Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Development and characterization of atmospheric pressure radio frequency capacitively coupled plasmas… Liang, Dong Cuan 1990

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

Item Metadata

Download

Media
831-UBC_1990_A1 L52.pdf [ 6.59MB ]
Metadata
JSON: 831-1.0059715.json
JSON-LD: 831-1.0059715-ld.json
RDF/XML (Pretty): 831-1.0059715-rdf.xml
RDF/JSON: 831-1.0059715-rdf.json
Turtle: 831-1.0059715-turtle.txt
N-Triples: 831-1.0059715-rdf-ntriples.txt
Original Record: 831-1.0059715-source.json
Full Text
831-1.0059715-fulltext.txt
Citation
831-1.0059715.ris

Full Text

DEVELOPMENT AND CHARACTERIZATION OF ATMOSPHERIC PRESSURE RADIO FREQUENCY C A P A C m V E L Y COUPLED PLASMAS FOR ANALYT ICAL SPECTROSCOPY By DONG CU AN LIANG M.Sc, Dalhousie University, 1986 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE STUDIES Department of Chemistry We accept this thesis as confonning to the required standard THE UNIVERSITY OF BRITISH COLUMBIA June 1990 © Dong Cuan Liang, 1990 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 ^ l A Q ^ ^ f ' r The University of British Columbia Vancouver, Canada DE-6 (2788) ABSTRACT An atmospheric pressure radio frequency capacitively coupled plasma (CCP) has been developed and characterized for applications in atomic emission spectrometry (AES), atomic absorption spectrometry (AAS) and gas chromatography (GC). The CCP torch was initially designed as an atom reservoir for carrying out elemental analysis using atomic absorption. Functionally, the device consists of two parts, the CCP discharge tube and the tantalum strip electrothermal vaporization sample introduction system. The torch design provides for very effective energy transfer from the power supply to the plasma by capacitive coupling. Therefore, the plasma can be generated at atmospheric pressure with a flexible geometry. The plasma can be operated at very low rf input powers (30-600 W) enabling optimal conditions for atom resonance line absorption measurements. Absorption by the analyte takes place within the plasma discharge which is characterized by a long path length (20 cm) and low support gas flow rate (0.2 L/Min). Both of these characteristics ensure a relatively long residence time. The device exhibits linear calibration plots and provides sensitivities in the range of 3.5-40 pg. A preliminary measurement gave a Fe I excitation temperature of approximately 4000 K for the discharge. At this temperature, potential chemical interferences are likely to be minimal. Chemical interferences for Fe, A l , As, Ca, Co, Cd, L i , Mo and Sr were negligible in the determination of silver. Chloride interference, which is prevalent in GF-AAS, was not found. The amount of Ag found in a SMR#1643b (NIST) water sample was 9.5 ± 0.5 ng/g which fell in the certified range of 9.8 ± 0.8 ng/g. Spikes of 30 ng/g and 60 ng/g of silver were added to the SRM and recoveries were found to be in a range from 105 % to 96.2 %. The RSD obtained for 7 replicates of 270 pg silver was 4.6 %. The results for the CCP AES are even more promising. The interferences of thirteen elements are negligible in the determination of silver. The chloride interference was not found. The detection limits for Ag, Cd, L i , Sb and B are 0.7, 0.7, 2, 80 and 400 pg i i respectively. The amount of silver found in a SRM#1643b (NIST) water sample was 9.3 ± 0.5 ng/g which also fell in the certified range of 9.8 ±0.8 ng/g. Spikes of 30 ng/g and 60 ng/g of silver were added into the SRM#1643b (NIST) samples; the recoveries were found to range from 97 % to 104 %. The RSD obtained for 7 analyses of 270 pg silver were 1.5 % for CCP-AES. It was also found that the signal to noise ratios (S/N) are higher in the AES mode than those in the AAS mode in the same CCP atomizer. In order to exploit advantages inherent in both GF-AAS and I CP-AES, an atmospheric pressure capacitively coupled plasma sustained inside a graphite furnace was developed. This source combines the high efficiency of atomization in furnaces and the high efficiency of the excitation in atmospheric pressure plasmas. In general, plasma sources are able to effectively excite high-lying excited states for most metals and non-metals and can also ionize vaporized elements. Therefore the possibility exists of using non-resonance lines to avoid the effects of self-absorption at high analyte concentrations. Ion lines may also be used in cases where they provide better sensitivity or freedom from spectral interferences. This source also offers the ability to independently optimize vaporization and excitation. However, the most important aspect of this new source is that it can be used for simultaneous, multielement determinations of small sized samples in a graphite furnace atomizer, a design which has been proven to be effective over many years of use. Preliminary quantitative characteristics of this new atmospheric pressure plasma emission source have been studied. The detection limit for Ag of 0.3 pg is lower than the value of 0.4 pg reported for GF-AAS. Variants of the CCP, including a gas chromatography (GC) detector, combinations of laser ablation - CCP, rf sputtering - CCP direct solid analysis, and its application as an intense spectral lamp have been developed and are reported in this dissertation. iii TABLE OF CONTENTS Page Abstract ii Table of contents v List of tables xii List of figures xiii List of patents arising from this study xvii List of publications arising from this study xix List of conference presentations arising from this study xxi Acknowledgements xxiii Chapter 1: INTRODUCTION 1 1.1 Overview and goals 2 1.2 Survey on atomizers and sources for analytical spectroscopy 2 1.2.1 Flames 2 1.2.2 Arcs and sparks 4 1.2.3 Low-pressure discharges 4 1.2.4 Plasmas 6 1.2.4.1 Inductively coupled plasmas 7 1.2.4.2 Microwave-inducted plasmas 8 1.2.4.3 Direct Current plasmas (DCP) 8 1.2.5 Electrothermal atomizers 8 1.2.6 Chemical vapor generators 9 1.2.7 Lasers 10 iv 1.3 Capacitively Coupled Plasmas (CCP) 11 1.3.1 Historical 11 1.3.2 Principles of plasma generation 18 1.3.3 Applications 19 1.4 General considerations of plasma source design 21 1.4.1 Temperature 21 1.4.2 Geometry and transport efficiency 21 1.4.3 Pressure 22 1.4.4 Effect of chemical reactions 22 1.4.5 Background 22 1.4.6 Others 23 References 24 Chapter 2 : ATMOSPHERIC PRESSURE CAPACITIVELY COUPLED PLASMA ATOMIZER FOR ATOMIC ABSORPTION SPECTROMETRY 29 2.1. Introduction 30 2.2 Experimental 32 2.1.1 Plasma torch 32 2.1.2 Experimental facilities 32 2.1.3 Analytical procedure 35 2.3 Results and discussion 38 2.3.1 The CCP and its characteristics 38 2.3.2 Background emission spectra in argon and helium CCP 40 2.3.3 Transmitance of argon and helium CCP 40 2.3.4 Effect of rf input power 40 v 2.3.5 Effect of plasma support-gas flow rate 45 2.3.6 CMbration curve 45 2.3.7 Sensitivities of absorption 45 2.3.8 Absorption enhancement effect and interferences 47 2.3.9 Precision and SRM sample analysis and analyte recovery 55 2.3.10 Iron excitation temperature 56 2.4 Summary 58 References 59 Chapter 3 : ATMOSPHERIC PRESSURE CAPACITIVELY COUPLED PLASMA AS A SOURCE FOR ATOMIC EMISSION SPECTROMETRY 60 3.1. Introduction 61 3.2 Experimental 63 3.2.1 Modification of CCP torch 63 3.2.2 Equipment and set-up 65 3.2.3 Analytical procedure 65 3.2.4 Standard solutions 65 3.3 Results and discussion 68 3.3.1 Effect of rf input power 68 3.3.2 Effect of plasma gas flow rate 73 3.3.3 Emission enhancement effect and interferences 77 3.3.4 Detection limits 83 3.3.5 Precision and signal-to-noise ratio in AES and AAS modes 83 3.3.6 Calibration curves and analyte recovery 85 v i 3.4 Summary References 87 88 Chapter 4: ATMOSPHERIC PRESSURE CAPACITIVELY COUPLED PLASMA SPECTRAL LAMP AND SOURCE FOR THE DIRECT ANALYSIS OF CONDUCTING SOLID SAMPLES - APPLICATION OF RF SPUTTERING 89 4.1. Introduction 90 4.2 Experimental 93 4.2.1 Lamp and source design 93 4.2.2 Equipment and facilities 96 4.2.3 Method for comparison of intensities of CCP spectral lamp and HCL 96 4.2.4 Analytical procedure 96 4.3 Results and discussion 98 4.3.1 Description of CCP spectral lamp and CCP direct solid analysis source 98 4.3.2 Effect of rf input power 98 4.3.3 Comparison of intensities of CCP spectral lamp and HCL 100 4.3.4 Ion spectra in the CCP lamps 103 4.3.5 Atmospheric pressure rf sputtering direct solid sampling 103 4.3.6 Calibration curves 108 4.3.7 Iron excitation temperature 108 4.4 Summary 110 References 111 vii Chapter 5 : ANALYSIS OF SOLID SAMPLES USING LASER ABLATION SAMPLING - ATMOSPHERIC PRESSURE CAPACITIVELY COUPLED PLASMA EMISSION SPECTROMETRY 112 5.1. Introduction 113 5.2 Experimental 115 5.2.1 Laser ablation cell 115 5.2.2 Equipmental facilities 115 5.2.3 Analytical procedures 117 5.3 Results and discussion 118 5.3.1 Effect of rf input power 118 5.3.2 Effect of plasma gas flow rate 118 5.3.3 Laser energy and focusing 121 5.3.4 Calibration curve 121 5.3.5 Detection limits 123 5.3.6 Accuracy 123 5.4 Summary 125 References 126 Chapter 6 : A N ATMOSPHERIC PRESSURE C A P A C n T V E L Y COUPLED PLASMA FORMED INSIDE A GRAPHITE FURNACE AS A SOURCE FOR ATOMIC EMISSION SPECTROSCOPY 128 6.1. Introduction 138 6.2 Experimental 132 6.2.1 APF-CCP source design 132 6.2.2 Free-run rf power supply 132 viii 6.2.3 Equipment and set-up 134 6.2.4 Analytical procedure 136 6.3 Results and discussion 137 6.3.1 Description of argon and helium plasmas 137 6.3.2 Background emission 137 6.3.3 Role of the CCP inside the graphite furnace 138 6.3.4 Effect of plasma support-gas flow rate 138 6.3.5 Deration limits 141 6.3.6 Calibration curves 141 6.3.7 Precision 141 6.3.8 Iron excitation temperature 141 6.3.9 Multielement capability 143 6.4 Summary 145 References 146 Chapter 7 : A CAPACITIVELY COUPLED PLASMA DETECTOR FOR GAS CHROMATOGRAPHY 147 7.1. Introduction 148 7.2 Experimental 150 7.2.1 The CCP GC detector design 150 7.2.2 Equipment and set-up 154 7.2.3 Analytical procedure 156 7.3 Results and discussion 157 7.3.1 Description of the CCP GC detector 157 7.3.2 Spectra of eluents 157 ix 7.3.3 Analytical performance 160 7.4 Summary 161 References 162 Chapter 8 : CONCLUDING REMARKS 163 x LIST OF TABLES Table 2.1 Experimental facilities and operating conditions for CCP-AAS 36 Table 2.2 Sensitivities and wavelengths for Ag, Cd, Cu, L i and Sb. 47 Table 2.3 Interference effects on atomic absorption for silver 55 Table 2.4 Detennination of silver in water samples by CCP-AAS (ng/g) 56 Table 3.1 Experimental facilities and operating conditions for CCP-AES 66 Table 3.2 Characteristic masses and detection limits for Ag, Cd, Cu, L i , Sb and B using CCP-AES, CCP-AAS, and GF-AAS. 83 Table 3.3 Determination of silver in water samples by CCP-AES (ng/g) 86 Table 4.1 Intensity ratio of resonance line of CCP lamp and HCL 102 Table 5.1 Quantitative analysis of chromium in steel samples by laser ablation-CCP (ng/g) 124 Table 6.1 Experimental facilities and operating conditions for APF-CCP 135 Table 7.1 Experimental facilities and operating conditions for GC detector 154 xi L I S T O F F I G U R E S Figure 1.1. Schematic diagram of the CCP with the inductance coil and sample-supplying procedure 12 Figure 1.2. Schematic diagram of the low power CCP torch 13 Figure 1.3. Schematic diagram of the A.RX. CMP source with nebulizer 17 Figure 2.1. Schematic diagram of the capacitively coupled plasma and sampling system 33 Figure 2.2. Block diagram of me experimental arrangement 34 Figure 2.3. Absorption and Emission signals for 0.1 ng Ag using the 328.1 nm line 39 Figure 2.4. Background emission spectrum of the CCP using argon as a support gas Conditions: Power - 200 W, support gas flow rate - 0.6 L/Min 41 Figure 2.5. Background emission spectrum of the CCP using helium as a support gas Conditions: Power - 200 W, support gas flow rate - 0.6 L/Min 42 Figure 2.6. Transmittance of the argon (squares) and helium (diamonds) CCP as a function of wavelength. 43 Figure 2.7. Effect of RF input power on Ag 1328.1 nm absorbance. 44 Figure 2.8. Effect of support gas flow rate on Ag 1328.1 nm absorbance. 46 Figure 2.9. Enhancement of sodium on Ag 1328.1 nm absorbance 48 Figure 2.10. Enhancement of lithium on Ag 1328.1 nm absorbance 49 Figure 2.11. Enhancement of gallium on Ag 1328.1 nm absorbance 50 Figure 2.12. Schematic representation of the enhancement on Ag 1328.1 nm absorbance 51 Figure 2.13. Influence of lithium on the effect of rf input power on Ag 1328.1 nm absorbance 53 xii Figure 2.14. Influence of lithium on the effect of support gas flow rate on Ag 1328.1 nm absorbance 54 Figure 3.1. Schematic diagram of the modified capaciuvely coupled plasma torch (a) Torch with fin structure, (b) Three round tube torch, (c) Square tube torch. 64 Figure 3.2. Effect of rf input power on Ag 1328.1 nm absorbance and emission intensity (a) Emission, (b) Absorption. 69 Figure 3.3. Effect of rf input power on B 1249.8 nm emission intensity 70 Figure 3.4. Effect of rf input power on background emission intensity (a) 328.1 nm, (b) 249.8 nm, (c) Band head at 247.9 nm. 72 Figure 3.5. Effect of support gas flow rate on Ag 1328.1 nm absorbance and emission intensity (a) Emission, (b) Absorption. 74 Figure 3.6. Effect of support gas flow rate on B1249.8 nm emission intensity 75 Figure 3.7. Effect of support gas flow rate on background emission intensity and peak to peak noise for boron measurement (a) Background intensity, (b) P-P Noise. 76 Figure 3.8. Enhancements of Ag 1328.1 nm due to sodium, lithium, calcium and gallium 78 Figure 3.9. Influence of ET£ on the effect of rf input power on Ag 1328.1 nm emission intensity (a) Without L i , (b) With L i . 79 Figure 3.10. Influence of EIE on the effect of support gas flow rate on Ag 1328.1 nm emission 81 Figure 3.11. Interference effect in CCP-AES 82 x i i i Figure 3.12. Absorption and Emission signals for 100 ng boron using the 249.8 nm line 84 Figure 4.1. Schematic diagram of the CCP spectral lamp 94 Figure 4.2. Schematic diagram of the CCP torch used for direct solids analysis 95 Figure 4.3. Effect of rf power on Zn 1213.9 nm emission intensity 99 Figure 4.4. Intensities of Zn 1213.9 nm in (1) HCL and (2) CCP spectral lamp (1) Perkin-Elmer Zn HCL :dc 15 mA, PMT 900 V, recorder 1 V/F.S.; (2) CCP spectral lamp with brass central electrode (Cu ~ 65 %, Zn ~ 35 % ): rf power 60 W, Ar flow rate 1.3 L/Min, PMT 500V, recorder 1 V/F.S.. 104 Figure 4.5. Manganese emission spectral from HCL and CCP spectral lamp HCL.: Perkin-Elmer Mn HCL, dc 12 mA, PMT 900V, Recorder 1 V/F.S. CCP spectral lamp : rf power 200 W, argon flow rate 0.7 L/Min, PMT 500 V, recorder 5 V/F.S. 105 Figure 4.6. Spectrum of the NBS 622 steel sample from the CCP spectral lamp 106 Figure 4.7. Calibration curve for manganese in steel 109 Figure 5.1. Schematic diagram of the laser ablation - CCP system 116 Figure 5.2. Effect of rf input power on Cr 1357.9 nm emission intensity 119 Figure 5.3. Effect of argon gas flow rate on Cr 1357.9 nm emission intensity 120 Figure 5.4. Analytical calibration curve of chromium 122 Figure 6.1. Schematic diagram of the APF-CCP source 133 Figure 6.2. Spectra of copper and zinc between 322 and 338 nm from the APF-CCP source 139 Figure 6.3. Comparison of Zn 1334.50 nm from (a) APF-CCP source at a dark red furnace temperature (approx. 800 C), (b) CFAAS at the same xiv furnace temperature as in (a), (c) Same as (b) except running at maximum furnace temperature (approx. 2800 °C) 140 Figure 6.4. Emission intensity of Cu 1324.75 nm as a function of the plasma support gas flow rate. 142 Figure 6.5. Multielement spectra from a helium APF-CCP source 144 Figure 7.1. Schematic diagram of the experimental system used to test the CCP as a GC detector. 151 Figure 7.2. Schematic diagram of the interface of GC column with the plasma torch 152 Figure 7.3. Schematic diagram of the CCP torches, (a) concentric geometry, and (b) parallel geometry 153 Figure 7.4. CCP emission spectrum between 270 and 306 nm. Power: 100 W, Frequency: 200 KHz, Carrier gas flow: 30 mL/Min. (a) Spectrum of (Q^^Sn detected using the concentric geometry Ar CCP, and (b) background spectrum. 158 Figure 7.5. (a) Gas chromatograms for three replicate injections of 150 pg of iodomethane {CH3I} into the GC. (b) Gas chromatogram of 0.1 mL of 0.1 ppm di-iodopropane {(CH2)3l2} in hexane detected using the CCP. 159 xv LIST OF PATENTS ARISING F R O M THIS STUDY 1. "Improved Atmospheric Pressure Capacitively Coupled Plasma Atomizer for Atomic Absorption and Emission Spectroscopy" (i.e. T-Torch). People's Republic of China: Filing Date - December 24,1987 Serial No. 87108198 Publish Date - June 21,1989 2. "Furnace Atomization Atmospheric Pressure Capacitively Coupled Excitation Source" (i.e. FAPES device). United States: Filing Date - May 19,1989 United States: Filing Date - July 11,1989 Serial No. 378263 Canada: Filing Date - December 11,1987 Serial No. 554178 Serial No. 354511 Patent Cooperation Filing Date - May 17,1990 Treaty (PCT): Serial No. PCT/CA90/00160 xvi 3. "A Capacitively Coupled Plasma Detector for Gas Chromatography" (i.e. GC Detector). United States: Filing Date - May 19,1989 Serial No. 354150 4. "Atmospheric Pressure Capacitively Coupled Spectral Lamp". United States: Filing Date - August 31,1989 Serial No. 401181 xvii L I S T O F P U B L I C A T I O N S A R I S I N G F R O M T H I S S T U D Y [9] "Interferences in determination of silver by atmospheric pressure capacitively coupled plasma atomic absorption spectrometry", Dong Cuan Liang and Michael W. Blades, in preparation. [8] "Atmospheric pressure capaciuvely coupled plasma for atomic emission spectrometry", Dong C Liang and M . W. Blades, Anal. Chem. in preparation. [7] "Application of weakly ionized plasma for materials sampling and analysis", M.W. Blades, D. Huang, P. Banks, Dong C. Liang, C. Gill and C. LeBlanc, IEEE Transactions on Plasma Science, in press. [6] "Analysis of solid samples using laser ablation sampling - atmospheric pressure capaciuvely coupled plasma emission spectrometry", Dong C. Liang and M.W. Blades, J. Anal. At. Spectrom, in press. [5] "Analytical characteristics of furnace atomization plasma excitation spectrometry (FAPES)", D. Lee Smith, Dong C. Liang, Doug Steel and M . W. Blades, Spectrochim. Acta 45B, 493 (1990). [4] "An atmospheric pressure capacitively coupled plasma formed inside a graphite furnace as a source for atomic emission spectroscopy", Dong C. Liang and M . W. Blades, Spectrochim. Acta 44B, 1059 (1989). [3] "A capacitively coupled plasma detector for gas chromatography", Degui Huang, Dong C. Liang and M.W. Blades, / . Anal. At. Spectrom, 4,789 (1989) [2] "Atmospheric pressure capacitively coupled plasma spectral lamp and source for the direct analysis of conducting solid samples", Dong C. Liang and M . W. Blades, Spectrochim. Acta 44B, p.1049 (1989). x v i i i "Atmospheric pressure capacitively coupled plasma atomizer for atomic absorption spectrometry" Dong Cuan Liang and Michael W. Blades, Anal. Chem., 60,27 (1988). xix LIST OF C O N F E R E N C E PRESENTATIONS ARISING F R O M THIS STUDY [14] "A radio-frequency capacitively coupled plasma formed at atmospheric pressure inside a graphite furnace", Dong C. Liang, D. Lee Smith, Doug Steel and M . W. Blades, Winter Conference on Plasma Spectrochemistry, January, 1990, St. Petersburg, Florida. [ 13] "Atmospheric pressure plasma excited atomic emission inside a graphite furnace", Dong C. Liang and M.W. Blades, 16th Federation of Analytical Chemistry and Spectroscopy Societies (FACSS), October, 1989, Chicago. [12] "Atmospheric pressure RF sputtering source for direct analysis of solid and liquid samples", Dong C. Liang and M.W. Blades, 16th Federation of Analytical Chemistry and Spectroscopy Societies (FACSS), October, 1989, Chicago. [11] "A novel capacitively coupled plasma detector for gas chromatography", Degui Huang, Dong C. Liang and M.W. Blades, 16th Federation of Analytical Chemistry and Spectroscopy Societies (FACSS), October, 1989, Chicago. [10] "Capacitively coupled plasma source - from AE to GC", M.W. Blades, D.C. Liang and D. Huang, 198th National American Chemical Society Meeting, September, 1989, Miami Beach. [9] "Plasma source development for atomic emission spectrometry", M.W. Blades, D.C. Liang, P. Banks and D. Huang, 72nd Canadian Chemical Conference and Exhibition, June, 1989, Victoria, B.C. [8] "An evaluation of a variety of sampling strategies for capacitively coupled plasma atomic emission spectrometry", M.W. Blades and Dong C. Liang, 15th Federation of Analytical Chemistry and Spectroscopy Societies (FACSS), October, 1988, Boston. xx [7] "Analytical and fundamental characteristics of a novel atmospheric pressure capacitively coupled plasma source", M.W. Blades and Dong C. Liang, 15th Federation of Analytical Chemistry and Spectroscopy Societies (FACSS), October, 1988, Boston. [6] "Can a plasma be used as an atomizer for atomic absorption spectroscopy?" Dong C. Liang and M.W. Blades, 39th Pittsburgh Conference & Exposition, Paper #1140, February, 1988, New Orleans. [5] "Considerations in the design of a capacitively coupled plasma torch for atomic absorption and emission spectrometry1', Dong C. Liang and M.W. Blades, 39th Pittsburgh Conference & Exposition, Paper #415, February, 1988, New Orleans. [4] "A novel capacitively coupled plasma for atomic absorption and emission spectrometry - characteristic and performance", Dong C. Liang and M.W. Blades, 1988 Winter Conference on Plasma Chemistry, January 1988, San Diego, California. [3] "New directions and new sources in plasma fundamental studies", M.W. Blades, L.L. Burton and D.C. Liang, 29th Rocky Mountain Conference, August, 1987, Denver, Colorado. [2] "A novel atmospheric pressure RF plasma atomizer for atomic absorption spectroscopy", M.W. Blades and Dong C. Liang, 31st International Congress of Pure and Applied Chemistry, July, 1987, Sofia, Bulgaria. [ 1] "A novel atmospheric pressure RF plasma atomizer for atomic absorption spectroscopy", Dong C. Liang and M.W. Blades, 25th Colloquium Spectroscopicum Internationale, June, 1987, Toronto xxi To my wife Kora and my daughters Fay and Sophia Thank you for your patience while waiting for the completion of my graduate work I love you xxii ACKNOWLEDGEMENTS I would like to thank my supervisor Mike Blades for his support and guidance during my stay in his research group. Thanks as well to all my colleagues in the group for their encouragement and valuable discussions. Special thanks to Lyle Burton, Peter Banks, Degui Huang, D. Lee Smith and Doug Steel for their cooperation and assistance. The experimental apparatus used in this study was constructed with the assistance of electronic shop, machine shop and glass shop in the Department of Chemistry, University of British Columbia. Martin Carlisle of the electronic shop, Brian Snapkauskas and Mark Vagg of the machine shop, and Steve Takacs and Steve Rak of the glass shop deserve special recognition for their contributions. Financial support from the Natural Sciences and Engineering Research Council of Canada (NSERC grant A-1525), the American Chemical Society Petroleum Research Fund (ACS-PRF grant 20514-AC5-SF89), the Canadian Government Industrial Research Assistance Program (IRAP-H grant), a Bruker Scholarship (Bruker Spectrospin Canada Ltd.) and an University of British Columbia Graduate Fellowship (UGF) is gratefully acknowledged. Acknowledgement is made to Perkin Elmer Corp. for the donation of an ICP power supply and Varian Associates Inc. for a Varian 875 atomic absorption spectrometer used in the CCP research project. I wish to thank Deborah and Ken Reimer of the Royal Roads Military College (Victoria, B.C.) for the use of the flameless atomizer, and Dr. A. Wade of UBC for providing me a Simplex program. Also I would like to thank J. Harnly of US Department of Agriculture, Beltsville Human Nutrition Research Center and D. Styris of Pacific Northwest Laboratory for providing us with the graphite electrodes used for the work reported in the Chapter 6. xxiii Chapter 1 INTRODUCTION 1 1.1 Overview and goals In spite of the success enjoyed by inductively coupled plasma (ICP) atomic emission spectrometry (AES) and the graphite furnace (GF) atomic absorption spectrometry (AAS) during the last decade, the development of new spectrochemical sources with improved analytical characteristics continues to be a challenge in the field of analytical atomic spectroscopy [1.1-1.3]. Some desirable characteristics for a source include: 1. complete dissociation of analyte molecules, 2. capability of analyzing a large number of elements, especially refractory and high excitation potential elements, 3. freedom from interferences, including spectroscopic interference, chemical interferences, ionization interferences and other physical interferences, 4. capability of independent optimization of vaporization and excitation, 5. low detection limits, 6. high precision, 7. capability of rnicro-sampling, 8. capability of direct solid sampling, 9. convenience of operation and high sample throughput rate, and 10. low cost of equipment and operation. The difficulty of designing such a source lies in the fact that these characteristics are not necessarily compatible. To date, no single source possesses all these properties. However they are the guidelines for the development of an excitation source. The goal of this study is the development of an atmospheric pressure capacitvely coupled plasma (CCP) which can challenge the performance of existing devices for applications in analytical spectroscopy. 2 1.2 Survey on atomizers and sources for analytical spectroscopy 1.2.1 Flames Combustion flames provide a remarkably simple means for converting analytes in solution into free atoms, and they will undoubtedly continue to be ubiquitous source for analytical atomic spectroscopy. Detection limits for flame AAS are 2 to 3 orders higher than for GF-AAS, which is mainly due to the fact that residence time in the flame is 2 to 3 order of magnitude less than that in the furnace. On the other hand, the interferences and systematic errors are less present in flame AAS. Therefore, flame AAS should always be given preference to GF-AAS if the analyte concentration in question is high enough. Flow injection analysis (FLA) is rapidly becoming accepted as a convenient means of introducing micro-samples into flame AAS [1.4]. Sinha [1.5] has developed a triple capillary aspiration system that enables the addition of a Lanthanum solution and implementation of a standard addition method for FIA-Flame-AAS. An on-line dilutor has been described [1.6]. Prudnikor [1.7] described a pulsed introduction system for microvolumes (~1 |iL) of sample solutions. Mukherjee [1.8], using a small sample cup, injected 1-3 ul samples into the nebulizer for flame spectrometric analysis. Stephens et al. [1.9-1.11] have proposed an instrumental method of solution preconcentration before flame AAS measurements. Aerosol was collected on a wire or in a liquid electrode by an electrostatic trapping technique. This technique was later applied in rf plasma emission spectrometry [1.12]. Instead of introducing samples via the nebulizer, direct micro-sample introduction into a flame has been developed by many analysts. Brandt et al. [1.13] proposed a Pt loop microsampling technique. The Delves' Cup method has been used for the determination of Pb in urban air particulates [1.14]. The furnace-in-flame atomizer combines an electrothermal graphite furnace with the flame [1.15-1.16]. This atomizer has improved 3 power of detection, and is as simple and convenient to operate as electrothermal atomizer (ETA)systems. It also has potential advantages for direct solid sample and oil analysis. Another application area of flame AAS is interfacing with either gas chromatography (GC) or liquid chromatography (LC). Reviews on this subject have been published.[1.17-1.18] 1.2.2 Arcs and sparks As one of the oldest excitation sources, arcs and sparks have again demonstrated their versatility for trace element analysis. Boumans' book Theory of Spectrocchemical Excitation gives an excellent account of the arc [1.19]. Arc and spark sources with different configurations have been re-investigated and developed in recent years. A double plasma arc within a graphite tube was developed by Nickel et al. [1.20]. The aim of this modification was to prolong the residence time of particles in the plasma and thus to increase the radiation intensity. Paksy et al. [1.21] have described an axial injection of argon into a conventional arc source through a hole drilled in the upper electrode. Self-reversal is practically eliminated. The analytical possibilities of high current d.c. argon arc (HCDC) were studied by the same author [1.22]. Aziz et al. [1.23] used a spark aerosol generator combined with excitation in an ICP to directly analyse solid samples. Green and Williams [1.24] have developed a microarc as a primary source for mutilelement exterminations. The possibility of using this source as a GC detector was also demonstrated. 1.2.3 Low-pressure discharges Discharges at low pressure have been closely studied since the second half of the nineteenth century and are thus among the best charaterized devices in physics. Caroli [1.25] has published a review on this subject. The three most important low pressure 4 discharges applied in analytical spectroscopy, to date, are the Grimm discharge (GD) lamp, the hollow cathode discharge (HCD) and the FANES (furnace atomization non-thermal excitation emission spectrometry) source. In a low pressure discharge, the electron temperatures are very high (>10,000K). Excitation temperatures however, are lower than electron temperatures and the values obtained differ depending upon the thermometric lines used for a temperature determination. Kinetic gas temperatures are low (<1000 K from the Doppler broadening of spectral lines and 800-1500 K from rotational and vibrational spectra).Therefore, it is apparent that low pressure discharges are not in thermal equilibrium. In GDs and HCDs, the analyte is volatilised uniquely by cathodic sputtering and therefore selective volatilization and related interferences are avoided. The typical penetration rate is about 3 p.m/min. Analyte atoms are excited in the negative glow of the discharge mainly by electron impact. The spectra consist mainly of the most sensitive atomic lines which indicates the largest population of analyte is in atomic state. High line-to-background intensity ratio is due to the high excitation temperature and lower gas temperature in low pressure discharges. The effect of Doppler and pressure broadening are small, and as a result, the line widths are very narrow. The discharge gas is usually a noble gas which, because of its high ionization energy, permits efficient excitation of high energy lines. On the other hand, low pressure operation is not only inconvenient, but also limits the residence time of analyte atoms and the atom density in the observation volume, consequently resulting in low absolute intensities of spectral lines. Ko [1.26] has designed several new glow discharge lamps. A preliminary investigation of this new family of lamps shows that they are well suited to different applications for the extermination of major and minor constituents, and for bulk and surface analysis. 5 Bubert [1.27] used a Grimm-type glow discharge lamp for atomic fluorescence spectrometry. A pulsed glow discharge has been used as an atom cell for laser excited atomic fluorescence by Smith et ai [1.28]. Lomdahl et al. [1.29-1.30] examined the factors influencing the determination of non-conducting materials by emission spectroscopy with a boosted-output glow discharge lamp. Precise and accurate determinations of metal in powdered rocks and metals can be obtained. One study has reported a HCD which was coupled to a microwave discharge. The intensities of the copper lines were higher in the presence of a 2450-MHz microwave discharge whereas the argon lines were reported to be weak [1.31]. In another experiment, the use of hollow cathode sources for the detection of Ag, Al, As, Bi, Cd, Cr, Cu, Fe, Ga, Mn, Ni, Pb, Se, Sn, Te, Ti and Zn by AES was reported [1.32]. A review summarizes recent work in this area [1.33]. Most conventional sources, such as arcs, sparks, combustion flames, ICP's and DCP's, do not permit independent optimization of vaporization, atomization, and excitation processes. Falk has [1.34-1.36] developed FANES in which the independent controll of sample vaporization and excitation processes has been achieved. The detection power of FANES is comparable to that of GF-AAS and it allows the possibility of use in mutilelement determinations. However the low pressure operation of the FANES is incapable of taking full advantage of the graphite furnace. Working at such a low pressure also reduces the gas temperature in the source. The question of whether this promising source will become important for analytical applications is not yet answered. 1.2.4 Plasmas Plasmas were first sustained at reduced pressure in the early 1990s. A quenching effect made them difficult to generate at atmospheric pressure. By the 1940s it was possible to sustain plasmas at atmospheric pressure using radio frequency (rf) generators. The ICP 6 and direct current plasma (DCP) have found applications in many areas, because of their simultaneous multielement deterrnination capability, low detection limits, wide dynamic range and relative freedom from chemical interferences. Today, commercial ICP instruments are very popular in analytical laboratories. The microwave-induced plasma (MIP) has been applied successfully as a detector for GC and was commercialized by Hewlett-Packard in 1989. 1.2.4.1 Inductively coupled plasmas The pioneering work of Reed [1.37], a physicist, in the 1960s began a new era in analytical atomic spectroscopy. He was the first to successfully generate a stabilized rf plasma by inductve coupling at atmospheric pressure in an open-ended tube with flowing argon. This device was originally designed for growing highly purified crystals. Although he recognized that the plasma offered possibilities for applications in AES, the initial studies of ICP AES were undertaken independently in groups lead by Fassel at the Ames Laboratory, USAEC, at Ames, Iowa [1.38] and by Greenfield at Albright and Wilson in Oldburg, England [1.39]. An ICP torch consists of three concentric quartz tubes. The plasma is sustained by a support gas such as Ar, He or N2- Normally, a plasma gas flow rate of 10-15 L/min is used. A water cooled copper load coil encircling the outer tube of the torch is used to couple the rf power to the plasma. Liquid samples are usually introduced into the plasma by pneumatic nebulizers. Recognized weaknesses of the pneumatic nebulizer have once again promoted interest in the development of alternative techniques of sample introduction for ICP. These techniques include graphite rod vaporization [1.40], graphite furnace vaporization [1.41-1.42], the use of a graphite rod for direct sample insertion [1.43-1.44], carbon cup vaporization [1.45-1.46], the microarc atomizer [1.47], RF arc [1.48] and spark [1.23] vaporization, laser ablation-ICP [1.49] and hydride generation [1.50]. 7 1.2.4.2 Microwave-induced plasmas Conventional microwave-induced plasma (MTP) sources generally do not reach the the detection limits of the ICP or DCP. The MTP has insufficient energy at low power to desolvate samples and the plasma tends to be quenched. However, it is potentially useful as a gas chromatography (GC) detector [1.51-1.52] in which the sample entering the MTP has already been vaporized. The manufacturing and operating costs of the MTP are lower than those of the ICP. 1.2.4.3 Direct Current plasmas A novel rotating DCP design was proposed by Hara et al. [1.53]. The calibration curve obtained in this device was linear over five orders of magnitude. In the only commercial DCP currently available (manufactured by Applied Research Laboratories, a subsidiary of Fisons), a Y-shaped plasma is formed between two graphite anodes and a tungsten cathode. Because it is more sensitive to interferences, the DCP has not enjoyed the same popularity as the ICP. 1.2.5 Electrothermal atomizers Since the introduction of electrothermal atomization in a graphite furnace in 1959 and the first commercial unit in 1969, AAS has become one of the most sensitive techniques for ultra-trace element analysis. Soon after its initial, extremely euphoric, period analysts began to recognize a serious interference problem which is due mainly to the very complex chemical reactions occurring during the thermal pretreatment and atomization steps. In recent years the importance of obtaining both temporal and spatial isothermal atomization in graphite furnace has been well recognized. In order to attain this condition, techniques such as the L'vov platform [1.51], the probe insert [1.52-1.53], capacitive 8 discharge heating [1.54], the constant temperature furnace [1.55], the double-walled furnace [1.56], and the integrated contact cuvette [1.57] have been developed. Significant progress in reducing chemical interferences has been realized by improving the graphite surface, e.g., by using pyrolytic-graphite or a metal carbide coating, glassy carbon tubes [1.58], or Ta foil - W wire lined graphite tubes [1.59]. Improved the background correction capability has led to a reduction of inherent interferences in GF-AAS. Continuum source correction systems proved quite adequate for flame AAS where background is modest Zeeman effect background correction using line splitting in a magnetic field is more effective for GF-AAS. The pulsed hollow-cathode lamp background correction system proposed by Smith and Hieftje [1.60], based upon the use of a hollow cathode lamp working in two modes, is another useful technique. 1.2.6 Chemical vapour generators Hydride generation was introduced in 1955 for use in d.c. arc AES [1.61], and later in 1969 for flame AAS [1.62]. The method is charaterized by its simplicity, low cost, high sensitivity and relative freedom from chemical interferences. It has found applications in the determination of As, Bi , Ge, Hg.Os, Sb, Se, Sn, Tl and Te etc. The main advantage of chemical vapour generation is that it involves analyte separation and preconcentration. This results in a superior sensitivity and a suppression of interferences during atomization. Hydride generation has been reviewed by Godden and Homerson [1.63], and by Dedina [1.64]. The methods have found many uses in AAS [1.65] and ICP [1.50] analysis. 1.2.7 Lasers Use of laser ablation to atomize the samples has been reported and reviewed [1.49,1.66-1.68]. The advantages of laser ablation and its applications in analytical atomic spectroscopy are discussed in the Introduction section of Chapter 6 of this thesis. 9 In recent years, there has been an increase in the use of lasers in other areas in analytical spectroscopy, e.g.laser-excited atomic fluorescence spectrometry (LEAFS) and laser enhanced ionization (LEI). The first report on laser-excited coherent forward scattering has appeared [1.69], though the results were not very encouraging. An electrothermal atomizer is more important than a flame for AFS, because an ideal atom cell for AFS would produce only ground-state atoms with no mechanisms for quenching the excited states. Dittrich [1.70] modified two commercially available atomizers, the Beckman 1268 and Perkin-Elmer HGA-500-EA3 graphite furnaces, for the use in LEAFS. Interferences in the determination of Pb were found to be reduced compared with those obtained with a laboratory-built carbon rod atomizer. A comprehensive comparison of atomizers and atmospheres has been made [1.71]. This study involved various combinations of a graphite rod, a graphite cup, a tantalum-foil liner and tantalum carbide, and pyrolytic graphite coating, Hydrogen-argon and argon atmospheres were employed at both normal and reduced pressures. Detection limits for A l , Cu, In, L i , Mn, Pb, Pt and Sn were found to be in the pg to sub-pg range with linear working ranges varying from three to seven orders of magnitude. Laser-enhanced ionization spectroscopy (LEIS) is a spectroscopic method from which an extremely good power of detection can be expected. Resonance radiation from a tunable laser can ionize elements within an atom reservoir very selectively. The ions can be detected by sequential optogalvanic means or with the aid of a thermionic diode, the latter being more sensitive by an order of magnitude or more. 10 1.3 CAPACITIVELY COUPLED PLASMA (CCP)* 1.3.1 Historical The principle of capacitive coupling to generate a plasma has been known for many years. In 1941 Cristescu and Grigorovici [1.72] reported that a CCP discharge would be produced between two circular plates, separated vertically by up to 15 cm. The plates formed a capacitor which was part of a resonance circuit and determined the frequency of the rf generator, typically in the range 60 to 90 MHz. The lower plate had a copper cone with a platinum tip at which a high field strength was produced By touching the tip with an isolated conductor to generate seed electrons, a discharge was formed at the tip; it was reported to have a temperature of 4000 K at an rf power of 650 W. In 1956, Badarau, Giurgea, Giurgea and Trutia [1.73] used a capacitor, consisting of a hollow cylinder and a coaxial electrode, to generate a brush discharge for spectrochemical analysis. The discharge was capacitively induced by a rectified rf source at 43 MHz. The calibration curves for lead and barium were found to be linear, and the sensitivities of thirteen elements were compared with those in flames. Another arrangement for capacitive coupling consists of an inductance coil of tubular copper with a coaxial electrode as shown in Fig. 1.1. The working gas, containing nebulized liquid, travels through the coil and emerges from small holes in the electrode tip. The coil is part of the L C loop of the oscillator circuit, but the discharge is sustained essentially by capacitive coupling [1.74-1.75], A low power (10-30 W) capacitively coupled plasma torch was developed for the evaluation of an ablation discharge aerosol generator [1.76]. The configuration of the plasma is shown in Fig. 1.2. The electrodes were made of tungsten or molybdenum. The * Part of this section has been submitted for publication in IEEE Transactions on Plasma Science, M.W. Blades, D. Huang, P. Banks, Dong C. Liang, C. Gill and C. Leblanc. 1 1 Quartz Tube Discharge He or H2 Carrier Gas and Atomized Sample Copper Tubing R.F. Coi Ground Plane J Figure 1.1. Schematic description of the oscillator circuit with the inductance coil and the sample-supplying procedure. The analytical solution is sprayed through the coil by a pneumatic atomizer operated with the discharge-sustaining gas (He). The droplets traverse the coil and exit from the four small holes in the molybdenum discharge tip. [From R. Mavrodineanu and R. C. Hughes, Spectrochim. Acta, 19, 1309-1317 (1963).] 12 6 Figure 1.2. Schematic diagram of the low power CCP torch 1) Quartz tube, 2) Argon inlet, 3) Connected to rf generator, 4) Center electrode, 5) Plasma, 6) Auxiliary electrode (from Dong C Liang, M.Sc. Thesis [1.76]) 13 argon flow rate was in the range 0.5-2.0 IVmin. The upper electrode was not essential, as the plasma can be sustained without an upper auxiliary electrode, i.e. a single electrode discharge. However, with a grounded auxiliary electrode the temperature of the plasma was increased and its stability was improved. Superimposing a dc component on the rf is an alternative to the Cristescu and Grigorovici CCP described above [1.72]. Rf power was applied across two electrodes, and an adjustable dc voltage was applied between them and a third electrode, movable between them [1.77-1.78]. Adjusting the dc current from 0.1-1.0 A changed the apparent excitation temperature and also the appearance of the plasma. Applications to spectrochemical analysis were also reported. A capacitively coupled discharge was formed by placing a water cooled, grounded electrode above a helium inductively coupled plasma (ICP) [1.79]. In comparison to the results of a helium ICP, the plasma stability was enhanced and the power required was reduced significantly. The detection limits of nonmetals obtained with this plasma at 500 W are either comparable or superior to those with the helium ICP at 1500 W. The electron number density for this discharge is greater than that for the helium ICP. Rf parallel plate capacitively coupled discharges have been widely used in plasma etching caused by rf sputtering at low pressure. Al l the CCP's mentioned above have a direct contact surface with the plasma that causes problems of electrode contamination for spectrochimcal analysis. The possibility of using a 13.56 MHz rf capacitively coupled low pressure oxygen plasma for spectrochemical analysis has been investigated by Winslow in 1980 [1.80]. The quartz plasma chamber was cylindrical in shape, with a length of 16 cm and radius of 4 cm. Two rf electrodes consisting of a condenser were located outside the plasma chamber. They were isolated from the plasma gas, so no electrode contamination existed. Capacitive coupling was the only mechanism of energy transfer. 1 4 Egorova [1.81] used a CCP with two external annular electrodes at each end of a water-cooled discharge tube for solution spectrochemical analysis. The plasma was run at atmospheric pressure in argon, and was viewed along its axis. This type of coupling had been used earlier for isotope analysis in plasmas at low pressure [1.82-1.83]. Two kinds of discharge were found to be useful for spectrochemical analysis. The first was a constricted plasma, like an arc channel with high excitation energy. The second was a continuous diffuse luminescence which filled the tube, in which spectra of low energy predominated. Zvyagintsev et al. [1.84-1.85] developed an electrodeless capacitive arc (plasmatron) which ran in air at atmospheric pressure. The discharge torch was a two external annular electrode type similar to Egorova's device. The capacitive arc was run at 150 MHz and a few kilowatts of power with an air flow rate of 30 - 600 IVmin. A U.S. patent [1.86] described a capacitively coupled plasma jet device. The main difference between this CCP jet and those of Egorova and Zvyagintsev was that one of the annular electrodes was smaller and was placed inside the discharge tube. An atmospheric pressure capacitively coupled radio frequency plasma discharge has been reported by Liang and Blades [1.87] for atomic absorption and emission analysis of small, discrete sample volumes (1-10 pX). This CCP device has been demonstrated as being potentially useful not only for both atomic absorption spectrometry (AAS), and atomic emission spectrometry (AES) but also for gas chromatagraphy (GC) and as a spectral lamp for spectroscopic measurements [1.88-1.90]. As a further development of the CCP technology, an atmospheric pressure furnace capacitively coupled plasma (APF-CCP) has also been developed by Liang and Blades [1.91-1.93]. This technology offers simultaneous multielement determinations and better detection limits than the graphite furnace AAS. The development of the CCP is described in the following chapters. Most recently, Duckworth and Marcus [1.94] have reported an rf powered glow dischange atomizer/ionization source for soils mass spectrometry (MS). This device was 15 essentially a low pressure (100-500 mTorr), rf CCP run at 13.56 MHz and with 5-50 W rf power. Mass spectra were presented for alloy, metal oxide, and glass matrix samples. Afterglows in nitrogen are often called "active nitrogen" (AN). During 1954-1955, Kenty observed that reactions usually observed in A N sources could be duplicated in electrical discharges of argon containing traces of nitrogen at low pressure (~300 Torr) [1.95]. In 1980, D'silva, Rice and Fassel used an atmospheric pressure active nitrogen (APAN) discharge for analytical spectroscopy. The A P A N torch consisted of two concentric fused silica tubes of 1.8 cm and 2.6 cm diameters, and approximately 40 cm in length. The copper foil electrodes were positioned inside the inner and outside the outer tube. An annular afterglow discharge can be formed in the gap of the two quartz tubes at 1800 Hz frequency, 500W power and 20 kV ac potential [1.96]. The APAN is essentially a low frequency capacitive discharge. Microwave plasmas can be classified into two groups, microwave induced plasmas (MTP), which has been discussed in the previous section, and capacitively coupled microwave plasmas (CMP). In the first type, the microwave energy is coupled to the plasma in a discharge tube within a resonant cavity, while in the second type, the microwaves generated from a magnetron are delivered to the coaxial waveguide by a rectangular waveguide, the plasma being formed at the tip of a central electrode. Most of the CMP sources are derived from the designs of Cobine and Wilbur [1.97] and Schmidt [1.98], which operate in the frequency range from 500-2450 MHz. A schematic diagram of a commercialized CMP is shown in Fig. 1.3 [1.99]. The discharge is a torch-like plasma. Recently, Winefordner et al. [1.100-1.101] have developed new torch designs for the CMP, including a high power CMP (1.6 kW). They reported that the high power CMP exhibits a number of advantages over the ICP and most MIPs. For example it is less sensitive to the introduction of aqueous solutions, has lower detection limits for non-16 Figure 1.3. Schematic diagram of the A.RX. CMP source with nebulizer I. Sample solution. 2. Pneumatic nebulizer (concentric type). 3. Fog chamber. 4. Drainage. 5. To exhaust 6. Aerosol injection tube (PTFE) to "burner." 7. Tuning stub. 8. Coaxial waveguide. 9. Magnetron. 10. Stabilized power supply. II. Replaceable two-piece burner tip 17 metals, is less expensive, is safer and is easier to operate. The detection limits for Cd, Mg, Sn, Ti, Ni, Fe, A l , Pb, Ca, Cr and Na were in the range from 0.05-12 ppb. 1.3.2 Principles of plasma generation Electrical discharges sustained by rf or microwave fields differ significantly from dc discharges in a number of ways. These include the initiation of the discharge and the conditions required to maintain the plasma. The interaction of the rf field, electrons or molecules, ions, and containing walls, determine the values of the plasma parameters in a gas which was subjected to an electric field. The breakdown electric field is a function of the ionization potential of the gas, of the collision properties of the electrons in the gas, of the plasma torch dimensions, and of the frequency of the applied field. When a rf or microwave electric field is applied across a gas, charged particles in the volume are accelerated. Electrons are always present in a gas due to cosmic radiation. In the case where there are no electrons in the gas during the time a rf field is applied, no energy transfer can take place and no discharge can be generated. The electrons are accelerated much more than the ions due to their difference in mass. When the direction of the field changes, the electrons oscillate, following the rf frequency within the volume of the gas. If the frequency is too low, the reversal of direction of acceleration does not take place before the electrons strike the wall of the gas container. This is the characteristic that distinguishes an rf or microwave discharge from the low frequency discharge. The electric field required to break down a gas is much higher than that required to maintain the discharge after it has started. Usually a Tesla coil is required to start a plasma at atmospheric pressure. In an ICP discharge, the rf currents flowing in the induction coil generate an alterating magnetic field which induces an eddy current in the plasma. This electron flow is analogous in behavior to the current flow in a short-circuited secondary of a transformer. Therefore, the ICP is a low voltage and high current discharge. In the case 18 of the CCP discharge, the rf potential is much higher than in the ICP arrangement. If the applied rf potential or the induced potential at the tip of the electrodes is high enough to break down the gases, a self-starting of the plasma becomes possible. While the electrons obtain energy from the field, they also lose energy via collision with gas molecules and atoms. The collisions play a major role in the processes of dissociation, excitation and ionization in a plasma. 1.3.3 Applications The applications of the CCP, CMP and APAN can be found in the areas of AAS [1.87, 1.102], AES [1.73, 1.79-1.81, 1.88-1.89], and MS [1.95], as well as GC [1.90]. Some of the applications and results have already been provided in the history section. The detection limits in an argon CCP reported by Egorova [1.73] for 24 elements were in the range of 0.5-30 ppm and 0.05-0.3 ppm using a pneumatic nebulizer and an ultrasonic nebulizer respectively. Boumans et al. [1.99] have compared the analytical performance of a CMP to an ICP in emission spectrometry. In 1983, Wunsch et. al. [1.102] used a nitrogen CMP as an atomizer for atomic absorption spectroscopy (AAS). An excitation temperature of 5600 to 4750 K and a kinetic temperature of 4450 K were reported for their plasma. The sensitivities for Ba, Ca, Co, Cr, Fe, Mg, Mo, Ni, Sr, W and Zn were given. Referring to equal optical path lengths the sensitivity in the CMP (and also the ICP) was not better than in reducing acetylene flames. Emission spectrometric determination of trace elements in aqueous solution by mantle-stabilized CMP was reported in 1982. Argon served as the plasma gas. By the addition of nitrogen as a second concentric gas mantle, the excitation source can be stabilized to give better excitation conditions. Detection limits for Be, Mg, Ca, Sr, B, Al, Ga, Tl, Si, Pb, Cu, Ag, Cd, Ti, V, Cr, Mo, Mn, Fe, Co, and Ni were in the range of 0.02-2.0 ppb 19 [1.103]. The determinations of arsenic [1.104], and tungsten [1.105] by CMP-AES were also reported. Compared to low pressure devices [1.80], an atmospheric pressure rf CCP has some definite advantages for spectrochemical analysis. Atmospheric pressure operation is not only convenient for changing samples but also provides for the possibility of high-yield rf sputtering sampling. Atmospheric pressure plasmas provide a relatively high thermal gas temperature, which should allow more complete dissociation of molecular species. This should reduce the occurrence of chemical interferences in a excitation source. The CCP sources developed by Liang and Blades [1.87-1.90] can be operated at atmospheric pressure and at very low RF input powers (30-600 W) which allows for optimal conditions for atom resonance line absorption measurements. Detailed description of the applications of the CCP is given through the chapter 2 to 7 of this thesis. Although the principle of using capacitive coupling to sustain a plasma is not new, it offers a challenging opportunity to develop new atomizers and sources for analytical spectroscopy and further novel developments can be expected. 20 1.4 General considerations of plasma source design 1.4.1 Temperature The number density of analyte atoms in a particular energy state in a source is determined by the dissociation equilibrium (Guldberg-Waage distribution), population factor (Boltzmann distribution), and degree of ionization (Saha equilibrium). Temperature is the most important parameter governing the above equilibria. Temperature also changes the full-width of half-maximum (FWHM) of line profile which affects the sensitivity and linear dynamic range of a calibration curve for AAS. Both temporal and spatial isothermal operation is important for GF-AAS to obtain freedom from interference, high sensitivity and precise analysis. Temperature is an important parameter in the control of the processes of diffusion, convection and gas expansion. Furthermore, background and signal-to-noise ratio are dependant on the temperature in the atomizer and source. 1.4.2 Geometry and transport efficiency Consider the following facts concering an atomizer or a source. The volatilized sample material is usually not transferred quantitatively to the observation zone, which results in transport losses. The length of optical path, the diameter of the source, the flow rate of the carrier gas and the fractional volatilization <kterrnine the residence time of analyte atoms in the observation volume. The dead volume of the source and the heating rate in the atomization step affect the peak shape of the analytical signals. A smaller dead volume and higher heating rate result in a narrow peak. Consequendy this results in higher sensitivity. Atoms are removed from the observation volume by the processes of diffusion, gas expansion, convection and recombination of the free atoms. Optimizing the geometric design of the source can increase the transport efficiency and the residence time. 21 1.4.3 Pressure Pressure is another important consideration in the design of a source. The following factors, determined by the pressure, should be taken into account in the development of a new source: 1) residence times decrease at low pressure, 2) the gas temperature is much lower than excitation temperature at low pressure conditions, 3) the analyte volatilization rates increase with decreasing pressure, 4) there is additional inconvenience associated with changing samples at non-atmospheric pressure, and 5) collision broadening increases at high pressure. 1.4.4 Effect of chemical reactions The number density of analyte atoms in an atomizer and a source can be changed by the chemical reactions. The following chemical reactions are typically observed in an atomizer: 1) Reactions on atomizer surfaces -reducing reactions in graphite furnaces, -carbide formation on graphite surfaces, -alloy formation at tantalum and tungsten surfaces; 2) Reactions in gas phases -halide formation, -reducing or oxidizing gases which change the dissociation equilibrium, -reactions of which molecular gases require extra dissociation energy and hence decrease the plasma temperature. 22 1.4.5 Background A continuum background is emitted in a source by incompletely volatilized particles and/or black body radiation from red hot electrodes. Molecular emission from the gas phase is another important source of background emission. In flames and plasmas, a continuum can also be emitted in the gas phase as a result of a recombination process between two particles. The energy released in the recombination process equals the sum of the discrete ionization or dissociation energy of the product particle and the initial translational energy of relative motion, which is continuously distributed. Ion-recombination is a major source of continuum emission in the ICP. The major sources of background absorption are scattering by smog or aerosol in the light path, and by molecular absorption. 1.4.6 Others Other considerations in the design of a source are: 1) uniform plasma medium, no self-absorption; 2) wide selection of working gases to provide either inert, reducing or oxidizing environment; 3) atmospheric pressure operation, easy to change samples; 4) independent optimization of vaporization and atomization as well as excitation; 5) no memory effect; 6) wide selection of samples eg. liquids, powders, solid including slurries; 7) Cost. 23 References [1.1] W. Slavin, D. C. Manning and G. R. Carnrick, / . Anal. At. Spectrom, 3, 13 (1988). [1.2] Gerhard A. Meyer, Anal. Chem., 59, 1345A (1987). [1.3] Gunther Tolg, Analyst, 112, 365, (1987). [1.4] M . Ni, Huaxue Shiji, 4, 253 (1982), Chem. Abstr., 98, 100299. [1.5] R. C. P. Shinha, Fert. Technol, 18, 99 (1981). [1.6] M . H. Ramsey and M . Thompson, Analyst, 107, 232 (1982). [1.7] E. S. Prudnikov, Zh. Anal. Khim, 37, 1133 (1982). Chem. Abstr., 97, 174022. [1.8] S. Mukherjee, Arch. Oral. Biol, 28, 45 (1983). [1.9] P. A. Michalik and R. Stephens, Talanta, 28, 37 (1981). [1.10] Idem, ibid., 28, 43 (1981). [1.11] Idem, ibid., 29, 443 (1982). [1.12] K. Haapakka and R. Stephens, Talanta, 31, 679 (1984). [1.13] H. Berndt, J. Messerschmidt, Spectrochim. Acta, 36B, 845 (1981). [1.14] D. G. Pachuta and L. J. Cline Love, Anal. Chem., 52, 438 and 444 (1980). [1.15] V. P. Borzov, B. B. L'vov abd G. V. Plyuschch, Zh. Prikl. Spektrosk., 11, 217 (1969). [1.16] T. Kanter, L. Bezur and E. Pangor, Mikrochim Acta, 1,289 (1981). [1.17] L. Ebdon, S. Hill , and R. W. Ward, Analyst, 111, 1113 (1986). [1.18] L. Ebdon, S. Hill and R. W. Ward, Analyst, 112, 1 (1987). [1.19] P. W. J. M . Boumans, Theory of Spectrochemical Excitation, Adam Hilger, London, 1966; Plenum, New York (1966). [1.20] H. Nickel, Ch. Arens and M . Mazurkiewicz, Spectrochim. Acta, 39B, 1431 (1984). [1.21] L. Paksy and I. J. Lakators, Spectrochim. Acta, 38B, 1099 (1983). [1.22] L. Paksy, Spectrochim, Acta, 39B, 1449 (1984). 24 [1.23] A. Aziz, J. A. C. Broekaert, K. Laqua and F. Leis, Spectrochim. Acta, 39B, 1091 (1984). [1.24] R. B. Green; R. R. Williams, Anal. Chim. Acta, 187, 301 (1986). [1.25] Sergio Caroli, / . Anal. At. Spectrom, 2, 661 (1986). [1.26] J. B. Ko, Spectrochim. Acta, 39B, 1405 (1984). [1.27] H. Bubert, Spectrochim. Acta, 39B, 1377 (1984). [1.28] B. W. Smith, N . Omenetto and J. D. Winefordner, Spectrochim. Acta, 39B, 1389 (1984). [1.29] G. S. Lomdahl and R. McPherson, Anal. Chim Acta, 148, 171 (1983). [1.30] G. S. Lomdahl and J.V. Sullivan, Spectrochim. Acta, 39B, 1395 (1984). [1.31] S. Caroli, A. Alimonti and F. Petrucci, Anal. Chim. Acta, 136, 269 (1982). [1.32] S. Caroli, O. Senofonte, N . Violante, F. Petrucci and A. Alimonti, Spectrochim. Acta, 39B, 1425 (1984). [1.33] S. Caroli, Prog. At. Sprctrosc, 6, 253 (1983). [1.34] H. Falk, E. Hoffmann, I. Jaeckel and Ch. Ludke, Spectrochim. Acta 34B, 333 (1979). [1.35] H. Falk, E. Hoffmann, and Ch. Ludke, Spectrochim. Acta, 36B, 767 (1981). [1.36] H. Falk, E. Hoffmann, and Ch. Ludke, Spectrochim. Acta, 39B, 283 (1984). [1.37] T. B. Reed, J. Appl. Phys., 32, 821, 2534 (1961). [1.38] R. H. Wendt and V. A. Fassel, Anal. Chem., 37, 920 (1965). [1.39] S. Greenfield, L L. Jones, and C. T. Berry, Analyst, 89, 713 (1964). [1.40] R. M . Barnes and P. Fodor, Spectrochim. Acta , 37B 1191 (1983). [1.41] G. Crabi., P. Carvalli, M . Acheilli, G. Rossi and N. Omenetto, At. Spectrosc, 3, 81 (1982). [1.42] A. Aziz, J. A. C. Broekaert and F. Leis, Spectrochim. Acta, 37B 369 (1982). [1.43] G. F. Kirkbright Z. Li-Xing, Analyst, 107, 617 (1982). 25 [1.44] Z. Li-Xing, G. F. Kirkbright, M . J. Cope and J. M . Watson, Appl. Spectrosc, 37, 250 (1983). [1.45] K. C. Ng and J. A. Caruso, Analyst, 108, 476 (1983). [1.46] K. C. Ng and J. A. Caruso, Anal. Chem., 55, 1513 (1983). [1.47] J. P. Keilsohn, R. D. Deutsch and G. M . Hieftje, Appl. Spectrosc., 37,101 (1983), [1.48] P. B. Farnsworth and G. M . Hieftje, Anal. Chem., 55, 1414 (1983). [1.49] T. Ishizuka and Y. Uwamino, Spectrochim. Acta, 38B, 519 (1983). [1.50] R. C. Hutton and B. Preston, Analyst, 108, 1409 (1983). [1.51] W. Slavin and D. C. Manning, Spectrochim. Acta, 35B, 701 (1980). [1.52] S. K. Giri, D. Littlejohn and J. M . Ottaway, Analyst, 107, 1095 (1982). [1.53] S. K. Giri, C. K. Shields, D. Littlejohn and J. M . Ottaway, Analyst, 108, 244 (1983). [1.54] C. L. Chakrabarti, H. A. Hamed, C. C. Wan, W. C. L i , P. C. Bertels, D.C. Gregoire and S. Lee, Anal. Chem., 52, 117 (1980). [1.55] D. R. Jenke and R. Woodriff, Appl. Spectrosc, 37, 470 (1983). [1.56] S. R. Rawson and R. Woodriff, Anal. Chem., 54, 2050 (1982). [1.57] W. Freeh, D. C. Baxter, and B. Hutsch, Anal. Chem., 58, 1973 (1986). [1.58] L. De Galan, M . T. C. De Loos-Vollebregt and M . A. R. Oosterling, Analyst, 108, 138 (1983). [1.59] Z. Wu and Y. Ma, Huanjing Huaxue, 1,228 (1982). [1.60] S. B. Smith, and G. M . Hieftje, Appl. Spectrosc, 37, 419 (1983). [1.61] K. Dittrich and R, Wennrich, Spectrochim. Acta, 35B, 731 (1980). [1.62] K. Surnino, R. Yamamoto, F. Matayama, S. Kitamura and H.Itoh, Anal. Chem., 52, 1064 (1980). [1.63] R. G. Godden and D. R. Thomerson, Analyst, 105, 1137 (1980). [1.64] J. Dedina, Prog. Analyt. Spectrosc. 11, 251 (1988). 2 6 [1.65] J. Dedina, I. Rubeska, Spectrochim. Acta, 35B, 119 (1980). [1.66] K. Dittrich and R. Wennrich, Spectrochim. Acta, 35B, 731 (1980). [1.67] K. Sumino, R. Yamamoto, F. Matayama, S. Kitamura and H. Itoh, Anal. Chem., 52, 1064 (1980). [1.68] K. Kagawa and S. Yokoi, Spectrochim. Acta, 37B, 789 (1982). [1.69] L. A . Davis, R. J. Krupa and J. D. Winefordner, Spectrochim. Acta, 41B, 1167 (1986). [1.70] K. Dittrich and H. J. Stark, J. Anal. At. Spectrom. 1, 237 (1986). [1.71] D. Goforth and J. D. Winefordner, Anal. Chem., 58, 2598 (1986). [1.72] G. Cristescu and R. Grigorovici, Bull. Soc. Roum. Phys.,A2, 37 (1941). [1.73] E. Badarau, M . Giurgea, Gh. Giurgea and A. T. H. Trutia, Colloq. Spectro. Internal., V I , 441, (1956). [1.74] R. Mavrodineanu and R. C. Hughes, Spectrochim. Acta, 19, 1309 (1963). [1.75] C. D. West and D. N. Hume, Anal. Chem, 36, 412 (1964). [1.76] Dong C. Liang, M. Sc. Thesis, Dalhousie Univ., Halifax, N.S., Can. (1986). [1.77] G.D. Cristescu, Ann. Physik., 6, 153 (1960). [1.78] G.D. Cristescu and M . Giurgea, Z. Chem., 7, 360 (1967). [1.79] Shi-Kit Chan, H. Tan and A. Montaser, Appl. Spectro., 43,92 (1989). [1.80] R. J. Winslow, PhD. Thesis, Univ. Massachusetts, (1980). [1.81] K. A. Egorova, Zh. Prikl. Spektrosk., 6,22 (1967). [1.82] H. P. Broida and J. W. Moyer, / . Opt. Soc. Am., 42, 37 (1952). [1.83] O. Botschkowa, S. Frich and E. Schreider, Spectrochim. Acta, 13, 50 (1958). [1.84] A. V . Zvyagintsev, R. V . Mitin and K. K. ftyaclkin, Zh. Tekh. Fiz., 45,278 (1975). [1.85] N . I. Gonchar, A. V . Zvyagintsev, R. V . Mitin and K. K. Pryadkin, Zh. Tekh. Fiz., 45, 657 (1975). [1.86] W. G. Elliott, U.S. Patent, 4,479,075 (1984). 27 [1.87] Dong C. Liang and M . W. Blades, Anal. Chem., 60,27 (1988). [1.88] Dong C. Liang and M . W. Blades, Anal. Chem., in press. [1.89] Dong C. Liang and M . W. Blades, Spectrochim. Acta, 44B, 1049 (1989). [1.90] Degui Huang, Dong C. Liang and M . W. Blades, / . Anal. At. Spectrom., 4,789 (1989) [1.91] M . W. Blades and Dong C. Liang, Canadian Chemical Conference and Exhibition, Paper #40 (1989). [1.92] Dong C. Liang and M . W. Blades, Spectrochim. Acta, 44B, 1059 (1989). [1.93] D. Lee Smith, Dong C. Liang, Doug Steel and M . W. Blades, Spectrochim. Acta 45B, 493 (1990). [1.94] D. C. Duckorth and R. K. Marcus, Anal. Chem., 61, 1879 (1989). [1.95] C. Kenty, Abstracts, J. Chem. Phys., 60, 5050 (1974). [1.96] A. P. D'silva, G. W. Rich and V. A. Fassel, Appl. Spectro., 34, 578 (1980). [1.97] J. D. Cobine and D. A. Wilbur, J. Appl. Phys., 22, 835 (1951). [1.98] W. Schmidt, Elektron. Rundschau., 13,404 (1959). [1.99] P. W. J. M . Boumans, F. J. de Boer, F. J. Dahmen, H. Hoelzel and A. Meier, Spectrochim. Acta, 30B, 449 (1975). [1.100] S. Hanaamura, B. W. Smith and J. D. Winefordner, Can. J. Spectrosc, 29, 13 (1984). [1.101] J. D. Hwang, W. Masamba, B. W. Smith and J. D. Winefordner, Can. J. Spectrosc, 33, 156 (1988). [1.102] G. Wunsch, G. Hegenberg and N. Czech, Spectrochim. Acta, 38B, 1135 (1983). [1.103] A. Disam, P. Tschoepel and G. Toelg, Z. Anal. Chem., 310, 131 (1982). [1.104] I. Atsuya and K. Akatsuka, Spectrochim. Acta, 36B, 747 (1981). [1.105] S. B. Zayakina, I. R. Shelpakova and I. G. Yudelevich, Chem. Abstr., 88 (14), 98580d. 28 Chapter 2 ATMOSPHERIC PRESSURE CAPACITIVELY COUPLED PLASMA ATOMIZER FOR ATOMIC ABSORPTION SPECTROMETRY * *Part of this chapter was published in Anal. Chem. 1988, 60,27, Dong C. Liang and M.W. Blades. 29 2.1. Introduction During the past two decades inductively coupled plasma optical emission spectroscopy (ICP-OES) has played an important role in the area of elemental analysis. ICP-OES possesses several distinct advantages over other atomic methods including simultaneous multi-element capability, relative freedom from chemical interferences, low detection limits, and a large linear dynamic range. In recent years the ICP has also been used as a source for multi-element atomic fluorescence spectrometry (AFS) [2.1,2.2] and plasma source mass spectrometry (ICP-MS) [2.3,2.4]. However, to date, the ICP has not been successfully exploited as an atomizer for atomic absorption spectrometry (AAS). The properties of the ICP as an atom reservoir for AAS have been investigated by Wendt and Fassel [2.5], Greenfield et al [2.6], and Veillon and Margoshes [2.7]. In addition, Magyar and Aeschbach [2.8] have studied the theoretical implications of using the ICP for AAS. They concluded that the ICP provided sensitivities a factor of ten poorer than those exhibited by a flame. The relatively low sensitivity of ICP-AAS can be attributed to several factors. A relatively high support gas flow rate is required to operate an ICP and this acts to dilute the sample atoms. The absorption volume in an ICP is not optimum for AAS, in particular, the absorption path length is.relatively short and this combined with the high aerosol transport rate means that the residence time of analyte atoms in the absorption volume is short. Moreover, traditional AAS primarily makes use of atomic resonance lines but in the ICP the high temperature favours the production of ionic species [2.9]. In spite of these factors, a plasma environment does provide several distinctive features which suggest that it could offer several advantages over flames and graphite furnaces for atomic absorption measurements. The relatively high temperature promotes efficient vaporization and dissociation and thus aids in the control of chemical interferences. In addition, radio frequency (rf) plasmas are relatively stable and easy to control. The atom reservoir temperature, and hence the characteristics of the absorption volume, can be 3 0 controlled by controlling the input power to the plasma. Also, since a plasma can be made to operate with a variety of gases (eg. Ar, He, N2, H2, etc) the gas phase chemistry can be easily modified. Finally, the shape and extent of a plasma can be controlled through appropriate design of the external electrodes used to couple the rf power into the plasma. . In this chapter a novel low-power rf plasma torch and sample introduction system designed primarily for atomic absorption spectrometry is reported. This torch operates at atmospheric pressure at very low support gas flow rates and makes use of capacitive power coupling to form the plasma. Sample introduction into the plasma is accomplished by using an electrically heated tantalum strip which offers the possibility that the sample vaporization and atomization steps can be separated and independently optimized. The long pathlength tube geometry of the discharge has been designed for atomic absorption measurements and this, in conjunction with the low support gas flow rates, maximizes analyte residence time in the absorption volume. 3 1 2.2 Experimental 2.2.1 Plasma torch A schematic diagram of the device is provided in Figure 2.1. Functionally, the device consists of two parts, the capacitively coupled plasma (CCP) discharge tube and the tantalum strip electrothermal vaporization sample introduction system. The main body of both parts is constructed of quartz glass and the two parts are joined through a narrow neck to form a T-shaped device. The plasma is contained in a quartz tube 20.0 cm in length and 0.4 cm inside diameter and 0.5 cm outside diameter. Power is coupled into the plasma using two stainless steel strips, 18.0 cm long and 0.5 cm wide which are placed parallel to each other on either side of, and in contact with, the outside of this quartz tube. These stainless steel electrodes are connected to the rf power supply. The plasma has been run using a fixed frequency 27.18 MHz rf supply and also with a 125 - 375 KHz variable frequency rf supply. A stable plasma can be sustained at rf powers ranging from 30 -600 W. Plasma support gas is introduced using an inlet on the side of the main body of the quartz container. The discharge will operate at gas flow rates ranging from 0.2 to 6 L/min. The plasma has been sustained using a variety of support gases including Ar, He, and mixtures of these gases with N2, H2, and air. Sample vaporization is accomplished using a tantalum strip which is fastened to two copper rod conductors which are connected to an electrothermal atomizer power supply. These electrodes are surrounded by a water cooled jacket. Samples are placed on the Tantalum strip vaporizer through one of two ground glass tapered inlets using a micro-pipette. Sample sizes vary from 1 to 10 ul. 2.2.2 Experimental facilities The experimental setup is schematically outlined in Figure 2.2 and details of the equipment used are provided in Table 2.1. The CCP discharge was mounted inside the model PT-2500 torch box. Power was coupled to the CCP by inserting a secondary coil into the normal ICP load coil. The leads from this secondary coil were attached to the 32 P l a s m a S a m p l e i n t r o d u c t i o n i n l e t (end v i e w ) T a n t a l u m s t r i p C o p p e r r o d s u p p o r t s 1 U — J — J — P l a s m a s u p p o r t —" gas i n l e t W a t e r i n l e t S t a i n l e s s s t e e l e l e c t r o d e s (to RF p o w e r s u p p l y ) Q u a rtz W a t e r o u t l e t T o DC p o w e r supply Figure 2.1 Schematic diagram of the capacitively coupled plasma and sampling system 33 HCL ^ {E=> 0 HCL POWER SUPPLY © © <§) ® ® RF GENERATOR L l PLASMA ATOMIZER z \ ETA POWER SUPPLY A E l L2 PRINTER MONOCHROMATOR RECORDER o o COMPUTER CHI CH2 RECORDER La LOCK-IN AMP CHI ANALOG DIGITAL x CONVERTER N Figure 2.2. Block diagram of the experimental arrangement stainless steel strip electrodes. The CCP has also been operated using a ENI Power Systems Inc. (Rochester, N.Y.) Model HPG-2 low power (0 - 200 W), variable frequency (125 - 375 KHz) power generator although all data reported in this paper were acquired using the equipment outlined in Table I. A plasma ignition system (tesla coil) was not required since the CCP automatically ignites upon application of approx. 100 W rf power. A 25 cm focal length fused silica lens was used to focus the HCL discharge at the middle of the CCP tube (50 cm object distance) and a 10 cm focal length fused silica lens was used to image the HCL and CCP onto the entrance slit of the monochromator with object and image distances of 27 and 17 cm respectively. At each end of the CCP discharge a stainless steel plate with a 0.4 cm hole was placed to cut down on the amount of un-absorbed HCL radiation and plasma background emission reaching the entrance slit. Analyte absorption and emission in the CCP discharge could be simultaneously measured by monitoring the output from both the preamplifier and the lock-in-amplifier. 2.2.3 Analytical procedure Al l absorption measurements were carried out using the following procedure. Using a positive displacement adjustable micropipette (Gordon-Keeble Ltd Model PDP 75) a 2 - 5 u.1 aqueous sample was placed on the Tantalum strip through the inlet port on the side of the quartz body. The plasma was off at this stage. The sample was dried, ashed, and atomized. The plasma is ignited at the end of the ash cycle. Data was collected through the atomize cycle. After each atomize cycle the atomizer was tested for a memory effect All analytical standards were prepared using Fisher Scientific 1000 ppm atomic absorption standards. The solutions were diluted to volume using 1% HNO3. 35 Table 2.1. Experimental Facilities and Operating Conditions Plasma Power Supply Perkin-Elmer ICP 5500 System consisting of a Plasma-Therm (Kreeson, N.J.), HFP-2500F rf generator, AMN-2500E automatic matching network, APCS-3 automatic power control system and PT - 2500 torch box. Sample Vaporization Tantalum strip 1 cm by 0.4 cm with a depression at the center. Power: Varian Model CRA-61. Normal operating cycle: Dry -105 °C for 60 s., Ash - 300 - 600 °C for 15 s, Atomize - 2000 - 2700 °C for 2 s. Spectrometer Schoeffel-McPherson (Acton, MA) Model 270,0.35 m Czerny-Turner mount scanning monochromator with 600 rulings/mm holographic grating. Reciprocal linear dispersion of 40 A/mm in the first order. Slits Entrance and exit slits set to 50 um, with spectral band width 200 pm. Hollow Cathode Lamps Hollow cathode lamps (HCL) were powered using a home-built, electronically modulated power supply. Modulation frequency was - 250 Hz and a duty cycle of 50%was used. Normal operating currents were used for the lamps. Detector Electronics The photocurrent from a Hammatsu R955 photomultiplier tube was amplified by a home-built preamplifier and fed to a Princeton Applied Research Model 121 Lock-In-Amplifier. The photomultiplier tube was powered using a McPherson Model EU-42A PMT power supply. 36 Data Acquisition Digital data acquisition was carried out using a Tulsa Computers (Owasso, OK) Telex Model 1280 B3M-AT compatible computer equipped with a RC Electronics (Santa Barbara, CA) Model ISC-16 analog-digital converter, running the RC Computerscope software package. Analog data were acquired using a Servocorder 210 chart recorder. 37 2.3 Results and discussion 2.3.1 The CCP and its characteristics An Ar-CCP can be generated as soon as the rf power is applied to the discharge tube. The plasma is light blue in colour and fills the discharge tube but does not enter the vaporization chamber nor is any arcing observed between the plasma and the Tantalum strip. The plasma appears stable without any observable flicker or modulation. At support gas flow rates lower than 4 L/Min the plasma is contained inside the discharge tube, however at support gas flow rates exceeding 4 L/Min a small plasma jet can be seen emanating from each end of the discharge tube. When a sample containing a relatively high concentration of L i is vaporized, a red band of L i emission can be observed to move from the junction of the tee down the length of each branch of the discharge tube. The device has also been found to operate with no difficulty on pure He and on Ar-H2, Ar-N2, Ar-air, He-H2, He-N2, and He-air rnixtures. The addition of H2, N 2 t and air to the argon support gas allows for the adjustment of excitation conditions in the plasma and permits the discharge to provide either an inert, oxidizing, or reducing environment. It is anticipated that this feature will prove to be very useful for the future application of the device to different sample types. For example, a reducing environment can be created by using an A r - H 2 mixture; this should help to control the formation of refractory oxides in the discharge. Typical time-resolved absorption and emission signals acquired from the CCP are provided in Figure 2.3. To record these signals, 0.1 ng of Ag was introduced onto the Tantalum strip, vaporized into the Ar plasma discharge and emission and absorption for the Ag 1328.1 nm line were measured. The power used was 125 W and the support gas flow rate was 0.6 L/Min. The origin on the time axis, which is marked in units of s, corresponds to the beginning of the atomization cycle. The vertical (signal intersity) axis is in arbitrary units. The apparent noise on the emission signal is from the hollow cathode 3 8 0.1 ng Ag 328 nm 2200 — 2000 Jt*^*^*^*^*^*: c ID I800H -Q 1600-] o " 1400 H £ I200H (D -E 1000 H Emission Absorption —i r 0.12 — i r 0.16 0.0 0.4 0.8 Time (sec) Figure 2.3. Absorption and Emission signals for 0.1 ng Ag using the 328.1 nm line. lamp modulation. The signals start to appear after about 0.6 s and persist for about 0.6 s following first appearance. The underlying background is relatively flat for both absorption and emission and is not appreciably affected by the vaporization step. 2.3.2 Background emission spectra in argon and helium CCP Background emission from the Ar-CCP and He-CCP recorded over the wavelength range 200-450 nm at an rf power of 200 W are provided in Figure 2.4 and 2.5 respectively. In both of these plasmas the main spectral features are OH emission in the 280-285 nm and 302-317 nm regions and NO emission in the 215-272 nm region. 2.3.3 Transmittance of argon and helium CCP The transmittance of the Ar-CCP and He-CCP over the wavelength range 200-380 nm at an rf power of 200W is recorded in Figure 2.6. This was recorded by using a D2 lamp and measuring the broadband %T at 10 nm intervals. The transmittance decreases with an increase in wavelength for both plasmas. For the He-CCP the transmittance is greater than 95% and for the Ar-CCP it is greater than 85% over this wavelength range. 2.3.4 Effect of rf input power The effect of changes in rf input power on absorbance of the Ag I 328.1 nm line was studied at a support gas flow rate of 0.6 L/m. The results over the power range 50-500 W are provided in Figure 2.7. The optimal rf power for this line was found to be between 100 and 200 W. At a power of 65 W the absorbance drops to 0 and at powers higher than 200 W the absorbance decreases steadily. At the low end of the power scale it is suspected that the formation of undissociated gas phase molecules reduces the sensitivity, and at the high end, the formation of Ag ions reduces the sensitivity. 40 CO c: ZJ _Q O CO c C D c O CO CO E LU 1 200 250 300 350 Wavelength (nm) 400 450 Figure 2.4 Background emission spectrum of the CCP using argon as a support gas. I ~ — ~ — — I [ I I I 200 250 300 350 400 450 Wavelength (nm) Figure 2 . 5 . Background emission spectrum of the CCP using helium as a support gas. 1 8 0 2 3 0 2 8 0 3 3 0 3 8 0 Wavelength (nm) Figure 2.6. Transmittance of the argon (squares) and helium (diamonds) CCP function of wavelength. 0 100 200 300 400 500 600 rf Power(W) Figure 2.7. Effect of RF input power on Ag 1328.1 nm absorbance. 44 2.3.5 Effect of plasma support-gas flow rate The effect of a variation in support gas flow rate on the absorbance signal for the Ag I 328.1 nm line at a rf power of 150 W is depicted in Figure 2.8. The support gas has a dual role to play in the operation of the CCP discharge. First, it acts to carry the analyte vapor into the discharge tube and second it supports the discharge itself. Therefore the response of analyte absorption is affected both by changes in transport rate, and hence residence time, and also by changes in the nature of the plasma as the support gas flow rate changes. From the figure it can be seen that the maximum absorbance is obtained at a support gas flow rate of 0.6 L/Min. At higher and lower support gas flow rates the absorbance decreases. 2.3.6 Calibration curves To check on analytical performance a support gas flow rate of 0.6 L/Min and an rf input power of 150 W were chosen as the working conditions. The CCP device exhibits good working curve linearity over the concentration range 0 -10 ng total analyte. 2.3.7 Sensitivities of absorption A listing of 0.0044 absorbance unit sensitivities is provided in Table 3.2 for Ag, Cd, Cu, L i , and Sb for the plasma system described in this paper, for conventional graphite furnace AAS [2.13] and for the ICP-AAS [2.8]. It can be seen that the sensitivities for the CCP are in the range of 3.5 to 40 pg depending on the element. These are superior to ICP-AAS and are comparable to those obtained with a graphite furnace. The sensitivities for the CCP are prebjrjinary results run under the compromise conditions listed above. It is expected that these values can be improved substantially when a full optimization study is completed. 45 Figure 2.8. Effect of support gas flow rate on Ag 1328.1 nm absorbance. Table 2.2 Sensitivities for and wavelengths for Ag, Cd, Cu, L i , and Sb. Element Wavelength (nm) Sensitivity CCP-AAS GF-AAS [2.13] ICP-AAS [2.8] (Pg) (Pg) (ppb) Ag 328.1 10 5 4306 Cd 228.8 3.5 1 4314 Cu 324.8 40 30 2746 Li 670.8 23 10 -Sb 217.6 24 20 _ 2.3.8 Absorption enhancement effect and interferences It was found that easy ionized elements (EE's) and other elements showed absorption enhancement effects in the CCP. The enhancement for silver by sodium, lithium, and gallium has been studied at a rf input power of 100 W and an argon gas flow rate of 2.5 L/min using a 4x4 mm square torch. The results are shown in Fig. 2.9.- 2.11. The emission intensities of Ag I 328.1 nm were normalized to the intensities without the presence of concomitant elements. The mass ratio of added concomitant to silver (270 pg) is used as a variable rather than the absolute mass of the element added. At low concentration of concomitant, the absorbances increased with an increase in the amount of sodium added, however, at higher concentrations of concomitant element a plateau effect is observed. The enhancement effects by lithium and gallium are similar to that of sodium. Fig. 2.12 is a schematic representation of the enhancement by ten other elements, i.e. A l , As, Ca, Cd, Co, Fe, Mg, Mo, Pb, Sb, Sn and Sr, in CCP-AAS in the extermination of silver. The mass ratio of the interference elements to silver (270 pg) was 3700. 47 1 -ft — I 1 1 1 1 1 1 1 1 — 200 400 600 800 Mass Ratio (Na/Ag) Figure 2.9. Enhancement of sodium on Ag 1328.1 nm absorbance 48 Figure 2.10. Enhancement of lithium on Ag 1328.1 nm absorbance 49 Figure 2.11. Enhancement of gallium on Ag 1328.1 nm absorbance 50 Co Fe Al Sr Sb Sn Ca Mo Pb Ql As Mg Dement Figure 2.12. Schematic representation of the enhancement on Ag 1328.1 nm absorbance 51 Al l absorbances are normalized to the absorbance of silver without the presence of concomitant elements. As indicated by Fig. 2.12, As, Mg, Cd, Mo, Ca, Pb, and Sn have strong enhancement effects on Ag 1328.1 nm absorbances. To investigate the influence of EJJEs as a function of the rf input power for Ag I 328.1 nm absorbance, the absorbances of 270 pg silver were measured in the presence and the absence of 1 u.g lithium (mass ratio : Li/Ag =3703 ) over a rf power range from 60 to 200 W using a 4x4 mm square torch. The argon flow rate was 1.9 IVmin. As shown in Fig. 2.13, the effect of rf power is similar in both cases. However, the higher enhancement was found in the low power range. To investigate the influence of argon gas flow rate on the EIE effect for Ag 1328.1 nm absorbance, the absorbances of 270 pg silver were measured in the presence and absence of 1 u.g lithium (mass ratio: Li/Ag =3703) at an rf power of 100 W using a 4x4 mm square torch. The results from 0.3 L/min to 6.4 L/min are shown in Fig. 2.14. Differences between the effects of argon gas flow rate in the presence and the absence of lithium are small. However, the lower enhancement was found in the higher flow rate range. A carrier effect is probably the dominant mechanism in the lower gas flow rate range. Interferences were studied in the presence of 1 u.g gallium as a "buffering" agent, and the results are shown in the Table 2.3. Aqueous solution of aluminium, iron and strontium were added in the forms of chloride, while the others were in the nitrate forms. After addition of gallium "buffer", the relative changes in absorbance for Ag I 328.1 nm, for most of the elements added, were in a range from 0.90 to 1.10. Chloride interferences are a notorious problem in GF-AAS. Metal chlorides in the furnace can cause gas phase interference, background interference and solid phase interference. The results shown in Table 2.3 indicate that at mass ratio of 3703 there is no significant chloride interference existing in the deterrriination of silver by CCP-AAS. 52 Figure 2.13. Influence of lithium (mass ration: Li/Ag=3703) on the effect of rf input power on Ag 1328.1 nm absorbance 0.0 2 4 Flow Rate (L/min) Figure 2.14. Influence of lithium (mass ration: Li/Ag=3703)on the effect of support gas flow rate on Ag 1328.1 nm absorbance 54 The EIE's and other elements also have an enhancement effect on the emission of Ag 1328. lnm. The enhancement and its mechanism and interference in CCP-AES are discussed in the chapter 4. Table 2.3 Interference Effects on Atomic Absorption for Ag 1328.1 nm Element Added Chemical Form Mass Ratio Relative Absorbance Fe FeCl 2 3700 0.90 Al A1C13 3700 1.03 As As(N03)3 3700 0.91 Ca Ca(N03)2 3700 1.13 Co Co(N03)3 3700 1.05 Cd Cd(N03)2 3700 1.01 L i LiN03 3700 1.07 Mo M02O3 3700 0.90 Sr S1CI2 3700 0.99 2.3.9 Precision and SRM sample analysis and analyte recovery The relative standard deviations (RSD) obtained for peak height measurements of seven replicates of 270 pg silver were 4.6 % for CCP-AAS. To check the accuracy of CCP-AAS, a stardard reference material SRM #1643b (NIST) water sample was analyzed and recovery examinations were carried out. The results are given in Table 2.4. The amount of silver found in the water sample was 9.5 ± 0.5 ng/g which fell in the certified range of 9.8 ± 0.8 ng/g. Spikes of 30 ng/g and 60 ng/g of silver were added to the SRM and recoveries were found to be in a range from 105 % to 96.2 %. Water samples 5 5 collected in Vancouver were analyzed and were found to be in a range from 2.5 ng/g to 4.2 ng/g. Table 2.4 DETERMINATION OF SILVER IN WATER SAMPLES BY CCP AES (ng/g) Sample Certified Value Amt. Found Error(%) Recovervf%) NBS 1643b 9.8 40.8 9.5 -3.0 NBS 1643b+30ng/g 41.5 105 NBS 1643b+60ng/g 67.5 96.2 rain water (Vancouver) 3.2 tank water 4.2 tap water 2.5 2.3.10 Iron excitation temperature An iron atom excitation temperature was measured using a method previously described [2.10]. A section of iron wire was introduced into the plasma at the junction of the tee to provide a source of iron atoms. The collection optics were set up to image the center of the discharge onto the entrance slit of the monochromator. Emission from a set of seven Fe (I) lines in the region 370-385 nm covering an energy range from 27000 to 35000 cm - 1 were used for this measurement. The lines used were the same as those which were outlined in reference 2.10. A Schoeffel-McPherson (Acton, MA) Model 2061 1-meter monochromator equipped with a linear photodiode array was used to carry out the measurement. The complete system has been described elsewhere [2.11]. The temperature was measured at an rf input power of 400W and a support gas flow rate of 0.6 L/Min. A 56 linear regression slope temperature indicated a temperature of 3960 ± 300 K at this power. We have previously measured Fe I excitation temperatures for a low-flow, low-power ICP system and found a temperature of 4000 K at an rf power of 400 W [2.12]. Also, for the ICP, the temperature was found to have a roughly linear relationship with power. An extrapolation to 100 W suggests that the temperature at this power should be on the order of 3000-3500 K, which is in the same range as that found in N 2 0 acetylene flames. 57 2.4 Summary The atmospheric pressure capacitively coupled plasma described in this chapter is a new atom reservoir and source for carrying out elemental analysis using atomic absorption. It has been designed for the analysis of small sample volumes of a size typically analyzed using furnace atomic absorption, however, it should also be possible to introduce dried aerosol or hydrides through the support gas inlet when continuous sample introduction is desired. The plasma discharge tube and sample introduction device allows separate control of the vaporization and atomization environments. This new spectrochemical source has a long absorption path length which provides extended analyte residence times when compared with a graphite furnace. The plasma can be operated at very low support gas flow rates which further enhances the analyte residence time. Also, since the analyte is contained in a plasma environment, vapor phase condensation is not a problem. Preliminary results of a temperature measurement yield a value of around 4000 K. At this temperature, potential chemical interferences should be minimized. The ability to operate on a variety of pure support gases and gas mixtures means that the atomization environment can be made inert, reducing, or oxidizing as the analysis situation demands. Preliminary results show that the device exhibits excellent sensitivities for many elements and these results warrant further exploration and applications of this device. 58 References [2.1] A. Montaser, and V. A. Fassel; Anal. Chem., 1976 ,48, 1490. [2.2] D. R. DcmcTsJSpectrochim. Acta, 1985,40B, 105. [2.3] R. S. Houk, V. A. Fassel, G. D. Flesch, H. J. Svec, A. L Gray and C. E. Taylor, Anal. Chem., 1980,52, 2283. [2.4] A.L.Gray, Spectrochim. Acta, 1985,40B, 1525. [2.5] R. H. Wendt and V. A. Fassel, Anal. Chem., 1966, 38, 337. [2.6] S. Greenfield, P. B.Smith, A. E. Breeze and N. M . D. Chilton, Anal. Chim. Acta, 1968, 41, 385. [2.7] C. Veillon and M . Margoshes, Spectrochim. Acta, 1968,23B, 503. [2.8] B. Magyar and F. Aeschbach, Spectrochim. Acta, 1980,35B, 839. [2.9] B. L. Caughlin and M . W. Blades, Spectrochim. Acta, 1985,40B, 1539. [2.10] M . W. Blades and B. L. Caughlin, Spectrochim. Acta, 1985,40B, 579. [2.11] Z. H. Walker and M . W. Blades,Spectrochim. Acta, 1986,41B, 761. [2.12] L. L. Burton and M . W. Blades, Appl. Spectrosc, 1986,40, 265. [2.13] C. W. Fuller/'Electrothermal Atomization for Atomic Absorption Spectrometry", Analytical Sciences Monograph, The Chemical Society, London, 1977. 59 Chapter 3 ATMOSPHERIC PRESSURE CAPACITIVELY COUPLED PLASMA SOURCE FOR ATOMIC EMISSION SPECTROMETRY* * Submitted for publication in Anal. Chem., Dong C. Liang and M. W. Blades. 60 3.1 Introduction The development of new spectrochemical sources with improved analytical characteristics continues to be a challenge in the field of analytical atomic spectroscopy. During the last two decades, several important sources such as the inductively coupled plasma (ICP), microwave induced plasma (MD?), direct current plasma (DCP), glow discharge (GD), and furnace atomic non-thermal excitation spectrometry (FANES) have been explored for spectrochemical applications [3.1-3.3]. It has long been recognized that in order to approach an ideal spectrochemical source, separation of the vaporization and excitation processes is advantageous. In recent years, a growing number of such sources, constructed from two independently energized parts, have attracted the attention of analytical spectroscopists. An electrothermal vaporizer (ETV) and an ICP combination has been used by some workers [3.4-3.5]. Other combinations include: 1) glow discharge sputtering - rf plasma [3.6], 2) ETV - MD? [3.7], 3) laser ablation - MIP [3.8], 4) ETV - GD, i.e. FANES [3.9], 5) spark vaporization - ICP [3.10] and 6) arc vaporization - ICP [3.11]. In chapter 2 an atmospheric pressure rf capacitively coupled plasma (CCP) has been reported as being potentially useful for atomic absorption spectrometry (AAS) [2.12]. Variants of this approach, including systems for gas chromatagraphic detecher (GC) [2.13], the combinations of laser ablation - CCP [3.14], GF-CCP, an atmospheric pressure CCP formed inside a graphite furnace [3.15], and rf sputtering - CCP direct solid analysis, and the application as an intense spectral lamp [3.16] have been reported in other chapters in this thesis. The torch design provides very effective energy transfer from the power supply to the plasma by capacitive coupling. As a result, the plasma can be generated at atmospheric pressure and the geometry has been shown to be quite flexible. The plasma can be operated over a wide range of rf input powers (10-600 W), which allows for optimum conditions for atom resonance line absorption and emission measurements. The 61 discharge is formed in an enlongated quartz tube and operates at a low support gas flow rate both of which allow a relatively long residence times of analyte atoms. Further investigation has indicated that the CCP is an excellent source for AES. This chapter describes the characteristics of CCP-AES, and compares them with those of CCP - AAS. 62 3.2 Experimental 3.2.1 Modification of the C C P torch The CCP torch described in the previous paper [3.12] provided the basic working model for generating a radio frequency plasma using capacitive coupling. When that torch was operated at high rf powers, the rf potential between the two electrodes was, at times, high enough to break down the air medium outside the plasma torch and cause rf arcing. Moreover, periodically, an arc could be observed between the plasma column and the tantalum strip vaporizer, especially when easily ionized elements (EJE) were vaporized from the tantalum strip into the plasma. Both of these types of arcing changed the plasma operating conditions substantially. To solve the latter problem a 5 cm long quartz capillary with an inner diameter of 0.08 cm was used to connect the tantalum strip housing to the plasma. The increased distance between the plasma column and the tantalum strip increased the potential required to initiate rf breakdown. An additional advantage of using a capillary is that the small dead volume reduces peak broadening. To prevent arcing between the two electrodes outside the plasma torch, several plasma tube structures have been tested. One design used a fin structure between the electrodes, as shown in Fig. 3.1a. This design could sustain higher rf input powers than the previous design [3.12], but it was not very effective in controlling the arcing. In another design, two rod-shaped electrodes were enclosed by quartz tubes, as shown in Fig. 3.1b. This was a very effective way to control the arcing, however, the plasma generated was not uniform. It was found that the most effective design was a square quartz tube for the plasma sandwiched between two rectangular quartz tubes which were used to enclose the strip electrodes, as shown in Fig. 3.1c. This design solved the problem of the design shown in Fig. 3.1a and Fig. 3.1b and was used to obtain all of the results reported in this study. 63 C r o s s s e c t i o n v i e w t h r o u g h T o DC p o w e r s u p p l y Figure 3.1. Schematic diagram of the modified capacitively coupled plasma torch (a) Torch with fin structure, (b) Tree round tube torch, (c) Square tube torch 64 3.2.2 Equipment and set up The equipment and experimental set up employed in this chapter are summarized in Table 3.1. 3.2.3 A n a l y t i c a l procedure All measurements were carried out using the following procedure. Using an adjustable micropipette (Eppendorf, 0.5 -10 ul Ultra Micro), a 5 -10 ul aqueous sample was placed on the Tantalum strip through the inlet port on the side of the quartz body. The rf was not applied at this stage. The sample was dried, ashed, and atomized. The plasma was ignited at the end of the ash cycle by application of the rf power. Data were collected through the atomize cycle. After each atomize cycle the atomizer was tested for memory effects. 3.2.4 S t a n d a r d solutions All analytical standards were prepared using Fisher Scientific (Nepean, Ont,) 1000 ppm atomic absorption standards. The solutions were diluted to volume using 1% HNO3. 65 Table 3.1. Experimental Facilities and Operating Conditions Plasma Power Supply Plasma gas Sample Vaporization Spectrometer Slits Detector Electronics Data Acquisition Perkin-Elmer ICP 5500 System consisting of a Plasma-Therm (Kreeson, N.J.), HFP-2500F RF generator, AMN-2500E automatic matching network, APCS-3 automatic power control system and PT - 2500 torch box. Argon. Tantalum strip 1 cm by 0.4 cm with a depression at the center. Power: Varian Model CRA-61. Normal operating cycle: Dry -105 °C for 60 s., Ash - 300 - 600 °C for 15 s, Atomize - 2000 - 2700 °C for 2 s. Schoeffel-McPherson (Acton, MA) Model 270,0.35 m Czerny-Tumer mount scanning monochromator with 1200 rulings/mm holographic grating. Reciprocal linear dispersion f 20 A/mm in the first order. Entrance and exit slits set to 50 um, spectral band width 100 pm. The photocurrent from a Hammatsu R955 photomultiplier tube was amplified by a home-built preamplifier. The photomultiplier tube was powered using a McPherson Model EU-42A PMT power supply. Digital data acquisition was carried out using a Tulsa Computers (Owasso, OK) Telex Model 1280 D3M-AT compatible computer equipped with a RC Electronics (Santa Barbara, CA) Model ISC-16 analog-digital converter, running the RC Computerscope software 66 package. Analog data were acquired using a Servocorder 2 1 0 chart recorder. 67 3.3 Results and discussion 3.3.1 Effect of rf input power The rf input power is an important parameter in the optimization of experimental conditions. The effect of changes in the rf input power on the absorbance of the Ag I 328.1 nm line using a round tube torch was reported in a previous chapter and elsewhere [3.12]. The effect of changes in rf input power on both absorbance and emission intensity of Ag 1328.1 nm was studied using an argon flow rate of 1 L/min and a 6 x 6 mm square tube torch. The results over an rf power range from 60 W to 300 W are provided in Fig. 3.2. The results for absorption are similar with those reported previously chapter and [3.12]. However, the effect of rf input power is less pronounced for emission compared to absorption. At rf powers below 100 W, both absorbance and emission signals decrease sharply. The absorbance signal is a maximsed at about 100 W, and the decreases steadily with increasing power. In contrast, the emission signal is fairly constant between 100 to 200 W, and thereafter decreases slightly with increasing rf power. The effect of rf input power on the B I 249.81 nm emission intensity is provided in Fig. 3.3. The measurements were taken at an argon gas flow rate of 1.25 L/min and rf input powers ranging from 30 to 500 W. Below an rf input power of 60 W, B1249.81 nm emission intensities were almost not distinguished from background With an increase in rf input power, the intensity increased Clearly, the rf input power has different effects for silver and boron. The excitation and ionization energies for boron are 4.91 eV and 8.3 eV and for silver, 3.66 eV and 7.57 eV respectively. The dissociation energies for the BO and AgO are 8.0 eV and 2.0 eV respectively. The dissociation energy for BO is very high, and dissociation is one of the dominant processes determining free atom ground state populations. Therefore boron requires a much higher rf input power to achieve a high degree of dissociation compared with silver. 68 Figure 3.2. Effect of rf input power on Ag 1328.1 nm absorbance and emission intensity (a) Emission, (b) Absorption. 69 600 rf Power(W) Figure 3.3. Effect of rf input power on B 1249.8 nm emission intensity 70 The effect of rf input power on the background emission intensity at 328.1 nm (corresponding to the Ag I wavelength) has been measured for an argon gas flow rate of 1.3 L/min and a spectral band width of 0.5 nm using a 6 x 6 mm square torch. The results are provided in Fig. 3.4a. The background intensities were found to have a linear relationship with input power. A measurement of the effect of rf input power on the background intensity was repeated at 249.8 nm (corresponding to the B I wavelength). The results in the range from 60 W to 500 W are shown in Fig. 3.4b. The background emission intensity increased from 60 W to 300 W and reached its maximum, then decreased with increasing rf power. A molecular emission band of NO with a band head at 247.9 nm was found. The effect of rf input power on the intensities of the NO molecular emission band head (Fig. 3.4c) has a similar behaviour to that in 0>) and it is likely that the background emission behavior at 249.8 nm is the result of this NO band. The effect of rf input power on the intensities of background emission is complex. In a plasma source, background emission mainly arises from blackbody radiation, molecular emission, ion-recombination continuum, chemical association continuum, and Bremsstrahlung. The background emission at the 328.1 nm and the 249.8 nm positions have different behaviours, probably due to the fact that there was no molecular band emission around the Ag I 328.1 nm position, whereas the B I 249.8 nm line fell on the wing of a NO molecular emission band with a band head at 247.9 nm. The background intensity increased when the power was changed from 60 to 300 W, then it decreased with increasing rf input power. The dissociation of the NO molecule reduces the intensity of the background at 249.8 nm at high rf input powers. 71 400 i • r Figure 3.4. Effect of rf input power on background emission intensity (a) 328.1 nm, (b) 249.8 nm, (c) Band head at 247.9 nm. 72 3.3.2 Effect of plasma gas flow rate Independent control over vaporization and excitation provides a mechanism for independent optimization of both processes in a source. The carrier gas not only acts as a plasma support gas, but also carries the analyte species from the vaporizer into the plasma in which the excitation takes place. The residence time of analyte atoms is determined mainly by the combination of carrier gas flow rate and diffusion effects. The effect of argon gas flow rate on the absorbance of the Ag I 328.1 nm line using a 4 mm diameter round torch has been discussed previously chapters and [3.12]. The effect of changes in argon gas flow rate on both absorbance and emission intensity of Ag I 328.1 nm was studied at an rf input power of 150 W using a 6 x 6 mm square torch. As indicated in Fig. 3.5, the results for absorbance are similar to those reported previously chapter and [3.12]. The effect of changes in argon gas flow rate on the emission intensity of B I 249.8 nm is provided in Fig. 3.6. The measurements were carried out at an rf input power of 500 W and a gas flow rates from 0.3 to 7.7 IVmin. The emission intensity is affected by both changes in the carrier gas flow rate, hence residence time, and also by changes in the plasma properties as the gas flow rate is changed. As indicated in Fig. 3.6, a maximum emission intensity was obtained at a support gas flow rate of 1.3 L/min. The effect of argon gas flow rate on the background emission intensity, and the peak to peak noise at the 249.8 nm wavelength position were also examined at an rf input power of 500 W. As indicated in Fig. 3.7, while the effect of gas flow rate on the background emission is negligible, the peak to peak noise increased with increasing gas flow rates from 0.3 L/min to 5 L/min and reached its maximum between 5 IVmin and 6.3 L/min, then decreased slowly. This phenomenon could be explained by turbulence in the gas flow pattern. In general, the lower the gas flow rate, the lower the noise. 73 Figure 3.5. Effect of support gas flow rate on Ag 1328.1 nm absorbance and emission intensity (a) Emission, (b) Absorption. Figure 3.6. Effect of support gas flow rate on B1249.8 nm emission intensity Figure 3.7. Effect of support gas flow rate on background emission intensity and peak to peak noise for B1249.8 nm measurement (a) Background intensity, (b)P-P Noise. 3.3.3 Emission enhancement effect and interference The enhancement effect by easy ionized elements (EIE) in ICP [3.17], DCP [3.18-3.19] and MB? [3.20] has been long known. The mechanism of the effect is different from source to source. The interferences probably arise from ionization suppression, changes in the rates of excitation and de-excitation processes, changes in sample introduction efficiency and transport processes. However a complete understanding is still elusive. Emission enhancement by EIE's and other elements was found in the CCP. The enhancement for silver by sodium, lithium, calcium and gallium has been studied at an rf input power of 100 W and an argon gas flow rate of 2.5 IVmin using a 4 x 4 mm square torch. The results are shown in Fig. 3.8. The emission intensities of Ag 1328.1 nm were normalized to the intensities without the presence of concomitant elements. The mass ratio of added concomitant to silver (270 pg) is used as a variable rather than the absolute mass of the element added. At low concentrations of concomitant, the emission intensity increased with an increase in the amount of sodium added. However, at higher concentrations of concomitant element a plateau effect is observed. The enhancement effects by lithium, calcium and gallium are similar to that of sodium. The mechanism of the enhancement effect is very complex and is, to a degree, dependent upon the nature of the element in question and the concomitant added. The resulting effect probably includes one or all of the following mechanisms: (1) assisted vaporization (a carrier effect), (2) shifts in ionization equilibrium (3) changes in excitation efficiency; and (4) changes in plasma parameters, such as temperature and electron number density. To investigate the influence of EJEs as a function of the rf input power for Ag I 328.1 nm emission, the intensity of 270 pg silver was measured in the presence and the absence of 1 u.g lithium (mass ratio : Li/Ag =3703 ) over an rf power range from 60 to 200 W using a 4 x 4 mm square torch. The argon flow rate was 1.9 L/rnin. As shown in Fig. 3.9, 7 7 CO c e o co CA E tt « tt 0 200 400 600 800 Mass Ratio (Na/Ag) cn C e o 1 10 E tt « cu 1000 2000 Mass Ratio (Ga/Ag) 3000 CO B 4> B .2 "to CC •mm E tt J5 "35 tt -i 1 1 1 • i 1-0 1000 2000 3000 4000 Mass Ratio (Li/Ag) CO E cy B O E tt cu CU tt 1000 2000 3000 4000 Mass Ratio (Ca/Ag) Figure 3.8. Enhancements of sodium, lithium, calcium and gallium on AgI328.1 nm 78 100 Figure 3.9. Influence of E E on the effect of rf input power on Ag 1328.1 nm emission intensity (a) Without L i , (b) With L i . 79 the difference between the effects of rf input power in the presence and the absence of lithium is small. The enhancement factor for lithium was about 1.5 which agreed with the results provided in Fig. 3.8 . The decrease in emission intensity at 60 W rf power is smaller in the 4 x 4 mm square torch than in the 6 x 6 mm square torch (see Fig. 3.2). It is expected that, for a given rf input power, the power density in the 4 x 4 mm square torch is higher than that in the 6x6 mm square torch since the plasma volume is smaller. To investigate the influence of the argon gas flow rate on the EIE effect for Ag 1328.1 nm emission, the intensities of 270 pg silver were measured in the presence of 1 u.g lithium (mass ratio : Li /Ag =3703) at an rf power of 100 W using a 4x4 mm square torch. The results from 0.3 IVrnin to 6.4 L/rnin are shown in Fig. 3.10. Compared with the results in Fig. 3.5, the difference between the effects of argon gas flow rate in the presence and the absence of lithium is very small. Fig. 3.11 is a schematic representation of the interferences of thirteen elements, i.e. Co, A l , As, Ca, Cd, Fe, L i , Mg, Mn, Mo, Sb, Sn and Sr, in CCP-AES in the deterrnination of silver. The results were obtained from 270 pg silver in the presence of 1 u.g gallium as a "buffering" agent. The mass ratio of the interferent elements to silver was 3700. Al l emission intensities are normalized to the intensity of the pure silver solution. Aluminium, iron, antimony, tin and strontium were added in the forms of chloride, while the others were in the nitrate forms. As indicated by Fig. 3.11, the relative change in intensity for Ag I, for all the elements added, were in a range from 0.90 to 1.10. Most of these, except for manganese and antimony, were in a range from 0.95 to 1.05. Chloride interferences are often seen in GF-AAS. Metal chlorides in the furnace can cause gas phase interference, background interference and solid phase interference. The results shown in Fig. 3.11 indicate that at mass ratio of 3703 there is no significant chloride interference existing in the (^termination of silver by CCP-AES. 8 0 ^ 120 C a o B Flow Rate (L/min) Figure 3.10. Influence of EIE on the effect of support gas flow rate on Ag 1328.1 nm emission 81 2 . 0 ' *? 1-5 H "5 B cu cu 0.5 H 0.0 IllililliBK Co Al As Ca Cd FB Li Mg Mn Mo Sb Sn Sr Element Figure 3.11. Interference effect of thirteen elements on Ag 1328.1 nm emission intensities by CCP-AES 82 3.3.4 Detection Limits The detection limits obtained by CCP-AES for Ag, Cd, Cu, L i , Sb, and B using peak height measurement are given in Table 3.2. The characteristic mases (1% absorption sensitivity) obtained by CCP-AAS and those obtained by GF-AAS are also given in the Table 3.2. Table 3.2. Characteristic mases and detection limits for Ag, Cd, Cu, L i , Sb and B using CCP-AES, CCP-AAS, and GF-AAS. Element Wavelength (nm) Characteristic mass (pg) Detection Limit (pg) CCP-AAS GF-AAS [3.13] CCP-AES Ag 328.1 10 5 0.7 Cd 228.8 3.5 1 0.7 Cu 324.8 40 30 I i 670.8 23 10 2 Sb 217.6 24 20 80 B 249.8 2000 1000 400 3.3 .5 Precision and Signal to Noise Ratio in AES and AAS Modes Fig. 3.12 displays the time resolved signals for emission and absorbance of 100 ng boron at 249.8 nm obtained using the time-resolved data acquisition protocol described previously (chapter 2 and [3.12]). The peak heights of both signals were adjusted to be almost the sameby changing the display factors. It is obvious that the baseline noise in the AAS mode is much higher than that in the AES mode. The same situation can be found for 83 Figure 3.12. Absorption and Emission signals for 100 ng Boron using the 249.8 nm line. other elements such as silver and magnesium, etc. The precision of CCP-AES and CCP-AAS has been compared by acquiring both emission and absorbance signals from the same atomization cycle. The relative standard deviations (RSD) obtained for peak height measurements of 7 replicates of 270 pg silver were 1.5 % for CCP-AES and 4.6 % for CCP-AAS. It was found that the signal to noise ratios (S/N) were higher in the AES mode than those in the AAS mode for the same CCP atomizer. 3.3.6 Calibration Curves and Analyte Recovery Emission calibration curves for silver exhibited very good linearity with a regression coefficient of 0.999 over the concentration range 0-1200 pg of total analyte. This is a indication of the low degree of self-absorption in the CCP torch. This characteristic can be explained by the fact that the CCP provides a fairly uniform plasma environment [3.12]. To check the accuracy of CCP-AES and CCP-AAS, a SRM #1643b (NIST) water sample was analyzed and recovery examinations were carried out. The results are given in Table 3.3. The amount of silver found in the water sample was 9.3 ± 0.5 ng/g which fell in the certified range of 9.8 ± 0.8 ng/g. Spikes of 30 ng/g and 60 ng/g of silver were added to the SRM and recoveries were found to be in a range from 97 % to 104 %. 85 Table 3.3 DETERMINATION OF SILVER IN WATER SAMPLES BY CCP-AES (ng/g) Sample Certified Value Amt. Found Errorf%) Recoverv(%) SRM 1643b 9.8 ±0.8 9.3 ±0.5 -5.0 SRM 1643b+30 ng/g 38.9 97 SRM 1643b+60 ng/g 72.2 104 86 3.4 S u m m a r y The atmospheric pressure capacitively coupled plasma described in this paper is a new atom reservoir and source for carrying out elemental analysis using not only atomic absorption but also emission spectroscopy. This new spectrochemical source has a long plasma path length which provides extended analyte residence times when compared with a graphite furnace. The plasma can be operated at very low support gas flow rates which further enhances the analyte residence time. Efficient coupling of the applied rf power using capacitive coupling enables the generation of a plasma at atmospheric pressure and in a flexible geometry. The chemical interference for Co, Al, As, Ca, Cd, Fe, Li, Mg, Mn, Mo, Sb, Sn and Sr, are negligible in the determination of silver if a buffering agent is used. A chloride interference, which is prevalent in GF-AAS, was not found. The plasma discharge tube and sample introduction device allows for the separate control vaporization and excitation. The giving detection power of the CCP is superior detection power when compared with ETV-ICP. Furthermore, the signal to noise ratio and the precision in the CCP emission mode are superior those in the absorption mode. These features combined with the simultaneous multielement deterrnination capability gives CCP - AES an advantage over CCP-AAS. 87 References [3.1] Gunther T61g, Analyst., 1987 , 112 365. [3.2] A. T. Zander, Anal. Chem., 1986, 58 1139A. [3.3] P. R. Banks and M . W. Blades, Spectrochim. Acta.,19&9, 44B 1117. [3.4] D. E. Nixon, V. A. Fassel and R. N . Kniseley, Anal. Chem., 1974, 46 210. [3.5] H. Matusiewicz, / . Anal. At. Spectrosc, 1986, 1.171. [3.6] P. E. Walters and H. G. C. Human, Spectrochim. Acta,l9Sl, 36B 585. [3.7] J. F. Alder and M . T. C. Da Cunha, Can. J. Spectrosc, 1980, 25 32. [3.8] T. Ishizuka and Y. Uwarnino, Anal. Chem., 1980, 52 125. [3.9] H. Falk, E. Hoffmann, and Ch. Ludke, Spectrochim. Acta , 1984, 39B 283. [3.10] A. Aziz, J. A. C. Broekaert, K. Laqua and F. Leis, Spectrochim. Acta,19&4, 39B, 1091. [3.11] R. Sing and E. D. Salin, ICP Inf. Newsletter 1982, 8 240. [3.12] Dong C. Liang and M.W. Blades, Anal. Chem. 1988, 60 27. [3.13] Degui Huang, Dong C. Liang and M.W. Blades , / . Anal. At. Spectrom., 1989, 4, 789. [3.14] Dong C. Liang and M . W. Blades, Appl. Spectrosc, in press. [3.15] Dong C. Liang and M . W. Blades, Spectrochim. Acta 1989,44B 1059. [3.16] Dong C. Liang and M . W. Blades, Spectrochim. Acta 1989,44B 1049. [3.17] M . W. Blades and G. Horlick, Spectrochim. Acta 1981, 36B 881. [3.18] L. J. Prell, C. Monnig, R. E. Harris and S. R. Koirtyohann, Spectrochim. Acta 1985 , 40B, 1401. [3.19] D. D. Nygaard and T. R. Gilbert, Appl. Spectrosc, 1981, 35 52. [3.20] M . H. Miller, Spectrochim. Acta, 1984, 39B.13. [3.21] S. Murayama, Spectrochim. Acta, 1970, 25B 191. 88 Chapter 4 ATMOSPHERIC PRESSURE CAPACITIVELY COUPLED PLASMA SPECTRAL LAMP AND SOURCE FOR THE DIRECT ANALYSIS OF CONDUCTING SOLID SAMPLES - APPLICATION OF RF SPUTTERING* * Part of this chapter was published in Spectrochim. Acta 1989,44B, 1049, Dong C. Liang and M.W. Blades. 8 9 4.1. Introduction In previous chapters the development and characterization of an atmospheric pressure, capacitively coupled plasma torch for atomic absorption spectrometry [4.1] and for atomic emission spectrometry [4.2] were described. The configuration described previously was designed for the analysis of small volumes of liquid samples of a size typically analyzed by electrothermal atomization AAS (5 - 50 pX), however, the CCP can also be combined with other sample introduction techniques including laser ablation [4.3]. Further exploration into the applications of this device has shown that the CCP has potential both as an intense spectral lamp and as a source for direct solid sample analysis. In this chapter the term "lamp" is used to describe devices primarily designed to emit spectra for use as primary sources in AAS and atomic fluorescence spectrometry (AFS) as -veil as other optical measurements whereas "source" describes devices primarily intended to be used as a means of vaporizing and exciting samples for analysis by emission methods. By far the most important commercial spectral lamp for AAS and AFS is the hollow cathode lamp (HCL). The main advantages of the HCL are its very small spectral line-width and its high signal-to-background ratio. However, the absolute intensity of emission from the HCL is relatively low compared with the radiation from other plasma sources. To overcome this problem techniques such as direct current (dc) boosted-HCL, rf boosted-HCL, microwave coupled HCL, and high current pulsed HCL have been developed [4.4]. Additionally, the intensities of ion lines in HCL's are very weak, due to the doininant population of ground state atom in glow discharges [4.5]. The factors contributing to the relatively low sensitivities of ICP-AAS have been discussed previously [4.1]. One of the factors is that traditional AAS primarily makes use of atomic resonance lines; however there is a large population of ground state analyte ions in ICP's even at relatively low powers [4.6]. The development of an intense ion line spectral source has some significance in this 90 area in that it could assist in the reduction of source induced shot noise, consequently improving the detection limits for plasma source AAS. There has recently been much interest in the application of sputtering sources in atomic spectrometry [4.7-4.9]. Sputtering is the ejection of material from a surface caused by bombardment with an energetic beam of particles [4.10]. Dc sputtering in a glow discharge source allows one to analyze solid samples by atarnizirig the analytes directly from the solid state. This approach offers some advantages. The time-consuming sample dissolution step can be omitted and analysis can be carried out without addition of reagents and without any separation and/or concentration steps. This allowsthe risks of introducing contaminants and of loss of the element to be determined to be considerably reduced and as a consequence an analysis can be carried out quite rapidly. It would appear that the sputtering rate should be a direct function of gas pressure, since the higher the pressure the more ions which would be available for sputtering. However, sputtering is usually carried out at pressures between 5xl0" 3 and 1 torr since glow discharges extinguish or switch over to arc discharges at higher pressures and the main sampling mechanism in arcs is thermal evaporation. Although rf sputtering is not widely used as a sample introduction method in atomic spectroscopy, it has long been recognized as an important technique in sputter etching and chemical vapour deposition [4.10]. Rf sputtering at low pressures first suggested by Wehner in 1955 [4.11] and demonstrated in 1962 by Anderson et al.[4.12] has become a standard method for etching materials in the semiconductor industry. Atmospheric pressure rf sputtering was previously used to supply Fe to the CCP discharge for the purpose of making temperature measurements [4.1]. More recently, Stephens [4.13] has described an rf discharge between two metal electrodes at atmospheric pressure, operating in helium at a power of 5-30 W. The sputtering effect of the discharge was deduced by observing atomic emission from the plasma and atomic absorption within the plasma. Stephens pointed out that this device offered a convenient means of observing either 91 emission or absorption for those elements for which sputtering was not inhibited by the presence of a stable oxide layer. The novel configuration of the CCP torch described previously [4.1] can be modified in order to carry out atmospheric pressure rf sputtering. In this chapter the original CCP design has been modified to provide an intense spectral lamp and an excitation source for direct solid sample analysis. Two different configurations are described, one which operates primarily as a lamp and one which operates primarily as a source. The intensities of the CCP spectral lamp have been found to be up to 2-3 orders of magnitude greater than those from an HCL. The CCP spectral lamp emits not only atom lines but also intense ion line spectra. Moreover, the lamp can be used as a source for the direct analysis of solid samples without any significant structural modification. Its analytical performance has been demonstrated through the determination of manganese in steel samples using standard (NBS) low-alloy steel samples. This modified CCP torch allows for a wide selection of plasma conditions, good control of sampling and excitation, and ease of interchange of samples for direct solids analysis applications. 92 4.2 Experimental 4.2.1 Lamp and source design A schematic diagram of the CCP lamp is provided in Fig. 4.1. The torch was constructed of quartz glass with an internal diameter of 0.65 cm and a wall thickness of 0.10 cm. A cylindrical stainless steel electrode 7 cm in length and 0.9 cm in diameter was placed around the outside of, and in contact with, the quartz tube. A central electrode was inserted into a male B8 standard tapered quartz joint and was sealed by two silicon rubber o-rings. The two quartz parts were connected together by a B8 standard quartz joint. The two electrodes were connected to the output of an rf power supply. The outside electrode and the central electrode were overlapped by about 15 mm in length (see Fig. 4.1). This arrangement serves to increase the current density on the surface of the central electrode and is important when high melting point metals are used as central electrodes. Plasma support-gas was introduced through an inlet which is connected to a tantalum strip vaporizer housing [4.1] by standard a B5 quartz tapered glass joint. This CCP lamp can also work as an excitation source for direct solid analysis by replacing the central electrode with electrically conducting sample rods. This lamp has a simple structure and can be easily made, however, the device has some inherent disadvantages. The temperature of central electrode is controlled by the power supporting the plasma itself. It was found that the central electrode would melt at a high rf input powers when low melting point materials (Cu, brass, etc.) were used as central electrodes. To overcome this problem an alternate CCP torch configuration, shown in Fig. 4.2, was developed. This design is a slight modification of the CCP torch described in the previous paper [4.1]. Sample electrodes were inserted into the plasma through a central sampling port. Teflon tape was wrapped on the sample rod to provide a gas seal. A 25 pF air-medium variable capacitor was placed in series with the sample electrode and was connected to the rf power supply ground. Since the plasma potential is higher than ground, the electrode provides a path for rf current. The main advantage of this design is that the rf \ Support Gas Inlet Figure 4.1. Schematic diagram of the CCP spectral lamp Sample electrode Plasma 25 pf Cross section view through the center Quartz torch Stainless steel electrodes (to rf power supply) Plasma gas inlet Figure 4.2. Schematic diagram of the CCP torch used for direct solids analysis 95 current in the sample rod can be controlled using the capacitor which in turn controls the sample rod temperature and the sampling rate thus preventing the sample rod from melting at high input powers. Another advantage is that the device can be used with both emission and absorption measurement systems. A disadvantage is that the torch works at gas flow rates higher than those in the CCP lamp (Fig. 4.1) because the torch has two open ends to the atmosphere. 4.2.2 Equipment a n d facilities The equipment and experimental set-up employed in this research were similar to those listed in Table 4.1 4.2.3 M e t h o d f o r comparison of intensities of C C P spectral lamp a n d H C L Because the light emitted from the CCP spectral lamp and the HCL have different spatial shapes, and in order to eliminate possible error caused by poor alignment of optical elements, the optical system used for intensity comparison was set up without using any lenses. Both the CCP spectral lamps and the HCLs are placed in the monochromator optical axis at a distance of 37 cm from the entrance slit. Before spectral measurements were made the positions of both lamps were optimized to give maximum signals. 4.2.4 A n a l y t i c a l procedure All the metal samples were fabricated with a diameter of 3.2 mm (1/8 in) and about 51 mm (2 in) long, and were polished by fine sand paper, then washed with deionized water. The sample rods were placed into the CCP spectral lamp with an overlap of 15 mm with the outside cylindrical electrode. The sample rods placed into the torch (Fig. 4.2) reached the plasma but did not block the optical aperture. After turning on the rf power for two rninutes and pre-heating the samples, the intensities of analytical lines and internal standard lines were recorded, using a chart recorder or a computerscope, and scanning the monochromator. The intensity ratios of the analytical line and the internal standard line were used to determine the concentration of analyte in the samples. A set of NBS low-alloy steel samples of 600 series were selected for establishing calibration curves for direct solid sample analysis. The NBS 600 series standards are 3.2 mm (1/8 in) in diameter and 51 mm (2 in) long. Five standards in the series, i.e. 661 through 665, cover a wide range of concentrations for many elements and are suitable for calibration for steel sample analysis. 97 4.3 Results a nd discussion 4.3.1 D e s c r i p t i o n of C C P spectral lamp and C C P direct so l i d analysis s o u r c e The plasmas in the CCP lamp and torch are very similar to those described in the previous chapters and in [4.1] except that the color of the plasma is modified by the components of the sample electrode. For the lamp, the plasma fills the entire volume of the quartz tube surrounded by the cylindrical electrode including the overlap region between the two electrodes. There is no wiring between the sample rod and the rf power supply. The sample rod is immersed in the plasma and picks up the rf current. The variable air medium capacitor was in series to ground. Changes of capacitance vary the impedance of the circuit, and control the rf current density on the surface of the sample rod, and consequently the sampling rate. At low rf power, the sample electrodes are dark or dark-red and atmospheric pressure rf sputtering is the dominant sampling mechanism. With an increase in rf power, the electrode color changes through orange to white-hot. Under this condition sampling takes place by a combination of both rf sputtering and thermal evaporation. 4.3.2 Effect of r f input power The effect of changes in rf input power on the emission intensity of the 213.9 nm Zn I line was studied at an argon flow rate of 0.63 L/Min. A CCP spectral lamp (Fig. 4.1) equipped with a brass central electrode was used in this study. The results over the power range 50-200 W are shown in Fig. 4.3. The emission intensity increases with an increase in rf power. When the rf power is increased above 75 W, the central electrode becomes hot enough to cause thermal vaporization. In this case, the intensity increases dramatically. If the rf power was continuously increased above 200 W, the brass electrode would melt down. However the modified CCP torch shown in Fig. 4.2 does not experience this problem. By decreasing the capacitance of the variable air medium capacitor, the CCP 9 8 60 0 H • 1 • 1 • 1 • r 0 50 100 150 200 250 rf Power(w) Figure 4.3. Effect of rf power on Zn 1213.9 nm emission intensity 9 9 torch shown in Fig. 4.2 can run at 500 W rf input power without melting the sample electrodes. 4.3.3 Comparison of intensities of CCP spectral lamp and HCL The intensities of resonance lines from the CCP spectral lamp and the HCL have been compared by the method outlined in section 4.2.3. The lamp operating parameters and the wavelengths of the resonance lines, as well as the emission intensity ratio (measured total intensity ratio) of the CCP spectral lamp and the HCLs, are listed in Table 4.1. Most of the HCL's were operated at the maximum currents specified by the manufacturers. The intensity measurements of zinc and copper for the CCP spectral lamp were carried out by placing a brass central electrode in the lamp; the intensity measurements of manganese, chromium and iron were carried out by using a stainless steel central electrode. A l l the resonance lines from the CCP spectral lamps studied in this research are one to two orders of magnitude stronger than those of the HCL's. Since the intensity differences of the CCP spectral lamp and the HCL are so large, the intensity measurements were carried out at different photomultiplier tube (PMT) supply voltages. The log gain of the PMT has a linear relationship with the PMT supply voltages. This relationship was examined under the experimental conditions, and was used for the normalization of the measurements. It is expected that larger intensity ratios can be obtained by using a pure metal rod as the central electrode. It is well known that using a primary source with broad band lines will lead to considerable sensitivity loss in AAS. This is due to the fact that the absorption decreases very rapidly when the source line profile is beyond the absorption profile. Magyar [4.17] derived a equation of sensitivity (SA) for AAS: (SA) = (2.175 x 10s cm/ umol) Pad-Y)P M A (b/D) (8V + 8^.s2)i/2 ( g f ) P e V (1) 100 where sensitivity SA is the slope of the calibration curve. The term (8X a 2 + 6XS2 )l/2 accounts for the effect of half-intensity width of source line profile SXg and half-intensity width of absorption profile Sk^ on the sensitivity. Other terms in the equation are refered in Magyar's paper [4.17], To estimate the physical band widths, 8A<x;pand 8X.HCL. die spectral lines emitted from the CCP and HCL were assumed to have pure Doppler profiles. Then the line width ratio of the CCP and the HCL for same spectral line is expressed as: For an extreme case, Doppler temperatures of 4000 K and 500 K were assumed for the CCP and the HCL respectively, giving a line width ratio of 2.8. Using this factor, unit line width (ULW) intensity ratios were calculated and are listed in Table 4.1. Considerable high intensity gains can still be found from the CCP lamp refering to the unit line width intensity. 1 0 1 Table 4.1. Intensity Ratio of resonance Line of CCP Lamp and HCL Element Wavelength (nm) Operating parameter for CCP lamp Operating parameter for HCL Intensity Ratio TTR* ULW# Z n l 213.86 rf power 60 w flow rate: 1.3IVmin P-E lamp dx.: 15 mA 369 130 F e l 248.33 rf power 250 w flow rate: 0.3 L/min P-E lamp dx.: 12 mA 246 87 M n l 279.48 rf power 250 w flow rate: 0.3 L/min P-E lamp dx.:12mA 84 30 CrI 357.87 rf power 250 w flow rate: 0.3 L/min P-E lamp dx.: 15 mA 13 4.6 C u l 324.75 rf power 60 w flow rate: 1.3 L/min Hamamatsu lamp d.c: 3 mA 13 4.6 * TTR - Total intensity ratio of the CCP to HCL. # ULW - Unit line width intensity ratio of the the CCP to HCL. Source induced PMT shot noise is one of the major noise sources in an atomic absorption spectrometer and contributes to the detection limit in AAS. In the case of shot noise, signal-to-noise ratios increase as the square root of source intensity [4.14]. Therefore, the development of intense spectral lamps is an important way to improve the detection limits for AAS. Rf electrodless discharge lamps (EDL) have an intensity advantage over the HCL. However, they are only suited to excite a few elements with relatively high vapor pressures such as As, Cd, Hg, Pb, Sb, Se and Zn etc.. For example, the zinc EDL is one of the best EDL's, and gives an approximate intensity gain of 30 relative to the HCL [4.15]. By comparison the CCP spectral lamp with a brass central electrode is 369 times more intense compared with the zinc HCL. The high yield of atmospheric pressure sputtering is probably one of main causes for the high intensities of the CCP spectral lamps. A significant improvement of signal to noise ratio for the Zn I 213.9 nm line was found in the CCP spectral lamp compared to the HCL as shown in 102 Fig. 4.4. The content of zinc in the brass is approximately 30-35 %. With a pure zinc central electrode and higher rf input power, higher emission intensity from zinc would be obtained. On the other hand, the intensity advantage of the CCP spectral lamp provides for the possibility of intense multielement lamps by using metal alloys for the central electrode. As can be seen in Table 4.1, the emission lines of Fe, Mn and Cr, from a stainless steel central electrode CCP spectral lamp, are significantly more intense than those of HCL's. When a multielement HCL is used, the detection limits usually become worse, due to the intensity loss in the multielement HCL's. 4.3.4 Ion spectra in the C C P lamps Typical emission intensities of manganese ion and atom lines in the CCP spectral lamp and the HCL are provided in Fig. 4.5. The experimental conditions are listed in the caption. The intensities of manganese ion lines are very weak in the HCL. The intensity ratio of Mn H 257.6 nm to the Mn I 279.5 nm line is 0.077 in the HCL and 0.77 in the CCP lamp. Since the material volatilized in a glow discharge is to a large extent present as a vapor cloud of free atoms [4.5], the small population of ions in HCL's results in very weak ion lines. Although the ion population in the CCP has not been measured, the intensity of Mn U 257.6 nm in the CCP spectral lamp is about three orders of magnitude higher than that in the manganese HCL. Thus this lamp has potential as a source for use in ion line absorption spectroscopy (ICP-IAS for example). 4.3.5 Atmospheric pressure r f sputtering direct solid sampling The spectrum of the NBS 662 steel sample from the CCP spectral lamp (Fig. 4.1) was recorded at an rf power of 150 W, and an argon gas flow rate of 1.3 L/Min. The spectrum in the wavelength range 253-300 nm is provided in Figure 4.6. Other than iron lines, the manganese triplet atom lines at 279.5,279.8, and 280.1 nm, the manganese triplet ion lines 103 © Figure 4.4. Intensities of Zn 1213.9 nm in (1) HCL and (2) CCP spectral lamp (1) Perkin-Elmer Zn H C L : dc 15 mA, PMT 900 V, recorder 1 V/F.S.; (2) CCP spectral lamp with brass central electrode (Cu ~ 65 %, Zn ~ 35 %): rf power 60 W, Ar flow rate 1.3 Umin, PMT 500V, recorder 1 V/F.S.. 104 ™ . C M csj CO CM ^ 00 cn to ro m K cn o m to CD CM CM CM c c 00 CM 00 cn cn T — "~ CM CM o oo CM c c 0 \ J HCL C C P - L o m p Figure 4.5. Manganese emission spectral from HCL and CCP spectral lamp H C L : Perkin-Elmer Mn HCL, dc 12 mA, PMT 900V, Recorder 1 V/F.S. CCP spectral lamp : rf power 200 W, argon flow rate 0.7 L/min, PMT 500 V, recorder 5 V/F.S. 105 901 at 257.6 nm, 259.4, and 260.6 nm, and the chromium ion lines at 283.6, 285.0, and 285.6 nm are present in this wavelength range. Mg 1285.2 nm, Pb 1283.3 nm and Sn I 284.0 nm lines were also observed. Due to the low resolution of the 0.35 m monochromator used in this research, it is difficult to eliminate the spectral overlap from these three lines. In addition, the chromium atom lines at 357.9,359.4, and 360.5 nm, the copper atom lines at 324.8 and 327.4 nm, the silver atom limes at 328.1 and 338.3 nm; and the zinc atom line at 330.3 nm were found to be very intense in NBS 662. The concentrations of manganese, chromium, copper, silver and zinc in the NBS 662 sample are 1.05%, 0.30%, 0.51%, 0.0010% and 0.0005% respectively. This indicates that it should be possible to use this configuration of the CCP for the direct analysis of impurities in conducting samples. A modified Simplex algorithm* was used for the optimization of operating parameters in the direct determination of Manganese in steel samples. The 279.5 nm, Mn I emission intensity was selected as the criterion to be optimized. The conditions used for the analysis were 150 W rf power and 0.3 L/Min support gas flow rate. Fe 1278.8 nm was selected as an internal standard line for manganese determination in steel samples. Iron is quite a good choice for an internal standard since the concentration of iron is approximately constant in all of the steel samples. Both manganese and iron are transition metals and have similar thermal properties. The first ionization potentials of manganese and iron are 7.43 eV and 7.87 eV respectively and the wavelength separation is very small, thus no spectral response calibration of the detector was needed. The spectrum in Fig. 4.5 shows that the line pair is free from spectral interference and is located in a relatively low background intensity spectral region. * Acknowledgement is made to Dr. A. P. Wade for providing the modified Simplex program. 107 4.3.6 Calibration curves NBS 661, 662, 663, 664 and 665 series standard samples were used to set up the calibration curve for the direct determination of Mn in steels. The intensity ratio of the Mn I 279.5 nm and Fe 1278.8 nm lines was used to plot the calibration curve which is provided in Fig. 4.7. The CCP exhibits a linear working curve over a concentration range covered by the NBS 600 series. The relative standard deviation for ten determinations of Mn in soft steel ([Mn] = 0.7 %) was found to be 4.2% 4.3.7 Iron excitation temperature Temperature measurements were based on the two-line method [4.16]. The two lines chosen were the Fe I 274.70 nm (E = 43321 cnr 1 , gf = 1.4) and Fe I 275.0 nm (E = 36767 cm - 1 , gf = 0.40) lines shown in Fig. 4.6. These two lines fall in the same spectral region as the line pair for the manganese determination. It is convenient to record these two lines when carrying out the Mn determinations, since all four of the lines fall in the same dynamic range of the PMT. The temperature was measured at an rf input power of 150W and a support gas flow rate of 1.3 IVmin. From the intensities of Fe 1274.7 nm and Fe 1275.0 nm, the calculated temperature was found to be 5080 ± 500 K. 108 Figure 4.7. Calibration curve for manganese in steel 109 4 . 4 Summary The CCP torch configuration [4.1] provides for very effective energy transfer, since capacitive coupling allows a plasma to be generated at atmospheric pressure. It offers a uniform and stable plasma medium which has potential applications in many areas. In this chapter the CCP has been used as a spectral lamp and a source for direct conducting solids analysis using atmospheric pressure rf sputtering. Rf sputtering at atmospheric pressure is characterized by a relatively high density of bombarding particles resulting in a high sputtering yield. This leads to high analyte emission intensities from the CCP lamp and source. At this point, it is felt that the atmospheric pressure rf sputtering CCP lamp and source show sufficient promise to merit further exploration. These devices have potential for the analysis of conductive solids direcdy and non-conductive solids by mixing with graphite or copper powders and then pressing into pins. It is also possible to analyze micro-liter volumes of liquid solutions by deposition on the surface of graphite or metal rods. Analysis of flat conducting sheets can be accomplished by modifying the CCP torch to accept flat samples. Also, it is possible that the CCP lamp could be used as a primary source for AAS and AFS although the relatively broad spectral line profiles would probably lead to non-linear working curves when used as a source for AAS. 1 1 0 References [4.1] D. C. Liang and M . W. Blades, Anal. Chem. 60,27 (1988). [4.2] D. C. Liang and M . W. Blades, Abstracts, The Pittsburgh Conference & Exposition, paper No.415 and 1140 (1988). [4.3] D. C. Liang and M . W. Blades, Journal in press. [4.4] Improved Hollow Cathode Lamps for Atomic Spectroscopy 1985, Ed. S. Caroli, Ellis Horwood Limited. [4.5] J. A. C. Broekaert, J. Anal. At. Spectrom., 2, 537 (1987). [4.6] G. Gillson and G. Horlick, Spectrochim. Acta 41B, 431 (1986). [4.7] P. Hannaford and A. Walsh, Spectrochim. Acta ,43B, 1053 (1988). [4.8] H. J. Kim and E. H. Piepmeier, Anal. Chem., 60, 2040 (1988). [4.9] A. E Bernhard, Spectroscopy, 2(6), 24 (1987). [4.10] B. Chapman, in Transactions of the Conference and School on the Elements, Techniques and Applications of Sputtering, 1 (1969). [4.11] G. K. Wehner, Advances in Electronics and Electron Phys., 7,239 (1955). [4.12] G. S Anderson, W. N . Mayer and G. K. Wehner, / . Appl. Phys., 33,2991 (1962). [4.13] R. Stephens, (J. Anal. At. Spectrom.,)3, 1137 (1988). [4.14] C. Th. J. Alkemade, W. Snelleman, D. D. Boutiler, J. D. Winefordner, T. L. Chester and N. Omenetto, Spectrochim. Acta,33B, 383 (1978). [4.15] in Atomic Absorption Supplies Catalog, No.AA-1,11, Varian Associates (1988). [4.16] P. W. J. M . Boumans, Theory of Spectrochemical Excitation, Hilger & Watts Ltd, Plenum, New York (1966). [4.17] Balazs Magyar, CRC Rev. Anal. Chem., 17, 145 (1987). 1 11 Chapter 5 ANALYSIS OF SOLID SAMPLES USING LASER ABLATION SAMPLING - ATMOSPHERIC PRESSURE CAPACITIVELY COUPLED PLASMA EMISSION SPECTROMETRY* *Submitted for publication in /. Anal. At. Spectrom, Dong C. Liang and M.W. Blades. 112 5.1 Introduction Methods for spectrochemical analysis using laser ablation are not new. Soon after discovery of the laser in ruby crystals in 1960 by Maiman [5.7], in 1962 Breck and Cross [8] developed the first laser microprobe for local analysis, and the Jarrell - Ash company marketed the first commercial equipment Since that time, the number of publications on the spectroscopic application of laser ablation has increased rapidly. Extensive applications of laser ablation can be found in spark discharge - AES [5.8], direct current plasma (DCP) - AES [5.9], capacitively coupled microwave plasma (CMP) - AES [5.10], microwave induced plasma (MLP) - AES [5.11], and inductively coupled plasma (ICP) -AES [5.12-5.15]. Its applications can also been found in the area of atomic absorption spectrometry (AAS) [5.16-5.17], atomic fluorescence spectrometry (AFS) [5.18-5.19], ICP mass spectrometry (ICP-MS) [5.20-5.21] and resonance ionization spectrometry (RIS) [5.22]. Analytical spectroscopy using laser atomizers has been reviewed in detail by Laqua [5.23] and Piepmeier [5.24]. Laser ablation has several advantages over other techniques of vaporization: 1) direct solid sample analysis becomes possible; 2) a time consuming decomposition step can be omitted; 3) analysis can be carried out without addition of reagents and without any separation and/or concentration steps; 4) samples do not necessarily need to be electrically conductive; 5) high spatial resolution, micro and local analysis is possible; 6) vaporization and excitation processes can be controlled independendy; and 7) high sample throughput is allowed. Most laser ablation analyses were done by using a pulsed solid laser in the normal or Q-switched modes. Ishizuka and Uwamino [5.11, 5.15] found that the normal mode resulted in larger vaporized mass, which gave better absolute detection limits and was less susceptible to sample inhomogeneity. Still, because of the micro and local sampling, sample homogeneity is a significant problem for bulk sample analysis. The number of atoms vaporized from the solid target differs from matrix to matrix, even when 1 1 3 laser energy remains constant. Continuous-wave laser atomization [5.25], complete laser vaporization [5.26] and primary plasma standardization [5.27] techniques have been developed to improve precision and accuracy in recent years. An atmospheric pressure radio-frequency (rf) capacitively coupled plasma (CCP) has been reported as being useful as an atomizer for atomic absorption spectrometry (AAS) [5.1], a source for atomic emission spectrometry (AES) [5.2], a detector for gas chromatagraphy (GC) [5.3], an atmospheric pressure plasma discharge formed inside a graphite furnace [5.4-5.5], and a CCP-rf sputtering device for direct solid analysis [5.6]. However, the CCP can also be combined with other sample introduction techniques including laser ablation. In this chapter, a laser ablation - CCP emission spectrometry system is described. Experimental parameters and analytical performance are examined, and the possibility of using this system for direct solid sample analysis is investigated. 1 1 4 5.2 Experimental 5.2.1 Laser ablation cell A schematic diagram of the laser ablation - CCP system is shown in Fig. 5.1. The cell was constructed of a brass cylindrical body with a quartz window and a sample adapter designed for NIST (formerly NBS) SRM 600 series pin (diameter of 3.2 mm) steel samples. Flat samples can be direcdy placed under the cell without using the adapter. An O-ring provided a gas seal between the cell body and samples. The inner diameter of the brass body was 1 cm and the length was 2 cm. The small cell volume reduced peak broadening of the signals. A quartz lens of 10 cm focal length was used to focus the laser beam onto the surface of the samples. No damage was seen in the quartz window after several hundred laser shots. 5.2.2 Experimental facilities A TRG 104A ruby laser (TRG Control Data Corporation) was used. The laser power supply may be operated in either normal or Q-switched mode. The laser input power can be varied from 0 - 1000 joules. No data on efficiency of lasing has been found. Laser firing was controlled either manually or by a remote control. In this study, only the normal mode with manual firing was used. The usable pulse repetition rate of the ruby laser was 1 minute. A modified tee torch [5.2] was used to generate a capacitively coupled plasma at atmospheric pressure. The torch, designed originally for tantalum strip vaporization sampling, was used in this study without further modification. The torch was constructed of a square quartz tube for the plasma, sandwiched between two rectangular quartz tubes which were used to enclose the two strip electrodes. A 5 cm long quartz capillary tube with an inner diameter of 0.08 cm was used to connect the tantalum strip housing to the 115 Plasma Atomizer 7^ rf Generator Ruby Laser n 0 M Sample Adapter y Laser Ablation Cell Figure 5.1. Schematic diagram of the laser ablation - CCP system 116 plasma. The outlet of the laser ablation cell was connected to the CCP torch plasma gas inlet with R-3603 Tygon tubing. Argon was chosen as plasma gas and carrier gas. Additional equipment and experimental details employed in this research were similar to those used previously [5.2] and are listed in Table 4.1. 5.2.3 Analytical Procedures Metal samples were fabricated either as a solid rod with a diameter of 3.2 mm and length of about 51 mm, or as a flat plate. Both types of sample were polished by fine sandpaper, and washed with deionized water. The sample rods were placed into the ablation cell adapter, and the bulk flat samples were placed under the ablation cell. The laser beam was focused near the edge of the sample rods, so that the rods could be rotated to expose a fresh surface to the laser beam. A low power helium neon laser was used for alignment and pre-focusing. Fine focusing was obtained by observing analytical signals to give a maximum intensity. The CCP was stable and adjusted for optimal conditions before the laser firing. A plume about 5 mm in height was produced from the sample surface by the laser beam. Both emission and absorption signals were simultaneously recorded after each firing. A set of NIST SRM 600 series low-alloy steel samples were selected for establishing calibration curves for direct solid sample analysis. Five standards in the series (661 through 665) cover a wide range of concentrations for many elements allowing calibration for steel sample analysis. 117 5.3 Results a n d discussion 5.3.1 Effect of r f input power The effect of changes in rf input power on the emission intensity of the Cr 1357.9 nm line at an argon flow rate of 2.2 L/Min and rf input power range of 100-350 W are provided in Fig. 5.2. The Cr 1357.9 nm emission intensities increased with an increase in rf power, and reached a maximum at 300 W. This power was used in the following experiments. 5.3.2 Effect of plasma gas flow rate For the laser ablation - CCP system used in this study, the laser acts as a sampling and vaporization device, and the CCP acts as an excitation device. Independent control over vaporization and excitation provides a mechanism for simultaneous optimization of both processes in this source. Devices of these types can be classified into two groups: in the first vaporization and excitation take place in the same observation volume, and in the second, vaporization and excitation take place in different volumes. For the latter group, vaporized analyte species are usually transported into the excitation volume using a carrier gas. The laser ablation - CCP system used in this study belongs to this group. The effect of carrier gas on the Cr I 357.9 nm emission intensity were studied in the laser ablation -CCP system. The results are provided in Fig. 5.3. High gas flow rate increases the analyte transport efficiency, but shortens the residence time. These are two competing processes in determining the sensitivity. An optimal flow rate of 2.2 L/min was found for Cr I 357.9 nm emission intensity at a rf input power of 300 W. The carrier gas not only acts as a plasma support gas, but also carries the analyte species from the vaporizer into the plasma in which the excitation takes place. The residence time of analyte atoms is mainly determined by the carrier gas flow rate and diffusion effects 118 80 u >> "35 B c 60 H 40 H o '55 20 H w E w J u J I L-100 200 300 rf Power (W) 400 Figure 5.2. Effect of rf input power on Cr 1357.9 nm emission intensity 119 30 Flow Rate (L/min) Figure 5.3. Effect of argon gas flow rate on Cr 1357.9 nm emission intensity 120 5.3.3 Laser Energy and focusing The amount of sample ablated by laser shots is affected by two parameters, laser focusing and pulse energy. Preliminary results show that the laser focusing changed the laser spot size, which changed the laser beam energy density, and consequentiy changed the analytical signal intensities. Before proceeding with other experiments.an optimization of laser focusing was carried out to give a maximum signal intensities, as mentioned in the experimental section. The mass loss resulting from a single shot was about 5 u.g for steel samples at a laser input power of 800 J per shot Each shot produced a pin hole on a 0.2 mm thick steel target when an optimal laser focusing position at the maximum laser power was used. The amount of the sampled ablated by the laser may be controlled by changing the laser energy. The effect of laser input energy on the emission intensities of Cr 1357.9 nm was examined at a rf input power of 300 W and an argon gas flow rate of 2.2 L/min. The results indicated that, as expected,the emission intensities increased with the laser input energy. The efficiency of the ruby laser used in this study is unknown, but an increase in lasing power with the laser input power can be expected. Changing the laser power allows quantitative analysis to be caned out over a wide range of analyte concentration. 5.3.4 Calibration curve A calibration curve for the Cr I 357.9 nm line, shown in Fig. 5.4, was obtained by using the NIST 600 series SRM's. Analytical conditions were set at 800 J laser input power, 300 W rf input power, and 2.2 L/min argon gas flow rate. The coefficient of regression of the slope of the analytical calibration curve was 0.995. The calibration curve was linear at least over a range of 0.007 - 1.4 % chromium. The linearity outside this range has not been examined. 121 y = 1.9341 * xA8.2038e-2 R A2 = 0.998 | i i i i i 111| i i i i i I I 11 i i i i i II11 i i i i i 1111 001 .01 .1 1 10 Concentration (%) Figure 5.4. Calibration curve for chromium using laser ablation CCP 122 5.3.5 Detection limits The detection limit was defined as the concentration or weight of element which gives a signal equal to 3 times the standard deviation of background noise. The CCP background emission measured at 357.9 nm for the calculation of standard deviation. The detection limit for chromium in steel samples was found to be 0.3 ppm. The average ablated mass resulting from a single laser shot was 5 ug. Therefore, the absolute detection limit was calculated to be 1.5 pg. These results were lower than those obtained in a laser vaporization - microwave induced plasma (MD?) system and in a laser ablation -inductively coupled plasma (ICP) system reported by Ishizuka and Uwamino [5.11,5.15]. In both systems, a ruby laser was used. They reported that the relative detection limit and absolute detection limit were 13 ppm and 10 pg respectively for the Cr 1357.9 nm line in the MIP system using the Q - switched mode, and 1 ppm and 30 pg respectively for the Cr D 267.7 nm line in the ICP system using the normal mode. 5.3.6 Accuracy To check the accuracy of the laser ablation - CCP spectrometry, five SRMs from the NIST were analyzed. The results are shown in the Table 5.1. Good agreement with the certified values was obtained. The errors fell within a + 1 . 6 % t o - l l % range. The concentration of chromium was found to be 0.0061 % in a soft steel sample. 123 Table 5.1. Q U A L I T A T I V E ANALYSIS OF CHROMIUM IN STEEL SAMPLES BY LASER ABLATION-CCP (ng/g) Sample Certified value (%) Found Ejrpr(%) SRM 661 0.69 0.634 ±0.04 -8.1 SRM 662 0.30 0.305 ±0.02 +1.6 SRM 663 1.31 1.37 ±0.07 +4.7 RSM664 0.06 0.0619 ± 0.004 +3.2 SRM 665 0.007 0.0069 ± 0.004 -11 solft steel - 0.0061 ± 0.004 1 2 4 5.4 Summary This chapter has described the first application of laser ablation sampling combined with a CCP source for direct analysis of solid samples. This compact plasma source can be used to replace the burner head in an atomic absorption spectrophotometer or can be coupled with a direct reading polychromator. The analytical calibration curve was found to be linear over a range of at least 0.007 -1 . 4 % for chromium in steel samples. Five standard reference materials (SRM) from the National Institute for Standards and Technology (NIST) were analyzed, and good agreement with the certified values was obtained. The relative and absolute detection limit for chromium are 0.3 ppm and 1.5 pg respectively, and are superior than those for similar laser ablation - MIP or ICP systems. The results obtained from this study suggest the possibility for a technique capable of the rapid direct analysis of minor or trace components in solid samples using a laser ablation - CCP system. 1 2 5 References [5.1] Dong C. Liang and M . W. Blades, Anal. Chem., 60, 27 (1988). [5.2] Dong C. Liang and M . W. Blades, Anal. Chem., in press. [5.3] Degui Huang, Dong C. Liang and M . W. Blades , / . Anal. At. Spectrom., 4,789 (1989) [5.4] Dong C. Liang and M . W. Blades, Spectrochim. Acta ,44B, 1059 (1989). [5.5] D. Lee Smith, Dong C. Liang, Doug Steel and M . W. Blades, Spectrochim. Acta, 45B, 493 (1990). [5.6] Dong C. Liang and M . W. Blades, Spectrochim. Acta ,44B, 1049 (1989). [5.7] T. H. Maiman, Nature, 187, 439 (1960). [5.8] F. Brech and L. Cross, Appl. Spectrosc, 16, 59 (1962). [5.9] P. G. Mitchell, J. Sneddon and L. J. Radziemski, Appl. Spectrosc.,41, 141 (1987). [5.10] L. G. Bachurina, V. M . Perminova and S. A. Savostin, Zavodsk. Lab.,45,1113 (1979). [5.11] T. Ishizuka and Y. Uwamino, Anal. Chem., 52, 125 (1980). [5.12] M . Thompson, J. E. Goulter and F. Sieper, Analyst ,106, 32 (1981). [5.13] J. W. Carr and G. Horlick, Spectrochim. Acta,37B, 1 (1982). [5.14] M . E Tremblay, B. W. Smith, M . B. Leong and J. D. Winefordner, Spectrosc. Utter ,20,311 (1987). [5.15] T. Ishizuka and Y. Uwamino, Spectrochim. Acta,3SB, 519 (1983). [5.16] R. M . Manabe and E. H. Piepmeier, Anal. Chem. 51, 2066 (1979). [5.17] K. Sumino, R. Yamamoto, F. Hatayama, S. Kitamura and H. Itoh, Anal. Chem., 52, 1064 (1980). [5.18] R. M . Measures and H. S. Kwong, Appl. Optics.lS, 281 (1979). 126 [5.19] H. S. Kwong and R. M . Measures, Anal. Chem., 51,428 (1979). [5.20] A. L. Gray, Analyst, 110, 551 (1985). [5.21] P. Arrowsmith, Anal. Chem., 59, 1437 (1987). [5.22] S. Mayo, T. B. Lucatorto and G. G. Luther, Anal. Chem., 54, 553 (1982). [5.23] K. Laqua, in Analytical Laser Spectroscopy, S. Martellucci and A. N . Chester, Plenum Press, New York, 1982, p. 159. [5.24] E. H. Peipmeier Ed., Analytical Application of Laser Spectroscopy, Wiley, New York, 1986, p.627. [5.25] J. Gagne, P. Pianarosa and F. Lafleur, / . Anal. At. Spectrom., 3, 683 (1988). [5.26] G. Su and S. Lin, / . Anal. At. Spectrom., 3, 841 (1988). [5.27] K. Kagawa, Y. Matsuda, S. Yokoi and S. Nakajima, / . Anal. At. Spectrom., 3, 415 (1988). 127 Chapter 6 AN ATMOSPHERIC PRESSURE CAPACITIVELY COUPLED PLASMA FORMED INSIDE A GRAPHITE FURNACE AS A SOURCE FOR ATOMIC EMISSION SPECTROSCOPY* * Part of this chapter was published in Spectrochim. Acta 44B, 1059 (1989), Dong C. Liang and M. W. Blades, and Spectrochim. Acta 45B, 493 (1990), D. Lee Smith, Dong C. Liang, Doug Steel and M. W. Blades. Acknowledgement is made to the donors of the National Research Council of Canada for suporting DLS and DS through an IRAP-H grant. Special thanks to DLS for preparing the calibration curve and measuring the detection limit, and to DS for assisting in modifying the rf power supply. 128 6.1. Introduction For many years, graphite furnace atomic absorption spectrometry (GFAAS) has been recognized as one of the most sensitive analytical techniques for elemental analysis [6.1]. On the basis of absolute mass, GFAAS detection limits are very low because sample volumes are small (5 to 100 uL), the analyte transport efficiency is high (90-100 %), and analyte residence time in the observation volume is relatively long (0.1 -0.5 s.). As a consequence, detection limits for GFAAS are in the sub-ng/mL range for many elements. A limitation of GFAAS arises as a result of interferences which have been classified as interferences due to background absorption, condensed phase interferences, vapour phase interferences, and effects due to gas expansion [6.1-6.3]. The combination of a thermal pre-treatment step; temporal and spatial isothermal atomization through the use of stabilized temperature platform furnaces (STPF), rapid heating cycles, probe insertion, and two-step furnaces; and background correction techniques such as Zeeman, Smith-Hieftje and the use of continuum lamps have enabled sensitive determinations in a variety of complex samples. However, interferences continue to limit the effectiveness of GFAAS and, although the use of hollow cathode lamps as primary sources provides high spectral selectivity, they introduce the limitation that restricts GFAAS to being essentially a single element technique. In the past, several approaches have been used to enhance the graphite furnace as a multielement source for atomic emission spectrometry (AES). Littlejohn and Ottaway [6.4] have described carbon furnace atomic emission spectrometry (CFAES) which is a sensitive technique for trace analysis using thermal excitation from normal furnace heating. This method is limited by the maximum temperature of the graphite furnace and is not very suitable for elements with high excitation energies. Falk and his co-workers [6.5 - 6.7] developed a low pressure glow discharge inside the graphite furnace; this technique has been termed FANES (Furnace Atomization Non-thermal Excitation Source). Detection 129 limits for FANES are generally similar or superior to those of GFAAS. The technique is attractive due to its large linear dynamic range, narrow atomic linewidth, multielement capability, and because there is the possibility for independent optimization of atomization and excitation. Recently, Ballou et al. [6.8] have described a similar device in which the graphite furnace serves as an anode of a glow discharge where the cathode is a graphite pin which runs down the centre of the furnace. This hollow-anode plasma excitation source is more flexible in terms of the electrical isolation requirements. This device has been used as an atomic emission source for the analysis of metals and non-metals. Both of the latter sources are essentially low pressure, direct current (dc) glow discharges. In a glow discharge the gas temperature is low (not in local thermal equilibrium), the residence time of analyte atoms is relatively short, analyte density in the gas phase is low, and perhaps most important from an analytical standpoint, it is not convenient to change samples at low pressure. An atmospheric pressure radio-frequency (rf) capacitively coupled plasma (CCP) has been reported in previous chapters of this thesis as being potentially useful for both atomic absorption spectrometry (AAS), atomic emission spectrometry (AES) and gas chromatography (GC) [6.9-6.12]. It should be noted that the geometric arrangement of the electrodes in the hollow-anode source was very similar to the geometric arrangement of electrodes in an atmospheric pressure CCP lamp and source described by Liang and Blades in Chapter 5 [6.11]. Originally sample introduction into this plasma was accomplished by using an electrically heated tantalum strip vaporizer [6.9]. The analyte atoms vaporized from the tantalum strip were carried by the plasma gas into the plasma through a quartz capillary. In this case, the transport efficiency is determined by the flow rate of the plasma gas. The greater the gas flow rate, the higher the transport efficiency, but the shorter the residence time of analyte atoms in the plasma. In order to increase the transport efficiency and residence time, an atmospheric pressure furnace capacitively coupled plasma (APF-CCP) has been developed. This device 130 combines the advantages of the graphite furnace with those of the CCP. The electrode arrangement in the APF-CCP is similar to the CCP lamp [6.11] and to that described by Ballou et al. [6.8], however, the plasma is formed between the graphite tube and a central electrode by rf capacitive coupling at atmospheric pressure. This is in contrast to the hollow-anode source which is a low pressure, dc glow discharge. With the APF-CCP, conventional thermal graphite tube atomization is still possible but atmospheric pressure rf sputtering can also act as an atomization mechanism [6.9, 6.11]. The plasma, which was operated in argon or helium at a frequency of 27 MHz and an rf power of 20 - 100 W, uniformly filled the graphite tube and could be utilized to excite an atomic vapour produced from a normal graphite furnace atomization heating cycle. By analogy with FANES this source has been given the anachronym FAPES (Furnace Atomization Plasma Excitation Spectrometry). The potential of this device for carrying out simultaneous multielement analysis in a manner similar to the use of the FANES source is obvious. This device provides a new dimension to the use of graphite furnaces for analytical atomic spectroscopy. This chapter describes the preliminary characteristics of this device. 131 RF Filter RF Power Supply Furnace Power Supply Fiugure 6.1. Schematic diagram of the APF-CCP source as a free-running oscillator such that the operating frequency, when connected between the graphite furnace and the central electrode, was dictated by the load impedance. By adjusting the tuning elements in the Model DX-60 the transmitter was made to operate resonandy between 27.02 and 27.32 MHz. The power delivered to the FAPES source was between 20 and 30 W. 6.2.3 Equipment and setup The equipment and experimental set up employed in this research are outlined in Table 6.1. For the analysis of aqueous samples, a Schoeffel-McPherson (Acton, MA) Model 270, 0.35 m monochromator equipped with a holographic grating with 2400 lines/mm was used to isolate the spectral lines of interest The entrance and exit slit widths were set at 100 um. An image of the plasma was formed on the entrance slit using a plano-convex fused silica lens with a focal length of 750 mm. The entrance slit height was set to 2 mm in order to exclude black body emission from the walls of the graphite tube. 134 Table 6.1 Experimental Facilities and Operating Conditions Plasma Power Supply 1) Power Amplifier: Ehrhorn (Canon, CO), Model Alpha 86 hf Amateur Linear Power Amplifier. Oscillator modified Heathkit (Benton Harbor, MI), Model DX-60 Phone and CW Transmitter. Working frequency - 27 MHz Impedance matching: Wm. M . Nye (Bellevue, WA), Model MB-V-A Antenna Tuner. 2) Modified Heathkit (Benton Harbour, MI) Model DX-60 Phone and CW Ham Radio Transmitter. This transmitter was modified to operate as a free-running oscillator (for the analysis of aqueous samples). Graphite Furnace Modified Instrumentation Laboratory (Wilmington, MA), Model 455 flameless Atomizer. Spectrometer Varian (Springvale, Australia), Model AA-875 Atomic Absorption Spectrophotometer operating in emission mode, integrate repeat 0.1 sec., fast recorder mode. Bandwidth 0.05 nm for spectral scans and 0.2 nm for intensity measurements. Data Acquisition Servocorder 210 chart recorder, 1 volt/ full scale, 3 cm/min. 135 6.2.4 A n a l y t i c a l procedure For this paper, spectra from the APF-CCP were obtained by placing a small solid piece of brass (about 5 mg) into the furnace through the furnace sample introduction port. The plasma was ignited. The graphite tube was the heated to a suitable temperature to provide atomic vapor from the solid sample (approx. 800 °C), and spectra were recorded. For quantitative analysis, liquid samples of 2-10 pL volumes of Ag(I) were introduced into the graphite furnace using an Eppendorf 0.5 -10 uL UltraMicro pipette. The stock solutions of Ag(I) were prepared fresh prior to the analytical studies by dissolving AgNC*3 salt in dilute HNO3. Care was taken to inject the sample on the inner wall of the graphite furnace. Analytical blanks were recorded using dilute HNO3. The atomizer control module was used to cycle the graphite furnace through the dry (45 s at 130 °C), ash (5 s at 500 °C), and atomize stages (5 s ramped to 2000 °C) under automatic control. The plasma was ignited during the middle of the ash cycle by application of rf power. The plasma was self igniting immediately upon application of this power. The plasma was automatically extinguished at the end of the automatic post atomize cooling cycle when the pneumatic sample introduction port opened adrnitting the room atmosphere which quenched the plasma. At this stage the RF power was turned off. 136 6.3 Results and discussion 6.3.1 Description of argon and helium plasmas An rf generator with an auto-matching network was used , for both argon or helium plasma gases. The plasma forms inside the furnace as soon as rf power is applied; a Tesla coil is not required for ignition. If one is using a fixed frequency rf power supply, impedance matching is accomplished by using an antenna tuner for matching the output impedance of the power supply to the impedance of the source. Them, the plasma can be ignited from thermionic emission during the vaporization step when the matching network is initially tuned for the plasma running position. Using this arrangement it was found that during the furnace atomization step the reflected power would increase, this causing the forward power to the source to vary. The free-running variable frequency method overcomes this difficulty since, as the APF-CCP source impedance changes, the frequency varies in a manner such that good impedance is maintained When the plasma is ignited, the tungsten rod is dark or dark-red at low rf powers; atmospheric pressure rf sputtering is the dominant sampling mechanism. With an increase in rf power, and /or heating up the furnace tube by application of dc current, the central rod changes through orange to white-hot Under this condition sampling takes place by both rf sputtering and by conventional thermal vaporization. The helium plasma is easier to be ignited than the argon plasma. While the argon plasma is light blue and uniform by eye, the helium plasma is pine in colour, with a bright layer adjoined on the surface of the central electrode. 6.3.2 Background emission The background spectra of argon and helium plasmas from the APF-CCP source are similar to those from the CCP torch shown in Fig. 3.4 and Fig. 3.5. Prominent emission bands are observed for OH, NO and CN, and arise from impurities in the plasma gases and 137 the ingresses of ambient atmosphere in the opened sources. This problem can be overcome by using a completely enclosed source, and using highly pure plasma gases. 6.3.3 Role of the CCP inside the graphite furnace Typical emission spectra are shown in Fig. 6.2. Fig. 6.2(a) was recorded at a lower gain setting relative to the the gain in Fig. 6.2(b). The spectra were obtained by placing a small brass chip (about 5 mg) on the inside of the graphite tube. The rf power was set to 20 W and the argon flow was 0.94 L/Min. The spectra cover a range from 322 nm to 338 nm. The concentrations of zinc and copper in the brass are approximately 30-35 % and 65-70 % respectively. As can be seen from Fig. 6.2(a), the intensities of Zn I 334.50, 330.26 and 328.23 nm lines with excitation energies of Ca. 7.78 eV are effeciently excited as are the Cu 1324.75 and 327.40 nm lines (excitation potentials are 3.82 eV and 3.78 eV respectively). The higher emission intensity for Zn relative to Cu can be explained on the basis of the boiling points of these metals which are 907 and 2595 °C respectively. To evaluate the effect of the CCP inside the graphite furnace on the emission intensity, the intensity of the Zn I 334.50 nm line was measured from the APF-CCP, i.e. eith the plasma on, running at the dark red furnace temperature (approx. 800 °C). The signal is shown in Fig. 6.3.(a) and was very strong. No signal was found at the same furnace temperature if the plasma was off (Fig. 6.3.(b)). When the furnace temperature was increased to its maximum (2700 - 3000 K) a small pure furnace emission (CFAES) signal was observed (Fig. 6.3 (c). The results shown in the Fig. 6.3 show that the plasma formed inside the furnace acts to excite atomic species in the gas phase. 6.3.4 Effect of plasma support-gas flow rate Although the plasma gas does not flow directly through the furnace itself, an increase in the plasma gas flow rate will decrease the residence time of analyte atoms in the furnace, 1 3 8 (a) ( b ) CO 3 <N 00 c N VO ON r s <N d d ro ro ro ro c c N N ro ro ro ro c c N N 322.0 336.0 324.0 VO ON CN <N d d ro ro ro ro Go »n ro ro ro ro (5 N ro <N oo <N ro c N Jl 338.0 Wavelength (nm) Wavelength (nm) Figure 6.2. Spectra of copper and zinc between 322 and 338 nm from the APF-CCP source (a) (b) (O RF on E o < RFoff V K E o _A_L t t v N E o < 1 min. Figure 6.3. Comparison of Zn 1334.50 nm from (a) APF-CCP source at a dark red furnace temperature (approx. 800 °C), (b) CFAAS at the same furnace temperature as in (a); (c) in the same as (b) except running at maximum furnace temperature (approx. 2800 °C) 140 and consequently reduce the emission intensity. This result has been observed by measuring the effect of argon flow rate on the intensity of the Cu I 324.75 nm line. The results are provided in Fig. 6.4. 6.3.5 Detection limits The detection limit for Ag was deterrained from 10 replicate area measurements of the background emission intensity and calculating the standard deviation in the background from these measurements. From this the 3a detection limit was determined to be less than 0.3 pg. The Ag detection limit reported for the FANES technique was 0.4 pg [6.7]. 6.3.6 Calibration curves In order to test the quantitative capabilities of the FAPES source a calibration plot for Ag was obtained over the concentration range 0 to 1 ng/mL. Each point on the plot was determined from 5 replicate injections of the relevant Ag(I) solution. The plots for both peak height and peak area were linear over this range. 6.3.7 Precision In order to test the precision of this source, replicate twelve signal acquisitions of the Cu I 324.75 nm emission intensity were acquired. The signals were obtained using a 1 s integration time and plotting the signals using a bar chart recording mode. The intensities were set at the middle of the dynamic range (full scale = 100). The average of thirteen signals was 54.4 with a relative standard deviation of 1.9 %. 6.3.8 Iron excitation temperature Iron atom excitation temperatures were measured using a method previously described [3.10]. These measurements were carried out using both cold (dark red) and hot (about 2700 K) furnace conditions with argon and helium plasma gases. For cold furnace 1 4 1 10 Row Rate(L/m) 20 Figure 6.4. Emission intensity of Cu 1324.75 nm as a function of the plasma support gas flow rate. 142 conditions, a section of iron wire was introduced into the graphite furnace to provide a source of iron atoms. In the hot furnace case, 5 uL of 10 ppm Fe solution was injected on the wall of the furnace using an adjustable microlitre pipette, followed by normal furnace operating cycle. The collection optics were set up to image the gap of the graphite tube and the graphite rod onto the entrance slit of the monochromator. Emission from a set of seven Fe (I) lines in the region 370-385 nm covering an energy range from 27000 to 35000 cm*1 were used for this measurement. The lines used were the same as those which were outlined in reference 3.10. A Schoeffel-McPherson (Acton, MA) Model 2061 1-meter monochromator equipped with a linear photodiode array was used to carry out the measurement. The complete system has been described elsewhere [3.11]. The temperature was measured at an rf input power of 30W. Linear regression slope temperatures were found to be 2980 ± 200 K and 3740 ±500 K for the cold and hot furnace conditions respectively with helium plasma. The argon plasma, however, gave 4420 ± 500 K and 4670 ± 460 K for the cold and the hot furnace conditions respectively. Since, the iron excitation temperature is higher in the argon plasma than in the helium plasma, it is expected that the sensitivities for most metals should be higher in argon plasma than those in helium plasma. 6.3.9 Multielement Capability A 5 uL multielement standard solution of 0.8 ppm Ag, 5 ppm Cd, 5 ppm A l , 10 ppm Zn, 100 ppm Cu, 100 ppm Ca and 200 ppm Cr was injected into the FAPES source. The spectrum obtained for this solution was recorded using the linear photodiode array spectrometer using experimental conditions outlined in section 6.3.8., and shown in Fig. 6.5. Strong spectral lines from all these elements were found in a single window of 50 nm.width. Although the sensitivity of the linear photodiode array is much lower than the PMT, this experiment shows that similtaneous multielement analysis is also feasible using the FAPES source. 1 4 3 o 00 I n t e n s i t y (x 1.00) (Jl CD C r i 301.757 All 308.215 o -Ci. CD CJ1-CD CD Al I 309.271 Al 1 309.284 a 11 315 887 an3H.993 Cu I 324.754 Cu 1 327.3% Ag I 328.068 Zn 1 328.233 Zn 1 33026 1 Cd I 326.105 6.4 S u m m a r y In this chapter an atmospheric pressure plasma sustained inside a graphite furnace has been described This source combines the high efficiency of atomization in furnaces and the high efficiency of the excitation in atmospheric pressure plasmas. Atmospheric pressure operation is not only convenient for changing samples but also provides for the possibility of high-yield rf sputtering. Atmospheric pressure plasmas provide a relatively high thermal gas temperature which should allow more complete dissociation of molecular species. This should reduce the occurrence of gas phase chemical interferences inside the furnace. This source also offers the ability to independently optimize vaporization and excitation. However, the most important aspect of this new source is that it can be used for simultaneous, multielement determinations of small sample sizes in an atomizer which has been proven to be effective over many years of use. 145 References [6.1] W. Slavin, Trends in Analytical Chemistry,.^, 194 (1987). [6.2] J.P. Matousek, Prog. Analyt. Atom. Spectrosc. ,4, 247 (1981). [6.3] W. Freeh, E . Lundberg and A. Cedergren, Can. J. Spectrosc, 30,123 (1985). [6.4] D. Litdejohn and J. M . Ottaway, Analyst ,104, 208 (1979). [6.5] H. Falk, E . Hoffmann, I. Jaeckel and Ch. Ludke, Spectrochim. Acta ,34B, 333 (1979). [6.6] H. Falk, E . Hoffmann, and Ch. Ludke, Spectrochim. Acta ,36B, 767 (1981). [6.7] H. Falk, E . Hoffmann, and Ch. Ludke, Spectrochim. Acta ,39B, 283 (1984). [6.8] N . E . Ballou, J. M . Handy, and D.L. Styris J . Anal. At. Spectrom., 3, 1141(1988) [6.9] Dong C. Liang and M . W. Blades, Anal. Chem.,60, 27 (1988). [6.10] Dong C. Liang and M . W. Blades, Anal. Chem.,in press. [6.11] Dong C. Liang and M . W. Blades, Spectrochim. Acta ,44B, 1049 (1989). [6.12] Degui Huang, Dong C. Liang and M . W. Blades, J. Anal. At. Spectrom., 4,789 (1989) 146 Chapter 7 A CAPACITIVELY COUPLED P L A S M A DETECTOR FOR GAS CHROMATOGRAPHY* *This chapter was published in / . Anal. At. Spectrom. 4,789 (1989), Degui Huang, Dong C. Liang and M.W. Blades. DH was a visiting scholar from the People's Republic of China when he was cooperating in this project and carrying out most of the experiments described in this chapter. 1 4 7 7.1. Introduction The use of atomic emission spectroscopic detectors for gas chromatography (GC) is an important methodology due to its simplication in the interpretation of complex chromatograms by providing element specific information about each eluting peak. In this context plasma sources are playing a very important role. Since the first paper [7.1] based on plasma emission spectroscopy was published, several kinds of plasmas, including microwave induced plasmas (MD?), direct current plasmas (DCP) [7.2-7.4], alternating current plasmas (ACP) [7.5,7.6], and inductively coupled plasmas (ICP) [7.7] have been utilized as GC detectors. The MD? has been the most successful because a relatively low power is required to sustain the plasma, a relatively small quantity of gas is consumed during routine operation and the detector volume can be kept small. In the initial work using a MIP as a GC detector McCofmach et al. [7.1] used a 2450 MHz, atmospheric pressure argon discharge to detect the elution of halogen and other non-metal containing compounds. The detection limits for most of the non-metal elements were found to be in the range of IO - 1 2 to IO - 9 g/s. Bache and Lisk [7.8-7.11] used a helium MIP at low pressure (5-10 mm Hg) and found that better power coupling and hence better atomizaton characteristics could be obtained. This situation was improved with the development of the Beenakker [7.12] TMoiO cylindrical resonance cavity which allowed efficient power coupling to a MB? at atmospheric pressure and at a relatively low power level (40-lOOw). Although the MB? is an excellent excitation source for element-specific GC detection it is constrained by some operational limitations. MD? cavities must be resonant with the driving frequency. According to Beenakker and Boumans [7.13], coupling of power using a fixed loop as originally described by Beenakker [7.12] is adequate for a helium plasma but not for an argon plasma. Additionally, discharge conditions in an MD? depend on both the inner diameter of the cavity and the dielectrics inside the cavity. MIP's are normally ignited using a tesla discharge. 148 This chapter describes an atmospheric pressure capacitively coupled plasma (CCP) torch which can be used as an element-specific, spectroscopic detector for GC. This plasma source is similar in design to a CCP source which has been used for atomic spectroscopy [7.14, 7.15]. One of the main differences between the configuration described in this paper and that previously described is that the plasma torch was made smaller to reduce peak broadening and to enhance the sensitivity for GC detection. Using the torch described in this paper, a stable self-igniting plasma can be sustained over a wide range of operating frequencies, input powers, and at very low carrier gas flow rates. The major advantages of this torch over a MTP is that the discharge can be sustained over a very wide range of conditions; power from 10 to 500 W, frequency from 200 KHz to 30 MHz, and carrier gas flow rates as low as 20 rnL/min. In addition, the torch configuration is simple, the plasma is stable, and a separate ignition system is not required. 149 7.2 Experimental 7.2.1 The CCP GC detector design A schematic diagram of the experimental system is provided in Fig. 7.1. Details of the experimental facilities are outlined in Table 7.1. The outlet of the column was connected to the gas inlet of the CCP through a i m length of 1.5 mm i.d. Copper tubing which was maintained at a slightly higher temperature than the GC column using heating tape (Electrothermal Engineering Ltd.), which surrounded the interface tubing. Details of this interface are provided in Fig. 7.2. Two different plasma discharge geometries have been used. Schematic diagrams for each of these are provided in Fig. 7.3. Fig. 7.3a shows a concentric design which is made from a single piece of quartz. The electrodes which couple the rf power into the discharge are arranged concentrically in that the inner electrode is housed inside a hollow quartz shaft which is sealed such that there is no electrode contact with the plasma. The outer electrode is simply a stainless steel cylinder which is wrapped around the outside of the quartz tube in the manner shown in Fig. 7.3a. When the rf power is applied, the plasma fills the annular space between the electrodes as well as the small space at the end of the torch. The second geometry (Fig. 7.3b) consisted of three rectangular quartz tubes which had dimensions 2 mm by 4 mm. The central tube containing the plasma was sandwiched between two quartz tubes, each of which enclosed a stainless steel electrode of slightly smaller dimensions than the interior demensions of the quartz tubes. Carrier gas from the GC column was directed into the plasma through one end of the torch. A B5 joint tube served as a support but did not act to introduce any extra flow of gas. 1 5 0 POWER SUPPLY COMPUTOR PMT ... l i V R (P) Plasma torch (L) Focusing lens (M) Monochromator (V) PMT h.v. power supply (A) Amplifier (R) Chart recorder Figure 7.1. Schematic diagram of the experimental system used to test the CCP as a GC detector 151 HEATING TAPE I COPPER FITTING PLflSMR TORCH I TEFLON TAPE COPPER TUBE Figure 7.2. Schematic diagram of the interface of GC column with the plasma torch 152 F R O M C O L U M N 90mm 80mm E N D V I E W 2, (a) 0o=3mm Di = 1.5mm D0=7mm Di=5mm SUPPORT ARM 80mm E N D Y I E W FROM COLUMN E L E C T R O D E S SUPPORT ARM 2X4mm (b) Figure 7.3. Schematic diagram of the CCP torches, (a) concentric geometry and (b) parallel 153 7.2.2 Equipment and setup The equipment and experimental setup employed in this research are summarized in Table 7.1. Table 7.1. Experimental Facilities and Operating Conditions Gas Chromatogaph Plasma Power supply Spectrometer: Detector electronics Data Acquisition Varian 6000 Gas Chromatograph; Columns: OV-101 and OV-07; Carrier gas flow rate: 30-80ml/Min. (a) Perkin-Elmer ICP 5500 system consisting of a Plasma-Therm (Kreeson, N.J.), HFP-2500F RF generator, AMN-2500E automatic matching network, APCS-3 automatic power control system and PF2500 torch box (b) ENI power systems Inc. (Rochester,N,Y,), Model HPG-2 RF power supply; frequency 125-375 KHz, output power 0-200W. Schoffel-McPherson (Acton.MA) Model 270, 0.35 Czerny-Turner mount scanning monochromator with 1200 lines/mm holographic grating, reciprocal linear dispersion of 2 nm/mm in the first order, entrance and exit slits set to 50 mm. The photocurrent from a Hammatsu R955 photomultiplier tube was amplified by an amplifier made by the department electronics shop, the photomultiplier tube was powered by a McPherson Model EU-42A PMT power supply. Digital data acquisition: Zenith computer, Model W-248-82, (512K, Zenith Electronics Corp.), IBM-AT compatible 154 computer equipped with a RC Electronics (Santa Barbara, CA) Model ISC-16 analog-digital converter running the RC computerscope software package. Analog data recording: Fisher Recordall 5000 chart recorder 155 7.2.3 Analytical procedure The spectra of organic compounds were obtained by using a 200 mL plastic botde as a sample diluting container, in which the carrier gas was mixtured with the injected volatile organic compounds and introduced to the plasma. Using this apparatus a continuous flow of volatile compound could be introduced into the plasma so that emission spectra could be obtained. After selection of an appropriate analysis line, the wavelength was fixed, the bottle was removed, and the GC interface was arranged according to Fig. 7.2 so that the GC eluent could be introduced into the discharge. Samples were prepared in hexane or toluene solvents. The sample solution was injected using a Hamilton Series 7000 (Reno, Nevada) 1 u.1 syringe. Using the current configuration, liquid samples larger than 0.2 u.1 could not used because the discharge was extinguished by the solvent Although the discharge would automatically re-ignite after the solvent front passed through the discharge, deposition of soot inside the discharge tube could be observed. It is possible that this problem could be alleviated by using a make-up gas such as oxygen to create an oxidizing environment inside the plasma discharge. 156 7.3 Results and discussion 7.3.1 Description of the CCP GC detector When operating at 27.12 MHz, energy was lost through heat radiation more easily in the concentric torch (Fig. 7.3a) than the planar, parallel plate design (Fig. 7.3b). This is probably because turbulence is much more serious in the cylindrically shaped torch. Also, when the input power was increased, the discharge in the concentric torch became unstable. Both argon and helium gases were used in this work. It was found that the helium plasma was more suitable for nonmetal elements (I, Cl , and Br); whereas a metallic element (Sn) gave a much larger response in the argon plasma. It is possible that there are some selective energy transfer processes between excited helium and the non-metals which are not operative for Sn. Throughout this work, helium was used when non-metal elements were being detected while argon was used for metal elements. 7.3.2 Spectra of eluents To show that it is feasible to use the CCP to dissociate organic compounds and excite the atoms we have continuously introduced tetramethyl tin (Me4Sn) into the concentric plasma. The tin emission spectrum from Me4Sn between 270 and 306 nm is reproduced in Fig. 7.4a. A toluene background spectrum is provided in Fig. 7.4b. The rf input power was 100 W at 200 KHz and the carrier gas flow was 30 mL/Min. The presence of Sn I lines indicates that this CCP is an effective dissociation and excitation source. Fig. 7.5 showy the emission spectrometric gas chromatograms of iodomethane (CH3I) {Fig. 7.5a} and l,3-diiodopropane(I(CH2)3l) {Fig. 7.5b} solution in hexane. The planar parallel electrode plasma (Fig. 7.3b) was operated at an input power of 50W at a frequency of 27.18 MHz and a GC carrier gas flow rate of 30 mL/min. The monochromator was fixed at the iodine 206.2 nm line. The column temperatures for CH3I and I(CH2)3l were 90 °C and 120 °C respectively. The injector block temperature was maintained at 20 °C 157 3 if) 270.0 Wavelength (nm) Figure 7.4. CCP emission spectrum between 270 and 306 nm. Power 100 W, Frequency: 200 KHz, Carrier gas flow: 30 mlVmin. (a) Spectrum of (G^^Sn detected using the concentric geometry Ar CCP, and (b) background spectrum. 158 0 40 60 RETENTION TTMEtfEQ Figure 7.5. (a) Gas chromatograms for three replicate injections of 150 pg of iodome thane {CH3I} into the GC. (b) Gas chromatogram of 0.1 mL of 0.1 ppm di-iodopropane I{ (012)3) I in hexane detected using the CCP. 1 5 9 higher than the oven temperature in each case.. The interface transfer line temperature was maintained at 10 °C above the column temperatures. The CH3I was injected as the headspace vapor, whereas I(CH2)3l was injected as a solution. From Fig. 7.4 it can be seen that the solvent (hexane) also gave response because of an increase in background emission. This problem could be overcome with background correction by using a polychromator or multichannel photodiode array instead of a single channel PMT detector. 7.3.3 Analytical performance The 3o detection limit for the GC determination I in I(CH2)3l was determined to be 7.2 x 10"14 g/s. 160 7.4 Summary The discharge design described in this chapter provides a stable plasma by capacitive coupling. A He or AT plasma can be generated at atmospheric pressure at any frequency between 0.20 and 27.18 MHz (and probably higher although it has not been tested in this laboratory). Preliminary work using this plasma as an element selective detector for gas chromatography is promising and will be continued. The advantages of the C C P when compared with the MIP are its relatively simple construction, ease of ignition, flexibility in choice of operating conditions (rf power and frequency) and, as Figs. 7.3 as b demonstrate, flexible geometry. 161 References [7.1] A. J. McCormack, S. C. Tong, and W. D. Cooke, Anal. Chem., 1965, 37,1470. [7.2] R. S. Braman and A. Dynako, Anal. Chem., 1968, 40, 95. [7.3] P. C. Uden, R. M . Barnes, and F. P. Disanzo, Anal. Chem., 1978, 50, 852. [7.4] R. J. Lloyd, R. M . Barnes, P. C. Uden, and W. G. Elliott, Anal. C/iem.,1978, 50, 2025. [7.5] R. B. Costanzo and E. F. Barry, Anal. Chem., 1988, 60, 826. [7.6] R. B. Costanzo and E. F. Barry, Appl. Spectrosc.,1988,42, 1387. [7.7] D.L. Windsor and M.B. Denton^p/. Spectrosc, 1978, 32, 366 [7.8] C. A. Bache and D. J. Lisk, Anal. Chem., 1965, 37, 1477. [7.9] C. A. Bache and D. J. Lisk, Anal. Chem.,1966, 38, 783. [7.10] C. A. Bache and D. J. Lisk, Anal. Chem., 1966,38, 1757. [7.11] C. A. Bache and D. J. Lisk, Anal. Chem., 1967, 39, 786. [7.12] C. I. M . Beenakker, Spectrochim. Acta., PartB, 1976, 31B, 483. [7.13] C. I. M . Beenakker, and P. J. W. M . Boumans, Spectrochim. Acta., 1978, 33B, 56. [7.14] Dong C. Liang and M . W. Blades, Anal. Chem., 1988, 60, 27. [7.15] Dong C. Liang and M . W. Blades, Spectrochim. Acta., PartB, in press. 162 Chapter 8 CONCLUDING R E M A R K S 163 In an effort to take the advantages of an electrothermal atomizer and a plasma source, the CCP has been developed and characterized for applications in atomic emission spectrometry (AES), atomic absorption spectrometry (AAS) and gas chromatography (GC). In a graphite furnace, the analyte transport efficiency is high (90-100 %), and analyte residence time in the observation volume is relatively long (0.1 - 0.5 s.). As a consequence, detection limits for GF-AAS are in the sub-ng/mL range for many elements. A limitation of GF-AAS arises as a result of interferences. Moreover, the use of hollow cathode lamps as primary sources restricts AAS to being essentially a single element technique. ICP-AES possesses several distinct advantages over other atomic methods including simultaneous multielement capability, relative freedom from chemical interferences, and a large linear dynamic range. In recent years the ICP has also been used as a source for multielement atomic fluorescence spectrometry and plasma source mass spectrometry. However, to date, the detection limits of the ICP are a few orders of magnitude poorer than those exhibited by GF-AAS. A relatively high support gas flow rate is required to operate an ICP and this acts to dilute the sample atoms. Consequendy the residence time of analyte atoms in the observation volume is short. The CCP torch was initially designed as an atom reservoir for carrying out elemental analysis using atomic absorption. Functionally, the device consists of two parts, the CCP discharge tube and the tantalum strip electrothermal vaporization sample introduction system. The torch design provides for very effective energy transfer from the power supply to the plasma by capacitive coupling. Therefore, the plasma can be generated at atmospheric pressure and in a flexible geometry. The plasma can be operated at very low rf input powers (30 - 600 W) which allows optimal conditions for atom resonance line absorption measurements. Absorption by the analyte takes place within the plasma discharge which is characterized by a long path length (20 cm) and low support gas flow 164 rate (0.2 L/Min). Both of these characteristics provide relatively long residence times. The device exhibits linear calibration plots and provides sensitivities in the range of 3.5 - 40 pg. A preliminary measurement gave a temperature of approximately 4000 K for the discharge. At this temperature, potential chemical interferences are likely minimal. The chemical interference for Fe, A l , As, Ca, Co, Cd, L i , Mo and Sr, were negligible in the determination of silver. Chloride interference, which is prevalent in GF-AAS, was not found. The amount of Ag found in a SMR#1643b (NIST) water sample was 9.5 ± 0.5 ng/g which fell in the certified range of 9.8 ± 0.8 ng/g. Spikes of 30 ng/g and 60 ng/g of silver were added to the SRM and recoveries were found to be in a range from 105 % to 96.2 %. The RSD obtained for 7 replicates of 270 pg silver was 4.6 %. The results for the CCP AES are even more promising. The interferences of thirteen elements are negligible in the determination of silver. The chloride interference was not found. The detection limits for Ag, Cd, L i , Sb and B are 0.7, 0.7, 2, 80 and 400 pg respectively. The amount of silver found in a SRM#1643b (NIST) water sample was 9.3 ± 0.5 ng/g which fell in the certified range of 9.8 ± 0.8 ng/g. Spikes of 30 ng/g and 60 ng/g of silver were added into the SRM#1643b (NIST) samples; the recoveries were found to range from 97 % to 104 %. The RSD obtained for 7 analyses of 270 pg silver were 1.5 % for CCP-AES. It was also found that the signal-to-noise ratios (S/N) are higher in the AES mode than those in the AAS mode in the same CCP atomizer. The CCP torch takes advantage of atmospheric pressure rf sputtering to introduce atoms from a conducting sample into the plasma discharge. Atmospheric pressure rf sputtering generates high density bombarding particles, resulting in a high yield of sputtering. In low pressure discharges, the number density of analyte vapour is limited, as is the resident time of analyte atoms in the source. This is one of the important reasons why low pressure glow discharge sources have relatively low absolute emission intensities. A modified atmospheric pressure capacitively coupled plasma torch has demonstrated its use as a high intense lamp and an excitation source for direct solid sample analysis by AES. 165 The device operates at support gas flow rates between 0.15 to 1 L/Min and at rf input powers between 50 and 500 W. The torch has been used for the determination of manganese in steel samples. A linear calibration plot was obtained over the concentration range 0.005 to 2% Mn. Laser ablation has several advantages over other techniques of vaporization. Using laser ablation, direct analysis of conducting and non-conducting solid samples becomes possible. Analysis can be carried out without addition of reagents and without any separation and/or concentration steps. High spatial resolution, micro and local analysis can also be done. Another advantage is that vaporization and excitation processes can be controlled independently. The CCP torch configuration developed for AAS and AES offers a uniform and stable plasma medium which can be coupled to a laser ablation cell. The experiment reported in this thesis is the first application of laser ablation to the CCP for direct solid analysis. The results obtained from this study suggested the possibility for rapid direct analysis of minor or trace components in solid samples by a laser ablation -CCP system. The detection limit for chromium obtained from this system is superior than those in the laser vaporization - MTP or laser ablation - ICP systems. In order to take advantage of both GF-AAS and ICP-AES, an atmospheric pressure capacitively coupled plasma sustained inside a graphite furnace was developed. This source combines the high efficiency of atomization in furnaces and the high efficiency of excitation in atmospheric pressure plasmas. In general, plasma sources are able to effectively excite high-lying excited states for most metals and non-metals and can also ionize vaporized elements. Therefore the possibility exists of using non-resonance lines to avoid the effects of self-absorption at high analyte concentrations. Ionic lines may also be used in cases where they provide better sensitivity or freedom from spectral interferences. This source also offers the ability to independently optimize vaporization and excitation. However, the most important aspect of this new source is that it can be used for simultaneous, multielement deterrninations of small sample sizes in a graphite furnace atomizer which has 166 been proven to be effective over many years of use. Preliminary quantitative characteristics of this new atmospheric pressure plasma emission source have been studied. The detection limit for Ag of 0.3 pg is lower than the value of 0.4 pg reported for GF-AAS. The use of atomic emission spectroscopic detectors for GC is an important methodology because it can simplify the interpretation of complex chromatograms by providing element specific information about each eluting peak. In this context plasma sources are playing a very important role. An He and Ar CCP, which were generated at atmospheric pressure at any frequency between 0.20 and 27.18 MHz, has been used as detectors for GC. The advantages of the CCP when compared with the MIP are its relatively simple construction, ease of ignition, and flexibility in choice of operating conditions. Preliminary work shows that the CCP is an excellent device for many applications in analytical spectroscopy. The results reported in this thesis show considerable promise. The technique merits further exploration and application. 167 

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:
https://iiif.library.ubc.ca/presentation/dsp.831.1-0059715/manifest

Comment

Related Items