu, and L. Bese, Fresnius J. Anal. Chem., 355, (1996), 252-253. 59. A. E. Cordos, S. D. Anghel, T. Frentiu, A. M. Rusu, and M. Ponta, Fresnius J. Anal. Chem., 355, (1996), 254-255. 60. A. E. Cordos, S. D. Anghel, T. Frentiu, A. Simon, and A. Popescu, J. Anal. Atom. Spectro., 14, (1999), 541-545. 61. A. E. Cordos, S. D. Anghel, T. Frentiu, A. M. Rusu, A. Fodor, and M. Ponta, Talanta, 48, (1999), 827-837. 62 J. D. Cobine and D. A. Wilbur, J. Appl. Phys., 22, (1951), 835-841. 63. W. Kessler and U. Jecht, Z Anal. Chem., 198, (1964), 133. 64. W. Kessler and U. Jecht, Z. Physik, 178, (1964), 133. 65. W. Kessler and F. Gebhardt, Glastech. Ber., 40, (1967), 194. 66. W. Kessler, Glastech. Ber, 44, (1971), 479. 67. S. Murayama, J. Appl. Phys., 39, (1968), 5478-5484. 68. S. Murayama, H. Matsumo, and M. Yamamoto, Spectrochim. Acta, 23B, (1968), 513-520. 69. S. Murayama, Spectrochim. Acta, 25B, (1970), 191-200. 70. M. Yamamoto and S. Murayama, Spectrochim. Acta, 23A, (1967), 773. 71. S. Hanaamura, B. W. Smith, and J. D. Winefordner, Anal. Chem., 55, (1983), 2026-32. 32 72. S. Hanaamura, B. W. Smith, and J.D. Winefordner, Can. J. Spectros., 29, (1984), 13-18. 73. J. D. Hwang, W. Masamba, B. W. Smith, and J. D. Winefordner, Can. J. Spectros., 33, (1988), 156-160. 74. B. M. Patel, E. Heithmar, and J. D. Winefordner, Anal. Chem., 59, (1987), 2374 -2377. 75. H. Uchida, P. A. Johnson, and J. D. Winefordner, J. Anal. Atom. Spectro., 5, (1990), 81-85. 76. A. H. Ali, K. C. Ng, and J. D. Winefordner, J. Anal. Atom. Spectro., 6, (1991), 211-13. 77. A. H. Ali, K.C. Ng, & J. D. Winefordner, Spectrochim. Acta, 46B, (1991), 1207 78. A. H. Ali and J. D. Winefordner, Analytica Chim. Acta, 264, (1992), 319-325. 79. A. H. Ali and J. D. Winefordner, Analytica Chimi. Acta, 264, (1992), 327-332. 80. A. M. Pless, B. W. Smith, M. A. Bolshov and J. D. Winefordner, Spectrochim. Acta, 51B, (1996), 55-64. 81. A. H. Ali and J. D. Winefordner, Analytica Chimi. Acta, 264, (1992), 327-332. 82. G. I. Babat, J. Inst. Bee. Eng. (London), 94, (1947), 27. 83. B. Platzer, E. Leitner, G. Knapp, A. Schalk and A. Grillo, Amer. Lab., Aug., (1990), 12-17. 84. R. J. Winslow, Ph. D. Thesis, University of Mssachusetts, 1980. 85. O. Auciello and D. L. Flamm, Eds. Plasma Diagnostics. . Vol. 1. 1989, Academic Press: San Diego. 86. H. R. Griem, in Plasma Spectroscopy, (McGraw-Hill Book Co., NY, 1964) 33 87. H. R. Griem, in Principles of Plasma Spectroscopy, (Cambridge University Press, Cambridge, 1997) 88. R. H. Huddestone and S. L. Leonard, in Plasma Diagnostic Techniques, (Academic Press, NY, New York, 1965) 89. I. M. Podgornyi, Topics in Plasma Diagnostics, (Pienum Press, NY.US, 1971). 90. J. M. Mermet, Spectroscopic Diagnostics: basic Concepts, in Inductively Coupled Plasma Emission Spectroscopy, Ed. By P.W.J.M. Boumans, 1987, John Wiley and Sons: New York. p. 368. 91. P. W. J. M. Boumans, Theory of Spectrochemical Excitation. 1966, London: Hilger and Watts. 92. M. W. Blades, Excitation Mechanisms and Discharge Characteristics, in Inductively Coupled Plasma Emission Spectroscopy, Ed. By P.W.J.M. Boumans, 1987, John Wiley and Sons: New York. 93. A. Thorne, U. Litzen, and S. Johansson, Spectrophysics. 1999, Berlin: Springer. 94. I. Ishii and A. Montaser, Spectrochim Acta, 46B, (1991), 1197. 95. C. I. M. Beenakker, Spectrochim. Acta, 32B, (1977), 173-187. 96. R. L Smith, D. Serxner, and R. Hess, Anal. Chem., 61, (1989), 1103-1108. 97. K. R. Hess, and W. W. Harrison, Anal. Chem., 60, (1988), 691-696. 98. J. W. Carnahan and G. M. Hieftje, Spectrochim. Acta, 47B, (1992), 731-739. 99. K. J. Jones and J. W. Carnahan, Spectrochim. Acta, 47B, (1992), 1229. 100. M. M. Rahman and M. W. Blades, J. Anal. Atom. Spectro., 15 (2000), 1313. 34 Chapter - 2 Experimental One of the objectives of this research was to set up an experimental system to characterize the parallel plate capacitively coupled plasma operating at atmospheric pressure. This chapter describes the experimental system designed for this purpose. A description of this experimental system is also available in previous publications [1,2]. 2.1 Instrumentation The experimental system reported in this thesis was designed for the acquisition of spatially, temporally, and spectrally resolved emission signals from the P P - C C P source. A schematic diagram of the overall experimental system is provided in Fig. 2.1. The main components of this experimental system were: an rf power supply, an rf matching network, the discharge torch, an electrothermal vaporizer (ETV), and an power supply for resistance-heating the ETV. The electrothermal vaporizer was housed in a quartz body, which was connected to the discharge torch by ground joint quartz fitting. A 1:1 image of the plasma discharge was formed on the monochromator entrance slit by using a two lens system and the spectra were detected using a suitable detector (PMT or IPDA). 35 R. F. Generator Impedance Matching Network Monochromator ^ r \ Gas Supply J > r -> Gas Flow Controller J 9-6- Plasma Workhead IPDA IPDA Controller Pyrometer Pyrometer Amplifier Pulse Generator Atomizer Power Supply Digital Oscilloscope Fig.2.1 : Block diagram of the overall experimental system 3 6 2.1.1 The Plasma Source A schematic diagram of the discharge torch is provided in Fig. 2.2. Unless otherwise noted, a 5 cm long torch was fabricated using a 4X4 mm fused-silica tube with a wall thickness of 1 mm (Wilmad Glass Company, Buena, N.J., USA). The rf power was coupled into the plasma using two water-cooled copper electrodes with dimensions of 4 x 1 x 0.5 cm, unless otherwise noted, which were placed on either side of, and in contact with, the outside of the torch, forming a transverse plasma discharge. The plasma torch and the electrodes were housed in an aluminum support structure. Macor™ provided the electrical insulation between the electrodes and aluminum support structure. A helium plasma was initiated and sustained using a 13.56 MHz rf Generator (Model RFX-600, Advanced Energy, Fort Collins, CO, USA), an automatic power tuner (Model ATX-600, Advanced Energy, Fort Collins, CO, USA), and an impedance matching network (Model 5017-000-G, Advanced Energy, Fort Collins, CO, USA) with an output power range of 0 - 620 W. The output of the impedance matcher was coupled to the electrodes through a variable 1-10 \iH inductor to facilitate impedance matching. Although this CCP type of discharge can be sustained using either argon or helium, the latter is more common because, at atmospheric pressure, the voltage required to breakdown and sustain a He-CCP is lower than that required for Ar-CCP. This is due to the fact that at atmospheric pressure, and for the dimensions normally used for CCPs, the ionization rate of a higher magnitudinal orders for helium than for argon [3]. An added advantage of helium is that, because of its high ionization energy (24.6 eV), it has been proven to efficiently excite and ionize certain analytes 37 through resonant Penning and charge transfer processes [4-6]. The helium gas (Union Carbide, Toronto, Ont., Canada) was supplied to the plasma source from a pressurized cylinder. The gas flow rate was regulated using a Side=Trak™ mass flow controller (Sierra Instruments Inc., Carmel Valley, C A , USA). Fig. 2 . 2 : Schematic diagram of the experimental system. L- lens, W-cooling water entry and exit points, l-Macor® insulator, E- copper electrodes, K-clamping knob, T- quartz torch, G- plasma gas inlet, H- atomizer housing, S- sample inlet, A-graphite platform, and C- carrier gas inlet. T and H are connected using a ground joint. The "eye' indicates the direction of observation of emission. 38 2.1.2 Electrothermal Vaporizer Two electrothermal vaporizers (ETVs) were used throughout this research, one constructed from a tantalum sheet and the other from pyrolytic coated graphite. A 1.9 x 1.0 cm tantalum strip vaporizer, with a small circular depression at the center to contain the liquid sample, was constructed from a 0.25 mm thick tantalum sheet (Fig. 2.3 A). The tantalum strip was narrowed to 0.7 cm adjacent to the depression in order to obtain maximum heating at the point of sample deposition. Another geometry of tantalum strip was also tried as shown in Fig. 2.3 B. In this case a longer metal sheet was bent on either side to form an upside down "U" shape. The sheet was narrowed at the lower arms (as is shown in the figure) to obtain the maximum possible electrical resistance and hence, an optimum hot platform. Both strips were attached to two copper rod conductors, surrounded by a water-cooled jacket, which were then connected to an electrothermal power supply. (A) (B) Figure 2.3: Two different designs of the tantalum strip platforms. 39 Unfortunately, the lifetime of each tantalum strip was limited to 15 -20 heating cycles which limits the applicability of this ETV source. To address this problem, the tantalum strip was replaced with a specially designed pyrolytic graphite platform. A 1.5 x 0.8 x 0.2 cm pyrolytic coated graphite platform (Fig. 2.4), with a small circular depression at the center to confine the liquid sample, was custom fabricated by SGL Carbon Group, Ringsdroff Werke, Germany. The platform was secured to two copper rod conductors surrounded by a water-cooled jacket, using molybdenum blocks. The electrothermal vaporizer and the copper rod conductors were housed in a quartz body, which was connected to the discharge torch using ground joint quartz fitting. The plasma torch and the ETV were placed on a vertically moveable plate, so that the position of the emission source was precisely controllable with respect to the monochromator entrance slit. 2.1.3 Optical Configuration A two lens (collimating and focusing) system was employed to collect the light from PP-CCP source and to focus it onto the entrance slit of the monochromator. This simplified the experimental reconfiguration needed when changing the observed wavelength. The refractive index (TI) of a given material is a function of the wavelength (k) of the incident radiation, and the focal length (f) of a lens is a function of the refractive index (n). By combining these two laws, one finds that the focal length of a lens is directly proportional to the wavelength of the incident radiation. By using a collimating and a focusing lens, refocusing (when the observed wavelength 40 was changed) of the light source could be done by simply changing the lens positions, rather than by changing the position of the light source or the monochromator. Figure 2.4: Schematic diagram of the electrothermal (made from graphite) vaporization system employed in PP-CCP source. A- graphite platform, B- circular concavity, c-molybdenum screw, D-molybdenum block, E-copper conductor, H & G - cold water in & out. 41 The light collection efficiency for several commonly used lens configurations was studied by Farnsworth ef al. [7]. A collection system consisting of two plano-convex lenses with their curved surfaces facing each other, was found to be the optimum configuration in terms of collection efficiency. Plano-convex lenses also reduce the spherical aberration. For all studies reported in this thesis, a 1:1 image of the plasma discharge was formed on the monochromator entrance slit by using this arrangement. The collimating lens had a focal length of 250 mm and the focusing lens had a focal length of 150 mm. A He -Ne laser (Melles griot, Carlsbad, CA, USA) was used to help in the alignment of all optical components, including the positioning of the plasma source with respect to the monochromator entrance slit. 2.1.4 Spectral Dispersion and Detection A 1.0 m Czerny-Turner scanning monochromator (Model 2061 Scoeffel-McPherson, Acton, MA, USA) with a 1200 lines mm'1 holographic grating (unless otherwise stated) was used to disperse emission from the plasma source for the studies described in Chapter 4, 5, 6 and 7. The entrance and exit slits were set at 50 um. The signals for the study described in these chapters were detected using an intensified photodiode array (IPDA) detector (Princeton Instruments, Model IRY-700/G/B/PAR, Trenton, NJ, USA); it had a 14-bit dynamic range and was thermoelectrically cooled to below -25°C to minimize dark current. The spectral range for the IPDA was 180-910 nm with a 500:1 variable gain. The detector was controlled using an ST-120 controller (Princeton Instruments, Model S-120A, 42 Trenton, NJ, USA), enabling up to 1000 s integration time, 33 ms array readout time, and external triggering. For the study described in Chapter 3, a 0.35 m Czerny -Turner scanning monochromator (Model 270, Schoeffel-McPherson, Acton, MA, USA) with a 2400 lines mm"1 holographic grating was used to spectrally resolve the light collected from the PP-CCP source. The entrance and exit slits were set at 50 um The signal was detected using a Hammatsu R955 photomultiplier tube (PMT) (Middlesex, NJ, USA). The PMT was operated at 750 V using a high voltage Kepko power supply (Kepko Inc. Flushing, NY, USA). The spectral response of all systems were calibrated using an Electro Optics Associates (Palo Alta, CA, USA) model L-10 quartz -iodine tungsten-filament standard lamp [8]. 2.1.5 Spatially Resolved Intensity Measurements To determine the spatial distribution, the plasma work-head was put on a horizontally moveable platform, which was mounted on a synchronous motor (Type: M062-FD03, Superior electric company, Bristol, CT, USA). A controller from Daedal Inc. (Harrison City, PA, USA, Model No. PC-300) was used to control the platform position, and hence the plasma position, with respect to monochromator entrance slit. The movement of the platform was monitored using a precision displacement indicator gauge (Model 2047-11, Mitutoyo, Japan). Spatial resolution was obtained by placing an appropriate aperture on the collimating lens. In chapter 4, the process of obtaining the desired spatial resolution is discussed in detail. 43 2.1.6 ETV Temperature Measurements The temperature of the electrothermal vaporizers (ETVs) was measured using an optical pyrometer (Ircon Series 1100, Model 11x30, IL, USA) which monitored the radiation emitted from ETV surface. The pyrometer was positioned on a tripod (Manfrotto, Made in Italy) and viewed at the center of the ETV. The output from the pyrometer was amplified (Electrical Service Shop, Department of Chemistry, UBC) and stored into a digital real-time Oscilloscope (Model TDS 380, Tektronix Inc., Beaverton, Oreagone, USA). The digitized data were converted into absolute temperature using calibration data provided by the pyrometer manufacturer. The temperature below the range of the pyrometer was determined by extrapolating the ETV temperature back to the ash temperature. The ash temperature was measured by using a thermocouple (Model 80TK, John Fluke, Everett, WA, USA). The atomization temperature mentioned in this thesis refers to the maximum temperature reached by the ETV during the atomization cycle. 2.2 Data Acquisition and Processing Two data acquisition systems were employed in the study described in this thesis. The data acquisition system used for the study described in Chapter 3 is discussed first. The modified carbon rod atomizer transmits a TTL pulse at the beginning of its atomization cycle, which triggered the analog-to-digital converter (ADC) to commence data acquisition. The output of the PMT was amplified using a current amplifier (Model 426, Keithley, Middlesex, NJ, USA) with a rise time of 0.3 44 ms. The output from the current amplifier was digitized with 12-bit resolution using a 16-channel Model ISC-16 (Santa Barbara, CA, USA) analog to digital (ADC) card with a sampling rate of 100 Hz and was then stored using a PC (Model PB8529V, Packard Bell). The data was then analyzed using Igor (version2.2) graphic software. For the study described in chapter 4, 5, 6, and 7 the data acquisition was controlled using a ST-120 controller (Princeton Instruments, Model S-120A, Trenton, NJ, USA), enabling up to 1000 s integration time, 33 ms array readout time, and external triggering. To manipulate the controller and for data acquisition, CSMA (CCD Spectrometric Multi-channel Analysis) software from Princeton Instruments Inc. was used. Triggering of the ST-120 controller was done via a pulse generator FG-100 (Princeton Instruments, Trenton, NJ, USA) triggered by the ETV power supply. The data was then stored in a PC. Once the data had been acquired, emission intensity profiles were calculated and analyzed using GRAMS (Graphic Relational array Management System) and Igor software. 2.3 Experimental Procedure A 5 to 10 fjL aliquot of sample was placed into the depression of the platform using an Eppendorf 0.5-10 y\ micro pipette (Brinkmann Instruments Co., Rexdale). A 2-3 cm long piece of 1 mm diameter polyethylene tube was fitted to the end of the pipette tip to enable sample delivery. Samples were subject to drying, followed by a "char" stage at a specific temperature with the helium flow on. After a two minute waiting period, radio frequency power was applied, whereupon the plasma 45 spontaneously ignited. Following a plasma stabilization period, the atomization stage was activated, reaching a selected temperature for all determinations. The blank determinations were carried out in the same manner without the analyte. Both peak heights and areas were used during the analysis of the data. Background subtraction was done from either side of the desired peak by baseline subtraction in the spectral dimension, thus removing any contribution to the peak heights from background continuum radiation due to ion-electron recombination. It should be noted here that the operating parameters and experimental procedures used in the studies reported in this thesis are given in more detail in the experimental sections of the appropriate chapters. 2.4 References 1. M. M. Rahman and M. W. Blades, Spectrochim. Acta, 52B, (1997), 1983- 93. 2. M. M. Rahman and M. W. Blades, Spectrochim. Acta, 55B, (2000), 327-38. 3. M. W. Blades, Spectrochim. Acta,, 49B, (1994), 47. 4. S. K. Chan and A. Montaser, Spectrochim. Acta, 40B, (1985), 1467. 5. P. G. Brandl and J. W. Carnahan, Spectrochim. Acta, 49B, (1994), 105-115. 6. P. G. Brandl and J. W. Carnahan, Appl. Spectrosc, 49B, (1995), 1781 -1788. 7. P. B. Fransworth, B.W. Smith, and N. Omenetto, Spectrochim. Acta, 45B, (1990), 1151. 8. R. Stair, W. E. Schneider, and J. K. Jackson, Appl. Optics, 2(11), (1963), 1151. 46 C h a p t e r - 3 Temperatures and Some Analytical Figures of Merit 3.1 Introduction An investigation of the fundamental properties of plasma sources helps lead to an improved understanding and aids in the development, optimization, and comparison of different kinds of plasmas as spectroscopic sources. Common fundamental properties that are investigated include electron number density, temperature, specific heat, viscosity, and conductivity [1]. Usually, spectroscopic methods are preferred for these investigations as they are non-invasive, the methodology is fairly well developed, and, if the temperature is very high, may be the only applicable method. Temperature in high frequency inert gas plasmas has frequently been studied in an effort to understand the excitation mechanisms, which give rise to the various population states of introduced analyte species. Temperature is an important fundamental property for CCP discharge. A high temperature provides a "robust" plasma environment with fairly low matrix interference effects, and, in addition, makes the discharge able to atomize and excite the analyte resulting in an intense emission signal and, consequently a high analytical sensitivity. In this chapter, the electronic (Text) and rotational (T,^ ) temperatures are determined and discussed for a parallel plate capacitively coupled plasma operating at atmospheric pressure. In addition, the effect of plasma gas flow rate and rf power on analyte emission signals, background signals, and limits of detection for lead and silver are also presented in this chapter. 47 3.2 Experimental 3.2.1 Instrumentation A schematic diagram of the discharge torch and electrothermal vaporizer was already provided in Fig. 2.1. A block diagram of the overall experimental system used in this section of the study is provided in Fig. 3.1. The main components of the PP-CCP assembly were a torch and a tantalum-strip electro -thermal vaporizer used as the sample introduction system. The 5 cm long torch was fabricated from 4X4 mm fused-silica tube with a wall thickness of 1 mm ( Wilmad Glass Company, Buena, N.J, USA ). The RF power was coupled into the plasma using two water cooled copper electrodes with dimensions of 4 x 1 x 0.5 cm, which were placed on either side of, and in contact with, the outside of the torch, forming a transverse plasma discharge. A 1.9 x 1.0 cm tantalum strip vaporizer, with a small depression at the center to contain the liquid sample, was constructed from a 0.25 mm thick tantalum sheet. The tantalum strip was narrowed to 0.7 cm adjacent to the depression in order to maximize heating at the point of sample deposition. The strip was attached to two copper rod conductors, surrounded by a water-cooled jacket, which were connected to an electrothermal power supply. The electrothermal vaporizer and the copper rod conductors were housed in a quartz body, which was connected to the discharge torch by a ground joint quartz fitting. The plasma was initiated and sustained using a 13.56 MHz rf generator (Model RFX-600, Advanced Energy, Fort Collins, CO, USA), an automatic power tuner (Model ATX-600, Advanced Energy, Fort Collins, CO, USA), and an impedance matching network (Model 5017-000-G, Advanced Energy, Fort Collins, CO, USA) with an output power range of 0 - 620 W. The output of the impedance matcher was coupled to the electrode through a variable 1-10 \iH inductor to facilitate impedance matching. The tantalum strip was heated using a carbon rod atomizer (Model 61 Varian Techtron, Melbourne, Australia) which was modified to provide a synchronization trigger. A 0.35 m Czerny -Turner scanning monochro -48 mator (Model 270, Acton, MA, USA) with a 2400 lines mm1 holographic grating was used to acquire spectra. The entrance and exit slit were set at 50 urn. A 1:1 image of the plasma discharge was formed on the monochromator entrance slit by using a two lens system consisting of two piano - convex fused silica lenses with their curved surfaces facing each other. The collimating lens had a focal length of 250 mm and the focusing lens had a focal length of 150 mm. The signal was detected using a Hammatsu R955 photomultiplier tube (Middlesex, NJ, USA). The output of the PMT was amplified using a current amplifier (Model 426, Keithley, Middlesex, NJ, USA) with a rise time of 0.3 ms. The output from the current amplifier was digitized with 12-bit resolution using a 16-channel Model ISC-16 (Senta Barbara, CA, USA) analog to digital (ADC) card with a sampling rate of 100 Hz, and then stored using a PC (Model PB8529V, Packard Bell). The spectral response of the system was calibrated using an Electro Optics Associates (Palo Alta, CA, USA) model L-10 quartz -iodine tungsten-filament standard lamp [2]. Lead and silver solutions were prepared by dissolving analytical grade Pb(N03)2 and AgN03 (BDH, Toronto, Canada) in 1% HN03 solution. Analytical grade ferrocene powder (Aldrich Chemical Company, Inc., USA) was used directly. 3.2.2 Procedure 3.2.2.1 Helium Excitation Temperature The He I excitation temperature was determined from a Boltzmann plot using four neutral atom lines ranging in wavelength from 388.87 to 492.19 nm. Corrections were made for spectrometer throughput and detector response. In order to implement background correction, steady state intensity measurements were made at the center of each line and at positions approximately 1 A to either 49 R. F. Generator Gas Supply Impedance Matching Network Gas Flow Controller i Monochromator -< Plasma Workhead PMT J Pyrometer Amplifier Pyrometer Current Amplifier Atomizer Power Supply ADC Computer Fig. 3.1 : Block diagram of experimental system. 50 side of the peak. The mathematical equation used for this temperature measurement has already been discussed in section 1.5.2. The physical data used for He I line are summarized in Table 3.1 [3,4] Table 3.1: Spectral data for Hel lines [3, 4] X. ( n m ) E e x c (e V ) 388.87 23.01 0.1934 396.47 23.74 0.051 438.79 24.04 0.1290 447.15 23.74 1.1059 492.19 23.74 0.3595 501.57 23.09 0.1514 3.2.2.2 Pb and Fe Excitation Temperatures Excitation temperatures were calculated using iron and lead as thermometric species. The iron excitation temperature was derived from a set of eight Fe I lines ranging from 3700 A to 3780 A. Spectral data used for this purpose, as summarized in Table 3.2, was obtained from Wiese et al. [4] and the NIST Web Site [5]. A uniform flow of iron into the plasma torch was maintained by introducing ferrocene vapor into the plasma carrier gas. This was accomplished by placing a small amount of ferrocene into the test tube and allowing the plasma gas to pass 51 CO E OJ Q . O i n CM I 03 3 O O 5= (0 OT D_ JO o CT Q- E CL fc c - 1 I i I ! 5 8 (0 Q . •<2 X3 £ l CM CO <D k_ g> u_ /4!sueju| 9AiiB|ey 52 through the tube, carrying sublimed ferrocene to the plasma torch. The test tube and the transfer line to the plasma torch were wrapped with heating tape to maintain a uniform temperature of 80 °C during the experiment. An example of the iron spectrum at 250 W plasma forward power is provided in Fig. 3.2. Table 3.2: Spectral data for Fe I lines [4, 5] X (nm) (eV) 371.99 3.33 0.3719 373.49 4.18 2.0746 373.71 3.37 0.2675 374.56 3.40 1.1693 374.82 3.42 0.0964 374.95 4.22 1.4491 375.82 4.26 0.9396 376.38 4.28 0.5776 The lead excitation temperature was derived from the two line method. The wavelengths used for this experiment were 280.19 and 283.3 nm. Lead was introduced to the system as a solution droplet on the tantalum platform atomizer, and was transported by the plasma gas into the plasma after being vaporized. 53 3.2.2.3 Rotational Temperature Rotational temperatures have been determined using the molecular spectra of two different species, N* and OH which are created in the plasma. N* rotational temperatures were calculated by measuring emission line intensities using the band head at 391.4 nm from the R branch of the B22l -*X 22g first negative system of N* (0 - 0). From the experimental emission line intensities le m a straight line may be obtained by plotting in ^ v s - (K"+lXK"+2) assuming a Boltzmann distribution between the rotational levels [6] where a=1 for even lines and a=2 for odd lines, owing to even-odd intensity alternation [7]. The slope of the plot is -B'hc/kT where B' is the rotational constant of the upper state. For this study measurements were made using seven lines spanning R6 to R18 in the 3870 -3920 A region. The assignments of the rotational lines as a function of the quantum number K" of the lower state of the transition against wavelength was taken from different sources [8-10]. Figure 3.3 is a typical spectrum obtained for N* in the CCP. In the case of the OH band, five main branches (O, P, Q, R, and S ) can be observed from A 2 2+ -+ X 2n transition with a total of twelve branches : 0 1 2, P1 t P2, P12, Qv Q^, Q2, Q 1 2, R^  R,v r\, and S21[10]. Line assignments have been dealt with in detail by Dieke and Crosswhite [11] and transition probabilities can be found in cited texts [11,12]. Using the (0 - 0) band we have selected six lines from Q, branch spanning Q,2 to QJ0 in the 3070 - 3115 A region. From the experimental emission line intensities (lem), assuming a Boltzmann distribution between the I A rotational levels, a plot of l o g - 5 3 5 - vs. E yields a straight line with a slope of -.A. hc/kT .^ An example of a typical PP-CCP spectrum for OH is shown in Fig. 3.4. 54 A)!SU9}U| U0ISSJUJ3 S A i i e s y 56 3.3 Results and Discussion 3.3.1 Excitation Temperature Calibration of spectral response as a function of wavelength allowed the use of a wide wavelength range, from 3888.7 A to 4921.9 A, for the measurement of He excitation temperature. Values of gf were taken from several sources [3, 4]. For the transitions used in this work, the uncertainties in the transition probabilities estimated by Martin and Wiese [4] are ±3 %. A "typical" Boltzmann plot for He I is shown in Fig. 3.5A 12.5 3.0 I F " 1 -3.2 T Fe I Excitation Energy, eV 4 3.6 3.8 4.0 1 1 1 I 1 1 1 I 1 1 1 I 4.2 1~T~r 4.4 12.0 (A) § 11-5 11.0 10.5 10.0 J fct I I I L 22.5 I L 23.0 _ J i i i i I 24.0 24.5 Fig. 3.5: Boltzmann plot for (A) He I and (B) Fe I line for a 200 W plasma with a gas flow rate of 0.2 L min'1. Correlation coefficient, r(A)=0.95 and r(B)=0.99. 57 The effect of plasma forward power on the measured He I excitation temperature is shown in Fig. 3.6. The uncertainty associated with each data point is the standard deviation for four independent measurements and varied from 0.3% for 150 W forward power to maximum of 0.9% for 100 W plasma forward power. The figure indicates that increasing the plasma forward power from 100 to 250 W increases the He excitation temperature from 3255 K to 3900 K. The values are similar to those measured for a FAPES (Furnace Atomization Plasma Emission Spectroscopy) source [13] and a He microwave induced plasma (MIP) [14] source operating at 100 W forward power and 1.5 L min1 gas flow rate, but are lower than the 6100 K reported by Long et al. [15] for a He MIP (operating at 150 W forward power and 1 L min1 gas flow rate), and significantly lower than 9300 K He excitation temperature reported by Falk et al. [16] for a FANES (Furnace Atomic Non-thermal Emission Spectroscopy) source operating at low (75 torr) pressure. 5000 r -* 4500 I...--""" H 4000 c o 03 0 3500 3000 Fe Excitation Temp. Pb Excitation Temp. He Excitation Temp. 2500 j l_ 100 J i 150 J i i i I 200 J i . . l_ 250 Plasma Forward Power, W Fig. 3.6: Influence of applied rf power on measured excitation temperatures 58 The iron excitation temperature was measured using eight Fe I emission lines ranging from 3700 A to 3780 A. The uncertainties in the transition probabilities are within ±10 % except for the 3737.1 A line where the uncertainty is ±25 % [4, 5]. It should be noted that the Fe levels used for this temperature measurement are levels whose excitation energy is below 5 eV. Blades et al. have observed that in the ICP, the excitation temperature for Fe I yields no unique value but is dependent upon the excitation energy [17, 18]; the higher the energy level, the higher the calculated excitation temperature. This situation has been interpreted as being due to significant radiative contributions to excited state de-population in which the radiative de-excitation rate becomes comparable to the collisional de-excitation rate for lower atomic levels [17], causing departures from LTE. Since both the ICP and PP-CCP are relatively low frequency atmospheric pressure plasma sources, a similar situation may exist for the PP-CCP. A typical Boltzmann plot for Fe I is shown in Fig. 3.5B, and the effect of plasma forward power on Fe I excitation temperature is shown in Fig. 3.6. The uncertainties indicated by the error bars on each temperature are the standard deviations for five independent temperature measurements and varied from ± 1.3 % to ± 3.2 % for 250 W and 125 W plasma forward power respectively. Figure 3.6 shows that the Fe I excitation temperature changes from 4300 K to 4890 K over the power range of 100 - 250 W. At a power of 100 W, Sturgeon ef al. [13] reported a temperature of 2920 K whereas Blades etal. [19] reported a temperature of 4200 K at a power of 50 W in FAPES source. The Fe I excitation temperature found in capacitively coupled microwave plasma is in the range of 3000 K - 4500 K [14, 20, 21]. These values are similar to those obtained from the PP-CCP source. When compared with the CMP, which operates in the power range of 700 - 1000 W, the PP-CCP has a similar temperature at a much lower power. The reason behind this may be that the radiative and convective heat losses from a PP-CCP source is 59 lower than that from CMP or, that the field strength is higher in PP-CCP. This may also be a combination of both factors. Excitation temperatures using lead were also determined using the two line method in the PP-CCP source. The lines used in this experiment were 2801.9 A and 2833.0 A. Spectral data used for this purpose was obtained from Corliss and Bozman [22] and are summarized in Table 3.3. Each data point was the average of four replicates of 25 ng of Pb in 1 % HN03. Based on the precision of the measured intensities, the uncertainties associated with each measured temperature is from ±1.2 % to ± 3.4 % for 250 and 100 W plasma forward power, respectively. The effect of plasma forward power on the Pb excitation temperature is shown in Fig. 3.6. The excitation temperature changes from 3540 ± 120 K to 4500 ± 55 K when the plasma forward power is increased from 100 to 250 W. Pb excitation temperatures from another capacitively coupled source, FAPES, reported by Hettipathirana [23], are 3990 ± 80 K, 4560 ± 100 K, and 4840 ± K for 20, 50, and 100 W plasma forward power, respectively. Table 3.3: Spectral data for Pbl lines [22] X(nm) E e x c (eV) gA (108 s 1 ) 280.19 5.7439 43 283.31 4.3749 1.8 60 3.3.2 Rotational Temperature It is generally accepted that the rotational temperature (Trot) is indicative of the gas temperature. The rotational temperature, T r o t, was determined using both N* and OH lines. Emission line intensities with a band head at 391.4 nm from the R branch of first negative system of N2 (0-0) were used for Nj rotational temperature. The uncertainties, determined from five independent temperature measurements, are from ± 2.3 % to ± 8.0 % for 225 and 100 W plasma forward power respectively based on the precision of four measurements. The Q t branch of the (0-0) transition of the OH radical in the region of 3070 A to 3115 A was used to determine the OH rotational temperature. The uncertainties associated with each data point were from ±2.1 % to ± 6.0 % for 275 and 100 W plasma forward power, respectively. 1100r-1000 r 900 Q. E 0) t 800 «o c . 0 I 700 o rx + £ 600 500 400 b: N24" Rotational Temp. OH Rotational Temp. -i i_ _i 1 i_ _i 1 1 i_ 50 100 150 200 Plasma Forward Power, W 250 300 Fig. 3.7 Effect of the plasma forward power on both N 2 + and OH rotational temperatures with a fixed gas flow rate of 0.2 L min'1. 61 The effect of plasma forward power on both N* and OH rotational temperature is shown in Fig. 3.7. It can be seen from the figure that Trot has only a weak dependence on rf power. The Trot measured from N* increases from 845 K to 956 K whereas it increases from 828 K to 911 K for OH rotational temperature by increasing the plasma forward power from 75 to 275 W. These values are below the reported Trot values from different He CMP sources, which generally fall in the range of 1300 K - 2500 K [14, 20, 21, 24-26] yet agree with LeBlanc's [27] measurements on a FAPES source. The temperature obtained from N* is higher by about 6 % compared with that from OH radical. LeBlanc pointed out that N* is excited by a charge transfer reaction between N 2 and He*, which causes the higher rotational temperature for N* than that for OH radical, which is probably excited collisionally. 3.3.3 Analytical Figures of Merit In order to determine the emission characteristics of analytes in the PP-CCP source, the effect of plasma gas flow rate and plasma forward power on analyte emission, as well as the noise in the measurements were investigated. Lead and silver were used as analytes for this study. Independent control over vaporization and excitation provides a mechanism for independent optimization of both processes in PP-CCP source. The plasma gas not only acts as a plasma support gas, but also carries the analyte species from the vaporizer into the plasma, where excitation and emission take place. The residence time of analyte atoms is determined mainly by the combination of plasma gas flow rate, diffusion effects and the length of torch. The effect of changes in plasma gas flow rate on emission intensity of Ag I 328.1 nm and Pb I 283.3 nm were studied at 250 W plasma forward power and are shown in Fig. 3.8 and 3.9 along with the background intensity, signal - to - noise ratio, and signal - to -62 0.0 0.2 0.4 0.6 0.8 1.0 25 20 15h 1 0 r -ok 0.0 - • - Signal-to-Noise Ratio • - o - Signal-to-BackgroundRatio 0.2 0.4 0.6 0.8 Gas Flow Rate, L min"1. 1.0 r- 0 Fig. 3.8: Effect of gas flow rate on (A) lead analyte and background and (B) SNR and SBR for a 250 W plasma for a deposition of 1.0 ng of lead. 63 03 c g> co c o "I E LU < i <-» CD CO 54 24 04 (A) ©--tr -•-Analyte Intensity -G- Background Intensity A - Background Noise •©-0.0 0.1 -0 A 0.5 0.4 \- 0.3 r- 0.2 \- 0.1 0.0 CD o CQ o c Q . CO CQ' 0) g C/>" CD 0.2 0.3 0.4 16-1 1 4 -o 1 2 -03 rx CD 1 0 -CO O 8 -Z 6 i 03 6 -c O) CO 4 -2 -0 -Fig. 3.9: 0.0 ~ T ~ 0.1 -•-Signal-to-Noise Ratio -G- Signal-to-Background Ratio \- 80 0.2 0.3 Gas Flow Rate, L min"1. 0.4 cp ca' 60 I CD 0) 4 0 * o c 20 S r- 0 Effect of gas flow rate on (A) silver analyte and background and (B) SNR and SBR for a 250 W plasma for a deposition of 1.0 ng of silver. 64 background ratio. As indicated in Fig. 3.8A and 3.9A, the emission intensity for both lead and silver are affected by changes in the gas flow rate. This is mainly due to a change in the residence time as well as the radiative and mass transport properties of the plasma. Initially the Pb intensity increases with gas flow rate, reaching maximum at 0.2 L min 1 , then flattening out up to 1.0 L m in 1 gas flow rate. The background intensity and the noise in the background increases very slightly with increasing gas flow rate. The phenomenon of increasing noise can be explained by turbulence in the gas flow inside the torch. In general, the lower the gas flows rate the lower the noise. The silver emission intensity shows a higher rate of increase with increasing flow rate, but the background intensity is relatively independent of gas flow rate up to 0.2 L m in 1 and then slowly decreases with increasing flow rate, yielding higher signal-to-background ratio, as shown in Fig. 3.9B. In the case of lead analyte, the background intensity remains almost constant up to 0.3 L min"1 gas flow rate and then increased with increasing flow rate resulting lower signal-to-background ratio which is illustrated in Fig. 3.8B. It can also be seen from Fig. 3.8B and 3.9B that the noise in the signal for Ag emission increases to a greater extent with increasing gas flow rate than that for Pb. When the plasma is operated at a 250 W forward power, the maximum signal-to-noise ratio obtained for silver and lead is at 0.15 L m in 1 and 0.2 L m in 1 gas flow rate, respectively. Plasma forward power is another important parameter in the optimization of experimental conditions. The effect of changes in plasma forward power on emission intensity of both Ag I 328.1 nm and Pb I 383.3 nm were studied using helium gas flow rates of 0.15 L min"1 and 0.2 L m in 1 respectively. The results over the power range from 50 to 300 W are shown in Fig. 3.10 and 3.11 for lead and silver analyte respectively. The effect of power is less pronounced for lead emission compared with silver emission. It can be seen from Fig. 3.11 A that the relative emission intensity for Ag increases slowly from 50 to 150 W and then increases sharply from 150 to 300 W, whereas the background emission shows a steady increase from 50 to 300 W plasma forward power. The noise in the silver 65 Plasma Forward Power, W Fig. 3.10: Effect of rf power on atomic emission behavior of Pb analyte: (A) emission signal and (B) Signal-to-Background and Signal- to- Noise Ratios at 0.2 L min 1 gas flow rate for a deposition of 1.0 ng of Pb. 66 50 100 150 200 250 300 1 1 1 1 1 T" 50 100 150 200 250 300 Plasma Forward Power, W. Fig. 3.11: Effect of rf forward power on atomic emission behavior of Ag analyte: (A) emission signal and (B) Signal-to-Background & Signal-to-Noise Ratios at 0.2 L m in 1 gas flow rate for a deposition of 1.0 ng of Ag. 67 emission signal also increases with increasing power resulting in a maximum signal-to-noise ratio at 250 W plasma forward power, as indicated in Fig. 3.11B. In the case of Pb, initially the emission intensity increases with increasing power, reaching a maximum at 200 W and then decreases above this level. This phenomenon of decreasing signal intensity is likely the result of an increase in ionization with increasing plasma forward power. This will be further discussed in chapter 6 of this thesis. The maximum signal-to-noise ratio for Pb is also obtained at 250 W whereas maximum signal-to-background ratio is obtained at 200 W plasma forward power. With the optimized value of plasma gas flow rate and plasma forward power, the detection limit, based on 3 times the standard deviation of the background, for lead and silver were determined to be 0.33 ng and 24 pg respectively. Our Pb detection limit is lower than the value found in the C M P source which is 2 ng [28], but higher than values (2.4 - 21 pg) reported for the F A P E S source [29-31]. This is also higher than the values (0.84 - 5.7 pg) found for the MIP [32, 33]. The silver detection limit is comparable with the reported value of 12 pg and 210 pg [28, 34] in C M P source and 1 .6 -10 pg [35-37] in MIP source, whereas our value is slightly higher than the values (0.35 -1 .2 pg) found in F A P E S source [29, 30, 38]. 3.4 Conclusions Atomic emission intensities from helium, lead and iron, and molecular emission intensities from OH and N 2 + were measured in an attempt to calculate the excitation and rotational temperatures of P P - C C P source. Lead excitation temperatures were calculated using the two line method by measuring the Pb I (280.19 and 283.31 nm) lines at different rf powers. Helium and iron excitation temperatures were determined from a Boltzmann plot using four (for helium) and seven (for iron) neutral atom lines. All three excitation temperatures were found to increase monotonically with an rf power over the range of 100 to 250 W. The 68 measured rotational temperatures from OH and N 2 + were also found to increase monotonically with rf power over the range of 75 to 250 W. The difference in measured excitation and rotational temperatures suggests the absence of local thermodynamic equilibrium (LTE) conditions in the PP-CCP. This chapter has also demonstrated that some important fundamental and analytical characteristics of the parallel plate He plasma are similar to those of a microwave induced plasma however, we feel that the CCP offers several operational advantages. First, the plasma is self ignighting and hence does not require a tesla discharge. Second, because a tuned cavity is not required, the dimensions of the CCP can be quite flexible. Third, compared to a capacitive microwave plasma, the CCP exhibits similar excitation temperatures but at significantly lower rf power. Unfortunately, the lifetime of the tantalum strip is limited to 15 - 20 heating cycles which limits the applicability of this ETV source. To address this problem the tantalum strip was replaced with a specially designed pyrolytic graphite platform and used for the remainder of the study, which is described in the next chapters. 3.5 References 1. W. Lochte-Holtgreven, Evaluation of Plasma Parameters. W. Lochte -Holtgreven, Ed., Chap.3, (John Wiley and Sons, New York 1968). 2. R. Stair, W. E. Schneider, and J.K. Jackson, Appl. Optics, 2(11),(1963), 1151. 3. L. Goldberg, Astrophysical Journal, 90, (1939), 414. 4. W. L. Wiese and G. A. Martin, Wavelength and Transition Probabilities for Atoms and Atomic Ions, J. Reader, Ed., p. 395, (Washington, U.S. Dept. of Commerce, National Bureau of Standards), (1980) 69 5. NIST Website, Database version 1.3 http://physics.nist. gov/physRefData/ ASD1/nist-atomic-spectra.html. 6. I. Ishii and A. Montaser, Spectrochim. Acta, 46 B (8), (1991), 1197. 7. G. Herzberg, Spectra of Diatomic Molecules, (Van Nostrand, NJ, 1950). 8. W. H. J . Childs, Proc. R. Soc. London Ser., A137, (1932), 641. 9. D. Coster and F. Brons, Z Physik, 73,(1932), 747. 10. J . M. Mermet, Spectroscopic Diagnostics: Basic Concepts, in Inductively Coupled Plasma Emission Spectroscopy, P. W. J . M. Boumans, Ed., (John Wiley & Sons : New York. 1987). 11. G. H. Deike and H. M. Crosswhite, J. Quant. Spectros. Radiant. Transfer., 2, (1962), 97-199. 12. A. K. Hui, M. R. McKeeven, and J . Tell inghuisen, J. Quant. Spectros. Radiant. Transfer., 21, (1979), 387. 13. R. E. Sturgeon, S. N. Willie, and V.T. Luong, Spectrochim. Acta, 46B, (1991), 1021. 14. K. Tanabe, H. Haraguchi, and K. Fuwa, Spectrochim. Acta, 38B, (1983), 49 15. L. D. Perkins and G.I. Long, Appll. Spectroscopy, 43(3), (1989), 499. 16. H. Falk, E. Hoffmann, and C. Ludke, Prog. Anal. Spectros, 11, (1988), 417. 17. Z. Walker and M. W. Blades, Spectrochim. Acta, 41 B(8), (1986), 761. 18. M. W. Blades, B. L. Caugh l i n , Z. H. Wa lke r , and L L Prog. Anal. Spectros., 10, (1987), 57. 19. T. D. Hettipathirana and M. W. Blades, Spectrochim. Acta, 47B, (1992), 493. 20. M. H. Abdallah and J . M. Mermet, Spectrochim. Acta, 37B(5), (1982), 391. 21. W. R. L. Masamba, A. H. Ali, and J . D. Winefordner, Spectrochim. Acta 47B(4), (1992), 481. 22. C. H. Corl iss and W. R. Bozman, Experimental Transition Probabilities for Spectral Lines. Vol. N B S Monograph 53. 1962: US Department of Commerce. 289. 23. T. D. Hettipathirana, Ph. D. Thesis, 1993, University of British Columbia. 70 24. P. Fleitz and C. Seliskar, Appll. Spectrosc, 41 , (1987), 679-682. 25. J . M. Workman, P. A . F l e i t z , H. B. Fannin, J . A . Caruso , and C. J . S e l i s k a r . , Appll. Spectrosc, 42, (1988), 96. 26. B. M. Spencer, B. W. Smith, and J . D. Winefordner, Appll. Spectrosc, 48(3), (1994), 289. 27. C. W. LeBlanc, Ph. D. Thesis, 1996, University of British Columbia. 28. A. H. AN, K. C. Ng, and J.D. Winefordner, J. Anal. Atom. Spectro., 6, (1991), 211. 29. R. E. Sturgeon, S. N. Willie, V. T. Luong, and S. S. Berman, Anal. Chem., 62, (1990), 2370-2376. 30. R. E. Sturgeon, S. N. Willie, V. T. Luong, and S. S. Berman, J. Anal. Atom. Spectro., 6, (1991), 19. 31. R. E. Sturgeon, S. N. W i l l i e , V . T. Luong, and R. K. Markus, Spectrochim. Acta, 48B, (1993), 893. 32. A. T. Zander, R.K. Williams, and G. M. Hieftje, Anal. Chem., 49, (1977), 2372. 33. A. T. Zander and G. M. Hieftje, Anal. Chem., 50, (1978), 1257. 34. A. H. Ali and J . D. Winefordner, Analytica Chimica Acta, 264, (1992), 327. 35. H. Kawaguchi, M. Hasegawa, and A. Mizuike, Spectrochim. Acta., 27B, (1972), 205-210. 36. H. Kawaguchi and B. L Val lee, Anal. Chem., 47, (1975), 1029. 37. C. I. M. Beenakker, B. Bosman, and P. W. J . M. Boumans, Spectrochim. Acte.,33B, (1978), 373-381. 38. R. E. Sturgeon, S. N. Willie, V. T. Luong , and J . G. Dunn, Appll. Spectrosc, 45,(1991), 1413-1418 . 71 C h a p t e r - 4 Spatial Distribution of Plasma Background Species 4.1 Introduction The spatial distribution of plasma species and their response to changes in operating parameters is largely unknown for the PP-CCP source. This being the case, we have measured the spatially resolved emission for He, N 2 +, and OH (which are the major plasma background species) while varying the applied plasma forward power and the gas flow rate. From these measurements some spatially resolved spectroscopic temperatures were also calculated. Obtaining spatially resolved distributions of plasma gas species is important because the strong spatial inhomogeneity of the plasma can affect the quality of the analytical signal. Thus, this information is useful not only for diagnostic purposes, but it may also serve as an aid in determining the optimal observation zone for analytical determinations. 4.2 Experimental 4.2.1 Instrumentation A complete description of the experimental system employed to acquire atomic and molecular emission signals was given in Chapter 2 of this thesis. In this study a 3600 lines mm 1 holographic grating was used in a 1.0 m Czerny-Turner scanning monochromator (Model 2061 Scoeffel - McPherson, Acton, MA, USA) to 72 disperse emission from the plasma. The entrance and exit slits were set at 50 \im. A 1:1 image of the plasma discharge was formed on the monochromator entrance slit using a two lens system consisting of two plano-convex fused silica lenses with the curved surfaces facing each other. The signal was detected using an intensified photodiode array (IPDA) detector (Model IRY-700/G/B/PAR, Princeton, NJ). The detector was controlled with a ST-120 controller from the same supplier. To manipulate the controller and for data acquisition, CSMA (CCD Spectrometric Multichannel Analysis) software from Princeton Instruments, Inc. was used. To measure the spatial distribution of emission intensity, the plasma work-head was put on a horizontally moveable platform, which was mounted on a synchronous motor (Type: M062-FD03, Superior Electric Company, Bristol, CT, USA). A controller from Daedal Inc. (Harrison City, PA, USA, Model No. PC-300) was used to control both the platform position and the plasma position, with respect to the monochromator entrance slit. 4.2.2 Procedure The spatial distribution of plasma species (e. g. He I, N2+) was measured for a PP-CCP source using a 4 cm long electrode. After trying several F/numbers by using different diameter apertures in front of the collimating lens (with a focal length of 250 mm), F/80 was decided upon as it yielded the best balance between light throughput and spatial resolution. Since the electrode was 4 cm long and the lens, with the aperture, was positioned to focus the center of the image onto the monochromator's entrance slit, resulting in a spatial resolution of 0.25 mm. The 73 width of the entrance slit of the monochromator was set to 50 nm. Throughout the experiment, reflected power was tuned manually (the matching network was not permitted to auto tune) and kept at 1, 2 and 3 W for 50, 100 and 200 W forward power, respectively. The measurement of rotational temperature in PP-CCP source has been discussed in detail in Chapter 3. In short, it is determined using the molecular spectra of the N* and OH, which is formed in the plasma. N* rotational temperatures were calculated by measuring emission line intensities using the band head at 391.4 nm from the R branch of the B22l ~^X22l first negative system of N* ( 0 - 0 ). From the experimental emission line intensities, le m, a straight line may be obtained by plotting In ^ ° — vs. (K" + l)(K" + 2), assuming a Boltzmann a2(K +1) distribution between the rotational levels [1], where a = 1 for even lines and a = 2 for odd lines, owing to even-odd intensity alteration [2]. The slope of the plot is -B'hc/kT where B' is the rotational constant of the upper state. Measurements were made using seven lines, spanning R6 to R18, in the 3870 - 3920 A region. The assignments of the rotational lines, as a function of the quantum number K" of the lower state of the transition against wavelength, were taken from a reliable source [3]. In the case of the OH band, five main branches (O, P, Q, R, and S ) can be observed from A 2 2 + - » X 2 f l transition, with a total of twelve branches : 0 1 2 , P,, P2, P 1 2, Q 1 t Q 2 1 , Q 2, Q 1 2 , R„ R2 1, R2, and S 2. Using the (0 - 0) band we have selected six lines from Q, branch spanning Q,2 to 0,10 in the 3070 - 3115 A region. From the experimental emission line intensities (lem), assuming a Boltzmann distribution 74 between the rotational levels, a plot of log-22— vs. E yields a straight line with a A slope of -hc/kTrot. 4.3 Results and Discussion 4.3.1 Spatial Distribution The spatial distributions of He I (388.86 nm), OH (band head at 310 nm) and N 2 + (band head at 392.4 nm), for a 150 W plasma with a gas flow rate of 0.5 L min1, have been measured and are illustrated in Fig. 4.1. It is evident from the plot, that the He atomic emission signal maxima is observed adjacent to the torch wall. As expected, the spatial distribution of N 2 + molecular emission was similar to that of He atomic signal, where the plasma has the greatest excitation ability, or the highest excitation temperature, as defined by the maxima in the emission from helium. From this observation, it is possible to assert that the center of the analytical region for this plasma source is different than the geometrical center of the discharge, where the He I emission signal was undetectable. In contrast to He I and N 2 +, the molecular OH behaved differently in the discharge. The OH emission intensity increases as we move from the edge with maxima at the center of the discharge. The OH emission intensity remains fairly constant at the central region (± 1.0 mm from center) of the discharge, and hence represents a uniform distribution of emission from the OH molecule. In general, the emission intensity from species in spectroscopic plasmas is an exponential function of excitation temperature and a linear function of the population 75 density of the emitting species. Since the population of the helium atom, at a given discharge volume and gas flow rate, is fairly constant, the emission profile from helium is representative of the excitation temperature of this source. Rahman and Blades reported that the excitation temperature in the PP-CCP source varies with the change of applied forward power [4]. An increase or decrease of applied rf power results in changes in the available energy in the discharge, as well as the density of the colliding particles and the number of collisions. These changes, in turn, alter the excited state population of N2 + and OH species, which no doubt cause a change in their relative emission intensities. 2.5 -A in c 2 0 CD c 1.0 E LU i £ 0.5 o.o4 ...o 4-... • o He I • • • • « • • • • OH -©- N 2 + I I I I I I I I I I I I I I I I I I I I I I I 1 I I I I I I I I I I I I I I I I I 2.0 -1.5 -1.0 -0.5 0.0 0.5 1.0 1.5 2.0 Spatial torch position, mm Figure 4.1: The spatial distribution of plasma species in PP-CCP source for a plasma of 150 w forward power and at 0.5 L min1 gas flow rate. 76 Figure 4.2A, 4.2B and 4.2C illustrate the effects of plasma forward power on the spatial distribution of He I, OH, and N2 + species at a given gas flow rate of 0.1 L min1. The "0.0" in the X-axis represents the horizontal center of the torch. As shown in Fig. 4.2A and Fig. 4.2B, at the measured spatial resolution, the distribution of He I and OH emission in the discharge remains unchanged with an increased application of rf power, from 75 to 200 W. However, in Fig. 4.2C, the signal maxima for N2 + molecular emission shifts towards the discharge wall as we increase the applied plasma power from 75 to 200 W. It is also evident from the plot, that the signal intensities from all three species increase with higher applied plasma power, however, the maximum emission from the OH molecule does not increase as much as the helium atomic emission upon increasing the rf power. The maximum emission intensity from the N 2 + molecular band increases approximately 4 fold as applied plasma power is increased from 75 to 200 W, in comparison to a 6 fold increase for helium atomic emission. In the case of the OH molecule, this enhancement was less than 2 fold. Figure 4.3 is a plot of the normalized emission signal from He I, OH, and N2 + as a function of applied rf power. All the signals were measured at 1.75 mm away from the center of the discharge, where the intensity from He I is at its maximum. This plot illustrates a relatively linear relationship between the applied rf power and the He I and N2 + emission intensities. This increase in N2 + emission is likely due to the mechanism by which N 2 + is formed. It has been shown [5] that in helium discharges, N2 + is often formed via a charge transfer reaction with He 2 +: He! + N 2 = 2He + Nj 77 _cg 1 1 1 1 1 1 1 1 1 ® -2.0 -1.5 -1.0 -0.5 0.0 0.5 1.0 1.5 2.0 5 -4 -3 -2 -1 -0 --2.0 -1.5 (C) - • - 75 W, - © - 100 W -A- 125 W, - A - 150W - s - 200 W ~ 1 — -1.0 T T -0.5 0.0 0.5 Torch Position, mm 1.0 1.5 2.0 Figure 4.2: The spatial distribution of (A) He I (388.87 nm), (B) OH (band head at 310 nm), and (C) N 2 + (band head at 391.4 nm) at 0.1 L min'1 gas flow rate and at different applied rf powers ranging from 75 to 200 W. 78 This reaction is possible because the wide recombination reaction energy of He2+ (18.3 - 20.5 eV) is resonant with the B22u+ energy level of N2+. It can also be seen in Fig 4.3 that the emission intensity from the OH molecule also posseses a linear relationship with the applied rf power up to 125 W. When the applied power exceeds 125 W, the emission response curve bends towards the rf power axis. This attenuation of OH emission over N2 + emission, suggests that the OH is dissociated by the plasma at a higher applied power. This observation positively correlates with the dissociation energies of these two molecules. The OH molecule with a dissociation energy of 4.39 eV is easier to atomize than N2+, whose dissociation energy is 8.71 eV [6]. 1.0-, 0.8 4 3 0.6 4 ca c g> CO c g '</> W 'E LU •o OJ) N "to E b 0.4 4 0.2 4 0.0 + Hel -E3- OH -©- N2+ T T T T 80 100 120 140 160 Plasma Forward Power, W 180 200 Figure 4.3: Effect of applied rf power on the emission intensity of He I, OH, and N2 + molecule measured at 1.75 mm off the center of the discharge. 79 The effect of gas flow rate on the spatial distribution of plasma species has also been studied at a constant applied rf power. Figure 4.4A, 4.4B and 4.4C illustrate the influence of gas flow rate on Hel, N2 + and OH emission response, at a 150 W plasma power. As can be seen from Fig. 4.4A, the helium atomic emission intensity increases slightly with increasing gas flow rates from 0.1 to 0.5 L min1 without changing its spatial distribution in the discharge. Unlike helium atomic emission, N2 + molecular band intensity, at a position adjacent to the torch wall, remains the same up to a 0.3 L min1 gas flow rate and then begins to decrease as the flow rate increases, as illustrated in Fig 4.4B. The emission intensity from OH (shown in Fig. 4.4C) also remains almost unchanged up to 0.3 L min1 gas flow rate before it starts decreasing at higher flow rates. In both cases, the spatial distributions of these molecules remain unchanged. Higher gas flow rates might be responsible for an increase in the number of collisions in the discharge due the to "thermal pinch" effect, resulting in an increase in He I intensity. It is well known that, for a free-burning d.c. arc, when the current carrying portion of the arc is constricted either mechanically, hydrodynamically, by magnetic pressure or by other means, the electron density (at the constant current level) increases in comparison to a conventional atmospheric pressure discharge [7-9]. For a DCP source, Zander and Miller [10] reported a dramatic increase in electron density in the confluence region due to nebulizer flow with the presence of a ceramic flow director, as well as due to the electrode coolant flow. It is possible that in the PP-CCP source, higher flow rates are responsible for plasma constriction (which is visually evident at very high flow rate such as 1.0 L min'1 and at the outer 80 co g> CO c <D +-« C c o CO CO 'E LU i CD or 1.2-0 . 8 -0 . 4 -0 . 0 -1 — -2.0 4.0- j 3.0-1 2.0 1.0 0.0 1 — -2.0 •1.5 -1.5 0.1 Lmin" 1 --©- 0.2 L min" 1 0.3 L min" 1 — A — 0.5 L min" 1 -1.0 -0.5 0.0 0.5 - • - 0.1 L min - A - 0.2 Lmin -A- 0.3 L min - O - 0 . 5 L min 1.0 -1.0 -0.5 0.0 0.5 1.0 1.5 1.5 2.0 2.0 3.0-I 2.0 1.0-0 . 0 -- A - 0.3 Lmin" 1 , 0.1 Lmin" 1 , - A - 0.2 Lm in " 1 0.5 L min" 1 1 1 1 1 1 1 1 1 -2.0 -1.5 -1.0 -0.5 0.0 0.5 1.0 1.5 2.0 Torch Position, mm Figure 4.4: Effect of different gas flow rates on the spatial distribution of (A) He atomic, (B) N 2 + , and (C) OH molecular emission in a 150 W plasma. 81 region of the plasma torch) and hence, a rise in the electron number density and the number of collisions. At the same time, an increased mass transfer into the discharge may cause the discharge to cool slightly, resulting in a decrease in the molecular emission intensity at higher flow rates. Despite these intensity variations, the spatial distribution of different plasma species remains completely unchanged at different plasma gas flow rates. 4.3.2 Spatially Resolved Rotational Temperature Spatially resolved rotational temperatures were calculated from the OH and N2 + molecular emission measurements discussed in the previous section. The effect of plasma forward power on the spatially resolved rotational temperatures is depicted in Fig. 4.5A and 4.5B with a gas flow rate of 0.1 L min1. The error bars represent the precision in the measured temperatures that are based on the reproducibility of five repetitive measurements. At a given applied power, the N2 + rotational temperature is at its maximum 1.25 mm from the center. At this position, the temperature increases from 845 K to 928 K as applied rf power increases from 100 to 200 W. As can be seen in Fig. 4.5A, from 1.25 mm outwards, the temperature decreases for each applied forward power. 82 CD 3 03 <D Q-E <D ca c g '§ o + CM Z 1000 900 4 800 700 600 4 03 O DC X O 500 r -2 1000- , 2 900 H CO CD Q. E CD H 75 c o 800 700 600 H 500 X----1' (A) -| 1 1 r -1 " T 0 T 1 r -i 1 1 1 2 (B) T i i—r | I I I I [ I I I I | I I I I | I I I I | I I I I | I I I I | I I I l | - 1 0 1 2 Torch Position, mm Figure 4.5: Spatially resolved (A) N 2 + and (B) OH rotational temperatures profile at different plasma forward powers and for a plasma gas flow rate of 0.1 L min 1 . Error bars represent ± 1 standard deviation (5 samples).. 83 The spatially resolved rotational temperature profile measured from the OH molecule is different from that of the N2 + molecule; as is plotted in Fig. 4.5B, the profile resembles a "W". At the center of the discharge, the rotational temperature has a higher value, which decreases as it moves outward, reaching its minimum at 1.25 mm, before beginning to increase again. The rotational temperature varies from 807 K to 872 K for a 100 W plasma forward power. As is expected, the temperature increases slightly with increasing applied power from 100 to 200 W, but the spatial profile remains unchanged. The effect of plasma gas flow rate on spatially resolved molecular rotational temperatures for N2 + and OH has also been studied, and the results are illustrated in Fig. 4.6A and 4.6B, with the applied rf power being 150 W for all the represented data. The error bars represent the precision in the measured temperatures that are based on the reproducibility of five repetitive measurements. These plots show that the molecular rotational temperatures decrease throughout the discharge volume of the source as the gas flow rate is increased from 0.1 to 0.5 L min1, but that no change in spatial profile is observed. This confirms our previous argument that the cooling of the plasma is related to increasing gas flow rate. 84 1000- , § 900 800 H 700 A 600 500 1000-n 900 800 700 600 500 ( A ) 0.1 Lmin 0.2 L min 0.3 L min — 0.5 L min I i i i 2 T i | i i r -1 I ' ' I 1 1 1 0 I I I I I I (B) . . - -I--I- I -z ...JL -I- ~x 1 - I - ' Flow =0.1 Lmin" 1 Flow =0.2 L min' 1 Flow =0.3 L min"1 — Flow =0.5 L min 0 1 Torch Position, mm Figure 4.6: Spatially resolved (A) N 2 + and (B) O H rotational temperatures profile for a 150 W applied plasma power and at different plasma gas flow rates. Error bars represent ± 1 standard deviation (5 samples). 85 4.4 Conclusions The spatial distributions of plasma background species were studied using helium atomic and OH and N2 + molecular emission. The spatial emission structure, observed for the source, is due to the transverse power coupling geometry. The electromagnetic field is attenuated during passage through the plasma, so that at the center of the discharge there is no detectable emission from Hel and N2+. The "analyte emission" or "emission sensitive" zone is situated adjacent to the discharge wall where He I and N2 + signals are dominant. The spatial distributions of plasma species remain unchanged, however, the signal intensities from all three species increases with an increase in applied plasma power. As is the case for changes in rf power, the spatial distribution of different plasma species remains totally unchanged at different plasma gas flow rates. At a given applied power, the N2 + rotational temperature is at its maximum, 1.25 mm from the center. At this position, the temperature increases from 845 K to 928 K, with increasing power from 100 to 200 W. From 1.25 mm outwards, the temperature decreases for each applied forward power. The rotational temperature profile calculated from OH molecular emission has a "W" shape in the discharge. At the center of the discharge, the rotational temperature has a higher value, which starts decreasing as we move outward, and reaches its minimum at 1.25 mm before it begins to increase once again. Changing the rf power or gas flow rate changes the absolute value of rotational temperature, but does not have an effect on its spatial profile. 86 4.5 References 1. I. Ishii, and A. Montaser, Spectrochim. Acta, 46B (8), (1991), 1197. 2. G . Herzberg, in Spectra of Diatomic Molecules, (Van Nostrand, NJ, 1950) 3. J . M. Mermet, Spectroscopic Diagnostics: Basic Concepts, in Inductively Coupled Plasma Emission Spectroscopy, P.W.J.M. Boumans, Ed., (John Wiley & Sons: New York, 1987). 4. M. M. Rahman and M. W. Blades, Spectrochim. Acta, 52B, (1997), 1983- 93. 5. M. Endoh, M. Tsuji, and Y. Nishimura, J. Chem. Phys., 79, (1983), 5368-75. 6. R. E. Sturgeon, S. N. Willie, V. T. Luong, and J . G . Dunn, Appll. Spectrosc, 45, (1991), 1413- 1418. 7. H. Gerdien and A. Lotz, Z.Tech. Physik, 4, (1923), 157. 8. G . M. Giannini, Sci. Amer., 197 (2), (1957), 80. 9. H. Hugel, IEEE Trans. Plasma Sci., PS8 (4), (1980), 437. 10. A. T. Zander and M. H. Miller, Spectrochim. Acta, 40B, (1985), 1023-1037. 87 C h a p t e r - 5 Effects from Easily lonizable Elements on Silver Analyte 5.1 Introduction All spectrochemical methods are susceptible to interference caused by concomitants present in the sample matrix; however, certain methods may prove more or less susceptible than others. In emission spectrochemical methods based on a graphite platform for analyte vaporization and a plasma source for analyte atomization and excitation, the analyte vaporization, atomization, and excitation characteristics may be affected by matrix effects found in graphite furnace methods, as well as those found in plasma methods. Interference effects due to the presence of concomitants in the sample can be classified into two categories: (1) spectral interferences which are due to incomplete isolation of the analyte signal from that of the interferent; and (2) non-spectral interferences caused by change in the atomic concentration being measured and therefore directly affecting the analyte signal [1]. Because of their dependence on excitation conditions and spectral resolution, spectral interferences in emission methods are highly instrument dependent. The advent of high-resolution wavelength selection devices provide an efficient means for the correction of spectral interferences. Non-spectral interferences in plasma emission spectrometry, directly affect the analyte signal by altering the atom concentration in the plasma, and/or by changing the plasma characteristics. For methods that use electrothermal sample 88 introduction, matrix volatilization or condensed phase interferences affect the analyte element up to the point at which it is released from the atomization surface into the gas phase. Vapor phase interferences on the other hand, influence subsequent gas-phase processes and/or plasma characteristics. Non-spectral interferences in graphite furnace AAS have been extensively studied and the proposed causes are: (I) volatile compound formation in the condensed phase, (ii) occlusion of the analyte in the bulk of the matrix, (iii) formation of refractory compounds or change in the rate of analyte supply, (iv) analyte vapor expulsion from the furnace due to rapid gas evolution by the matrix, (v) a shift in the dissociation equilibria due to stable compound formation, and (vi) a shift in ionization equilibria or change in the rate of analyte removal. All of the above mentioned mechanisms, excepting iv, could be encountered for the graphite platform used for the PP-CCP. in addition to these potential interference effects, the presence of an interferent can also alter the plasma discharge characteristics. Examples of this would be: (i) changes in collisional excitation due to a change in the electron energy distribution or density as a result of the production of excess electrons from easily ionizable elements (EIE) present in the sample, (ii) shifts in the ionization equilibrium of the analyte species due to ionization of matrix elements, (iii) a decrease in gas kinetic temperature, (iv) alteration of the power coupling efficiency of the plasma as a consequence of changes in the load impedance in the presence of EIE, and (vi) changes in the plasma gas composition, in which a significant amount of helium is displaced by matrix elements due to the presence of EIE. 89 The effects of easily ionized elements from different plasma sources used in atomic emission spectrometry have been reported extensively in the atomic spectroscopy literature [2-17]. Both P P - C C P and F A P E S (Furnace Atomization Plasma Emission Spectrometry) are capacitively coupled plasmas, primarily using He, operating at similar rf powers, and with relatively similar plasma characteristics [18]. In addition, the P P - C C P shares some fundamental and analytical character -istics with the microwave induced plasma (MIP)[18]. Smith et al. [19] observed a slight enhancement in Ag signal at low concentration in a F A P E S source and a depression at higher concentration with both NaCI and N a N 0 3 as the interferent. It was suggested that at low concentrations of interferent the signal was enhanced by a shift in the ionization equilibrium, and, at higher concentrations, the plasma was being "quenched". Sturgeon et al. [20] reported a reduction in Pb emission when NaCI was present in the sample and found that atomization from a platform significantly decreased the degree of the interference in their F A P E S source. Hettipathirana et al. [10] noted a significant suppression in the Pb and Ag emission signals for a F A P E S source, even in the presence of trace amounts (i.e. 162 ng) of NaCI and N a N 0 3 . For a given amount of Na, the magnitude of the reduction in Ag emission was much greater for N a N 0 3 than that for NaCI. Gilchrist and Liang [21] reported suppression of Tl emission signal in the presence of NaCI for the commercial (Aurora Instruments Ltd.) F A P E S source. For the F A P E S source Imai and Sturgeon [11] have suggested that radiative power losses from the plasma, due to excitation of the EIE matrix species, as well as 90 alteration of the electron energy distribution function, as a result of EIE ionization and collisional dissociation of molecules, are the major causes of EIE interferences. For low pressure helium microwave induced plasmas (He-MIP) it was found that the presence of mM concentrations of KCI, produced enhancements in emission intensity of as much as 1000 fold for Cu and Mn [2]. Alder et al. [22] found that the presence of KCI enhanced the emission response, but in this case, a low pressure Ar-MIP was employed and signal enhancement were less, a 10 -15 fold enhance -ment for Mn. For an atmospheric pressure He-MIP, enhancements due to alkali salts appear to be generally less than a factor of 10 [6, 23]. Falk et al. [24] noted a significant influence of NaCI on analyte signals when its concentration was 1000 times higher than the analyte concentration in a F A N E S (Furnace Atomization Non-thermal Emission Spectrometry) source, in which the plasma is a DC glow discharge. They reported an initial slight enhancement followed by a 20% suppression of Cu signal in the presence of 100 ug/m\ Na, which was 2000 times higher than the analyte concentration. Interference effects have not been reported for P P - C C P to date. Since the design and operation of the P P - C C P for atomic emission spectrometry is somewhat different than that for MIP, F A P E S , and F A N E S , the nature and/or mechanism of interferences may likewise differ for the P P - C C P . This work examines the influence of easily ionized elements (Na as NaCI and NaN0 3 ) on Ag emission response in a P P - C C P . This influence is a matter of significant practical relevance, as E lEs, especially Na and K, are present in a wide variety of natural matrices, such as animal and plant materials, blood, seawater and geological samples. 91 5.2 Experimental 5.2.1 Instrumentation Details of the construction and operation of the P P - C C P have been given in chapter 2 of this thesis. In this study a 1.0 m Czerny-Turner scanning monochromator (Model 2061 Scoeffel-McPherson, Acton, MA, USA) with a 3600 lines mm"1 holographic grating was used to disperse emission from the plasma source. The signals were detected using an intensified photodiode array (IPDA) detector (Princeton Instruments, Model IRY-700 /G/B / PAR, Trenton, NJ, USA). The graphite platform temperatures were measured using an optical pyrometer (Ircon Series 1100, Model 11x30, IL, USA). 5.2.2 Reagents All analyte solutions were prepared by the serial dilution of a 1000 ppm stock solution prior to use. Silver and Na from NaCI solutions were prepared by dissolving analytical grade A g N 0 3 and NaCI ( Fisher Scientific, Ontario, Canada) in 1% H N 0 3 solution and distilled water respectively. Analytical grade N a N 0 3 (Anachemia, Vancouver, Canada) was used to prepare another Na solution in 1% H N 0 3 solution. Purified He (Paraxair Products Inc., Mississauga, O N , Canada) was used as the plasma gas. 92 5.2.3 Procedure A 10 uL sample aliquot was placed into the depression of the platform using an Eppendorf 0 . 5 - 1 0 uL micro pipette fitted with polyethylene tip. Matrix solutions were mixed on the platform. Samples were dried for 40 s at 90 °C followed by a 30 s "char" stage at 250 °C with the He flow on. After a two minute waiting period, radio frequency power was applied, whereupon the plasma spontaneously ignited. Following a 30 s plasma stabilization period, the atomization stage was activated reaching a temperature of 2000 °C for all determinations. The atomization cycle was on for 5 seconds. The blank determinations were carried out in the same manner without the analyte or analyte and interferent. Four replicate measurements for each determination were made using the resonance lines of Ag (328.07 nm), Na (330.23 nm), and He (388.87 nm). 5.3 Results and Discussion During the atomization cycle, compositions of the vapor phase changes with or without the presence of matrix. In addition to altering the atomic population in the vapor phase, EIE matrix (molecular and elemental) may change and/or influence further vaporization of the analyte (from the graphite platform) and/or the characteristics of plasma, resulting in an observed interference (both enhancements and depression of analyte signal). Figure 5.1 is a plot of the temporal behavior of the Ag signal with and without the presence of 150 ng of Na as NaCI, at 150 w plasma forward power, and 0.1 L min 1 plasma support gas flow. In this plot all the emission signals are normalized with respect to the signal for 1.0 ng of Ag without any Na. 93 The figure also shows the temporal response of Na as NaCI and the Na response in the presence of Ag analyte, as well as the temporal profile of the electrothermal vaporizer. 3.0 r -.1 1-0 Q) rr 0.5 0.0 - , 2000 Temperature _i i i i_ Ag Alone Ag with Na Na Alone — Na with Ag Temp. (Pyrometer) — Temp. (Extrapolated) -i i i i • • - I L . 1500 TJ o> =* o 3 -H 1000 § •o CD c CD 500 O 0.0 0.5 1.0 1.5 Time, Sec. 2.0 2.5 J J 0 3.0 Fig. 5.1: Effect of 150 ng of Na from NaCI on the temporal response of 1.0 ng of Ag at 150 W plasma forward power and 0.1 L min 1 gas flow rate. The temperature profile on the electrothermal vaporizer is also shown. As can be seen, there are differences in temporal behavior between pure Ag and Ag mixed with NaCI, as well as between Na and Na mixed with Ag. For pure Ag, there is a single well-defined peak with an appearance temperature of 1240 K, 94 consistent with the literature value [25-27]. When NaCI is present with Ag there are two peaks for Ag, a small peak appearing earlier at a temperature of 1150 K, and a sharp peak appearing latter in time which has similar characteristics to Ag temporal emission from pure A g N 0 3 . Sturgeon [28] suggested a first order release process for Ag whereby the atomization mechanism proceeds by the following reactions [29-33]: - A g 2 0 ( S ) + 2 N 0 2 + 1 /2 0 2 (5.1) :> 2Ag ( S ) + 1/2 O a (5.2) ^ 2Ag ( S ) + C O (5.3) ^ A g ( g ) (5.4) In the presence of NaCI, Ag may also be formed by the decomposition of volatile AgCI. The existence of two peaks for an Ag signal is indicative of Ag being formed by two different mechanisms. In the presence of C l , the early peak is likely due to the vaporization and gas-phase dissociation of AgCI and the second peak from thermal and/or carbon reduction of A g 2 0 to Ag, which appears in the plasma latter in time. In Fig.5.3 when no chloride anion is present, Ag produces a well-defined single peak with or without the presence of Na as N a N 0 3 . This also supports the hypothesis that the formation of the earlier peak of Ag occurs as a result of the dissociation of volatile gas-phase AgCI. In Fig. 5.1 it can also be seen that Na, when present as NaCI, produces a single well defined peak with an appearance temperature of 1200 K, consistent with the published value of 1250 K [25, 26]. The most probable mechanism of Na atom 2 A g N 0 3 ( S ) A g 2 0 ( S ) ^ A g 2 0 ( S ) + C < Ag (s) + 95 formation is the gas-phase dissociation of volatile NaCI. The existence of two peaks for the Na signal from N a N 0 3 (Fig. 5.2) and also from NaCI in the presence of A g N 0 3 (Fig. 5.1) is an indication of Na being formed by two different mechanisms. Campbell and Ottaway [25] have suggested that in ETAAS, Na can be formed from carbon reduction of N a 2 0 at 1200-1300 K: 2 N a N 0 3 • N a 2 0 + 2 N 0 2 +1/2 0 2 (5.5) N a 2 0 + C 4 • 2Na + C O (5.6) Na may also be formed by the thermal decomposition of N a 2 0 , resulting in the observed double peak. In the presence of 150 ng of Na as both NaCI and N a N 0 3 > the Ag signal is significantly enhanced, as can be seen in Fig. 5.1 and 5.2. As in Fig. 5.1, all emission signals in Fig. 5.2 are also normalized with respect to the Ag signal without the presence of Na. In the case of NaCI, the appearance time for the Ag second peak is the same as that for Ag without the presence of interferent, and the peak signal appears at almost the same time as that of pure analyte. In the presence of N a N 0 3 , the Ag signal appears at the same time as without N a N 0 3 , but the peak maximum occurs earlier. The appearance and peak time for the analyte signal points to an enhancement in the excitation ability of the plasma and/or enhancement in atomization efficiency as the cause for this signal enhancement. It should be noted that Na, when present as NaCI, leaves the platform 0.07 s earlier than the pure Ag, 96 whereas as a mixture of both Na and Ag comes off at the same time (Fig. 5.1). In Fig. 5.2 it can be seen that Na always comes off earlier than Ag. For this particular set of operating conditions, the signal enhancement is less in the presence of N a N 0 3 than that of NaCI. This suggests that part of the enhancement, in the presence of chloride, is a result of atomization efficiency enhancement due to the formation of a volatile chloride compound. 2 . 5 - , " y — i —n—r— ] — i — i —i—i—| i i — i i | i — n — r - | — n — i i | — r — 1 — n — [ — i — i — i — r - j 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 Time, Sec. Fig. 5.2: Effect of 150 ng of Na from N a N 0 3 on the temporal response of 1.0 ng of Ag for a 150 W plasma forward power and 0.1 L min 1 gas flow rate. 97 All excitation processes are ultimately influenced by the density and energy of free electrons in the plasma [34]. Collisional excitation and ionization of analyte and matrix elements will cause a decrease in energy for plasma electrons relative to the situation without the presence of analyte and matrix elements. Figure 5.3 illustrates the effect of vaporization of different amounts of NaCI into the plasma on emission from He I (388.8 nm) during an atomization cycle. In all three figures, 5.3A-5.3C, the bottom trace represents the plasma background signal (measurements at positions on either side of the He I line) with the presence of 1.0 ug, 100 ug and 300 ug of Na respectively. It is clear, from Fig. 5.3A, that the He I signal increases gradually as the temperature of the platform increases without any matrix present. This gradual increase is probably due to the preheating of plasma gas as it enters the torch, yielding an enhancement in excitation ability. The He I emission signal remains unchanged up to 1.0 ug of added Na. However, when the matrix amount is increased to 5.0 ug and above, the emission intensity is significantly suppressed. This observation suggests a decrease in energy in the plasma due to the presence of matrix species. The effect becomes more pronounced with an increase in the amount of matrix species (Fig. 5.3B and 5.3C), and the plasma is perturbed, particularly during the introduction of matrix into the plasma. If one assumes that the excitation of He is the result of collisions between the He atom and high-energy electrons (>23 eV), there might be several factors causing the decrease in He I response. Plasma induced dissociation of matrix molecules (i.e. NaCI) via impact with high-energy electrons decreases the population of these electrons, resulting in a decrease in the effective collision frequency between helium atoms and high -98 10000 8000 6000 4000 2000 h ( A ) — He, No Atomization, No Na He with Atomization, No Na — - He with 0.25 ng Na — He with 1.0 ug Na — Background with 1.0 ug Na 0 0.0 • ' ' • I I I I I I I I I I L. 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 ca c gi co c g OT OT 'E LU a> X i rr 8000 6000 4000 2000 h ( B ) — He with 5 ug Na He with 25 ug Na - - He with 100ugNa . - - Background with 100 ug Na 0 B- i ! I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 10000r-8000 6000 4000 2000 (C) — He with 200 ug Na - - He with 300 ug Na Background with 300 ug Na i i i i i i J_ J i i i l i i i i l i j i i i i i i i_ OIL 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 Time, Sec. Fig. 5.3: Influence of added amounts of EIE (Na as NaCI, as an example) on He atomic emission signal for a plasma forward power of 150 W and gas flow rate of 0.1 L m in 1 in a P P - C C P source. 99 energy electrons and, hence, a decrease in He I emission intensity. The relatively low ionization potential of Na (5.14 eV) favors collisional ionization between Na and high energy electrons, resulting a decrease in the population of latter. On the other hand, the ionization of Na floods the plasma with low energy electrons shifting the E E D F (electron energy distribution function) to lower energy [35] as the energy input into the plasma remains constant. The plasma can also change its characteristics due to an excitation of the atomic Na vapor, resulting in an increased amount of power radiated from it in the form of photon loss [11, 34]. Falk [34] has estimated the magnitude of this effect by considering the excitation of the Na 589 nm doublet. For example, a 10% radiation power loss can occur when Na constitutes as little as 0.01% of the plasma gas in a 50 W source. Such power loss and changes in E E D F would certainly have an impact on the excitation of analytes in plasma at higher concentrations of matrix species. The temporal response of an emission signal from a deposition of 1.0 ng of Ag with increasing amounts of EIE matrix, where Na was added as NaCI, is shown in Figure 5.4 for a 150 W plasma forward power and 0.1 L min'1 gas flow rate. The emission signals are normalized relative to the signal from 1.0 ng of Ag without any matrix element added. It can be seen from this figure that, with small amounts of matrix species, the signal appearance time is much earlier than that for a pure analyte. There is also an earlier first peak for Ag emission, supporting the idea of early vaporization of volatile chloride compounds into the plasma. The signal appearance time is earliest for 0.125 ug of Na, which is 0.14 s earlier than the Ag without any matrix addition. The disappearance of the earlier peak with increasing 100 amounts of matrix species is probably due to a decrease in plasma power via radiative loss, and thus shifts the E E D F to a lower energy which delays the plasma induced dissociation of Ag-CI bond. The figure also shows that the appearance time gradually increases with increasing amounts of NaCI. For 40, 100, and 250 ug of Na, the signal is delayed by 0.07, 0.07, and 0.14 s, respectively, in comparison to the pure analyte. 6 - , 0.0 0.5 1.0 1.5 2.0 2.5 3.0 Time, Sec. Fig. 5.4: Temporal response of Ag atomic emission from a deposition of 1.0 ng of Ag with different amounts of Na (as NaCI) for a 150 W plasma power and 0.1 L m in 1 gas flow rate. 101 As illustrated in the figure below (Fig. 5.5), in the case of NaN03 there was no delay in signal appearance time with the addition of higher amounts of Na. The signal does however appear at 0.07, 0.14, and 0.07 s earlier for 0.005, 0.05, and 1.0 ug of added Na, respectively; there was no correlation between the maximum signal enhancement and the signal appearance time. Maximum signal enhancement occurs at 40.0 ug of Na as NaCI, where the signal maximum is delayed by 0.07 s compared with pure analyte, and for NaN03 the maximum signal enhancement occurs at 0.05 ug of Na where signal appears 0.14 s earlier than pure analyte. 8 - i 0.0 0.5 1.0 1.5 2.0 Time, Sec. Fig. 5.5: Temporal response of silver atomic emission for a deposition of 1.0 ng of Ag with different amounts of Na (as NaN03) for a150 W plasma power and 0.1 L min1 gas flow rate. 102 Figures 5.6 and 5.7 illustrate the effect of increasing the amount of added matrix component, Na as NaCI and N a N 0 3 respectively, on 1.0 ng silver emission response at a constant plasma forward power of 150 W and a gas flow rate of 0.1 L min 1 . All the signals are normalized with respect to the analyte emission signal without any added matrix and peak heights are used for all calculations. As can be seen from the figures, the emission response increases gradually, reaching a maximum of 5.4 times enhancement with the addition of 40.0 ug of Na as NaCI. The enhancement with the addition of 0.05 ug of Na as N a N 0 3 was about 8 times. With increasing amounts of added Na, the Ag emission signal decreased steadily and ultimately reached 80% of the analyte only signal with the addition of 250.0 and 200.0 ug of Na as NaCI and N a N 0 3 , respectively. The peak height analysis reflects the convolution of vaporization and atomization efficiency, residence time, and excitation efficiency of the source. Changes to the integrated (peak area) response of silver atomic emission as depicted in Fig. 5.8 and 5.9 were also considered in order to see the changes in the kinetics of sample introduction and loss in the system. The signals were normalized to the analyte emission signal without any matrix added. The figures illustrate that the emission response increases gradually and reaches a maximum of 6.3 times enhancement with the addition of 40.0 ug of Na as NaCI, whereas the enhancement was about 10 times with the addition of 0.05 ug of Na as NaN0 3 . With an increasing amount of added Na, the Ag emission signal decreased steadily and ultimately reached 90% of the analyte only signal with the addition of 250 and 200 ug of Na as NaCI and N a N 0 3 respectively. It is clear that N a N 0 3 enhances the signal more than NaCI does. This strongly suggests that the 103 5 -4 -3 -2 1 0 4 T 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0.0 0.5 1.0 1.5 2.0 6 - , 5 4 -3 -2 -1 -0 T — r "T* 50 T — i — r T T T — i — i — r 100 150 Added Amount of Na, 200 250 Fig. 5.6: Effect of varying amounts of EIE (Na from NaCI) on Ag emission signal (from 1.0 ng sample introduction) for a 150 W plasma forward power and 0.1 L min 1 gas flow rate. 104 84 64 44 24 04 I I I I I I I I I I I I I I I I I I I I I I I I I I I 0.0 0.2 0.4 i i i i 11 i i i i i i 11 i 11 11 i i i i i i | 0.6 0.8 1.0 ^ 8 64 44 24 T i r ~t 1 1 1 1 1 1 1 1 1 1 1 1 r 50 100 150 Amount of added Na, ug 200 Fig. 5.7: Effect of varying amount of EIE (Na from NaN0 3 ) on Ag emission signal (from 1.0 ng sample introduction) in a 150 W plasma and 0.1 L min' 1 gas flow rate. 105 7 - i 5 -4 -3 -2 -0 4. —I—i—i—i—i—|—i—i—i—i—|— 0.0 0.5 1.0 T 1 1 r 1.5 2.0 7 - , 6 5 4 3 2 1 0 ' • 1 , , 1 1 1 1 1 | 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 50 100 150 200 250 Amount of Added Na, ug Fig. 5.8: Effect of varying amounts of EIE (Na from NaCI) on Ag emission signal (from 1.0 ng sample introduction) at 150 W plasma forward power and 0.1 L min 1 gas flow rate. 106 N H 8-4 6-\ 4-^  0 i i i i 11 i i i 1111 i i 11 0.0 0.2 T T T T -0.4 | I I I I | I I I I | I I I I | I I I I | I I I I | 0.6 0.8 1.0 10^ 8^ S 6-I - 4 24 OA Fig. 5.9: ~I 1 1 r i 1 r T 1 1 r 50 100 150 i 1 r "~I 200 Amount of added Na, fig Effect of varying amounts of EIE (Na from NaN03) on Ag emission signal (from 1.0 ng sample introduction) for a 150 W plasma forward power and 0.1 L min1 gas flow rate. 107 difference in behavior is due to a molecular dissociation effect in the plasma. As a result of the lower bond energy of Ag-0 (2.55 eV) compared with that of Ag-CI (3.2 eV), the extent of plasma induced dissociation of Ag-0 is greater, giving a larger signal enhancement. The addition of an EIE can also increase the electron density, thereby shifting the ionization equilibrium between analyte atoms and ions toward the atom, resulting in an enhancement of the atomic emission signal. The effects of the magnitude of the plasma power on the interference effect from NaCI and NaN03, on Ag atomic emission, are provided in Fig. 5.10 -13. All signals are normalized relative to that of Ag at 100 W plasma power without any added Na matrix. For each sample injection 1.0 ng of Ag or 1.0 ng of Ag and 150 ng of Na as NaCI or NaN03 were present. Fig. 5.10 is the temporal response of Ag atomic emission with (Fig. 10A) and without (Fig. 10B) the presence of Na as NaCI at a plasma gas flow rate of 0.1 L min1. It can be seen in these figures, that the temporal response is independent of the magnitude of plasma power. For each applied forward power, the Ag atomic emission signal appears first, as a smaller peak; 0.14 s earlier than those without the presence of Na matrix. The effect of NaN03 matrix, at different plasma powers, on silver analyte emission from a deposition of 1.0 ng of Ag is depicted in Fig. 5.11 A and 5.11B. Figure 5.11 A represents the temporal response with the presence of the matrix whereas Fig. 5.11B represents the response without the presence of matrix. As can be seen from the plot, in the presence of NaN03, the Ag analyte signal appears at the same time as without NaN03, but the peak maximum occurs earlier, independently from applied plasma power. 108 03 c g> co c o CO (0 E LU CD > co 20 CD N "ca I 15 o z p = 100 w p = 125 W P = 150 W - - - p = 200 W — p = 250 W (B) 1 0 4 5 4 — P=100W P=125W — P=150W — - P=200 W •— P=250 W n — i — i — i — | — i — i — i — i — | — i — i — i — i — | — i — i — i — r O^ O 0.5 1.0 1.5 2^0 i — i — i — i — | — i — i — i — i — | 2.5 3.0 Fig. 5.10: Effect of different plasma applied powers on the temporal emission response of 1.0 ng silver: (A) with (150 ng) and (B) without the presence of matrix (Na as NaCI). The gas flow rate is 0.1 L min 1 . 109 c C o "I E Hi | CO T3 <D N « E k-o z 2 0 - , 15-J 10-^ 2 0 - , 15 10-J 5-J P=100w P=125 W P=150W P=200 W (B) f I p = 100 w p = 125 W p = 150 W - - - p = 200 W — p = 250 W •7/ v - \ N 1—i—i—i—i—|—i—i—i—i—|—i 0.0 0.5 1.0 I 1 1 1 1 I 1.5 2.0 T - T - T - I — r - T - r — | 2.5 3.0 Time, Sec. Fig. 5.11: Effect of different applied rf powers on the temporal emission response of 1.0 ng silver: (A) with (150 ng) and (B) without the presence of matrix (Na as NaN0 3 ) . The gas flow rate is 0.1 L min 1 . 110 The emission intensity, on the other hand, is very much dependent on the magnitude of plasma power, as can be seen in Fig. 5.12 and 5.13, due to the presence of NaCI and NaN03, respectively, as matrix species. Without the presence of matrix species, the silver emission increases monotonically as power is increased from 100 to 250 W, with a constant gas flow rate of 0.1 L min"1. With increasing power, the electron density, as well as the electron temperature increases in the plasma [36]. Assuming that the excitation of emitting species occurs predominantly 35-, 100 150 200 250 Plasma Forward Power, W Fig. 5.12: Effect of different applied rf powers on analyte signal enhancement from a deposition of 1.0 ng of silver with (150 ng) and without sodium matrix (Na as NaCI). The gas flow rate is 0.1 L min1. Ill through electron impact, increased power would have the net effect of increasing emission output. When Na is present as a matrix element, the Ag emission is enhanced for each increase in applied power, as opposed to Ag alone. The integrated emission signals presented in both Fig. 5.12 and 5.13 were normalized to the signal at 100 W plasma power without the presence of any matrix species. 30-j 100 150 200 250 Plasma Forward Power, W Fig. 5.13: Effect of different applied rf powers on analyte signal enhancement from a deposition of 1.0 ng silver with (150 ng) and without sodium matrix (Na as NaN03). The gas flow rate is 0.1 L min1. 112 As shown, for a 100 W plasma power, the Ag signal increased 22 times with the presence of NaN03 in comparison to the signal without NaN03. On the other hand, the presence of NaCI causes the signal to increase 18 times in comparison to the signal without any matrix. In both cases, the degree of enhancement decreases with increasing plasma power. At a 250 W plasma power, and in the presence of NaN03 as matrix species, the analyte signal was depressed by 15% rather than enhanced. It seems that the ionization of the small amount of EIE produces additional electrons which, although not energetic enough to ionize the He plasma gas, may acquire sufficient energy in the rf field to excite the Ag analyte (3.8 eV), giving rise to an increased response over that obtained in the absence of Na matrix. This enhancement may also be due to increased ionization of EIE, which enhances neutral atomic emission via shifts in the ionization equilibrium from ion-electron recombination reactions. Sturgeon et al. [ 3 7 ] reported that the degree of ionization increases with increasing plasma power in a FAPES source, which is closely related to the PP-CCP source. It is possible that, at a higher power, and with the presence of EIE, the enhancement decreases due to higher ionization, which in turn causes a shift in the ionization equilibrium from ion-electron recombination reactions. 5.4 Conclusions The effect of NaCI and NaN03 on emission from Ag for an atmospheric pressure PP-CCP has been studied. Both enhancement and suppression of analyte line emission was observed depending upon the amount of matrix species. 113 Compared with the emission signal obtained when there is no interferent, the presence of either NaCI or NaN03 (up to 40 and 0.05 ug respectively) as a concomitant in the sample, caused an interference effect by enhancing the emission intensity. With increasing amounts of matrix species the signal enhancement decreased, resulting in a 10% signal depression in the presence of 250 ug of Na as NaCI and 200 ug of Na as NaN03. The interference effect also depends upon the plasma input power. At 125 W plasma power, the signal enhancement is at its maxi -mum and then begins to decrease in signal enhancement with increasing power. The results of these studies suggest two opposing mechanisms for interference from EIE in the PP-CCP source, one causing an enhancement of the analyte signal and the other causing a depression. Plasma induced dissociation of volatile chloride compound, a shift in the atom-ion equilibrium towards the atom, and enhanced electron impact excitation due to an alteration of both the electron density and energy distribution are the most probable reasons for the enhancement of the analyte signal in the presence of smaller amounts of matrix species. When there is a higher amount of EIE matrix, the factors causing the enhancement of the analyte signal outweigh the factors acting to suppress the signal. Plasma induced dissociation of matrix molecules and collisional ionization of the matrix element decrease the population of energetic electrons in the plasma. Subsequently, ionization of higher amounts of EIE shifts the EEDF to a lower energy by flooding the plasma with low energy electrons. Signal suppression may also be caused by a loss of plasma energy due to an increased amount of power radiated from the plasma in the form of photon loss as a result of the excitation of atomic sodium vapor. "1 114 5.6 References 1. IUPAC, Nomenclature, symbols, units and their usage in spectro - chemical analysis-lll. Spectrochim. Acta, 33B, (1978), 247-269. 2. H. Kawaguchi and B. L. Vallee, Anal. Chem., 47, (1975), 1029. 3. H. Kawaguchi, I. Atsuya, and B. Vallee, Anal. Chem., 49(2), (1977), 266-270. 4. I. Atsuya, C. Veillon, H. Kawaguchi & B. Vallee, Anal. Chem, 49, (1977), 1489 5. M. W. Blades and G. Horlick, Spectrochim. Acta, 36B, (1981), 881-900. 6. J . P. Matousek, B. J . Orr, and M. Selby, Spectrochim. Acta, 41B, (1986), 415 7. C. Le Blanc and M. W. Blades, J. Anal. At. Spectrom, 5, (1990), 99-107. 8. K. Kitagawa and G. Horlick, J. Anal. At. Spectrom., 7, (1992), 1221-1229. 9. P. J . Galley, M. Glick, and G. M. Hieftje, Spectrochim. Acta, 48B, (1993), 769. 10. T. D. Hettipathirana and M. W. Blades, J. Anal. Atom. Spectro., 8,(1993), 995. 11. S. Imai and R. E. Sturgeon, J. Anal. Atom. Spectrom., 9, (1994), 765-772. 12. N. N. Sesi and G. M. Hieftje, Spectrochim. Acta, 51B, (1996), 1601-1628. 13. M. Wu and G. M. Hieftje, Spectrochim. Acta, 49B(2), (1994), 149-161. 14. X. Romero, E. Poussel, & J . M. Mermet, Spectrochim. Acta, 52B, (1997), 495. 15. J . M. Mermet, J. Anal. Atom. Spectrom., 13 (5), (1998), 419-422. 16. C. Dubuisson, E. Poussel, J . L. Toboli, and J . M. Mermet, Spectrochim. Acta, 53B(4), (1998), 593-600. 17. C. Dubuisson, E. Poussel, and J . M. Mermet, J. Anal. Atom. Spectrom., 13, (1998), 1265-1269. 18. M. M. Rahman and M. W. Blades, Spectrochim. Acta, 52B, (1997), 1983- 93. 115 19. D. L. Smith, D. C. Liang, D. Steel, and M. W. Blades, Spectrochim. Acta, 45B, (1990), 493. 20. R. E. Sturgeon, S. N. Willie, V. T. Luong, and S. S. Berman, J. Anal. Atom. Spectrom., 6, (1991), 19. 21. G. F. R. Gilchrist and D. C. Liang, Am. Lab., 25, (1993), 34U. 22. J. F. Alder and M. T. DA Cunha, Can. J. Spectros., 25, (1980), 32-38 23. M. Zerezghi, K. Mulligan, & J. A. Caruso, Anal. Chim. Acta, 154, (1983), 219 24. H. Falk, E. Hoffmann, and C. Ludke, Prog. Anal. Spectros, 11, (1988), 417. 25. W. C. Campbell and J. M. Ottaway, Talanta, 21, (1974), 837-844. 26. B. V. UVov, Spectrochim. Acta, 33B, (1978), 153-193. 27. H. Falk, E. Hoffmann, and C. Ludke, Spectrochim. Acta, 39B, (1984), 283-94 28. R. E. Sturgeon, Fresenius Z Anal Chem, 324, (1986), 807-818. 29. W. Freeh, A.. Persson, & A. Cedergren, Prog. Anal. At. Spectro., 3(1980), 279 30. W. Freeh, E. Lundberg, & A. Cedergren, Prog. Anal. At. Spectro., 8(1985),257 31. K. H. Stern, J. Phy. Chem. Ref. Data, 1(3), (1972), 747-772. 32. J. G. Jackson, A. Novichikhin, R. W. Fonseca, and J. A. Holcombe., Spectrochim. Acta, 50B, (1995), 1423-1426. 33. J. G. Jackson, R. W. Fonseca, and J. A. Holcombe, Spectrochim. Acta, 50B, (1995), 1449-1457. 34. H. Falk, J. Anal. Atom. Spectrom, 6, (1991), 631-635. 35. D. Fang and R. K. Marcus, Spectrochim. Acta, 46B, (1991), 983-1000. 36. I. Lin, J. Appll. Phys., 58, (1985), 2981-2987. 37. R. E. Sturgeon and R. Guevremont, J. Anal. Atom. Spectro., 13 (1998), 229. 116 C h a p t e r - 6 A n a l y t e Ionization in the P P - C C P S o u r c e 6.1 Introduction The effect of increasing applied rf power to the electrode on the analyte (silver and lead) emission and the atomic excitation and molecular rotational temperatures were discussed in Chapter 4, whereby it was shown that all temperatures were increased with increasing applied power. This increase in temperature effects the emission intensity of atomic and molecular species. The temporally integrated lead signal was found to increase with increasing rf power up to 200 W. Further increases in applied power resulted in a decrease in the atomic emission. One possible explanation for this observation is that at higher rf powers, there may be an increase in the ionization of analyte atoms. In Chapter 5, the effects of the magnitude of the plasma power on the interference of NaCI and N a N 0 3 (as EIE) on Ag emissions were presented. It was shown that when Na was present as a matrix element, the Ag emission was enhanced for each applied power, in comparison to Ag alone. The degree of enhancement decreased with increasing plasma forward power. It is possible that at higher power (with the presence of EIE), the enhancement decreases due to higher ionization of the analyte, causing a shift in the ionization equilibrium from ion-electron recombination reactions. Details of analyte ionization were recently reported for a F A P E S source [1, 2]. The F A P E S source has similar plasma properties to those of the P P - C C P ; both 117 operate at closely comparable rf power ranges and frequency, as well as sharing a similar means by which power is coupled to the plasma in both these torches. Since 1981, after Douglas and French [3] first reported an Ar MIP-MS system, there have been a growing number of reports on microwave plasmas as ion sources for mass spectrometry. As the reported rotational and excitation temperatures of the P P - C C P source are comparable to those from F A P E S , MIP, and M C P [4], one would expect similar analyte ionization characteristics for a P P - C C P source. Also, considering the power availability in the central channel of an Ar ICP, of the order of only 100 W at 1 kW forward power [5], one might expect that the power density within the P P - C C P source when operated at a forward power of 100-200 W is comparable to an ICP. A project was thus undertaken to investigate the degree of analyte ionization in P P -C C P source and the effect of applied plasma forward power and plasma gas flow rate on the degree of ionization. A part of the work described in this chapter has been published as a full paper [6]. 6.2 Experimental 6.2.1 Background The ionization efficiency of analytical plasmas is frequently evaluated by measuring the degree of ionization. The degree of ionization is often utilized as a diagnostic tool as it can be very sensitive to changes in operating conditions. Once the plasma has been optimized, the degree of ionization can be used to monitor, and in fact verify, the operating parameters of spectrochemical analytical rf plasmas. The 118 degree of ionization can be obtained by assuming that the plasma is in a p-LTE (partial local thermal equilibrium) state, by measuring the relative intensity (I) of an atomic and ionic line, and by then applying the following relations: No ion _ pAQ(T)\ ( g, A \ v j E l o n -Efltm),*flr N0atom " { 9 A J \\XQ(T) j X e " a"m> B (6 1 ) atom From the value of this ratio the degree of ionization of the analyte can be calculated using: ' N0 ion \ % Ionization (a) = V o atom;— ^ -JQO (6.2) 1 + K |N ° I O N ) V N 0 atom) where: N 0 is the total number density of the species under study; I, the total intensity of a spectral line of wavelength X; Q(T), the temperature dependent partition function; g„ the statistical weight of the excited state of the species; A, the Einstein transition probability; E l o n and E a t o m , the excitation energies of the atomic and ionic lines; kB, the Boltzmann constant; and T, the ionization temperature. In order to apply this measurement technique to determine the degree of ionization, a number of experimental requirements need to be fulfilled. These requirements are the availability of accurate transition probabilities for the lines used, use of transitions having similar energy levels for E a t o m and E l o n, an optically thin plasma (i.e., no self-absorption), the simultaneous measurement of ionic and atomic emissions, and freedom from spectral interferences. Magnesium and cadmium were 119 chosen as the spectrometric species in this experiment both for the availability of several parameters used in eq. (6.1)[7, 8], and for the "convenient" appearance temperatures, which were within the operating range of the platform we used as an atomization source. These two elements also possess both atomic and ionic emission lines, with wavelengths that fall within one spectral window of the intensified photodiode array (IPDA) used for these experiments. 6.2.2 Instrumentation Details of the construction and operation of the P P - C C P have been given in Chapter 2 of this thesis. In this study, a 1200 lines mm"1 holographic grating was used to disperse emission from the plasma source using a 1.0 m Czerny-Turner scanning monochromator (Model 2061 Scoeffel-McPherson, Acton, MA, USA). The entrance and exit slits were set at 50 um. A 1:1 image of the plasma discharge was formed on the monochromator entrance slit by using a system consisting of two planoconvex fused silica lenses with their curved surfaces facing each other. The signal was detected using an intensified photodiode array (IPDA) detector (Model IRY-700/G/B/PAR, Princeton, NJ). To manipulate the controller and for data acquisition, C S M A software from Princeton Instruments, Inc. was used. The data was then stored in a P C for analysis and manipulation. 120 6.2.3 Reagents All analyte solutions were prepared using a serial dilution of A. A. standard 1000 ppm stock solution (BDH Inc. Toronto, Ontario, Canada) with deionized water prior to use. Analytical grade NaN03 (Anachemia, Vancouver, Canada) was used to prepare the Na solution in 1% (vA/) HN03 solution. Purified He (Paraxair Products Inc., Mississauga, ON, Canada) was used as the plasma gas. The helium was 99.998% pure containing 5 ppm oxygen and less than 3 ppm moisture. 6.2.4 Procedure A 10 u\ sample aliquot of working standard(s) for a given analyte was placed manually into the basin of the platform using an Eppendorf 0.5-10 p\ micro pipette fitted with a polyethylene tip. Samples were dried for 40 s at 90 °C, followed by a 30 s "char" stage at 300 °C for Mg and 200 °C for Cd analyte with the He flow on. After a two minute waiting period, the radio frequency power was applied and the plasma spontaneously ignited. Following another 30 s plasma stabilization period, the atomization stage was activated to reach the maximum temperature of 2000 °C. The atomization step was active for 6.5 s for the Mg analyte; during the Cd study it was active for 5 s. The blank determinations were carried out in the same manner without the analyte or analyte and interferent. Four replicate measurements for each determination were carried out. For each sample injection, 60 time-resolved spectra were collected. The data acquisition time for each signal was 0.068 s for Cd and 0.102 s for Mg, respectively. This data acquisition system enabled us to study the temporal emission behavior of the PP-CCP source. Throughout the experiment, 121 reflected power was tuned manually (the matching network was not permitted to auto-tune during an atomization) and kept to 2, 3 and 4 W for 100, 200 and 300 W forward power, respectively. The effect of forward power was examined for the range of 50 -300 W. The influence of the flow rate of the plasma carrier gas, and hence the carrier gas, on the degree of ionization was also investigated by measuring the atomic and ionic signals with a flow rate of 0.1 - 1.2 L min"1. Na as NaNOg was introduced into the system to determine whether there would be any effect of easily ionizable elements on analyte ionization. The Na solution was mixed with the analyte into the sample depression of the graphite platform. The physical characteristics of the atomic and ionic lines used in eq. (6.1) are summarized from reliable sources [7, 9] in Table 6.1. Partition functions of atoms and ions could have been calculated from the tables of deGalan et. al. [10], but we used the common low-temperature approximation of equating the electronic partition function with the statistical weight of the ground state instead. 6.3 Results and Discussion In eq. (6.1), for T we have used the excitation temperature derived from Pbl rather than an ionization temperature. Careful examination of the equation reveals that a close match in the energy levels of the atomic and ionic lines makes the calculated degree of ionization weakly dependent on T values. Calculation of the degree of ionization, using different temperatures, shows that a variation of the T 122 value by ±1000 K from the used value of 4500 K varies the degree of ionization by ±1%. The uncertainty in calculated ionization arises primarily from variation in the intensities of atomic and ionic lines, from replicate measurements, and can be reduced by measuring the lines simultaneously. The other variable that introduces uncertainty into the final calculated degree of ionization is the uncertainty in the transition probability or the "A" term. By carefully considering all the variables and functions, as well as the uncertainty involved with them, it is reasonable to say that eq. (6.1) should provide a fair estimate of the degree of ionization in the plasma. Table- 6.1: Physical characteristics of atomic and ionic lines Species X(nm) IP (eV) 9o 9i E/cm-1 A(107s-1) Unc* Mgl 285.21 7.64 1 3 35051 5.3 D Mgll 279.55 2 4 35761 2.6 C Mgll 280.27 2 2 35669 2.6 C Cdl 228.8 8.99 1 3 43692 5.3 C Cdll 226.5 2 2 44136 3.0 C Uncertainties: C<25%; D<50%. 123 6.3.1 Temporal Response By examination of the temporal response of the analyte, as well as the background, species can provide insights into excitation, vaporization, and atomization characteristics of the source as a function of time. In GFAAS, the temporal response of the analyte signal is an important diagnostic tool and has often been combined with kinetic and thermodynamic calculations to study analyte atomization mechanisms during the high temperature atomization step of the analyte [11-16]. The temporal response of the analyte signal in our source is likely to be different from that in GFAAS. In this PP-CCP source, the vaporized sample is transported into the plasma after being vaporized using a graphite platform for subsequent atomization, excitation and signal detection. During the transportation of the gaseous sample into the plasma, and also during the atomization, the plasma characteristics may change. Any change in plasma characteristics may also change the ionization behavior of the analyte and hence the degree of ionization. Figure 6.1 is the time resolved emission spectrum of 5.0 ng Mg for a plasma operating at 250 W forward power and 1 L min'1 gas flow rate. The three dimensional graph consists of a series of sixty sequential frames which represent the emission intensities for the entire atomization cycle of 6 s. For each frame (spectrum) the data acquisition time was 0.102 s. From the plot, the temporal emission behavior of Mg atomic (at 285.21 nm) and ionic (at 279.55, 280.27 nm) lines, as well as the background species (an OH band exists around 284 nm), can be observed during the atomization cycle. 124 E 3 z CD c o iri H— o c CO o Q . CD " O CD E o CD -f—• O CO CD CJ) c E _co CO c CJ) CO o "c o " O c "o a CD C D o CL " D CD i— o CD E co CD J D 0 CL 1 I CD to CJ) CM o c <D O co ~ o o Q - co % O) CD -o O CD CD CJ) 125 0 1 2 3 4 5 6 Time, Sec. Fig. 6.2: Temporal response of Mg I and Mg II and calculated ratio of Mg ll/Mg I as a function of time for a deposition of 5.0 ng of Mg in a 250 W plasma with a gas flow rate of 1.0 L min 1 . At the beginning of the atomization cycle, no analyte emission signals were observed. As the atomization cycle proceeded to 1.6 s, the emission of atomic Mg at 285.21 nm appeared, as well as the ionic lines at 279.55 and 280.27 nm. All three reached their maximum intensity at 2.4 s. Throughout the entire atomization cycle the O H band signal remained almost unchanged. A two dimensional representation of Fig. 6.1 is depicted in Fig. 6.2, with Mg ionic and atomic signals and their ratio as 126 a function of time. As can be seen from Fig. 6.2, the ionic line at 279.55 nm was more intense than the atomic line throughout the entire atomization cycle, and, hence, yielded a higher ion - atom ratio. It can also be observed from this figure that the ion - atom ratio (using both ionic lines) remains relatively constant during the atomization cycle. Figure 6.3 is a plot of the time-resolved emission spectrum for cadmium atomic and ionic species at 250 W plasma forward power and 1 L m in 1 gas flow rate, with the deposition of 5.0 ng Cd on the platform. The graph consists of a series of sixty sequential frames, with a data acquisition time of 0.068 s for each frame. Although both atomic and ionic signals appeared at the same time (0.6 s after the atomization cycle started) the weak ionic signal can be observed for only 0.6 s, whereas the atomic signal can be observed for the entire atomization cycle. The difference in appearance time for Mg and Cd lies in the difference in their reported appearance temperatures. Due to the lower appearance temperature of Cd (730 K), Cd came off the platform and was transported into the plasma earlier than Mg, whose appearance temperature is 1530 K [17]. Cd has a higher ionization potential (8.99 eV) than Mg (7.65 eV), which is why the cadmium did not produce as large an ionic signal as Mg. Figure 6.3 also indicates the presence of a C O + band around the 230 nm region. It is well known that C O + is readily excited in helium discharges as a result of the selective excitation of the B 2 Z + state of C O + by He 2 + (18.3 - 20.3 eV) through a resonant charge-transfer mechanism [18, 19]. It is most likely that C O + is formed predominantly from the precursor C O , which itself is formed from the oxidation of carbon on the graphite platform and carried into the plasma along with 127 u s s Z, C D C o 4— o c g '•53 CO O C L Q) •o CO _C0 CO c CJ) CO g 'c o T J c T O O CO -4—« CO o CO CO C D c e " D C CO i— CO <: o C L T D CO = P co E co CO JO g C L 11 CO i n CM CO c o o CO o •o CO CO c o C L CO 0 I— "CO O CO E 1 |2 I c o CD CX) LL 128 the analyte. CO may also be formed by the direct combination of oxygen atoms with carbon atoms, evolved from the graphite platform at a high temperature. In Fig. 6.3 it can be seen that the C 0 + intensity increases with time, indicating that the number density of the CO molecule increases as the atomization cycle proceeds and hence the temperature of the platform increases. 100-. 80 4 c o 'c 60 4 o 40 4 Cd Degree of Ionization --G- Using Mgll, 279.55 nm Using Mgll, 280.27 nm P= 250 W, F=1.0Lmin"1. 20 4 04 Atomization Time, Sec. Figure 6.4 Degree of ionization (for Mg and Cd) as a function of time at 250 W plasma forward power with a gas flow rate of 1.0 L min1. 129 The degree of ionization for Mg and Cd calculated from the emission response (shown in Fig. 6.1 and Fig. 6.3) is presented in Fig. 6.4. Although the relative signal intensities of Mg ionic lines (279.55 and 280,27 nm) are different from each other, the degree of ionization calculated from these lines is consistent. Figure 6.4 indicates that, within these plasma operating parameters (250 W forward power and 1.0 L min1 gas flow rate), the Mg analyte has a temporally uniform degree of ionization, with a value of approximately 83%. The degree of ionization for Cd varies slightly as a function of time. At the time of signal appearance (0.6 s), the degree of ionization was 38%, which increased to a maxima of 47% at 0.8 s and then decreased to ~ 32% subsequently. 6.3.2 Effect of Plasma Power The results of a study of the influence of applied plasma power on analyte ionization are presented in Fig. 6.5A and 6.5B. Both figures depict a typical atomic and ionic response as a function of plasma forward power using peak height and peak area, respectively. The amount of sample introduced for each experiment was 5.0 ng of Mg. The gas flow rate used in this study was 1.0 L min1. The error bars represent noise in the signal from four repetitive measurements for each applied power. It can be seen from these graphs that both the atomic and ionic signal increased very slowly with increased power up to 175 W, at which point there was an accelerated signal increase. It is interesting to note, that although both ionic lines (279.55 and 280.27 nm) have a relatively lower response than that of the atomic line 130 200x103 -n 50 100 150 200 250 300 Plasma Forward Power, W. Figure 6.5: Response of magnesium atomic and ionic emission signals (A: using peak height, B: using peak area), from a deposition of 5.0 ng of Mg, as a function of applied rf power and 1.0 L min1 gas flow rate. 131 up to 200 W, the ionic signal from 279.55 nm is stronger than the atomic line when the applied power is greater than 200 W. 100 80 .2 60 -A o 40 20-1 :.S--.-.-.-.-.-;S ' SB" •0- - Mgll, 279.55 nm • • - Mgll, 280.27 nm Mgl 1(279.55 nm) / Mgl Ratio • - - Mgl 1(280.27 nm) / Mgl Ratio 0 4 ^ 1 l 1 ' 100 i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — r 150 200 250 300 1.2 1.0 CQ 0.8 = V- 0.6 0.4 0.2 0.0 CQ 0) Plasma Forward Power, W Figure 6.6: Effect of plasma forward power on magnesium ion : atom ratio and the degree of ionization measured from a deposition of 5.0 ng of Mg with a gas flow rate of 1.0 L min 1. The ion : atom ratio and the degree of ionization, using both Mg ionic lines, were calculated and are presented in Fig. 6.6. It can be seen from this figure that the degree of ionization increased rapidly from 36% to 80% when power was increased from 125 to 200 W, leveling off at 83% at higher than 200 W. It is possible that the 132 degree of ionization reaches a threshold at some higher power level due to limitations in the matching efficiency of the rf power. During the experiment, the auto tuner of the matching network was kept in the "off" position and tuning was done manually before the atomization cycle started. Reflected power was ~4 W with the application of 200 - 300 W forward power. Figure 6.7: Integrated atomic and ionic emission signals from a deposition of 5.0 ng of Cd as a function of plasma forward power with a gas flow rate of 1.0 L min1. 133 For a system at equilibrium, the degree of ionization (a) of an analyte can be calculated from the Saha equation and a plot of log [a/(1 + a)] versus ionization potential (IP) should yield a straight line. Sturgeon et al. [2] reported an exponential relationship with a negative slope, between the degree of ionization and ionization potential (IP). An analyte with a lower IP has a higher degree of ionization in the FAPES source, which is quite similar to PP-CCP in nature. To compare the results, the degree of ionization of Cd, with an IP of 8.99 eV, was also studied at different plasma powers and gas flow rates. The response of cadmium atomic and ionic emission as a function of plasma power was measured and illustrated in Fig. 6.7. Unlike the magnesium atomic signal, the cadmium atomic signal increased very rapidly up to 100 W and then leveled off with higher applied powers. The ionic signal on the other hand, kept increasing with increased applied forward power. As presented in Fig. 6.8, the ratio of ionic to atomic signal from cadmium analyte is much lower than that of magnesium analyte, due to the higher ionization potential of cadmium yielding a lower degree of ionization. As can be seen from this figure, the degree of ionization increased from 20% to 45% when the plasma power has increased from 100 to 200 W. At powers exceeding 200 W, the degree of ionization leveled off at-48%. In a glow discharge, analyte atomic excitation, ionization and ion excitation occur primarily in the negative glow region by electron impact, Penning and/or charge transfer ionization. For a given population of ground state analyte atoms, the rate of production of an excited analyte atom and/or ion will be proportional to the 134 concentration of the relevant species. Seven major species exist in helium discharges: free electrons (e), ground state helium atoms (He), metastable helium (Hem*), helium ions (He+), excited helium ions (He+*), excited diatomic helium (He2*) and diatomic helium ions (He 2 +). Carnahan and Hieftje [20] showed that, for an atmospheric pressure helium MIP (at 5000 K), the energies associated with free Figure 6.8: Effect of plasma forward power on cadmium ion - atom ratio and the degree of ionization measured from a deposition of 5.0 ng of Cd with a gas flow rate of 1.0 L min 1 . 135 electrons and ground state helium atoms are less than 0.5 eV. It is thus reasonable to assume, that as the PP-CCP operates at atmospheric pressure with an excitation temperature similar to that of a helium-MIP, the free electrons and ground state helium would be similarly energetic. The energies associated with the other five species are: Hem*, 19.8 and 20.6; He+, 24.58; He +\ 40.82; He2", 13.3-15.9; and He2 +, 18.3-20.5 eV [21]. The energy differences between most of the analyte excitation (< 8 eV) and helium species are too high to account for the direct population of the excited states of analyte atoms and ions via the efficient exchange of internal energy during collisions. It is most probable that high energy electron impact (for excited atoms) and Penning and/or charge transfer (for excited ions) result in highly excited states, which subsequently undergo radiative and/or collisional cascade from an upper level, to the level in question. For Penning ionization and subsequent excitation, metastable helium atoms with 19.8 and 20.6 eV and diatomic helium with 13.3-15.9 eV would be involved in the process, whereas, for charge transfer ionization, He+ with 24.58 and He 2 + with 18.3-20.5 eV would be involved. Baltayan et al. [22] reported that Penning ionization for Cd is not significant, but that the charge transfer between He+ and Cd is, leading to the population of upper Cd II levels. They also reported, that in those charge transfer reactions, the fraction of Cd + ions excited in the five high-lying 9p 2P, 8d 2D, 6g 2 G , 6f 2 F and 9s 2 S states is almost half of the product of Cd+ ions; the remaining half is distributed between other excited levels of Cd + down to a few eV below the energy of the He+ ion. Collisional equilibration serves to populate the measured 5p 2P° 1 / 2 {(8.99 + 5.47) eV, 226.5 nm} level. Sun and Sturgeon [23] suggested that Mg ionization in the He plasma may arise as a 136 result of Penning reactions resulting in higher Mg II levels followed by collisional equilibration to the measured levels. 6.3.3 Gas Flow Effect A typical example of the effects of plasma gas flow rate on cadmium atomic and ionic signal transients is shown in Fig. 6.9A and Fig. 6.9B, respectively. All signals were recorded for the introduction of a 5.0 ng sample on the platform with a plasma forward power of 250 W. It is evident from these figures that, independent of signal intensities, both atomic and ionic signals appeared earlier with increasing gas flow rate. The cadmium atomic signal at 0.2 L min1 gas flow rate appeared at 0.75 s, whereas it appeared at 0.54 s when the gas flow rate was increased to 1.0 L min1. The ionic signal also appeared 0.14 s earlier with an increase in gas flow rate from 0.4 to 1.0 L min1. This shift in temporal signal is due to the faster and more effective transportation of the analyte from vaporizer to plasma. Unlike cadmium, the magnesium signal appears significantly later for a given gas flow rate, mainly due to the higher appearance temperature of the latter. For example, with a gas flow rate of 1.0 L min1, the Mg atomic signal appeared at 1.53 s, whereas the Cd atomic signal appeared at 0.54 s. The temporal behavior of magnesium atomic and ionic emission signals are depicted in Fig. 6.1 OA and 6.10B, respectively. As with cadmium, both magnesium atomic and ionic emission signals appear earlier with increasing gas flow rate. This observation further strengthens our earlier assertion of faster sample transport from the vaporizer into the plasma with 137 12x1(f -A 03 c g> co E o < > a) cr 2000-1 1500 03 C g> •2 1000-1 .1 05 a> 500 4 (B) 0.0 Flow = 1.0 L min"1 Flow = 0.8 L min"1 Flow = 0.6 L min'1 Flow = 0.4 L min"1 — Flow = 0.2 L min"1 i — - i — i — r T — i — r 2.0 2.5 Flow = 1.0 L min"1 Flow = 0.8 L min"1 Flow = 0.6 L min"1 — Flow = 0.4 L min"1 - j — i — i — i — T — r -0.5 1.0 i — i — i — | — i — i — i — i — | — i — i — r 1.5 2.0 Atomization Time, Sec 2.5 Figure 6.9: Influence of plasma (also the carrier) gas flow rate on Cd (A) atomic and (B) ionic transient signals in a 250 W plasma. The signals were generated from the deposition of 5.0 ng of Cd standard sample. 138 increased flow rate. It is also evident from figure 6.9B and 6.1 OB that at a given gas flow rate the Mg ionic signal spanned a longer time period than that of the C d ionic signal. For example, with a gas flow rate of 1.0 L min' 1, the magnesium ionic signal stayed for ~ 4 s whereas for cadmium, that time was only ~ 0.4 s. Higher ionization potential of C d is probably responsible for the short-lived ionic signal. Figure 6.11 depicts the effect of gas flow rate on the total atomic and ionic signals (A: using peak height and B: using peak area) from a 5.0 ng cadmium deposition in a 250 W plasma. It is clear from this figure, that the cadmium signal intensity, for both atomic and ionic emissions, increased almost monotonically with an increasing flow rate when peak height is being considered. Considering the peak area, the cadmium atomic signal increased monotonically up to 0.6 L min 1 gas flow rate, followed by the rate of signal increase decreasing slightly from 0.6 L min 1 to 1.0 L min 1 , whereas the ionic signal increased linearly up to 1.0 L min~1gas flow rate. An identical situation could be observed for the magnesium analyte with the exception that the atomic signal (using peak area) decreased with a gas flow rate of over 0.75 L min 1 . The influence of plasma gas flow rate on Mg analyte is illustrated in Fig. 6.12 (A: using peak height and B: using peak area) from a 5.0 ng magnesium deposition in a 250 W plasma. Increasing the plasma gas flow changes the radiative and mass transport properties of the source, and may also be responsible for increasing the signal. High flow rates of He serve to maintain the plasma gas in a state of high purity, diluting both the gaseous products of decomposition of analyte salts and any water or acid 139 Figure 6.10: Influence of plasma (also the carrier) gas flow rate on Mg (A) atomic and (B) ionic transient signals in a 250 W plasma. The signals were generated from the deposition of a 5.0 ng of Mg standard sample. 140 CO O CO CD CO c CD c 14 —. 12-10-8-6-4 -2 -0 -¥ 70-, c g CO co E LU g 'E o E x I 50H to ° 40H 30 20-10-0 0.0 0.0 - • - Fitted Atom Signal ©- Fitted Ion Signal 3500 3000 2500 2000 1500 1000 I- 500 h 0 O 0) Q. | 3 0.2 0.4 0.6 0.8 1.0 -#- Fitted Atom Signal ©- Fitted Ion Signal T T o o" m 3 10000 § 3 3 f-i-CD 3 CO CD' to 8000 h 6000 \- 4000 \- 2000 0.2 0.4 0.6 0.8 Gas Flow Rate, L min"1 1.0 Figure 6.11: Influence of gas flow rate on integrated Cd atomic and ionic signals, (A) using peak height and (B) using peak area, in a 250 W plasma discharge. The amount of deposited cadmium sample was 5.0 ng. 141 vapor desorbed from the heated surface during the vaporization and atomization stages, as well as minimizing the ingress of atmosphere into the plasma torch [24]. These factors might well be responsible for increasing the atomic and ionic signals. Higher gas flow rates may also increase the turbulence in the flow, causing more collisions, and hence, making the plasma hotter. Hotter plasma explains the linear increase in the ionic signal and decrease (for Mg) or increase with a lower slope (for Cd) in the atomic signal with a higher gas flow rate. The influence of the gas flow rate on Cd and Mg ionization is illustrated in Fig. 6.13. The results shown in this plot were calculated from four repetitive measurements of the atomic and ionic signal, from 5.0 ng of Cd and Mg deposition, in a 250 W plasma operating at different gas flow rates. The error bars represent the error in the calculated degree of ionization due to noise in the signal. The degree of ionization increased initially with flow rate and then leveled off. For Mg it increased slowly from 60% to 78% with an increase of flow from 0.1 to 0.5 L min 1 and reached a plateau of ~ 82% with a higher flow. The Cd ion signal could not be detected at less than 0.4 L min 1 gas flow rate and the degree of ionization increased from 27% to 48% for a flow rate of 0.4 - 1 . 0 L min 1 . Figure 6.12: Influence of gas flow rate on Mg atomic and ionic signals, (A) using peak height and (B) using peak area, in a 250 W plasma discharge. The amount of deposited cadmium sample was 5.0 ng. 143 100 80 4 o 60 4 - 404 20 4 04 >-Using Mgll, 279.55 nm - © - Using Mgll, 280.27 nm - Cd Ionization IIIII11111111111111111111111111111111 II 111111111111111111111111 0.0 0.2 0.4 0.6 0.8 1.0 1.2 Gas Flow Rate, L min -1 Fig. 6.13: Influence of gas flow rate on cadmium and magnesium degree of ionization for a plasma operating at 250 W forward power. 6.3.4 Effect of Easily lonizable Element (EIE) The effect of easily ionizable elements from different plasma sources used in atomic emission spectrometry has been extensively reported [25-40]. Smith et al. [41] observed a slight enhancement in Ag signal at low concentrations in a FAPES source and a depression at higher concentrations for the presence of Na as the EIE. It was suggested that, at low concentrations of EIE, the signal was enhanced by a shift in the ionization equilibrium, and, at higher concentrations, the plasma was 144 being "quenched". Imai et al. [34] suggested that radiative power losses from the plasma, due to excitation of the EIE species and alteration of the electron energy distribution function (due to EIE ionization and collisional dissociation of molecules), are the major sources of EIE interferences. The effect of EIE on analytes in a P P -C C P source has recently been reported by Rahman and Blades [42], where they suggested that EIE shifted the ionization equilibrium between analyte atoms and ions toward the atom, resulting in an enhancement of the atomic emission signal. Although both atomic and ionic signals initially increased with the addition of EIE (Na in this case), the increase in the ionic signal is stronger than that of the atomic signal; this is depicted in Fig. 6.14. This figure illustrates the influence of EIE on the Mg ion to atom response ratio for a 250 W plasma with a gas flow rate of 0.5 L min 1 . It can be observed in this figure, that the ion : atom ratio increased with increasing amounts of Na from 0 to 0.05 ug and then started decreasing with amounts of Na exceeding this level. It is likely that the presence of small amounts of EIE increases the density of the species in the plasma, resulting in a higher number of collisions, thus making the plasma hotter and, hence, increasing the ion signal more than the atom signal. Further increases in EIE amounts decrease both the ionic and atomic signal with a higher decreasing rate for ions, i.e., resulting in a lower ion-atom ratio. There is strong evidence that added electrons from EIE, shift the atom-ion equilibrium towards the atom, resulting in lower ion-atom ratios. However, increasing the number density of electrons must alter the electron energy distribution function (EEDF), decreasing both the atomic and ionic signals. 145 0.8 -, 0.6 4 OA A £ 0.2 c g 'to co m 0.0 E o 1ST ©--•-MgII:MgI (279.55/285.21) •© -MgllrMgl (280.27/285.21) i i i i l i i i i I i i i i i i i i i I i i i i I i i i i I i i i i I i i i i I 0.0 0.2 0.4 0.6 0.8 g 0.8 n E CO CO & 0.6 CO 0.4 4 0.2 4 0.0 4 H-o o. - -o i i i i i I i i i i i i i i i I i i i i i i i i i I i i i i i i i i i I i i i i i i i i i I 2 4 6 8 10 Added Amount of Na, \ig Fig. 6.14: Mg II - Mg I ratio as a function of added Na mass for a 250 W plasma forward power and 0.5 L min"1 gas flow rate. 146 100-i 80 H 60 4 40 20-1 I 0 cs N C O 4-111111111111111111111111111111111111111111111111111111 0.0 0.1 0.2 0.3 0.4 0.5 S 100-, CD OJ CD Q 80^ 60-40-20-0 -Using Mgll, 279.55 nm ©- Using Mgll, 280.27 nm i i i i i i i I i i i i i 2 I i i i i i i i i i I i i i i i i i i i I i i i i i i i i i I 8 10 Added Amount of Na, ug, Fig. 6.15: Effect of Na amount on measured degree of ionization for magnesium analyte at 250 W plasma forward power and 0.5 L min'1 gas flow rate. 147 Figure 6.15 depicts the effect of easily ionizable elements on the measured degree of ionization for the magnesium analyte. The results shown here were calculated from four repetitive measurements of the atomic and ionic signal, from 5.0 ng of magnesium deposition, in a 250 W plasma operating at a 0.5 L min 1 gas flow rate. The error bars represent the error in the calculated degree of ionization due to noise in the signal. As can be seen from the figure, the degree of ionization increased nominally from 72% to 74% with the addition of 0.05 ug of Na but further addition of Na, up to 10 jL/g, decreased the degree of ionization by 15%, consistent with a shift in the ionization equilibrium. Unlike magnesium, cadmium reacts very strongly to the presence of sodium. The presence of very small amounts of Na as EIE decreased the ion - atom ratio and hence, the calculated degree of ionization, which is illustrated in Fig. 6.16B. Fig. 6.16A represents the integrated atomic and ionic signal from a 2.5 ng of cadmium sample deposition with varying amounts of Na; the atomic emission increased in the presence of Na, but the ionic signal decreased. It is possible, that the higher number of collisions, due to the presence of EIE, does not make the plasma hot enough to ionize the Cd , which has a higher ionization potential than that of Mg. On the other hand, the secondary electrons from Na suppress the analyte ionization by shifting the atom - ion equilibrium towards the atom, resulting in a lower ionic signal, and hence, lower ion - atom ratio and degree of ionization. With a significantly higher amount of sodium, (250 ng or more), this suppression is so strong that the ionic signal became negligible or undetectable by the present detection system. 148 30x1 cr -, 25 4 CO c CD CO o E o < CD 20 4 154 2 104 CO CD c £ 54 04 0) (A) -I - • - Cd •©- Cd r- 1400 h 1200 10001 CQ 800 £ I- 600 | co 400 <§' h 200 11111111111111111111111111111111111111111111111111111 0 10 20 30 40 50 x10"3 "[ 60 h-10H-(B) - e - Degree of Ionization • ©- Cdll -Cdl Ratio 0 rrl i 11111 1111 11111 11111 1111 11111 1111111111 11 1111111 11-0 0.20 O Q_ 0.15 = I O Q. m 0.10 3 to CO o' 3 -TJ 0.05 §. o 0.00 10 20 30 40 Added Na Amount, \ig 50x10" Fig. 6.16: Effects of Na amount on (A) Cd atomic and ionic emission and, hence on (B) the measured degree of ionization for a 250 W plasma power and 0.5 L min1 gas flow rate. 149 6.4 Conclusions The degree of ionization in a parallel plate capacitively coupled plasma has been studied under various operating parameters. The applied forward power in the plasma has shown a significant effect on the degree of ionization for both magnesium and cadmium analyte. Analyte ionization increases with an initial increase in power from 100 to 200 W and, hence, the degree of ionization increases. The plasma gas (also used as the carrier) flow rate also affects the degree of ionization. For Mg it increased slowly from 60% to 78% between 0.1 to 0.5 L min' 1 and reached at a plateau of ~ 82% for higher flow. The C d ion signal could not be detected at less than 0.4 L min 1 gas flow rate and the degree of ionization increased from 27% to 48% between 0.4 - 1.0 L min' 1. A small amount of EIE increases magnesium ionization slightly but higher amounts decrease the degree of ionization due to ionization equilibrium shift. Due to the higher ionization potential, cadmium analyte does not show any ionization improvement in the presence of EIE, but rather decreases the degree of ionization. The present set of experiments demonstrate that this parallel plate source has a significant degree of ionization of 83% and 48% for Mg and Cd , respectively, for a 250 W plasma with 1.0 L min"1 gas flow rate. These values are higher than those reported for the F A P E S source [2] and are comparable to the measured Cd value (49%) from an MIP source [43]. This study leads us to believe that this P P - C C P would be a potentially unique ion source for mass spectrometry. 150 6.5 References 1. C. Le Blanc, Ph.D. Thesis, 1996, University of British Columbia. 2. R. E. Sturgeon and R. Guevremont, J. Anal. Atom. Spectro., 13, (1998), 229 3. D. J . Douglas and J.B. French, Anal. Chem, 53, (1981), 37. 4. M. M. Rahman and M.W. Blades, Spectrochim. Acta, 52B, (1997), 1983-93. 5. J . W. Olesik, Anal. Chem., 68, (1996), 469A-474A. 6. M. M. Rahman and M. W. Blades, J. Anal. Atom. Spectro., 15, (2000), 1313 7. W. L. Wiese and G.A. Martin, Wavelength and Transition Probabilities for Atoms and Atomic Ions, J . Reader, Ed., (Washington : U.S. Dept. of Commerce, National Bureau of Standards. p395,1980). 8. NIST WebSite, Atomic Spectroscopic Database, http://aeldata.phy. nist. gov /nist_atom ic_spectra. htm I. 9. NIST Website, Database version 1.3 http://physics.nist.gov/physRefData/ A S D 1 / nist-atomic-spectra.html. 10. L. DeGalan, R. Smith, and J . D. Winefordner, Spectrochim. Acta, 23B, (1968), 521-525. 11. S. L. Paveri-Fontana, G. Tessari, and G. Torsi, Anal. Chem., 46, (1974), 1032 12. G . Torsi and G. Tessari, Anal. Chem., 47, (1975), 839-842. 13. D. A. Bass and J . A. Holcombe, Spectrochim. Acta, 43B, (1988), 1473 - 1483. 14. S. J . Cathum, C. L. Chakrabarti, and J . C. Hutton, Spectrochim. Acta, 46B, (1991), 3 5 - 4 4 . 15. E. Masera, P. Mauchien, and Y. Lerat, Spectrochim. Acta, 51B, (1996), 1007. 151 16. A. Lebihan, H. Legarrec, J . V. Cabon, and V. Guern, Spectrochim. Acta, 53B, (1998), 1347-1353. 17. B. V. L'Vov, Spectrochim. Acta, 33B, (1978), 153-193. 18. C. B. Collins and W. W. Robertson, J. Chem. Phys., 40, (1964), 701. 19. M. Endoh, M. Tsuji, and Y. Nishimura, J. Chem. Phys., 79, (1983), 5368-5375 20. J . W. Carnahan and G. M. Hieftje, Spectrochim. Acta, 47B, (1992), 731-739. 21. C. F. Bauer and R. K. Skogerboe, Spectrochim. Acta, 38B, (1983), 1125. 22. P. Baltayan, J .C . Pebay-Peyroula, and N. Sadeghi, J. Phys. B: At. Mol. Phys, 18, (1985), 3615-3628. 23. F. Sun and R. E. Sturgeon, J. Anal. Atom. Spectrom., 14, (1999), 901-912. 24. R. E. Sturgeon and H. Falk, Spectrochim. Acta, 43B, (1988),. 25. H. Kawaguchi and B. L. Vallee, Anal. Chem., 47, (1975), 1029. 26. H. Kawaguchi, I. Atsuya, and B. Vallee, Anal. Chem., 49(2), (1977), 266-270. 27. I. Atsuya, C. Veillon, H. Kawaguchi, and B. Vallee., Anal. Chem., 49,(1977), 1489 28. M. W. Blades and G. Horlick, Spectrochim. Acta, 36B, (1981), 881-900. 29. J . P. Matousek, B.J. Orr, and M. Selby, Spectrochim. Acta, 41B, (1986), 415. 30. C. Le Blanc and M. W. Blades, J. Anal. Atom. Spectrom., 5, (1990), 99-107. 31. K. Kitagawa and G. Horlick, J. Anal. Atom. Spectrom., 7, (1992), 1221-1229. 32. P. J . Galley, M. Glick, and G. M. Hieftje, Spectrochim. Acta, 48B, (1993), 769. 33. T. D. Hettipathirana and M. W. Blades, J. Anal. Atom. Spectrom., 8, (1993), 995 34. S. Imai and R. E. Sturgeon, J. Anal. Atom. Spectrom., 9, (1994), 765-772. 152 35. N. N. Sesi and G. M. Hieftje, Spectrochim. Acta, 51B, (1996), 1601-1628. 36. M. Wu and G. M. Hieftje, Spectrochim. Acta, 49B (2), (1994), 149-161. 37. X. Romero, E. Poussel, and J .M. Mermet, Spectrochim. Acta, 52B, (1997), 495-502. 38. J . M. Mermet, J. Anal. At. Spectrom., 13 (5), (1998), 419-422. 39. C. Dubuisson, E. Poussel, J . L. Toboli, and J . M. Mermet, Spectrochim. Acta, 53B, (1998), 593-600. 40. C. Dubuisson, E. Poussel, and J . M. Mermet, J. Anal. Atom. Spectrom., 13(11), (1998), 1265-1269. 41. D. L. Smith, D. C. Liang, D. Steel, and M. W. Blades, Spectrochim. Acta, 45B, (1990), 493. 42. M. M. Rahman and M. W. Blades, Spectrochim. Acta, 55 B, (2000), 327-338. 43. C. Prokisch and J . A. C. Broekaert, Spectrochim. Acta, 53B, (1998), 1109. 153 Chapter 7 The Effect of Electrode Length on Fundamental and Some Analytical Characteristics 7.1 Introduction In previous chapters, work on the parallel plate capacitively coupled plasma focused on some fundamental properties namely excitation temperature (T e x c) and rotational temperature (T r o t). Some analytical characteristics of this source are also presented in the previous chapters such as detection limits, effect of easily ionizable elements on the analyte, and the degree of analyte ionization. The data presented in these chapters has shown that this source has a relatively low detection limit for lead and silver, and fairly high degree of analyte ionization. A number of possible reactions can contribute to the excitation and ionization of analyte species in He plasma[1]. The most important of these reactions, are collisional excitation and de-excitation via electron impact, Penning excitation and ionization, and charge transfer ionization. Any changes in the plasma parameters, which can in turn cause a change in the number density of collision partners, would cause a change in observed plasma characteristics. The effect of applied plasma power (typically 100 - 300 W), which contributes a reasonable change in excitation temperature, and the effect of plasma gas flow (on some analytical characteristics) have already been studied and presented before. It is reasonable to assume that an additional instrumental parameter is the electrode length, which changes the plasma 154 volume as well as the rf power density and the analyte residence time. It is the influence of electrode length on the operating and analytical characteristics that is reported herein. Analytical figures of merit are compared for atomic and ionic lines of Ag and Mg, as well as some background species for electrode lengths ranging from 2 to 6 cm. 7.2 Experimental 7.2.1 Instrumentation A complete description of the experimental system employed to acquire atomic and ionic emission signals is given in Chapter-2 of this thesis. In this study a 1200 and a 3600 (during measurement of plasma background species) lines m m 1 holographic grating were used in a 1.0 m Czerny -Turner scanning monochromator (Model 2061 Scoeffel - McPherson, Acton, MA, USA) to disperse emission from the plasma source. The entrance and exit slit were set at 50 fxm. A 1:1 image of the plasma discharge was formed on the monochromator entrance slit by using a two lens system consisting of two piano - convex fused silica lenses with their curved surfaces facing each other. The signal was detected using an intensified photodiode array (IPDA) detector (Model IRY-700/G/B/PAR, Princeton, NJ). The detector was controlled with a ST-120 controller from the same supplier. To manipulate the controller and for data acquisition C S M A software from Princeton Instruments, Inc. was used. 155 7.2..2 Reagents Mg analyte solutions were prepared by a serial dilution of an A. A. 1 0 0 0 ppm standard stock solution (BDH Inc. Toronto, Ontario, Canada), using deionized water prior to use. Analytical grade A g N 0 3 (Fisher Scientific, Nepean Ontario, Canada) was used for the Ag standard in 1% (vA/) H N 0 3 solution. Purified He (Paraxair Products Inc., Mississauga, O N , Canada) was used as the plasma gas. The helium was 9 9 . 9 9 8 % pure, containing 5 ppm oxygen and less than 3 ppm moisture. 7.2.3 Procedure A 1 0 JL/L sample aliquot of working standard(s) of a given analyte was placed manually into the depression of the platform using an Eppendorf 0 . 5 - 1 0 uL micro pipette fitted with a polyethylene tip. With the He flow on, samples were dried for 4 0 s at 1 1 0 °C, followed by a 3 0 s "char" stage at 3 0 0 °C for the Mg and 2 5 0 °C for Ag. After a two minute waiting period, the radio frequency power was applied and the plasma spontaneously ignited. Following another 3 0 second plasma stabilization period, the atomization stage was activated to reach the maximum temperature of 2 0 0 0 °C. The atomization step was active for 6 .5 seconds for the Mg analyte and 5 seconds for the Ag analyte. The blank determinations were carried out in the same manner without the analyte present. Four replicate measurements for each determination were also carried out. For each sample injection, 6 0 subsequent signals were collected. Data acquisition time for each signal was 0 . 0 6 8 s for Ag and 0 . 1 0 2 s for Mg analyte. This data acquisition system enabled us to study the temporal behavior of the analyte. Throughout the experiment, reflected power was 156 tuned manually (the matching network was not permitted to auto tune during atomization) and kept at 2, 3 and 4 W for 100, 200 and 300 W applied forward power, respectively. The measurement of rotational temperature (T r o t) in P P - C C P source has been discussed in detail in Chapter 3. The measurement procedure for the degree of ionization described in this chapter has also been discussed in detail in the previous chapter. 7.3 Results and Discussion 7.3.1 Different Electrode Lengths and Plasma Power The analyte emission signal could be enhanced by improving the atomization and excitation characteristics of plasma and also by improving analyte transport efficiency. Another important parameter that significantly affects the emission signal is the analyte residence time. The residence time of analyte atoms is determined primarily by the combination of plasma gas flow rate, diffusion effects, and the length of plasma (the length of electrodes). In order to determine the effect of electrode length on the analytical characteristics of the P P - C C P source, different analytical figures of merit were investigated under the condition of different applied powers and gas flow rates. Magnesium and silver were used as "typical" analytes for this study. 157 Figure 7.1: Influence of plasma forward power on (A) He I and (B) total N 2 + emission intensity for different electrode lengths at a constant gas flow rate of 1.0 L min' 1. 158 Plasma forward power is one of the important parameters in the optimization of experimental conditions. The effect of changes in plasma power and the electrode length on emission characteristics of Ag I (328.1 nm), Mg I (285.21 nm), Mg II (279.55 nm), and Mg II (280.27 Mm) were studied using a helium gas flow rate of 0.2 L m in 1 (for Ag) and 1.0 L min 1 (for Mg). The plasma background was also studied at different plasma powers for different electrode lengths. The effect of electrode length, as well as the rf forward power, at a constant gas flow rate of 1.0 L min 1 , on the helium atomic and N 2 + molecular emission is depicted in Fig. 7.1 A and 1B. All the data present in this plot is the average of four repetitive measurements. The noise in the signal was so small that it appears undetectable in the plot: As can be seen in the plot, both background signals increase with increasing electrode length and reach a maximum for a 4 cm long electrode before it begins to decrease. It is also evident from this plot that the He I and N 2 + emission increases monotonically as a function of plasma power irrespective of the electrode length. However, the situation is different for the analyte signals, which is illustrated in Fig. 7.2A, 2B and 2C, where all signals are normalized with respect to the signal from 250 W plasma power for each electrode. As can be seen in Fig. 7.2A, no signal was detected for silver analyte at a power below 100 W, even though the plasma could be ignited at as low as 25 W plasma forward power. When the electrode length was 6 cm long, no signal could even be detected for a 100 W. In this plot all signals were collected for a 3 ng silver deposition on the ETV where a plasma gas flow rate of 0.2 L min 1 was used. The error bars represent the noise in the signal from four sets of experiments. 159 ca c CX) co c o CO CO 'E LU o 'E o < T3 CD CO 16 E o 1.2-, 1.0 0.8-^  0.6 0.4-0.2 0.0 ( A ) 2cm • • • © • • • 3cm - B - 4cm - A - 5cm -<e>- 6cm I 50 1.2n 1.0-1 0.8-^  0.6 0.4 0.2 0.0 4 —1—I— 100 T 1 1 r 150 — l — r 200 —1—l 250 2cm ...©... 3cm - - E - 4cm - A - 5cm -<3>- 6cm -T—i—i—i—i—r mn isn r—r—r—T- — i ?5n 1.2-, 1.0-(C) 0.8- - • - 2 cm • G - 3 cm 0.6- - A - 4 cm - A - 5 cm 0.4- -<s>- 6 cm 0.2-0.0-50 - i—|—i—i—i—i—|—i—i—i—i—|—i—i—i—i—| 100 150 200 250 Applied Plasma Power, W Figure 7.2: The effect of rf forward power on (A) silver, (B) magnesium atomic and (C) magnesium ionic emission. The plasma gas flow rates were 0.2 and 1.0 L min1 for a deposition of 3.0 ng Ag and 5.0 ng Mg, respectively. 160 In the case of magnesium, the atomic emission signal could be detected at as low as 50 W rf power for 2, 3, and 4 cm long electrodes but not for the 5 and 6 cm long ones, which is illustrated in Fig 7.2B. This figure represents the effect of plasma forward power on the magnesium atomic signal for different electrode lengths, whereas Fig. 7.2C represents the same, but on magnesium ionic signal. A gas flow rate of 1.0 L min 1 was used in these experiments. All emission signals were collected for a deposition of 5.0 ng of standard magnesium solution on a graphite platform and are normalized with respect to the signal at 250 W plasma power for each electrode. The error bars represent the noise present in the signal from four sets of measurements. As it can be seen from the plot, the magnesium ionic (Mg II) signals did not appear at a power lower than 100 W, and there was no detectable Mg II signal for electrodes longer than 3 cm for a 100 W forward power. For a 6 cm long electrode, the Mg II signal was first detected at 150 W. It is also evident from the plot that the silver atomic emission increases linearly with power, whereas both the magnesium atomic and ionic emission increases exponentially with power irrespective of the electrode length. The effect of electrode length on analyte emission is presented in Fig. 7.3A, B and C for an introduction of 3.0 ng of silver and 5.0 ng of magnesium analyte. Peak area is used for this illustration and the error bars represent the noise present in the signal from four sets of measurements. At higher plasma powers (i.e. 200 and 250 W), the silver atomic emission increases from 2 to 3 cm long electrode and then levels off up to the 5 cm electrode, before it shows a decreasing trend. For lower applied power, such as 100 W, there is no detectable signal for a 6 cm long 161 electrode, which is shown in Fig. 7.3A. In the case of magnesium, the atomic signal increases with increasing electrode length up to 5 cm before it leveling off (Fig. 7.3B). The magnesium ionic signal behaves in almost the same manner as the ionic signal, with an exception of no detectable signal for an electrode longer than 3 cm with a 100 W plasma forward power. All these observations suggest that there are two opposing factors at play as we increase the electrode length for any given plasma forward power. As we know, increasing electrode length increases the residence time and hence the analyte signal. Increasing the electrode length also increases the volume of the discharge. At a constant applied power, the power density per unit volume of the discharge decreases with increasing electrode length. This decreasing power density is responsible for the decrease in signal intensity for longer electrodes. Increasing the electrode length also provides a better analyte signal-to-noise ratio for a given power and flow, which is discussed in section 7.3.3. As the electrode length increases with a given plasma gas (also the carrier) flow, the flow becomes less turbulent causing less noise in the signal. The influence of plasma forward power on the signal noise measured at different plasma forward powers and electrode lengths for silver and magnesium analytes is depicted in Fig. 7.4A and 7.4B. With a few exceptions, signal-to-noise ratio (SNR) increased with increasing applied plasma power. Although the emission signal increased monotonically (for silver) or exponentially (for magnesium) with increased powers, the signal-to-noise ratio increased non-linearly. As can be seen in Fig. 7.4A and 7.4B, all electrodes, with a few exceptions, resulted in better signal-to-noise ratio at higher powers. 162 50-1 40 WC80. 20 10 04 250-200 « 1504 CO 1 o *100^ 50 0 (A) 100 W, - © - 150 W 200 W, - A - 250 W 2 T 3 4 T 5 6 (C) 100 w, 200 W, •©- 150 W • A - 250 W n 1 r 3 4 5 Electrode Length, cm Figure 7.3: The effect of electrode length on (A) silver, (B) magnesium atomic and (C) magnesium ionic emission. The plasma gas flow rates were 0.2 and 1.0 L min 1 for a deposition of 3.0 ng Ag and 5.0 ng Mg, respectively. 163 Figure 7.4: The effect of plasma power on signal-to-noise ratio (SNR) from (A) Ag and (B) Mg analyte for different electrode lengths. The plasma gas flow rates were 0.2 and 1.0 L min 1 for Ag and Mg, respectively. 164 The influence of plasma applied power on signal-to-background ratio (SBR) is provided in Fig. 7.5A and 7.5B for silver and magnesium, respectively. All the data presented in this plot was collected from a deposition of 3.0 ng Ag and 5.0 ng Mg. The applied plasma power was varied from 50 to 250 W with a plasma gas flow rate of 0.2 and 1.0 L min 1 for silver and magnesium, respectively. With an insignificant increase in background signal, signal-to-background ratio increased with increasing plasma power irrespective of the electrode lengths and analytes. For a given plasma power and gas flow rate, the background signal was higher for longer electrodes, resulting in a lower signal-to-background ratio. 7.3.2 Different Electrode Lengths and Gas Flow One of the important features of P P - C C P source is the possibility of independent optimization for vaporization, excitation and ionization. The plasma gas acts, not only as a plasma support gas, but also carries the analyte species from the vaporizer into the plasma, where excitation, ionization and emission take place. The length of torch, diffusion, and gas flow rates all determine the residence time of the analyte atoms and thereby influence the signal intensity. Changes in the gas flow rate also change the mass transfer characteristics of the source, which might, in turn, affect the analytical characteristics of plasma. For a plasma source used in F A P E S , Sturgeon et al. [2] suggest that high helium flow rates serve to maintain the plasma gas in a state of high purity, diluting both the gaseous products of decomposition of the analyte salts and any water or acid vapor desorbed from the heated surface during the vaporization stage, as well as minimizing the ingress of 165 Figure 7.5: The effect of plasma power on signal-to-background ratio (SBR) from (A) Ag and (B) Mg analyte for different electrode lengths. The gas flow rates were 0.2 and 1.0 L min 1 for Ag and Mg, respectively. 166 atmosphere into the torch. We have studied the effect of different plasma gas flow rates on emission intensity of Ag I (328.1 nm), Mg I (285.21 nm), and He I (388.87 nm) at 250 W plasma forward power for different electrode lengths. The results are shown in Fig. 7.6, 7.7, 7.8 and 7.9. All data presented in these plots are the average of four repetitive measurements where the error bars represent the noise in the signal. Figure 7.6 depicts the influence of plasma gas (also the carrier gas) flow rates on the silver analyte for different electrode lengths from a deposition of 2.5 ng of silver standard solution on the ETV platform for a 250 W plasma forward power. The signal intensity (using both peak height, Fig. 7.6A, and peak area, Fig. 7.6B) is normalized with respect to the intensity at a 0.75 L min"1 gas flow rate. The effect of gas flow rate on the magnesium analyte is illustrated in Fig. 7.7 using both peak height (Fig. 7.7A) and peak area (Fig. 7.7B) for a deposition of 5 ng magnesium standard solution with a fixed plasma forward power of 250 W. The magnesium signal is normalized with respect to the intensity at a 1.0 L min 1 gas flow rate. As can be seen from the plots, different analytes behaved differently for shorter (2 and 3 cm long) electrodes, whereas for longer electrodes (4, 5, and 6 cm long) the analytes behaved similarly with the increase of the gas flow rates. As is illustrated in Fig. 7.6 and 7.7, for shorter electrodes the silver analyte signal increased linearly up to 0.2 L m in 1 flow rate leveling off at higher flow rates; however, magnesium emission increased linearly with increasing gas flow rate without leveling off. The silver analyte is vaporized from the platform earlier in time than the magnesium; its reported appearance temperature is 1100 K compared to that of 1500 K for magnesium. It could be hypothesized that the effects of residence time outweigh the 167 Figure 7.6: The plasma gas flow rate and the emission intensity from silver analyte, (A) using peak height and (B) using peak area, for different electrode lengths. The applied plasma forward power is 250 W. 168 Figure 7.7: The plasma gas flow rate and the emission intensity from magnesium analyte, (A) using peak height and (B) using peak area, for different electrode lengths. The applied plasma forward power is 250 W. 169 effects of mass transport efficiency for shorter electrodes. For longer electrodes, the signal from both analytes increased linearly with increasing gas flow rate. This observation suggests a higher and more effective transport efficiency of analyte atoms and/or molecules into the plasma (for a given electrode length) with a higher gas flow rate. Figure 7.8: The plasma gas flow rate and the emission intensity from He I for different electrode lengths, with a plasma forward power of 250 W. 170 The effect of different electrode lengths on the plasma background signal (using He I emission) is shown in Fig. 7.8 for different gas flow rates. The He I emission signal presented in this plot is the normalized one with respect to the maximum signal obtained at 250 w power and 0.1 L min 1 gas flow rate for a 4 cm long electrode. It is evident from the plot, that at a given set of plasma parameters, the background intensity increases linearly up to a 4 cm long electrode before it starts decreasing with increasing electrode length. It can also be seen from Fig. 7.8 that for a given electrode length, He I intensity remained almost unchanged with changing flow rates. This observation leads one to believe that ingress of atmosphere is not a problem in this source within the operating flow range. Unlike He I emission, both silver and magnesium analytes behave differently, which is illustrated in Fig. 7.9A and 7.9B. The signal presented in this plot is the average peak area for four successive measurements. Although different electrode lengths do not have any significant effect on the silver emission signal, the magnesium emission signal responds sharply with electrode length. At any given plasma operating parameter, the magnesium atomic signal increases linearly with increasing electrode length up to 5 cm before displaying a decreasing trend, which is depicted in Fig. 7.9B. It is reasonable to assume that "diffusion" plays an important role in determining the residence time between these two analytes for any given gas flow rate and electrode length. Due to lower molar mass, magnesium has a higher diffusion rate than that of silver. Increasing electrode length provides longer residence time for both analytes, however, for magnesium, the "residence time" effect outweighs the "lower power density" effect up to a 5 cm long electrode (Fig. 171 Figure 7.9: The effect of electrode length on (A) silver and (B) magnesium analyte for a deposition of 2.5 and 5.0 ng of Ag and Mg sample, respectively. The applied plasma power is 250 W with different gas flow rates. 172 7.9B). For electrode lengths exceeding 5 cm, the "plasma power density" outweighs the "residence time" effect, causing a lower emission signal for magnesium. In the case of silver, at higher gas flow rates (e. g. 0.5, 0.75 L min 1) the analyte residence time and power density effects cancel each other out as the electrode length increases. However, at lower flow rates (e.g. 0.1, 0.2 L min 1) the residence time effect is outweighed by the power density effect which is clearly depicted in Fig. 7.9A. Figure 7.1 OA and 7.1 OB are plots of the influence of plasma gas flow rate on signal - to - noise ratio (SNR) using both silver and magnesium as the analyte, where "noise" represents the noise in the signal. For silver, the signal - to - noise ratio (SNR) increased sharply with increasing gas flow rate up to 0.2 L min 1 , beyond this point, the SNR increased, but with a much lower slope, as shown in Fig. 7.1 OA. Unlike silver, the signal - to - noise ratio for magnesium increased linearly with increasing gas flow rate up to 1.0 L min 1 . At a lower gas flow rate there was hardly any correlation between the flow rate and the electrode length, but at a higher flow rate (0.5 and 0.75 L m in 1 for silver and 1.0 L min 1 for Mg) longer electrodes provided a higher signal - to - noise ratio than the shorter ones. This observation further strengthens our earlier suggestion that a better signal - to - noise ratio is due to a more efficient mass transfer from the electrothermal vaporizer (with increasing flow rate) to the plasma discharge, as well as a longer residence time of analyte (because of higher discharge volume). 173 Figure 7.10: The effect of gas flow rate on signal-to-noise ratio (SNR) using (A) silver and (B) magnesium as analyte for different electrode lengths. The applied plasma forward power is 250 W. 174 7.3.3 Analytical Figures of Merit In order to determine the analytical characteristics of different electrode lengths, the analyte emission and the noise in the signal were both analyzed at different plasma conditions, such as applied power and gas flow rate. Detection limit is one of the most important figures of merit that reflects the power of an analytical method and/or instrument. For the measurement of detection limits, either of two approaches may be followed: the SNR approach, which uses the signal-tc-noise ratio, or the S B R - R S D B approach, which uses the signal-to-background ratio (SBR) and the relative standard deviation of the background (RSDB) [3]. We have used the SNR approach, where a detection limit ( c j is experimentally defined as the analytical concentration that yields a net analyte signal ( x j equal to k times the standard deviation (oB) of the background (xB) as in the following equation: C L = - * 2 L (7.1) x A / c 0 where c 0 is the concentration yielding a net signal x A, (xA/c 0) representing the sensitivity of the instrument or the slope of the calibration curve. In eq. (7.1), it is assumed that near the detection limit, the system is limited by the background noise. The numerical value of the factor k used in this equation is important in the context of the statistical interpretation of the detection limit as a practical analytical figure of merit [4], but for the sake of uniformity, the use of k=3 is generally recommended. The precision, expressed in terms of the relative standard deviation (RSD) or the confidence interval of the concentration, is also a useful tool to measure the power 175 of an analytical method. Precision depends primarily on the R S D of the measured net line signal (RSDN) if the statistical error in the calibration is neglected. R S D N , in turn, is dictated by the fluctuations in the measured background and gross line signals. The latter includes the fluctuations in the net emission line signal, hence, the analyte flicker noise. The detection limits and precision (%RSD) were determined using both silver and magnesium analytes for this P P - C C P source. The effects of electrode length on signal-to-noise ratio (SNR), where noise represents the "noise" in the signal and signal-to-background ratio (SBR) are provided in Fig. 7.11 A and 7.11B, respectively. All data was collected for a 250 W plasma, with a gas flow rate of 1.0 L m in 1 for Mg analyte. For the silver analyte, the gas flow rates were 0.2 and 0.75 L min 1 . It is clearly evident from this figure that, although there is no linear relationship between electrode length and signal - to -noise ratio (SNR), the latter improved with increasing electrode length for both analytes. The residence time of all species in the plasma increases with increasing electrode length, which in turn results in a higher background signal. This causes a poor signal- to-background ratio (SBR), for a given plasma parameter, as the electrode length increases. This behavior is clearly illustrated in Fig. 7.11B. The study of rf forward power and plasma (also the carrier) gas flow rate on silver and magnesium analytes gives an insight of optimum power and flow rate to be used in parallel-plate capacitively coupled plasma for a best possible result. With the optimized value of power and flow rate, limits of detection were calculated, according to equation 7.1, for Ag and Mg (using both atomic line and ionic lines). The k value of 3 is used in this equation. Figure 7.12 illustrates the limits of detection 176 1 I i i r 2 3 4 5 6 Electrode Length, cm Figure 7.11: The effect of electrode length on (A) signal-to-noise and (B) signal-to-background ratio for a 250 W plasma and 0.2 and 0.75 L min 1 (for Ag) and 1.0 L min 1 (for Mg) gas flow rates. 177 as a function of electrode length. The results presented in this plot indicate that although the shortest electrode resulted in good detection limits, the 3 cm long electrode caused the situation to worsen; however, the situation improved slightly with a further increase in electrode length. In general, the longer electrodes (i.e. 6 cm) produced the best (lowest) detection limits for the optimized experimental parameter used. 5-, Electrode Length, cm Figure 7.12: The effect of electrode lengths on the analyte (for both silver and magnesium) detection limit for an optimized plasma operating conditions. 178 The lowest detection limits obtained were 0.9 pg for Ag, 0.7 pg for Mg (using atomic line), and 0.3 pg for Mg (using ionic line). The silver detection limit is better than the reported value of 12 pg and 210 pg [5,6] in C M P source and 1.6-10 pg [7-9] in MIP source and compares well with the values (0.4 - 1.2 pg) found in F A P E S source [10-12]. Our Mg detection limit is slightly better than the value ( 1 - 1 5 pg) found in C M P and AAS [5, 13, 14]. The precision (% RSD) was approximately 4 -10% for both elements mentioned here at a concentration of 100 times the limit of detection. 7.3.4 Rotational Temperature and Degree of Ionization The rotational temperature (T r o t), which reflects the gas kinetic temperature of the source, was also studied for different electrode lengths. The effect of rf power and plasma gas flow rate on the N 2 + rotational temperature have been measured for various electrode lengths and are illustrated in Fig. 7.13 and 7.14. Fig. 7.13A demonstrates that the rotational temperature has only a very weak dependence on plasma forward power for a given gas flow rate of 0.2 L min 1 , independent of the electrode length. Increasing plasma forward power from 100 to 250 W increases the measured rotational temperature from 854 to 886 K, from 849 to 875 K, from 798 to 844 K, from 795 to 836 K, and from 770 to 813 K for a 2, 3, 4, 5, and 6 cm long electrode, respectively, which is illustrated in this plot. The uncertainties, determined from five independent temperature measurements, were from ± 0.1% to ± 0.6% for a power range of 100 - 250 W. 179 CD t_ CO I— CD CL E CD h-75 c o "g o + CM z 1000-j 900 800-700-^  600 500-1000n 900-800 700-600 500-(A) -o- _4 o-- • - 2 cm, --©- 3 cm - A - 4cm, •V" 5cm - o - 6 cm T 1 1 1 i 1 1 1 1 1 1 1 1 1 1 1 1 1 r 100 150 200 250 Plasma Forward Power, W (B) • O §T...r...~$r...nr...„jp „ . . . U , . • W • • » • • « ' • • , M • • , , , 2 cm, --©- 3 cm - A - 4cm, - v - 5cm - o - 6 cm 111111111 111 i 111 0.2 0.4 i i i i i i i i i i i i i i i i i i i i | i i i i i i i i i | i i i i i i i i i | 0.6 0.8 1.0 Gas Flow Rate, L min"1 Fig. 7.13: The effect of (A) rf power (at a constant gas flow rate of 0.2 L min1) and (B) gas flow rate (at a constant rf power of 250 W) on measured N2 + rotational temperature for different electrode lengths. 180 Figure 7.13B illustrates that the gas flow rate only weakly influences the rotational temperature. With an increase in the flow rate from 0.1 to 1.0 L min 1 , the rotational temperature increased about 40 K, regardless of the electrode length. The uncertainties, determined from five independent temperature measurements, were from ± 0.1% to ± 0.6% for a gas flow range of 0.1 - 1 . 0 L min 1 . 1000- , 2 3 4 5 6 Electrode Length, cm Figure 7.14: The effect of electrode length on calculated N 2 + rotational temperature at a constant applied rf power of 250 W and two different gas flow rates of 0.2 and 1.0 L min 1 . 181 The effect of electrode length on the measured rotational temperature for a given set of plasma operating conditions is isolated from Fig. 7.13A and 7.13B and then depicted in Fig. 7.14. As can be seen from this plot, for a 250 W plasma and either 0.2 or 1.0 L min 1 gas flow rate, the rotational temperature decreases slightly with increasing electrode length. This, as well as the aforementioned observations in this chapter, suggest that although the available power density decreases with increasing electrode length, at a given set of operating parameters, the analytical characteristics of P P - C C P change not dramatically, but slightly and steadily, as we change the electrode length, and hence the discharge volume. Helium capacitively coupled plasmas could serve as an ion source for elemental mass spectrometry (MS), since this plasma contains sufficiently high energetic species (i.e. metastable He and He ions) to efficiently ionize elements with an ionization potential as high as 10.45 eV (iodine) [15]. In the previous chapters, the degree of ionization in this P P - C C P source has been demonstrated. To improve the sensitivity of this source for MS application, one needs to optimize and / or enhance the degree of ionization; hence this study. We have investigated the effects of electrode length on analyte ionization (using Mg as an example) at various plasma conditions, and the results are illustrated in the following plots. Figure 7.15 represents the influence of electrode length on the degree of ionization (a) at different applied powers and at a constant gas flow rate of 1.0 L min 1 . For shorter electrodes (2 and 3 cm long), an ionic signal was detected with a power as low as 100 W; the degree of ionization increases from 42% to 81% as the power increases from 100 to 250 W. Where longer electrodes are concerned (5 and 6 cm long), the 182 minimum power required to detect the ionic emission signal is 150 W, due to a decrease in the power density. It is evident from this plot that at a relatively high rf power (e. g. 200, 225 or 250 W) the degree of ionization increased linearly with increasing electrode length from 2 to 5 cm and then began decreasing. For a lower rf power (e.g. 150 W) the degree of ionization decreased for a electrode length longer than 4 cm. The degree of ionization reached its maximum value of 87% for the 5 cm long electrode at 250 W plasma power. Figure 7.15: The effect of electrode length on the degree of ionization (using Mg as a typical analyte) for different plasma forward powers and at a fixed gas flow rate of 1.0 L min 1 . 183 Plasma gas flow rate also influences the ionization, a fact which is depicted in Fig. 7.16. At a constant plasma power of 250 W, the degree of ionization increased with increasing gas flow rates. This enhancement in ionization depends on the electrode length. For a 2 cm long electrode, the degree of ionization increased nominally with increasing gas flow rate from 0.1 to 0.5 L min 1 before it started decreasing with higher flow rates. The degree of ionization enhanced significantly, for 4 and 5 cm long electrodes, with increasing flow rate from 0.1 to 0.25 L min 1 and, at a higher flow rate, the degree of ionization increased slowly. With a 6 cm long Figure 7.16: The effect of plasma gas flow rate on magnesium ionization using different electrode lengths for a 250 W plasma forward power. 184 electrode, the ionic signal could not be detected below 0.25 L min' 1 gas flow rate. The degree of ionization increased linearly with increasing flow rate, from 0.25 to 1.0 L min 1 . From these observations, it is reasonable to say that for longer electrodes, where the analyte residence time is high, higher flow rates yield higher degrees of ionization due to better transport efficiency and longer residence time. In the case of shorter electrode length, hence a shorter discharge volume, the transport efficiency and the higher collisions are outweighed by the analyte residence time factor, resulting in a decrease in the degree of ionization with higher gas flow rates. Figure 7.17 illustrates the effect of electrode length on the degree of ionization for different gas flow rates. For a plasma parameter of 250 W power and 0.1 L m in 1 flow rate, the degree of ionization decreased linearly from 74% to 60% by increasing the electrode length from 2 to 5 cm, due to a decrease in available power density in unit volume of discharge. For an electrode longer than 5 cm, the ionic signal was undetectable at 0.1 L min 1 gas flow rate. The degree of ionization increasedfrom 80% to 87% with increasing electrode length from 2 to5 cmand then began decreasing for a 250 W plasma with a gas flow rate of 1.0 L min' 1. Plasma with 0.5 and 0.25 L min 1 gas flow rates also behaved in a similar fashion, but with a lower absolute value for the degree of ionization. From the above mentioned observations (from Fig. 7 . 1 5 - 1 7 ) , it is obvious that for a given plasma condition two opposing factors, namely "residence time" and "power density", become active when the electrode length changes. At the plasma conditions used in this study, the residence time emerged as a dominant factor when the electrode measured up to 5 cm long; with a longer electrode, the power density outweighed the residence time factor. 185 90-, o - • - F = 1.0 Lmin -©- F = 0.5 L min"1 - • - F = 0.25 L min F = 0.1 Lmin"1 Q -1 50 4 2 3 4 5 6 Electrode Length, cm Figure 7.17: The effect of electrode length on the degree of ionization (using Mg as 7.3.5 Conclusions The effect of electrode length, and hence plasma volume, on some analytical characteristics of this parallel-plate capacitively coupled plasma source is studied and the results are presented in this chapter. The results indicate that independent of electrode length, the transport efficiency increases with increasing plasma gas flow rate. With higher flow rates, longer electrodes give lower noise in the signal, due to less turbulence in the flow, and hence result in better signal-to-noise ratios. Although power density decreases with increasing plasma (electrode) length, for a given power and flow rate, longer electrodes yield better signal-to-noise ratios. The detection limit improves with increasing electrode length, reaching as low as 0.96 pg analyte) for a 250 W plasma at different gas flow rates. 186 for silver and as low as 0.34 pg for magnesium for 6 cm long electrode. The rotational temperature (T r o t) has a very weak dependence on plasma forward power for a given gas flow rate. Increasing plasma forward power from 100 to 250 W increases the measured rotational temperature (T ro t) by about 50 K irrespective of the electrode length. The plasma gas flow rate also influences the rotational temperature very weakly. With an increase in flow rate from 0.1 to 1.0 L min 1 , the rotational temperature (T r o t) increases about 40 K, regardless of the electrode length. For a given set of plasma conditions, N 2 + rotational temperature decreases slowly with increasing electrode length, suggesting that although the available power density decreases with increasing electrode length, for a given set of parameters, the plasma characteristics do not change dramatically, but slightly and steadily. The degree of ionization increases with increasing electrode length for a 250 W plasma at a gas flow rate of 1.0 L min' 1 and reaches its maximum value of 87% for a 5 cmlong electrode. In contrast, for a 250 W plasma with a low flow rate, 0.1 L min 1 for example, the degree of ionization decreases linearly from 74% to 60% by increasing the electrode length from 2 to 5 cm probably due to a decrease in available power density per unit volume of discharge. This rather unique analyte ionization characteristic may be attributed to two opposing factors, namely, "residence time" and "power density", which become active as we change the plasma length. At the plasma conditions used in this study, the residence time emerged as a dominant factor when the electrode measured up to 5 cm; with longer electrodes, the power density outweighed the residence time factor, causing a lower degree of ionization. 187 7.5 References 1. C. I. M. Beenakker, Spectrochim. Acta, 32 Part B, (1977), 173-187. 2. R. E. Sturgeon and H. Falk, Spectrochim. Acta, 43 Part B, (1988), 421-38 . 3. P. W. J . M. Boumans, Spectrochim. Acta, 46 Part B, (1991), 917-939. 4. P. W. J . M. Boumans, Spectrochim. Acta, 33 Part B, (1978), 625-634. 5. A. H. Ali and J . D. Winefordner, Analytica Chimica Acta, 264, (1992), 327. 6. A. H. Ali, K. C. Ng & J . D. Winefordner, J. Anal. Atom. Spectrom.,(1991), 211. 7. H. Kawaguchi and B. L Vallee, Anal. Chem., 47, (1975), 1029. 8. H. Kawaguchi, M. Hasegawa, and A. Mizuike, Spectrochim. Acta., 27B, (1972), 205. 9. C. I. M. Beenakker, B. Bosman, and P. W. J . M. Boumans, Spectrochim. Acta., 33B, (1978), 373. 10. R. E. Sturgeon, S. N. Willie, V. T. Luong, and J . G. Dunn, Appll. Spectrosc, 45, (1991), 1413-1418. 11. R. E. Sturgeon, S. N. Willie, V. T. Luong, and S. S. Berman, J. Anal. Atom. Spectrom., 6, (1991), 19. 12. R. E. Sturgeon, S. N. Willie, V. T. Luong, and S. S. Berman, Anal. Chem., 62, (1990), 2370. 13. O. G. Koch, P. D. LaFleur, G. H. Morrison, E. Jackwerth, A. Townshend, and G. Tolg, Pure and Appl. Chem., 54, (1982), 1565 - 1577. 14. A. H. Ali, K. C. Ng, and J . D. Winefordner, Spectrochim. Acta, 46B, (1991), 1207-1214. 15. R. E. Sturgeon and R. Guevremont, J. Anal. Atom. Spectrom., 13(1998), 229. 188 C h a p t e r - 8 C o n c l u s i o n s 8.1 Summary In an effort to understand the basic characteristics of the helium capacitively coupled plasma in parallel plate geometry, a number of temporally and spatially resolved spectroscopic studies have been performed and are presented in this thesis. Investigation of the plasma fundamental properties (such as excitation temperature, rotational temperature, etc.) and analytical characteristics (such as detection limits, matrix interferences, degree of ionization etc.) of PP-CCP source were the major objectives of the studies described in this thesis. In this chapter, a brief summary of these studies, and their results, are given along with some suggestions for further research on the PP-CCP source. It has been demonstrated throughout this thesis that some important fundamental and analytical characteristics of the parallel plate He-CCP are similar to those of a microwave induced plasma, however, it is evident that the CCP offers several operational advantages. First, the plasma is self-igniting and hence does not require a tesla discharge. Second, because a tuned cavity is not required, the dimensions of the CCP can be quite flexible. Also, compared to CMP, the PP-CCP exhibits similar excitation temperatures but at a significantly lower rf power. Finally, it is a compact source and could lead to the development of a portable analytical tool for trace and ultra-trace elemental analysis in the near future. 189 Optical emission from plasma background species (such as He I, N 2 + and OH) in the P P - C C P source was found to have a spatial emission structure, which is due to the transverse power coupling geometry. Both He I and N 2 + emissions were found to be most intense adjacent to the plasma torch wall, however, the OH emission intensities remained fairly constant at the central region (± 1.0 mm from center) of the discharge. It is well known that the most important excitation mechanisms involved in an atmospheric pressure plasma source are electron impact excitation and ionization, Penning excitation and ionization, and charge transfer ionization (and excitation) involving He + or He 2 + , which along with neutral ground and excited states of He atom are major plasma species. The spatial distribution of these plasma background species is the predominant factor that determines the analyte emission zone in the source. The "analyte emission" or "emission sensitive" zone in a P P -C C P source is thus situated adjacent to the discharge wall where the He atomic signal is dominating. The spatial distribution of plasma species remained unchanged, however the signal intensity from all three species increased with increasingly applied rf power. As with rf power, the spatial distribution of different plasma species also remained totally unchanged at different plasma gas flow rates. Atomic emission intensities from He, Pb, and Fe, and molecular emission intensities from O H and N 2 + were measured in order to calculate the excitation and rotational temperatures of P P - C C P source. Pb excitation temperatures were calculated using a two-line method by measuring the Pb 280.19 and 283.31 nm lines at different rf powers. He and Fe excitation temperatures were determined from a Boltzmann plot using four (for He) and seven (for Fe) neutral atom lines. All three 190 excitation temperatures were found to increase monotonically with rf power over a range of 100 - 250 W. The measured rotational temperatures from OH and N 2 + were also found to increase monotonically with rf power over a range of 75 - 250 W. The difference in measured excitation temperature suggests the absence of local thermodynamic equilibrium (LTE) conditions in P P - C C P . The matrix effects of NaCI and N a N 0 3 on the silver emission were investigated to understand the interference mechanism and their effects on the plasma. Both enhancement and suppression of analyte line emission were observed depending upon the amount of matrix species present in the system. Compared to the emission signal obtained when there was no interferent, the presence of either NaCI or N a N 0 3 (up to 40 and 0.05 /vg respectively) as a concomitant in the sample caused an interference effect by enhancing the emission intensity. With increasing amounts of matrix species the signal enhancement decreased, giving a 10% signal depression in the presence of 250 jjg of Na as NaCI and 200 /vg of Na as N a N 0 3 . The interference effect also depends on the plasma input power. For a 125 W plasma forward power the signal enhancement was found to be at it's maximum, above which power, the signal enhancement decreased. The temporal response of the analyte showed that the presence of a small amount (e. g. 0.125 ug) of NaCI caused the analyte signal to appear earlier than the signal without any NaCI. This supports the idea of early vaporization of a volatile chloride compound. The earlier peak disappeared with increasing amounts (e.g. 100 fjg of Na) of matrix species, probably due to a decrease in plasma power via radiative loss and shifts of the E E D F to a lower energy. 191 Temporally resolved emission measurements from ionic and atomic lines showed that the applied rf power in the plasma has a significant effect on the degree of analyte ionization; cadmium and magnesium were used in this study. The degree of ionization from the Cd analyte increased from 20% to 45% by increasing the rf power from 100 to 200 W. Ionic signal form Mg analyte could not be detected below 125 W rf power, and its degree of ionization increased from 36% to 80% with increasing power from 125 W to 200 W. At a power higher than 200 W the degree of ionization leveled off at 48% and 83% for Cd and Mg, respectively. The plasma gas (also used as the carrier) flow rate also effected the degree of ionization. For Mg it increased slowly from 60% to 78% with an increase of gas flow rate from 0.1 to 0.5 L m in 1 and reached a plateau of ~ 82% for higher flow rates. The C d ionic signal could not be detected at less than 0.4 L min 1 gas flow rate and the degree of ionization increased from 27% to 48% for a flow rate of 0.4 - 1.0 L min*1. A small amount of EIE increased magnesium ionization slightly, but higher amounts decreased the degree of ionization due to an ionization equilibrium shift. Due to higher ionization potential, the cadmium analyte did not show any ionization improvement in the presence of EIE, rather it decreased the degree of ionization. Independent of electrode length (hence the discharge volume), the transport efficiency increases with increasing plasma gas flow rate. With higher flow rates, longer electrodes yielded lower noise in the signal, due to less turbulence in the flow, and resulted in a better signal - to - noise ratio. Although power density per unit volume of discharge decreases with increasing electrode length, for a given power and flow rate, longer electrodes resulted in a better signal - to - noise ratio. The 192 detection limits improved with increasing electrode length, reaching as low as 0.96 pg for silver and 0.34 pg for magnesium, for a 6 cm long electrodes. The rotational temperature showed a very weak dependence on applied rf power for a given gas flow rate. Increasing rf power from 100 to 250 W increased the measured T r o t by only about 50 K irrespective of the electrode length. The plasma gas flow rate also influenced the rotational temperature very weakly. With an increase in flow rate from 0.1 to 1.0 L min1, the T r o t increased about 40 K, regardless of the electrode length. For a given set of plasma conditions, N 2 + rotational temperature decreased slowly with increasing electrode length. The degree of ionization (a) increased with increasing electrode length for a 250 W plasma with a gas flow rate of 1.0 L min1 and reached its maximum value of 87% for a 5 cmlong electrode due to higher residence time and mass transport efficiency. In contrast, for a 250 W plasma with a low gas flow rate of 0.1 L min1, the degree of ionization decreased linearly from 74% to 60% with increasing the electrode length from 2 to 5 cm. This is evidence that the decrease of available power density per unit volume of discharge overpowered the combined effect of residence time and mass transport efficiency. All studies relating to electrode length suggest that although the available power density decreases with increasing electrode length, under a given set of plasma conditions, the plasma characteristics do not change dramatically. 8.2 Suggestions for Future Research Most of the experiments described in this thesis were limited to rf powers of 100 W or more, although the plasma could be ignited and sustained at a power as 193 low as 10 W. It is likely that a portion of the applied rf power was dissipated through the system. A modified impedance matching network with a shortest possible distance between impedance coil and the parallel plate electrode could enhance the power delivery to the discharge. Also, the electrical characteristics of this source are not well studied yet. A comprehensive study of the current-voltage characteristics and the electron energy distribution function of the plasma source would definitely help to better understand the characteristics of this plasma. In this thesis it has been shown that in comparison to MIP, M C P and F A P E S , this P P - C C P source exhibits similar excitation temperatures, detection limits, and degrees of analyte ionization. Perhaps it is time to look into the viability of this instrument for the analysis of real samples (as opposed to analytical standards). In this research direction, complete analytical methods should be devised for the simultaneous multielemental analysis. The use of higher atomization temperatures (as opposed to 2000 - 2100 K used in the present studies) and faster heating rates of the ETV platform would be beneficial to the improvement of the analytical characteristics and to reduce the matrix effects in analytical determination. Using a cross gas flow in front of the plasma torch should be looked into to improve the linear dynamic range. Last but not least, the adaptation of the automatic sampling device could enhance the precision and throughput of the analysis. This P P - C C P source was operated with a torch as small as 1 mm width. It is time to investigate the possibility of further miniaturization of this source, perhaps in a capillary column as small as 100 nm which would enable this source to be better coupled with a gas chromatography (GC) column. 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0059599/manifest

Comment

Related Items