Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Studies of analyte ionization in a furnace atomization plasma excitation spectrometry source Lu, Shengyong 2002

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

Item Metadata

Download

Media
831-ubc_2002-750442.pdf [ 7.29MB ]
Metadata
JSON: 831-1.0061282.json
JSON-LD: 831-1.0061282-ld.json
RDF/XML (Pretty): 831-1.0061282-rdf.xml
RDF/JSON: 831-1.0061282-rdf.json
Turtle: 831-1.0061282-turtle.txt
N-Triples: 831-1.0061282-rdf-ntriples.txt
Original Record: 831-1.0061282-source.json
Full Text
831-1.0061282-fulltext.txt
Citation
831-1.0061282.ris

Full Text

STUDIES OF ANALYTE IONIZATION IN A FURNACE ATOMIZATION PLASMA EXCITATION SPECTROMETRY SOURCE by Shengyong Lu B.Sc, Tsinghua University, China, 1991 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE STUDIES Department of Chemistry We accept this thesis as confirming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA June 2002 © Shengyong Lu, 2002 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 of ^<£/^ S7>y The University of British Columbia Vancouver, Canada Date DE-6 (2/88) 11 ABSTRACT Furnace Atomization Plasma Excitation Spectrometry (FAPES) is a relatively new atomic emission spectrochemical method which employs a conventional graphite furnace for analyte atomization and an atmospheric-pressure helium plasma sustained inside the furnace for analyte excitation. The generation of the plasma is achieved by applying radio frequency (rf) power to an electrode located inside, and coaxial with, the graphite furnace. The primary objective of this thesis was to characterize the fundamental properties of the plasma, study analyte excitation and ionization processes, and seek ways to improve analyte ionization efficiency. Background emission characteristics have been observed in a new FAPES source, and temporally resolved emission profiles of background species (He, N 2 + , OH and CO+) have been measured. The ionization mechanisms of major background species are also discussed. Plasma temperatures have been measured in order to characterize the helium plasma. During an atomization cycle, rotational temperatures for N 2 + and OH at 40 W have been found to be 1300 K and 1400 K, respectively, and the excitation temperature for He at 40 W is about 3600 K. Plasma temperatures can be substantially affected by plasma operating conditions. Thus, the effects of conditions such as rf power, gas flow rate, atomization temperature and the dimensions of the center electrode on plasma temperatures and analyte ionization were studied. The temporal atomic and ionic emission behaviors of Cr, Mg, Cd, Fe and Zn have been measured, and analyte atomization mechanisms have been proposed based on the measurements. The effects of operating conditions on analyte ionization have been studied, and an appropriate atomization temperature was found for optimum analyte ionization. An Il l optimum gas flow rate can also maximize analyte ionization. Compared with a "continuous-flow" mode, a "stop-flow" mode can improve analytical sensitivity. Increasing rf power was found to be the best way to achieve a high degree of ionization. A variety of center electrodes with different physical dimensions were used to modify the FAPES source in an effort to improve the ionization capability. However, the larger electrode size acted to reduce the voltage drop across the plasma sheath compromising the analytical performance. The effects of varying the counter ion (MgC^ , MgO, MgN03 and M g S 0 4 ) on Mg atomization, excitation, and ionization were studied, and it was found that these compounds exhibit different atomization mechanisms. It was also observed that Mg ionization in the FAPES source could be improved by the addition of a minute amount of Pd modifier. Figures of merit for magnesium demonstrate improved analytical performance for this new FAPES source. An ionization temperature of about 7000 K at 80 W rf power level was measured. The calculation of electron number densities in the FAPES source shows that in addition to helium ionization, secondary electron emission and thermal or thermionic electron emission from the graphite walls and center electrode inside the source also contribute to the total electron number density. iv TABLE OF CONTENTS Abstract ii Table of Contents iv List of Tables x List of Figures xi List of Symbols and Abbreviations xv List of Publications and Presentations Arising From This Study xviii Acknowledgements xx Chapter 1 Introduction 1 1.1 Evolution of Furnace Atomization Plasma Excitation Spectrometry (FAPES) ....2 1.1.1 Plasma 2 1.1.2 Glow Discharge 3 1.1.2.1 Gas Discharge 3 1.1.2.2 Glow Discharge 5 1.1.3 Atmospheric Pressure, Radio-frequency, Capacitively Coupled Plasma ; 6 1.1.4 FAPES: Recent Developments 10 1.1.5 FAPES: A Potential Ion Source 14 1.2 Thesis Objective 15 1.3 Analyte Atomization 16 1.3.1 Electrothermal Atomizer 16 V 1.3.2 Atomization Efficiency 17 1.4 The rf Discharge in the FAPES Source 19 1.5 Excitation and Ionization 21 1.5.1 Excitation and Ionization Mechanisms 21 1.5.2 Plasma Temperature Measurements 26 1.5.3 Ionization Efficiency 28 1.6 Thesis Overview 29 1.7 References 30 Chapter 2 Experimental System 36 2.1 Instrumentation 36 2.1.1 The FAPES source 38 2.1.2 Light Collection System 41 2.1.3 Optical Detection System 41 2.1.4 Data Acquisition System 42 2.2 Operating Conditions Control 43 2.3 Analyte Atomization 43 2.4 References 45 Chapter 3 Measurement of Temporal Emission and Plasma Temperatures 46 3.1 Introduction 46 vi 3.2 Experimental 49 3.2.1 instrumental Set-up 49 3.2.2 Procedure 51 3.3 Results and Discussion 51 3.3.1 Background Emission Spectrum 51 3.3.2 Temporal Emission Measurements 53 3.3.3 Plasma Temperature Measurements 57 3.3.3.1 Time-resolved Temperature Measurements of Background Species 57 3.3.3.2 Excitation Temperature of Pb 58 3.3.4 Effect of Operating Conditions on Plasma Temperatures 60 3.3.4.1 rf power 61 3.3.4.2 Gas Flow Rate 61 3.3.4.3 Atomization Temperature 63 3.3.4.4 Center Electrode Size 64 3.4 Summary 67 3.5 References 68 er 4 Temporal and Operating Characteristics of Analyte Emission 69 4.1 Introduction 69 4.2 Experimental 69 4.3 Results and Discussion 73 4.3.1 Temporal Emission Spectra and Profiles 73 4.3.2 Atomization Temperature Effects 81 4.3.2.1 Atomic and Ionic Emission Intensities 81 4.3.2.2 Temporal Ionic Emission Profile 84 4.3.3 rf Power Effects 86 4.3.3.1 Temporal Ionic Emission Profile 86 4.3.3.2 Ionic Emission Intensities 86 4.3.4 Gas Flow Rate Effects 88 4.3.4.1 Atomic and Ionic Emission Intensities 88 4.3.4.2 Comparison of Stop-flow Mode and Continuous-flow Mode 90 4.4 Summary 93 4.5 References 95 er5 Effects of Operating Conditions on Analyte Ionization 96 5.1 Introduction 96 5.2 Experimental 98 5.3 Results and Discussion 98 5.3.1 Effect of rf Power 98 5.3.2 Effect of Atomization Temperature 102 5.3.3 Effect of Gas Flow Rate 106 5.3.4 Effect of Center Electrode Size 108 vni 5.3.5 Effect of Center Electrode Size with Constant Electrode Volume 112 5.3.6 Effect of Center Electrode Length 112 5.4 Summary 114 5.5 References 116 Chapter 6 Studies of Analyte Matrices and Palladium Modifier Effects 117 6.1 Introduction 117 6.2 Experimental 119 6.3 Results and Discussion 119 6.3.1 Effect of the Analyte Matrix 119 6.3.2 Effect of Pd Modifier 123 6.4 Summary 126 6.5 References 126 Chapter 7 Analytical Performance of The FAPES Source 128 7.1 Introduction 128 7.2 Experimental 130 7.3 Results and Discussion 130 7.3.1 Analytical Figures of Merit of Mg 130 7.3.2 Relationship between Degree of Ionization and Ionization Potential ..131 ix 7.3.3 Calculation of Ionization Temperature and Electron Number Density 132 7.4 Summary 137 7.5 References 138 er 8 Conclusions 139 8.1 Summary 139 8.2 Suggestions for Future Research 144 8.3 References 146 X LIST OF TABLES Table Page Table 1.1 A comparison of different plasma sources 3 Table 1.2 Excitation and ionization mechanisms in the FAPES Source 24 Table 3.1 Spectroscopic data used to calculate the rotational temperature of N 2 + 50 Table 3.2 Spectroscopic data used to calculate the rotational temperature of OH 50 Table 3.3 Spectroscopic data used to calculate the excitation temperature of Pb(I) 50 Table 4.1 Data for atomic and ionic lines of selected elements 71 Table 4.2 Experimental parameters for atomic and ionic emission measurements 73 Table 5.1 Experimental conditions for each operating condition experiment 99 Table 7.1 Figures of merit for Mg at 80 W 131 Table 7.2 Partition functions of several elements at 6976 K 134 Table 7.3 The calculated electron number densities at 80 W 134 Table 7.4 Calculated electron number density corresponding to the assumed LTE temperature 135 xi LIST OF FIGURES Figures Page Fig. 1.1 Schematic representation of a dc electrical discharge at reduced pressure 5 Fig. 1.2 Schematic illustration of various atmospheric pressure rf CCP sources 8 Fig. 1.3 Schematic representation of an end-view of the FAPES plasma source 20 Fig. 1.4 Schematic representation of helium atomic energy levels 25 Fig. 2.1 Schematic illustration of the experimental system 37 Fig. 2.2 Schematic diagram of the FAPES source 39 Fig. 2.3 Schematic representation of the rf connector interface 40 Fig. 3.1 Schematic illustration of rovibronic transitions of N 2 + ( B 2 Z + U - X 2 E + g ) 48 Fig. 3.2 Background emission spectrum of the FAPES plasma (40 W, 30 °C, 50 ml/min) .52 Fig. 3.3 Temporal emission spectrum of He and N 2 + (80 W, 2000 °C, 50 ml/min) 54 Fig. 3.4 Temporal emission spectrum of OH (2000 °C, 80 W, 50 ml/min) 56 Fig. 3.5 Temporal emission spectrum of CO + (2000 °C, 80 W, 50 ml/min 57 Fig. 3.6 Temporally resolved temperatures of background species 58 Fig. 3.7 Temporally resolved atomic emission and excitation temperature profiles of Pb (50 ml/min, 80 W, 2600 °C) 59 Fig. 3.8 rf power effect on the excitation temperature of Pb (50 ml/min, 2600 °C) 60 Fig. 3.9 rf power effect on plasma temperatures of He, OH and N 2 + 61 Fig. 3.10 Gas flow effect on rotational temperatures of OH and N 2 + 62 Fig. 3.11 Gas flow effect on the excitation temperature of He 62 Xll Fig. 3.12 Atomization temperature effect on rotational temperatures of OH and N2 63 Fig. 3.13 Atomization temperature effect on the excitation temperature of He 64 Fig. 3.14 Center electrode size effect on rotational temperatures of OH and N2+ (80 W, 2600 °C, 50 ml/min) 66 Fig. 3.15 Center electrode size effect on helium emission intensity and excitation temperature (80 W, 2600 °C, 50 ml/min) 66 Fig. 4.1 Temperature dependence of degree of ionization for Mg, Cd, Fe, Zn, and Cr at an actual line emission intensity ratio (peak emissions at 80 W) 72 Fig. 4.2 Temporal emission spectra of Mg at 80 W 74 Fig. 4.3 Temporal emission spectra of Cd at 80 W 75 Fig. 4.4 Temporal profiles of atomic emission, ionic emission and degree of ionization for Mg at 80 W 76 Fig. 4.5 Temporal profiles of atomic emission for Mg at different temperatures (80 W).. .76 Fig. 4.6 Temporal profiles of atomic emission, ionic emission and degree of ionization for Cd at 80 W 79 Fig. 4.7 Temporal profiles of atomic emission, ionic emission and degree of ionization for Feat 80 W 80 Fig. 4.8 Temporal profiles of atomic emission, ionic emission and degree of ionization forZnat80W 80 Fig. 4.9 Temporal profiles of atomic emission, ionic emission and degree of ionization forCrat80W 81 Fig. 4.10 Effect of atomization temperature on Cd atomic and ionic emission intensities .. .83 Fig. 4.11 Effect of atomization temperature on Cr atomic and ionic emission intensities .. ..84 X1U Fig. 4.12 Mg ionic temporal emission profiles at different atomization temperatures 85 Fig. 4.13 Mg ionic temporal emission profiles at different rf powers (Best fit curves) 87 Fig. 4.14 Effect of rf power on ionic emission intensities of Fe and Zn 88 Fig. 4.15 Effect of gas flow rate on Fe atomic and ionic emission intensities 89 Fig. 4.16 Effect of gas flow rate on Cr atomic and ionic emission intensities 90 Fig. 4.17 Comparison of stop-flow mode and continuous-flow mode on Cr emission intensities 91 Fig. 4.18 Comparison of stop-flow mode and continuous-flow mode on Cd emission intensities 91 Fig. 4.19 The enhancement of emission intensities under stop-flow mode for Cr, Fe and Mg 93 Fig. 5.1 Temporal profiles of degree of ionization at different rf power levels 100 Fig. 5.2 rf power effect on the intensity ratio of ionic and atomic emission as well as the degree of ionization for Fe 101 Fig. 5.3 rf power effect on the intensity ratio of ionic and atomic emission as well as the degree of ionization for Zn 102 Fig. 5.4 Atomization temperature effect on the intensity ratio of ionic and atomic emission as well as the degree of ionization for Mg 103 Fig. 5.5 Atomization temperature effect on the intensity ratio of ionic and atomic emission as well as the degree of ionization for Cd 104 Fig. 5.6 Atomization temperature effect on the degree of ionization of Fe, Cr, and Zn 105 Fig. 5.7 Gas flow effect on the degree of ionization of Fe, Cr, and Mg under continuous-flow mode 107 xiv Fig. 5.8 Gas flow effect on the degree of ionization of Fe, Cr, and Mg under stop-flow mode 108 Fig. 5.9 The effect of electrode diameter on atomic emission and ionic emission 110 Fig. 5.10 The effect of electrode diameter on the degrees of ionization of analyte 110 Fig. 5.11 The effect of center electrode size with constant electrode volume on the degrees of ionization of analyte 113 Fig. 5.12 The effect of center electrode length on the degrees of ionization of analyte 113 Fig. 6.1 The effect of different matrices on atomic emission, ionic emission and degree of ionization of Mg 120 Fig. 6.2 Pd modifier effect on the ratios of emission intensities and degree of ionization of Mg (80 W, 2600 °C, 50 ml/min) 124 Fig. 7.1 Mg calibration curve 131 Fig. 7.2 Degrees of ionization as a function of ionization potentials for elements - Cr (6.76 eV), Mg(7.64 eV), Fe(7.87 eV), Cd(8.99 eV) and Zn(9.39 eV) 132 Fig. 7.3 A plot for the calculation of ionization temperature 133 LIST OF SYMBOLS AND ABBREVIATIONS A Einstein transition probability or surface area AAS Atomic Absorption Spectrometry ac alternating current arb arbitrary arm atmosphere °C degree Celsius CCP Capacitively Coupled Plasma CMP Capacitively Coupled Microwave Plasma d the separation dc direct current DCP Direct Current Plasma D.L. Detection Limit E excitation energy Ei the first ionization potential of element EIE Easily Ionized Element ETV Electrothermal Vaporizer eV electron volt FAPES Furnace Atomization Plasma Excitation Spectrometry g gram(s) or statistical weight GD Glow Discharge h the Planck's constant Hz Hertz I the total intensity of a spectral line IP Ionization Potential ICP Inductively Coupled Plasma I.D. Inner Diameter IPDA Intensified Photodiode Array xvi j the rotational quantum number K Kelvin or electronic quantum nun kB the Boltzmann constant LTE Local Thermal Equilibrium M(I) atom M(II) ion min minute MIP Microwave Induced Plasma ml milliliter N a the number of free gaseous atoms n e the number density of electrons Ni the number of ions ng nanogram nm nanometer OES Optical Emission Spectrometry P pressure Pg picogram ppb parts per billion ppm parts per million rf radio frequency s second S/B Signal to Background Ratio S/N Signal to Noise ratio T Temperature t time M-g micro gram micro liter V vibrational quantum number V voltage drop V B the breakdown voltage W Watt xvii X molecular species Z partition function a degree of ionization pa the fraction atomized p v the fraction volatilized ea atomization efficiency r\ vapor transport efficiency X wavelength XV111 LIST OF PUBLICATIONS AND PRESENTATIONS ARISNG FROM THIS STUDY 1. S. Lu, and M. W. Blades, Operating Characteristics of Analyte Ionization in a Furnace Atomization Plasma Excitation Spectrometry Source, in preparation. 2. S. Lu, and M. W. Blades, Studies of Plasma Characteristics in a Furnace Atomization Plasma Excitation Spectrometry Source, in preparation. 3. S. Lu, C. W. LeBlanc and M. W. Blades, Analyte Ionization in the Furnace Atomization Plasma Excitation Spectrometry Source - Spatial and Temporal Observations, J. Anal. At. Spectrom. 16(3): p. 256-262 (2001). 4. S. Lu, and M. W. Blades, Analytical Characteristics of a Furnace Atomization Plasma Ionization Source, Winter Conference on Plasma Spectrochemistry (Fort Lauderdale, FL, 2000). 5. M. W. Blades, S. Lu and M. Rahman, Progress in the Development of Capacitively Coupled Plasmas for Spectrochemical Analysis, Winter Conference on Plasma Spectrochemistry (Fort Lauderdale, FL, 2000). 6. M. W. Blades, A. Bass, M. Rahman, and S. Lu, Miniature Capacitively Coupled Plasmas, The 27th Annual Conference of the Federation of Analytical Chemistry and Spectroscopy Society (Nashville, Tennessee, 2000). 7. S. Lu, and M. W. Blades, Excitation and Ionization Characteristics of Background Species in A Furnace Atomization Plasma Excitation Spectrometry Source, The 26th Annual Conference of the Federation of Analytical Chemistry and Spectroscopy Society (Vancouver, Canada, 1999). xix M. W. Blades, M. M. Rahman, and S. Lu, Temporal Emission Characteristics of Capacitively Coupled Plasmas Using A Gated-IPDA, The 24th Annual Conference of the Federation of Analytical Chemistry and Spectroscopy Society (Texas, 1998). XX ACKNOWLEDGEMENTS I would like to express my sincere appreciation to my research supervisor, Dr. M. W. Blades, for his invaluable guidance and advice as well as his assistance throughout the course of this project. Acknowledgement also goes to my group members for their valuable suggestions. I thank Dr. M. Gerry and Dr. A. J. Merer for their valuable discussions. I would like to extend my thanks to Mr. D. Lovrity, Mr. C. Neale, Mr. K. Love and Mr. O. Greiner in the Mechanical Services Shop of the Chemistry Department for constructing the FAPES source. Thanks also go to Mr. M. Carlisle of the Electronics Shop for his technical assistance. Acknowledgement also goes to Dr. Don Douglas for his suggestions. Acknowledgement is also made to the Natural Sciences and Engineering Research Council of Canada (NSERC), the National Research Council of Canada (NRC) and MDS-SCIEX for funding and financial support. Finally, I would like to thank my wife Amy and my daughter Eileen for their encouragement and patience during the completion of this work. Special thanks are given to my mother and my mother-in-law for their loving support. 1 Chapter 1 INTRODUCTION Analytical chemistry is a branch of chemistry that deals with the separation, identification, and determination of constituents in a sample. Among approaches to quantitative and qualitative measurement of a particular sample, spectrochemical methods are widely used to determine chemical species by measuring the interaction of chemical species with light. Plasma atomic emission spectrometry is one spectrochemical method that is capable of providing high atomization efficiencies, high degrees of excitation, and low detection limits for many elements, and it has become a very powerful and widely used elemental analysis technique. Furnace Atomization Plasma Excitation Spectrometry (FAPES) is a relatively new plasma atomic emission spectrometry method. Since FAPES was first described [1,2], it has been under development by several research groups as an alternative elemental analysis technique. The FAPES source consists of a conventional atomic absorption graphite furnace combined with a coaxial electrode located in the interior of the cylindrical atomizer volume. The graphite furnace acts as a desolvator and atomizer for solid or liquid samples, and samples are dried, ashed, and atomized in the graphite cuvette, which is heated in the same manner as conventional graphite furnaces used for atomic absorption spectrometry. A radio frequency (rf) potential is applied to the coaxial electrode generating an atmospheric-pressure helium plasma within the volume of the furnace. This plasma excites the analyte species, which then emit their characteristic radiation thus providing the analytical signal in this atomic optical emission source. 2 1.1 Evolution of Furnace Atomization Plasma Excitation Spectrometry (FAPES) 1.1.1 Plasma A plasma, also referred to as an electrical discharge, is usually sustained by either a direct current (dc), alternating current (ac), or radio-frequency (rf) electric field. A plasma is a partially ionized gas with an equal number of positive and negative charges. The major species present in the plasma are neutral atoms, ions, and unbound or free electrons. In order to sustain a steady-state plasma, a source is required to provide energy for the plasma gas. The degree of ionization, the fraction of the original neutral species which has become ionized, is an important plasma parameter. Plasmas with a degree of ionization much less than unity are referred to as weakly ionized plasmas. The behavior of these types of weakly ionized plasmas is mainly dominated by the presence of a relatively large population of neutral species. Driven by the needs of environmental protection, product quality control, and industrial process monitoring, weakly ionized plasmas have been widely used as spectroscopic sources over the past decades. Plasma sources commercially available for spectrochemical analysis mainly include dc (arc) and ac (spark) plasmas, direct current plasmas (DCP), microwave-induced plasmas (MTP), glow discharges (GD), inductively coupled plasmas (ICP), capacitive microwave plasmas (CMP), and capacitively coupled plasmas (CCP). Among these the DCP, ICP, MTP, GD, and CCP have enjoyed widespread interests for a variety of analytical applications. An overview of the characteristics of these plasma sources is provided in Table 1.1. These plasma sources, except for the GD, have the common fact that their temperatures are high enough to provide complete vaporization and dissociation of analytes to maximize excited state production. It is beyond the scope of this thesis to discuss the details of each plasma source, however, a brief introduction to gas discharges and glow discharges may be helpful to understanding the processes which occur during the formation ofa FAPES plasma. Table 1.1 A comparison of different plasma sources Sources DCP-plasma jet ICP MTP GD-planar CCP-FAPES Electric Field Type dc rf rf dc rf Operational Power (W) 300 - 1000 1000-1500 20-150 20-300 20-100 Support Gas Ar Ar Ar/He Ar He Operating Pressure (torr) 760 760 760 1-3 760 Electron Number density (cm"3) 1015 - 1016 5xl014-5xl015 1014-1015 10"-1013 1013-1014 Excitation Temperature (K) 5000-7000 5000-8000 3000-4000 5000 3000-5000 Sample Introduction Liquid Liquid Liquid Solid Liquid/Solid Problems EIE effects and interference Spectroscopic and non-spectroscopic interferences Mass loading effects Matrix effect caused by differential sputtering rate EEE effects 1.1.2 Glow Discharge 1.1.2.1 Gas Discharge A gas discharge refers to a phenomenon of an electrical current flowing through a gaseous medium. When two electrodes are immersed in a gaseous medium and an electric potential is applied, the gas medium will break down electrically at some potential. This critical electrical potential is called the breakdown potential of the gas. After the initial breakdown, positively charged ions and electrons are formed, and they are then accelerated toward the cathode or the anode, respectively. During the migration processes, these species may collide with background neutral molecules or atoms creating more charged species. As a result, the gaseous medium may experience a transition from a poor conductor to a good conductor. If sufficient energy is continuously added to the gaseous medium, the discharge can be self-sustained. The breakdown potential of a gas is dependent on the properties of the electrodes, the support gas, and the applied electric field. For a dc discharge, the breakdown voltage is described as [3] V B = A(Pd)/(C+ln(Pd)) (1.1) where V B is the breakdown voltage, A, C are constants which depend on the gas, P is the pressure, and d is the separation between the two electrodes. From equation (1.1), we may know that for large Pd, the breakdown voltage is linearly proportional to the product of the separation between electrodes and the pressure. This relationship suggests that a gas discharge can be induced at a lower breakdown voltage by reducing the gas pressure and/or decreasing the spacing between the electrodes. After breakdown, the electric field provides energy for electrons to ionize the support gas through collisions. The increasing number of charged carriers drops the resistance between the electrodes and thereafter causes a decrease in the discharge voltage. 5 An rf discharge may improve the rate of ionization. In an rf electric field, prior to collisions with a wall, electrons undergo oscillatory motion causing electrons to collide with neutrals and ionize these species, leading to an enhanced rate of ionization. 1.1.2.2 Glow Discharge The spatial structure of the CCP discharge can be best understood by beginning with an introduction to the glow discharge. The glow discharge is one of the most common types of discharges, and it is created by inserting two electrodes in a cell filled with a gas at low pressure. Because of its low operating power and collision-rich negative glow region, the glow discharge has been widely used for a variety of applications. A schematic diagram of a "normal" glow discharge plasma is shown in Fig. 1.1. Aston Dark Space -Cathode Dark Space I— Faraday Dark Space Cathode Glow Positive Column Negative Glow + Anode Glow Anode Dark Space Fig. 1.1 Schematic representation of a dc electric discharge at reduced pressure. Reprinted from Chapman [4]. Starting from the cathode to the anode, eight distinct regions are distinguishable. In the Aston dark space, positively charged gas ions collide with the cathode surface and cause electron emission, leading to a net negative space charge. Some of the electrons emitted from the cathode are accelerated to leave the cathode area, while others in the vicinity of the cathode undergo inelastic collisions with gaseous species. The cathode dark space is known as such because of its noticeably low luminosity, and the majority of the potential difference between the electrodes is dropped across this narrow region. The negative glow is a region where electrons excite and ionize gas atoms most efficiently, and it is characterized by its luminosity. Adjacent to the negative glow is the Faraday dark space. In this region, electrons have lost almost all of their kinetic energy through collisions and are not sufficiently energetic to cause excitation or ionization. A plasma can be seen in the positive column since the net space charge is equal to zero in this region. However, it is difficult for electrons and ions to reach the region from the cathode, and thus the positive column is rarely used for the study of analyte atomic emission and ionization. In the vicinity of the anode, the electrons are accelerated, and thus the phenomena of anode dark space and anode glow resemble those of the Aston dark space and cathode glow. The actual number and size of the regions depends on the glow discharge configuration. In a typical glow discharge employed in analytical chemistry, the cathode dark space, the negative glow, the Faraday dark space, and the positive column can be observed. The spacing of two electrodes and the support gas pressure are two major factors that may affect the sizes of the regions. 1.1.3 Atmospheric Pressure. Radio-frequency. Capacitively Coupled Plasma Studies of analytical CCPs began in the mid 1950s [5, 6]. The first CCP device consisted of a hollow cylinder and a coaxial electrode, and the discharge was capacitively induced by a rectified rf source at 43 MHz. In the following decades, several types of CCPs 7 were investigated for spectrochemical analysis [7-20]. However, these CCPs were basically operated at low pressures, and thus it was inconvenient to operate and maintain these plasma sources and change samples. In 1967, Egorova [21] designed an atmospheric pressure, annular rf-CCP based on an existing annular electrode discharge [22]. This device consisted of a water-cooled cylindrical quartz tube around which two separated annular electrodes were placed. The plasma was operated at atmospheric pressure in argon, and it was used for determination of a group of 24 elements with detection limits of 0.05-30 ppm. The atmospheric pressure, radio-frequency parallel plate capacitively coupled plasma (PP-CCP) was described by Liang and Blades in 1988 [1] (Fig. 1.2a). The device consists of two parts, the capacitively coupled plasma (CCP) discharge tube and the tantalum strip electrothermal vaporization sample introduction system. The discharge is formed in a long quartz tube and runs at low support gas flow rates, and as a result a relatively long residence time of analyte atoms can be expected. An effective energy transfer from the power supply to the plasma can be achieved by capacitive coupling. Therefore, the plasma can be generated at atmospheric pressure in a flexible geometry. The PP-CCP source can be used as an atomizer for atomic absorption spectrometry, a plasma emission source for atomic emission spectrometry, and a detector for gas chromatography [23, 24]. In order to increase the transport efficiency, an atmospheric pressure capacitively coupled plasma, later known as the FAPES plasma source, was also developed by Liang and Blades [25] (Fig. 1.2b). Functionally, this FAPES source consisted of an electrothermal atomizer and an rf discharge. The plasma was formed between the graphite tube and a central electrode by rf capacitive coupling at atmospheric pressure. This device combined the 8 Piesme Semple introduction tnlet p«Uiriiiic.Samplc •;. Introducltoti Pon •. (end vie v.') Ten.lo.lum strip Copper.rod supports Plesfne support gcs inlet : Stainless steel electrodes' (to RF power supply) • Qiitrtz • l V s W6i,er cutlet . ToDC-ppwer.'Sjppiy / To RF Filie; *nd -. Piismj Gis loict. Cnphiic Furnace Power Supply Oriptute Furaice : : 93> .>Hing eiqctrode; Ouriii^-tpbn ' RJ, discharge' Quartz Tutting Microrruichiried Qtiartx. Plate Fig. 1.2 Schematic illustration of various atmospheric pressure rf CCP sources. Reprinted from References[l, 51, 26-27]. advantages of the graphite furnace with those of the CCP, and offered simultaneous multielement determinations with lower detection limits than those for graphite furnace AAS. 9 More details about the FAPES will be discussed in next section. In addition to the above PP-CCP and FAPES, the electrode geometry for an rf-CCP has also been obtained using a tip-ring electrode geometry configuration [28, 29](Fig. 1.2c). The rf discharge is unipolar, and it is obtained at the tip of a sharp platinum electrode placed inside a quartz tube. A second electrode is a ring, and it is placed outside the tube and connected to ground. The outer electrode can increase the temperature and the stability of the plasma. The discharge is ignited by means of a Tesla coil, and then it can be easily maintained at atmospheric pressure in air or argon. This plasma can be operated at a low to medium power, and it may be used for atomic emission analysis of both pneumatically nebulized sample solutions and conductive solid samples [26, 30]. As a further development of the PP-CCP technology, an atmospheric pressure capacitively coupled microplasma (CCpP) is under development in our laboratory [27](Fig. 1.2d). This optical emission source with dimensions as small as 0.25 x 0.25 x 5 mm is implemented on a micro-machined fused silica chip, and the plasma discharge can be generated and sustained on the chip. This device is potentially useful as a detector for a chip based gas chromatograph. Microwave plasma torches (MPT), also referred to as capacitively coupled microwave plasmas (CMP), may also be considered CCP's. Microwave plasmas are produced by the interaction of electric fields at microwave frequency (usually 2450 MHz) with gases. Microwave plasmas are usually classified into two types according to the method of power coupling to the plasma gas. In the first type, the microwave energy is either coupled by an electric field (E coupling) or a magnetic field (H coupling) into the gas stream 10 in a discharge tube within a resonant cavity or other microwave supporting structure. This type is called a microwave induced plasma (MIP). In the second group, an inner conductor forms a capacitance against ground and transfers energy into the working gas stream through its tip, and the plasma is sustained at the tip of the central electrode of a cavity. This type of plasma is called a CMP. The CMP was first developed by Cobine in 1951 [31], and was later modified by Schmidt in 1959 [32], and applied to the analysis of sample solutions by Mavrodineanu and Hughes in 1963 [11]. A CMP can be operated at atmospheric or reduced pressure, and it may be formed over a wide power range (up to 1500 W). In addition, CMP sources have a fairly high tolerance for solvent and sample vapors. However, the CMP usually suffers problems of higher emission backgrounds, and analytical performance is poorer than that obtained with an MIP. 1.1.4 FAPES: Recent Developments FAPES has undergone a decade of development and characterization as a plasma atomic emission spectrometry method, and it has been recognized as a promising multielemental analytical technique. A summary of FAPES research to date is presented below. Several papers have reported some fundamental characteristics of the FAPES plasma [33-42]. Excitation temperatures derived from a Boltzmann plot of He(I) and Fe(I) were found to be 3300 K [34] and in the range of 3000-4500 K [36] respectively, while an excitation temperature measured with Fe(II) lines was 7610 K. The rotational temperature measurements of background species showed that there is a significant thermal gradient in 11 the source with OH rotational temperatures ranging between 680 K and 1050 K and N, + rotational temperatures ranging between 580 K and 1920 K with 60 W rf power applied to the center electrode [42]. Based on the measurement of the width of the Be(I)-234.861nm line profile, a kinetic or Doppler temperature was measured to be 7030 K [34]. Electron number densities were derived from the Stark broadening of the Hp line and found to be 9.6, 8.1, and 6.7 X 1013 cm" with an operational rf power of 100, 75, and 50 Watts, respectively. The helium ionization temperature, calculated from the measured electron number density, was found to be 10,920 K. The above temperature data suggest that a complete LTE does not exist in the FAPES plasma, and plasma temperatures may depend on many factors such as the thermometric species, and the operating conditions. Hettipathirana et al. [36] and Le Blanc et al. [42] described spatial distributions of atomic emission intensity from He and molecular emission intensity from OH and N 2 + . The emission intensities of these species were found to be most intense adjacent to the center electrode. Sturgeon et al. [43] reported spatially and temporally resolved images of the emissions from analyte species of Ag(I), Ca(I) and Ca(II) and plasma background species of He(I) by using a CCD camera system. The result showed that plasma and thermally induced processes have some influence on vaporization-atomization and ionization processes, leading to a convolution of the analyte density distribution and the excitation distribution. Several publications focused on the effect of easily ionized elements (E1E) on analyte excitation have shown that EIE's might affect analyte excitation in the FAPES source [33, 40, 44, 45]. The presence of an EIE could attenuate the analytical signal by physical expulsion of analyte from the excitation or observation volume as a result of a decreased plasma power by a greater production of photons, or the loss of plasma power coupling 12 efficiency due to changes in the load impedance of the plasma caused by the EIE. A study from Pavski et al. [45] also showed that, with the addition of NaCl, NaN03 and CsCl, analyte emission from Ag(I), Cu(I) and Ca(II) in the region between the center electrode and the tube wall is strongly suppressed. The degree of suppression relies on the (temporal) extent of vapor cloud overlap between analyte and EIE. Furthermore, the suppression brought about by the addition of Fe as an interfering matrix also suggests that the primary cause of suppression is the loss of the energy from the plasma due to excitation and ionization of matrix vapor. The effects of some operating conditions were also investigated in several papers [33, 35, 45-50]. The rf bias potential may cause enhancements in absolute sensitivity and detection limits for a number of less volatile elements [46]. Pavski et al. [45] studied the effect of dc bias on the spatial distribution, and found that the dc bias of the center electrode significantly affects the spatial distribution of He(I), Cu(I), Ag(I), Cs(I) and Ca(II) emission. An increase in the frequency of the applied rf power to the center electrode did not lower detection limits due to an increase in the background [35]. A better signal-to-noise ratio was found for an Ar plasma in the FAPES source [47], and thus Sun et al. [48] studied the influence of plasma gas composition on the operating and analytical characteristics of a FAPES source. With the addition of Ar to He, He(I) and Ar(l) excitation temperature increase 30% whereas argon ion excitation temperatures decrease from 33000 K to 26000 K in the presence of helium. Collisional exchange of internal energy between excited states of Ar and He accounts for these changes. A study of plasma support gas at controlled pressure demonstrated that an optimum pressure for analytical work may be unique for each element and is likely to be slightly higher than atmospheric pressure as a result of the competition 13 between increasing density, longer residence time, adsorption-desorption processes and decreasing excitation energy for collision [49]. When a Massmann-type FAPES system employing a L'vov platform and Pd modifier was used to atomize the volatile elements, the detection limits, sensitivity, and precision were significantly improved [33]. A two step furnace was found to permit a stable plasma in the tube prior to the introduction of analyte vapor, and figures of merit were similar to those obtained in a Massman-type FAPES with atomization from a platform and chemical treatment of the sample with Pd [50]. Some analytical figures of merit have been investigated in several papers [33-35, 46, 47, 50-54]. The linear dynamic range has been found to be 4 orders of magnitude with detection limits on the order of picograms and precision of roughly 5% [34]. As a further development, the FAPES source also has been investigated for a variety of other applications. Chan et al. [55] developed a new technique of determining lead in a chloride matrix by using atomic absorption spectrometry with electrothermal vaporization and FAPES atomization. The precision was found to be better than that obtained from the most widely used method - electrothermal vaporization atomic emission spectrometry with chemical modifiers and electrochemical methods. Jimentez et al. [56] determined the speciation of methyl- and inorganic-mercury in biological tissues using ethylation and gas chromatography (GC) coupled with a FAPES detector. The accuracy of the technique was validated by the analysis of certified samples from the National Research Council of Canada. A performance comparison was also made between FAPES and MIP-atomic emission spectrometry for the determination of mercury species in gas chromatography effluents [57]. The result showed that, had an optimal interface been used with the FAPES system, better performance could be achieved for GC effluent detection by means of FAPES. 14 Clearly, FAPES has been shown to be a promising multielement ultra-trace analytical technique that has detection limits comparable to graphite furnace atomic absorption spectrometry (GFAAS) [54]. As a result, it has been commercialized as an atomic emission source by Aurora Instruments (Vancouver, Canada). 1.1.5 FAPES: A Potential Ion Source During the past ten years, there has been a rapid growth in the analytical use of plasma atomic mass spectrometry. Since its advent, the inductively coupled plasma (ICP) has greatly increased the speed and sensitivity of analytical determinations, and the ICP has dominated trace elemental analysis. However, ICP-MS has some weaknesses. Sample introduction for ICP-MS usually employs solution nebulization technique, and only 1-3% of the nebulized sample actually reaches the plasma. A characteristic of ICP-MS is that it is less tolerant of dissolved solids in the samples. The ICP is sustained by an electrical discharge in argon caused by radio frequency (rf) power applied through a load coil. The use of argon as the plasma gas leads to some isobaric polyatomic ion interferences such as 4 0 A r 1 6 O , which affect the determination of certain analytes such as 5 6Fe, and plasmas formed with argon do not efficiently ionize some elements with high ionization potentials such as P and S. The use of helium for ICP is an alternative [68], but has been ignored by analytical community. The ETV nature of the FAPES source coupled with a helium plasma and high analyte transport efficiency makes it a very attractive candidate as an ion source for mass spectrometry. As a combined source, FAPES can also vaporize and atomize solid or liquid samples. The He plasma within the cuvette of the FAPES source would reduce nearly all of the isobaric interferences from the support gas, because helium has only a single low mass 15 isotope, which results in fewer polyatomic ions. The ionizing ability of a He plasma is greatly increased by virtue of the higher ionization potential of helium compared to that of argon, 24.2 eV versus 15.7 eV. Also, helium has a suitable electrical resistivity to allow effective coupling and power dissipation. Based on the above potential advantages of the FAPES plasma, it appears that the FAPES source may be developed into an effective ion source of elemental mass spectrometry. In recent years, this kind of ion source has been under development in the laboratories of Dr. Ralph Sturgeon and Dr. Mike Blades, and some interesting results has been reported so far [61, 65-67]. 1.2 Thesis Objective FAPES has been shown to offer the advantage of combining the ultratrace analysis capability derived from graphite furnace atomic absorption spectrometry with the simultaneous multielement analysis capability derived from optical atomic emission spectrometry. As was previously stated, the high ionization energy of the He plasma in the FAPES source gives it the potential to be an excellent ionizing plasma. Other advantages such as combined atomization and ionization processes as well as potential mass spectral simplicity also make the FAPES source an attractive ion source for a low cost, compact, elemental mass analyzer. A more complete study of the operational characteristics of background species may help us thoroughly understand the fundamental and analytical characteristics of the FAPES plasma. Gaining more information about spectroscopic temperatures such as the excitation 16 temperature and the rotational temperature of plasma background species will also help us better understand the nature of the plasma source. In addition, the plasma temperatures (excitation temperature of He and rotational temperatures of N 2 + and OH) can be used as diagnostic tools to acquire more information about fundamental properties of the FAPES source. In order to develop FAPES as an ion source for elemental mass spectrometry, knowledge of the temporal behavior of analyte ionization in the source could be very useful for improving our understanding of its ionization characteristics. The analyte ionization efficiency in the FAPES source may be improved by the optimization of a variety of operating conditions such as rf power, gas flow rate, and atomization temperature. The effects of matrices, Pd modifier, and dimensions of the center electrode on plasma temperatures and analyte ionization efficiency were studied with the goal of improving the ionization efficiency of the FAPES source. In summary, the objective of this thesis work was to develop a compact FAPES source, and perform studies of characterization and improvement of analyte ionization in the FAPES system. The primary interest behind these investigations is to try to produce an elemental mass analyzer with low cost, simple operation and familiar environment to the users of GFAAS. 1.3 Analyte Atomization 1.3.1 Electrothermal Atomizer FAPES employs a conventional atomic absorption graphite furnace for the vaporization and atomization of analytical samples, and thus the atomization process and mechanisms are similar to those of the conventional graphite furnace. 17 As one of the popular electrothermal atomizers, the graphite furnace has been widely used for trace and ultratrace elemental analysis in combination with atomic absorption spectrometry. The graphite furnace as an atomization device for analytical AAS was proposed and developed by L'vov in 1961 [58], and later Massmann developed a considerably simpler furnace [59], referred to as a Massmann graphite furnace. Due to its simplicity, Perkin-Elmer commercialized the Massmann graphite furnace in 1971. This furnace consists of a graphite tube and electrical contacts. The graphite tube is 55 mm long with an inner diameter of 6.5 mm, and the tube is heated resistively from the ends. Resistive heating permits easy control of the temperature for thermal pretreatment, volatilization, and atomization of the analyte. To prevent the oxidation of graphite at high temperature, the furnace is made of polycrystalline electrographite with a thin layer of pyrolytic-graphite and purged with an inert gas. The furnace can be heated up to 3000 °C. 1.3.2 Atomization Efficiency The atomization efficiency is often used to measure the effectiveness of free atom production, and it is defined as follows [60]. (£a)t = (Na + Ni)/N = r|pvpa (1.2) where (sa )t is the atomization efficiency, N a is the number of free gaseous atoms, N; is the number of ions, N is the deposited amount, r\ is the vapor transport efficiency from the vaporizer to the atomizer, pv is the fraction volatilized from the condensed phase to the vapor phase, and pa is the fraction of the analyte that is atomized. 18 Equation (1.2) tells us that the atomization efficiency is proportional to the vapor transport efficiency, the volatilization fraction, and the atomization fraction. Since the FAPES source is a combined source in which a graphite tube acts as desolvator and atomizer for analytical samples, the vapor transport efficiency is very high. Analyte volatilization refers to a process in which analytical sample (liquid or solid) is transformed from the condensed phase into the vapor phase. In electrothermal atomization, this conversion is usually realized by a drying stage and an ashing (or pyrolysis) stage. The drying stage is to evaporate the solvent at an appropriate rate leaving the analyte in a dried form. Thus, the drying temperature should be selected appropriately so that the analyte droplet is heated carefully and slowly without causing any violent boiling and sample expulsion as aerosol droplets. The ashing stage is to eliminate as much of the matrix as possible before atomization initiation, and thus the selected ashing temperature should be high enough to evaporate the bulk of the matrix before the atomization stage. However, a relatively high temperature may result in a loss of the analyte through evaporation, particularly in the case of the more volatile elements which have relatively high vapor pressures. In general, analyte volatilization could be improved through optimization of drying temperature and ashing temperature. The atomization fraction depends on the analyte atomization process and the analyte atomization mechanisms involved in the atomization stage. The atomization process for an analyte should be rapid so that almost all of the analytical sample is confined to the observation volume before the gaseous sample diffuses out of that volume, causing a significant loss. Therefore, a fast heating rate could help produce a free atom reservoir instantaneously for the excitation and ionization processes, leading to an enhancement of 19 analytical performance. Since atomization mechanisms are both analyte and matrix dependent, analytes and their chemical environment will play important roles in the atomization fraction. 1.4 The rf Discharge in the FAPES Source The atmospheric pressure rf helium discharge is self-ignited by applying sufficient (~9 W) rf power on the center electrode. The discharge is essentially capacitively coupled to the rf generator with a blocking capacitor connected in series so that no net current flows through the circuit. The plasma is heated by the rf electric fields, and it fills the furnace volume. The operating frequency plays a major role in the formation and operation of the rf discharge. At low frequencies, where ions can follow the changing electric fields, the discharge will behave similarly to the dc glow discharge. As the frequency increases, the ions will no longer be able to follow the instantaneous electric field, but will instead respond to the time averaged fields. As a result of the difference in the mobility of positive ions and electrons as well as an unequal electrode discharge configuration, an rf self-bias potential will arise which produces a time-average negative voltage on the smaller electrode. Ions will be accelerated by the difference between the time-average plasma potential and the time-average bias potential, and consequently a very bright region surrounding the center electrode is formed. The FAPES plasma is not spatially homogeneous due to an unequal electrode configuration. An end-on view of the FAPES plasma is presented in Fig. 1.3. Five distinct regions are observed in the FAPES source, and moving from the center electrode towards the 20 furnace wall, they are the inner negative glow, the inner Faraday dark space, the positive column, the outer Faraday space, and the outer negative glow. The inner negative glow is the Fig. 1.3 Schematic representation of an end-on view of the FAPES plasma source. Adapted from [61]. The relative emission intensity is represented by a gray scale. 1-Graphite Furnace, 2-Negative Glow, 3-Faraday Dark Space, 4-Positive Column, 5-Center Electrode brightest region where there is the highest excitation and rotational temperature and the highest degree of ionization. The outer negative glow is associated with the reversal of polarity of the applied rf power during which time the wall behaves as a cathode, moderated by the electric field strength much weaker over this large surface area. In the Faraday dark spaces, although the electric field strength is relatively low, electrons can be accelerated toward the positive column. Compared with the negative glow region, the positive column is less bright since fewer excitation collisions occur here. 21 The rf discharge may have plasma sheaths in which strong electric fields accelerate ions and electrons. Since the plasma sheath may be only a few tens of pm, it may be invisible. One plasma sheath is next to the center electrode and another is next to the furnace wall. In those regions, the electron number densities are lower, but the power density and the electric field are much greater. It is in the plasma sheath that the bulk of the rf power is transmitted to the plasma electrons. There are three possible mechanisms that are responsible for energy input to the rf discharge. First, energetic ions striking the center electrode, and the impact may cause the formation of secondary electrons. These electrons can gain energy from the electric field in the plasma sheath near the center electrode. Second, the oscillating electric field may accelerate electrons in the negative glow regions. These electrons may be accelerated through the plasma sheath and acquire additional energy. Finally, the local oscillating electric field can input energy directly into the electrons in the positive column. 1.5 Excitation and Ionization 1.5.1 Excitation and Ionization Mechanisms In the FAPES source, the species present may come from four main sources: plasma gas, solvent, dissolved analytical sample, and entrained atmosphere. The principal species present are neutral atoms or molecules, ions, and unbound or free electrons. Although three major processes may contribute to the excited state population, i.e. collisional processes, radiative excitation and decay processes, collisional processes are generally considered as the dominant analyte excitation and ionization mechanisms. Collisions fall into two categories depending on whether the internal energy of the colliding partners is changed by the collisions. A particle includes two types of energies: internal energy and kinetic energy. 22 Internal energy refers to electronic excitations in atoms or electronic, vibrational, and rotational excitations in molecules. An elastic collision only changes the kinetic energy of the colliding partners while an inelastic energy collision alters both the kinetic energy and internal energy. Since the effect of elastic collisions is to redistribute energy of the particles based on their masses, the elastic collisional procress is negligible for particles with much different masses. Few studies on excitation and ionization process mechanisms in the FAPES source have been reported over the years, but for an introduction to excitation and ionization mechanisms, the following discussion is provided. Based on the category of species involved in the inelastic collisions, Table 1.2 summarizes possible excitation and ionization processes in the FAPES source. a. Ionization and Excitation of Helium Atoms In FAPES, helium is employed as the plasma gas. Initially, electrons and helium ions come from the breakdown of helium gas by the rf discharge. Thereafter, three important processes may occur in the FAPES plasma for the excitation and ionization of helium atoms: electron impact, double Penning ionization and helium ion impact. Electron impact is a very important inelastic collision process in which an internal energy transfer from energetic electrons to helium atoms occurs, and as a result helium ions (He+), helium metastable atoms (Hera), or excited state helium atoms (He) are produced. An illustrative representation of atomic energy level for He is presented in Fig. 1.4. The ionization energy of He is about 24.6 eV while the first electronic excited state is 20.6 eV, and thus electrons involved in these electron impact processes should be very energetic electrons. In electron impact processes, 23 the production of electrons caused by ionization leads to electron multiplication and thus sustains the operation of the plasma. The lifetime of He* is very short, and thus the excitation is often followed by a radiative decay. However, metastable helium atoms have long radiative lifetimes because radiative transitions to lower-energy levels are spin-forbidden, and the lifetimes are longer than the collisonal lifetimes of gaseous atoms. There are two types of metastable atoms i.e. He(2'S) and He(23S), and their energies are 20.6 eV and 19.8 eV respectively. However, He(2'S) is usually not considered because this state can quickly convert to He(23S) state by superelastic collisions [62, 63]. When helium metastable atoms collide with each other, a double Penning ionization may occur due to their considerable energy. As a consequence, a helium ion and an energetic electron are created, and the electron may further collide with other species. Helium ions may be accelerated in the oscillating field and gain more energy. When these energetic ions collide with helium neutral atoms, a compound helium molecule may be created during a helium ion impact process. Although the electron number density originating from ionization of helium gas is low, the production of electrons through thermionic emission from the heated center electrode may be substantial. Those electrons may gain energy from the strong electric field in the plasma sheath, leading to an increased probability of electron impact process. Therefore, electron impact process may play a dominant role in helium excitation and ionization among these processes. 24 Table 1.2 Excitation and Ionization Mechanisms in the FAPES Source I. Ionization and Excitation of Helium Atoms A. Electron impact He0 + e- -» He+ + 2e" He0 + e" —> He7 He*+ e" B. Double Penning ionization Hem+Hem^He ++ He" + e C. Helium ion impact He° + He + ^He 2 + + hv II. Ionization and Excitation of Plasma Background Species A. Electron impact X + e" - » X * + e" X + e" -» X7X+* + 2e" B. Charge transfer X + He2+ -» X7X+* + 2He° C. Penning ionization X + He" -» X7X+* + He0 + e" III. Ionization and Excitation of Analyte Atoms A. Electron impact M + e" —> M* + e M + e' -» M + /M + * + 2e" B. Charge transfer M + X + -> M7M +* + X M + He2+ -» M + /M + * + 2He° C. Penning ionization M + Hem -» M+/M+* + He0 + e" Note: X, molecular species; M, analyte atom; He'/X'/M', excited atoms or molecules; Hem, helium metastable atoms including He (2'S) and He (23S). 25 Fig . 1.4 Schematic representation of helium atomic energy levels. Adapted from [3]. b. Ionization and Excitation of Plasma Background Species Electron impact also excites, and/or ionizes plasma background molecular species. Furthermore, charge transfer and Penning ionization also contribute to the excitation and ionization of background species. Since the ionization potentials of most molecular species are much lower than that of helium, less energetic electrons can also cause efficient ionization of molecular species. As a result, a greater extent of ionization of molecular species is produced. In addition, the potential energy of a helium molecular ion (He2+) may be transferred to molecular species such as N, and CO [64], and thus the transfer of charge between ions and molecules also leads to ionization and excitation of molecular species. 26 c. Ionization and Excitation of Analyte Atoms For analyte ionization and excitation, electron impact, charge transfer, and Penning ionization are the most likely candidates. The electron impact process is almost the same as that for plasma background species. Since there are many different kinds of molecular ions and the number density of these molecular ions is much higher than that of helium ions in the FAPES source, a charge transfer process involving an analyte atom and a molecular ion most likely occurs. The collision can lead to the transfer of an electron from the atom to the molecular ion if the energy difference between the molecular ion ground state or metastable level and the energy levels of the resulting analyte ion is very small. Although Penning ionization may contribute to analyte ionization, the production of metastable atoms is not favorable at atmospheric pressure, and thus this mechanism only plays a minor role in analyte ionization and excitation. 1.5.2 Plasma Temperature Measurements The plasma temperature is a major consideration in evaluating the robustness of a plasma source. A high temperature may promote atomization, excitation, ionization, and the reduction of matrix interferences. Thus, plasma sources with relatively high temperature are often desirable because of their analytical merit. Since a plasma is a mixture of electrons, ions and neutrals, one temperature is not sufficient to describe the plasma characteristics. In addition, since complete thermal equilibrium is not established in plasma sources, a number of different temperatures each of which is described by a thermodynamic function may be required. The electron temperature (T.) can be derived from continuum intensity measurements or measured by the use of an electrostatic probe (Langmuir probe). The gas 27 temperature (T ) is usually determined from Maxwell velocity distribution whose average velocity is derived from a spectral line broadening due to the Doppler effect. A rotational temperature (Trol) is traditionally calculated from diatomic rovibronic line intensity measurements. An excitation temperature (Texc) can be measured from a Boltzmann distribution of the number density. A Saha equation may be used to determine the ionization temperature (Tion) if the electron number density is known in advance. The "usual" relationship among these plasma temperatures is illustrated as follows. In this thesis, rotational temperatures and excitation temperatures are measured to characterize the plasma. The rotational temperatures of molecular species are derived from their rovibronic spectra. Different molecular species require different methods to calculate their respective rotational temperatures. Chapter 3 will describe how the rotational temperatures of N 2 + and OH were measured. The excitation temperature of an element is derived from the transitions between electronic states. The line pair intensity ratio method is used to calculate excitation temperatures of elements. With two spectral lines from the same ionization state, an element excitation temperature can be obtained using the following equation which is derived from a Boltzmann distribution: T > T = T > T > T e ion exc * rol " j gas (1.3) T exc _ kB(ln (1.4) 28 where Ii and I2 are emission intensities, X\ and X2 are wavelengths, Ai and A2 are Einstein transition probabilities, gi and Q2 are statistical weights, Ei and E2 are excitation energies, and ka is the Boltzmann constant. 1.5.3 Ionization Efficiency Experimental measurements of the degree of ionization is usually used to evaluate the ionization efficiency of an analytical plasma. This can be obtained by measuring the relative intensities of atomic and ionic emission line and applying the following equation [65]: N0ion N0atom M Q l [ 9 A 1 e ( E i o n - E a t o m ) / k B T (1.5) gA J ion \.\XZ(T)J atom where N 0 is the total number density of the species under study, I is the total intensity of a spectral line at wavelength X, Z(T) is the partition function, g is the statistical weight of the excited state of the species, A is the Einstein transition probability; E i o n and E a t o m are the excitation energies of the atomic and ionic lines respectively, ke is the Boltzmann constant, and T is the atomic excitation/ionic excitation temperature. From the value of this ratio the degree of ionization (in %) of the analyte can be calculated using: r Nnion VrNLatom, „ ^ % Ionization = 7 2 r X 100 (1.6) \ , NJon \ N0atomy The degree of ionization of an analyte in an analytical plasma can be utilized as a diagnostic tool since it can be very sensitive to changes in operating conditions. Once the 29 plasma has been optimized for analytical use, the degree of ionization can be used to monitor, and in fact verify, the operating parameters for analytical rf plasmas. 1.6 Thesis Overview In this thesis, we mainly focus on further characterization and understanding of the FAPES plasma, investigation of analyte excitation and ionization processes, and enhancement of ionization efficiency. The fundamental properties such as excitation temperature, rotational temperature, and temporal emission characteristics of background species have been investigated. The temporal atomic and ionic emission characteristics, effects of operating conditions on analyte ionization, analytical figures of merit, and effects of matrix, modifier and dimensions of the center electrode on analyte ionization have been studied. A new FAPES source and an experimental system, developed for the measurements of plasma characteristics and analyte ionization, is described in chapter 2. Chapter 3 deals with characterization of the FAPES source. The results include plasma emission characteristics, the temporal profiles of He (I), N, + and OH, and the measurements of excitation temperature of He(I) and rotational temperatures of N, + and OH. The excitation and ionization mechanisms of N 2 + and CO + are also discussed. Chapter 4 reports temporal atomic and ionic emission characteristics of Mg, Cd, Fe, Zn and Cr. The effects of gas flow rate, rf power, and atomization temperature on emission characteristics are also described. Chapter 5 discusses the improvement of analyte ionization by optimizing the operating conditions such as gas flow rate, atomization temperature, rf power, and the dimensions of the center electrode. In Chapter 6, the effects of analyte matrices and Pd modifier on 30 emission characteristics and analyte ionization are discussed, and the atomization mechanisms of magnesium are also presented. Analytical performance of this FAPES source is described in Chapter 7, and the ionization temperature and electron number density are also reported. Conclusions of this research are summarized in Chapter 8, and future research suggestions are provided. 1.7 References 1. D.C. Liang and M.W. Blades, Anal. Chem. 60: p. 27 (1988). 2. R.E. Sturgeon, S.N. Willie, V. Luong, S. S. Berman and J. G. Dunn, J. Anal. At. Spectrom.4: p. 669 (1989). 3. S.M. Rossnagel, J.J. Cuomo and W.D. Westwood, Handbook of Plasma Processing Technology (Noyes Publications, 1990). 4. B. Chapman, Glow Discharge Processes: Sputtering and Plasma Etching (John Wiley & Sons, New York, 1980) 5. A.L. Stolov, Uch. Zap. Kap. Kazan. Gos. Univ. 116: p. 118 (1956). 6. E. Badarau, M. Giurgea, G. H. Giurgea and A. T. H. Trutia, Spectrochim. Acta 11B: p. 441 (1957). 7. H.A. Schwab and C K . Manka, J. Appl. Phys. 40(2): p. 696 (1969). 8. H.A. Schwab and R.F. Hotz, J. Appl. Phys. 41(4): p. 1503 (1970). 31 9. H.A. Schwab, Proc. I. E. E. E. 59(4): p. 613 (1971). 10. G.D. Cristescu, Ann. Physik. 6: p. 153 (1960). 11. R. Mavrodineanu and R.C. Hughes, Spectrochim. Acta 19B: p. 1309 (1963). 12. V. Trunecek, Z Chem. 4: p. 358 (1964). 13. H. Dunken, G. Pforr and W. Mikkeleit, Z. Chem. 4: p. 237 (1964). 14. H. Dunken and G. Pforr, Z. Phys. Chem. 230: p. 48 (1965). 15. C D . West and D.N. Hume, Anal. Chem. 36: p. 412 (1964). 16. G. Pforr and K. Langner, Z Chem. 5: p. 115 (1965). 17. G.D. Cristescu and M. Giurgea, Z Chem. 7: p. 360 (1967). 18. K.A. Egorova, Zh. Prikl. Spektrosk. 6: p. 12 (1967). 19. A.V. Zvyagintsev, R.V. Mitin and K.K. Pryadkin, Sov. Phys. Tech. Phys. 20: p. 177 (1975). 20. N.l. Gonchar, Sov. Phys.-Tech. Phys., 1975. 20: p. 407 (1975). 21. K.A. Ergova, Zh. Prikl. Specktrosk., 1967. 6: p. 12 (1967). 22. G.I. Babat, J. Inst. Elec. Eng. (London) 94: p. 27 (1947). 23. D. Huang, D.C. Liang and M.W. Blades, J. Anal. At. Spectrom. 4: p. 789 (1989). 24. D. Huang and M.W. Blades, J. Anal. At. Spectrom. 6: p. 215 (1991). 32 25. D.C. Liang and M.W. Blades, Spectrochim. Acta 44B(10): p. 1059 (1989). 26. E.A. Cordos, S. D. Anghel, T. Frentiu, A. Popescu, J. Anal. At. Spectrom. 9(5): p. 635 (1994). 27. A. Bass, C. Chevalier and M.W. Blades, J. Anal. At. Spectrom. 16 (9): p. 919 (2001). 28. E. Tataru, S.D. Anghel and A. Popescu, Rev. Roum. Phys. 6: p. 29 (1991). 29. E.Tataru, S.D. Anghel and A. Fordor. Proceedings of XXVII Colloquium Spectroscopicum Internationale. (Norway, 1991) 30. S.D. Anghel, et al., Fresenius J. Anal. Chem. 355: p. 252 (1996). 31. J.D. Cobine and D.A. Wilber, J. Appl. Phys. 22: p. 835 (1951). 32. W. Schmidt, Elektron. Rundschau. 13: p. 404 (1959). 33. R.E. Sturgeon, S.N. Willie, V. Luong and S. S. Berman, J. Anal. At. Spectrom. 6: p. 19(1991). 34. R.E. Sturgeon, S.N. Willie and V.T. Luong, Spectrochim. Acta 46B: p. 1021 (1991). 35. R.E. Sturgeon, S.N. Willie, V. Luong and J.G. Dunn, Appl. Spectrosc. 45: p. 1413 (1991). 36. T.D. Hettipathirana and M.W. Blades, Spectrochim. Acta 47B(4): p. 493 (1992). 37. R.E. Sturgeon and S.N. Willie, J. Anal. At. Spectrom. 7: p. 339 (1992). 38. T.D. Hettipathirana and M.W. Blades, J. Anal. At. Spectrom. 8: p. 955 (1993). 33 39. S. Imai and R.E. Sturgeon, J. Anal. At. Spectrom. 9(4): p. 493 (1994). 40. S. Imai and R.E. Sturgeon, J. Anal. At. Spectrom. 9: p. 765 (1994). 41. V. Pavski, C L . Chakrabarti and R.E. Sturgeon, J. Anal. At. Spectrom. 9: p. 1399 (1994). 42. C.W. LeBlanc and M.W. Blades, Spectrochim. Acta 50B: p. 1395 (1995). 43. R.E. Sturgeon, V. Pavski and C L . Chakrabarti, Spectrochim. Acta 51B: p. 999 (1996) . 44. T.D. Hettipathirana, Ph. D. Dissertation, University of British Columbia (1993). 45. V. Pavski, R.E. Sturgeon and C L . Chakrabarti, J. Anal. At. Spectrom. 12: p. 709 (1997) . 46. R.E. Sturgeon, S.N. Willie, V. Luong and R. K. Marcus, Spectrochim. Acta 48B: p. 893 (1993). 47. G.F.R. Gilchrist, P.M. Celliers, H.Yang, C. Yu and D.C. Liang, J. Anal. At. Spectrom. 8: p. 809 (1993). 48. F.S. Sun and R.E. Sturgeon, Spectrochim. Acta 54B: p. 2121 (1999). 49. S. Imai, R.E. Sturgeon and S.N. Willie, J. Anal. At. Spectrom. 9: p. 759 (1994). 50. K.E.A. Ohlsson, R.E. Sturgeon, S.N. Willie and V. Luong, J. Anal. At. Spectrom. 8: p. 41 (1993). 34 51. D.L. Smith, D.C. Liang, D. Steel and M.W. Blades, Spectrochim. Acta 45B: p. 493 (1990). 52. R.E. Sturgeon, S.N. Willie, V. Luong and S. S. Berman, J. Anal. At. Spectrom. 5: p. 635 (1990). 53. R.E. Sturgeon, S.N. Willie, V. Luong and S. S. Berman, Anal. Chem. 62: p. 2370 (1990). 54. M.W. Blades, Spectrochim. Acta 49B: p. 47 (1994). 55. G.C.Y. Chan and W.T. Chan, J. Anal. At. Spectrom. 13: p. 209 (1998). 56. M.S. Jimenez and R.E. Sturgeon, J. Anal. At. Spectrom. 12: p. 597 (1997). 57. W. Freeh, J.P. Snell and R.E. Sturgeon, J. Anal. At. Spectrom. 13: p. 1347 (1998). 58. B.V. L'vov, Spectrochim. Acta 17B: p. 761 (1961). 59. H. Massmann, Spectrochim. Acta 23B: p. 215 (1968). 60. R.E. Sturgeon and S.S. Berman, Anal. Chem. 55: p. 190 (1983). 61. C.W. LeBlanc, Ph. D. Dissertation, University of British Columbia (1996). 62. D.R. Bates, Phys. Rev. 77: p. 718 (1950). 63. C B . Collins and W.W. Robertson, J. Chem. Phys. 40: p. 701 (1964). 64. M. Endoh, M. Tsuji and Y. Nishimura, J. Chem. Phys. 79: p. 5368 (1983). 35 65. S.Y. Lu, C.W. LeBlanc and M.W. Blades, J. Anal. At. Spectrom. 16(3): p. 256 (2001). 66. R.E. Sturgeon and R. Guevremont, Anal. Chem. 69: p. 2129(1997). 67. R.E. Sturgeon and R. Guevremont, J. Anal. At. Spectrom.13: p. 229(1998). 68. E.Hy wei Evans, Inductively Coupled And Microwave Induced Plasma Sources for Mass Spectrometry (Cambridge, Royal Society of Chemistry, Great Britain, 1995). 36 Chapter 2 EXPERIMENTAL SYSTEM As mentioned in Chapter 1, one of the objectives of this research was to design an improved FAPES source and to set up an experimental system for the characterization of the FAPES plasma and for studies of analyte excitation and ionization in this FAPES source. The experimental system described in this chapter was used for most of the research described in the remainder of this thesis unless otherwise specified. 2.1 Instrumentation A schematic diagram of the experimental system used in this thesis is presented in Figure 2.1. The system consists of an rf power generator, an rf matching network, a FAPES source, a furnace power supply, a light collection system, an optical detection system, and a data acquisition system. Radio frequency power was provided by a 13.56 MHz rf generator with an impedance matching network (Model RF5S and Model AM5 respectively, RFPP Inc., Voorhees, NJ, USA). The light collection system employed two plano-convex fused silica lenses which were used to focus an image of the negative glow region of the plasma into the optical fiber, which transferred the light to the entrance slit of a lm Czerny-Turner monochromator (Model 2061, Scoeffel-McPherson, Acton, MA, USA) which was equipped with a holographic grating for selecting analytical lines. An intensified photodiode array (IPDA) (Model IRY-700G/B/PAR, Princeton, NJ) was used as the detector, and the Optical Spectrometric Multi-channel Analyzer (OSMA) program for the OSMA detector controller ST-120 system was used to collect, display and manipulate data. R F Generator Tuner AM-5 Matching Network 0*""| F A P E S Source C L HGA-500 Furnace Power Supply Fig. 2.1 Schematic illustration of the experimental system 38 With this instrumental arrangement, it was possible to collect spectrally and temporally resolved atomic and ionic emission data simultaneously. 2.1.1 The FAPES source The FAPES source built for this project was based on the modification of a Perkin-Elmer Model HGA-500 graphite furnace. A schematic diagram of the FAPES source is provided in Fig.2.2. The graphite furnace was assembled on an adjustable mount which provided horizontal and vertical adjustment of the furnace position. The furnace assembly could be opened pneumatically in order to change the graphite tube. When the furnace was heated, a high dc current is provided to the graphite tube through the contact cylinders. The graphite tube could be heated to a preselected atomization temperature using a rapid ramping rate. The end of the ramping period was detected using an optical sensor, which contains a silicon photodiode with associated electronic sensor. The internal gas stream, which was set at some gas flow rate, was led from both ends into the graphite rube and exited through the sample introduction hole in the graphite tube. The external gas stream was led through holes in the graphite contact cylinders into the space between the graphite tube and the contact cylinders, and it exited through the sample introduction port and the space between the contact cylinders. Helium was used for the internal stream for the purposes of plasma generation, to provide an inert atmosphere to prevent oxidation of the graphite tube, to minimize oxide formation, and to remove sample vapors and fumes formed during an atomization cycle. Graphite Cooling Ring External Gas Stream Internal Gas Stream HgMSJg Fig. 2.2 Schematic diagram of the FAPES source 40 The right-hand window of the furnace was removed to mount an rf connector which held the graphite electrode which was placed coaxially in the graphite tube. A schematic illustration of the rf connector interface is provided in Fig. 2.3. At the left in the diagram is the electrode holder, and at the right is an N-type male connector to which the matching network was connected. Different sizes of electrode holders could be inserted into the left copper adapter of the rf connector interface. A Vespel spacer was used to insulate the copper adapter from the graphite tube, and an "O" ring was employed to seal the system. Fig. 2.3 Schematic representation of the rf connector interface Compared with the source previously used in our laboratory [1-6], the new HGA-500 based source has the following improved characteristics. • Reduced void volume providing more efficient and uniform heating and cooling of the graphite tube. 41 • A faster heating rate allowing the atomic population to build up rapidly creating a high instantaneous atomic concentration in the atomizer volume increasing the analytical sensitivity. • The addition of a transistor-transistor-logic (TTL) circuit inside the power supply enabling us to trigger the data acquisition system at any time during an atomization cycle. • Improved ability to control and change the settings of temperature and gas flow rate at different stages of an atomization cycle. 2.1.2 Light Collection System An optical train, consisting of two plano-convex lenses with their convex surfaces facing each other [7], was found to be the optimum configuration for collecting and conveying light emission from the furnace. A collimating lens (f = 150 mm, Melles Griot Inc., Irvine, California, USA) and a focusing Lens (f = 100 mm, Melles Griot Inc., Irvine, California, USA) was used to collect the light and to focus the 1:1 image onto an optical fibre (200 pm I.D., Polymicro Technologies, Phoenix, Arizona, USA), which was used to transmit the light to the entrance slit of the monochromator. A He-Ne laser (Melles Griot, Carlsbad, California, USA) was employed to assist in the alignment of the light collection system. 2.1.3 Optical Detection System A 1.0 m Czerny-Turner scanning monochromator (Model 2061 Scoeffel-McPherson, Acton, MA, USA), with either a 1200 or 3600 lines mm"1 holographic grating, was employed 42 to spectrally resolve the light from the FAPES source. The 3600 lines mm"1 grating was used to resolve molecular emission spectra of the plasma background species while the 1200 lines mm"1 grating was used for the spectral resolution of other analytical lines. The entrance and exit slits of the monochromator were both set at 50 pm. An intensified photodiode array (IPDA) detector (Princeton Instruments, Model IRY-700/G/B/PAR, Trenton, NJ, USA) was mounted at the exit plane of the monochromator using a coupling ring built in-house. The IPDA A/D board has a 14-bit dynamic range, and its spectral range is 180-910 nm with a 500:1 variable gain enabling up to a 1000 s integration time, 33 ms array readout time, and external triggering. The IPDA detector was controlled using an ST-120 controller (Princeton Instruments, Model S-120A, Trenton, NJ, USA), and was thermoelectrically cooled to below -25°C to minimize dark current. This allowed simultaneous measurement of emission intensity for an approximately 20 nm spectral window. 2.1.4 Data Acquisition System The ST-120 controller was interfaced to a Pentium II computer running the Optical Spectrometric Multi-channel Analysis (OSMA) software from Princeton Instruments Inc.. After a TTL signal from the HGA-500 power supply triggered the ST-120 controller, a real time emission profile was collected and displayed on screen. The emission data could be imported into GRAMS (Graphic Relational Array Management System, Galactic Industries Corporation, USA) or Igor (WaveMetrics Inc., USA) for post-acquisition processing. Peak emission intensities were corrected for using a baseline subtraction method which removed any contribution to the peak heights from the background continuum radiation due to ion-43 electron recombination and from blackbody radiation of the graphite furnace and center electrode. Both peak height and peak area were used for the peak emission characterization. 2.2 Operating Conditions Control The HGA-500 furnace can manage the application of a temperature setting to its graphite tube by means of a temperature feedback control assembly. In addition, the applied temperatures were further calibrated using an optical pyrometer (Ircon Series 1100, Model 11x30, IL, USA) which was positioned on a tripod (Manfrotto, Italy) and focused onto the graphite tube wall through the dosing hole. The output from the pyrometer was collected and stored using a digital, real-time, oscilloscope (Model TDS 380, Tektronix Inc., Beaverton, Oregon, USA), and the furnace temperature was extracted from the digitized data using the calibration curve provided by the pyrometer manufacturer. Temperatures below the detection range of the pyrometer were measured using a thermocouple (Model 80TK, John Fluke, Everett, WA, USA). The HGA-500 furnace controller can be used to regulate the gas flow rates of the external streams and internal streams. An rf forward power level was set through the front panel of the rf generator, and its output value was regulated by an automatic power tuner (Model 7600001010, RFPP Inc., Voorhees, NJ, USA) with a reflected power value of 0 W under most cases. 2.3 Analyte Atomization As was discussed in Chapter 1, an analyte atomization cycle consisted of three stages: a drying stage, an ashing stage, and an atomization stage. During the drying stage, solvent 44 was evaporated. The major objective of the ashing stage was to remove any organic residues in the sample by decomposition and vaporization and to minimize the formation of gas-phase species that may chemically or spectrally interfere with analytical determination. During the atomization stage, the analyte and the remainder of the sample form a vapor cloud inside the furnace. Because of the functional difference of these stages, different combinations of temperature, ramping time and hold time were used for these three stages. Thus, different heating rates were actually employed during an atomization cycle. For the analyte atomization stage, higher heating rates can be used to improve atomization processes as a result of the promotion of free atom formation at higher temperatures. On the other hand, a higher heating rate would potentially increase the rate of gas expansion in an atmospheric pressure atomizer such as FAPES, leading to an increased diffusion rate and a decreased residence time. Generally, we would choose an atomization temperature such that the furnace is heated as rapidly as possible to ensure that the atom generation is rapid compared to the atom loss. 45 2.4 References 1. D.C. Liang and M.W. Blades, Spectrochim. Acta 44B: p. 1059 (1989). 2. T.D. Hettipathirana and M.W. Blades, Spectrochim. Acta 47B: p. 493 (1992). 3. T.D. Hettipathirana and M.W. Blades, J. Anal. At. Spectrom. 7: p. 1039 (1992). 4. T.D. Hettipathirana and M.W. Blades, J. Anal. At. Spectrom. 8: p. 955 (1993). 5. C.W. LeBlanc and M.W. Blades, Spectrochim. Acta 50B: p. 1395 (1995). 6. C.W. LeBlanc and M.W. Blades, Applied Spectroscopy 51: p. 1715 (1997). 7. P.B. Farnsworth, B.W. Smith and N. Omenetto, Spectrochim. Acta 45B: p. 1151 (1990). 46 Chapter 3 MEASUREMENT OF TEMPORAL EMISSION AND PLASMA TEMPERATURES 3.1 Introduction Among the diagnostic tools used for the fundamental studies of plasmas, spectroscopic methods are very commonly utilized because they are relatively easily implemented and are non-invasive and, as a result, do not perturb the plasma. For these reasons we have used spectroscopic measurements to investigate the fundamental characteristics of the FAPES plasma. Background emission from an analytical plasma can impact its analytical characteristics, such as signal-to-background ratios and hence detection limits. Therefore, an investigation of spectral emission characteristics in the FAPES source can be used to help optimize and fully utilize the FAPES plasma for analytical determinations. Since FAPES is a transient source, a knowledge of the temporal emission characteristics during an atomization cycle may allow a choice of appropriate time window for analytical measurements. In addition, since plasma temperatures such as electronic excitation temperature and rotational temperature are important measures of the robustness of plasma sources, these have also been measured for the FAPES source. Electronic excitation temperature is a measure of the ability of the plasma to excite atoms and ions since it is indicative of the population distribution of excited states. Excitation temperature can be determined from a measurement of the relative intensities of spectral transitions between electronic states of an atom or ion "embedded" in the plasma. In this 47 thesis, the excitation temperature of He (I) was measured from the relative intensities of two electronic transitions-He(I)-501.57 nm ('P, —>1S0) and He(I)-492.19 nm (1D2-> 'P,) using Equation 1.4. A rovibronic transition of a molecular species involves two energy levels that may occupy different electronic states, vibrational states and rotational states. Therefore, the temperature, if it is used to characterize this type of transition, should consist of three parts: excitation temperature, vibrational temperature and rotational temperature. The rotational temperature of a molecular species thus refers to the rotational transition part of one rovibronic transition, and its value can be calculated from molecular band emission. In this thesis, we measured the molecular spectra from N 2 + and OH to determine the rotational temperatures for the FAPES device. For N 2 + , the rotational temperature can be calculated by using the R branch emission of its first negative system - B22TU - X 2 S + g . A schematic illustration of rovibronic transitions for N 2 + is presented in Fig. 3.1. Each electronic state (K or K') consists of many vibrational states (v or v'), and each vibrational state includes many rotational states (j or j'). These rovibronic transitions involve not only electronic states (ground electronic state X vs the second excited electronic state B) but also different vibrational states (v' vs v) and rotational states (j' v s j)- By measuring the emission intensities of R branch and plotting the curve ln[Ij/2(j+l)] versus (j+\)(j+2) (L, - emission intensity, j- the rotational quantum number), the rotational temperature T r o tof N 2 + can be derived from the slope of the plot whose value is equal to -2.983/Trol [1]. Similarly, the rotational temperature of OH can be determined by using the Q, branch A 2 Z + - X 2 n transition. The rotational temperature can be derived by plotting the curve ln(lX/A) versus E k (I - emission intensity, X - wavelength (nm), A - transition probability (s' '), Ek - the energy level of the upper state (cm'1)) and calculating the value from the slope of the curve, whose value is -1.4388/Trot [1]. v' = 1 K' (B) v' = n v= 1 K(X) v = 0 Fig. 3.1 Schematic illustration of rovibronic transitions of N 2 +( B 2 S + U - X 2 E + g ) 4 3 2 1 j' = 0 4 3 2 1 j =0 49 3.2 Experimental 3.2.1 Instrumental Set-up The FAPES source and experimental system described in Chapter 2 were used for most of the measurements described in this chapter. A 3600 lines/mm holographic grating was used for the measurement of N 2 + , OH and CO + spectra, and a 1200 lines/mm holographic grating was used for the measurement of Pb(I) spectra. For simultaneous measurement of the background spectrum, an Ocean Optics fiber optic spectrometer (SD2000, Ocean Optics Inc., Dunedin, FL, USA) was used to collect spectra within a spectral window. The light collection system used two plano-convex fused silica lenses (Collimating Lens f = 150 mm, Focusing Lens f = 100 mm) to focus the image into the optical fiber which transferred the light into the fiber optic spectrometer. Data acquisition was performed using an ADC 500 A/D card and OOIBase32 software (Ocean Optics Inc., Dunedin, FL, USA) installed in a Pentium II computer. The helium plasma gas (99.9999%, Union Carbide, Toronto, Ont., Canada) was supplied to the source from a pressurized cylinder. The HGA-500 graphite furnace power supply unit was used to control the gas flow rate and atomization temperature. The graphite center electrodes (diameters 0.9, 1.0, 1.4 and 1.9 mm) were purchased from Ringsdorff-Werke GmbH company (German). The spectral data for the rotational temperature measurements of N 2 + and OH are listed in Table 3.1 and Table 3.2, respectively. Spectroscopic data for the atomic lines used for Pb excitation temperature measurements are provided in Table 3.3 [6-7]. 50 Table 3.1 Spectroscopic data used to calculate the rotational temperature of N 2 + R branch R18 R16 R14 R12 R10 R8 R6 j 18 16 14 12 10 8 6 20+1) 38 34 30 26 22 18 14 Q+m+2) 380 306 240 182 120 90 56 Table 3.2 Spectroscopic data used to calculate the rotational temperature of OH Qi branch Q.2 Q.4 Q |5 Q.6 Q ,9 Q.io X (nm) 307.995 308.328 308.52 308.734 309.534 309.859 A(S ' ) 17 33.7 42.2 50.6 75.8 84.1 E k (cm"1) 32453 32779 32948 33150 33952 34283 Table 3.3 Spectroscopic data used to calculate the excitation temperature of Pb(I) Species Wavelength (nm) g k A k ( 10V ) E k (cm"1) Pb(i) 280.20 43 46329 Pb(i) 283.31 1.8 35287 51 3.2.2 Procedure Lead sample solution was prepared by dissolving Pb(N03) in deionized water and then by diluting the stock solution to reach the required concentration. For Pb atomic emission measurements, 5pL of the Pb solution was pipetted into the graphite tube using an Eppendorf 1 - 10 pi 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 ensure reproducible sample delivery. The graphite furnace temperature program consisted of a drying stage which heated the tube from room temperature to 110 °C using a 10 s ramp then maintained the temperature at 110 °C for 20 s, an ashing stage which included a 10 s ramp to 250 °C and then held this temperature for another 10 s, and an atomization stage which took 1 s to reach 2600 °C and maintained that temperature for 4 s. The rf power was applied to the center electrode at the beginning of the atomization stage. The IPDA detection system was automatically triggered by a TTL signal from the graphite furnace power supply at the beginning of the atomization stage. 3.3 Results and Discussion 3.3.1 Background Emission Spectrum A typical background emission spectrum is shown in Fig. 3.2. This spectrum was collected within a 200-550 nm spectral window using the fiber optic spectrometer. The most prominent features in the background emission spectrum are those of CO +, OH, N 2 , N 2 + and He. Of course, the helium plasma certainly produces the intense He (I) atomic emission. 52 Since the lower electronic states (3S, and 'S0) of these electronic transitions of He (I) such as He(I)-388.87 nm ( 3 P ° U ->3S,) and He(I)-501.57 nm ( , P ° I ^ , S 0 ) are the metastable states g 2 0 0 0 -"5. § 1 5 0 0 -o « 1 0 0 0 -CD C o 'to to E Ul 500H —I 1 1 1 2 5 0 3 0 0 3 5 0 4 0 0 Wave leng th (nm) 4 5 0 5 0 0 5 5 0 Fig. 3.2 Background emission spectrum of the FAPES plasma (40 W, 30 °C, 50 ml/min) with relatively long lifetimes and considerable energies, Penning ionization and excitation processes may occur if these metastable atoms collide with neutrals. He,n (23S, 2'S) + X -> X + * + He (l'S) + e (3.1) (X= Neutrals) The source of OH is the dissociation of H 2 0 which was probably derived from the water adsorbed on the surface of the graphite tube. The primary source of N 2 is diffusion of ambient air from the sample introduction port, and CO comes from the reaction of carbon sputtered from the center electrode with the atmospheric oxygen. Because the excitation 53 energies of the lowest vibronic states of CO + (B2S+) and N 2 + (B2XU+) are about 19.7 eV and 18.8 eV respectively, the existence of CO + and N 2 + emission spectra indicates that a charge transfer process contributes to the production of these species [2, 3]. He2 + which has a recombination energy between 18.3 eV and 20.5 eV can overlap with low-lying electronic states of CO + and N 2 + bands to produce their excited vibronic states - B 2 Z + and B 2 Z + U , respectively. The following reaction may occur. He2+ (X2 Zu+) + X -> X + * + 2He (3.2) (X=CO,N2) As mentioned above, the metastable energies of He and the He2 + potential energy are capable of ionizing analytes with high ionization potentials, and thus the helium plasma inside the FAPES source is a potentially powerful ionizing source. Since the spectral region with wavelength below 290 nm has lower background emission, it was the preferred region used for analyte emission measurements performed in this thesis. 3.3.2 Temporal Emission Measurements During an atomization cycle, as the furnace temperature changes, the plasma may exhibit different spatial and temporal emission characteristics. LeBlanc et al. [4] have already described the spatially inhomogeneous features of the molecular species throughout the plasma volume, and thus this section mainly focuses on the temporal response of the molecular species. 54 Fig. 3.3 is the temporal emission spectrum of N 2 + and He in the spectral region 387 -392 nm. This 3-D graph consists of a series of frames, each of which sequentially integrated the emission intensities for 0.1 second during an atomization cycle. The emission intensities of the R branch of the first negative system of N 2 + (0-0) rotational band basically decreased during the atomization cycle. The reason is that gas thermal expansion inside the FAPES source during the atomization cycle caused decreased number densities of species leading to reduced emission of N 2 + . Another important reason is the loss of the probability of charge transfer reaction with He2+ to CO. During an atomization cycle, a high temperature promoted the production of CO molecules due to the reaction of carbon from graphite tube wall with oxygen molecules entrained from atmosphere. The increased number density of CO may increase the charge transfer reaction probability of CO with He2+ and, at the same time, He Wavelength (nm) Fig. 3.3 Temporal emission spectrum of He and N2+(80 W, 2000 °C, 50 ml/min) 55 decrease the charge transfer reaction probability of N 2 with He 2 + since the lowest vibronic states of CO + are closer to the recombination energy of He 2 + than that of N 2 + and thus the charge transfer reaction between He 2 + and CO is more favorable. The observation that emission intensity of CO + was increased during the atomization cycle will be discussed later. Unlike that of N 2 + ; the temporal emission behavior of He(I)-388.86 nm during the atomization cycle was complicated. Because emission intensity from a species is a function of excitation temperature and the total population of the emitting species, the helium atomic emission decreased as a result of its decreasing population caused by gas thermal expansion, and then it increased due to the higher excitation temperature (The enhanced excitation temperature is shown in Fig. 3.6). After more helium atoms were electronically excited to the upper electronic states, electronic transitions from the upper electronic state (3P°) to the lower electronic state He (2 3Si) gave off more light at the wavelength of 388.87 nm, leading to increased atomic emission intensity. Fig. 3.4 shows the temporal emission spectrum of OH. The molecular emission intensities of the OH band are determined by the population density of OH and its rotational temperature. The population density of OH increased first due to the dissociation of H 2 0 vapor molecules entrained from atmosphere under high temperature, and then it decreased as a result of gas thermal expansion and the dissociation of OH in the plasma. Thus, the population density of OH led to a variation of molecular emission intensity of OH band during the atomization cycle. Furthermore, an increased rotational temperature (see Fig. 3.6) also complicated the OH temporal behavior. 56 Fig. 3.4 Temporal emission spectrum of OH (2000 °C, 80 W, 50 ml/min) The temporal spectrum of CO + during the atomization cycle is illustrated in Fig. 3.5. It consists of 13 frames, each with an integration time of 0.1 s, separated by 0.5 s. As previously discussed, the formation of CO during the atomization cycle increased the population density of CO, and thus more CO molecules could be promoted to the excited vibronic states (B2 Z+) by charge transfer reaction and Penning ionization processes. As a result, the emission intensities of the CO + band increased and reached a maximum at about 3.6 s into the atomization cycle. 57 3000-, 226 228 230 232 234 236 Wavelength (nm) Fig. 3.5 Temporal emission spectrum of CO + (2000 °C, 80 W, 50 ml/min) 3.3.3 Plasma Temperature Measurements 3.3.3.1 Time-resolved Temperature Measurements of Background Species The time-resolved temperature profiles of He, OH andN2+ during an atomization cycle are shown in Fig. 3.6. The excitation temperature of He and the rotational temperatures of OH and N 2 + gradually increased during the atomization cycle, reaching peak values of 3650K, 1500K and 1520K, respectively. These plasma temperatures were about 200-400 K higher at the peak position than at the beginning of the atomization cycle. In addition to the contribution of thermionic electron emission from the graphite walls to the temperature enhancements, gas thermal expansion and thermal heating during the atomization cycle may also play a role in the temperature enhancements. Gas thermal expansion inside graphite tube reduced species densities which thus resulted in a decrease in the collision frequency, leading to a greater production of energetic particles. These energetic particles led to more 58 0) 03 i CD CL E -*—' c o o 4 0 0 0 - j 5 3 © 3 8 0 0 H 3600-3400-3200-3000 He (left axis) N 2 + (right axis) O H (right axis) (40 W, 50 ml/min, 2000 C ) ^ A—7!r-n 1 1 1 r 2 3 4 5 6 Time (second) r-2000 jj o h1800 I o r- 1600 i . 1400 3 3-1200 1 - 1 0 0 0 5 - 8 0 0 ^ Fig. 3.6 Temporally resolved temperatures of background species effective collisions with background species to distribute kinetic energy in the plasma. 3.3.3.2 Excitation Temperature of Pb In order to more fully characterize the FAPES source, we also used Pb, an introduced species, to measure the temporal emission profiles and electronic excitation temperature. The reason we chose Pb is that it has two emission lines close in wavelength but widely separated in energy. We began with a study of temporally resolved atomic emission and excitation temperature profiles of Pb during an atomization cycle which are shown in Fig. 3.7. Starting from the beginning of the atomization cycle, the Pb excitation temperature increased and reached a constant value at about 1 s. 59 The effect of rf power on the Pb excitation temperature is presented in Fig. 3.8. Within the normal operating range of rf power (with zero or little reflected power), an increase in rf power enhanced Pb excitation temperature, and therefore enhanced the plasma robustness. Pb (l)-280.2 nm 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 Atomization time (Second) Fig. 3.7 Temporally resolved atomic emission and excitation temperature profiles of Pb (50 ml/min, 80 W, 2600 °C) 60 7000 | 5 0 0 0 4 r | 6000H 5 Q. C o 'ra 4000 o X LU 3000-1 20 30 40 50 RF power (W) 60 70 80 Fig. 3.8 rf power effect on the excitation temperature of Pb (50 ml/min, 2600 °C) 3.3.4 Effect of Operating Conditions on Plasma Temperatures The excitation temperature of He and the rotational temperatures of OH and N 2 + are often used to characterize He plasmas. These temperatures may be affected by plasma operating conditions. Thus, it is useful to understand the relationship between the plasma temperatures and operating conditions such as rf power, gas flow rate, atomization temperature, and center electrode size. Plasma temperatures in an analytical source can indicate the robustness of the source. Although the excitation temperature and the rotational temperature are different from the ionization temperature, an increase in excitation temperature and rotational temperature is generally accompanied by an increase in ionization temperature, and hence, the degree of ionization. Therefore, operating parameters that yield an increase in plasma temperatures also probably provide an enhancement of plasma ionization capability. 61 3.3.4.1 rf power Application of higher rf powers is one approach to increase plasma temperatures. Fig. 3.9 shows the effect of rf power on the rotational temperatures of OH and N 2 + as well as on the excitation temperature of He. When the forward power was increased above 80 W, a further increase in rf power didn't cause an increase in these temperatures. Two reasons might be responsible for the reduction of plasma temperatures, i.e. extensive thermal emission of electrons from the heated tube wall and the reflected rf power from the FAPES source. Consequently, less forward power was actually applied to the plasma, and more power was consumed on the matching network or reflected, and as a result, the plasma temperatures were decreased. to 0) CL E < c o •*—» CC •*—' "o ILU 4 0 0 0 -3 8 0 0 -3 6 0 0 -3 4 0 0 ' 3 2 0 0 ' 3 0 0 0 ' A He (left axis) © N 2 + (right axis) IS — O H (right axis) (50 ml/min, no atomization ) 6 0 R F power (W) 331 o 03 Fig. 3.9 rf power effect on plasma temperatures of He, OH and N 2 + 3.3.4.2 Gas Flow Rate 62 Plasma temperatures can be influenced by the gas flow rate since helium acts as both analyte carrier gas and plasma support gas. The effects of gas flow rate on the rotational temperatures of OH and N 2 + as well as on the excitation temperature of He are presented in Fig. 3.10 and Fig. 3.11. These results show that these plasma temperatures have maxima at © 50 100 150 200 250 300 Gas flow rate (ml/min) Fig. 3.10 Gas flow effect on rotational temperatures of OH and N 2 + 3800-] 3750-| 3650' CD E 3600-1 CD l _ 3550-1 3500-1 He(l)(40 W, 2000 C ) 50 100 150 200 Gas flow rate (ml/min) 250 300 Fig. 3.11 Gas flow effect on the excitation temperature of He 63 a gas flow rate of 50-100 ml/min. These phenomena suggest that there is an optimum gas flow rate and higher gas flow rates cool the plasma, leading to decreased plasma temperatures. 3.3.4.3 Atomization Temperature Fig. 3.12 and Fig. 3.13 displays the effect of atomization temperature on the rotational temperatures of OH and N 2 + as well as on the excitation temperature of He. There was an enhancement of about 150 - 250 K for the plasma temperatures of these species. The reason is probably due to the higher atomization temperature providing the plasma with more thermal energy and further resulting in an increase in plasma temperatures. As atomization temperature was increased to some extent, thermionic electrons emitted from the heated graphite tube absorbed extra thermal energy and lowered average kinetic energy, causing a detrimental effect on plasma temperatures. 1600 1800 2000 2200 2400 2600 0 Atomization temperature ( C ) Fig. 3.12 Atomization temperature effect on rotational temperatures of OH and N 2 64 3900 a s s e -s' ~ 3800-k _ ffl 3750-1 CD O L I 3700-3650-3600-ar i 1 600 He(l)(40 W, 50 ml/min) T T 1800 2000 1 2200 Atomization temperature ( C ) 2400 2600 Fig. 3.13 Atomization temperature effect on the excitation temperature of He On the other hand, these two graphs also demonstrate the atomization temperatures at which these three plasma temperatures reached maxima were in the increasing order. This relationship may result from their different excitation and ionization characteristics as well as physical properties. Although OH and N 2 + are both molecular species, OH has a dissociation energy of 4.39 eV while the lowest vibrational state of N 2 + has an excitation energy of about 18.7 eV. As a result, they exhibited different characteristics. Since helium atomic emission involves an excited electronic state whose metastable energy is around 23 eV, its excitation temperature increase is a little greater than those of OH and N 2 + . 3.3.4.4 Center Electrode Size Since the rf electric field is dependent on the spacing between the graphite tube and center electrode, changing the center electrode size may affect the plasma characteristics of 65 the FAPES source. The effects of center electrode size on the plasma temperatures and the helium atomic emission intensity are shown in Fig. 3.14 and Fig. 3.15. Increasing the electrode size reduced the rotational temperatures of OH and N 2 + as well as the excitation temperature of He. These results suggest that an increase in the center electrode diameter lowers the robustness of the FAPES source. The lower helium atomic emission intensity also supports this premise. In the FAPES source, the two electrodes are the center electrode and the graphite tube respectively, and they are of different surface areas. According to the scaling law [5], the relationship between electrode voltage drops V, and V 2 and surface areas of electrodes A, and A 2 is as follows. v, T A V v2 l 2 V A i J (3.3) When the center electrode diameter is increased under constant electrode length, its surface area is increased. Since increasing the electrode size actually didn't cause more rf power to be applied to the plasma between these two electrodes, the voltage drop was distributed more evenly between the two electrodes. Thus, the electric field strength of the negative glow that is close to the center electrode became lower so that electrons and ions might gain less energy from the weakening electric field. Consequently, increasing the electrode size didn't help improve plasma temperatures. Another possible reason is due to the reduction of the total voltage drop between the center electrode and graphite tube. According to equation 1.1, the breakdown voltage decreases with decreasing spacing between two electrodes. When the center electrode size of the FAPES source was increased, the separation between center electrode and graphite rube 66 1500-1 1.0 1.2 1.4 1.6 1.8 Electrode diameter (mm) Fig. 3.14 Center electrode size effect on rotational temperatures of OH and N 2 + (80 W, 2600 °C, 50 ml/min) Electrode diameter (mm) | Fig. 3.15 Center electrode size effect on helium emission intensity and excitation temperature (80 W, 2600 °C, 50 ml/min) 67 was decreased, leading to a decreased breakdown voltage. After breakdown, the total voltage correspondingly decreased, and as a result, the electric field strength was reduced. 3.4 Summary Plasma characteristics in a newly designed FAPES source have been studied. Because of different excitation and ionization mechanisms, the species He, N 2 + , OH and CO + exhibit different temporal emission behaviors during the atomization cycle. The measurement of the time-resolved electronic temperature (for He) and rotational temperatures (OH and N 2 + ) show that these plasma temperatures are enhanced during an atomization cycle since some thermal energy might boost the plasma temperatures. The excitation temperature of helium and rotational temperatures of N 2 + and OH are a little bit higher than or consistent with the previously reported plasma temperatures [4, 8-15]. Increasing rf power up to 80 W provides more energy to the FAPES source and enhances plasma temperatures, which causes enhanced excitation and ionization. Optimizing the gas flow rate and atomization temperature can improve plasma temperatures to some extent, and potentially, these effects may be helpful to the improvement of plasma ionization capability. The investigation of the center electrode size effect on the FAPES source has demonstrated that increasing electrode size resulted in the reduction of the power density, and consequently less voltage was dropped across the plasma sheath where the greatest excitation and ionization occurs, leading to a reduction of plasma robustness. 68 3.5 References 1. P.W.J.M. Boumans, Inductively Coupled Plasma Emission Spectroscopy Part 2. p. 368-372 (John Wiley & Sons, USA, 1987). 2. M. Endoh, M. Tsuji, and Y. Nishimura, / . Chem. Phys. 79: p. 5368 (1983). 3. B. Lescop, M. Ben Arfa, G. Le Coz, M. Cherid, G. Sinou, G. Fanjoux, A. Le Nadan and F. Tuffin, J. Phys. II France 7: p. 1543 (1997). 4. C.W. LeBlanc and M.W. Blades, Spectrochim. Acta 50B: p. 1395 (1995). 5. J.R. Roth, Industrial Plasma Engineering: Volume I Principles: p. 438 (Institute of Physics Publishing, USA, 1995). 6. J. Lotrian, Y. Guern, J. Cariou and A. Johannin-Gilles, J. Quant. Spectrosc. Radiat. Transfer 21: p. 143 (1979). 7. C.H. Corliss and W.R. Bolzman, Experimental Transition Probabilities for Spectral Lines of Seventy Elements, NBS Monograph #53 (US Dept. of Commerce, 1962). 8. R.E. Sturgeon, S.N. Willie and V.T. Luong, Spectrochim. Acta 46B: p. 1021 (1991). 9. R.E. Sturgeon, S.N. Willie, V. Luong and J.G. Dunn, Appl. Spectrosc. 45: p. 1413 (1991). 10. T.D. Hettipathirana and M.W. Blades, Spectrochim. Acta 47B(4): p. 493 (1992). 11. R.E. Sturgeon and S.N. Willie, J. Anal. At. Spectrom. 7: p. 339 (1992). 12. T.D. Hettipathirana and M.W. Blades, / . Anal. At. Spectrom. 8: p. 955 (1993). 13. S. Imai and R.E. Sturgeon, J. Anal. At. Spectrom. 9(4): p. 493 (1994). 14. S. Imai and R.E. Sturgeon, J. Anal. At. Spectrom. 9: p. 765 (1994). 15. V. Pavski, C L . Chakrabarti and R.E. Sturgeon, / . Anal. At. Spectrom. 9: p. 1399 (1994). 69 Chapter 4 TEMPORAL AND OPERATING CHARACTERISTICS OF ANALYTE EMISSION 4.1 Introduction The temporal emission profiles and plasma temperatures of He, OH and N 2 + in the FAPES source were discussed in Chapter 3. The results show that the FAPES plasma exhibits different temporal characteristics during an atomization cycle. In this chapter, the use of temporal profiles to study analyte excitation and ionization characteristics in the FAPES source will be described. Analyte temporal behavior in a conventional graphite furnace has been widely discussed. However, the presence of the center electrode could provide a potential site for condensation and secondary release of analyte species, and the atmospheric pressure rf helium plasma could become an additional source of thermal energy. As a result, analyte temporal characteristics in the FAPES source need to be investigated. In order to further develop FAPES as an ion source for elemental mass spectrometry, knowledge of the temporal behavior of analyte ionization could be very useful both for improving our understanding of the ionization characteristics of the source and for insight into choosing and optimizing the mass analyzer. This chapter describes the temporal behavior and operating characteristics of analyte atomic emission and ionic emission. 4.2 Experimental A complete description of the experimental system used for analyte atomic and ionic emission measurements was given in Chapter 2. With this experimental set-up, it was possible to simultaneously collect spectrally and temporally resolved emission data. A 70 feature of the HGA-500 based FAPES source compared to previously developed FAPES sources in our lab is its faster heating rate which allows the analyte atomic population to build up rapidly, creating a high atomic concentration in the atomizer volume and thus increasing analytical sensitivity. During the rapid heating step, the HGA-500 unit controls the temperature which was monitored using an optical feedback system to maximize the heating rate and minimize possible overshoot. The elements used for these investigations were chosen to satisfy several requirements. First, it was necessary to choose elements whose appearance temperatures in the FAPES source were within the operating range of the instrumentation. Second, only those elements that possessed both atomic and ionic emission lines with wavelengths that fell within a single spectral window of the detector were useable. Third, the selected elements should have typical ionization energies. Finally, calculations of the degree of ionization required reliable spectral data such as the partition function of the species, Q(T), the statistical weight of excited state of the species, g, the Einstein transition probability, A, and the excitation energies of the atomic and ionic lines Ei0„ and E a t o m . Based on the above considerations, magnesium, cadmium, iron, zinc and chromium were chosen as the spectrometric species. Table 4.1 lists the relevant fundamental spectral data for these elements [1-4]. The temperatures used in the equations for these measurements were derived from Pb(I) excitation temperatures which were discussed in Chapter 3. The assumption made is that the atomic excitation temperature is approximately equal to the ionic excitation temperature. While there is no evidence that this assumption is valid, a close look at Equations 1.5 and 1.6 shows that the degree of ionization has only a weak dependence on 71 Table 4.1 Data for atomic and ionic lines of selected elements Species Wavelength gk A E k IP Appearance temperature (nm) (x 108 /s'1) (cm'1) (eV) (K) Mg(I) 285.2 3 4.95 35021 7.64 1530 Mg(II) 279.6 4 2.68 35761 Cd(I) 228.8 3 5.3 43692 8.99 730 Cd(II) 226.5 2 3.0 44136 Fe(I) 252.3 9 2.8 39626 7.90 1580 Fe(II) 259.9 10 2.2 38459 Zn(I) 213.9 gA = 18.91 46745 9.39 1050 Zn(II) 206.2 gA = 6.575 48481 Cr(I) 278.1 gA = 9.8 35951 6.76 1740 Cr(II) 284.3 gA = 6.4 35161 the temperature used to calculate it from measured ion/atom emission intensity ratios. For illustration, Fig. 4.1 is a plot of the temperature dependence of degree of ionization for Mg, Cd, Fe, Zn and Cr for a "typical" measured ion/atom emission intensity ratio. From the plot, it can be seen that a 2000 K change in temperature (from 4000 K to 6000 K) only changes the derived degree of ionization by about 5%. The sample solutions were prepared by dissolving their respective salts (MgSCu, CdCb, ZnCb, and FeS04) in 1% HNO3 and then by diluting the stock solutions. Cr sample solution was obtained from commercial stock solution (SPEX chemical, 1000 pg/ml in 2% HNO3). For analyte emission measurements, 5 pL of solution was pipetted into the graphite tube. The temperature program, consisting of a drying stage, an ashing stage, and an 72 Fig. 4.1 Temperature dependence of degree of ionization for Mg, Cd, Fe, Zn, and Cr at an actual line emission intensity ratio (peak emissions at 80 W) atomization stage, was initiated. During the drying stage, the tube was heated from room temperature to 110 °C using a 10 s ramp, and maintained at 110 °C for 20 s. The ashing stage used a 10 s ramp to reach a temperature of 250 °C which was held for another 10 s. The atomization stage used a Is ramp to reach 2600 °C (or other values as noted in the text.) which was maintained for 4 s. The rf power was applied to the electrode 5 s before the atomization stage and the data collection system was triggered at the beginning of the atomization stage. Table 4.2 lists the experimental conditions for the analyte atomic and ionic emission measurements. 73 Table 4.2 Experimental parameters for atomic and ionic emission measurements Element Sample amount (ng) Atomization temperature (K) Mg 0.05 2600 Cd 50 2600 Fe 50 2600 Zn 50 2600 Cr 50 2600 4.3 Results and Discussion 4.3.1 Temporal Emission Spectra and Profiles Analyte atomic and ionic emission profiles were simultaneously collected over an approximately 8 nm spectral window using the IPDA detection system. Fig. 4.2 shows temporal emission spectra of Mg. The three dimensional graph consists of a series of spectra, in which each spectrum is the integrated emission intensity using a 0.2 s integration time. The interval between spectra is 0.2 s. The temporal emission behaviors for Mg(I), 285.2 nm, Mg(II), 279.6 nm and Mg(II), 280.3 nm as well as the background species ( there is a OH band around 285 nm ) are identified in the figure. The first few frames contain spectral information from the plasma background only. As the atomization cycle proceeded to about 0.6 second, Mg atomic and ionic emission began to appear. During the appearance of Mg emission, background emission from OH is depressed, reappearing in the last 10 frames at the end of the atomization cycle. The depression in OH emission intensity could arise either 74 Mgl-285.2 nm ~"i r r f r-278 280 282 284 286 Wavelength (nm) Fig. 4.2 Temporal emission spectra of Mg at 80 W from a reduction in the excitation temperature during the atomization cycle or from a reduction in the density of OH radical during the application of the temperature ramp. It is most certainly the latter that causes the reduction in OH emission intensity since the evidence in the previous chapter of this study suggests that the excitation temperature remains relatively constant or is enhanced a little during the atomization cycle. Figure 4.3 shows the temporal emission sequence for the atomization of Cd. The Cd(I), 228.8 nm and Cd(II), 226.5 nm, emission lines are clearly seen in the figure. Because of the lower appearance temperature, Cd emission appears earlier in time than that for Mg. The emission behavior for CO + , which has bands at 222, 230, 233 and 235 nm, is similar to that for OH in that the emission is depressed during the application of the temperature ramp. 75 Using the temporal spectra of Mg (shown in Fig. 4.2), we can obtain the temporal profiles of analyte atomic and ionic emission as well as the degree of ionization. This is provided in Fig. 4.4. The double peak for Mg atomic emission is clearly observed in the temporal profile. For this experiment, MgS04 was used to prepare the Mg solution. Fig. 4.3 Temporal emission spectra of Cd at 80 W Normally, during the atomization ramp time period, MgS04 is first decomposed to MgO and then reduced to Mg. MgS04 (s) = MgO (s) + S03 (g) (4.1) MgO (s) + C (s) = Mg (g) + CO (g) (4.2) However, a variety of potential mechanisms can be involved in atomization and excitation of Mg leading to the appearance of double peaks. This experiment was repeated at different atomization temperatures, and the double peaks weren't found to appear at atomization temperatures under 2200 °C (Refer to Fig. 4.5). There are several possibilities that could 76 Fig. 4.4 Temporal profiles of atomic emission, ionic emission and degree of ionization for Mg at 80 W o 5000-| x 4000' 3 O 3000H CO I 2000-1 c o 1000-1 w w E ILU 0 A 1500 C 1800 C 2000 C 2200 C n 1 r 2 3 4 Time (Sec.) Fig. 4.5 Temporal profiles of atomic emission for Mg at different atomization temperatures (80 W) 77 cause this phenomenon. The first is that the double peaks result from sequential atomization of the furnace wall and center electrode respectively. Because the electrode is passively heated radiatively and through convection by heat loss from the furnace wall, its temperature lags that of the furnace wall. At early stages in the atomization cycle, the electrode can act as a condensation site for atomized analyte that can subsequently re-atomize when the electrode reaches a higher temperature. Hettipathirana has observed this phenomenon for Pb atomic emission [5]. An additional cause for double peaks might be the formation of intercalation compounds of Mg with graphite. The loose layered structure of graphite allows many molecules and ions, especially alkali and alkaline earth elements, to penetrate the layers to form intercalation or lamellar compounds [4]. As a result, during early stages of the heating cycle, surface Mg is atomized and at latter stages, the intercalated Mg is released. A third possibility is the formation of a metal carbide during the early stages of the atomization cycle. At lower temperature, heating a mixture of magnesium oxide and atomic carbon results in the reduction of the oxide to magnesium metal as described in equation 4.2. However, an additional reaction can occur at higher temperature because alkaline earth metals can form carbides when their oxides are heated in the presence of carbon. MO + 3C = MC 2 + CO ( M = Mg, Ca, Sr, Ba) (4.3) The reaction is endothermic, and a higher temperature is required for this reaction to occur. A temperature of 2200 °C is required for CaO. No temperature data is available for the formation of MgC2 by the reduction of MgO, but the temperature value is possibly around 2200 °C since the heat of carbide formation of Mg is approximately close to that of Ca [4]. The production of atomized Mg from reaction (Equation 4.2) is greatly reduced by the reaction (Equation 4.3) so that only a small emission peak is seen at its "normal" temporal 78 position. Thereafter, MgC2 can be further converted to Mg2C3 at higher temperatures. At the latter stages of the atomization cycle, Mg2C3 can thermally dissociate as follows: r , . Vaporization Dissociation . A . Mg2C3(s) Mg2C3(g) 2Mg(g)+ 3C(g) (4.4) > > For Mg, these three mechanisms are possible, but the formation of carbide seems most reasonable because it is consistent with the fact that there were no double peaks under 2200 °C although detailed data on the passive heating rate of the center electrode is required to rule out the contribution of condensation completely. In addition, Mg may be produced through direct vaporization of MgO and subsequent dissociation to form Mg and O2 causing an early appearance of the atomic and ionic emission. In Figure 4.4, the dotted line is the temporal curve for the degree of ionization for Mg. It appears that, between 1 s and 2s, the degree of ionization is changing rapidly and after that period of time it is stable at about 60%. The initial change is most likely due to disequilibrium or spatial segregation of atomic and ionic emission during the initial application of the temperature ramp. Fig. 4.6-4.9 show data on temporal behavior of the atomic and ionic emission lines and derived degree of ionization for Cd, Fe, Zn and Cr. Because the appearance temperature of Cd is the lowest of this group of elements, the atomic and ionic emission of Cd occurred earliest and the emission transient lasted only one second. Its degree of ionization increased sharply, reaching a maximum value of about 50% at the peak of atomic emission and ionic emission. Although the ionization energy and appearance temperature of Fe are almost the same as those of magnesium, the ionic emission had only a single peak, possibly because the interstitial carbide of Fe requires a higher temperature for formation. The degree of ionization 79 for Fe shows the same trend as that for Mg, increasing initially and reaching a maximum value of between 70 and 80%. The data after 4 s are not considered accurate since the emission intensities are very weak. The degree of ionization for Zn exhibits an initial spike and levels off at about 20%. For Cr there is some variation in the degree of ionization depending on the time during the atomization cycle. Interestingly, for both Cr and Fe there is a displacement in the peak maxima for the ionic and atomic lines. It is possible that this is due to spatial segregation. Compared with Cr and Fe, Cd and Zn have lower appearance temperatures and higher ionization energies; their degrees of ionization are lower and the time needed for reaching the maximum ionic emission is shorter. Thus, we may conclude that the appearance temperature of an element generally determines the temporal position of the maximum ionic emission. 1 0 0 0 0 - ! Cd (l)-228.8 nm C d (II)- 226.5 nm Degree of ionization 3 - 3 0 O 0 T o 0 1 2 3 Atomization Time (Sec.) 4 5 Fi; g. 4.6 Temporal profiles of atomic emission, ionic emission and degree of ionization for Cd at 80 W 80 2 0 0 0 - — Fe (l)-252.3 nm r- 1 0 0 CD -— - Fe (II) -259.9 nm o (Counts/pi) 1 5 0 0 -1 0 0 0 -Degree of ionization nization Degree (% o o o o 00 to CM 1 I I I ensity 5 0 0 -/ \ '* / *" *•*• V * I / - ' A ' * / •* / \ J / ' ' \ nization Degree (% o o o o 00 to CM 1 I I I c 0 - I I I I 0 1 2 3 4 5 Atomization Time (Sec.) Fig. 4.7 Temporal profiles of atomic emission, ionic emission and degree of ionization for Fe at 80 W Atomization Time (Sec.) Fig. 4.8 Temporal profiles of atomic emission, ionic emission and degree of ionization for Zn at 80 W 81 5 0 0 -— Cr( l) Cr( l l ) — Degree of ionization 1 - 1 0 0 a> X 4 0 0 - / I I - 8 0 nts/pi 3 0 0 -\ l \ /•' / ' i i i a - eocS —* (Cou 2 0 0 -i V i i i CD CD - 4 0 ° ntensity 1 0 0 -i *. i \ i "***. 1 ionization o CM c i i — I i I > 1 - 1 D 1 2 3 4 5 Time (Second) Fig. 4.9 Temporal profiles of atomic emission, ionic emission and degree of ionization for Cr at 80 W 4.3.2 Atomization Temperature Effects 4.3.2.1 Atomic and Ionic Emission Intensities Since the emission intensity from a species in a plasma source is a linear function of the number density of the emitting species, a good way to enhance emission intensity is to increase the number density. For an optimization of atom and ion number densities in the FAPES source, the atomization temperature must be high enough to ensure that complete atomization is achieved, however, excessively high temperature will increase the diffusion rate and decrease the residence time, leading to a net reduction of emission intensity. Therefore, an appropriate atomization temperature is desired for an enhancement of atomic and ionic emission intensities. 82 Fig. 4.10 shows the effect of atomization temperature on atomic and ionic emission intensities of Cd. An increase in the atomization temperature increased atomic and ionic emission of Cd up to around 2000 °C and then emission intensities decreased. This is consistent with Sturgeon's observation that Cd has stronger atomic emission intensity at 2100 K compared with that at other temperatures [6]. For analyte atomization, the optimum atomization temperature should meet the requirement of producing a maximum atom vapor concentration for better signal measurement. Since helium has high thermal conductivity, this extra contribution from thermal conduction of the helium plasma heating inside the graphite tube causes the optimum atomization temperature for an analyte in FAPES source to be less than that in GFAAS [7]. However, the temperature distribution is not uniform along the length of the graphite tube. For example, the temperature at both ends of the graphite tube is 1100-1200 K lower than that at tube center (when heated to 2700 K) [8]. In addition, the temperature sensor, positioned to the center of the graphite tube, regulates the nominal temperature and the emission intensity is longitudinally integrated, and thus the optimum temperature for analyte to achieve maximum longitudinally-integrated intensity will be much higher than its appearance temperature. As a result, although the Cd appearance temperature is 730 K, a higher atomization temperature is required so that Cd atomization can be maximized at any longitudinal point in the graphite tube. On the other hand, the optimum atomization temperature is limited by the rate of diffusion, which causes analyte loss. The rate of diffusion of analyte is proportional to Tn(n=l.5-2.0) [4], The higher the atomization temperature, the greater the analyte loss. Therefore, there is an optimum atomization temperature for emission intensities of Cd. It was also observed that increasing the 83 >< C L C o in w E 0 o E o 3 8000 H 6000 •E 4000 H 2000 H O H .A . j O - - . - K - Cd(l) r- 300 - ® - Cd(ll) h 250 - 200 § - 150 f CO - 100 ^ o c - 50 5f -3" - o £ 1600 1800 2000 2200 2400 2600 Temperature ( C) Fig. 4.10 Effect of atomization temperature on Cd atomic and ionic emission intensities atomization temperature did not enhance ionic emission. The reason is that Cd has a high ionization energy (8.99 eV), and, as a result, increasing the atomization temperature does not produce more energetic particles required for the production of Cd ions by collision processes. Since Cr has a higher appearance temperature (1740 K) than Cd, the optimum atomization temperature of Cr in this longitudinally heated FAPES source should be higher than that of Cd. Thus, it's reasonable for atomic and ionic emission intensities of Cr to consistently increase with increasing temperature (See Fig. 4.11). Although Cr has no optimum atomization temperature available in the GFAAS for reference, it has been reported that the maximum intensity was obtained at a vaporization temperature of 3000 °C for Cr [9] in an ETV-ICP source. Unlike ionic emission for Cd, increasing atomization temperature enhanced the amplitude of ionic emission for Cr. The lower ionization energy of Cr (6.76 eV) probably contributes to its different emission behavior. As atomization temperature 84 increased, thermionic electron emission from the graphite tube increased, leading to a greater electron number density. These electrons might gain moderate energy in the rf electric field resulting in increased analyte excitation and ionization. Therefore, the ionic emission of Cr was enhanced. 500 - 1 - H - Cr(l) CD X -•€>- Cr(ll) CL i2 400-o o. 300->, GO inter 200- ,JSr'" o "GO GO 100- H . — -E3 ' £ LU o -— - — H — • " . . a — ' " ~ * I 1600 I 1800 I 2000 Temperature I 2200 o C) I 2400 I 2600 Fig. 4.11 Effect of atomization temperature on Cr atomic and ionic emission intensities 4.3.2.2 Temporal Ionic Emission Profile Since our long-term intent is to develop the FAPES source into an ion source for elemental mass spectrometry, the information about analyte ionization characteristics is very important for ion extraction and ion sampling from the FAPES source. Because the temporal ionic emission profile is a good reflection of analyte ionization characteristics during the atomization cycle, the temporal ion emission profile can be used as a diagnostic tool to investigate how analyte ionization characteristics can be affected by operating conditions such as atomization temperature and rf power. 85 Fig. 4.12 shows the effect of atomization temperature on Mg ion emission. During atomization cycles with different atomization temperatures, Mg ionic emission intensity changed with atomization time and finally reached a peak value. This phenomenon is consistent with that of atomic emission (Refer to Fig. 4.5). On the other hand, Mg integrated ionic emission intensity increased with increasing atomization temperature. For example, Mg had greater integrated ion emission intensity at 2200 °C than at 2000 °C. In addition, the appearance time at which ion emission occurred became shorter and shorter with increasing atomization temperature. This was possibly due to an incomplete atomization process occurring at lower atomization temperatures, leading to a delayed, and hence, weak ionic emission. However, when a sufficiently high temperature was applied for analyte atomization, the analyte atom population density was increased and built up very shortly, and correspondingly the analyte ion population density appeared earlier and became greater, resulting in stronger ionic emission intensity. cu 3 5 0 - i X Q. 3 0 0 -In c o 2 5 0 -O >- 2 0 0 -'co c CD 1 5 0 -c 1 0 0 -c o CO GO 5 0 -E LU 0 -IV n / . — / 1500 C — 2000 C i 1 r 2 3 4 Time (Sec.) 1800 C 2200 °C T 5 1 6 Fig. 4.12 Mg ionic temporal emission profiles at different atomization temperatures 86 The temporal behavior of ionic emission in the FAPES source suggests that the ion production is not constant during the atomization cycle, and analyte ionization process may proceed for a few seconds. The time at which analyte ion production has the maximum yield can be derived from the appearance time of the analyte ionic emission peak. 4.3.3 rf Power Effects 4.3.3.1 Temporal Ionic Emission Profile Since rf power may potentially affect analyte ionization characteristics, the effect of rf power on temporal ionic emission is also necessary for a thorough understanding of plasma ionization characteristics. Fig. 4.13 plots the temporal ion emission characteristics of Mg at power levels ranging from 20 W to 60 W. Clearly, ionic emission intensity was enhanced by increasing the rf power level. Additionally, the time of the ionic emission peak at different rf power levels only shifted a little. Unlike atomization temperature, rf power greatly increased ion production and slightly affected the temporal emission profile. As far as the development of the FAPES source as an ion source is concerned, this feature is desirable since it may be helpful to the sampling of analyte ions at an estimable time. 4.3.3.2 Ionic Emission Intensities 87 0 1 2 3 4 5 Time (Sec.) Fig. 4.13 Mg ionic temporal emission profiles at different rf powers (Best fit curves) Increased rf power is often used to enhance analyte ionization. For a potential ion source, the application of high rf power may produce sufficient ion density. However, a continuous increase in rf power level does not necessarily cause an enhancement of analyte ionization. Fig. 4.14 shows the effect of rf power on ionic emission intensities of Fe and Zn. Within the operating rf power range from 25 W to 125 W, the ionic emission intensity of Zn kept increasing, but the ionic emission intensity of Fe reached maximum at 80 W, and then decreased with increasing rf power level. The difference in behavior of Fe and Zn is no doubt due to the different ionization energies and other physical characteristics. In Chapter 5, we'll discuss this phenomenon in detail. 88 1000-] .1 800-1 Q. g. 600 H c o "co 400 CO 'E co •2 200-1 o cu OH 40 ~ i 1 — 60 80 rf power (W) i - 300 100 & - 250 o' o' - 200 CD 3 CO CO - 150 o' 3 "o" - 100 o c 3 5T - 50 "5" x' CD '—' h o 120 Fig. 4.14 Effect of rf power on ionic emission intensities of Fe and Zn 4.3.4 Gas Flow Rate Effects 4.3.4.1 Atomic and Ionic Emission Intensities This FAPES source incorporates an external gas stream and an internal gas stream. Atmospheric pressure helium is used for both gas streams. The external gas flow, which purges the contact cylinders and the outside of the graphite tube, is maintained at a fixed rate of 900 ml/min. The internal gas stream acts as not only a carrier gas as in GFAAS but also as the plasma support gas. In this thesis, the gas flow rate refers to internal gas flow rate. Furthermore, once the gas flow rate was set, it was kept constant through the entire atomization cycle except for stop-flow mode. Since the gas flow may influence several processes such as those of transport, atomization, excitation, and ionization, it's very important to investigate the effects of gas flow on analyte emission intensities. Fig. 4.15 shows the effect of gas flow on Fe emission. It can be seen that the ionic emission for Fe reached a maximum at a much lower gas flow rate 89 than the atomic emission. For both atomic and ionic emission, lower gas flow was insufficient to maintain the normal operation of helium plasma, and as a result, analytical signals suffered a loss from unstable analyte excitation and ionization processes. When higher gas flow made analyte excitation stable, atomic and ionic emission intensities for Fe reached the maximum. Further increases in gas flow cool the plasma, leading to a reduced excitation temperature. Also, a high gas flow does not allow enough time for complete condensation of Fe particles on the center electrode nor is there sufficient residence time for optimum Fe ionization in the plasma. A similar situation is observed for Cr (Fig. 4.16). However, the optimum gas flow rate for Cr ionic emission is higher than that for Fe ionic emission. The reason is that compared with Fe, Cr has lower ionization energy. At higher gas flow rates, decreased plasma temperatures caused by plasma cooling affect emission intensities for Cr less than that for Fe. 1200-1 5. 1000- j ® - B - Fe(l) Fe(ll) c 3 o o 800 H E LU 'co co 200 H o - L 50 100 150 200 Gas flow rate (ml/min) 250 300 Fig. 4.15 Effect of gas flow rate on Fe atomic and ionic emission intensities 90 600-i \jL 500-1 I § 400-| | M 300 - | o CO . • | 100 -J LU OH 50 -a—H-I I I 100 150 200 Gas flow rate (ml/min) Cr(l) - Cr(ll) 250 300 Fig. 4.16 Effect of gas flow rate on Cr atomic and ionic emission intensities 4.3.4.2 Comparison of stop-flow mode and continuous-flow mode From the above discussion, we find that sufficient gas flow is required for complete atomization, and, at the same time, the gas flow rate is also limited by its effect on analyte loss and plasma cooling. In order to avoid these adverse effects, we explored the use of a stop-flow mode for the atomization stage. The gas flow rate under stop-flow mode refers to the flow which was applied during the drying stage and ashing stage of the atomization cycle, while the gas flow rate under continuous-flow mode refers to the flow which was applied during the whole atomization cycle. Fig. 4.17 and Fig.4.18 depict the comparison of emission intensities between continuous-flow mode and stop-flow mode for Cr and Cd. These graphs show that the atomic and ionic emission intensities of Cr and Cd were stronger under stop-flow mode than under continuous-flow mode. Compared with Cd, Cr had a greater enhancement of ionic emission intensity at high gas flow rates. 91 co x Q. 1200-1 1000 800 H 6 0 0 - | * 400-I . 200 JO- --a V 8 50 Cr(l)(continuous-flow) -x>- Cr(l)(stop-flow) --S3- Cr(ll)(continuous-flow) -L3- Cr(ll)(stop-flow) —a-~. 100 150 200 Gas flow rate (ml/min) 250 .--® —I 300 Fig. 4.17 Comparison of stop-flow mode and continuous-flow mode on Cr emission intensities a. 20x10 - i c o u m C CD c o Vi cn E CD O I o < Cd(l)(continuous-flow) Cd(l)(stop-flow) S;- Cd(ll)(continuous-flow) • - Cd(ll)(stop-flow) 15 H 10 5 H OH i - 800 600 o zs o " CD 3 CO cn o ' 3 h 400 3-200 h o o o cr =J I—. cn "5 X 150 200 Gas flow rate (ml/min) 300 Fig. 4.18 Comparison of stop-flow mode and continuous-flow mode on Cd emission intensities 92 In order to show this enhancement phenomenon, emission intensities were normalized by dividing emission intensities under stop-flow mode by their respective emission intensities under continuous-flow mode. Fig. 4.19 is a plot of the normalized emission intensities of Cr, Fe, and Mg. The data in the graph show that atomic and ionic emission intensities of Fe and Cr under stop-flow mode were higher than those under continuous-flow mode, and, strikingly, the ionic emission of Cr was greatly enhanced. Since there was no gas flow during the atomization stage under stop-flow mode, the plasma temperature could be maintained without being cooled down. Additionally, Welz et al [8] also found that the gas temperature in GFAAS is 200-500 K higher under stop-flow mode than under a continuous-flow mode. Therefore, the overall temperature interaction would contribute to the enhancement of emission intensities. Furthermore, the analyte transport mechanism under stop-flow mode is different from that under continuous-flow mode. Under continuous-flow mode, analyte atom vapor is carried out of the graphite tube, causing greater sample loss. In contrast, under stop-flow mode, analyte transport is determined by convection and diffusion, which are caused by the temperature difference. Because the temperature distribution in FAPES source is longitudinally dependent, analyte migrates from the tube center to the ends, and, as a result, the longitudinally integrated emission signals are enhanced. 93 Relative intensity ratio o ro cn co 1 1 1 1 1 - a - Cr(l) - © - Cr(ll) - S - Fe(l) ja—— 3 - ® - Fe(ll) - « - Mg(l) - • - Mg(ll) ••„• • a 1 1 1 1 1 1 50 100 150 200 250 300 Gas flow rate (ml/min) Fig. 4.19 The enhancement of emission intensities under stop-flow mode for Cr, Fe and Mg 4.4 Summary The investigations of temporal profiles for Mg, Cd, Fe, Zn, and Cr during the atomization cycle shows that analyte appearance temperature determines the time at which maximum atomic emission and ionic emission occur, and an increase in atomization temperature may reduce the appearance time of the maximum. An appropriate atomization temperature may maximize atomic and ionic emission intensities, and ionic emission profiles are more affected by the atomization temperature than by the rf power. For each analyte, an optimum gas flow rate can increase the atomic and ionic emission intensities. Operation in the stop-flow mode can provide an enhancement of both atomic and ionic emission intensities. There has been some recent interest in developing the FAPES source as an atomization and ionization source for analytical mass spectrometry [10-13]. The results 94 presented in this chapter suggest that this could be possible if certain experimental conditions were used. Namely, to maximize the ionization, the FAPES source should be operated at a fairly high rf power applied to the center electrode. Spatial selection of zones adjacent to the center electrode, where ionization is greatest, should provide the highest ion current to the mass spectrometer from the FAPES source, although this complicates the extraction because, in the current design, the center electrode is powered. The temporal behavior of ion production will help determine when to sample the ions so as to improve the signal to noise ratio. Based on the fact of temporal variation of ion production, a time of flight mass analyzer would probably be the best choice for simultaneous ion detection. 95 4.5 References 1. R.C. Weast and M.J. Astle, CRC Handbook of Chemistry and Physics 59th edition (CRC Press, U.S., 1978). 2. W.L. Wiese, M.W. Smith and B.M. Miles, Atomic transition probabilities: Volume II Sodium through calcium, a critical data compilation (Nat. Bur. Stand., U.S., 1969). 3. J. Kuba, Conincidence Tables for Atomic Spectroscopy (Elsevier, New York, 1965). 4. B.V. L'vov, Spectrochim. Acta 33B: p. 153 (1978). 5. T.D. Hettipathirana and M.W. Blades, J. Anal. At. Spectrom. 7: p. 1039 (1992). 6. R.E. Sturgeon, S.N. Willie, V. Luong and S. S. Berman, J. Anal. At. Spectrom. 5: p. 635 (1990). 7. G.F.R. Gilchrist, P.M. Celliers, H.Yang, C. Yu and D.C. Liang, J. Anal. At. Spectrom. 8: p. 809 (1993). 8. B. Welz, M. Sperling and G. Schlemmer, Spectrochim. Acta 43B: p. 1187 (1988). 9. L. Samuel, K. Nakagawa and T.Kimijima, Fresenius J. Anal. Chem. 356: p. 31 (1996). 10. C.W. LeBlanc, Ph.D. Dissertation, University of British Columbia (1996). 11. R.E. Sturgeon and R. Guevremont, Anal. Chem. 69: p. 2129(1997). 12. R.E. Sturgeon and R. Guevremont, J. Anal. At. Spectrom..13: p. 229(1998). 13. S.Y. Lu, C.W. LeBlanc and M.W. Blades, J. Anal. At. Spectrom. 16(3): p. 256(2001). 96 Chapter 5 EFFECTS OF OPERATING CONDITIONS ON ANALYTE IONIZATION 5.1 Introduction In recent years, there have been a few reports detailing the development of FAPES as an ion source for mass spectrometry [1-4]. The primary interest behind these investigations is to try to produce an elemental mass analyzer with low cost, simple operation, and environment familiar to the users of GFAAS. As previously described in Chapter 1, the analytical characteristics of a FAPES source make the exploration of developing the FAPES source as an ion source for mass spectrometry promising. First of all, according to a previous study [5], the analyte ionization temperature may reach more than 5000 K at 100 W. Based on the results from studies of plasma temperatures, reported in Chapter 3, the newly designed FAPES source used in our lab has higher plasma temperatures, for example, the Pb excitation temperature is about 6000 K at 80 W, and thus a higher than 6000 K ionization temperature is expected in this FAPES source, and at this temperature most elements can be ionized to a significant degree. Second, the FAPES source employs helium as plasma gas, and a helium plasma has some potential advantages relative to an argon plasma (commonly employed for the inductively coupled plasma (ICP)): for example, its monoisotopic nature and low mass (eliminating isobaric interference), high ionization energy (non-thermally ionizing some elements with high ionization potential [6]), good electrical resistivity (causing a high potential dropped between the electrode and furnace wall), and suitable thermal conductivity (allowing faster source heating through convection and radiation). Third, as a combined source, FAPES 97 source has very high analyte transport efficiency which improves atomization efficiency and analytical sensitivity for analytes. The analytical performance of an ion source is determined by a variety of factors, for example, the degree of ionization, the energy of the ions produced, and the efficiency with which ions can be extracted. The degree of ionization of an analyte in this source is our primary focus in this study. For a potential ion source, effective ionization is required to produce sufficient ion density so that ions can be sampled from the plasma with high sensitivity and low limits of detection. The effects of several operating conditions on atomic emission in the FAPES source have been discussed previously. Gilchrist et al. studied the effect of atomization temperature on atomic emission and showed that each element has an optimum atomization temperature [7]. Sturgeon and Guevremont demonstrated that gas flow rate might influence atomic and ionic emission intensities [3]. Hettipathirana and Blades observed that Pb has maximum atomic emission at 50 W rf power level [8]. LeBlanc and Blades described similar behavior for Mg atomic emission intensity and concluded that ionization is the main cause of the decrease in atomic emission intensity at higher rf power [9]. From the above discussion, it is apparent that operating conditions play a role in analytical figures of merit for atomic emission. Similarly, they may also influence analyte ionization. Therefore, an optimization of a variety of operating conditions may improve analyte ionization in the FAPES source. The primary purpose of this chapter is to investigate the effects of rf power, atomization temperature, gas flow rate, and center electrode size on analyte ionization. In this chapter, an experimental measurement of degree of ionization is used to characterize analyte ionization. 98 5.2 Experimental A complete description of the experimental system employed to acquire atomic and ionic emission signals is presented in Chapter 2. Five elements were chosen as spectrometric species. Stock solutions (1000 ppm) were prepared from the respective salts (MgO, CdCl2, ZnCl2, and FeS04) in 1% HN0 3, and Cr was obtained from a commercial stock solution (SPEX chemical, 1000 Ltg/ml in 2% HN03). Sample solutions were made by diluting the stock solutions. Because the analytical line pair of 285.2 nm and 279.6 nm chosen in this experiment for Mg are the strongest spectral lines of Mg (I) and Mg (II) respectively, 50 pg was deposited (5uL of a 10 ng/ml solution). However, for the other elements, the analytical lines which satisfy the previously mentioned requirements for the calculation of the degree of ionization (described in Chapter 4) are not the most sensitive lines, and therefore, 50 ng (5 uL 10 ppm solution) was used. The related parameters of these line pairs have been presented in Chapter 4. The experimental procedures have been described in Chapter 4. The operating conditions for each part of the experiment are listed in Table 5.1. For the experiment involving a constant center electrode volume, one electrode 42 mm in length and 0.9 mm in diameter and another electrode 34 mm in length and 1.0 mm in diameter were used to keep a constant electrode volume inside the graphite tube. 5.3 Results and Discussion 5.3.1 Effect of rf Power The temporal behavior of the effect of rf power on atomic and ionic emission is discussed in chapter 4. Now we can use this data to examine the effect of rf power on the degree of ionization of the analyte during an atomization cycle. Fig. 5.1 has a series of plots 99 that show the temporal behavior of the degree of ionization for Fe at rf powers ranging from 20 to 80 W. It can be seen that, as the rf power level increased, the appearance time for Fe ionization was earlier and, generally, the degree of ionization shows an increasing trend as the power is varied from 20 to 80 W. Table 5.1 Experimental conditions for each operating condition experiment Experimental parameter Atomization temperature (°C) Electrode diameter (mm) Gas flow rate (ml/min) Sample amount (ng) rf power (W) Atomization temperature variable 0.9 50 0.05 for Mg, 50 for others 80 Gas flow rate 2600 0.9 variable 0.05 for Mg, 50 for others 80 rf power 2600 0.9 50 50 for Fe, Zn variable Electrode size 2600 variable 50 0.05 for Mg, 50 for others 80 Fig. 5.2 and Fig. 5.3 illustrate the effect of rf power on the intensity ratio of ionic emission and atomic emission as well as the degree of ionization for Fe and Zn. The intensity ratio and degree of ionization for Fe (Fig. 5.2) increased with an increase in rf power up to a power of 80 W and then decreased at higher rf powers. The reason is that as rf power was increased up to 80 W, more energy would be input into the plasma due to an enhanced rf electric field. This surmise was supported by the studies of plasma temperature enhancement caused by an increase in rf power. LeBlanc [9] reported that the rotational temperatures of plasma background species (N2 + and OH) increase with rf power over the range of 20-200 W. 100 0 1 2 3 4 5 Atomization time (Second) Fig. 5.1 Temporal profiles of degree of ionization at different rf power levels Hettipathirana [8] demonstrated that the excitation temperature of Fe increases from 3100 to 4200 K over the rf power range of 15 to 50 W. In Chapter 3, our study also further confirms that plasma temperatures increase with increasing rf power, and then these temperatures level off at some power level. Although there is no report of ionization temperature in FAPES, it's reasonable to expect that an increase in rf power will increase ionization temperature leading to an improvement of the degree of ionization. When the forward power was above 80 W, higher rf power resulted in considerable thermionic emission from the heated tube wall. This decreases the resistance of the plasma and causes the impedance to be mismatched between the plasma and the matching network. As a result, less forward power is transferred to the 101 plasma and much more power is dissipated in the matching network or is reflected. Thus, the plasma temperatures are reduced above 80 W. On the other hand, the intensity ratio and degree of ionization for Zn increased over the range of 25 -125 W (Fig. 5.3). The differing behavior for Zn and Fe is perhaps due to less thermionic emission and the unchanged plasma impedance when Zn was analyzed. The degree of ionization for Zn was consistently increased within the rf power operating range, but its maximum degree of ionization may occur at an rf power level higher than 120 W. Power (W) Fig. 5.2 rf power effect on the intensity ratio of ionic and atomic emission as well as the degree of ionization for Fe 102 40 — i -3 Degree of ionization - ® - Zn(ll)/Zn(l) 100x10 5 40 60 80 Power (W) 100 120 Fig. 5.3 rf power effect on the intensity ratio of ionic and atomic emission as well as the degree of ionization for Zn 5.3.2 Effect of Atomization Temperature From previous studies in Chapter 4, we know that the atomization temperature influences vaporization and atomization of analyte, and thus it may affect analytical performance such as analyte ionization capability. Fig. 5.4 shows the influence of atomization temperature on the degree of ionization of Mg. At lower atomization temperatures, the intensity ratio and the degree of ionization for Mg were depressed due to an incomplete atomization process and/or spatial segregation of ionic emission and atomic emission. In graphite furnaces based on the Massmann design (or end-heated furnace), the graphite tube atomizer ends are in close contact with water-cooled graphite end pieces. Therefore, a temperature gradient exists over the length of the furnace. Since the analyte vapor distributed over the graphite tube is heated by radiation from the furnace wall, a high 103 1 OO—i ^ 8 0 -c S 6 0 -c g o 4 0 -CD CD e? 2-0-Q 0 -i Degree of ionization - X - Mg ( l l ) /Mg ( l ) I I 1 6 0 0 1 8 0 0 I 2 0 0 0 1 2 2 0 0 I 2 4 0 0 Temperature ( C) CQ h 0 . 8 - 5 —r 2 6 0 0 h 0 . 0 Fig. 5.4 Atomization temperature effect on the intensity ratio of ionic and atomic emission as well as the degree of ionization for Mg atomization temperature is necessary for all of analyte vapor to be atomized due to a temperature gradient inside the furnace. Thus, an enhancement of the intensity ratio and the degree of ionization occurred between 1800 °C and 2000 °C which is much higher than the appearance temperature of Mg. However, the increasing value for the intensity ratio and the degree of ionization of Mg became plateaued after 2000 °C, therefore, a further increase in atomization temperature wouldn't help improve analyte ionization. The effect of atomization temperature on the intensity ratio of ionic emission and atomic emission as well as the degree of ionization for Cd is presented in Fig. 5.5. Increasing atomization temperature didn't enhance the intensity ratio or the degree of ionization within the atomization temperature range from 1500 °C to 2600 °C. As was previously discussed in Chapter 4, the low appearance temperature of Cd corresponds to a low optimum atomization 104 3 0 - | 2 5 -c g 2 0 -CD N ' c 1 5 -g o CD 1 o -CD O) CD Q 5 -0 --•©•- Degree of ionization ->« -Cd( l l )/Cd( l) @ 0 . — x- -X-T — © -X- -X-r 0.20 h 0.15' -® <a o Q. h 0.10! 0.05" I I I 1600 1800 2000 2200 2400 2600 o Temperature ( C) J- 0.00 Fig. 5.5 Atomization temperature effect on the intensity ratio of ionic and atomic emission as well as the degree of ionization for Cd temperature. When Cd achieved optimum emission intensities and the degree of ionization due to a complete atomization under some atomization temperature, a further increase in atomization temperature can't enhance its atomic and ionic emission as well as the degree of ionization, and on the contrary, it may weaken ionic emission and decrease the degree of ionization. Apparently, the appearance temperature difference between M g (1530 K) and Cd (730 K) is one of the major factors contributing to their different behaviors. Fig. 5.6 shows the effect of atomization temperature on the degrees of ionization for Fe, Cr, and Zn. The trend observed is that the degree of ionization increased as the atomization temperature increased. Clearly, the degree of ionization of Cr increases over the i operating range of atomization temperature. Compared with Fe (1580 °C, 7.90 eV) and Zn (1050 °C, 9.39 eV), Cr has a higher appearance temperature and a lower ionization energy (1740 °C, 6.76 eV). Thus, a sufficiently high temperature would provide a complete 105 atomization of Cr and thus enhance Cr ionization process. As a result, the degree of ionization of Cr was increased. For Fe, the degree of ionization basically increased with increasing atomization temperature within the temperature range of this experiment. It should be noted that below 2000 °C, ionic emission for Zn couldn't be detected probably due to its high ionization energy (9.39 eV). At higher atomization temperatures, thermal expansion of neutral species inside graphite tube decreased species density by a factor of Ti IT2 (Ti -atomization temperature, T2 - room temperature) which thus increased electron mean free path by a factor of T1/T2. During the atomization cycle, electrons might gain more energy in the rf electric field before they collided with analyte atoms, and therefore there was an increase in the number density of electrons whose energy is sufficient to cause analyte ionization. As a result, analyte ionization was enhanced, leading to an increase in the production of Zn ions. Additionally, thermionic emission during the atomization cycle increased the electron density, and correspondingly, the electron-analyte collision frequency was increased, resulting in an enhanced Zn ionization process. 1600 1800 2000 2200 2400 2600 Temperature ( C) Fig. 5.6 Atomization temperature effect on the degree of ionization of Fe, Cr, and Zn 106 5.3.3 Effect of Gas Flow Rate During an atomization cycle, diffusion is the dominant analyte loss mechanism, however, thermal expansion of the gas inside the tube represents a further significant mechanism for expelling analyte from the furnace and is increasingly important at high heating rates and especially for the volatile elements. Apart from diffusion and expulsion, analyte atoms and ions are also swept from the analysis volume by the forced flow of helium gas. Therefore, gas flow rate can affect analytical performance. The effect of gas flow rate under continuous-flow mode is described in Fig. 5.7. With an increase in gas flow rate from 30 ml/min to 50 ml/min, the degrees of ionization of Fe and Cr were increased. The degree of ionization of Fe was reduced gradually at gas flow rates higher than 80 ml/min. However, the optimum gas flow rate for Cr with the greatest degree of ionization is higher than that for Fe, and a further increase above the optimum gas flow rate also decreased the degree of ionization of Cr. Similarly, an optimum gas flow rate of 30 ml/min was found to provide better analytical performance for Mg. Higher gas flow rates resulted in a shorter residence time for Mg, leading to the reduction of the degree of ionization. 107 100-n - e - Fe - X - Cr - A - Mg X ion (%; 8 0 - x * — ro N 6 0 -c o 4 0 -d) CO A 8 2 0 -o -I 50 I 100 I I 150 200 Gas flow rate (ml/min) I 250 I 300 Fig. 5.7 Gas flow effect on the degree of ionization of Fe, Cr, and Mg under continuous-flow mode As discussed in Chapter 4, the gas flow rate under stop-flow mode may affect analyte atomization process and /or excitation and ionization processes, and thus an investigation of the effect of stop-flow mode on the degree of ionization may be helpful to the optimization of analyte ionization capability. Fig. 5.8 is a diagram of the effects of gas flow rate on the degrees of ionization of Fe, Cr and Mg. Compared with those under continuous-flow mode (Fig. 5.7), the degrees of ionization of Fe and Mg under stop-flow mode were slightly enhanced. Without the introduction of a forced gas flow, thermal expansion greatly reduced gas density. The longer residence time and a smaller collision frequency allow analyte to be efficiently ionized in the plasma. However, as also indicated in Fig. 5.8, the degree of ionization for Cr was not enhanced above 100 ml/min. While the emission intensities of Cr 108 were both enhanced, stop-flow mode was more advantageous for Cr atomic emission than for Cr ionic emission, and hence there was no improvement in the degree of ionization of Cr. 100^ - @ - Fe Cr 0s- 8 0 --A - Mg —y • • x c • — , X @ tizatic 6 0 -A " i_ o \ o CD 4 0 - \ Degre 2 0 -0 -— — A A I 50 i 100 i i 150 200 I 250 I 300 Gas flow rate (ml/min) Fig. 5.8 Gas flow effect on the degree of ionization of Fe, Cr, and Mg under stop-flow mode 5.3.4 Effect of Center Electrode Size Since a greater percentage of the forward rf power is dissipated in the matching network, a study to increase actual rf power delivered to the plasma is desirable. Impedance matching has been proved to be the most important factor in determining the amount of the energy transfer from the rf power supply to the plasma [10-11]. According to Mohamed Salem [12], the surface areas of two electrodes of a plasma source play an important role in the plasma impedance. The plasma impedance consists of a resistive part and a reactive part, which includes an inductive component and a capacitive component. The resistive part is inversely proportional to the average area of the two electrodes of a plasma source; both the 109 inductive component and the capacitive component are also functions of the surface areas of two electrodes. From above discussion, we can expect that changing the dimensions of electrodes may modify the impedance of the FAPES source so as to cause more power to be applied on the plasma. Since it's difficult to change the dimensions of the graphite tube due to the commercial availability and instrumental limitations, a change in plasma impedance is expected to achieve by modifying the dimensions of the center electrode. A variety of center electrodes of 0.9, 1.0, 1.4, and 1.9 mm in diameters were chosen for this study, and the lengths of the electrodes were kept constant and extended along the full graphite tube in order to make the plasma fill the graphite tube. By employing these electrodes, we studied atomic emission and ionic emission under continuous-flow mode, and the effect of electrode diameter on atomic emission and ionic emission is shown in Fig. 5.9. By dividing emission intensities collected under the experiments of employing electrodes of 1.0, 1.4 and 1.9 mm by their corresponding emission intensities collected under the experiments of employing electrode of 0.9 mm, the atomic and ionic emission intensities employing electrodes of 1.0, 1.4 and 1.9 mm were normalized so that they can be easily compared with those employing electrode of 0.9 mm. The results show that as the electrode size increased from 0.9 mm to 1.9 mm, ionic emission intensities of all elements decreased, and atomic emission intensities were also decreased except for that of Cr. In Chapter 3, we found that increasing center electrode size decreases the plasma temperatures including excitation temperature, and thus as a function of excitation temperature, unsurprisingly, atomic and ionic emission intensities were reduced. The unusual behavior of Cr atomic emission may be due to its some unknown vaporization and atomization characteristics. 110 1.0 1.2 1.4 1.6 1.8 Electrode diameter (mm) Fig. 5.9 The effect of electrode diameter on atomic emission and ionic emission I l l Fig. 5.10 shows the effect of center electrode size on the degree of ionization of these elements. Generally, the degree of ionization decreased with increasing electrode size. Several reasons may be responsible for this observation. One reason is that due to the increasing electrode surface area, the current density and power density were reduced. In addition, the increasing electrode size may cause a worse impedance mismatch, leading to less rf power transfer to the plasma source. Consequently, less voltage was dropped across the plasma sheath where the greatest excitation and ionization occur. A third possible reason is a change in the plasma electrical characteristics. Since the FAPES source has an asymmetric electrode structure, i.e., the center electrode has a much smaller surface area relative to the furnace wall, the voltage drops in these two regions (one is close to the center electrode and the other is close to the furnace wall) are much different based on the scaling law - Equation 3.3, and as a result, a region with strong ionization and excitation is formed close to the center electrode. With increasing electrode size, the surface area of the center electrode was correspondingly increased, and thus the voltage drop within the plasma sheath around the center electrode would be smaller. Thus, the electric field in the zone close to the electrode was weakened, leading to a reduction in analyte ionization. Since the surface area of the furnace wall is very large when compared with that of the center electrode, the voltage drop in the zone close to the furnace wall would increase only slightly, which would not effectively enhance ionization and excitation capability of this region. J 112 5.3.5 Effect of Center Electrode Size with Constant Electrode Volume As discussed above, an increase in the center electrode size with constant length can lead to a diminution of analyte ionization capability. One reason might be due to a reduction of power density resulting from increased electrode volume, and thus in this section, we'll investigate the effect of different size electrodes with constant electrode volume on analyte ionization capability. Fig. 5.11 is the effect of center electrode size with constant electrode volume on the degrees of ionization of analyte. The electrodes of 0.9 mm and 1.0 mm were used for this study, and four elements - Cr, Mg, Fe and Cd were chosen for this experiment. The graph shows that for all the elements, the degrees of ionization are much lower with the use of a 1.0 mm center electrode than with the use of a 0.9 mm center electrode. Thus, the 1.0 mm electrode doesn't improve analyte ionization capability. The result suggests that the ionization capability of the analytical zone close to the center electrode was weakened due to an increase in the center electrode size, the causes of which were discussed in the previous section. If we want to maintain the voltage drop in the analytical zone close to the center electrode under the usage of a shorter, larger electrode, we also have to use a shorter, larger graphite tube. However, due to the physical and instrumental limitations, we cannot change the length and size of the graphite tube. 5.3.6 Effect of Center Electrode Length Since the measurements of emission intensities inside the FAPES source are longitudinally integrated over the whole plasma volume, it is reasonable to expect a reduction 113 7.0 7.5 8.0 8.5 Ionization energy (eV) Fig. 5.11 The effect of center electrode size with constant electrode volume on the degrees of ionization of analyte 100^  - a - 42 mm 80- 34 mm C O lizat 60-i— o iree of 40-iree of Deg 20-o-I 7.0 I 7.5 I 8.0 I 8.5 lonizatbn pontential (eV) Fig. 5.12 The effect of center electrode length on the degrees of ionization of analyte 114 in emission intensities as observed in the experiment when the center electrode fails to fill the entire graphite tube length. If the center electrode only partially fills the graphite tube length, what happens to the degrees of ionization? Fig. 5.12 shows the effect of center electrode length on the degrees of ionization of analyte. In this study, 1.0 mm center electrodes with different lengths were used. An electrode with 42 mm in length fully fills the entire graphite tube, and another electrode with 34 mm in length partially extends about % of the length of the graphite tube. The graph demonstrates that the degree of ionization is slightly lower when a 34 mm electrode was used than when a 42 mm electrode was used. A couple of reasons might contribute to this experimental result. Since the plasma is generated between the center electrode and the graphite tube, a reduction in the length of the center electrode leads to a decrease in the plasma volume. In addition, because the rf power is applied to the center electrode, a change in the length of the center electrode would cause an alteration of the impedance of the FAPES source, possibly affecting the rf power transferred to the source and voltage drop. Therefore, as a combined result of these reasons, decreasing the center electrode length resulted in the reduced ionization capability of the FAPES source. 5.4 Summary The effects of rf power, atomization temperature, gas flow rate, and dimensions of center electrode on analyte ionization in the FAPES source have been studied in this chapter. Increasing the applied rf power is a better way to increase analyte ionization, however, a higher rf power than an optimum value doesn't increase the degree of ionization. In addition, the maximum rf power that can be applied to the center electrode is also limited by the structure of the FAPES source and analyte atomization characteristics. Analyte ionization can 115 be enhanced by using an atomization temperature which ensures complete analyte atomization and seems to also add some extra energy to the plasma source. However, a very high atomization temperature may interfere with the normal plasma operation leading to deleterious side effects on the performance of the plasma source. Helium in the FAPES source plays several roles: it purges the graphite tube, it sustains the plasma, and it transports analyte for atomization. Thus, sufficient helium gas flow rate is required for the maintenance of the plasma and analyte atomization while a lower helium gas flow rate is preferred for a longer analyte residence time and for temperature stabilization. Therefore, an optimum gas flow rate can be found for analyte ionization. For some analytes such as Fe and Mg, analytical sensitivities and ionization capability can be improved using stop-flow mode. The observations of electrode size effect on anlyte excitation and ionization show that increasing the electrode size doesn't improve the analytical performance of the plasma source. An insufficient center electrode length not only decreases the atomic and ionic emission intensity but also lowers analyte ionization capability. The electrode diameter plays a more important role in determining the ionization capability of the FAPES source than the electrode length does. In sum, these operating conditions can influence analyte ionization by affecting one or more physical and electrical processes such as atomization, excitation, ionization, and current/voltage characteristics. 116 5.5 References 1. LeBlanc, C. W., Ph.D. Dissertation, University of British Columbia (1996). 2. Sturgeon, R.E. and R. Guevremont, Anal. Chem. 69: p. 2129(1997). 3. Sturgeon, R.E. and R. Guevremont, J. Anal. At. Spectrom.13: p. 229(1998). 4. Lu, S.Y., C.W. LeBlanc, and M.W. Blades, / . Anal. At. Spectrom. 16(3): p. 256 (2001). 5. Sturgeon, R.E., S.N. Willie, and V.T. Luong, Spectrochim. Acta 46B: p. 1021 (1991). 6. Beenakker, C.I.M., Spectrochim. Acta 32B: p. 179 (1977). 7. G.F.R. Gilchrist, P.M. Celliers, H.Yang, C. Yu and D.C. Liang, J. Anal. At. Spectrom. 8: p. 809(1993). 8. Hettipathirana, T.D. and M.W. Blades, Spectrochim. Acta, 47B: p. 493 (1992). 9. LeBlanc, C.W. and M.W. Blades, Spectrochim. Acta 50B: p. 1395 (1995). 10. John, R. and T. Alexis, Bell Techniques and Applications of Plasma Chemistry (John Wiley & Sons, 1974). 11. Bowick, C , RF Circuit Design (Howard W. Sams & Co., 1989). 12. M. Mohamed Salem, J.-F. Loiseau and B. Held, Eur. Phys. J. AP 3: p. 91 (1998). 117 Chapter 6 STUDIES OF ANALYTE MATRICES AND PALLADIUM MODIFIER EFFECTS 6.1 Introduction Interference is a common phenomenon encountered when using spectrometric methods. For plasma emission or mass spectrometry, interferences can be divided into two categories: spectroscopic interferences and non-spectroscopic interferences. Spectroscopic interferences are mainly caused by incomplete isolation of the analyte signal from that of interferent. For example, in plasma source mass spectrometry, some atomic or molecular ions that have the same nominal mass as the analyte can interfere with the analysis by causing an erroneously large signal at the mass-to-charge of interest. The polyatomic ion interferences result from molecular ions formed from precursors in the plasma gas, entrained atmospheric gases, acids used for dissolution, water, and the sample matrix. A non-spectroscopic interference is characterized by a change in analyte signal due to the suppression or enhancement of analyte signal by the sample matrix. Since the sample matrix can affect the analytical performance of plasma sources, an understanding of sample matrix effects is helpful in the characterization and optimization of the analytical performance of the FAPES source. The effect of easily ionized elements (EIE's) is a frequently encountered sample matrix interference in atomic spectroscopy and has been widely discussed and studied over the years [1-4]. When a large excess of an EIE element with a low first-ionization potential (e.g. Na or K) is present in the sample as a concomitant, a large excess of electrons and positive ions will be produced after the ionization of the EIE. The excess of electrons force the equilibrium for the analyte towards atom formation, resulting in a suppression of analyte 118 ion formation. In addition, EIE's may influence the analyte vaporization, atomization, and excitation processes. Several papers have discussed the effect of EIE's on analyte atomic emission and analyte distribution in the FAPES source [2-4]. In this source, power dissipation may occur as a result of the excitation of EIE's, and radiative power losses from the plasma due to excitation of the EIE matrix species can influence the plasma temperatures and appear to lead to an alteration of the discharge characteristics. Since previous studies have already described the EIE effect on the FAPES source in detail [1-4], the topic of EIE's effect will not be covered in this thesis. Instead, we have studied the effect of a variety of different analyte matrices on the FAPES source. Here, analyte matrices refer to a group of analytes with the same cation and a different anion. For example, MgS04, MgCl2, MgN03, and MgO have same cation Mg, but different anions S042" , Cl", N03", and O2" respectively. Due to the different chemical and physical properties of such a group of analytes, analyte matrices may have a profound effect on the atomization, excitation, and ionization characteristics of the plasma discharge. In addition, since chemical modification of the graphite furnace may alter the vaporization and atomization properties of the analyte, as well as enhance retention of analyte at a higher temperature, an addition of some modifier may change the atomization efficiency, leading to an alteration of sensitivity and analytical performance. Palladium has been extensively used for the investigation of chemical modification mechanisms. Yang et al. [5] have studied the atomization efficiency for indium and tin from Pd atomizer surfaces and found that there are strong interactions between these two elements and palladium. Indium can form an alloy with the palladium modifier in the graphite furnace and is atomized in the phases Pdln3 ~ Pd2In3 ~ Pdln ~ Pd2In. As a result, the atomization of 119 indium was hindered by the reactions of indium with the Pd on the surface. In contrast, the Pd modifier catalyzed the reduction of SnO or Sn02 to metallic tin which forms a solid solution in excess of palladium, thus changing the atomization process, reducing the volatility of tin oxide, and leading to an improvement in the atomization efficiency. Therefore, chemical modification with palladium inside the FAPES source is expected to cause some chemical or physical interactions among the gaseous species, Pd modifier and the graphite furnace, and potentially influence analytical performance of the FAPES source. 6.2 Experimental The experimental system and experimental procedures have been described in previous chapters. For the analyte/matrix experiments, stock solutions (1000 ppm) were prepared from the respective salts (MgS04, MgCl2, MgN0 3, MgO) in 1% HN0 3 . For the preparation of Pd stock solution, 0.2500 g Pd was first dissolved in aqua regia and then diluted in 250 ml deionized water. For these experiments, 50 pg of Mg was used (5uL 10 ng/ml). All data were measured under the same experimental conditions. For the Pd modifier experiments, the palladium solution was co-injected with MgS0 4 solution, and the ashing temperature was set at 500 °C. 6.3 Results and Discussion 6.3.1 Effect of the Analyte Matrix 120 Since the analyte matrix can affect the vaporization and atomization mechanisms of an analyte, it is necessary to investigate its effect on atomic and ionic emission intensities as well as the degree of ionization. Fig. 6.1 shows the effect of different magnesium matrices on atomic emission, ionic emission and degree of ionization of Mg. The atomic and ionic emission intensities were calibrated so that the modified emission intensities were given from equimolar amounts of magnesium. The atomic and ionic emission intensities and the degree of ionization for the different matrices were normalized to those of magnesium chloride. From the graph, we can see that atomic emission intensities decreased slightly in the order of MgCb, MgO, and MgN03, but decreased sharply for MgS04. 1.2 -i MgCI2 MgO MgN03 MgS04 • Atomic emission • Ionic emission • Degree of ionization Fig. 6.1 The effect of different matrices on atomic emission, ionic emission and degree of ionization of Mg 121 As far as ionic emission intensities are concerned, MgCl 2 showed the highest emission intensity, and MgS04 the lowest. Among these matrices, MgCl2 has the greatest degree of ionization, and the others have lower values. The difference is probably due to different atomization mechanisms of these matrices. The possible atomization mechanisms are as follows: MgCl2(s) •MgCl 2 (g) •MgCl(g) •Mg(g) (6.1) Mg(N03)2 (s) • MgO (s) + 2 N0 2 (g) + 0 2 (g) (6.2) MgS04 (s) • MgO (s) + S0 3 (g) (6.3) MgO (s) + C (s) • Mg (g) + CO (g) (6.4) Daminelli et al. [6] performed vaporized MgCl2 in a graphite furnace, and observed the appearance of molecular or atomic spectra of MgCl2(g), MgCl(g) and Mg(g). In addition, according to their study, the hydrolysis of MgCl2 is a slow process which requires the presence of water vapor during the atomization step, but the drying effect is faster with the use of a gas, such as helium, into which the diffusion of water can take place at higher rate. Thus, a hydrolysis reaction to produce Mg(OH)Cl and Mg(OH)2 in a helium atmosphere is unlikely to occur with high atomization temperatures and low support gas flow rates. Therefore, MgCl2 most likely follows the route shown in Equation 6.1 to produce atomic Mg. Katskov [7] showed that the presence of graphite induces a reduction of MgO in the condensed phase, which facilitates the vaporization. Therefore, MgO is directly reduced to Mg by the reaction shown in Equation 6.4. However, the decomposition processes of Mg(N03)2 and MgS0 4 involve formation of magnesium oxide through the reactions 122 Equations 6.2 and 6.3 respectively, and then the creation of Mg atoms by the reaction shown in Equation 6.4. Because thermal expansion reduces the number densities of gas species during the atomization stage, this results in a shift of the reaction to the right, and Mg atomization from MgCl2 could easily occur, but Mg atomization from MgO is more complex. The formation of CO vapor in the FAPES source resists the reaction, shown in Equation 6.4, from proceeding to the right, and thus the production of atomized Mg is affected. In addition, MgO can form an ionic carbide (MgC2) at higher temperatures [8] so that the production of atomized Mg from the reaction shown in Equation 6.4 is reduced. Therefore, the atomization efficiency of MgO is lower than that of MgCl2, resulting in a difference in their analytical sensitivities. MgN03(s) and MgS04(s) proceed through Equation 6.2, 6.3 and 6.4 to be atomized. As a result, their atomization efficiencies are reduced. Since the emission intensity is a linear function of the number density of the emitting species, the differential in atomized Mg number density among four magnesium matrices resulted in the difference of their analytical sensitivities. Since a high atomization temperature may provide a lot of thermal energy which facilitates the vaporization of MgN0 3 [7], the atomization efficiency of MgN0 3 is higher than that of MgS04. In spite of the different emission behavior, the degrees of ionization for the matrices of MgS04, MgN03, MgO were close because that they all experience the same atomization process of MgO as shown in Equation 6.4. The reason for a higher degree of ionization of MgCl2 than those of other matrices is possibly due to the even distribution of the analyte and less interference to the plasma ionization capability. In summary, it has been shown that, for the FAPES source, the analyte matrix can alter the vaporization and atomization processes and thus influence analytical performance. 123 6.3.2 Effect of Pd Modifier Chemical modifiers may alter, physically or chemically, the graphite furnace surface, and this kind of surface alteration is expected to affect analyte vaporization and atomization in three ways [9]: • Evaporation of the matrix during the ashing stage causing a decrease in interferences during atomization. • A reduction in analyte volatility, preventing analyte losses during the ashing stage and retarding the vaporization of the analyte during the atomization stage before a constant temperature in the graphite furnace is reached. • A transformation of all chemical forms of the analyte to one form thus improving the precision. Since a palladium modifier is regarded as a universal chemical modifier [13, 17], it was used to modify the FAPES source and to investigate the interactions between the gaseous species and the graphite furnace surface. Fig. 6.2 is a diagram of the Pd modifier effect on the ratios of emission intensities and degree of ionization of Mg. Atomic emission intensity, ionic emission intensity as well as the degree of ionization of Mg using different amounts of Pd are plotted. The data are normalized using data for MgSC>4 without the addition of Pd. The added Pd amount is 2 ng, 5 ng, 20 ng, 50 ng, 500 ng and 2000 ng respectively. 124 o 03 1.2H 1.6H 1.4H Relative ionic emiss ion intensity ratio Relative atomic emiss ion intensity ratio - B - Relative degree of ionization ratio > 1.0H1 0 DC 0 . 8 H 0.6H 0.4H 0.5 1.0 1.5 2.0 LoQi o(Pd amount in ng) 2.5 3.0 Fig. 6.2 Pd modifier effect on the ratios of emission intensities and degree of ionization of Mg (80 W, 2600 °C, 50 ml/min) Clearly, as the amount of Pd increased, atomic emission intensity, ionic emission intensity and degree of ionization of Mg were enhanced, reached a maximum at 5 ng Pd, and decreased thereafter. The reason for the enhancement is that Pd can prevent carbide formation so that magnesium analyte loss is avoided. This explanation was supported by the observation that no double atomic emission peaks were observed when Pd modifier was added. During the atomization cycle, palladium nitrate decomposes via the oxide to the metal at 870 °C [10] which melts at 1552 °C [11]. As a result, Pd may serve two functions. First of all, palladium can activate the pyrolytically coated surface of the graphite forming channels through the graphite substrate [18]. Second, at temperatures lower than 2000 °C, palladium forms a [Pd, Mg, O] compound with MgO released from MgS04, stabilizing the magnesium. As it is heated, this compound will decompose near the graphite surface, i.e. 125 [Pd,Mg,0] • Pd(g) + Mg(g) (6.5) Although no experiment has been done to prove the above mechanism for the interaction between magnesium and palladium, we may get some support for the above mechanism from Fischer's study that revealed that the presence of palladium modifier may stabilize selenium [12]. The physical and chemical interactions could allow atomized magnesium vapor to migrate away from the graphite surface into the central region of observation, and as a result improve the magnesium number density there, leading to enhanced analytical sensitivity. The addition of Pd might also stabilize the plasma by removing interferences and alleviate the impedance mismatch of the plasma during the atomization cycle, resulting in an improvement of the degree of ionization of magnesium at a minute amount addition, 5 ng, of Pd modifier. However, the addition of a greater amount of palladium modifier may lead to an rf power loss due to excessive loading of the surface of the furnace. Moreover, a greater amount of palladium in the graphite furnace could cause an increase in the appearance temperature and an increase in activation energy relative to analyte alone [13-16]. As a consequence of these phenomena, the atomic emission intensity, the ionic emission intensity and the degree of ionization of magnesium were reduced. 126 6.4 Summary The magnesium analyte matrix effect experiment demonstrates that the matrix influences analyte vaporization and atomization processes as well as atomization efficiency, and the effect is related to concrete atomization processes. When the analyte has fewer reactions to achieve complete atomization, it potentially has higher atomization efficiency. Furthermore, the reactions involved in an analyte atomization process can be influenced by the pressure and concentration of gaseous species. Analyte ionization efficiency can only be improved a little by choosing an appropriate analyte matrix since plasma temperatures are not significantly enhanced. The addition of an appropriate amount of palladium as a chemical modifier into the FAPES source may remove volatilization and vapor-phase interferences and prevent the formation of interstitial compounds, leading to better stabilization for analyte atomization and plasma maintenance during the atomization cycle. This kind of stabilization may be achieved either by physically increasing the number of active sites on the graphite surface or by the chemical formation of a compound that contains both analyte and stabilizer components. Analyte ionization efficiency can be improved with an addition of a minute amount of palladium modifier. 6.5 References 1. S. Imai and R.E. Sturgeon, J. Anal. At. Spectrom. 9: p. 765 (1994). 2. T.D. Hettipathirana and M.W. Blades, J. Anal. At. Spectrom. 8: p. 955 (1993). 127 3. V. Pavski, R.E. Sturgeon and C L . Chakrabarti, J. Anal. At. Spectrom. 12: p. 709 (1997). 4. G.C.Y. Chan and W.T. Chan, J. Anal. At. Spectrom. 13: p. 209 (1998). 5. W.M. Yang and Z.M. Ni, Spectrochim. Acta 52B: p. 241 (1997). 6. G. Daminelli, D.A. Katskov, R.M. Mofolo and T. Kantor, Spectrochim. Acta 54B: p. 683 (1999). 7. D.A. Katskov, G. Daminelli and P. Tittarelli, Spectrochim. Acta 54B: p. 1045 (1999). 8. S.Y. Lu, C.W. LeBlanc and M.W. Blades, J. Anal. At. Spectrom. 2001. 16: p. 256 (2001). 9. A.B. Volynsky, Spectrochim. Acta 53B: p. 139 (1999). 10. G. Schlemmer and B. Welz, Spectrochim Acta 41B: p. 1157 (1986). 11. R.C Weast and M.J. Astle, Eds, Handbook of chemistry and Physics 6F' Edn. (CRC Press, Boca Raton, 1980) 12. J.L. Fisher and CJ. Rademeyer, Spectrochim. Acta 53B: p. 1998 (1999). 13. D.L. Styris, L.J. Prell, D.A. Redfiled, J.A. Holcombe, D.A. Bass and V. Majidi, Anal. Chem. 63:p.508 (1991). 14. A. Mazzucotelli and M. Grotti, Spectrochim Acta 50B: p. 1897 (1995). 15. C L . Chakrabarti, S.J. Cathum, Talanta 38: p. 157 (1991). 16. J.L. Fisher and CJ. Rademeyer, Spectrochim Acta 53B: p. 537 (1999). 17. D.L. Styris, L.J. Prell and D.A. Redfield, Anal. Chem. 63: p. 503 (1991). 18. K.W. Jackson, Electrothermal Atomization for Analytical Atomic Spectrometry (John Wiley & Sons Ltd, USA, 1999). 128 Chapter 7 ANALYTICAL PERFORMANCE OF THE FAPES SOURCE 7.1 Introduction For a spectrochemical method, figures of merit of a group of analytes are usually used to demonstrate the method's analytical performance. These figures include several parameters, such as accuracy, precision, sensitivity, detection limit, dynamic range, signal to noise ratio (S/N), and signal to background ratio (S/B). In this chapter, we will use detection limit, dynamic range, signal to noise ratio (S/N) and signal to background ratio (S/B) to characterize analytical performance of magnesium in the FAPES source. The detection limit is a critical figure of merit since a spectrometric technique cannot be used without a preconcentration step if the analyte concentration in the analytical sample is below the detection limit. The detection limit is typically defined as the analyte concentration yielding an analytical signal equal to three times the standard deviation of the signal from the blank . Dynamic range can be defined as the concentration range over which the analytical curve is linear or the calibration slope is constant. Signal to background ratio (S/B) is used to represent the response function (or optimization criterion) of analytical signal. Here S is the analytical signal from the analyte and B is the sum of background spectral signals from a blank sample. Signal to noise ratio (S/N) is the ratio of analytical signal and the rms noise in the total signal that is equal to the sum of the noise contributions from the analytical and the blank signals. The noise in the blank signals can be subdivided into individual components: shot noise and flicker noise in the background signal as well as dark current rms noise. For analytical emission techniques, the background noise is mainly caused by background 129 emission. In a plasma emission source, the background signal noise is the dominant part of the blank noise. Since little information about ionization temperature and electron number density was available for the newly built FAPES source, we are also interested in characterizing these parameters. The relationship between these two parameters and the degree of ionization (an indication of ionization efficiency) is shown in Saha equation [9], and this equation can be rewritten as follows. _ 2 L = A = i A ( 2 m T l k B If exp(-EJkBT) (7.1) 1 - a n a n e Zg \f where a is the degree of ionization of the element of interest, n;, n^ and na are the number densities of the ions, free electrons, and atoms respectively, Zx and Z a are the ionic and atomic partition functions respectively, m is the electron mass, h is Planck's constant, ks is Boltzmann constant, T is the temperature, and E; is the first ionization potential of the element of interest. From above equation, we may derive a linear equation as follows. log [ct/(l-a)] = 15.684 + log(Zi/Za) + 1.5 logT -5040 (Ej/T) - log(ne) (7.2) Where the units of n e and Ej are cm"3 and eV, respectively. Based on Equation 7.2, a plot of log [a/(l-ct)] versus E; should yield a straight line from the slope of which (-5040/T) the ionization temperature can be derived without a priori knowledge of the electron density in the FAPES plasma. After we obtain the ionization temperature of the plasma, we may calculate the electron number density by substituting the degree of ionization, the ionization potential and partition functions of some specific element as well as ionization temperature into Equation 7.2. 130 7.2 Experimental The experimental set-up and procedures are almost same as those used in previous chapters unless otherwise stated. The measurements of detection limit, dynamic range, signal to noise ratio (S/N) and signal to background ratio (S/B) of magnesium were undertaken using an atomization temperature of 2100 °C. The blank determinations without the analyte were carried out in the same manner as those sample analyses with the analyte. In order to determine the ionization temperature, the measurements of the degrees of ionization for Cr, Mg, Fe, Cd and Zn were performed at the same operating conditions. Since the electronic partition functions of the atoms and ions of analyte are temperature dependent, the following equation [1] was employed to calculate the partition functions of Cr, Mg, Fe, Cd and Zn. Z(T) = a + b ( ^ 3 ) + c ( - ± 3 ) 2 + d (^ - ) 3 + e ( ^ 3 )4 + f ( - ^ )5 (7.3) 10 3 10 3 10 3 10 3 103 Where a, b, c, d, e and f are coefficients which can be obtained from de Galan et al. [1]. 7.3 Results and Discussion 7.3.1 Analytical Figures of Merit of Mg Since Mg is commonly used for the evaluation of the analytical performance of a plasma source such as the ICP [2], Mg was chosen as the spectrometric species for this study. The figures of merit of Mg are given in Table 7.1. The detection limits using the atomic arid ionic lines of Mg are 0.2 pg and 1 pg, respectively. The difference in detection limits between magnesium atoms and magnesium ions lies in the fact that the magnesium atomic emission line is more intense. The data of S/N and S/B were measured with a deposition of 50 pg. Clearly, the signal to noise ratio and signal to background ratio of magnesium atomic emission are superior to those of magnesium ionic emission. The calibration curve of magnesium is presented in Fig.7.1. At a much lower concentration of magnesium, the 131 calibration curve becomes non-linear due to a possible spatial segregation and fluctuations in rf power during the atomization cycle. At high concentrations, the use of resonance line, Mg (I, 285.2 nm), could lead to self-absorption causing the slope of the log-log plot to deviate from linearity and approach its high concentration limiting value. However, non-linearity at high concentrations is expected and useful analytical results can still be obtained in the region where self-absorption occurs, at the expense of lower sensitivity. Table 7.1 Figures of merit for Mg at 80 W Species D. L. (pg) S/N S/B Dynamic range Mg(I) 0.2 37 21 3 orders (10 pg - 10 ng) Mg (II) 1 25 8 - 3 orders (10 pg - 10 ng) * D. L.: detection of limit. S/N: signal to noise ratio. S/B: signal to background ratio T -1.0 -0.5 0.0 Log(amount in ng) Fig. 7.1 Mg calibration curve 7.3.2 Relationship between Degree of Ionization and Ionization Potential The relationship between the degree of ionization and the ionization potential of the element is illustrated in Fig.7.2. As shown in the graph, the degree of ionization is a strong function of the ionization potential of the element. As the ionization potential increases, the 132 degree of ionization decreases. This observation is consistent with the Saha Equation. According to Equation 7.1, the degree of ionization, a, is related to exp(-E/kBT). As E{ increases, the term exp(-Ej/kBT) becomes smaller, and thus the term oc/(l-a) on the left of the equation decreases, leading to decreased degree of ionization. From Equation 7.1, the degree of ionization is also characterized by an ionization temperature and electron number density, and thus a knowledge of these two parameters could be very helpful to our further understanding of the plasma discharge. ion (%) 1 0 0 -8 0 -C r Mg Fe ionizat 6 0 - Cd ® "o 4 0 - ® Zn Degree 2 0 -0 -I 7.0 I I 7.5 8.0 Ionization potential I 8.5 (eV) I 9.0 Fig. 7.2 Degrees of ionization as a function of ionization potentials for elements - Cr (6.76 eV), Mg(7.64 eV), Fe(7.87 eV), Cd(8.99 eV) and Zn(9.39 eV). 7.3.3 Calculation of Ionization Temperature and Electron Number Density 133 In order to calculate the analyte ionization temperature in the FAPES source, we can use Equation 7.2 to plot log[oc/(l- a)] versus Ej (ionization potential) by employing data collected for Cr, Mg, Fe, Cd and Zn. The graph is shown in Fig. 7.3. From the slope an ionization temperature of 6976 K was obtained. This result is in good agreement with Sturgeon's earlier reported ionization temperature of 6025 K at 50 W rf power level [3], which was obtained using the same method. Using the ionization temperature obtained at 80W, we may further derive the electron number density by substituting the relevant data of one element into Equation 7.1. cn 0 . 0 -o B - 0 .5H 0.5H 1 . 0 H 7.0 7.5 8.0 E Ionization potential (eV) 8.5 9.0 Fig. 7.3 A plot for the calculation of ionization temperature 134 The partition functions of atoms and ions at the temperature of 6976 K are calculated based on Equation 7.3, and the data are listed in Table 7.2. The elements Cr, Mg, Fe, Cd and Zn are chosen as spectrometric species, and the electron number densities of the FAPES source at 80 W are presented in Table 7.3. Since these data vary slightly, a typical electron number density is calculated by averaging these data, yielding a value of (3±1) xlO15 cm"3. Table 7.2 Partition functions of several elements at 6976 K Element Cr Mg Fe Cd Zn 7 atom 15.21 1.10 36.17 1 1 Zjon 9.55 2 51.58 2 2 Table 7.3 The calculated electron number densities at 80 W Element Cr Mg Fe Cd Zn Average Electron number density (xl0 1 5crn 3) 2.56 4.11 4.45 1.66 1.64 3 ± 1 Alternatively, we may use plasma modeling to estimate the electron number density corresponding to an assumed local thermal equilibrium (LTE) temperature. Based on the ICP modeling procedures described by Blades et al. [4], we have the following relations: ne = nH e + (7.4) P = (nHe+nHe++ne)k8T (7.5) nH e + ne/nHe = 2ZHe+(T)/ZHe(T) (27tmkBT/h2)3/2exp(-EABT) (7.6) where P is the pressure of the plasma (1 atm). The partition functions of helium atom and helium ion are temperature-independent, and the values are 1 and 2 respectively. Here we 135 have two assumptions. One is that the FAPES plasma is one-element plasma without an addition of analyte, and the plasma consists only of helium gas with only neutral (nHe) and singly ionized (nHe+) helium species as well as electrons (ne) present. Another assumption is that the plasma is very close to LTE, and the gas kinetic temperature, the ionization temperature and the electron temperature are approximately same. By substituting the assumed different ionization temperatures into these equations, we can derive the electron number densities at different temperatures, shown in Table 7.4. Table 7.4 Calculated electron number density corresponding to the assumed LTE temperature LTE Temperature (K) Electron Number Density (cm"3) 3500 4000 4500 5000 5500 6000 7000 8000 9000 9300 9500 11000 4.45x102 8.03xl04 4.61X10 6 1.19x10s 1.71X10 9 1.58x10'° 5.31x10" 7.49xl012 5.94X1013 l.OlxlO1 4 1.42xl014 1.23X10'5 136 According to Table 7.3 and Table 7.4, the electron number densities of the plasma at an ionization temperature of about 7000 K are (3± l )x l0 1 5 cm"3 and 5.31x10" cm"3 respectively. There is a huge gap between these two values. The similarity for these two calculation methods is that we assume the existence of LTE in the FAPES source, and the difference between these two methods is that except for those species stated, no other species is present in the FAPES source for the second calculation method. Actually, based on the previous studies, it is likely that LTE does not exist in the FAPES source. In addition, there are also some other background species (such as N 2 + , OH, CO+) present in the plasma, and especially, an addition of analyte during the atomization cycle further complicates the species behavior in the plasma. Thus, these two values may not be a real indication of the electron number density at such operating conditions in the plasma. However, since the values in Table 7.3 are derived from the degrees of ionization which were obtained experimentally, we may reasonably expect that the real electron number density at 80 W is closer to the value (3±1) xlO 1 5 cm"3 than the value 5.31x10" cm"3. This assumption is supported by the data reported in a previous study [5]. Sturgeon employed a Langmuir probe to measure the electron number density, and a value of 8.1x 1013 cm"3 at 50 W was reported. We can also calculate the electron number density of an argon ICP using the equations like Equations 7.4-7.6, and the value is 2.5 x 1014 cm"3 at an ionization temperature of about 7000 K. If we consider the ionization energies of argon (15.8 eV) and helium (24.5 eV), it seems impossible for a helium plasma to reach an electron number density approximating the value of the argon plasma. However, it may be reasonable if we think about the sources of electrons. The primary source of the electrons in the argon ICP is the ionization of argon, but the electrons in the FAPES source may come from several sources. 137 In addition to helium ionization, secondary electrons may be emitted when fast electrons or energetic particles collide with the center electrode, and thermionic electron emission from the graphite furnace wall may occur during the atomization cycle. As a result, these extra sources of electrons contribute to a greater increase in electron number density and perhaps dominate the electron number density. 7.4 Summary Some analytical figures of merit of magnesium have been studied and used to characterize the analytical performance of the newly built FAPES source. Compared with other FAPES sources previously built in this lab [6-8], this FAPES source has lower detection limits, which are 2-3 times better, and comparable dynamic ranges. LeBlanc has also studied analyte ionization in the FAPES source and reported about 30% for the degree of ionization of magnesium in his Ph.D. dissertation [8]. In this FAPES source, analyte ionization efficiency has been improved greatly, and, for example, the degree of ionization of magnesium can reach more than 70% under the optimum operating conditions. An analyte ionization temperature at an rf power level of 80 W has been derived from the Saha equation, and the temperature is approximately 7000 K. Combined with the results of the plasma temperatures reported in Chapter 3, we may conclude that there is no LTE present in the FAPES source. The electron number densities in the FAPES source have been derived experimentally and calculated by plasma modeling, and the difference between these two values suggests that unlike in the argon ICP, the ionization of helium gas is probably not the primary source of the electrons in the FAPES source. Instead, secondary electron emission from the center electrode and thermionic electron emission from the graphite tube wall may be the major sources of electrons. 138 7.5 References 1. L. de Galan, R. Smith and J.D. Winefordner, Spectrochim. Acta 23B: p. 521 (1968). 2. J.M. Mermet, Anal. Chim. Acta 250, 85 (1991). 3. R.E. Sturgeon and R. Guevremont, / . Anal. At. Spectrom.13: p. 229(1998). 4. M.W. Blades, B.L. Caughlin, Z.H. Walker and L.L. Burton, Prog. Analyt. Spectrosc.lO: p. 57 (1987). 5. R.E. Sturgeon, V.T. Luong and R.K. Marcus, Twenty-First Annual Conference of the Federation of Analytical Chemistry and Spectroscopy Societies paper No. 272 (St. Louis, Mo, USA, 1994). 6. D.C. Liang and M.W. Blades, Spectrochim. Acta 44B: p. 1059 (1989). 7. T.D. Hettipathirana and M.W. Blades, Spectrochim. Acta, 47B: p. 493 (1992). 8. C.W. LeBlanc, Ph.D. Dissertation, University of British Columbia (1996). 9. P.W.J.M. Boumans, Inductively Coupled Plasma Emission Spectroscopy Part 2 (John Wiley & Sons, USA, 1987). 139 Chapter 8 CONCLUSIONS 8.1 Summary The characteristics of analyte excitation and ionization in a newly-built FAPES source have been investigated and are presented in this thesis. Although the primary objective of this study was to characterize analyte ionization and to seek ways to improve the ionization efficiency, a number of studies of the fundamental properties and operating characteristics of the plasma were carried out leading to a better understanding of the plasma source characteristics. This chapter will briefly summarize these studies and then provide some suggestions for future research. The plasma background emission in the FAPES source is highly structured and mainly dominated by some molecular species - CO + , OH, N 2 , N 2 + and He. The existence of the molecular ion emission spectra of N 2 + and CO + indicates that, under certain circumstances, the FAPES source is a powerful ionization source. Since the potential energies of He2+ and metastable energies of He"1 are close to the energy levels of low-lying excited electronic states of N 2 + (B2Z+U) and CO + (B 2 £ + ) , in addition to electron impact ionization process, ionization by charge transfer and Penning ionization may occur and contribute to the production of these molecular ion species. In addition, the Penning ionization process is more favorable due to the production of a large population of metastable helium atoms (3S, and 'S0) resulting from helium atomic emission such as He(I)-388.87 nm (3P°, 2 -^S,) and He(I)-501.57 nm ('P 0 , -^) . During the atomization cycle, the temporal emission behaviors of He, N 2 + , OH and CO + exhibit different characteristics, corresponding to their respective excitation and ionization mechanisms. 140 Plasma temperatures of major background species (the rotational temperatures of OH and N 2 + as well as the excitation temperature of He) are enhanced during the atomization cycle, suggesting that thermal energy from an atomization cycle may participate in analyte excitation and ionization processes. The effects of rf power, gas flow rate, and atomization temperature on plasma temperatures demonstrate that these operating conditions can influence the robustness of the plasma source, and an optimization of these conditions can be used to increase plasma temperatures and thus improve the degrees of ionization of analytes. Using the current experimental set-up, increasing the center electrode diameter causes a decrease in plasma temperatures, leading to a reduction in analytical performance of the FAPES source as illustrated in Chapter 5. An investigation of the effect of rf power on the excitation temperature of Pb suggests that an improvement of ionization capability of the plasma can be better achieved by the application of a high rf power. An examination of the temporal profiles of Mg, Cd, Fe, Zn, and Cr during an atomization cycle indicates that the appearance temperature determines the temporal positions of the maxima of atomic emission and ionic emission. The discussion of double atomic emission peaks of magnesium at an high atomization temperature demonstrates that many factors such as appearance temperature, carbide formation and atomization mechanisms may affect analyte excitation and ionization processes. Background species such as CO + can influence analyte atomization processes by participating in some gaseous atomization reaction. The temporal emission spectra of Mg and Cd show that a temporal variation in the emission signals from the background species occurs during an atomization cycle, and if an analyte species emission line coincides with one of these background species, the S/B and hence the detection limit will be impacted. 141 Analytical characteristics of analyte atomic emission and ionic emission as a function of rf power, gas flow rate, and atomization temperature have been studied. An increase in atomization temperature can facilitate the early appearance of atomic emission and ionic emission during an atomization cycle, and may alter ionic emission temporal profiles. In general, a sufficiently high atomization temperature is required for complete analyte atomization, and a further temperature increase doesn't improve atomic and ionic emission intensities. Increasing the forward power to the plasma results in an increase in the temperature of the center electrode and a growth in the volume of the plasma within the tube, and thus analyte atomic and ionic emission intensities are enhanced. In addition, the temporal ionic emission profiles of analyte are basically not affected by an rf power change. However, the maximum rf power applied to the center electrode can be limited by the nature of the analyte and the physical characteristics of the FAPES source. Since helium transports the analyte vapor, maintains the plasma, purges the graphite tube, and prevents the graphite tube from being oxidized, a low gas flow rate is insufficient to support these operations. However, a high gas flow rate can cool the plasma and reduce the residence time of the analyte, leading to reduced emission intensities. Therefore, an optimum gas flow rate could be found for the maximization of analyte atomic and ionic emission intensities. Moreover, since the stop-flow mode during the atomization cycle may maintain the plasma temperatures and eliminate the analyte loss caused by a forced gas flow expulsion, this type of operating condition may enhance analyte atomic and ionic emission intensities. The effects of rf power, gas flow rate, and atomization temperature on the degree of ionization of analyte have been studied. All of these operating conditions can influence analytical performance by affecting one or more physical processes such as atomization, 142 current/voltage characteristics, excitation process, and ionization process. Since an increase in rf power can increase the plasma volume and electron number density, application of a high rf power is the best way to enhance analyte ionization although, the maximum rf power applied to the center electrode may be limited by the structure of FAPES source and analyte atomization characteristics. An appropriate atomization temperature is required for the maximization of analyte ionization, however, a much higher atomization temperature may cause a lot of thermionic emission from the graphite furnace wall and thus alter the impedance characteristics of the plasma interfering the normal operation. Due to the facts that a sufficient gas flow rate is required for the maintenance of the plasma and for analyte atomization, a lower gas flow rate is preferred for longer residence time and temperature stabilization, and an optimization of gas flow rate could help promote ionization efficiency. Although the stop-flow mode may increase analytical sensitivity for some elements, such as Fe, Cr, and Mg, it can't play a big role in the improvement of analyte ionization capability. A variety of center electrodes with different physical dimensions have been used to modify the impedance of the FAPES source. The experimental results show that an increase in the center electrode size decreases the voltage drop in the zone close to the center electrode and thus reduces the ionization capability of this zone, leading to lower degrees of ionization of analytes. Under the experimental condition of maintaining a constant power density, a shorter, larger electrode does not improve the degree of ionization of analyte due to the same reasons as it was mentioned above. The measurement of the effect of the center electrode length on analyte ionization demonstrates that a complete filling of the center electrode inside the full graphite tube length may maximize the degrees of ionization of 143 analytes, and a failure of filling the graphite tube length can cause a reduction in analyte ionization capability. The studies of different magnesium matrices show that analyte matrix may affect analyte vaporization and atomization processes, and atomization processes involving more reactions decrease the atomization efficiency. Analyte matrices may make vaporization and atomization occur through alternative processes, and as a result influence analytical performance. As is observed with GFAAS, chemical modification of the FAPES source by palladium may prevent the formation of interstitial compounds, produce activation sites, and alter atomization processes, leading to an increase in atomization efficiency. An improvement of the degree of ionization of analyte can be realized with an addition of a minute amount of palladium modifier. In the presence of large masses of palladium modifier, power loss, detuning of the plasma, and the recombination of analyte ion with electrons could cause a serious side effect on the plasma source. Magnesium was chosen the spectrometric species to characterize analytical figures of merit in the FAPES source. Compared with previous sources, this FAPES source has comparable detection limits, good S/N and S/B, comparable dynamic range, and improved ionization efficiency. The ionization temperature was about 7000 K at 80 W power level. Although helium has a high ionization potential, the calculation of electron number density shows that the electron number density in the FAPES source can reach approximately 1015 cm"3, which is not as low as we expect. The comparison of the electron number densities derived from two different methods suggests that the production of electrons sputtered from the center electrode and emitted from the graphite furnace wall may be the major sources of electrons in the FAPES source. 144 In sum, the FAPES source has considerable potential as being an ion source for an elemental mass analyzer. In addition, the compact nature of the FAPES source with the elimination of the void volume is also convenient for interfacing with a mass analyzer. Therefore, the FAPES source deserves further study and future development. 8.2 Suggestions for Future Research One of the major tasks in the future is to further improve the ionization capability of the FAPES source. Basically, we may concentrate on the following aspects: optimization of the physical dimensions of the FAPES source, enhancement of the voltage drops of plasma sheaths, and improvement of analytical characteristics of the plasma. Since analyte ionization processes mainly occur in the negative glow region, some factors relating to expanding the negative glow region should be more thoroughly investigated. For example, applying a negative dc bias voltage on the center electrode may make the electrode behave as the cathode for a greater period of time during a given half cycle, and the plasma will become more stable in that this voltage control affords robustness of plasma temperatures during atomization transient [1]. A better configuration of the FAPES source may be achieved by employing a shorter and larger graphite tube with a complete filling of a larger diameter electrode, which may carry more power, maintain electrical properties and increase plasma robustness. A higher generator frequency can increase plasma temperatures [2], and thus an appropriate frequency may potentially increase ionization capability of the FAPES source. The use of pulsed rf power could increase instantaneous ionization capability and improve signal to noise ratio [3-4], but we must take some measures to speed up the auto-tuning capabilities of the impedance matching networks because the impedance changes of the 145 FAPES source occur too rapidly for these mechanical systems to follow. A solution is to use a Free Running Oscillator (FRO) to track plasma impedance changes for which is compensated by a shift in the plasma excitation frequency regulated by the oscillator [5]. The addition of Ar to He may facilitate collisional exchange of internal energy between excited states of Ar and He and thus increase ionization temperatures by about 1000 K at 50 W [6], and therefore investigation of the optimum composition of plasma gas is expected to further improve ionization efficiency. Another major task is to look further into the viability of this FAPES source as ion source for mass spectrometric analysis. Since ionic signals in the FAPES source are time-dependent, a time-of-fiight mass analyzer is probably optimum for the detection of these transient signals. A practical interface providing high ion extraction efficiency and ion transmission efficiency between the FAPES source and the mass analyzer needs to be designed, constructed, and tested. 146 8.3 References 1. R.E. Sturgeon, S.N. Willie, V. Luong and R. K. Marcus, Spectrochim. Acta 48B: p. 893 (1993). 2. R.E. Sturgeon, S.N. Willie, V. Luong and J.G. Dunn, Appl. Spectrosc. 45: p. 1413 (1991). 3. Y. Su, Z. Zhou, P. Yang, X Wang and B. Huang, Spectrochim. Acta 52B: p. 633 (1997). 4. S. Ashida, M. R. Shim and M. A. Lieberman, J. Vac. Sci. Technol. A 14(2): p391 (1996) 5. I. L. Turner, Development ofAn Improved Plasma rf System for ICP-MS (Instruments at work, Varian Australia Pty Ltd, 1996). 6. F.S. Sun and R.E. Sturgeon, Spectrochim. Acta 54B: p. 2121 (1999). 

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.831.1-0061282/manifest

Comment

Related Items