摘要:

Journal of Analytical Atomic Spectrometry (Including Atomic Spectrometry Updates - Formerly ARAAS) JAAS Editorial Board" Chairman: L. C. Ebdon (Plymouth, UK) J. Egan (London, UK) M. S. Cresser (Aberdeen, UK) D. L. Miles (Wallingford, UK) B. L. Sharp (Aberdeen, UK) M. Thompson (London, UK) A. M. Ure (Aberdeen, UK) *The JAAS Editorial Board reports to the Analytical Editorial Board, Chairman J. D. R. Thomas (Cardiff, UK) JAAS Advisory Board F. C. Adams (Antwerp, Belgium) R. M. Barnes (Amherst, MA, USA) L. Bezirr (Budapest, Hungary) R. F. Browner (Atlanta, GA, USA) S. Caroli (Rome, Italy) L. de Galan (Delft, The Netherlands) J. 6. Dawson (Leeds, UK) K. Dittrich (Leipzig, GDR) W. Frech (UmeA, Sweden) K. Fuwa ( Tokyo, Japan) A. L. Gray (Guildford, UK) S. Greenfield (Loughborough, UK) G.M. Hieftie (Bloomington, IN, USA) G. Horlick (Edmonton, Canada) 6. V. L'vov (Leningrad, USSR) J. M. Mermet (Villeurbanne, France) N i Zhe-ming (Beijing, China) N. Omenetto (lspra, Italy) E. PIGko (Bratislava, Czechoslovakia) R. E. Sturgeon (Ottawa, Canada) R. Van Grieken (Antwerp, Belgium) A. Walsh,.,K. B. (Victoria, Australia) B. Welz (Uberlingen, FRG) T. S. West (Aberdeen, UK) Atomic Spectrometry Updates Editorial Board Chairman: *M. S. Cresser (Aberdeen, UK) R. M. Barnes (Amherst, MA, USA) N. W. Barnett (Plymouth, UK) *J. Egan (London, UK) *A. A. Brown (Cambridge, UK) J. C. Burridge (Aberdeen, UK) J. B. Dawson (Leeds, UK) J. R. Dean (Norwich, UK) *L. C. Ebdon (Plymouth, UK) H. J. Ellis (Ross-on-Wye, UK) J. Fijalkowski (Warsaw, Poland) D. J . Halls (Glasgow, UK) S.J. Haswell (London, UK) *D. A. Hickman (London, UK) G. M. Hieftje (Bloomington, IN, USA) S. J. Hill (Plymouth, UK) H. Hughes (Anglesey, UK) P. N. Keliher (Villanova, PA, USA) K. Kitagawa (Nagoya, Japan) K. W. Jackson (Saskatoon, Canada) F. J. M. J. Maessen (Amsterdam, The Nether- *J. Marshall (Middlesbrough, UK) *D. L. Miles (Wallingford, UK) J. M. Mermet (Villeurbanne, France) E. Norval (Pretoria, South Africa) I . Novotny (Brno, Czechoslovakia) P. E. Paus (Oslo, Norway) P. R. Poole (Hamilton, New Zealand) T. C. Rains (Washington, DC, USA) J. M. Rooke (Leeds, UK) G. Rossi (lspra, Italy) I. RubeSka (Prague, Czechoslovakia) A. Sanz-Medel (Oviedo, Spain) *B. L. Sharp (Aberdeen, UK) W. Slavin (Norwalk, CT, USA) R. Stephens (Halifax, Canada) J. Stupar (Ljubljana, Yugoslavia) A.Taylor (Guildford, UK) M. Thompson (London, UK) J. F. Tyson (Loughborough, UK) *A. M. Ure.!Aberdeen, UK) B. Welz (Uberlingen, FRG) J. B. Willis (Victoria, Australia) *D. Littlejohn (Glasgow, UK) lands) *Members of the ASU Executive Committee Editor, JAAS: Judith Egan The Royal Society of Chemistry, Burlington House, Piccadilly, London W I V OBN, UK. Telephone 01-734 9864. Telex No. 268001 US Associate Editor, JAAS: Dr. J. M. Harnly US Department of Agriculture, Beltsville Human Nutrition Research Center, BLDG 161, BARC-EAST, Beltsville, MD 20705, USA. Telephone 301-344-2569 Advertisements: Advertisement Department, The Royal Society of Chemistry, Burlington House, Piccadilly, London W I V OBN. Telephone 01-437 8656. Telex No. 268001 Journal ofAnalytical Atomic Spectrometry (JAAS) (ISSN 0267-9477) is published eight times a year b y The Royal Society of Chemistry, Burlington House, London WIVOBN, UK.All orders accompanied with payment should be sent directly to The Royal Society of Chemistry, The Distribution Centre, Blackhorse Road, Letchworth, Herts. SG6 IHN, UK. 1987 Annual subscription rate UK f180.00, Rest of World f202.00, USA $356.00. Air freight and mailing in the USA b y Publications Expediting Inc., 200 Meacham Avenue, Elmont, NY 11003. USA Postmaster: send address changes t o Journal of Analytical Atomic Spectrometry fJAAS), Publications Expediting Inc., 200 Meacham Avenue, Elmont, NY 11003. Second class postage paid at Jamaica, NY 11431. All other despatches outside the UK by Bulk Airmail within Europe, Accelerated Surface Post outside Europe.PRINTED IN THE UK. 0 The Royal Society of Chemistry, 1987. All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form, or by any means, electronic, mechanical, photographic, recording, or otherwise, without the prior permission of the publishers. Information for Authors Full details of how to submit material for publication in JAASare given in the Instructions to Authors in Issue 1. Separate copies are available on request. The Journal of Analytical Atomic Spectrometry (JAAS) is an international journal for the publi- cation of original research papers, short papers, communications and letters concerned with the development and analytical application of atomic spectrometric techniques.The journal is published eight times a year, includes com- prehensive reviews of specific topics of interest to practising atomic spectroscopists and incor- porates the literature reviews which were pre- viously published in Annual Reports on Analy- tical Atomic Spectroscopy (ARAAS). Manuscripts intended for publication must describe original work related to atomic spec- trometric analysis. Papers on all aspects of the subject will be accepted, including fundamental studies, novel instrument developments and practical analytical applications. As well as AAS, AES and AFS, papers will be welcomed on atomic mass spectrometry and X-ray fluoresc- ence/emission spectrometry. Papers describing the measurement of molecular species where these relate to the characterisation of sources normally used for the production of atoms, or are concerned, for example, with indirect methods of analysis, will also be acceptable for publication.Papers describing the development and applications of hybrid techniques (e.g., GC-coupled AAS and HPLC - ICP) will be parti- cularly welcome. Manuscripts on other subjects of direct interest to atomic spectroscopists, including sample preparation and dissolution and analyte preconcentration procedures, as well as the statistical interpretation and use of atomic spectrometric data will also be accept- able for publication. There is no page charge. The following types of papers will be con- sidered. Full papers, describing original work. Short papers: the criteria for originality are the same as for full papers, but short papers generally report less extensive investigations or are of limited breadth of subject matter.Communications, which must be on an urgent matter and be of obvious scientific importance. Communications receive priority and are usually published within 2-3 months of Veceipt. They are intended for brief descriptions of work that has progressed to a stage at which it is likely to be valuable to workers faced with similar problems. Reviews, which must be a critical evaluation of the existing state of knowledge on a parti- cular facet of analytical atomic spectrometry. Every paper (except Communications) will be submitted to at least two referees, by whose advice the Editorial Board of JAAS will be guided as to its acceptance or rejection.Papers that are accepted must not be published else- where except by permission. Submission of a manuscript will be regarded as an undertaking that the same material is not being considered for publication by another journal. Manuscripts (three copies typed in double spac- ing) should be addressed to: Judith Egan, Editor, JAAS The Royal Society of Chemistry, Burlington House, Piccadilly, London WIV OBN, UK Dr. J. M. Harnly US Associate Editor, JAAS US Department of Agriculture, Beltsville Human Nutrition Research Center, BLDG 161, BARC-EAST, Beltsville, MD 20705, USA or All queries relating to the presentation and submission of papers, and any correspondence regarding accepted papers and proofs, should be directed to the Editor or US Editor (addresses as above).Members of the JAASEditorial Board (who may be contacted directly or via the Editorial Office) would welcome comments, suggestions and advice on general policy mat- ters concerning JAAS. Fifty reprints are supplied free of charge.Journal of Analytical Atomic Spectrometry (Including Atomic Spectrometry Updates - Formerly ARAAS) JAAS Editorial Board" Chairman: L. C. Ebdon (Plymouth, UK) J. Egan (London, UK) M. S. Cresser (Aberdeen, UK) D. L. Miles (Wallingford, UK) B. L. Sharp (Aberdeen, UK) M. Thompson (London, UK) A. M. Ure (Aberdeen, UK) *The JAAS Editorial Board reports to the Analytical Editorial Board, Chairman J. D. R. Thomas (Cardiff, UK) JAAS Advisory Board F. C. Adams (Antwerp, Belgium) R. M. Barnes (Amherst, MA, USA) L.Bezirr (Budapest, Hungary) R. F. Browner (Atlanta, GA, USA) S. Caroli (Rome, Italy) L. de Galan (Delft, The Netherlands) J. 6. Dawson (Leeds, UK) K. Dittrich (Leipzig, GDR) W. Frech (UmeA, Sweden) K. Fuwa ( Tokyo, Japan) A. L. Gray (Guildford, UK) S. Greenfield (Loughborough, UK) G. M. Hieftie (Bloomington, IN, USA) G. Horlick (Edmonton, Canada) 6. V. L'vov (Leningrad, USSR) J. M. Mermet (Villeurbanne, France) N i Zhe-ming (Beijing, China) N. Omenetto (lspra, Italy) E. PIGko (Bratislava, Czechoslovakia) R. E. Sturgeon (Ottawa, Canada) R. Van Grieken (Antwerp, Belgium) A. Walsh,.,K. B. (Victoria, Australia) B. Welz (Uberlingen, FRG) T. S. West (Aberdeen, UK) Atomic Spectrometry Updates Editorial Board Chairman: *M. S. Cresser (Aberdeen, UK) R. M.Barnes (Amherst, MA, USA) N. W. Barnett (Plymouth, UK) *J. Egan (London, UK) *A. A. Brown (Cambridge, UK) J. C. Burridge (Aberdeen, UK) J. B. Dawson (Leeds, UK) J. R. Dean (Norwich, UK) *L. C. Ebdon (Plymouth, UK) H. J. Ellis (Ross-on-Wye, UK) J. Fijalkowski (Warsaw, Poland) D. J . Halls (Glasgow, UK) S. J. Haswell (London, UK) *D. A. Hickman (London, UK) G. M. Hieftje (Bloomington, IN, USA) S. J. Hill (Plymouth, UK) H. Hughes (Anglesey, UK) P. N. Keliher (Villanova, PA, USA) K. Kitagawa (Nagoya, Japan) K. W. Jackson (Saskatoon, Canada) F. J. M. J. Maessen (Amsterdam, The Nether- *J. Marshall (Middlesbrough, UK) *D. L. Miles (Wallingford, UK) J. M. Mermet (Villeurbanne, France) E. Norval (Pretoria, South Africa) I . Novotny (Brno, Czechoslovakia) P. E. Paus (Oslo, Norway) P.R. Poole (Hamilton, New Zealand) T. C. Rains (Washington, DC, USA) J. M. Rooke (Leeds, UK) G. Rossi (lspra, Italy) I. RubeSka (Prague, Czechoslovakia) A. Sanz-Medel (Oviedo, Spain) *B. L. Sharp (Aberdeen, UK) W. Slavin (Norwalk, CT, USA) R. Stephens (Halifax, Canada) J. Stupar (Ljubljana, Yugoslavia) A. Taylor (Guildford, UK) M. Thompson (London, UK) J. F. Tyson (Loughborough, UK) *A. M. Ure.!Aberdeen, UK) B. Welz (Uberlingen, FRG) J. B. Willis (Victoria, Australia) *D. Littlejohn (Glasgow, UK) lands) *Members of the ASU Executive Committee Editor, JAAS: Judith Egan The Royal Society of Chemistry, Burlington House, Piccadilly, London W I V OBN, UK. Telephone 01-734 9864. Telex No. 268001 US Associate Editor, JAAS: Dr. J. M. Harnly US Department of Agriculture, Beltsville Human Nutrition Research Center, BLDG 161, BARC-EAST, Beltsville, MD 20705, USA.Telephone 301-344-2569 Advertisements: Advertisement Department, The Royal Society of Chemistry, Burlington House, Piccadilly, London W I V OBN. Telephone 01-437 8656. Telex No. 268001 Journal ofAnalytical Atomic Spectrometry (JAAS) (ISSN 0267-9477) is published eight times a year b y The Royal Society of Chemistry, Burlington House, London WIVOBN, UK. All orders accompanied with payment should be sent directly to The Royal Society of Chemistry, The Distribution Centre, Blackhorse Road, Letchworth, Herts. SG6 IHN, UK. 1987 Annual subscription rate UK f180.00, Rest of World f202.00, USA $356.00. Air freight and mailing in the USA b y Publications Expediting Inc., 200 Meacham Avenue, Elmont, NY 11003.USA Postmaster: send address changes t o Journal of Analytical Atomic Spectrometry fJAAS), Publications Expediting Inc., 200 Meacham Avenue, Elmont, NY 11003. Second class postage paid at Jamaica, NY 11431. All other despatches outside the UK by Bulk Airmail within Europe, Accelerated Surface Post outside Europe. PRINTED IN THE UK. 0 The Royal Society of Chemistry, 1987. All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form, or by any means, electronic, mechanical, photographic, recording, or otherwise, without the prior permission of the publishers. Information for Authors Full details of how to submit material for publication in JAASare given in the Instructions to Authors in Issue 1.Separate copies are available on request. The Journal of Analytical Atomic Spectrometry (JAAS) is an international journal for the publi- cation of original research papers, short papers, communications and letters concerned with the development and analytical application of atomic spectrometric techniques. The journal is published eight times a year, includes com- prehensive reviews of specific topics of interest to practising atomic spectroscopists and incor- porates the literature reviews which were pre- viously published in Annual Reports on Analy- tical Atomic Spectroscopy (ARAAS). Manuscripts intended for publication must describe original work related to atomic spec- trometric analysis. Papers on all aspects of the subject will be accepted, including fundamental studies, novel instrument developments and practical analytical applications. As well as AAS, AES and AFS, papers will be welcomed on atomic mass spectrometry and X-ray fluoresc- ence/emission spectrometry.Papers describing the measurement of molecular species where these relate to the characterisation of sources normally used for the production of atoms, or are concerned, for example, with indirect methods of analysis, will also be acceptable for publication. Papers describing the development and applications of hybrid techniques (e.g., GC-coupled AAS and HPLC - ICP) will be parti- cularly welcome. Manuscripts on other subjects of direct interest to atomic spectroscopists, including sample preparation and dissolution and analyte preconcentration procedures, as well as the statistical interpretation and use of atomic spectrometric data will also be accept- able for publication.There is no page charge. The following types of papers will be con- sidered. Full papers, describing original work. Short papers: the criteria for originality are the same as for full papers, but short papers generally report less extensive investigations or are of limited breadth of subject matter. Communications, which must be on an urgent matter and be of obvious scientific importance. Communications receive priority and are usually published within 2-3 months of Veceipt. They are intended for brief descriptions of work that has progressed to a stage at which it is likely to be valuable to workers faced with similar problems. Reviews, which must be a critical evaluation of the existing state of knowledge on a parti- cular facet of analytical atomic spectrometry. Every paper (except Communications) will be submitted to at least two referees, by whose advice the Editorial Board of JAAS will be guided as to its acceptance or rejection. Papers that are accepted must not be published else- where except by permission. Submission of a manuscript will be regarded as an undertaking that the same material is not being considered for publication by another journal. Manuscripts (three copies typed in double spac- ing) should be addressed to: Judith Egan, Editor, JAAS The Royal Society of Chemistry, Burlington House, Piccadilly, London WIV OBN, UK Dr. J. M. Harnly US Associate Editor, JAAS US Department of Agriculture, Beltsville Human Nutrition Research Center, BLDG 161, BARC-EAST, Beltsville, MD 20705, USA or All queries relating to the presentation and submission of papers, and any correspondence regarding accepted papers and proofs, should be directed to the Editor or US Editor (addresses as above). Members of the JAASEditorial Board (who may be contacted directly or via the Editorial Office) would welcome comments, suggestions and advice on general policy mat- ters concerning JAAS. Fifty reprints are supplied free of charge.

ISSN:0267-9477    

DOI:10.1039/JA98702FX021

出版商:RSC    

年代:1987

数据来源: RSC

摘要:

JASPE2 2(6) 51 1-648 (1987) September 1987 Journal of Analytical Atomic Spectrometry 1987 WINTER CONFERENCE ON PLASMA AND LASER SPECTROCHEMISTRY, LYON, FRANCE, 12-16 JANUARY, 1987 51 1 51 3 527 533 537 543 549 553 557 561 567 573 579 505 59 1 595 599 607 61 1 61 5 623 629 637 645 CONTENTS Foreword-Jea n-M ic hae I M ermet Inductively Coupled Plasmas: Line Widths and Shapes, Detection Limits and Spectral Interferences. An Integrated Picture. Plenary Lecture-P. W. J. M. Boumans, J. J. A. M. Vrakking Atomisation Efficiency and Over-all Performance of Electrothermal Atomisers in Atomic Absorption, Furnace Atomisation Non-thermal Excitation and Laser-excited Atomic Fluorescence Spectrometry. Plenary Lecture-Heinz Fa I k, Joha n nes Ti Ich Molecular Non-thermal Excitation Spectrometry (MONES): a Procedure for the Determination of Non-metals Using Diatomic Molecules in the Non-thermal (FANES) Atomiser.Part 1. Determination of Fluoride and Chloride Ions by Magnesium Fluoride and Magnesium Chloride MONES. Plenary Lecture-Klaus Dittrich, H. Fuchs State of the Art of Glow Discharge Lamp Spectrometry. Plenary Lecture-Jose A. C. Broekaert Sample Introduction in Plasma Emission and Mass Spectrometry. Plenary LectureRichard F. Browner, Guangxuan Zhu Flow Injection Techniques in Inductively Coupled Plasma Spectrometry. Plenary Lecture-Cameron W. McLeod Trace Enrichment and Determination of Sulphate by Flow Injection Inductively Coupled Plasma Atomic Emission Use of a Glass Frit Nebuliser with a Helium Microwave-induced Plasma-Robert G. StahI, Katherine J.Timmins Relationship Between Detection Limits and Mechanisms in Inductively Coupled Plasma Atomic Emission Measurement of True Gas Kinetic Temperatures in an Inductively Coupled Plasma by Laser-light Scattering. Plenary Laser-enhanced lonisation Spectroscopy in Flames and Plasmas. Plenary Lectur-Gregory C. Turk Thermal Lensing Spectrophotometry of Uraniurn(V1) with Pulsed Laser Excitation-Nicolo Omenetto, Paolo Cavalli, Guglielmo Rossi, Giovanni Bidoglio, Gregory C. Turk Inductively Coupled Plasma Fourier Transform Spectrometry: A New Analytical Technique? Potentials and Problems. Plenary Lecture-Lynda M. Faires Preliminary Results with a High-resolution Inductively Coupled Plasma Fourier Transform Spectrometer-Dominic E. M. Spillane, Richard D. Snook, Anne P.Thorne, J. E. G. Wheaton Role of Aerosol Water Vapour Loading in Inductively Coupled Plasma Mass Spectrometry-Robert C. Hutton, Andrew N. Eaton System Optimisation and the Effect on Polyatomic, Oxide and Doubly Charged Ion Response of a Commercial Inductively Coupled Plasma Mass Spectrometry Instrument-Alan L. Gray, John G. Williams Studies of Metalloprotein Species by Directly Coupled High-performance Liquid Chromatography Inductively Coupled Plasma Mass Spectrometry-John R. Dean, Seumas Munro, Les Ebdon, Helen M. Crews, Robert C. Massey Automation in Element Pre-concentration with Chelating Ion Exchangers. Plenary Lecture-Gunter Knapp, Kurt Muller, Martin Strunz, Wolfhard Wegscheider Fully Automated Dissolution and Separation Methods for Inductively Coupled Plasma Atomic Emission Spectrometry Rock Analysis.Application t o the Determination of Rare Earth Elements. Plenary Lecture-Kuppusami Govindaraju, Guy Mevelle Application of the Grimm Glow Discharge Lamp (GDL) for the Analysis of Geological and Related Materials-Isaac B. Brenner, Kurt Laqua, Michael Dvorachek Inductively Coupled Plasma Atomic Emission Spectrometric Analysis of Cobalt-base Superalloys-A. Gomez Coedo, M. T. Dorado Lopez, A. Vindel Maeso Direct Trace Element Analysis of Tungsten Powders, Alloys and Related Materials by Inductively Coupled Plasma Atomic Emission Spectrometry (ICP-AEW-Isaac B. Brenner, Sarah Erlich, Gilbert Vial, J. McCormack, P. Grosdaillon, Aric E. Asher Determination of Trace Amounts of Palladium, Iron and Copper in Pure Gold by Inductively Coupled Plasma Atomic Emission Spectrometry-Montserrat Baucells, Gloria Lacort, Montserrat Roura Spectrometry-Alan G.Cox, Cameron W. McLeod, Douglas L. Miles, Jennifer M. Cook Spectrometry-Marthe Marichy, Monique Mermet, Jean-Michel Mermet Lecture-Kim A. Marshall, Gary M. Hieftje Typeset and printed by Black Bear Press Limited, Cambridge, EnglandJASPE2 2(6) 51 1-648 (1987) September 1987 Journal of Analytical Atomic Spectrometry 1987 WINTER CONFERENCE ON PLASMA AND LASER SPECTROCHEMISTRY, LYON, FRANCE, 12-16 JANUARY, 1987 51 1 51 3 527 533 537 543 549 553 557 561 567 573 579 505 59 1 595 599 607 61 1 61 5 623 629 637 645 CONTENTS Foreword-Jea n-M ic hae I M ermet Inductively Coupled Plasmas: Line Widths and Shapes, Detection Limits and Spectral Interferences.An Integrated Picture. Plenary Lecture-P. W. J. M. Boumans, J. J. A. M. Vrakking Atomisation Efficiency and Over-all Performance of Electrothermal Atomisers in Atomic Absorption, Furnace Atomisation Non-thermal Excitation and Laser-excited Atomic Fluorescence Spectrometry. Plenary Lecture-Heinz Fa I k, Joha n nes Ti Ich Molecular Non-thermal Excitation Spectrometry (MONES): a Procedure for the Determination of Non-metals Using Diatomic Molecules in the Non-thermal (FANES) Atomiser. Part 1. Determination of Fluoride and Chloride Ions by Magnesium Fluoride and Magnesium Chloride MONES. Plenary Lecture-Klaus Dittrich, H. Fuchs State of the Art of Glow Discharge Lamp Spectrometry. Plenary Lecture-Jose A. C. Broekaert Sample Introduction in Plasma Emission and Mass Spectrometry.Plenary LectureRichard F. Browner, Guangxuan Zhu Flow Injection Techniques in Inductively Coupled Plasma Spectrometry. Plenary Lecture-Cameron W. McLeod Trace Enrichment and Determination of Sulphate by Flow Injection Inductively Coupled Plasma Atomic Emission Use of a Glass Frit Nebuliser with a Helium Microwave-induced Plasma-Robert G. StahI, Katherine J. Timmins Relationship Between Detection Limits and Mechanisms in Inductively Coupled Plasma Atomic Emission Measurement of True Gas Kinetic Temperatures in an Inductively Coupled Plasma by Laser-light Scattering. Plenary Laser-enhanced lonisation Spectroscopy in Flames and Plasmas. Plenary Lectur-Gregory C. Turk Thermal Lensing Spectrophotometry of Uraniurn(V1) with Pulsed Laser Excitation-Nicolo Omenetto, Paolo Cavalli, Guglielmo Rossi, Giovanni Bidoglio, Gregory C.Turk Inductively Coupled Plasma Fourier Transform Spectrometry: A New Analytical Technique? Potentials and Problems. Plenary Lecture-Lynda M. Faires Preliminary Results with a High-resolution Inductively Coupled Plasma Fourier Transform Spectrometer-Dominic E. M. Spillane, Richard D. Snook, Anne P. Thorne, J. E. G. Wheaton Role of Aerosol Water Vapour Loading in Inductively Coupled Plasma Mass Spectrometry-Robert C. Hutton, Andrew N. Eaton System Optimisation and the Effect on Polyatomic, Oxide and Doubly Charged Ion Response of a Commercial Inductively Coupled Plasma Mass Spectrometry Instrument-Alan L. Gray, John G. Williams Studies of Metalloprotein Species by Directly Coupled High-performance Liquid Chromatography Inductively Coupled Plasma Mass Spectrometry-John R.Dean, Seumas Munro, Les Ebdon, Helen M. Crews, Robert C. Massey Automation in Element Pre-concentration with Chelating Ion Exchangers. Plenary Lecture-Gunter Knapp, Kurt Muller, Martin Strunz, Wolfhard Wegscheider Fully Automated Dissolution and Separation Methods for Inductively Coupled Plasma Atomic Emission Spectrometry Rock Analysis. Application t o the Determination of Rare Earth Elements. Plenary Lecture-Kuppusami Govindaraju, Guy Mevelle Application of the Grimm Glow Discharge Lamp (GDL) for the Analysis of Geological and Related Materials-Isaac B. Brenner, Kurt Laqua, Michael Dvorachek Inductively Coupled Plasma Atomic Emission Spectrometric Analysis of Cobalt-base Superalloys-A. Gomez Coedo, M. T. Dorado Lopez, A. Vindel Maeso Direct Trace Element Analysis of Tungsten Powders, Alloys and Related Materials by Inductively Coupled Plasma Atomic Emission Spectrometry (ICP-AEW-Isaac B. Brenner, Sarah Erlich, Gilbert Vial, J. McCormack, P. Grosdaillon, Aric E. Asher Determination of Trace Amounts of Palladium, Iron and Copper in Pure Gold by Inductively Coupled Plasma Atomic Emission Spectrometry-Montserrat Baucells, Gloria Lacort, Montserrat Roura Spectrometry-Alan G. Cox, Cameron W. McLeod, Douglas L. Miles, Jennifer M. Cook Spectrometry-Marthe Marichy, Monique Mermet, Jean-Michel Mermet Lecture-Kim A. Marshall, Gary M. Hieftje Typeset and printed by Black Bear Press Limited, Cambridge, England

ISSN:0267-9477    

DOI:10.1039/JA98702BX023

出版商:RSC    

年代:1987

数据来源: RSC

摘要:

Professor 0 ttaway Memorial Meeting Edinburgh, UK, November 5th and 6th, 1987 The Analytical Division of the Royal Society of Chemistry is holding a meeting to commemorate the late Professor John Ottaway. It will be held on Thursday and Friday, November 5th and 6th, 1987, at the Royal Society of Edinburgh, George Street, Edinburgh, UK. Programme “Fresh Perspectives in Analytical Atomic Spectroscopy: Don’t Believe Everything You Read,” Professor T. C. O’Haver (University of Maryland, USA). “Electrothermal Atomisation-A Successful Co-operation Between University and Industry,” Mr. L. Morris (Pye Unicam, Cambridge, UK). “Graphite Furnace Atomic Emission Spectroscopy: The Rediscovery of a Technique,” Dr. D. Littlejohn (University of Strathclyde, Glasgow, UK). “Evaluation of a Mathematical Model for Peak Interpretation in GFAAS Based on Free Analyte Atom Deposition,” Dr.B. Welz (Perkin-Elmer & Co. GmbH, Uberlingen, FRG). “The Role of Kinetics in Analytical Chemistry,” Dr. C. Fuller (CECB, Nottingham, UK). “Conditions, Kinetics and Mechanisms in the Titration of Hydrazine with Bromate and the Mechanism of Rosaniline Visual indication,” Professor E. Bishop (University of Exeter, UK). “Why Fluorescence?-Photon Emission Methods in Trace Molecular Analysis,” Professor J. N. Miller (University of Technology, Loughborough, UK). “Applications of Atomic Spectroscopy in Clinical Chemistry,” Dr. G. S. Fell (Royal Infirmary, Glasgow, UK). “Graphite Furnace AAS on the Way to Absolute Analysis,” Professor B. V. L’vov (Polytechnical Institute, Leningrad, USSR).“Furnace Atomisation for Multi-element AAS,” Dr. J. M. Harnly (US Department of Agriculture, Maryland, USA). “The ICP: Is It the Real Thing?” Dr. J. Marshall (ZCI, Middlesbrough, UK). “Zn situ Pre-concentration in Flame Atomic Spectrometry,” Professor T. S. West (Macaulay Institute for Land Use Research, A berdeen, UK) . “Furnace Probe for Molecular Emission Cavity Analysis of Solid Samples,” Professor A. Townshend (University of Hull, UK). “Determination of Metal Contaminants in Air by GFAAS Monitor,” Dr. Lazslo Bezur (Technical University, Budapest, Hungary). “Flame and Furnace; Emission and Absorption: A Historical Dialogue,” Dr. A. M. Ure (Macaulay Institute for Land Use Research, Aberdeen, UK). “Results of Joint Research with John Ottaway on FANES Techniques,” Professor H.Falk (Academy of Sciences of the GDR, Berlin, GDR). “Surfactant Formulations-Separations and Analysis,” Dr. W. C. Campbell (ICZ Petrochemicals Division, Middlesbrough, UK). “The Teaching of Analytical Science into the 21st Century,” Dr. J . F . Alder (UMIST, UK). “Spectroscopy-The Scottish Dimension ,” Professor D. T. Burns (Queen’s University of Belfast, U K ) . The meeting will also include poster sessions and there will be a Conference Dinner on the Thursday evening. There is a registration fee for this meeting. Further details can be obtained from: The Analytical Division, Royal Society of Chemistry, Burlington House, Piccadilly, London W1V OBN, UK 01 437-86561988 Winter Conference On Plasma Spectrochemis try San Diego, California, USA January 5 9 , 1988 The 1988 Winter Conference on Plasma Spectrochemistry, fifth in a series of biennial meetings sponsored by the ICP Information Newsletter, will feature developments in plasma spectrochemical analysis by inductively coupled plasma (ICP), d.c.plasma (DCP), microwave plasma (MIP) and glow and hollow-cathode discharge (GDL, HCL) sources. The meeting will convene Monday, January 4 to Saturday, January 9,1988 at the San Diego Princess resort and convention centre in San Diego. Expert short courses at introductory and advanced levels and an exhibition of spectroscopic instrumentation also will be included. ~ Programme and Objectives Symposia organised and chaired by recognised experts will include the following topics: (1) Sample introduction and transport phenomena; (2) Listrumentation and automation, including on-line analysis and remote systems; (3) Excitation mechanisms and plasma characteristics; (4) Interferometry; ( 5 ) Atomic fluorescence; (6) Glow and hollow-cathode discharges; (7) Flow injection analysis; (8) Chromatography and plasma detectors; (9) Plasma source mass spectrometry; (10) Industrial applications of ICP mass spectrometry; and (11) Sample preparation and pre-concentration techniques.Six plenary and 15 invited lectures will be presented. Three afternoon poster sessions will feature applications, automation and new instrumentation. Four panel discussions will address critical development areas. Plenary, invited and submitted papers will be published as the official conference proceedings following the meeting after peer review in Journal of Analytical Atomic Spectrometry , September 1988 issue.Instrument Exhibition A three day exhibition of spectroscopic instrumentation and chemicals , electronics, glassware, publications and software supporting plasma spectroscopy will complement the scheduled sessions. Expert Short Courses Introductory and advanced four-hour short courses will be offered January 2-3 and 9,1988. Designed to provide background and intensive training in popular topics of plasma spectrochemistry, these will cover analytical applications, instrumentation, samples introduction and various techniques (e.g., plasma diagnostics, scientific writing, chemical and physical pre-concentration and applications of isotope dilution and tracers). Registration The conference registration fee includes a copy of the conference proceedings, abstracts, a tee-shirt and conference dinner. The pre-registration fee is $275 until October 16, 1987, after which time it will be $375. On-site registration will be $400. Discounts are provided for students, and no registration fee is required for spouses. Short-course pre-registration fee $75 for each four-hour short course, after October 16 this will be $100. Further details on all aspects of the Conference can be obtained from: Dr. Ramon M. Barnes Department of Chemistry, GRC Towers, University of Massachusetts, Amherst, MA 01003-0035, USA (413) 545-2294

ISSN:0267-9477    

DOI:10.1039/JA98702FP037

出版商:RSC    

年代:1987

数据来源: RSC

摘要:

511 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, SEPTEMBER 1987. VOL. 2 Foreword 1987 Winter Conference on Plasma and Laser Spectrochemistry: Lyon, France, January 12th-16th, 1987 Following the previous “Winter Confer- ences on Plasma Spectrochemistry” organised by R. M. Barnes in Puerto-Rico (1980), Orlando (1982), San Diego (1984), a European Winter Conference was organised in Leysin, Switzerland in 1985, by J. M. Mermet and E. Janssens. About 170 scientists participated in the Leysin meeting (15 invited lectures and 40 submitted contributions). In 1986 the Winter Conference was held in Kona, Hawaii. In 1987, the Conference went back to Europe and was organised in Lyon, France. The name “Winter” was certainly apt as the weather conditions during the week were the most severe experienced in Europe in recent years.Fortunately, Lyon is a city where it is easy to survive and the “Lyonnaise” gastron- omy was one of the highlights of the week. The Lyon Conference was organised by J. M. Mermet, E. Janssens and N. Omenetto. In addition to the plasma field, the subject area was extended to include laser spectrochemistry . Thirteen invited lectures and 98 contributions were effectively presented. The number of participants in the Conference was 220. The invited lectures were given by T. Berthoud (Fontenay-aux-Roses, France), M. Blades (Vancouver, Canada), P. W. J. M. Boumans (Eindhoven, The Nether- lands), J. Broekaert (Dortmund, FRG), R. F. Browner (Atlanta, USA), K. Dit- trich (Leipzig, GDR), L. Faires (Los Alamos, USA), H. Falk (Berlin, GDR), K.Govindaraju (Nancy, France), G. M. Hieftje (Indiana, USA), G. Horlick (Edmonton, Canada), C. W. McLeod (Sheffield, UK) and G. C. Turk (Gaithersburgh, USA). Most of the sub- mitted contributions were given in the form of posters during three half-day sessions. It was unanimously agreed that the scientific and artistic quality of the posters was high. This high standard was no doubt encouraged by an award for the best poster presentation from one of the sponsors, P hilips Analytical, the winner being J. W. M. Kocken (The Nether- lands). Other contributing sponsors were Applied Research Laboratories, Jobin- Yvon and Perkin-Elmer. A conference should be a reflection of the present developments of the tech- nique and it was no surprise to observe that 20 papers were devoted to ICP-MS, confirming that it really is a growing field.Other sessions dealt with “Recent advances in plasma sources and exotic plasmas,” “Sample introduction, torches and HF generators,” “Diagnostics and processes,” “Fourier transform spec- trometry,” “Spectrometers ,” “Laser spec- trometry” and “Applications. ” Traditionally, a large number of the papers presented at the various “Winter Conferences” have been published either in the form of a book,l or as special issues of Spectrochirnica Acta, Part B.2-5 For the Lyon Conference, the Royal Society of Chemistry kindly agreed to publish a selection of the papers presented at the conference in a special issue of the Jour- nal of Analytical Atomic Spectrometry (JAAS). Thanks to the co-operation of the authors, the referees and Judith Egan (Editor of the journal), it has been possible to publish the present issue only nine months after the meeting.This issue contains 23 papers, following the normal refereeing procedures of JAAS. This combination of reviews, fundamental and applied papers efficiently illustrates the state of the art in this field and the strong activity of researchers, companies and users, which confirms the value of such meetings where an efficient exchange of information can be obtained. That is why another Winter Conference will be organ- ised by R. M. Barnes in 1988, in San Diego, USA. A selection of papers presented at this Conference will also be published in JAAS. References 1. Barnes, R. M . , Editor, “Developments in Atomic Plasma Spectrochemical Analysis,” Heyden, London, 1981. 2. “Plasma Spectrochemistry I,” Spectro- chim. Acta, Part B , 1983, 38, NO. 112. 3. “Plasma Spectrochemistry 11,” Spectro- chim. Acta, Part B, 1985, 40, No. 112. 4. “Plasma Spectrochemistry III,.’ Spectro- chim. Acta, Part R, 1986, 41, No. 112. 5. “Plasma Spectrochemistry IV,” Spectro- chim. Acta, Part B , 1987, 42. No. 112. Abstracts of the Lyon Conference are avail- able in ICP Information Newsletter, 1987. 12, 704-766 and 794-821. Jean-Michel Mermet Conference Chairman

ISSN:0267-9477    

DOI:10.1039/JA9870200511

出版商:RSC    

年代:1987

数据来源: RSC

摘要:

JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, SEPTEMBER 1987, VOL. 2 513 Inductively Coupled Plasmas: Line Widths and Shapes, Detection Limits and Spectral Interferences. An Integrated Picture* Plenary Lecture P. W. J. M. Boumans and J. J. A. M. Vrakking Philips Research Laboratories, P.O. Box 80.000, 5600 JA Eindhoven, The Netherlands This paper discusses the significance of the basic knowledge of atomic spectra for the interpretation of detection limits, the optimisation of line selection and the design of interference libraries for universal application in inductively coupled plasma atomic emission spectrometry (ICP-AES). Specifically, the paper reviews (i) the measurement of the physical widths and shapes of 350 prominent ICP lines of 65 elements, (ii) the use of these results for the breakdown of detection limits in general and the assessment and comparison of three comprehensive sets of detection limits in particular, (iii) the development and application of a new rational criterion for line selection in dependence on sample composition and spectral resolution and (iv) the prospects of using physically resolved spectral data in future compilations of spectral information for ICP-AES.Although the topic is treated with reference to ICP-AES, many aspects of the discussion are of general interest in AES. Keywords : Inductively coupled plasma; atomic emission spectrometry; line widths and shapes; spectral band width; spectral interferences The availability of basic spectroscopic data, such as transition probabilities, excitation and ionisation energies, cross-sections for a variety of processes and line widths, is indispensable in plasma diagnostics to describe the processes and mechanisms that govern the generation of the signals used for analytical purposes.Insight into these processes and mechanisms, in turn, helps clarify observations made under analytical conditions and, more importantly, may be used to rationalise analytical procedures. One of the areas where such a rationalisation is still badly needed is the part of atomic emission spectroscopy concerned with line selection and spectral interferences. In a recent review, “A century of spectral interferences in atomic emission spectroscopy-Can we master them with modern apparatus and approaches?,”’ it was indicated that, in spite of various recent attempts to produce new data compila- tions, the real progress is still rather meagre, because the capabilities of computers and modern spectroscopic instru- ments have not yet been sufficiently exploited.This situation can be expected to change rapidly in the forthcoming years. In this light the present text discusses (i) some recent measure- ments of basic spectroscopic data, (ii) the direct application of these data to analytical atomic emission spectroscopy and (iii) the perspectives of extending the data base and using it in a rational way to face the problem of line selection and spectral interferences in inductively coupled plasma atomic emission spectrometry (ICP-AES) . Specifically, the following topics will be addressed: (a) The measurement of the effective shapes and physical widths of about 350 prominent lines, including lines with hyperfine structure (HFS), of 65 elements emitted from an ICP.(b) The use of the results of the line width measurements (a) in a breakdown of three extensive sets of AES detection limits for argon ICPs compiled with three different ICP sources and spectrometers in three laboratories. (c) The development of a novel analytizal figure merit as a rational criterion for line selection, that is, the “true detection limit” for real samples, formulated in such a way that both the selectivity and the “conventional detection limit” for smooth background are taken into account. * Presented at the 1987 Winter Conference on Plasma and Laser Spectrochemistry, Lyon, France, 12th-16th January, 1987.(d) The interference library of the 1990s: a library in the software domain based on physically resolved spectral data, from which the effective data relevant to a particular spectroscopic apparatus can be derived by convolution with the instrumental function. Physical Widths of Lines Emitted from an ICP Data Available up to the End of 1986 In the textbook “Inductively Coupled Plasma Emission Spectroscopy”2 Boumans summarises the literature con- cerned with determinations of physical line widths in ICPs, as follows. (a) Human and Scott3 found a width of 3.6 pm for Ca 1422.673 nm using a Fabry - Perot interferometer. (b) Also using interferometry Kawaguchi et al.4 measured the widths of 15 lines of 10 elements. Their results are based on the assumption of a Lorentzian instrumental profile and Gaussian profiles for both hollow-cathode and ICP lines, yielding a Voigt profile as experimental profile.For the de-convolution Kawaguchi et al. used an approximation proposed by Posener,s but Boumans and Vrakking6 recalcu- lated their results using a more accurate formula. (c) Hasegawa and Haraguchi7 reported widths for 34 lines of 19 elements obtained with an echelle monochromator with crossed dispersion incorporating a refractor plate for wavelength scanning. Their results are based on the assump- tion that the experimental profile is a Voigt profile of which the Gauss (or Doppler) component of the half-width and the a parameter can be determined with a curve-fitting procedure. The Gauss and Lorentz components are then split separately, into an instrumental and a physical part, using quadratic de-convolution for the Gauss components and linear de- convolution for the Lorentz components.(d) Batal and Mermetg calculated line widths for nine atomic or ionic lines of Ca, Mg and Sr on the basis of Doppler broadening ( T = 5000 K) and van der Waals interaction. (e) Broekaert et al.9 determined the widths of 18 lines of 16 rare earths using a spectrograph provided with an order sorter. They used the same method and equipment as had been used previously by Laqua et ~ 1 . 1 0 for the measurements of the widths of 67 lines of 40 elements emitted from four types of d.c. arc and a soft spark.514 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, SEPTEMBER 1987, VOL. 2 (f) Using high-resolution Fourier transform spectrometry (FTS) Faires et af.11 determined the widths and shapes of 81 Fe I lines in the spectral range 290-390 nm.(8) Boumans and Vrakking6 determined line widths with a 1.5-m echelle monochromator with pre-disperser. They covered 16 Fe I lines (364-384 nm) for comparison with the results of Faires et al. 11 and subsequently 13 lines of nine other elements. Recently, Boumans and Vrakking12 extended the latter approach to ca. 350 prominent lines of 65 elements. This work will be considered more extensively in the next section. New Data Published at the End of 1986 As in the initial approach6 for the line-width measurements Boumans and Vrakking used a 1.5-m echelle monochromator with pre-disperser, the characteristics of which have been described previously.6.13 If such an instrument is used for line-width measurements, two conditions must be fulfilled: (i) it should be possible to adjust the spectral band width so that it matches the physical line widths, and (ii) the band width should be accurately known. Generally, the practical spectral band width of a dispersive spectrometer consists of three components: resultant spectral slit, diffraction width and a contribution from aberrations.The resultant spectral slit is the product of the largest slit width and the reciprocal linear dispersion. The latter quantity can be calculated according to the procedure reviewed in recent literature. 6,1&16 At the slit widths used with the echelle monochromator,h the diffraction width is negligible, but the contribution from aberrations is large, which appears to be a feature of all dispersive spectroscopic instruments.14 Initially Boumans and Vrakking6 determined the contribution from aberrations by measuring the widths of hollow-cathode lines, but experienced the problem that the value of the aberration correction varied with the kchelle angle.This difficulty could be solved by avoiding small echelle angles and using the echelle then at off-blaze settings at the opposite angle in the next high order.12 The contribution from aberrations then enters the formula for the band width as a constant correction of the slit width. The value of the correction was finally adjusted so that the results of line width measurements for iron lines were, on average, identical with those of Faires et ~1.11 who used high-resolution FTS with negligible instrumental broadening.The widths subsequently measured for the lines covered by Hasegawa and Haraguchi’ and Kawaguchi et af.4 were found to be intermediate between the results reported by the latter two groups of authors. In view of these agreements the approach was considered to be reliable enough to be extended to about 350 prominent lines of 65 elements, of which the four to five “most” prominent lines were usually covered. More lines were included for Au, Cr, Eu, Hf, Ho, Lu, Mo, Nb, Ni, Pd, Ru, V and W. The measurements were made in the analytical zone (12-mm observation height) of a conventional 50-MHz argon ICP, operated at 1.15 kW. The spectral band width varied between 1.3 and 3.8 pm, depending on wavelength, and was almost always smaller than 1.5 times the found physical line widths.Each measurement comprised a computer-controlled step- wise scan of the effective line profile. The data were stored on disks and subsequently plotted using a five-point, second- order Savitzky - Golay fit.17 Examples are shown in Figs. 1-3. The effective line width (EFW) was determined “manually” in the plot. The physical line width (PHW) was calculated from PHW = d(EFW*-BW2) . . . . (1) where BW is the practical band width. The Doppler width was computed for a temperature of 6300 K, that is, the temperature which, according to the measure- ments by Faires et al.,11 is likely to prevail under the Wavelength Fig. 1. Examples of effective line profiles with virtually Gaussian shapes for which the physical width found is close to the Doppler width if a Doppler temperature of 6300 K is assumed.12 Spectral band width, 2.35 pm.( a ) Tm I1 313.126 nm, PHW = 1.4 pm; ( b ) Be I1 313.042 nm, PHW = 6.4 Dm. (ReDroduced with Dermission from Boumans, P. W. J. M., andbrakkini, J. J. A. M., Sp’ectrochirn. Acta, Part B , 1986,41, 1235) Wavelength Fig. 2. Effective profile of Co I1 238.892 nm as an example of a line with unresolved HFS.12 The Co line has an apparent physical width of 5.7 pm (a = 2.4) and is compared with a simple Fe line of about the same wavelength having a physical width of 2.1 pm. Spectral window is 16.7 pm. (Reproduced with permission from Boumans, P. W. J . M., and Vrakking, J. J. A. M., Spectrochim.Acta, Part B , 1986,41, 1235) c Ho379.675 I Wavelength Fig. 3. Effective profiles of four lines for which complex HFS is evident. Horizontal lines indicate the apparent physical widths. a-f refer to the peaks of HFS components which were identified and whose wavelengths are included in the “atlas of spectral scans” in reference 12. (ReproducedwithpermissionfromBoumans, P. W. J. M., and Vrakking, J. J. A. M., Spectrochim. Acta, Part B , 1986,41,1235)JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, SEPTEMBER 1987, VOL. 2 515 experimental conditions used. The calculation of the a parameter from the measured physical line width and the Doppler width then provided, in principle, an estimate of the contribution from Lorentz broadening. Actually, however, the values of the a parameter served as an indication of the presence or absence of HFS.It was known from previous w0rk4~63~Jl that the lines emitted from an argon ICP generally show little Lorentz broadening, H and Ar lines being the exceptions. This was confirmed by the present measurements in that, for the majority of the lines, the values of the a parameter varied between 0 and 0.5. Therefore, an a value of 0.5 was chosen as a criterion for assessing the plausibility of HFS. Evidently this is a reasonable, but not rigorous, criterion. HFS became just discernible when a was larger than ca. 2 and was clearly revealed when a exceeded a value of 3, as is illustrated in Fig. 3. In these instances the components of the HFS composites were identified using the tabulated wavelengths18 of narrow lines of other elements as references.The main results are presented12 in the following forms: (a) A table with wavelengths, physical line widths, Doppler widths and a values, of which a sample page is reproduced as Table 1. (b) An atlas of some 90 effective profiles of HFS compo- sites, of which Fig. 3 provides examples. The last column of Table 1 refers to the figure numbers of the profiles in the atlas. (c) A tabulation of the wavelengths of HFS components which are of potential interest in spectrochemical analysis as separate prominent analyses and/or interfering lines. The entirety of the results can be basically summarised as follows: (i) Many prominent ICP lines show chiefly Doppler broadening (a < 0.5). The physical widths of these lines lie between 0.9 and 1.2 pm for heavy elements and between 4 and 8 pm for light elements.Fig. 1 provides two examples; the narrow profile of Tm I1 313.126 nm, for which PHW = 1.4 pm (a = 0.05), and the broad profile of Be I1 313.042 nm, for Table 1. Sample page from the complete table in reference 12 listing the measured physical widths (Ahphys), the Doppler widths (Ah,) calculated for T = 6300 K, and the corresponding a values for the 350 prominent lines covered. Column 2 states the concentration at which the measurement was made, while the last column refers to the figure if included in the atlas of spectral scans. (Reproduced with permission from Boumans, P. W. J. M., and Vrakking, J. J. A. M., Spectrochirn. Acta, Part B , 1986,41, 1235) Spectral line/ nm Ag I1 224.641 Ag I1 243.779 Ag I 328.068 Ag I 338.289 A1 I 237.312 A1 I 308.215 A1 I 309.271 A1 I 396.152 Ar I 415.859 As I 193.696 As I 197.197 As I 200.334 As I 228.812 Au I 197.745 Au I1 200.081 Au I1 208.209 Au I1 211.068 Au I 242.795 Au I 267.595 B I 208.893 B I 208.959 B I 249.678 B I 249.773 Ba I1 230.424 Ba I1 233.527 Ba I1 455.403 Ba I1 493.409 Be I 234.861 Be I 249.473 Be I1 313.042 Be I1 313.107 Bi I 206.170t$ 206.148t 206.152-l 206.155t Bi I 223.061 Bi I 222.825 Bi I 306.772t 306.776 t 306.775T Concentration/ pg ml-I 300 300 30 30 100 100 100 100 - 300 1000 1000 1000 300 300 300 300 300 300 300 300 20 20 25 100 1 25 2.5 30 2.5 2.5 1000 - - - 300 300 300 - - Ahphysf Pm 1.4 1.5 2.1 1.8 3.6 4.2 4.0 5.1 5.3 1.3 1.5 1.5 1.8 0.9 0.9 1.3 1.2 1.9 3.1 4.3 4.4 5.1 5.0 1.2 1.5 3.6 3.4 4.7 5.5 6.4 6.2 13.5 0.7 1.3 1.5 1.5 1.3 11.5 1.6 1.2 AhD/ Pm 1.2 1.3 1.8 1.8 2.6 3.4 3.4 4.3 3.7 1.3 1.3 1.3 1.5 0.8 0.8 0.8 0.8 1 .o 1.1 3.6 3.6 4.3 4.3 1.1 1.1 2.2 2.4 4.4 4.7 5.9 5.9 0.8 0.8 0.8 0.8 0.9 0.9 1.2 1.2 1.2 U 0.20 0.20 0.30 0.00 0.60* 0.40 0.25 0.25 0.60* 0.05 0.30 0.20 0.30 0.25 0.20 0.75* 0.55* 1.2* 2.1* 0.30 0.35 0.25 0.25 0.10 0.45 0.85* 0.60* 0.10 0.25 0.15 0.10 13.8* 0.30 0.75* 1.1* 0.90* 0.70* 7.9* 0.45 0.05 * HFS is perceptible if a > 2; otherwise the line only appears broadened or asymmetrically distorted.1- Partly or completely resolved HFS: the wavelengths and widths of completely resolved components are listed here; the wavelengths of all $ Wavelength of the centre of the quintet structure (Fig. 3 in this paper) determined to be 206.155 nm using Cr 206.149 as reference.identifiable components are stated in the relevant figure caption and in Table 6 of the original paper.516 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, SEPTEMBER 1987, VOL. 2 which PHW = 6.4 pm (a = 0.15). The Be line can be used to measure changes in Doppler temperature because the absolute change in line width with temperature is relatively large, owing to the small atomic mass of Be. Thus it was found, for example, that the Doppler temperature in the. ICP increased from the assumed 6300 K to 7400 K when the power was increased from 1.15 to 1.65 kW. (ii) Substantial broadening, most probably resulting from unresolved HFS, was found for lines of Co, Mn, Pb, Pt, Sb and V. As an example, Fig. 2 contrasts Co I1 238.892 nm (PHW = 5.7 pm, a = 2.4) and Fe I1 239.562 nm (PHW = 2.1 pm, a = (iii) Very broad structures with partly or wholly resolved HFS were found for Bi, Eu, Ho, In, La, Lu, Nb, Pr, Re, Ta and Tb. The separate components of these structures showed mainly Doppler broadening. A complete analysis of the complex structures is possible with curve-fitting techniques.19 Fig.3 shows four examples. The horizontal line in each frame represents the “apparent line width,” that is, the width of the line when the structure is unresolved. Analytically this width is of interest with regard to the behaviour of the signal to background ratio (SBR) of the line, as is discussed below. 0.20). Two main conclusions were drawn from this work: (1) The knowledge of the physical widths of a great number of prominent lines will permit an accurate calculation of the effect of spectral band width on “conventional” detection limits (smooth background) and thus allow unambiguous comparisons of detection limits in real situations.(2) It appears possible to set up an interference library (software) for ICPs based on physically resolved spectra from which, for any spectroscopic instrument, the effective spectra can be computed to assess the interferences and to select the best analysis lines a priori or a posteriori for any specified sample type, if an appropriate criterion is specified. These points will be elaborated on in the following sections. Breakdown of Detection Limits Rigorous Comparison of Detection Limits The knowledge of the physical widths of a great number of prominent lines permitted (i) a rigorous comparison20 of three extensive sets of detection limits reported for different argon ICPs and (ii) the establishment of “standard” values for a 50-MHz ICP and 15 pm spectral band width.These values may be considered as being representative of the performance of the conventional argon ICP. The principle of this approach originates from the work of Laqua and co-workers.10.21 Recently, Boumans and Vrakking applied this approach to ICP detection limits, first tentatively, using empirical approxi- mations,22 then rigorously,20 using the values of the physical line widths discussed above. The approach implies the breakdown of detection limits into the factors contributed by the SBR characteristics of the source, the spectral band width of the spectroscopic apparatus and the noise characteristics of the complete system (source and spectroscopic apparatus).The following sections discuss the relationships involved and their application to an assess- ment and comparison of detection limits reported by Winge and co-workers,23.24 Wohlers2S and Boumans and Vrakking.20 Relationships 10,13,20,21,2628 For a smooth, featureless background the detection limit (q) can be formulated as cL = 2V2 (0.01 RSDB) c@BR . . (2)* * For fundamental reason^,*^^^* equation (2) is written with the factor 2 ~ 2 instead of the commonly used factor 3; numerically the difference is immaterial. SBR is the signal to background ratio for an analyte concentration cO: SBR=XA/XB .. . . * (3) where xA is the net analyte signal and xB the background signal. RSDB is the relative standard deviation of the background signal (“/o), which can be written as RSDB = (G + P / x B + Y/X;)’” . . . . (4) where the terms in parentheses account for flicker noise, shot noise and detector noise, respectively. In ICP-AES, the flicker noise coefficient (aB) is usually 0.5-1%. The value of the shot noise coefficient (P) depends on the units in which XB is expressed.26 In the case of photon counting P = lo4 if RSDB is expressed as a percentage.13 For a good photomultiplier, detector noise is negligible ( y = 0). A measured SBR can be written as the product of two factors: (SBR)meas = fopt(SBR)source * - * * ( 5 ) where (SBR),ource is the SBR at infinite resolution (BW = 0) andfop, is the factor by which this “source SBR” is modified by the spectroscopic apparatus (BW > 0).In view of equations (2) and ( 5 ) , a ratio R(I/JI) of two detection limits obtained with methods I and I1 can be written as the product of three factors: R(I/II) = FnoiseFoptFsource * - . - (6) where Fnoise = RSDB(I)/RSDB(II) . . . . (7) Fopt = fopt(IIYfopt(1) * . . . * . (8) and Fsource = SBRso”rc,(II>/SBR,ou~c~(I) * (9) Equation (6) was applied to data sets for which R(I/II) and Fnoise are known and Fop, can be computed from the known spectral band widths and physical line widths. The latter calculation is based on the following, experimentally established dependence of fop, on BW and PHW: fopt = 1 if PHW > BW . . . . (10) . . (11) . .(12) fop, a l/m if PHW < BW < 2 PHW fopt a 1/EFW if BW > 2 PHW where EFW is the effective line width [cf., equation (l)]: EFW = d(PHW2+BW2) . . . . (13) The significance of equations (10)-(12) is illustrated by the data in Table 2 and Fig. 4 (see also the Appendix). Table 2 shows in a self-explanatory way the calculated effect of a change in spectral band width from 20 to 3 pm in terms of the ratio of the effective line widths and the ratio of the corresponding values of fopt for spectral lines with different physical widths. Fig. 4 depicts the profiles of the narrow Mo line at 317.0 nm and the broad Nb line at 309.4 nm, measured at two spectral band widths. For the Mo line, an increase in band width by a factor of ca. 3 results in a reduction of the SBR by a factor of 2.6, whereas for the Nb line hardly any change in SBR is observed.The results for the Nb line clearly demonstrate that a line behaves as a continuum when the spectral band width is smaller than the physical line width. Comparison of Three Sets of Detection Limits20 The experiment involving the line-width measurements12 discussed above also covered the measurement of the SBRs and background signals for the 350 prominent lines emitted by the 50-MHz ICP. The latter measurements were not made atJOURNAL O F ANALYTICAL ATOMIC SPECTROMETRY, SEPTEMBER 1987, VOL. 2 517 Table 2. Effect of a change in spectral band width from 3 to 20 pm on the effective width (EFW) and fopt for prominent lines with different physical widths (PHW). The effect is expressed in terms of the ratios EFW(20)/EFW(3) and f0pt(3)/f0pt(20).[Reproduced with permission from Boumans, P. W. J. M., Labstract, 1986, No. 3, 8 (Philips Nederland, Eindhoven)] Spectral line nm Au I1 200.081 U I1 263.553 Mo I1 202.030 Nb I1 269.706 Ag I 328.068 Lu TI 261.542 V I1 311.071 La I1 408.672 B I 249.773 Be I1 313.107 Co I1 237.862 V I1 290.882 Co I 345.350 Lu I1 307.760 La I1 398.852 Nb I1 309.418 Ho I1 347.426 PHW*/ Pm 0.9 1.1 1.2 1.9 2.1 2.5 2.7 2.9 5.0 6.2 6.4 7.0 8.6 9.1 14.6 14.8 21.2 * From reference 12. EFW(3)/ Pm 3.1 3.2 3.2 3.6 3.7 3.9 4.0 4.2 5.8 6.9 7.1 7.6 9.1 9.6 14.9 15.1 21.4 EFW (20)/ Pm 20.0 20.0 20.0 20.1 20.1 20.2 20.2 20.2 20.6 20.9 21.0 21.2 21.8 22.0 24.8 24.9 29.1 EFW(20)/ fopt(3Y EFW(3) fopt(20) 6.4 6.4 6.3 6.3 6.2 6.2 5.7 5.1 5.5 4.8 5.2 4.3 5.0 4.1 4.8 3.9 3.5 2.3 3.0 1.9 3.0 1.8 2.8 1.7 2.4 1.4 2.3 1.4 1.7 1.1 1.6 1.1 1.4 1 .o PHW = 2.3 BW = 4.1 t il SBR = 3.7 PHW = 2.3 BW = 12.5 L SBR = 1.4 r -1 I 1 Nb It 309.4 PHW = 4-.- 14.8 I B W = i, 4.7 I I EFW = 15.5 I I I I SBR = 78 It 309.4 PHW = I ‘I Nb EFW = 20.3 \ ,LA t i \ SBR = 81 Fig.4. Examples illustrating the effect of an increase in spectral band width (BW) by a factor of about three on the SBRs of two lines with widely differing physical widths (PHW). They include the values of the effective line width (EFW) and SBR. All line widths are expressed in pm. The SBR of the Nb line has been measured with respect to a blank, because a back round measurement within the narrow spectral window would inclufe a substantial contribution from the line wing the spectral resolution used in the line-width measurements, so-called “physical high resolution” (PHR), but at a resolution level compatible with analytical measurements, for conven- ience designated “analytical high resolution’’ (AHR) .This is the resolution achieved at the optimised slit width, thus the resolution at which the best compromise is found between the effect of the spectral band width on the shot noise contribution to RSDB and its effect on the SBR, as has been discussed previously. 1328 The practical spectral band width correspond- ing to AHR ranges from 4 to 12 pm, depending on wavelength. 13.27 The availability of numerical values of SBR and background signal for the 350 lines implies the availability of numerical values of the detection limits, because for the ICP spec- trometer used, RSDB is a unique function of the background signal (XB)13: RSDB = 10.52 + 104/~B]1’2 .. . . (14) where RSDB is expressed in % and XB in counts [cf. , equation Detection limits determined thus for the 50-MHz ICP and AHR are listed in column 2 of Table 3, which is a sample page from the complete data set.20 These detection limits were compared with those reported by Wohlers25 for a 27-MHz ICP, measured at a 24-pm band width, and those of Winge and co-worker~23~~~ for another 27-MHz ICP, determined at a 17-pm band width. Columns 4 and 5 of Table 3 list these detection limits, while columns 6 and 7 show the ratios RWoH and RWIN with respect to the detection limits in column 2 for the 50-MHz ICP and AHR, thus for a band width between 4 and 12 pm.Obviously, both sets of detection limits for 27-MHz ICPs lag substantially behind those for the 50-MHz ICP. To what extent are the differences attributable to a difference between the sources? This question can be rigorously answered by application of the breakdown approach outlined in the previous section. In practice this involves the calculation of Fsou,c- from equation ( 6 ) , explicitly: (411. . . . . . R( 1/11) Fsource = r noiseropt where for R(I/II) the experimental values of RWoH or RWrN (Table 3) must be substituted and the values of Fnoisc and Fopt are found from equations (6) and (7), as follows. (a) For both the Wohlers25 and Winge and co-workers23J4 data sets RSDB has been assumed to be 1%; therefore, in equation (6), RSDB(1) = 1%.The value of RSDB for the 50-MHz ICP, thus RSDB(I1) is calculated from equation (14) from the measured values of xB. (b) Fopt is calculated from the known values of the band widths and the physical line widths, as discussed in the previous section. The results are listed as FWoH and FWrN (“F factors” or “source SBR ratios”) in columns 8 and 9 of Table 3. From the complete results20 it was concluded that the 50-MHz ICP gave a SBR advantage of a factor of 3-15 with respect to the ICP used by Winge and co-workers and a factor of 2-6 with respect to the ICP used by Wohlers, on average factors of 6.5 and 3, respectively, Interestingly, the scatter of the source SBR ratios ( F factors) among the elements was found to be reasonably small in the comparisons considered, namely a factor of 1.5 on the basis of one standard deviation for a log-normal distribution.This indicates that, aside from a constant factor, SBRs can be transferred from the one ICP to another, if all SBRs are measured under ICP compromise conditions, as happened in the considered situations. “Standard” Detection Limits for the Conventional ICP To facilitate comparisons between detection limits the results for the 50-MHz ICP and AHR were also converted into values that apply at medium resolution, defined by a spectral band width of 15 pm. These normalised values are listed in italics in column 3 of Table 3. They may be considered as standard values for the conventional argon ICP. The consideratioin of these detection limits may throw a different light on results that have been published in recent years for high-efficiency ICPs29 or ICPs generated with novel types of torches.30 Authors have gladly used the data of Winge and co-workers for comparison, either as they appeared in the publications from Iowa State University23.24 o r in the form518 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, SEPTEMBER 1987, VOL.2 Table 3. Sample page from the complete table in reference 20 listing detection limits (c,), ratios of detection limits (R) and ratios of source SBRs ( F factors), as detailed in the text. The detection limits are defined on the basis of 2 f i I e q u a t i o n (2)]. (Reproduced with permission from Boumans, P. W. J. M., and Vrakking, J. J. A. M., Spectrochim. Acta, Part B, 1987, 42, 553) Detection limit/ng ml-I Spectral line/ nm Ag I 328.068 Ag I 338.289 Ag I 243.779 Ag I1 224.641 A1 I 309.278 A1 I 396.152 A1 I 308.215 A1 I 237.312 As I 193.696 As I 197.i97 As I 228.812 As I 200.334 Au I 242.795 Au I 267.595 Au I1 208.209 Au I 197.745 Au I1 211.068 Au I1 200.081 B I 249.773 B I 249.678 B I 208.959 B I 208.893 Ba I1 455.403 Ba I1 233.527 Ba I1 230.424 Be II 313.042 Be f 234.861 Be I1 313.107 Be I 249.473 B:* I1 493.409 50 MHz 27 MHz Ratio F factor CL,AHR 0.39 1.1 30 55 1.5 1.6 2.4 7.8 11 21 22 49 0.92 1.1 9.9 11 21 33 0.54 1.1 2.1 3.9 0.040 0.088 0.37 0.56 0.037 0.064 0.065 1.9 ~L,15pm 1.1 2.0 47 63 2.8 3.0 4.3 6.4 7.2 I1 18 29 1.3 1.7 5.9 6.5 12 19 0.52 1.00 1.3 2.4 0.090 0.14 0.38 0.51 0.050 0.065 0.080 1.7 CL,WOH 4.0 7.6 230 - 9.9 11 18 33 65 74 59 - 8.5 14 24 27 40 - 3.1 6.2 5.7 0.28 0.57 3.7 5.1 0.28 0.28 0.57 11 11 CL,WIN 6.6 12 110 120 22 27 42 28 50 73 78 110 16 30 40 36 59 88 4.5 5.4 9.4 1.2 2.1 3.8 3.8 0.25 0.29 0.69 3.5 11 RWOH 10 6.7 7.6 6.8 7.3 7.6 4.2 5.7 3.6 2.7 9.2 2.4 2.4 1.9 5.8 5.9 2.8 2.9 7.1 6.4 9.1 7.6 4.4 8.7 5.7 - - 12 - 10 RWIN 17 11 3.8 2.2 15 18 17 3.6 4.4 3.5 3.5 2.3 17 27 4.0 3.2 2.9 2.7 8.4 5.1 4.6 2.9 31 3 10 6.7 6.8 4.6 1.8 11 FWOH 2.4 2.3 3.0 2.3 2.4 2.8 3.2 5.7 4.2 2.0 4.0 5.0 2.5 2.6 2.1 3.9 4.0 2.8 3.1 2.0 2.6 6.1 6.2 3.7 2.8 4.6 4.2 - - - FWIN 5.6 5.3 2.1 1.7 7.0 8.1 8.9 3.9 6.2 5.8 3.8 3.5 11 16 5.9 4.9 4.4 4.2 7.8 4.8 6.5 4.3 12 14 8.9 6.6 4.6 4.0 7.8 1.9 stated in Boumans' line coincidence tables31 or a related publication.32 The latter detectim limits were directly based on the measurements of Winge and co-workers and therefore do not differ essentially from their results. Doubtless, it is encouraging to use rather poor data as standards of compari- son, but it may put innovations into a biased perspective.We have shown on various occasions that a conventional argon ICP can yield detection limits that are much better than those of Winge and co-workers and not only in combination with high-resolution apparatus. 13&%,28,33 The fact that we have used such apparatus in the past few years may have confused the situation to some extent, but in reality, the profit of high resolution (HR) compared with medium resolution (MR) is less than a factor of 3 for detection limits in pure aqueous solution.1-13,27,28 The detection limits at HR may be even worse than at MR, as c>n be seen in Table 3 for lines of low wavelengths.This is due to increased shot noise.lJ3.28 The publication of the extensive set of new data20 aims at putting this point into the correct perspective, also in comparisons of detection limits obtained with either disper- sive spectrometry or FTS.3"36 Here, too, the use of rather poor data as being representative for dispersive spectrometry and a conventional ICP may easily bias the appreciation. Alternatively, the availability of data on line widths12 and the description of how these can be used to account for the contribution of the spectrometer to the detection limits20 may be helpful towards interpreting experimental values of detec- tion limits more rigorously. As a further aid, the Appendix in this paper gives a simple algorithm for the band width conversion of SBRs (and detection limits). Generally, a rigorous interpretation of detection limits requires the availability of more information than the mere values of these detection limits. Either the SBRs of the lines or the RSDs of the background signals should also be known, in addition to an accurate value of the practical spectral band width, not just the dispersion and the slit width, or the theoretical spectral band width (= spectral slit).Aberrations or trivial and unnoticed optical misalignments tend to make an important contribution to the practical spectral band width. In fact, the latter can be easily determined with the aid of relatively narrow lines emitted from the ICP.Therefore, one measures the effective line width and calculates the band width according to BW=d(EFW2-PHW2) . . . . (16) The use of narrow lines minimises the error, but generally it will not be necessary to use a hollow-cathode lamp. Lines emitted from an ICP by elements such as Ni, Mo or W are narrow enough12 to be used for reliable band width measure- ments. In addition, these elements provide lines over an extended wavelength range so that the variation of the band width with wavelength, common to kchelle spectrometers, but also inherent in MR to HR grating instruments,6J416 can be conveniently covered.519 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, SEPTEMBER 1987, VOL. 2 Detection Limit Including Selectivity as a Criterion for Line Selection Quantification of the Effect of Line Overlap on the Detection Limit Detection limits for pure water are indispensable figures of merit, but do not necessarily apply to real-sample analysis.It is well known that the concomitants of a sample can drastically worsen the detection limits, in particular, as a result of line overlap. This section concerns the quantification of this effect. 37 Fig. 5 serves as the starting-point and shows two spectral scans. The lower scan represents the profile of an interfering line as obtained on a blank solution of the interferent. The upper scan represents the profile for the solution of the interferent spiked with analyte. At the wavelength h, of the analysis line we have three signals: the background signal xg, the net interfering signal x I and the analyte signal xA.(i) The background signal XB is the signal that can be rationally considered as the smooth background level in that wavelength region so that its magnitude can be unambiguously determined by a measure- ment outside the structure. (ii) The net interfering signal XI is that part of the signal contributed by the interferent at the wavelength h, of the analysis line which cannot be covered by the measurement of xB, but must be derived from an additional measurement. The latter may be either a measure- ment at h, during the aspiration of a blank solution or a measurement at a wavelength of another line of the interferent during the aspiration of the sample (cf. Fig. 2 in reference 37). In the former instance it must be ensured that the blank solution yields a spectrum identical with that of the sample blank; in the latter, the ratio of the two signals of the interferent must be constant.The equations for the detection limit, the SBR and the RSD of the background signal can now be written as (SBR)BI = xA/xBI . . . . . . (18) RSDBl = (a: + + ([j/xBI + Y/X;,)’” . . (19) These equations differ from equations (2)-(4) in two respects: (a) XB has been replaced by xBI (B1 = blank); and (b) the expression for the RSD of the background signal [equation (19)] contains an additional flicker noise term a:, associated with the determination of the net interfering signal. It is this term which introduces an essential difference between the situation of smooth background and the case of line overlap. To appreciate the effect of a: on the detection limit one must distinguish between the two approaches for measuring XI xB I I A - Wavelength Fig.5 . Profile of an interfering line (blank) and resultant profile of interfering and analysis lines (spiked) to illustrate the meaning of the quantities: net analyte signal xA, net interfering signal x I and background signal xB.3’ The arrow marks the peak wavelength of the analysisline. (ReproducedwithpermissionfromBoumans, P. W. J . M., and Vrakking, J. J. A. M.. Spectrochim. Acta, Part B , 1987,42,819) referred to above: the direct determination using a blank solution and the indirect determination using a reference signal of the interferent. (i) In the direct determination the blank is measured with a fixed wavelength setting of the spectrometer (“static measure- ment”).Under these idealised conditions, the additional flicker noise term will be small; moreover, its contribution to RSDBI will be partly compensated for by a reduction of the shot noise term. Therefore, a series of successive intensity measurements based on 10-s integrations, for instance, will generally yield a value of RSDBI that is only slightly higher than the value of RSDB found for the pure solvent.13 Hence the chief effect of the line interference on the detection limit would be a marginal effect from a decreased SBR. We have called this the “Conventional detection limit.” I Although it can be measured by an unambiguous procedure, it is in general an unrealistic quantity, because its determination underlies the postulate that the blank of all samples is exactly the same as the separately measured blank.This situation will hardly ever be found in real sample analysis. Commonly the concentrations of the concomitants vary for each sample, which necessitates a separate indirect determination of the interfering signal in each sample. (ii) The indirect determination of the net interfering signal requires measurements at various wavelength positions for each sample separately. This “dynamic measurement” will yield a substantially higher value of at than the “static measurement,” since ( ~ f will now reflect the effects of variations in sample composition, variations in the excitation conditions between samples and instabilities in the wavelength setting such as those occurring in slew-scan spectrometers.It is the latter value of at which dictates RSDBl and thus the “true detection limit” in the case of line overlap. Unfortunately, this realistic value of a: cannot be measured simply in daily practice and actually can only be approximately determined via model experiments. Using the latter approach, Boumans and Vrakking27-3s arrived at an approximation, which makes it reasonably possible to take the effect of a: into account and at the same time can be used in common practice. The starting-point of this approach is the limit of determination, cD, which, as usual, was defined as the concentration which can be determined with an RSD of 10%. The concentration CD is a function of the errors in the background and the net interfering signals.Boumans and Vrakking made it plausible that with line overlap cD can be approximated by . . . where SA is the sensitivity of the analyte signal. The first term in equation (20) is twice the analyte concentration equivalent of xI. The value of 2 for the coefficient is not based on a rigorous treatment, but is merely an estimate,27,28 which, however, links up well with practical experience.lJ9 The first term in equation (20) links the limit of determination with the selectivity, the quotient SA/xI being proportional to the “line selectivity,” that is, the ratio of the net line signal and the net interfering signal, or generally the sum of the interfering signals.27 The selectivity term in equation (20) replaces a complex expression which, in a rigorous treatment, would be necessary to account for the various error sources in the case of line overlap.The second term in equation (20) is five times the conventional detection limit, that is, the detection limit defined by equations (2)-(4) but conventionally determined by static measurements on a blank solution. This second term has been introduced for reasons of continuity: when the net interfering signal decreases to zero, the equation should yield the classical relationship between the limit of determination and the limit of detection.’.2”40.4’ As it is customary to look primarily at limits of detection rather than limits of determination, Boumans and Vrakking37c3n JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, SEPTEMBER 1987, VOL. 2 finally defined the “true limit of detection” with line overlap to be where cD is given by equation (20).Substitution then yields In the event that more than one component contributes an interfering signal, xI should be understood as the sum of the various net interfering signals. To conclude, equation (22) takes into account both the selectivity and the detection limit that is classically reached when the selectivity becomes infinite. This approach clarified two points: (i) It brings a quantitative understanding of the effect of spectral resolution on analytical performance.l.2J7-38-42343 (ii) It provides a quantitative criterion for line selection for any sample composition and spectral resolution.37 The use of equation (22) for line selection will be illustrated in the next section for the determination of traces of indium in binary mixtures of tungsten and molybdenum.The prime target of that discussion is to show how the problem of line selection can be quantified, what data are required and how these data can be derived from spectral scans. The complete treatment of the problem in the original paper37 also covers the effects of the spectral resolution, the solid concentration in the solution and the ICP operating conditions. Line Selection for the Determination of Traces of In in a Binary Mixture of W and Mo: Classical Table The analyst is asked to choose the best In line(s) for the determination of In traces in binary mixtures of W and Mo, the composition of which may vary from pure W to pure Mo. Consultation of lists of prominent ICP lines2”-25Jl will primarily lead to the four In lines listed in Table 4, which includes the conventional detection limits and the spectral band widths for HR and MR.Unfortunately, none of the present wavelength tables contains sufficient information on the ICP spectra of W and Mo to select a line which is “free from interference.” If such data were available in the form of a classical table, they would look as shown in Table 5 , which lists the wavelengths and ICP sensitivities of all relevant lines as derived from HR spectral scans for pure In, W and Mo solutions. The reader may attempt, as we did, to select optimum analysis lines on the basis of the data in Tables 4 and 5. Doubtless the reader will then experience the uneasy feeling that, in spite of the abundance of relevant data, he or she cannot make a rigorous decision. We state the problem in this form because it clearly reveals that with the classical approach there is no firm basis for line selection even if the available data are entirely complete and relevant.It is only by rigorous calculations that line selection can be lifted from the qualitative “yesho domain” to the domain of unambiguous quantification, which not only tells whether an analysis “goes” or “does not go,” but also indicates accurately what is analytically still possible under the circumstances given. Actually, to quantify the problem, data different from those in Tables 4 and 5 should be derived from the spectral scans. * A critical examination of the approach made by Boumans and Vrakkingz’ indicates that they could also have adopted a different starting-point to arrive at essentially the same results.They chose the limit of determination (10% error) as the prime figure of merit to assess the adverse effects of line overlap on analytical performance under realistic, dynamic measurement conditions. Alternatively, they could have adopted the “true limit of detection” ( c ~ , ~ ~ ~ ~ ) as a starting-point, defining it as the concentration which can be determined with an RSD of 50”/0.~O Afterwards, it appears that this starting-point was avoided for “psychological reasons.” Whatever it was, the results would have been entirely the same. Table 4. Detection limits of In lines in pure aqueous solutions for the “soft” conditions specified37 at high (HR) and medium resolution (MR), as defined by the spectral band widths stated.The detection limit is defined by equation (2) and refers to a 10-s integration time. (Reproduced with permission from Boumans, P. W. J . M., and Vrakking, J . J . A. M., Spectrochim. Actu, Part B , 1987, 42, 819) Spectral line/ nm In I1 230.559 In I1 230.606 In I1 230.612 In I 325.609 In I 303.936 In I 451.131 Detection limit/ Spectral ng ml- band widthlpm HR MR HR MR 21 24 13 3.6 10.9 29 4 7 5.0 14.9 8 10 5.1 14.7 11 30 7.3 22.1 - - - - - - Table 5. Classical table of data relevant to the selection of the best In line for the determination of traces of In in a binary mixture of W and Mo.” The table lists the peak wavelengths h (in nm) of the lines, the distance Ah (in pm) of potentially interfering lines to the respective In lines and the peak sensitivities S (in counts per 10 s per kg ml-1) for “soft” ICP conditions and HR.(Reproduced with permission from Boumans, P. W. J . M., andvrakking, J. J. A. M., Spectrochim. Acta, Part B , 1987,42, 819) h Ah S MO 230.585 -21 130 W 230.590 -16 8.4 W 230.596 -10 4.5 MO 230.596 -10 4.7 MO 230.600 - 6 3 .O In 230.606* 0 2400 W 230.614 + 8 1.1 W 230.626 +20 1.1 MO 303.906 -30 640 W 303.933 - 5 500 In 303.936* 0 37000 W 303.958 +22 390 Ar 451.073 -58 - In 451.131* 0 37000 W 451.122 - 9 45 * Analysis lines. h Ah MO 230.585 -27 W 230.590 -22 W 230.596 -16 MO 230.596 -16 MO 230.600 -12 In 230.612* 0 W 230.614 + 2 W 230.626 +14 In 325.609* 0 Mo 325.622 +13 W 325.623 +14 W 325.596 -13 S 130 8.4 4.5 4.7 3.0 1.1 1.1 2000 210 90 000 2 600 360 Line Selection for the Determination of Traces of In in a Binary Mixture of W and Mo: Equation (22) As an example, Fig.6 shows spectral scans of In I1 230.606 nm; a detailed explanation of the figure is given in the legend. At HR the In line is a resolved triplet: 230.598,230.606 and 230.612 nm. 12,13,18 In 230.606 nm does not experience line interference from W, while In 230.612 nm is free from line interference from Mo. At MR, both W and Mo contribute net line signals at the peak (230.606 nm) of the unresolved In line. From the printouts of the scans it followed that W and Mo each produce, at both HR and MR, background enhancements due to line wings.44 We now consider how the data implied in these scans can be quantitatively used, where “quantitative use” means the calculation of the true detection limit as a function of the sample composition according to equation (22): The net interfering signal x I can be written as XI = c ~ S I , ~ + CZS~,~ .. . . . . (23) where c is concentration and SI the sensitivity of the net interfering signal with subscript 1 for W and 2 for Mo. In agreement with experimental evidence,44 all signals are assumed to be additive.JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, SEPTEMBER 1987, VOL. 2 521 Wavelength Fig. 6. Spectral scans in the vicinity of In 230.606 nm. Spectral window 0.050 nm.37 Each frame contains two scans: the lower ones are for the pure matrix, the upper ones for the matrix spiked with the analyte. (a)-(d) The W matrix and (e)-(h) the Mo matrix. The matrix concentration is 1 mg ml-I.Analyte concentration c is expressed in pg ml- l . Where appropriate, an arrow marks the wavelength position of the analysis line. The scans are for "soft" excitation conditions. (Reproduced with permission from Bournans, P. W. J. M., and Vrakking, J. J. A. M., Spectrochim. Acta, Part B , 1987, 42, 819) The measurement of the three sensitivities SA, SI,l and SI,2 for the pure solutions of the analyte and the two interferents permits the calculation of the first term in equation (22) for any composition of the binary mixture.* For the calculation of the second term we proceed as follows. We write for the conventional detection limit CL,BI = ~ ~ ~ ( O . O ~ R S D B ~ ) X B I / S A . . . . (24) This is essentially equation (17) in which C~I(SBR)BI has been replaced by the equivalent expression xBI/SA.The total background signal xB1 is made up of three contributions : xBI = x B + XW + X I . . . . . . (25) where xB is the background signal for the pure solvent and xw the background contributed by the two interferents as continua and/or line wings. This contribution can be built up additively from the separate contributions44: where Sw.l and SW,* are the corresponding sensitivities, which can also be determined for pure solutions of the interferents. Thus xB1 in equation (25) can be calculated for any sample composition. There remains the last quantity, RSDBI. Generally it may be calculated using an equation of the form of equation (4) and not equation (19) because the first term in equation (22) accounts for the additional noise from the interfering signal.Therefore, for the system used in this work: RSDBl = (0.52 + lO4/xBI)li2 . . . . (27) in analogy with equation (14). In summary, the experimental measurement of the solvent background xB, the analyte sensitivity SA and the sensitivities SI.], S1.2, Sw,l and Sw.z of the interfering signals provides all the essential data needed for the calculation of the true limit of detection for any composition of the binary mixture. These data can be derived from spectral scans. Table 6 lists such data for the four In lines as derived from MR spectral scans. The sensitivities and the background signal XW = c ~ S W , ~ + C ~ S W , ~ . . . . (26) * The original paper37 also covers possible contributions from the pure solvent plasma (Ar lines and band components). Table 6.Experimental values of the sensitivities (S) of the analytc signals, net interfering line signals, and wing signals and thc background contribution ( x B ) from the pure solvent p l a ~ r n a . 3 ~ The sensitivities and background signal refer to MR and are stated in counts per 10 s. (Reproduced with permission from Bournans, P. W. J. M., and Vrakking, J . J. A . M., Spectrochim. Acta, Part B , 1987,42, 819) Wavelengt hlnm Parameter 230.606 325.609 303.936 451.132 SA(pg-lml) . . . . 16300 440000 208000 212000 SI,l (pg-l ml) . . . . 2.42 10.2 1785 125 S,.,(pg-:ml) . . 17.0 120.5 89.1 5.92 124.0 14.0 0.00 Sw,2(pg- ml) . . 12.0 x B . . . . . . . . 4820 186000 130000 182000 Sl,2(pg-1ml) . . . . 20.3 1023 58.3 0.00 are stated in counts per 10 s, with the understanding that the eventual analysis is based on 10-s integrations at fixed wavelength positions rather than on scans.Clearly, the integration time is relevant only in the calculation of RSDBI. Fig. 7 shows the detection limit of In for the four In lines as a function of the Mo concentration, whereby the total concen- tration of W and Mo has been assumed to be 1 mg ml-1. The curves have been calculated with the aid of equations(22-27) using the data in Table 6. Evidently, up to about 0.2 mg ml-1 Mo, In 325.6 nm is the best line, while for higher Mo concentrations In 451.1 nm is optimum. Note the position of the curve for In 303.9 nm high in the diagram. This line cannot compete because it suffers a severe overlap from a W line, as is illustrated by the scans in Fig. 8.The effect of this coincidence is unambiguously revealed and quantitatively expressed in a diagram of the type shown in Fig. 7. As a further illustration of the approach we consider what happens if the spectral resolution is changed from MR to HR. Qualitatively the effect of this change is seen in the spectral scans (Figs. 6 and 8 and the corresponding figures for In 325.6 and In 450.1 nm in reference 37). Quantitatively the effect is expressed in terms of a diagram similar to that of Fig. 7. Such a diagram is included and discussed in reference 37. Here we confine discussion to the final result, shown in Fig. 9, which contrasts the curves for the best lines (In 325.6 and In 450.1 nm) at MR and HR. The detection limit is expressed here in522 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, SEPTEMBER 1987, VOL.2 MR 1000 r I Y lob 1 0 0.2 0.4 0.6 0.8 1.0 Molybdenum concentrationimg ml- 10 1 I I I I 0 0.2 0.4 0.6 0.8 1.0 Fig. 7. Dependence of detection limit of indium (ng ml-1) on the composition of a binary mixture of W and Mo with a total metal concentration of 1 mg ml-l for four lines: A, In I , 303.9; B, In I, 325.6; C, In 11, 230.6; and D, In I, 451.1 nm. The detection results refer to medium resolution and “soft” excitation conditions.37 (Reproduced with permission from Boumans, P. W. J. M., and Vrakking, J. J. A. M., Spectrochim. Acta, Part B , 1987,42, 819) pg g-1 with respect to the dissolved solid for a 1 mg ml-1. metal concentration in solution. * Fig. 9 demonstrates (i) a drastic effect of the resolution on the optimum line choice and (ii) an appreciable difference in detection power associated with the difference in spectral resolution.In summary, the application of the true detection limit [equation (22)] as the criterion for line selection implies the quantitative and realistic evaluation of a basic analytical figure of merit for real samples. Therefore the approach permits the calculation, of what is analytically possible with each analysis line of interest, given the composition of the sample and the spectral band width of the spectroscopic apparatus. The result of this calculation then provides a quantitative basis for decisions on line selection. Thus the approach transfers line selection from the nebulous domain of “interferencelno interference” to a domain in which the seriousness of an interference is quantified on a continuous scale and is numerically expressed in terms of an analytical figure of merit with a well defined meaning. Suitable software permits application of the criterion to multi-component samples of whatever degree of complexity.Although the present discussion covers only the application of the criterion to a priori line selection, it is implicit that the same criterion can be used in a posteriori line selection, and thus in “multiple line analysis,” as used in d.c.-arc methods for general survey analysisl.45 and proposed for It appears reasonable to determine the concentration of an element as the weighted mean of the concentrations found with the various analysis lines, where the weighting factor for each line could be rationally defined as the ratio of the pertinent concentration and the “true detection limit” for that line.As this “true detection limit” depends primarily on the selectivity, such an approach automatically implies the rejection of lines with poor selectivity, i . e . , lines suffering severe line overlap from concomitants. ~~ * The reasons for changing from ng ml-1 to pg g-1 are given in the original paper.37 The change is irrelevant in this context because for a 1 mg ml-1 metal concentration, the two detection limits are numerically equal. The approach should also be of interest in comparisons of dispersive and Fourier transform spectrometry. It has been shown34,48 that the detection limit in FTS worsens in the presence of concomitants as a result of both an increase in the flicker noise due to strong lines of concomitants and a decrease in analyte sensitivity.In dispersive spectrometry, concomitants may also produce an increase in flicker noise due to line overlap. However, this point is hardly ever quantified. The proposed approach may be helpful towards treating the problem more rigorously and easing comparisons between the two types of spectroscopy. The Interference Library of the 1990s The criterion for line selection [equation (22)] discussed in the preceding section also indicates how the ideal interference library should look: (i) it must contain the sensitivities of interfering lines and line wings at the wavelength positions of the analysis lines, and (ii) be in such a form that the data can be universally applied to any spectroscopic equipment.Data such as those presented in Table 6 fulfil condition (i) but lack universality, in that they carry the seal of the spectral instrumental function of the apparatus with which they were collected. Obviously, interference data in the form of a classical table, as shown in Table 5 , are still further remote from the ideal. The principle of the approach used in Boumans’ line coincidence tables31 meets the conditions for an ideal tabula- tion, as is discussed in references 2 and 37, but the result was not ideal because the line coincidence model and the data base were not entirely adequate.1.2 Recent developments indicate that the ideal interference library is coming within reach, that is, an interference library in the software domain which contains information about analysis and interfering lines in the form of physically resolved spectral data so that the computer can reconstruct actual spectra for any spectroscopic apparatus by convoluting physically resolved spectra with the instrumental func- It was concluded from the measurement of physical line widths in ICP spectral2 that the physical profiles of all simple lines can be approximated by Doppler profiles, if necessary, extended to Voigt profiles with a d 0.5.This approximation can also be used for complex structures if the wavelengths and relative intensities of the separate components are known. Both simple lines and the separate components of HFS composites can be convoluted with the instrumental function to yield effective line profiles.The total effective profile of complex structures can then be found by superposition of the effective profiles of the components. In this way, the basic data primarily required for the characterisation of ICP spectra can be reduced to peak wavelengths, peak sensitivities, atomic masses and a values. It must be investigated whether a single a value for all lines or separate values for the individual lines are required. Additional data on line wings and recombination continua may also be needed. Either of two approaches can be adopted for collecting the necessary data: (i) registration of the complete ICP spectra of the elements, or (ii) registration of spectral data in narrow spectral windows centred about prominent ICP lines.Considering the satisfactory results obtained hitherto with the 1.5-m kchelle monochromator referred to earlier, the present authors are investigating the application of this instrument to approach (ii). 193) For approach (i), high- resolution FTS appears the most promising.sl The future will have to show to what extent these efforts will eventually lead to interference libraries that can be universally applied in ICP-AES. In view of the large amount of work involved in the acquisition, testing and handling of the data, it seems realistic tion.l,l2,43.49JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, SEPTEMBER 1987, VOL. 2 523 Wavelength Fig. 8. with permission from Boumans, P. W. J . M., and Vrakking, J .J. A. M., Spectrochim. Acra, Part B , 1987, 42, 819) Spectral scans in the vicinity of In 303.936 nm.37 Spectral window 0.067 nm. See legend to Fig. 6 for further explanations. (Reproduced r L l E u, 5 . c .- E .- - c 0 a a .- 4- c n I 1 0 20 40 60 80 100 M o l y bde n u m concentration , ‘10 Fig. 9. Dependence of detection limit of indium (pg g-1) with the best available lines, (A, In 1325.6 and B, In 1451.1 nm) on the composition of a binary alloy of W and Mo at high (HR) and medium (MR) res0lution.3~ Total metal concentration in the solution is 1 mg ml-I. (Reproduced with permission from Boumans, P. W. J. M . , and Vrakking, J. J. A. M., Spectrochim. Acta, Part B, 1987,42,819) to refer to these data bases as “the interference library of the 1990s. l7 Conclusions The conventional detection limit for smooth background can be formulated as a function of (i) the relative standard deviation (RSDB) of the background signal, (ii) the signal to background ratio (SBR) in the source and (iii) an optical factor v) by which the source SBR is modified by the spectroscopic apparatus.Factor f is a function of the physical width of the line and the practical spectral band width of the spectrometer. The availability of numerical values for the physical widths of some 350 prominent ICP lines permits the calculation off for these lines and therefore allows a rigorous comparison of detection limits measured with different spectrometers in different ICPs if, in addition, the values of RSDB and the practical spectral band width are known.Owing to possible contributions from aberrations andlor optical misalignment, the practical spectral band width does not necessarily equal the resultant spectral slit as calculated from the slit width and the reciprocal linear dispersion. However, the practical spectral band width can be easily measured with the aid of narrow lines emitted from the ICP, e . g . , Ni, Mo or W lines. Reliable and conservative values of RSDB can be conve- niently determined if RSDB is recognised as a function of flicker noise, shot noise and detector noise. It has been definitely shown that the source SBRs for the 27-MHz argon ICP used by Winge and co-~orkers*3,24 are a factor of 3-15 poorer than those of a conventional 50-MHz argon ICP. The use of the values produced by Winge and co- workers as standards of comparison thus tends to introduce optimistic bias in assessments of the detection capabilities of novel ICP sources. Detection limits recently determined for a 50-MHz ICP and normalised to a spectral band width of 15 pm are therefore recommended as standard values for the conventional argon ICP.The effect of line overlap on the detection power can be quantitatively expressed in terms of the “true detection limit,” defined as the sum of the conventional detection limit for smooth background and a selectivity term. The “true detec- tion limit” is dictated by the sample composition and the spectral resolution; therefore, it is an adequate and effective criterion for line selection. Prominent ICP lines, including the separate components of hyperfine structure composites but excluding Ar and H lines, show chiefly Doppler broadening with a small contribution from Lorentz broadening (a < 0.5).If this feature applies generally, then the data required for an efficient storage of physically resolved ICP spectra can be limited to peak wavelengths, peak sensitivities, atomic masses and u values. Recombination continua must be separately accounted for, while line wings may be described by a single function.44 Modern spectroscopic instruments, such as high-resolution Cchelle spectrometers or high-resolution Fourier transform spectrometers, permit the measurement of physically resolved spectra. The further application of such instruments for data acquisition brings the ideal universal interference library within reach, that is, a compilation of physically resolved spectra, which can be convoluted with the spectral instrumen- tal function of whatever spectroscopic apparatus.A new data base of this type is the only adequate way to solve an old basic problem of atomic emission spectrometry: dealing with spectral interferences.524 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, SEPTEMBER 1987, VOL. 2 Appendix Band Width Conversion of SBRs in Practice For converting SBR values obtained at a particular spectral band width BW1 to another band width BW2 the following practical procedure is recommended. One calculates, for each of the two band widths, the factorf by which the SBR increases when the band width is reduced to the value where it equals the physical line width. If we denote the two SBRs as SBRl and SBR2 and the factors asfl andf2, respectively, then F = f l / f 2 .. . . . . (Al) is the factor by which SBRl has to be multiplied to obtain SBR2: S B R 2 = F x SBRl . . . . (A2) As the same procedure is followed for the calculation offl and f2, we formulate this procedure in terms of the single factorf. (a) The effective line width is calculated from the known band width BW and the listed12 value of the physical line width PHW: EFW= dPHW2+BW2 . . . . (A3) (b) To formulate the algorithm we define the transition EFWL= P H W a . . . . . . (A4) points: and E F W H = P H W f i . . . . . . (A4) (c) The algorithm then takes the following form: If BW d PHW, then f = 1 . . . . . . . . (A5) and no further calculation is required.If BW d 2PHW, then f = ~EFWIEFWL or f = dEFW/(PHWfi) . . . . (A6) else f = (EFW/EFWH) dEFWWEFWL f = E F W / ( P H W a ) . . . . (A7) (d) Procedure: calculate EFW for situations 1 and 2 [equation (A3)]; apply the algorithm to calculate fr and f2 [equations (A5)-(A7)]; calculate F [equation (Al)]; and calculate SBR2 [equation (A2)]. (e) For example, for Mn I1 257.610 nm SBRl was measured to be 100 for BW1 = 5.0 pm. What is SBR 2 if BW2 = 15 pm? From reference 12 we find PHW = 3.7 pm. The procedure yields F = 0.46 and therefore SBR2 = 46. If the RSD of the background signal is the same for both situations, then the detection limit at 15 pm band width will be a factor of 100/46 = 2.17 higher than at 5 pm band width. For hypothetical lines with physical widths of 1, 2, 4 , 6 and 10 pm, one calculates F to be 0.34, 0.36, 0.49, 0.66 and 0.89, respectively.For other examples refer to Table 2 in reference 20. or References 1. Boumans, P. W. J. M., “Proceedings of 24th Colloquium Spectroscopicum Internationale, Garmisch-Partenkirchen, 1985”; Fresenius Z . Anal. Chem., 1986, 324, 397. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31, 32. Boumans, P. W. J. M., “Line Selection and Spectral Interfer- ences,” in Boumans, P. W. J. M., Editor, “Inductively Coupled Plasma Emission Spectroscopy. Part 1, Methodology, Instrumentation, and Performance ,” Wiley, New York, 1987, Chapter 7, pp. 358-465. Human, H. G. C . , and Scott, R. H., Spectrochim. Acta, Purt B , 1976, 31, 459.Kawaguchi, H., Oshio, Y., and Mizuike, A , , Spectrochim. Acta, Part B , 1982, 37, 809. Posener, D. W., Aust. J . Phys., 1959, 12, 184. Boumans, P. W. J . M., and Vrakking, J . J. A. M., Spectrochim. Acta, Part B , 1984, 39, 1239. Hasegawa, T., and Haraguchi, H., “Proceedings of 1984 Winter Conference on Plasma Spectrochemistry, San Diego”; Spectrochim. Acta, Part B , 1985, 40, 123. Batal, A . , and Mermet, J. M., Spectrochim. Acta, Part B , 1981, 36, 993. Broekaert, J. A. C., Leis, F., and Laqua, K., Spectrochim. Acta, Part B , 1979, 34, 73. Laqua, K., Hagenah, W.-D., and Waechter, H., Fresenius 2. Anal. Chem., 1967, 225, 142. Faires, L. M., Palmer, B. A . , and Brault, J. W., “Proceedings of 1984 Winter Conference on Plasma Spectrochemistry, San Diego”; Spectrochim.Acta, Part B , 1985,40, 135. Boumans, P. W, J. M., and Vrakking, J. J. A. M., Spectro- chim. Acta, Part B , 1986, 41, 1235. Boumans, P. W. J. M., and Vrakking, J. J . A. M., Spectro- chim. Acta, Part B , 1984, 39, 1261. McLaren, J. W., and Mermet, J. M., Spectrochim. Acta, Purt B , 1984, 39, 1307. Maessen, F. J. M. J., and Tielrooij, J . A . , Fresenius 2. Anal. Chem., 1986, 323, 490. Olesik, J. W., “Spectrometers,” in Boumans, P. W. J . M., Editor, “Inductively Coupled Plasma Emission Spectroscopy, Part 1, Methodology, Instrumentation, and Performance,” Wiley, New York, 1987, Chapter 8, pp. 466-335. Savitzky, A., and Golay, M. J . E . , Anal. Chem., 1964, 36, 1627. Harrison, G. R . , “MIT Wavelength Tables,” MIT Press, Cambridge, MA, 1969. Vrakking, J.J . A. M., and Boumans, P. W. J . M., paper presented at the 25th CSI, Toronto, 1987; to be published in Spectrochim. Acta, Part B. Boumans, P. W. J. M., and Vrakking, J . J . A. M., Spectro- chim. Acta, Part B , 1987, 42, 553. Laqua, K., “Emissionsspektroskopie,” in “Ullmanns Encyklo- paedie der technischen Chemie,” Fourth Edition, Volume 5 , Verlag Chemie, Weinheim, 1980, pp. 441-500. Boumans, P. W. J. M., and Vrakking, J. J. A. M., Spectro- chim. Acta, Part B , 1985, 40, 1437. Winge, R . K., Peterson, V. J., and Fassel, V. A., Appl. Spectrosc., 1979, 33, 206. Winge, R. K., Fassel, V. A., Peterson, V. J., and Floyd, M. A., “Inductively Coupled Plasma Atomic Emission Spectroscopy. An Atlas of Spectral Information,” Elsevier, Amsterdam, 1984. Wohlers, C. C . , “ICP-AES Wavelength Table,” ICP Inf. Newsl., 1985, 10, 601. Boumans, P. W. J . M., McKenna, R. J., and Bosveld, M., Spectrochim. Acta, Part B , 1981, 36, 1031. Boumans, P. W. J. M., and Vrakking, J . J. A. M., Spectrochim. Acta, Part B , 1985, 40, 1085. Boumans, P. W. J. M., “Basic Concepts and Characteristics of ICP-AES,” in Boumans, P. W. J . M., Editor, “Inductively Coupled Plasma Emission Spectroscopy, Part 1, Methodology. Instrumentation, and Performance,” Wiley, New York, 1987, Chapter 4, pp. 100-257. Boumans, P. W. J. M., and Hieftje, G. M., “Torches for Inductively Coupled Plasmas,” in Boumans, P. W. J . M., Editor, “Inductively Coupled Plasma Emission Spectroscopy,” Part 1, Methodology, Instrumentation, and Performance,” Wiley, New York, 1987, Chapter 4, pp. 258-295. Davies, J., and Snook, R. D., J . Anal. A t . Spectrom.. 1986, 1 , 195. Boumans, P. W. J. M., “Line Coincidence Tables for Induc- tively Coupled Plasma Atomic Emission Spectrometry,” Second Edition, Pergamon Press, Oxford, 1984. Boumans, P. W. J. M., Spectrochim. Acta, Part B , 1981, 36, 169.JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, SEPTEMBER 1987, VOL. 2 525 33. 34. 35. 36. 37. 38. 39. 40, 41. 42. 43. Boumans, P. W. J. M., Spectrochim. Acta, Part B , 1983, 38, 747. Faires, L. M., Spectrochim. Acta, Part B , 1985, 40, 1473. Ng, R. C. L., and Horlick, G . , Appl. Spectrosc., 1985,39,834. Stublcy. E. A . , and Horlick, G . , Appl. Spectrosc., 1985, 39, 811. Boumans, P. W. J. M., and Vrakking, J. J. A. M., Spectro- chim. Acta, Part B , 1987,42, 819. Boumans, P. W. J. M., and Vrakking, J. J . A. M., Spectrochim. Acta, Part B , 1985, 40, 1107. Botto, R . I., Spectrochim. Acta, Part B, 1983, 38, 129. Boumans, P. W. J. M.. Fresenius 2. Anal. Chem., 1979, 299, 337. Boumans, P. W. J . M., Spectrochim. Acta, Part B , 1978, 33, 625. Boumans, P. W. J. M., and Vrakking, J. J. A. M., Spectrosc. Spectral Anal. (in Chinese), 1986, 6 , No. 5 , 26; No. 6, 28. Boumans, P. W. J. M., "Analytiktreffen 1986, Neubranden- burg. GDR," 1987, in the press. 44. 45. 46. 47. 48. 49. 50. 51. Boumans, P. W. J. M., and Vrakking, J. J. A. M.. Spectro- chim. Acta, Part B , 1984, 39, 1291. Witmer, W. A . , Jansen. J . A. J . , van Gool, G . H.. and Brouwer, G., Philips Tech. Rev., 1974, 34, 322. Faires, L. M., Anal. Chem., 1986, 58, 1023A. Thorne, A., Anal. Proc., 1985, 22, 63. Stubley, E. A . , and Horlick, G . , Appl. Spectrosc., 1985, 39. 805. Faires, L. M., invited lecture presented at the 1986 Winter Conference on Plasma Spectrochemistry, Kona, Hawaii. Boumans, P. W. J . M., and Vrakking, J . J . A. M., paper presented at the 25th CSI, Toronto 1987: to be published in Spectrochim. Acta, Part B. Faires, L. M., ICP Znf. News/., 1984, 10, 449. Paper 57121 Received March 3rd, I987

ISSN:0267-9477    

DOI:10.1039/JA9870200513

出版商:RSC    

年代:1987

数据来源: RSC

摘要:

527 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, SEPTEMBER 1987, VOL. 2 Atomisation Efficiency and Over-all Performance of Electrothermal Atomisers in Atomic Absorption, Furnace Atomisation Non-thermal Excitation and Laser-excited Atomic Fluorescence Spectrometry* Plenary Lecture Heinz Falk and Johannes Tilch Zentralinstitut fur Optik und Spektroskopie der Akademie der Wissenschaften der DDR, Berlin, GDR The atomisation efficiency of electrothermal atomisers using cup or tube geometries is estimated by model calculations. Tube atomisers operated a t atmospheric pressure have an efficiency of more than 50% whereas that of the low-pressure FANES system and the cup atomiser is typically less than 5%. The results of the calculations are supported by atomic absorption measurements. Using these data, the minimum number of analyte atoms detectable within the atomisers using LAFS, FANES or GFAAS has been calculated t o be in the region of 104-109.The concentration of the matrix species to be expected in the gas phase of the ETAs is estimated. Gas-phase interferences should be less using the cup and low-pressure atomisers compared with the usual tube furnace because of the longer residence time in the latter. Keywords: Electrothermal atomisation; residence time of atoms; laser-excited atomic fluorescence spectrometry; graphite furnace atomic absorption spectrometry; furnace atomisation non-thermal excitation spectrometry The over-all performance of analytical atomic spectroscopic methods such as graphite furnace atomic absorption spec- trometry (GFAAS), furnace atomisation non-thermal excita- tion spectrometry (FANES) and laser-excited atomic fluores- cence spectrometry (LAFS) is determined by both the efficiency of the atomisation and the performance of the measurement procedure used.In this paper the analytical results achieved by GFAAS, LAFS and FANES are dis- cussed, particularly from the point of view of the atomisation efficiency of the different atomiser systems used. Considering the maximum analyte atom density during the transient signal that is delivered by electrothermal atomisers (ETA), the atomisation efficiency at this moment may be defined as followsl.2: where Nap and N,, are the number of analyte atoms within the observation volume at the peak maximum and that which is introduced, respectively;fsl is the supply-loss function depend- ing on time t , volatilisation time tl and residence time ~ 2 ; -qt is the fraction of sample transported into the atomiser itself, which is unity for most ETAs; (3(7‘) is the fraction of the analyte in the atomic form and is assumed to be unity for the further considerations, i.e. , molecule formation and ionisation will be ignored. (SI units are used throughout.) If, for the sake of simplicity, a constant volatilisation rate is assumed, the atomiser efficiency can be expressed as2 fsl = rlap = - exp (-T,/t2)1T2/T1 f * (2) The residence time t2 for the tube atomiser can be calculated using a relatively simple model formulated by L’vov et al. leading to satisfactory agreement with experimental AAS results.At gas-stop conditions L’vov et al.3 found t 2 = P/(8D) . . . . . . . . (3) where I is the tube length and D the diffusion coefficient (m2 s-I); D is inversely proportional to the gas pressure and is temperature dependent.3 The atomic number density within a tube atomiser then becomes (1 - exp(- T ~ / - C ~ ) ] . . . caVsNA n, = f r 2 M t l where c, and V , are the analyte concentration and the sample volume, respectively, NA is Avogadros’ number, r the tube radius and M the molar mass of the analyte. Correspondingly, the number of atoms present within the atomiser can also be calculated from equation (4). A cup atomiser, such as is shown in Fig. 1 , is usually operated with a purge gas flow to prevent it from being oxidised. When the residence time of the analyte within the observation zone is short compared with the volatilisation time, then the atomic density can be written as where S, is the cross-section of the atom cloud perpendicular to the direction of the gas flow and 7, is the average velocity of the analyte atoms within the observation zone.For the sake of simplicity it is assumed that S, equals the area of the cup bore. Sample vapour Observation volume CUP Electrode I 1 1 I Protective t t t t gas t 1 1 [r----Sample * Presented at the 1987 Winter Conference on Plasma and Laser Spectrochemistry, Lyon, France, 12th-16th January, 1987. Fig. 1. Cross-section of cup atomiser528 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY. SEPTEMBER 1987, VOL. 2 Apart from convection of the purge gas flow the analyte atoms are also transported by diffusion.The latter process can be represented by the velocity4 where I, is the distance from the atom source, which in this instance is the bottom of the cup, to the observation zone. The velocity of the purge gas is * * (7) - V Ta v a = - . . . . . . SOT, where Vg is the purge gas flow-rate (m3 s- I ) , T, and T, are the atomiser and room temperatures, respectively, and So is the cross-sectional area of the purge gas flow. Depending on the distance from the cup opening, the real velocity of the analyte atoms will vary between vdjff and Fa. Usually, the observation zone is high enough above the cup for the relationship Vdiff << v, to hold. Recently, results on the use of a cup atomiser within a vacuum chamber have been reported.5 In this instance the velocity of the analyte atoms is simply determined by gas kinetics and the atomic density becomes - n, = (6.87 X 10-3) c , V s N A I [ t l S a ~ ~ ] .. The atomic density within the atomiser can be determined by measuring the atomic absorption signal3 (8.69 x 1 0 7 ) ~ ~ ' ( T $ Z j z(T,) n, = * * (9) lyH(c0,a) f exp(-EllkTa) where A is the absorbance, Z(7') is the state sum at temperature T , y is a coefficient accounting for hyperfine splitting, H ( o , a ) is the Voigt integral for the point of the absorption line contour a distance from the line centre of LO = 0.72a, f is the oscillator strength, hZ1 is the wavelength of the absorption line, El is the energy of the lower level involved and k is the Boltzmann constant. Experimental Atomic Absorption Measurements A conventional AAS system was used for the determination of the atomic density above the cup.The arrangement consisted of a hollow-cathode lamp, a medium-resolution monochroma- tor and a detection system with a time constant of 10 ms. The cup bore and its depth were both 4 mm. The maximum sample volume was 20 vl. The cross-section of the nozzle determining the dimensions of the purge gas flow was So = 3.88 X 10-4 m2. Therefore, the ratio between the cross-sections of the gas flow and the atom cloud of the analyte was So/S, = 30.9 . . . . . . (10) The absorption length was assumed to be equal to the diameter of the cup bore (4 mm). The position of the measuring beam for AAS was 6 mm above the bottom of the cup. The maximum heating rate and temperature of the cup were 2 kK s-1 and 3.3 kK, respectively. Argon was used as purge gas.The absorbance values for the tube atomiser at atmospheric pressure were taken from the experimental characteristic masses published by L'vov et al.3 By using equation (9) the atomic density within the furnace at peak maximum and then from that the efficiency of the atomiser were calculated. The same procedure was applied t o the FANES atomiser, whose cross-section is shown in Fig. 2. The FANES workhead can be evacuated by a vacuum pump while the lid is closed. To initiate a hollow-cathode discharge within the graphite tube (the cathode), the system is operated at low pressure, typically of 2-3 kPa. At low pressure the diffusion coefficient is 3 \ 1 \ Gas I 4 Fig. 2. Cross-section of FANES workhead.1, Carrier gas port; 2, pump port; 3, anode; 4, window; 5 , removable lid for sample injection; 6, graphite electrode; 7. graphite furnace and hollow cathode; 8, gasket; and 9, pivot. Water cooling omitted. U, and Uh are anode and heating voltages, respectively correspondingly increased relative to the atmospheric pres- sure and, consequently, the atomic density is decreased. The validity of this relationship between atomic density within the tube furnace and pressure has been shown elsewhere by AAS measurement . f ~ Atomic Fluorescence Apparatus A block diagram of the experimental arrangement for the LAFS measurements is shown in Fig. 3 . 7 The dye laser covered a spectral range of 400-650 nm without and 265-300 nm with second harmonic generation (SHG), delivering a maximum power of 80 and 15 kW, respectively. The pulse duration was typically 6 ns and the repetition rate 10 IIz.A cross-sectional area of 4 x 4 mm2 above the cup was illuminated by the laser beam, which was also imaged on to the entrance slit of the monochromator cf1/2.S). Power supply for the graphite cup atomiser, triggering of the pulse laser and data read-out were computer controlled. A specially designed hollow-cathode lamp constructed in our laboratory was used as a resonance monochromator to adjust and control the wavelength position of the dye laser. Furnace Atomisation Non-thermal Excitation Details of the FANES system have been given elsewhere.8 Results for the limits of detection achieved with FANES have been taken from a previous publication.8 The same ramp rates of the furnace temperature during atomisation, about 2 kK s-1, was used in the FANES experiments as in AAS measurements at atmospheric pressure.Consequently, the volatilisation time was the same in both instances.6 The dimensions of the graphite tubes were the same for FANES and AAS (length, 28 mm; diameter, 6 mm) as was the carrier gas, argon. Results and Discussion Comparison of Detection Limits The absolute limits of detection (LOD) for a range of elements with different volatilities are shown in Table 1. The LODs for easily volatilised elements are of the same order of magnitude for AAS and FANES but with LAFS the LOD is much lower than the theoretical limit that can be achieved by AAS. Refractory elements determined by AAS have a lowerJOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, SEPTEMBER 1987, VOL.2 529 Fig. 3. lenses; F. filter; S , shutter; HCL, hollow-cathode lamp; and BCI boxcar integrator Block diagram of the LAFS system. P, removable total reflecting prism; M, and MS, mirrors; M2, removable mirror; L,-LS, Table 1. Absolute detection limits (pg) Atomic absorption Laser AFS. Graphite Graphite graphite tube cup FANES Element CUP Ag TI Pb Au c u V C O Pd Rh Ir Na . . . . . . 0.9 . . . . . . 0.8 . . . . . . 2.5 . . . . . . 17 . . . . . . 5 . . . . . . 357 - . . . . . . 30 . . . . . . 100 . . . . . . SO00 - . . . . . . 0.1 2.2 1 .x 1 0.7 10 1 5 10 SO 0.4 * From reference 9; atmospheric pressure. t From reference 5; vacuum. 0. 1 0.4 0.0007 0.2 0.005 3 3 0.8 1.2 17u0 6 0 .2 0.06* 6t 0.7 2.0 0.6 0.04 - - - - - 300 sensitivity when a graphite cup is used than with the tube atomiser. This lower atomisation efficiency of the cup atomiser for higher temperatures can be partly compensated for by using LAFS for the detection of the atoms. Obviously, the considerable drop in the atomisation efficiency of the cup at elevated temperatures is not observed for the tube atomiser. This difference can be explained by the temperature gradient above the cup within the observation zone, which causes additional loss of analyte atoms due to molecule formation and condensation. This effect, expected to be much less for the tube atomiser, could be confirmed by LAFS experiments using the tube atomiser.10 Direct evidence of the strong temperature gradient above the cup is given by the magnitude of the fluorescence signal cis a function of the observation height, shown in Fig.4. Atomiser Efficiency From the atomic absorption signals measured for cup as well as for tube atomisers at atmospheric pressure, the atomic density can be calculated by using equation (9). The atomisa- 0 2 4 6 8 1 0 Observation heightimrn Fig. 4. the LAFS signal Influence of the observation height above the graphite cup on tion efficiency of the atomisers at peak maximum was calculated directly from this density using equation (1) and the known observation volume. For the cup atomiser the atomic density of the analyte determined experimentally was used to calculate the average velocity of the analyte atoms within the observation zone using equation ( 5 ) . This velocity enables the calculation of the residence time of the analyte atoms. The residence times for the tube atomiser at atmospheric and low pressures were obtained from equation (3) using the diffusion coefficients given by L’vov et al.3 These residence times are valid only for gas-stop conditions, in so far as they are upper limits for real experiments where additional losses by thermal convection and diffusion through the injection port are present.The residence times for several elements are shown in Table 2. Typically, the residence times for the tube atomiser operated at atmospheric pressure are of the order of 100 ms, but for FANES and the cup these are 30-100 times lower. Consequently, the condition ‘cI < ‘c2 can be met for the tube atomiser at atmospheric pressure which means qlap $ 63‘1/0.This is not so for the FANES system and the cup atomiser,530 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, SEPTEMBER 1987, VOL. 2 Table 2. Residence times of sample species in atomisers (ms) Atomisation temperature/ Graphite Graphite tube FANES Element K CUP Ag . . . . . . 2000 2.9 280 9.0 T1 . . . . . . 2100 7.4 250 7.9 Pb . . . . . . 2100 6.7 280 9.0 Au . . . . . . 2400 0.9 220 6.8 Cu . . . . . . 2600 0.9 140 4.4 V . . . . . . 3000 0.14 110 3.6 Co . . . . . . 2600 1.6 150 4.8 0.004* * From reference 5 ; vacuum, tl = 10 s. Table 3. Efficiency (la) of electrothermal atomisers from atomic absorption measurements Atomisa- Velocity Velocity r a , "/o tion ofcon- ofdif- tempera- vectiun/ fusion/ Graphite Graphite Element ture/K ms-1 ms--l cup tube* FANES? Ag .. T1 . . Pb . , Au , . cu . . v , . co , . c o t . . COEj . . 2100 0.52 2100 0.08 2100 0.13 2400 2.32 2600 2.23 3000 14.75 2600 1.24 - 1.24 - 967.0 0.17 2.5 72 4.5 0.20 4.8 69 3.9 0.17 4.2 72 4.5 0.23 0.6 65 3.4 0.35 0.6 54 2.2 0.43 0.03 23 0.7 0.32 1.0 56 0.32 0.41 - - - 5 x 10-5 - - - * Atmospheric pressure. t Pressure = 3 kPa; atomisation temperature is 200 K lower. 3 From reference 9; atmospheric pressure, tl = 1 s, atomisation 0 From reference 5; vacuum, 'tl = 10 s. temperature = 1700 K. resulting in correspondingly lower efficiencies. Vacuum volat- ilisation gives rise to very short residence times of a few ps. Results for the calculation of atomisation efficiencies together with characteristic velocities are given in Table 3.This table shows that tube atomisers at atmospheric pressure come close to the theoretical limit for efficiency when the elements studied are volatile at the atomiser temperatures achievable and when they do not tend to form stable molecules, e.g., carbides. Typical efficiency values for cup and FANES atomisers are of the order of 1-5%. The efficiency of atomisation in a vacuum is extremely low. It is worth noting that the volatilisation time for the experiments with the vacuum atomiser was 50 times greater than those used in most other experiments. Therefore, under comparable conditions with tl = 0.2 s the efficiency of the vacuum atomiser should be of the order of 0.002%, as is expected from the given gas kinetic velocity of the atoms.Minimum Number of Atoms Detectable The analyte atom density at the detection limit using the same cup atomiser for AAS and LAFS and geometrically identical tube atomisers for AAS and FANES, respectively, are given in Table 4. The densities at the detection limit are comparable for AAS experiments using the cup or the tube. The FANES emission method is able to detect typically one order of magnitude lower densities. For several elements (K, Na and Li) FANES can detect 100-times lower analyte densities than AAS. In these instances the detection capability of AAS could only be increased by using a smaller cross-section of the furnace, which is not a practical proposition. The best values for the minimum atom density detectable by FANES and LAFS are essentially the same.Only for the vacuum atomiser is the minimum atom density detectable by LAFS consider- ably lower. Table 4. Atomic density at the detection limit (m--3) Atomic absorption Laser AFS, Graphite Graphite graphite Ag . . . . 3.70 x 1O15 8.00 x l O I 4 4.31 x 1014 1.27 x lot4 'rl . . . . 1.21 x 1016 6.80 x loi4 1.09 x 1013 2.46 x 1013 Pb . . , . 2.23 x loi6 6.69 x 1015 4.44 x 10l3 5.00 x 1014 Na , . . . - 3.70 x 1015 3.24 x 1015 3.09 x 1013 AU . . . . 8.73 x 1015 3.80 x 101s - 3.94 x 1014 Cu . . . . 8.41 X 1015 8.40 X 1015 1.35 x 1015 3.19 x 1014 V . . . . 4.55 X 1016 1.46 X l0l7 2.17 X 10'7 6.39 X 1014 c o . . . . - 1.29 x 101" 6.56 x 1014 - CUP FANES* Element CUP tube 7.87 x 10'3-t 5.65 x loll$ * Pressure = 3 kPa. t From reference 9; atmospheric pressure, t1 = 1 s.$ From reference 5 ; vacuum, T~ = 10 s, atomisation temperature = 1700 K. Table 5. Minimum number of analyte atoms detectable Atomic absorption Laser AFS, Element Ag . . . . TI . . . . Pb . . . . Na . . . . Au . . . . cu . . . . v . . . . co . . . . Graphite CUP 1.18 X 108 3.88 x 108 7.15 x 108 2.79 x 108 2.69 x 108 1.46 x lo" - - * Pressure = 3 kPa. Graphite tube 6.33 x 108 5.40 x lo' 5.29 x lo" 2.93 x 10' 3.01 x lo' 6.63 x lo" 1.16 x 1011 1.02 x 10"' graphite CUP 1.38 x 107 3.50 x 10' 1.42 x 106 1.04 x 108 4.33 x 107 6.95 x 10" 2.10 x 107 2.52 x 106i 1.81 x 104i - FANES* 1.01 x 108 1.95 x 107 3.93 x 108 2.44 x 107 3.12 x 108 2.52 x 108 5.06 x 108 t From reference 9; atmospheric pressure, t, = 1 s. $ From reference 5; vacuum, tl = 10 s, atomisation temperature = 1700 K.The absolute number of atoms present within the observa- tion volume at the detection limit are given in Table 5. The difference in the number of atoms detectable by AAS using the cup and the tube, respectively, corresponds roughly to the 25-times higher observation volume of the tube atomiser. Typical FANES values are comparable to those with the cup and AAS. It should be noted that the absolute numbers of atoms detectable by LAFS are often two orders of magnitude lower than those by FANES. The lowest values have been achieved by using LAFS with a vacuum atomiser, which gives results that are three orders of magnitude lower than the best results of the non-laser methods. It is worthwhile mentioning that the degree of excitation that can be achieved by a glow discharge or an ICP is 0.01-0.1% whereas with laser excitation approximately 50% of the analyte atoms can be brought into an excited state during the laser pulse.8 Density of Matrix Species It is one of the benefits of the electrothermal atomisation that the volatilisation of analyte and matrix may occur at different temperatures.Consequently, it is not a requirement that the concentrations of the analyte relative to that of the matrix in the sample and within the gas phase of the atomiser are the same. This fact is in contrast to other types of atomisers, such as a flame or an ICP. From the analytical point of view a high matrix concentration in the gas phase should be avoided while the analyte conwntration is being measured. The high atomisation efficiencies found for the tube atomiser at atmospheric pressure means long residence times for the analyte as well as for the matrix species.In other words, theJOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, SEPTEMBER 1987, VOL. 2 531 Table 6. Maximum density of matrix species (Na) relative to the carrier gas (”/.) for 0.1% mlV NaCl solution Atomisat ion Graphite Graphite tempera- Volatilisation cup (20-1.11 tube (50-1.11 ture/K timels sample) sample) FANES 2000 0.2 11* 23 44 2400 0.2 2.9t 29 38 3000 0.5 0.23$ 14.7 14.8 * Convection velocity = 0.52 m SKI (Ag). + Convection velocity = 2.32 m S K I (Au). $ Convection velocity = 14.75 m s-l (V). Table 7. Matrix influence on LAFS detection limits Limit of detectiodpg Matrix Pb TI Pd Rh Ir 4% NaCl solution , , ., . .0.055 0.63 0.65 1.9* 140 Acetone(Merck) . . . . . . - - 2.4 3.0 25 2.6 1.5 45 1.6 3.0 40 1.0 3.0 45 1.8 2.0 - De-ionisedwater . . , . . .0.005 0.0007 0.4 3.8 40 0.57 - - Ethanol (technical grade) . . . . - - Methanol(Merck) . . . . . . - - Ethanol(purified) . . . . . . - - Acetone(purified) . . . . . . - - Methanol(purified) . . . . . . - - * 2.5% NaCl solution. separation in time of matrix and analyte peak concentrations during the atomisation can be achieved more easily for a shorter residence time of the species within the atomiser. As an example, the density of the matrix species caused by a 0.1% mlV NaCl solution is shown in Table 6. The cup system with the relatively high convection velocity of the carrier gas contains the lowest matrix concentration whereas the relative concentrations of tube and FANES atomisers are comparable.However, for both cup and FANES the lifetime of a given species within the observation volume is less than 10 ms (see Table 2) in contrast to the usual graphite furnace, which is more than 100 ms. The influence of matrix species on the limit of detection for several elements using LAFS are given in Table 7. The matrix effect is not usually very severe for elements with high volatilisation temperatures. A strong matrix effect is observed when T1 is determined in the presence of NaCl; Pb also shows a signal depression due to the NaCl matrix. The vapour pressure curves as a function of temperature for T1 and NaCl show almost identical behaviour whereas the Pb is shifted by about 150 K to a higher temperature.11 The volatilisation of Pd starts at temperatures of more than 1000 K higher, with the result that matrix and analyte concentrations within the atomiser are fairly well separated in time.This can be seen in Fig. 5 where the LAFS signals with and without NaCl matrix are shown. The strong early matrix signal is due to scattered laser radiation that could not be completely rejected by the 3 L 1 I 1- L o 1 2 3 4 - Atomisation time/s Fig. 5 . LAFS signals for Pd. Excitation and observation wavelengths were 276.3 and 351.7 nm, respectively. 5 1.11 solution of: 0, 4% mlV NaCl with 20 p.p.b. Pd; X , 4% mlV NaCl; A, 20 p.p.b. Pd; and 0, de-ionised water monochromator. As expected, the falling edge of the scatter peak is relatively steep according to the short residence time for cup atomisers.At higher temperatures the pulse shape is mainly determined by the speed of volatilisation while the heating rate of the cup is decreasing. In practice the NaCl can be volatilised with the help of a suitable ashing stage to avoid laser scatter by matrix species during the atomisation of Pd. 1. 2. 3. 4. 5 . 6. 7. 8. 9. 10. 11. References Sturgeon, R. E., and Berman, S . S., Anal. Chem., 1983, 55, 190. Falk, H., “Graphite Furnaces as Atomisers and Emission Sources in Analytical Atomic Spectrometry,” CRC Press, New York, in the press. L’vov, B. V., Nikolaev, V. G., Norman, E. A,, Polzik, L. K., and Mojica, M., Spectrochim. Acta, Part B , 1986, 41, 1043. Falk, H., Spectrochim. Actu, Part B , 1978,33, 695. Bolshov, M. A., Zybin, A. V., Koloshnikov, V. G., Mayorov, I . A., and Smirenkina, I. I., Spectrochim. Acra, Part B, 1986, 41, 487. Falk, H . , Hoffmann, E . , and Ludke, Ch., presented at “Congress on Advances in Spectroscopy and Laboratory Sciences, CASALS ’86,” Toronto, Canada, 1986. Tilch, J., Falk, H., Paetzold, H.-J., Mon, P. G., and Schmidt, K. P., presented at “Colloquium Spectroscopicum Internation- ale XXIV,” Garmisch-Partenkirken, FRG, 1985, paper 067. Falk, H., Hoffmann, E., and Ludke, Ch., Spectrochim. Acta, Part B , 1984,39, 283. Bolshov, M. A , , Zybin, A. V., and Smirenkina, I. I., Specrrochim. Actu, Part B , 1981, 36, 1143. Dittrich, K., and Stark, H.-J., presented at “6th International Symposium on High-purity Materials in Science and Technol- ogy,” Dresden, GDR, 1985, poster C93. Weast, R. C., Editor, “Handbook of Chemistry and Phyqics,” 44th Edition, Chemical Rubber Co., Cleveland, 1962. Paper J711 9 Received February IOth, 1987

ISSN:0267-9477    

DOI:10.1039/JA9870200527

出版商:RSC    

年代:1987

数据来源: RSC

摘要:

JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, SEPTEMBER 1987, VOL. 2 533 Molecular Non-thermal Excitation Spectrometry (MONES): a Procedure for the Determination of Non-metals Using Diatomic Molecules in the Non-thermal (FANES) Atomiser Part lm Determination of Fluoride and Chloride Ions by Magnesium Fluoride and Magnesium Chloride MONES* Plenary Lecture Klaus Dittrich and H. Fuchs Karl-Marx-University, Section of Chemistry, Analytical Division, DDR 70 10 Leipzig, Talstr. 35, GDR A method for the determination of the halides F- and CI- by non-thermal excitation of MgF and MgCl molecules in a glow discharge established in a hot graphite tube (2000 "C, 2 kPa) has been developed. The MgX molecules are formed by simultaneous evaporation of Mg- and F-containing species (Mg being the additive and F the sample).The detection limits are 0.5 ng of fluoride and 0.24 ng of chloride, respectively. The method is called ETE-MONES (molecular non-thermal excitation spectrometry with electrothermal evapora- tion) by analogy with the FANES technique (furnace atomisation non-thermal excitation spectrometry). Keywords: Furnace atomisation non-thermal excitation spectrometry; molecular non-thermal excitation spectrometry; non - th erm al o ve r-excita tion; molecule form a tion; n on -metals de term in ation The early analytical applications of the emission spectra of diatomic molecules utilised flames. 1-3 Later, following the trends of microanalysis, molecular emission cavity analysis (MECAj4 was developed, which is a chemiluminescence method using cavities and flames.The MECA technique has been used for the determination of halides through the measurement of InC1, InBr and In1 molecular emission.'-8 The detection limits reported for these methods were 5-50 ng for bromine and 1.3-SO ng for chlorine. Electrically heated graphite tubes have also been used for the measurement of molecular emission, and Gutsche and co-workers9%10 successfully determined halides by measuring the emission of InBr and InCl. Molecular absorption measurements have also been used for the determination of halides using diatomic molecules. Fuwa and co-workers determined chloride by molecular absorption of InCI'1 and AIC1.12 Dittrich and co-workers have also determined fluoride and chloride by MX molecular absorption (where M = Al, Ga, In or Mg) in an electrothermal atomiserl3-17 and reported detection limits in the ng region. The advantages of molecular absorption spectrometry with electrothermal evaporation (ETE-MAS) for the determination of non-metals have been reviewed,lS--'Y and the most attractive features are the simplicity of the methods and the low detection limits achieved for various elements.The development of furnace atomisation non-thermal excitation spectrometry (FANES) by Falk et al. 20321 provided a different approach to the determination of trace metals in micro-samples by atomic emission. However, in real sample analysis, depressions of the atomic emission signals of the Group 111 elements in the presence of halides was observed.z2 This was probably due to the relatively low vapour tempera- ture within the FANES atomiser tube at the time of volatilisation of the analyte and matrix species.As spectro- scopic studies of InF,23 GaF24 and MgCP molecules have been performed using conventional hollow-cathode dis- charges, it seemed reasonable to investigate the formation of these molecules in the FANES discharge. The initial results from this study have been reported recently26 for the * Presented at the 1987 Winter Conference on Plasma and Laser Spectrochemistry, Lyon, France, 12th-16th January, 1987. determination of F-, C1- and Br- ions by InX emission in the FANES source. These investigations gave rise to a new pseudonym, MONES, molecular non-thermal excitation spectrometry. The detection limits obtained for F-, C1- and Br- by MONES were superior to the values obtained earlier by ETE-MAS and other emission techniques.In this paper, the possibility of determining halides using MgX MONES measurements will be described. Experimental Apparatus A diagrammatic representation of a two-channel FANES/ MONES spectrometer is shown in Fig. 1. The two-channel 13A 15 Fig. 1. Diagram of the FANES/MONES spectrometer: 1 , power supply for heating and discharge with microcomputer control; 2, FANESMONES source (see Fig. 2); 3 A and B , two identical monochromators, SPM 2 (Zeiss Jena, GDR), with gratings of 1300 grooves mm-1 for 750-360 nm, resolution 2 nm mm-I and 2600 grooves mm-' for 360-180 nm, resolution 1 nm mm-I; 4, teletyper for programming of the microcomputer (1); 5 , Oscilloreg chart recorder (Siemens, Karlsruhe, FRG); 6, recorder, working in A, B or A-B mode; 7, two identical power supplies for PM tubes A and B, Type 4213 (Statron, Furstenwalde, GDR); 8, vacuum pump; 9, thermostat to ensure constant cooling; 10, gas; 11, quartz lenses, f = 80 mm; 12, HCL/EDL for wavelength adjustment only; 13A and €3, PM tubes, M12 FQC51 (Werk fur Fernsehelektronik, Berlin, GDR); 14, power supply for HCL; and 15, beam splitter534 v) C 3 4- .- 2 4 0 - P c .- I] 30- i .- v) C 20 .- JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, SEPTEMBER 1987, VOL.2 - 4 6 11 6 Fig. 2. Cross-section of the FANEYMONES excitation source: A, fixed parts and B, moveable parts; 1, closure with hole for sample support; 2, graphite tube and calhode for glow discharge; 3, sample support; 4, quartz window; 5 , anode for glow discharge; 6, power for tube heatitig; 7, graphite contact rings; 8, outlet to vacuum pump; 9, argor.inlet for inner Ar flow; 10, water cooling system; and 11, electrical insulator -4 .g 70 3 L- F .c 50 -P i 4- .- 30 al C c 0 c .- .; 10 .- I ( a ) 359.4 nm '1 5 t 268.94 nrn E L I I 359.0 360.0 268.0 269.0 w Wavelengthhm Fig. 3. Light emission of MgF molecules in glow discharges: T,,, . 2400 "C (MgF MONES); Mg2+ concentration, 2 pg per 10 p1; f-' concentration, A 0.1 pg per 10 pi and B 0.5 pg per 10 pi; ( u ) A band system and ( b ) B band system; broken line, non-specific background spectrometer can be used for simultaneom determination of two ions or for simultaneous background correction using the two lines (band) method. For routine simultaneous multi- elemental analysis, a multi-channel spectrometer is required.A cross-section of the FANEYMONES excitation source is shown in Fig. 2. Procedure The sample :lolution containing F- or C1- ions is mixed with a solution containing Mg2+ ions as the additive. The final concentration of Mg2+ should be higher than that of the X- ions in order to shift the equilibrium in the plasma. M g + X e M g X . . . . . . (1) A micro-volume (10-50 PI) of the solution is deposited in the graphite tube and the temperature - time - discharge pro- gramme is started: 1, drying phase; 2, first ashing phase at normal pressure; 3, evacuation phase (up to 5&100 Pa); 4, second ashing phase at low pressure; 5 , argon filling phase (1000-2000 Pa); 6, generation and stabilisation of the glow discharge (3-5 s); 7, evaporation phase (formation and excitation of the diatomic molecules); 8, light emission of the diatomic molecules and its measurement.The time for the cycle is 1-2 min, similar to that for ETA-AAS, as are the duration and the temperatures of the phases. The Zclration of the emission signal is relatively short, in the range 0.1-0.5 s. The intensity of emission is propor- tional to the concentration of diatomic molecules and hence to the concentration of X- ions in the original solution. 0 ' ' I -' 371.5 375.0 380.0 Wave1ength:nrn Fig. 4. 1900 "C (MgCI MONES); Mg2+ concentration, 3 pg per 10 pi; Cr-' concentration, 0.1 pg per 10 pl; A band system; broken line, non-specific background Light emission of MgCl molecules in glow discharges: T,,, u1 E 8 1 2 3 4 5 6 Mg2+ concentrationipg per 10 pl Fig.5. Influence of the amount of Mg2+ on the MgF MONES. Conditions: 359.4 nm; F-, 50 ng per 10 1.11; broken line, non-specific background Table 1. Optical properties of MgX m o l e ~ u l e s ~ ~ ~ 2 ~ Band Wavelength/ Excitation Molecule system Transition nm energyleV MgF . . . . A A2n+X2Z+ 368.6-346.8 3.5 B B2Z+ -+ X'Z' 274.2-263.0 4.6 C C'Z+ + X2Z+ 24Ch225 MgCl . . . . A A Z n j X 2 384-369 3.4 B B'X+X2C+ 273-266 4.6 Results and Discussion Optical Properties of MgX Molecules Table 1 shows the band systems for the MgX molecules. Using these known values, the spectra of MgX molecules were measured point by point with wavelength intervals (Ah) of 0.1 nm. The results are shown in Figs. 3 and 4.In Fig. 3, the band spectra of MgF for the A and B systems are shown. The band heads lie at 359.4 and 268.94 nm, respectively. These correspond to the 0 , O transitions. No results were found for the C system. For MgCl molecules, intense emission was found only for the A system (Fig. 4). The band head (0,O transition) lies at 377.9 nm. As the sample residue is evaporated at low pressure relatively low evaporation temper- atures are possible. These low temperatures are useful for quantitative molecule formation. No significant influence of the discharge on the molecule dissociation was observed, but no quantitative calculations were performed to establish the efficiency of molecule formation.JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, SEPTEMBER 1987, VOL.2 535 Optimisation of the Experimental Conditions Influence of Mgz+ concentration The influence of the Mg2+ concentration on the MgF MONES signal is shown in Fig. 5 . The mass of F- used in this experiment was 0.05 pg. The MgF MONES signal increased with the addition of up to 2 pg of Mg*+ to the sample, which corresponds to an ionic ratio of Mg2+ to F- of 30: 1 in the solution. It was impossible to calculate or estimate the ratio of the species in the plasma at the time of signal measurement, because of fractionated evaporation of the Mg- and F-contain- ing species. A further increase in the Mg*+ concentration did not lead to an additive enhancement of the molecular signal in contrast to MgF MAS.16 In ETE-MAS, 20-50 pg of Mg2f per sample volume gave the best analytical results.Similar results were obtained for solutions containing CI - (100 ng C1- per 10-pl sample volume). The optimum mass of Mg*+ for 100 ng of CI- was 3 pg. This corresponds to an ionic ratio in the solution for Mg2+ to C1- of 45 : 1. An increase in the mass of Mg2+ to 50 pg per 10 pl caused a large enhancement in the background emission signal for both halides. Influence of' thermal, guseous and electrical conditions The optimum conditions for the analytical determination of F- and C1- -concentrations are summarised in Table 2. As can be seen, none of the parameters have to be critically controlled. It is obviously necessary to avoid losses of Mg2+ or X- in the ashing phases, but such losses were found only at temperatures above 1200 and 1000 OC, respectively. Analytical Results for Pure Solutions Analytical results for the determination of trace amounts of F- and CI- are shown in Table 3 along with a comparison with our previous results for MgX MAS.16 The sample species were NaF and NaCI, and Mg(N03)* was the additive.The solutions were adjusted to pH 6. The detection limits were defined as the mass of analyte giving a signal equivalent to 0.25 mV in MONES and 0.01 A in AAS. From past experience, detection limits derived on these bases are similar to values obtained from 30 calculations where o is the standard deviation of repeated signals at concentrations close to the detection limits. Analytical Results for Fluoride and Chloride Determinations in the Presence of Other Halides Fig. 6 shows the dependencies of the MgF ( a ) and of the MgCl ( b ) MONES signals on the concentration of other halides. For MgF MONES there are two stages of interference.In the first interval of interferent concentration, for between 1O-I1 and 10-9 mol X-, it is clear that different halides have different interference effects, and these are in the order CI->Br->I-. This order corresponds to the dissociation energies of the molecules (Edi, MgF 4.7; MgC13.4; MgBr 3.0; and MgI 2.9 eV). The reason for this interference can probably be explained by MgXl +X,SMgX,+X, . . . . (2) In the second interval of interferent concentration (when the concentration of NaX is higher than 10-9 M) there are no differences between the interference of the different halides. In this instance the interference is probably caused by the equilibrium where M2 = Na.For MgCl MONES, the interference of fluoride added as NaF salt in the concentration range from 10-1' to 10-7 M is mainly caused by fluoride ions [as shown by equation (2)J. Fluoride forms the most stable MgX molecule. Because there is only a small difference between the NaBr and NaI interference, it can be concluded that the interference is mainly caused by Na+ ions [as shown by equation ( 3 ) ] . The results shown in Fig. 6 indicate that in the presence of other halides, interference free determination of fluoride and chloride by MgF- and MgCl- MONES is not possible. In spite of this fact it is possible to determine fluoride and chloride in the presence of other halides, when the interference concen- trations are known approximately.Table 4 gives the relative MgX, +M,=M,X, +Mg . . . . (3) Table 2. Optimum conditions for determination of fluoride and chloride ions by MgF and MgCl by ETE-MONES. Pyrolytically coated graphite tubes were used for all measurements Parameter MgF* MgCl Discharge current1mA . . 35 (25150) 30 (25135) Ar pressure1kPa . . . . 2 (1.312.8) 2 (1.312.8) Ashing temperature (10-100 Pa)/"C . . . . 1100 (90011200) 950 (90011000) Heatingrate/"Cs-1 . . . . > 2000 > 2000 Wavelength1nm . . . . 359.04 377.6 Evaporation temperature (2 kPa)/"C . . . . . . 2400 (230012450) 2100 (200012200) 268.44 * Values in parentheses are the conditions under which the MONES signal decreases to 90% of the maximum value. ~~ Table 3. Analytical results f o r MgX MONES compared with those from MgX Detection Ion Wavelength/ Excitation limit determined Molecule nm cncrgy1eV of X 1pg MONES- F- .. MgF 359.4 3.5 460 F - . . MgF 268.9 4.6 2400 CI . . MgCl 377.9 3.4 240 F - . . MgF 268.9 4.6 7500 CI- . . MgCl 377.9 3.4 9700 MAS- F - . . MgF 359.4 3.5 2300 Table 4. Analytical results for MgF and MgCl MONES in the prewnce of other halides Re la t ive detection limit, ofX in P.PJy Ion determined Matrix NaX F - . . . . . . . . NaCl 5000 F . . . . . . . . NaBr 1000 F-- . . . . . . . . NaI 600 C1 . . . . . . . . NaF 1400 CI . . . . . . . . NaBr 70 CI . . . . . . . . Nal 40 Q) U 0 - I I I L I 1 I I 10-10 10-9 10-8 10-7 10-10 10-9 1 0 - 8 10 7 Halide concentrationimol per 20 ul Interference of halide matrices on the intensity of ( a ) MgF Fig.6. and ( h ) MgCl ETE-MONES signals536 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, SEPTEMBER 1987, VOL. 2 detection limits for such determinations. The calculations of these detection limits are based on the following: concentra- tion of the interferent matrix at which the MgX MONES signal is depressed to 50%; and twice the measured detection limit (compare with Table 3). Conclusion Molecules of MgF and MgCl are formed at low pressure (2000 Pa) in hot graphite tubes (2000-3000°C) if both the metallic and non-metallic components are present. These molecules can be excited in the glow discharge of a FANES instrument and the intensity of the emitted light is proportional to the concentration of F- and C1-, if the metal concentration is high enough (an excess ratio of about 30 : 1).Comparing the results obtained with those previously found for F- and C1- determinations by InX MONES2” it can be seen that both types of molecules give similar possibilities for analytical applications. This means that the most appropriate metal component can be selected depending on the composi- tion of the sample. The MONES results with In or Mg are more than one order of magnitude more sensitive as the C1 FANES results obtained with our equipment (DL 8 ng C1-).26 So far we have not been able to detect an analytical signal for fluoride by FANES. As regards the interference found, it can be concluded that determinations of minor halides contents (10-1-10-2% F- and lO-*-lO-3% Cl-) in the presence of other halides are possible, but determinations of trace amounts of halides (<10-2O%) in the presence of other halides are not possible.Hence chemical separation procedures have to be combined with the MONES measurements, but these separation or enrichment methods ( e . g . , extraction or ion chromatography) do not need to be highly efficient, because the minor contents can be determined. A further major advantage of MONES measurements is the spectral selectivity between the halides and other elements. 1. 2. 3. 4. References Miller, W. A . , Philos. Mag., 1845, 27, 81. Salet, G., C.R. Acad. Sci., 1869, 68, 404. Salet, G., Bull. SOC. Chim. Fr., 1869, 11, 302. Belcher, R., Bogdanski, S. L., and Townshend, A., Anal. Chim. Actu, 1973, 67, 10. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20.21. 22. 23. 24. 25. 26. 27. 28. 29. Belcher, R., Bogdanski, S. L., Kassir, L. M., Stiles, D. A . , and Townshend, A., Anal. Lett., 1974, 7, 751. Abdel-Kader, M. H. K., Peach, M. E., Ragab, M. H. T., and Stiles, D. A., Anal. Lett., 1979, 12, 1399. Osibanjo, O., and Aiyai, S. O., Anal. Chim. Acta, 1980, 120, 371. Abdel-Kader, M. H. K., Peach, M. E . , and Stiles, D. A . , J. Assoc. Off. Anal. Chem., 1979, 62, 114. Gutsche, B., Rudiger, K., and Herrmann, R . , Fresenius 2. Anal. Chem., 1977, 285, 103. Gutsche, B . , and Rudiger, K., Chromatogruphia, 1978, 11, 367. Yoshimura, E., Tanaka, Y., Tsunoda, K . , Toda, S., and Fuwa, K., Bunseki Kagaku, 1977, 26, 643. Tsunoda, K . , Fujiwara, K., and Fuwa, K., Anal. Chem., 1978, 50, 861. Dittrich, K., Anal. Chim. Acta, 1978, 97, 59 and 69. Dittrich, K., Anal. Chim. Acta, 1979, 111, 123. Dittrich, K., and Meister, P., Anal. Chim. Acta, 1980, 121,205. Dittrich, K., and Vorberg, B., Anal. Chim. Acta, 1982, 140, 237. Dittrich, K., Vorberg, B., Funk, J . , and Beyer, V., Spectro- chim. Acta, Part B, 1984, 39, 349. Dittrich, K., Prog. Anal. At. Specirosc., 1980. 3, 209. Dittrich, K., Crit. Rev. Anal. Chern., 1986, 26, 223. Falk, H., Hoffmann, E., and Ludke, Ch., Fresenius Z. Anal. Chem., 1981,307, 362. Falk, H . , Hoffmann, E., and Ludke, Ch., Spectrochim. A m , Part B, 1981, 36, 767. Dittrich, K., Hanisch, B . , and Stark, H. J., “Proceedings of the Second Hungaro-Italian Symposium on Spectrochemistry,” 1985, p. 361. Barrow, R. F., Glaser, D. V., and Zeemann, P. B., Proc. Phys. SOC. London, Sect. A, 1955, 67, 962. Barrow, R . F., Jaquest, J . A. T., and Thomson, E. W., Proc. Phys. SOC. London, Sect. A , 1954, 66, 528. Joffe, R. B., and Korovin, J. I., Zh. Prikl. Spektrosk., 1978.29, 197. Dittrich, K., Hanisch, B., and Stark, H. J., Fresenius Z. Anal. Chem., 1986,324, 497. Rosen, B., Editor, “International Tables of Selected Con- stants, Volume 7. Spectroscopic Data Relative to Diatomic Molecules,” Pergamon Press, Oxford, 1970, p. 250. Vorberg, B., Dissertation, Karl-Marx-University, 1981. Welz, B., “Fortschritte in der analytischen Atomspektrosko- pie,” Volume 11, Verlag Chemie, Weinheim, 1986, p. 97. Paper J7l7 Received January 13th, 1987 Accepted May I4th, I987

ISSN:0267-9477    

DOI:10.1039/JA9870200533

出版商:RSC    

年代:1987

数据来源: RSC

摘要:

JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, SEPTEMBER 1987, VOL. 2 537 State of the Art of Glow Discharge Lamp Spectrometry* Plenary Lecture Jose A. C. Broekaert lnstitut fur Spektrochemie und angewandte Spektroskopie, Postfach 778, 4600 Oortmund 1, FRG The development of glow discharge sources for atomic spectroscopy is traced. The electrical characteristics of the analytically relevant sources, sample volatilisation and the predominant excitation and ionisation processes are discussed. The possibilities of using the Grimm-type glow discharge lamp with a flat cathode for the spectral emission analysis of compact metallic samples, for non-conducting powders, after mixing with metal powder and making into pellets and for in-depth profiling are described. New types of lamps and the features of glow discharges with hollow cathodes are evaluated.The use of glow discharges as atom reservoirs for atomic absorption and fluorescence spectrometry and recent advances in glow discharge mass spectrometry are also covered. Keywords: Glow discharge; atomic absorption spectrometry; atomic emission spectrometry; atomic fluorescence spectrometry; mass spectrometry Different types of glow discharges have been used for a long time as radiation sources in atomic spectroscopy. Not only because of their abilities to excite atomic spectra but also because of the volatilisation processes involved, they have been found to be interesting alternatives to arc and spark sources operated at atmospheric pressure. At 10-1000 Pa, the filler gas pressure range for the analytical sources being discussed, the electrical characteristic of the gas discharge (Fig.1)’ commences with a Townsend discharge. There is only a small amount of ion and free electron production and a transition range where, owing to the increased energy exchange by collisions, the current even increases at decreasing discharge voltages. In the glow discharge region, the current increases at constant current density, as here the cathode surface covered by the discharge grows with the discharge current. Once the discharge covers the whole cathode (restricted glow discharge), the current can only increase with increasing current density which requires an increase in the discharge voltage. In this “abnormal” region, charged particles acquire high energies by passing through the high electric fields produced by the high discharge voltage.t .g Q, gk! ‘v, I- I - 1 - - _ _ - - - - I I I 1 I 10-9 10-7 10-5 10-3 10-1 10 CurrentiA - Fig. 1. Characteristic of a self-sustaining gas discharge [V = f(log i)]: Vb = breakdown voltage; V, = normal cathode fall of potential; and Vd = arc voltage (similar to reference 1) * Presented at the 1987 Winter Conference on Plasma and Laser Spectrochemistry, Lyon, France, 12th-16th January, 1987. This particularly applies for the energy-rich cathode region. Accordingly, filler gas species impact on the cathode surface and may cause “sputtering” of the cathode material. When the discharge current increases further, the current density becomes so high that intensive heating of the cathode through bombardment with filler gas species may cause thermal evaporation.Then, the availability of high number densities of analyte perturbs the high fields by which the characteristic becomes normal, i.e., the current then increases at decreasing discharge voltage as is the situation for a d.c. arc. Sample volatilisation in glow discharges is mainly due to cathodic sputtering (Fig. 2 ) . Indeed, because of the acquisi- tion of high energies in the cathode region, positive ions, through the transition of mechanic momentum, may release cathode material on impact. When the sample to be analysed is brought into the cathode region or the sample itself acts as the cathode, the analyte volatilises by cathodic sputtering. In the analytically important glow discharges, hard-sphere colli- sions mainly occur.As is known from sputtering experiments with an ion source under high vacuum, the sputtering yield S then is given by K m.M s=-.-. E . . . . . . ( 1 ) ( m + W 1 h(E) = - . . . JI: R2no where K is a constant, rn the mass of the incident particles, M the mass of the static particles, E the energy of the incident particles, no the number of target atoms per unit volume and Ro the distance of closest approach under hard-sphere conditions.2 Accordingly, the influence of various parameters on the ablation rates in analytical glow discharges, as discussed later, can be investigated. However, as several ionic and neutral species at pressures of 10-1000 Pa coexist with very different energies and collide with the filler gas, it is impossible to derive quantitative conclusions from such formulae.Selective \ Mv Photoionisation and electron emission Cathodic sputtering Thermal evaporation Fig. 2. Analyte volatilisation in atomic spectrometry538 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, SEPTEMBER 1987, VOL. 2 volatilisation, which occurs when the sample material enters the discharge plasma by thermal evaporation, is typical of arc discharges at atmospheric pressure but is less so for the glow discharges being discussed here. Excitation and ionisation in glow discharges are mainly due to electron impact, but other processes also contribute.3 1. Electron impact (mainly with high energy electrons) : Excitation e + Ar+ Ar* + e e + M+ M* + e Ionisation e + Ar -+ Ar+ or AT+* 2. Collisions of the second kind: Excitation Ar* + M-+ M* + Ar Ar*+M+-+M+*+Ar Ar* + Ar+ Ar+ + Ar* Ionisation Arm + M + M+ * + Ar (also Arm) (analyte excitation by electron impact) (for cathodic sputtering, emission of secondary electrons) (analyte excitation by collisions of the second kind) (analyte excitation by collisions of the second kind) (charge transfer, sputtering by neutrals) excitation by Penning ionisation) (analyte 3.Recombination (mainly with slow electrons): M++e+M* (radiative recombina- tion) Glow discharge plasmas are known not to be in local thermal equilibrium (LTE). Indeed, electron temperatures, as measured with probes, are very high (>lo000 K). The electrons, because of their small mass, may acquire high velocities at the high field gradients. Excitation temperatures, as derived from spectroscopic measurements, are lower and differ from the thermometric lines used, This also relates to deviations from the Boltzmann distribution.The existence of temperatures of ca. 5000 and 10000 K in a hollow-cathode discharge plasma, for instance, can be explained by the non-Maxwellian velocity distribution of the electrons as a result of the high electric field and by the existence of two groups of electrons, namely a low-energy group involved in recombination and a high-energy group responsible for the excitation of high energy levels. Hollow cathode Glow discharge lamp FANES Fig. 3. spectrometry Discharges under reduced pressure for atomic emission Kinetic gas temperatures in glow discharges are low. They may be measured from the Doppler broadening of spectral lines or approximated by the rotation - vibration tempera- tures.For a glow discharge with a flat cathode, gas tempera- tures of <lo00 K resulted from Doppler width measure- ments.4 For a hollow-cathode glow discharge, values from rotation - vibration spectra5 are of the same order of magni- tude (800-1500 K). Glow Discharge Atomic Emission Spectrometry Discharges under reduced pressure, from the point of view of sample volatilisation and excitation, have unique features as radiation sources for atomic emission spectrometry. When the sample is cooled, material volatilisation is solely based on cathodic sputtering, which excludes selective volatilisation and related interferences. However, it is also possible to adjust the cathode temperature carefully by regulation of the discharge current and/or external heating.The latter provides ideal conditions for the selective volatilisation of elements or species from a complex matrix. Owing to the non-LTE character of the discharge, high line to background intensity ratios are obtained for the analyte lines. The spectra hence consist mainly of the most sensitive lines, which are narrow because of the low Doppler and pressure broadening and do not suffer significantly from interferences by molecular bands. Indeed, the discharge gas is nearly always a noble gas which, moreover, because of its high ionisation energy, also permits an efficient excitation of high-energy terms. Three main types of discharges under reduced pressure (Fig. 3) are used as radiation sources for atomic emission spectrometry and are discussed below.(GD-AES) Glow Discharge Lamp In 1968 the first practical glow discharge lamp for spectro- chemical analysis with a flat cathode was described by Grimm.6 In his lamp (Fig. 4) the sample, which must be electrically conducting, is taken as the cathode and the discharge is restricted to the sample surface. This is achieved by keeping the distance between the anode tube on one side and the cathode block and the sample on the other below the mean pathway of the free electrons, and by evacuating the interspace with a supplementary vacuum pump. The analyte is volatilised uniquely by cathodic sputtering (typical penetra- tion rate 3 pm min-1) and is excited in the negative glow of the discharge mainly by electron impact.The radiation is measured by end-on observation, As indicated by the black area in the right-hand part of Fig. 4, the material is partly deposited on the anode tube, which normally has a diameter of between 6 and 10 mm, and the remainder is sucked away into the vacuum system. The lamp is normally operated in argon at a pressure of 10-300 Pa. The vacuum supply includes a dual-vacuum pump and a needle valve for admitting the filler gas (see, for example, references 7 and 8). In commercial instrumentation interlocks and automatic rinsings are pro- vided. To pump 2 Cathode body). , Sample 0.2 mm thick ' 1 min- PTFE sheath TO pump I CatLode Fig. 4. Sample volatilisation and excitation in a Grimm-type glow discharge lamp539 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, SEPTEMBER 1987, VOL.2 In the glow discharge lamp, the material volatilises uniquely by cathodic sputtering. The influences of the sample composi- tion and the discharge parameters on the ablation rate have been studied extensively7.9 and they can be explained by the impulse theory. Indeed, this clearly indicates that the ablation rate will increase in the sequence: helium < neon < argon < krypton (for aluminium samples), and C < Al < Fe < Cu < Zn (in argon gas). In most instances argon is used as the filler gas, as the heavier noble gases are expensive and their ionisation energies are lower. When initiating the glow discharge and bringing the operation voltage and current to their nominal values of about 700 V and 80-120 mA, respectively, a certain burning time is required before the sputtering and excitation conditions become stable.This time, for example, for steel samples, may be as short as 20 s. It may be reduced by high-energy pre-burning. When equilibrium conditions are reached, the composition of the ablated material becomes equal to that of the sample, which is a pre-requisite of low matrix interferences and for the exclusion of structural effects. As shown by the burning spot records (Fig. 5 ) , no melting of the surface and accordingly no selective volatilisation occur in the glow discharge in contrast to spark erosion. As shown in many publications (see, for example, refer- ences 10-14) very different types of alloys (e.g., pure aluminium and different A1 alloys*s) can be analysed with GD-AES using the same calibration function.This demon- strates the potential interest in glow discharge lamp emission spectrometry by users who have to analyse a wide variety of samples. As shown by results for steel,lh.'7 the detection limits of GD-AES are in the range 1-10 yg g-1. Results are often higher than for spark atomic emission spectrometry. However, as will be discussed later, progress with respect to this point is still being made. Apart from compact metallic samples, non-conducting powders can also be analysed with the glow discharge lamp. They can be mixed with a metal powder and the mixture can be briquetted into pellets. By using appropriate techniques, vacuum-tight and mechanically stable pellets can be obtained even with small amounts of sample (Fig.6).*,18 As is known from the analysis of slags, for instance, analyses can be performed by calibrating with synthetic samples. The method is of topical interest because, in contrast to X-ray fluorescence spectrometry, light elements such as beryllium and boron, which may be important for the quality of new ceramic materials, can also be determined. It has been shown that the graininess of the material plays an important role in the analytical performance. 19 A major feature of the glow discharge lamp lies in the easily controllable layer by layer ablation. The penetration rate depends on the voltage, the current and the gas pressure and is material specific.20-22 By integrating the analytical signals with a small time window, information on the in-depth variation of the elemental composition is obtained.2" The penetration rate of the glow discharge lamp operated at about 100 Pa of argon and 90 W (i = 50-70 mA and V = ca.800 V) is of the order of 3.5 ym min-1. As discussed by Quentmeier and Laqua,22 RSDs of the intensities at a concentration level of 10 mg g-1 for an integration time of 1 s are below 0.1. Accordingly, for many applications integration times down to 0.1 s are still tolerable and the resolution of in-depth profiling by GD-AES is of the order of 5 nm. The technique has been proven to be useful for the control of the thickness as well as for the study of the composition of coatings on technical surfaces. Compara- tive studies of several surface analysis techniques showed that the in-depth resolution of GD-AES is lower than in secondary ion mass spectrometry and Auger electron spectroscopy, the latter being a real "surface" technique.Therefore, GD-AES has a high power of detection and a high multi-element capacity and it just may solve problems where low in-depth resolution is required.23 Applications to the quality control of coatings on steel24 and aluminium23 have been described. As in all known surface techniques, with GD-AES it is still difficult to derive quantitative information from in-depth resolved intensity profiles. This applies particularly to infor- mation from layers near to the surface where the excitation conditions in the glow discharge are often perturbed by gases released from the surface during the initiation of the disc- harge.For deeper layers, it has previously been shown22 that it is possible to quantify elements by referring to the elemental line intensities from the bulk. This idea has been pursued by Bengston ,25 who introduced model sputtering constants and discharge parameters derived from measurements with the pure metals. For galvanisation layers his glow discharge values agreed well with those obtained by electropolishing. Despite the fact that the problem of quantification has not been satisfactorily resolved, GD-AES is already used for solving many practical problems, particularly in metallurgy. 0.2 rnrn Copper + 'sample Cobper Fig. 6. Preparation of pellets from electrically non-conducting powders for glow discharge lamp spectrometry (according to refer- ence 18) Fig- 5.medium voltage spark ( b ) , respectively; sample, aluminium Burning spots obtained with a glow discharge ( a ) and aJOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, SEPTEMBER 1987, VOL. 2 540 In addition to the application areas mentioned above there is still considerable development in GD-AES. As described in papers by Ko26 and Ferreira et al., 27 a lamp in which the anode is further away from the sample and where a ceramic restrictor defines the burning spot is advantageous. Indeed, the field across the sample is then more uniform, which results in a flatter burning spot and, accordingly, in a better in-depth resolution. It was found that the analytical performance of this lamp in terms of precision and freedom from matrix interfer- ences is even better than that of the original Grimm-type lamp.Special efforts are being made to improve the power of detection of GD-AES. This is required by the demands for elemental characterisation of metals in the sub-yg g-1 range in technological applications. Some success could be obtained by increasing the ablation rate. This is possible to a certain extent with the magnetic field lamp where a strong Co - Sm magnet is used to focus the sputtering beam on the sample and at the same time to increase the length of the negative glow. As shown for stee1,28,2’ both the ablation rate and the power of detection with the magnetic lamp could be improved. It was also found that it was no longer necessary to use a second vacuum line to restrict the discharge to the sample surface.Therefore, the rinsing of the lamp became difficult for ferromagnetic samples. A further method of improving the power of detection lies in the use of a cross excitation. Indeed, as shown by Ferreira and Human,30 the greater part of the sputtered analyte in a conventional glow discharge lamp is present as ground-state neutral atoms. The use of a d.c.-boosted discharge has been propagated in the work of Lowe and other workers.31-33 It has been shown that the net intensities are considerably increased by the boosting. This helps to overcome detector noise and counting error limitations, which for the low radiative output in glow discharge work often limit the power of detection. Similar experience has been reported for h.f. boosted lamps.34 Leis et al.35.36 showed with a microwave-boosted lamp that not only a gain in net intensities but also a gain in signal to background ratio could be obtained. The detection limits for steel samples could be improved by a factor of up to five as compared with those of a conventional lamp while preserving the same analytical precision and a linear dynamic range of more than three orders of magnitude.Hollow-cathode Glow Discharge Lamps and FANES Glow discharges with hollow cathodes have long been important as radiation sources for emission spectrochemical analysis. It has been shown by Mandelstam and Nedler’7 that, in particular, owing to the high residence times of the analyte in the excitation zones which results from the cathode geometries, they are the most sensitive emission techniques.This fact is reinforced by the departures from LTE, by which the line to background ratios increase as compared with those of LTE sources.3* Both demountable cooled and hot hollow- cathode lamps have been used for spectrochemical analysis. In cooled hollow cathodes, the sample is placed as a disc at the bottom of the cathode or the hollow cathode is machined from the sample. Volatilisation takes place by cathodic sputtering only. In contrast with the glow discharge lamp with a plane cathode, however, no equilibrium in sample ablation is attained. Similarly, with the flat cathode the technique can be used for the analysis of metals and of electrically non- conducting powders. For the latter, Caroli39 showed that detection limits are in the pg g-1 range. Microanalyses with liquid sample aliquots can be performed by transferring them into a metal or graphite cathode and exciting the dry solution residues .40 The hot hollow cathode has continued to be of interest for emission spectrochemical analysis.Here, the sample volatilisation is due mainly to thermal evaporation. This has been shown by measurements on brass chips which have been excited for short periods.41 As the temperature of the cathode can be controlled very efficiently by selecting the proper discharge conditions, volatile elements can be evaporated reproducibly from refractory matrices. Accordingly, the hot hollow cathode is still being used for the analysis of high- temperature alloys, which can be used in the form of chips in graphite cathodes.Indeed, on the one hand these alloys are often difficult to dissolve which could hamper their analyses by ICP-AES but on the other hand the emitted matrix spectra are very line rich. However, in the hot hollow cathode volatile elements such as arsenic, bismuth, lead and selenium can be volatilised selectively from the matrix, hence spectral inter- ferences are circumvented. Owing to the high excitation temperatures in the hollow cathode these elements, which have high excitation energies, can be effectively excited and detection limits down to the sub-yg g-1 level are 0btained.42-~~ Hot hollow cathodes can also be used for applications such as the direct determination of toxic elements in airborne dust collected in graphite electrodes.45 The hot hollow cathode is considered to be the emission spectrochemical radiation source with the highest absolute power of detection.46 Indeed, absolute detection limits are in the picogram range in many instances, but only in the absence of a matrix.An effective separation of volatilisation and excitation is realised in furnace atomic non-thermal excitation spec- trometry (FANES), as developed by Falk et al. (see, for example, references 47 and 48). Here the samples are transferred into a graphite furnace, similar to those used in graphite furnace atomic absorption (GFAAS) work. The furnace is operated under reduced pressure and the analyte that is released by heating is excited in the negative glow between the furnace and an additional electrode. For dry sample aliquots, detection limits for a series of elements are in the picogram range.In contrast to GFAAS, simultaneous multi-element determinations are possible and calibration graphs are linear over several decades of concentration. The technique has been shown to improve considerably the power of detection as compared with furnace emission spectrometry at atmospheric pressure.49 Applications in the analysis of biological fluids have been described”) and direct solids sampling of pulverised biological substances has also recently been shown to be possible.51 Glow Discharges as Atom Reservoirs As the material volatilised in a glow discharge is to a large extent present as a vapour cloud of free atoms, glow discharges are suitable atom reservoirs for atomic absorption and atomic fluorescence work.By using a Grimm-type glow discharge as an atomiser, direct analyses of compact metallic samples can be performed by atomic absorption spectrometry (see, for example, refer- ence 52). By optimising the discharge conditions at the highest sample ablation rates, high powers of detection can be obtained. Further, the use of a reference signal has been shown to enable high precision analyses. Atomic fluorescence can also be performed on the vapour cloud. The fact that decay of the excited species by quenching, because of the reduced pressure, is low is a distinct advantage. The non-dispersive resonance fluorescence spectrometer de- scribed by Human et a1.53 makes use of one glow discharge as the primary radiation source and a second glow discharge, for which the sample is used as the cathode, as the atom reservoir.For a series of applications spectral resolution of the analytical signal has been shown not to be required.54 Multi-elemental determinations can be performed by using a multi-element primary source and a spectrometer. Laser induced fluor- escence with a glow discharge as the atom reservoir has alsoJOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, SEPTEMBER been described.55.56 In experiments with dry solution residues on a flat cathode, detection limits in the picogram range are obtained (see, for example, reference 57). Glow discharges are also of potential interest as atom reservoirs for laser enhanced ionisation techniques. Glow Discharges as Ion Sources Apart from their use as sources for optical atomic spec- trometry, glow discharges also have been recognised as being powerful ion sources for inorganic mass spectrometry. Owing to the material volatilisation by sputtering, selective volatilisa- tion as compared with classical spark source or vacuum arc ion sources is low, which is a pre-requisite for reducing the matrix interferences.As steady-state conditions in sample ablation can be obtained the use of sequential systems, such as low-priced quadrupole systems, became possible. A glow discharge mass spectrometry system thus consists of the glow discharge source, which is operated at a pressure in the range of one to several hundred Pa, a pump-backed skimmer unit to extract the ions at a suitable location and to bridge the pressure difference between the source and the mass spec- trometer and the ion optics and the mass spectrometer.Systems using discharges with one or two pin electrodes but also hollow-cathode plumes have been described by Harrison et al.57 Detection limits in the sub-pg g-1 range and a better analytical precision and accuracy than with conventional spark source systems have been reported .58 Apart from quadrupole mass filters, double focusing mass spectrometers providing higher spectral resolution have been used. It is also possible to use a modified Grimm glow discharge with a flat cathode as an ion source,59 by which, in addition to bulk analyses, in-depth profiling is possible. It has been shown that by suitable selection of the ion sampling location the intensities of the analyte and the cluster signals and accordingly the interfer- ences in the mass spectrum can be greatly influenced.Glow discharge mass spectrometry is now commercially available and may be expected to become of considerable interest for trace analyses of solid samples. Again, the technique may be applied to metals and electrically non-conducting powders, as discussed for GD-AES. Dry residues can also be analysed. The possibility of using isotopic dilution enables high ana- lytical precision and accuracy to be realised. The highest selectivity can be obtained by resonant ionisa- tion of the atoms released in the glow discharge with the aid of laser radiation. Resonance ionisation mass spectrometry thus enables it to suppress considerably the signals produced by the matrix.6” As an analytical method, however, it is still in the development stage. Conclusion Glow discharge lamp spectrometry has been shown to be a powerful method in analytical atomic spectrometry.In emission spectrometry it is an interesting alternative to spark excitation especially for in-depth analyses. Its powers of detection are still being improved. From this point of view, its use as an ion source for mass spectrometry could be an important development and, when properly optimised, high precision and accuracy as well as the capability of getting in-depth resolved information can then be realised. This work was supported by the Ministerium fur Wissenschaft und Forschung des Landes Nordrhein-Westfalen and by the Bundesministerium fur Forschung und Technologie. References 1 .2. Penning, F. M . , “Electrical Discharges in Gases,” Philips Technical Library, Eindhoven, 1957, p. 41. Kaminsky, M., “Atomic and Ionic Impact Phenomena on Metal Surfaces,” Springer-Verlag, Berlin, 1965, p. 227. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20, 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 1987, VOL. 2 541 Flugge, S., Editor, “Encyclopedia of Physics,” Volume XXII, Springer-Verlag, Berlin, 1956. Ferreira, N. P., Human, H. G . C., and Butler, L. R. P., Spectrochim. Acta, Part B, 1980, 35, 287. Broekaert, J. A . C., Bull. SOC. Chim. Belg., 1977, 86, 895. Grimm, W., Spectrochim. Acta, Part B, 1968, 23, 443. Dogan, M., Laqua, K., and Massmann, H., Spectrochim. Acta, Part B , 1971, 26, 631.El Alfy, S . , Laqua, K., and Massmann, H., Fresenius Z . Anal. Chem., 1973, 263, 1. Boumans, P. W. J . M., Anal. Chem., 1972, 44, 1219. Dogan, M., Laqua, K., and Massmann, H., Spectrochim. Acta, Part B, 1972, 27, 65. Klockenkamper, R., Laqua, K., and Dogan, M., Spectrochim. Acta, Part B, 1980, 34, 527. Kruger, R. A., Butler, L. R. P., Liebenberg, C. J . , and Bohmer, R. G . , Analyst, 1977, 102, 949. Jager, H., Anal. Chim. Acta, 1972, 60, 303. Ferreira, N. P., and Butler, L. R. P., Analyst, 1978, 103, 607. KO, J. B., and Laqua, K., XVIII Colloquium Spectroscopicum Internationale, Grenoble, Abstracts 11, 1975, p. 543. Wagatsuma, K., and Hirokawa, K., Anal. Chem., 1984, 56, 908. Rademacher, H . W., and de Swardt, M. C., Spectrochim.Acta, Part B, 1975, 30, 353. El Alfy, S . , PhD Dissertation, Dortmund, 1978. Mai, H., and Scholze, H., Spectrochim. Acta, Part B, 1986,41, 797. Berneron, R., Spectrochim. Acta, Part B., 1978, 33, 665. Waitlevertch, M. E., and Hurwitz, J . K., Appl. Spectrosc., 1976, 30, 510. Quentmeier, A . , and Laqua, K., in Koch, K. H., and Massmann, H . , Editors, “13. Specktrometertagung,” W. de Gruyter, Berlin, 1981, p. 37. Quentmeier, A., Bubert, H., Garten, R. P. H., Heinen, H. J., Puderbach, H., and Storp, S., Mikrochim. Acta, Suppl., 1985, 11, 89. Koch, K. H., Kretschmer, M., and Grunenberg, D., Mikro- chim. Acta, 1983, 2, 225. Bengston, A., Specrrochirn. Acta, Part B, 1985, 40, 631. KO, J . B., Spectrochim. Acta, Part B, 1984, 39, 1405. Ferreira, N. P., Strauss, J.A., and Human, H. G. C., Spectrochim. Acta, Part B , 1983, 38, 899. Kruger, R . A., Bombelka, R. M., and Laqua, K., Spectrochim. Acta, Part B, 1980, 35, 581. Kruger, R. A., Bombelka, R. M., and Laqua, K., Spectrochim. Acta, Part B, 1980, 35, 589. Ferreira, N. P., and Human, H. G. C., Spectrochim. Acra, Part B, 1981,36, 1981. Lowe, R. M., Spectrochim. Acta, Part B, 1978, 31, 257. Lomdahl, G . S . , McPherson, R . , and Sullivan, J. V., Anal. Chim. Acta, 1983, 148, 171. Gough, D. S . , and Sullivan, J. V., Analyst, 1978, 103, 887. Walters, P. E . , and Human, H . G . C., Spectrochim. Acta, Part B , 1983,36, 585. Leis, F., Broekaert, J. A. C., and Laqua, K . , XXIV Col- loquium Spectroscopicum Internationale, Garmisch-Partenkir- chen, Book of Abstracts, Volume 4, 1985, p.640. Leis, F., Broekaert, J . A. C., and Laqua, K., Spectrochim. Acta, Part B, in the press. Mandelstam, S. L., and Nedler, V. V., Spectrochim. Acta, 1961, 17, 885. Falk, H., Spectrochim. Acta, Part B, 1977, 33, 437. Caroli, S . , Progr. Anal. At. Spectrosc., 1983, 6, 253. Harrison, W. W., and Prakash, N. J., Anal. Chim. Acta, 1970, 49, 151. Broekaert, J . A. C . , Spectrochim. Acta, Part B, 1979, 34, 11. Thornton, K., Analyst, 1969, 94, 958. Thelin, B., Appl. Spectrosc., 1981, 35, 302. Berglund, B., and Thelin, B., Analyst, 1982, 107, 867. Broekaert, J. A. C., Bull. SOC. Chim. Belg., 1976, 85, 755. Zil’bershtein, Kh.I., “Spectrochemical Analysis of Pure Sub- stances,’’ Adam Hilger, Bristol, 1977. Falk, H., Hoffmann, E., and Ludke, Ch., Spectrochim. Acta, Part B, 1981, 36, 767. Falk, H . , Hoffman, E., and Ludke, Ch., Spectrochim. Acta, Part B, 1984, 39,283. Falk, H., Hoffmann, E., Ludke, Ch., Ottaway, J. M., and Giri, S. K., Analyst, 1983, 108, 1459.542 50. 51. 52. 53. 54. 55. 56. JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, SEPTEMBER 1987, VOL. 2 Falk, H., Hoffmann, E., Ludke, Ch., Ottaway, J. M., and Littlejohn, D., Analyst, 1986, 111, 285. Falk, H., Hoffmann, E., Ludke, Ch., and Schmidt, K. P., Spectrochim. Acta, Part B, 1986, 41, 853. McDonald, D. C., Anal. Chem., 1977, 49, 1337. Human, H. G. C., Ferreira, N. P., Kruger, R. A., and Butler, L. R . P., Analyst, 1978, 103, 469. Bubert, H., Spectrochim. Acta, Part B, 1984,39, 1337. Smith, B. W., Omenetto, N., and Winefordner, J. D . , Spectrochim. Acta, Part B, 1984, 39, 1389. Patel, B. M., and Winefordner, J. D., Spectrochim. Acta, Part B, 1986,41,469. 57. Harrison, W. W. , Hess, K. R. , Marcus, R. K., and King, F. L., Anal. Chem., 1986, 58, 341A. 58. Bruhn, C. G., Bentz, B. L., and Harrison, W. W., Anal. Chem., 1979,51, 673. 59. Jakubovsky, N., Stuwer, D., and Tolg, G., Znt. J. Mass Spectrom. Ion Proc., 1986, 71, 183. 60. Hess, K. R., and Harrison, W. W., Anal. Chem., 1986, 58, 1696. Paper 5715 Received January 12th, 1987

ISSN:0267-9477    

DOI:10.1039/JA9870200537

出版商:RSC    

年代:1987

数据来源: RSC

摘要:

JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, SEPTEMBER 1987, VOL. 2 543 Sample Introduction in Plasma Emission and Mass Spectrometry* Plenary Lecture Richard F. Browner and Guangxuan Zhut School of Chemistry, Georgia Institute of Technology, Atlanta, Georgia 30332-0400, USA The differences between sample introduction strategies for inductively coupled plasmas when used as either photon sources for atomic emission spectrometry or ion sources for mass spectrometry are discussed. The concept of a coupled interface in mass spectrometry detection is contrasted with the uncoupled interface existing in emission spectrometry detection. Differences in the roles that analyte mass transport and solvent mass transport may play in the two types of systems are compared. The influence of various voltages in the mass spectrometer interface on optimum signal response are compared for a representative selection of elements.Keywords: Inductively coupled plasma; inorganic mass spectrometry; atomic emission Spectrometry; sample introduction In recent years a great deal of interest has focused on the role of sample introduction in atomic emission spectrometry. 1-4 It has become clear that the sample introduction process remains a limiting factor in two analytical areas of considerable importance. These concern (1) the use of emission spectrometry for analysing difficult samples with good freedom from matrix interferences and (2) the achievement of low detection limits. The recent advent of inductively coupled plasma mass spectrometry, with its generally enhanced detection limits compared with atomic emission spectrometry, calls for a re-evaluation of the role of sample introduction.A key point of interest is whether the factors that are significant and limiting in atomic emission spectrometry remain the same in the case of mass spectrometry. This paper reviews some basic principles of sample introduction both in inductively coupled plasma atomic emission spectrometry (ICP-AES) and in inductively coupled plasma mass spectrometry (ICP-MS) and discusses points of similarity and difference. Results and Discussion The four key analytical benchmarks which the sample introduction process influences are: (1) detection limits; (2) degree of interference from matrix components or solvents; (3) measurement precision; and (4) speed and ease of analysis.One clear difference between the two techniques, from a mechanistic standpoint, relates to the species population which determines the magnitude of the analytical signal. In atomic emission spectrometry, this is proportional to the population of excited state atoms and ions. However, in ICP-MS the signal is related to the population of ground-state ions. From this simple standpoint, it should be possible to generalise to a statement of the optimum environment for analyte reaching the plasma which will provide optimum populations for ICP-AES and ICP-MS. Further, moving one step back in the process, it should also be possible to consider what procedures can be used to produce samples in a form most appropriate for each technique.There are many quantitative factors which can be used to evaluate the quality of the sample introduction process. One of the most useful is a measurement of the analyte mass * Presented at the 1987 Winter Conference on Plasma and Laser t On leave from Dalian Institute of Chemical Physics, Dalian, Spectrochemistry, Lyon, France, 12th-16th January, 1987. People’s Republic of China. transport rate,2 W,, which is a measurement of the analyte mass per second reaching the analytical source in a form which can be efficiently converted into the measurement form of interest, i.e., excited atoms and ions in ICP-AES, and ground- state ions in ICP-MS. A second factor of great importance is the solvent loading on the plasma. It has been shown, and will be discussed later, that solvent loading can have a critical effect on the plasma in both atomic emissions and mass spectrometry.6 However, it is very important to consider the different role that the solvent plays in these two spectrometric techniques because this may give rise to different conclusions on the optimum solvent loading.Finally, in order to minimise any possible matrix interaction effects in the plasma, it is clearly necessary to maintain the mean drop size of the aerosol which reaches the plasma ( i e . , the tertiary drop size) below some critical value, d,,,. When the drop size exceeds d,,,, a reduction in the rate of atom formation in the plasma will occur, which will lead to matrix effects of the type well documented in atomic absorption spectrometry (AAS).’ While values ford,,, have been determined for AAS, they are as yet unknown for ICP-AES and ICP-MS.Analyte Mass Transport It is clear, in principle, that a higher rate of analyte delivery to the plasma should lead to a higher analyte signal if there is a linear relationship between W, and the optical emission signal in ICP-AES or the ion population in ICP-MS. However, in practice, it is difficult to accomplish this increase in mass transport without an accompanying increase in solvent trans- port, and this has given rise to the attempted development of many desolvation systems for various sample introduction devices.sI0 It is notable, in this respect, that the improve- ments which have been typically found in atomic emission spectrometry using ultrasonic nebulisation have not really been of any great magnitude, except when desolvation devices have been used.Without desolvation a typical improvement might be a factor of 2.8,lO With desolvation, the improvement can rise to a factor of 10, which is roughly proportional to the increase in analyte mass transport to the plasma observed with such a systems There are of course many additional practical difficulties associated with the ultrasonic nebulisation - desolvation process which have hindered its more widespread application. 11 Other approaches which have achieved favour for various reasons include the fritted-disc’* and grid-type nebulisers, which allow a more efficient production of aerosol. However, these devices do not give rise to very significant improvements544 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, SEPTEMBER 1987, VOL.2 in detection limits because they are limited by the analyte flow-rate which is ca. 100 pl min-1. Consequently, although the net efficiencies of the devices may be very high, typically %-60°/~, the net analyte transport rate, W, to the plasma remains comparable to that which is found with conventional nebulisers and spray chambers. Thus, these devices may prove highly useful for coupling to low flow, e . g . , microbore liquid chromatography separations but are unlikely to achieve any particular advantage in general applications where sample volume is not restricted. A more direct approach has been applied by Lawrence et aZ.,13 in which essentially 100% analyte transport is achieved through direct introduction of the entire effluent of a micropore LC column to a nebuliser built into the injector tip of the plasma torch.Similarities and Differences Between Sample Introduction in Similarities between the two techniques are primarily in the physical construction and operation of the sample introduc- tion device and the plasma. The plasma system, comprising plasma gases, r.f. power generator, nebuliser and spray chamber are essentially identical for both systems, with only a few minor differences of geometry. These systems also have a requirement that the sample be efficiently converted into an aerosol form and be transported to the plasma through the well established route of sample aerosol leading to desolvated sample aerosol, salt crystals and ultimately atoms (Fig.1). The primary differences between ICP-AES and ICP-MS obviously relate to differences between both the species of interest sampled from the plasma, and the manner in which they are sampled. For ICP-AES, photons emitted by the plasma and detected by the photomultiplier (through a spectrometer) may be considered as uncoupled from the sample introduction process and even from the plasma itself (Fig. 2). This implies that there is no interaction between the detection system and the plasma per se, which is a key difference in relation to ICP-MS where there is the existence of direct coupling between the plasma and the detection system. This arises because the manner in which the plasma is operated directly influences the interface and the detection system. It has been determined that ions sampled at the interface can influence the operation of the various focusing and extraction voltages that exist at the interface between the skimmer cone and the quadrupole mass filter.Some of these ICP-AES and ICP-MS Bulk analyte solution Aerosol Sa I t cry st a Is Molecular vapour I Atomic and ionic vapour Excited atoms and ions 1 I 1 4 I CP-AE S ICP-MS Fig. 1. Sample introduction routes for ICP-AES and ICP-MS voltage seem to be critical, whereas others appear to be much less sensitive to small voltage changes away from the optimum. It has been observed that variations in the voltages of the lens and optical focusing elements of the mass spectrometer can exert a major influence on the extent to which the technique is free from interferences.While it may not be necessary to optimise these voltages for each sample type, there exists the potential for a noticeable sensitivity response in the mass spectrometer to sample type. However, ICP-MS occupies a largely superior position to ICP-AES with respect to spectral interferences. With complex sample matrices, ICP-AES has the potential for substantial spectral overlaps, except where spectrometers of exceptionally high resolution are used, such as UV region Fourier transform and echelle spectrometers. The much simpler spectra of ICP-MS have fewer potential spectral overlaps, but have less options for selection of non-overlapped isotopes should an isobaric interference occur. One aspect of ICP-MS where the sample introduction process may have a significant effect on potential interferences is in its influence on oxide and polyatomic-ion interferences.In ICP-MS it is important to ensure that the percentage of oxides and other polyatomic ions formed is kept to an absolute minimum, because these species may give rise to the type of unwanted isobaric interferences mentioned above. Conse- quently, controlling the volume of solvent introduced to the ICP may play a key role in controlling these interferences. Detection Limits in ICP-AES and ICP-MS From published values of detection limitsl4J5 it appears that ICP-MS detection limits are typically between 10- and 100-fold lower than ICP-AES detection limits. Approximately 52 elements may be determined with detection limits in the range 0.01-0.1 mg ml-1. A further 14 elements have detection limits in the range 0.1-1 ng ml-1, and 4 have detection limits between 1 and 10 ng ml-1.Only 8 elements, namely carbon, nitrogen, oxygen, fluorine, silicon, phosphorus, sulphur and chlorine, have detection limits greater than 10 ng ml-1. It is of some interest that although there are some slight variations in published performance characteristics between the two com- mercially available ICP-MS instruments, these are relatively minor and, in general, the detection limits quoted are remarkably uniform for the two systems. However, it should be noted that the detection limits are also dependent on the sample type studied and will, generally, degrade in the Interface Quadrupole Detector Plasma Lenses etc. Fig. 2. Sources of analytical signals in ICP-AES and ICP-MS: (a ICP-AES, no direct coupling of photons and spectrometer; and (b] ICP-MS, direct coupling of ions and spectrometer through interfaceJOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY.SEPTEMBER 1987, VOL. 2 545 presence of high concentrations of matrix elements. This type of behaviour is also typical of ICP-AES systems, where high background levels result from high matrix concentrations. One of the key driving forces in efforts to improve sample introduction in ICP atomic emission spectrometry has been the need to improve detection limits substantially, without causing a consequent degradation in either precision or its freedom from interferences. The question, with regard to ICP-MS, is whether the generally better detection limits, in fact, make the requirement for improved sample introduction throughput redundant, in contrast to ICP-AES. The tentative answer is that at the present stage of development, provided that freedom from matrix interferences can be maintained in ICP-MS, this may well be so.In general, detection limits in ICP-MS are sufficiently low that they are probably adequate for most trace and ultra-trace elemental analysis purposes at the present time. When operating close to the current detection capabilities of ICP-MS, considerable strain may be placed on the entire chain of sample handling, dissolution and dilution. This necessitates extreme care to avoid contamina- tion from various sources, and also a judicious choice of solvents to minimise polyatomic-ion interferences. Because of its excellent detection capabilities, it may well be that with ICP-MS the most important aspects of sample introduction to be considered will relate to factors other than the need to increase analyte throughput to the plasma.This may lead to a significant divergence in the sample introduction goals for ICP-MS compared with ICP-AES. Aspects of Signal Generation for ICP-AES and ICP-MS It has been well documented in many studies, especially those of Blades and Horlick.16 that the spatial emission properties of inductively coupled plasmas are in fact complex when significant amounts of matrix are present. However, the aspect of signal measurement in ICP-AES that reduces the impact of these height profile effects is the ability to use a mean measurement height of 15 mm, with an optical window extending a few millimetres above and below this value.This results in an averaging of spatial shifts in the plasma which typically cancels out any net shift which may otherwise be observed. It may be anticipated that ion sampling would be more sensitive to position with ICP-MS than with ICP-AES, because the height-averaging effect is not present. Ions are effectively sampled from only a very restricted spatial volume in the plasma. In practice, it appears that with appropriate instrument design the influence of sampling depth is not a critically important factor.17 For a given nebuliser gas flow - r.f. power combination the signal will generally decrease as the sampling depth is increased above its optimum value. However, this is generally a smooth trend and not accompan- ied by any significant variation in the influence of nebuliser gas flow.This is a very important factor because it means that essentially one set of conditions may be selected for a particular series of elements without the need to optimise individually for each element. The general validity of this relationship appears to be instrument dependent and other workers, using different instrumentation,lx have concluded that height optimisation may be necessary for each set of gas flow conditions. The effect of the nebuliser gas flow on ICP-AES generally results in an incredse in the optimum height of measurement for atom lines and relatively little change in the optimum height of measurement for ion lines. With ICP-MS, the effect of the nebuliser gas flow has a different effect.Signals from all ions show common trends, i.e., that there is an optimum gas flow for any particular r.f. power. For example, with the VG PlasmaQuad instrument, this leads to an optimum gas flow of approximately 0.68 1 min-1 at a forward power of 1.35 kW.17 These conditions are generally applicable for a wide range of elements. The optimum value for the nebuliser gas flow appears to shift to higher values as the forward power to the plasma is raised, but reaches a plateau at about 1.35 kW, which is generally selected as an optimum value. Clearly some balance between power, gas flow and position of measurement is critical, and this will also depend upon the choice of auxiliary gas flow. These trends have been shown to occur for elements covering the mass range from 24 to 208 (e.g., Mg, Ba and Pb), which emphasises the reproducible nature of these trends and the ability to achieve an optimum combination of nebuliser flow and forward r.f.power over a wide range of masses. It is also fortunate that optimum nebuliser flow and optimum power for maximising the ion count for the element of interest generally corresponds to a minimum in both the M2+ and MO+ populations. The absolute populations of these other ions will vary from element to element. 17-19 For elements with very high second ionisation potentials and low oxide bond energies such as Pb, values for the M2+/M+ ratio would typically be 0.04% and for the MO+/M+ ratio would typically be 0.001%. For Ce, which has a much lower second ionisation potential and a much stronger metal-oxide bond, both these numbers will be significantly higher, e.g., in the region of 1-5%.Extraction, Lens and Pre-filter Voltages Because the mass spectrometer is tightly coupled to both the plasma and the ion-sampling region, control of various voltages in the system involved with ion extraction and focusing is an important factor. This is a significant point of difference between ICP-MS and ICP-AES. However, the behaviour appears to be reasonably reproducible and straightforward. Fig. 3 shows schematically the positioning of extraction and focusing voltages in the PlasmaQuad interface. Fig. 4 shows how control of these voltages may or may not have a significant influence on the ion counts observed for each element.The actual range of voltages appropriate for each of the components differs substantially, and vary from a few volts to more than 100 V. However, as can be seen from the figures, the system response is relatively insensitive to certain of the voltages but fairly sensitive to others. Clearly these must be set with some care. In particular, the lens voltages for L 1-4 (Fig. 3), can have a substantial effect on the properties of the ion current. However, it can be observed that a peak position may be obtained for each of these which is broadly the same for the ions measured. In fact, there is some variation from element to element, in that the relative position of the peak currents may vary substantially from lens to lens. Nonetheless, once the peak position of each lens voltage is established, this should provide an acceptable compromise condition for a wide mass range of elements.The optimum voltages selected should take account of the particular group of elements selected for measurement. Skimmer Collector Extractor I aperture cone Sampling Differential Fig. 3. Extraction and focusing plates in PIasmaQuad interface546 - X-x---x-x I I I JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, SEPTEMBER 1987, VOL. 2 t al Y \ \ 4.5 7.5 10.5 13.5 16.5 -199 -208 -217 -226 -235 -20 -19.3 -18.6 -17.9 -17.2 A x-x- x-x I I I I I -199 -208 -217 -226 -235 Lens 1 C ( b ) U .. . . 2.7 5.7 8.7 11.7 14.7 Lens 2 (el .J. : - A x-x-x-x-x 1 I 1 I I -35.4 -32.4 -29.4 -26.4 -23.4 I Differential aperture (h) I .. u. c] ......C'" c .... D 'A- - (2-- - - <;>+ :@22 :.+ ~---d B \ A x-x-x-x-x I I I I I I 1 - 139.8 - 136.8 - 133.8 - 130.8 - 127.8 ( k ) C Front pole A x-x-x-x-x I 1 I I I -150 -147 -144 -141 -138 Applied volt ag e N -1.2 1.8 4.8 7.8 10.8 Lens 4 ( f l U .. . . : p *. 0 \ . -5.4 -2.4 0.6 3.6 6.6 I A I x-x-x-x-x -4.2 -3.6 -3.0 -2.4 -1.8 ~ ( I) C Pre-filter A X- x-x-x-x -1.32 -0.72 -0.12 0.48 1.08 Fig. 4. (g-f) locations less sensitive to voltage setting. A, 'Li; B, W o ; C, '3'jLa; and D, zOyBi Influence of extraction, focusing and pre-filter voltages on system response: (u-fl locations sensitive to voltage setting; and Role of Solvent Loading on Plasma Properties The importance of solvent in determining plasma excitation and ion-generation properties has been referred to earlier.One of the key factors here is clearly the extent to which the solvent influences two important properties, namely plasma temperature and plasma chemistry. The situation is compli- cated by the likelihood that these processes are interactive. In ICP-AES, the introduction of water to solvent-free analyte initially increases the electron density. For a water input level of approximately 20 mg 1-1 of gas, typical for a pneumatic nebuliser - Scott spray chamber, the electron density reaches a plateau region at about 2 x 101 * electrons cm-3. As the water loading is increased above this value, the temperature of the plasma decreases significantly, which can result in a substantial decrease in the atomic emission signa1.2O.2' However, in ICP atomic emission, the decrease in the emission signal from the plasma is accompanied by a corresponding decrease in plasma background emission.As the detection limits of the system are set by the ratio of signalJOURNAL O F ANALYTICAL ATOMIC SPECTROMETRY, SEPTEMBER 1987. VOL. 2 547 to background, the net effect of increased water loading on detection limits is relatively slight. However, the influence on the raw signal may be substantial. For example, doubling the water loading from 12 to 24 mg 1-1 can result in a 100-fold reduction in the emission signal from Mg 1.2’ In ICP-MS, water loading has a less dramatic influence on the population of ions sampled by the mass spectrometer, hence the primary importance of water loading may be in the formation of unwanted oxide species.As has been shown by Gray6 the introduction of sample to the mass spectrometer system through laser sampling can result in a dramatic reduction in oxide levels, typically by two orders of magni- tude. Clearly there is a maximum level of water which can be tolerated in ICP-MS sampling, if oxide levels are to be kept to acceptable levels, as water forms the primary route to oxide formation in the plasma. Another form of sample introduction which would accomplish the same goal of minimum water vapour introduction to the plasma would be the introduction of sample by use of an electrothermal vaporiser. These devices have been shown to have value in ICP atomic emission spectrometry when working with very small samples, and are capable of producing very low absolute detection limits.20J2 Their use in ICP-MS would likewise appear to form a natural link.This paper is based on work supported by the National Science Foundation under Grant No. CHE-8503090. The loan of a PlasmaQuad instrument from VG Isotopes is gratefully acknowledged. 1. 2. References Boorn, A. W., and Browner, R. F., Anal. Chem., 1982, 54, 1402. Browner, R. F,, Boorn, A. W., and Smith, D. D., Anal. Chem., 1982, 54, 533. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. Browner, R. F.; and Boorn, A. W., Anal. Chem., 1984, 56, 786A. Browner, R. F., and Boorn, A. W., Anal. Chem., 1984, 56, 875A. Long, S. E., and Browner, R. F., Spectrochim. Acta, Part B , 1986, 41, 639. Gray, A. L., Analyst, 1985, 110, 551. Smith, D. D., and Browner, R. F., Anal. Chem., 1984, 56, 2702. Olson, K. W., Haas, W. J., and Fassel, V. A., Anal. Chem., 1977, 49, 623. Boumans, P. W. J . M., and deBoer, F. J . , Spectrochim. Acta, Part B , 1976, 31, 355. Mermet, J . M., and Trassy, C . , in Barnes, R. M., Editor, “Developments in Atomic Plasma Spectrochemical Analysis,” Heyden, London, 1981, p. 245. deGalan, L., Anal. Chem., 1986, 58, 697A. Layman, L. R., and Lichte, F. E., Anal. Chem., 1982,54,638. Lawrence, K. E . , Rice, G . W., and Fassel, V. A., Anal. Chem., 1984, 56, 289. Douglas, D. J . , and Houk, R. S . , Prog. Anal. At. Spectrosc., 1985, 8, 1. Gray, A. L., Spectrochim. Acta, Part B , 1985, 40, 1525. Blades, M. W., and Horlick, G . , Spectrochim. Acta, Part B , 1981, 36, 861. Zhu, G., and Browner, R. F., Appl. Spectrosc., 1987,41,349. Vaughan, M. A . , and Horlick, G., Appl. Spectrosc., 1986,40, 434. Gray, A. L., and Williams, J . G., J. Anal. At. Spectrom., 1987, 2,81. Long, S. E., Snook, R. D., and Browner, R. F., Spectrochim. Acta, Part B , 1985, 40, 553. Long, S . E., Kull, R., and Browner, R. F., Spectrochim. Acta, Part B , 1987, submitted. Matusiewicz, H., and Barnes, R. M., Appl. Spectrosc., 1984, 38,745. Paper J 71 45 Received April 6th, 1987

ISSN:0267-9477    

DOI:10.1039/JA9870200543

出版商:RSC    

年代:1987

数据来源: RSC

摘要:

JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, SEPTEMBER 1987, VOL. 2 549 Flow Injection Techniques in Inductively Coupled Plasma Spectrometry* Plenary Lecture Cameron W. McLeod Department of Chemistry, Sheffield City Polytechnic, Sheffield S l I WB, UK The development and current status of flow injection techniques in inductively coupled plasma atomic emission spectrometry is reviewed. Topics such as direct sample introduction, dilution, calibration and on-line sample pre-treatment are used to illustrate the versatility of the combined experiment. The analytical potential of flow injection techniques in inductively coupled mass spectrometry is given brief consideration. Keywords: Inductively coupled plasma spectrometry; flow injection; trace element analysis; automated an a I ysis The last two decades have witnessed extensive developments in plasma spectrochemical methods, in particular, inductively coupled plasma atomic emission spectrometry (ICP-AES) and more recently inductively coupled plasma mass spectrometry (ICP-MS).Much of this development has been based on the measurement of steady-state signals after pneumatic nebulisa- tion of liquid samples, although attention has been given to discrete sample introduction techniques such as pulse nebuli- sation,] electrothermal vaporisation2 and flow injection (FI)3,4 which are characterised by signals of a transient nature. Flow injection is based on the injection of a discrete sample plug into a non-segmented carrier stream, and since its inception in 1975 the technique has emerged not only as a novel sample introduction tool but also as a versatile sample handling facility for atomic spectrometry.Therefore, features such as high precision, high sampling rates, microsample manipulation, calibration techniques, ease of automation and ability to perform rapid on-line pre-treatment chemistry have all provided the necessary stimulus for development of FI-based methodology. Authoritative reviewss-8 are available and it is of interest to note that publications in the atomic spectrometry field account for less than 10% of a substantial FI literature.8 It is clear that the original pioneers of ICP (Greenfield and Fassel) and FI techniques, (Rfiiitka, and Hansen) have made significant contributions to FI - ICP-AES. The aim of the present paper is to trace the evolution of FI - ICP systems and to indicate the potential advantages of the combination.Some emphasis is given to microcolumn techniques as a means of improving method sensitivity and overcoming spectral interferences in ICP-AES. The scope for the use of FI methods in ICP-MS is addressed and brief consideration is given to results for the multi-element analysis of blood serum. Principles of Flow Injection Methods The underlying principle of all FI methods is the controlled dispersion of the sample in the carrier stream as the sample moves from the injection point to the detector. Provided exact timing and synchronisation of injection with the measurement routine have been arranged, then reproducible measurements of the dispersed sample zone can be achieved.It is normal practice, as a measure of dispersion, to relate the maximum concentration of analyte (Cmax) in the sample zone at the detector to the original (undispersed) analyte concentration (0) before injection, through measurement of the respective * Presented at the 1987 Winter Conference on Plasma and Laser Spectrochemistry, Lyon, France, 12th-16th January, 1987. transient and steady-state signals (i. e., dispersion coefficient, D = Co/Cmax). (In FI - ICP-AES measurements it has been standard practice to integrate the transient signals.) The power of the FI method lies in the ability to manipulate sample dispersion to suit analytical requirements and this is achieved by control of key experimental parameters such as injection volume, carrier stream flow-rates and tubing dimensions. Flow injection systems have been classified, as shown in Fig.1, according to the degree of dispersion.9 A limited dispersion (D = 1-3) system is used for direct sample introduction applications. High sensitivity is preserved as mixing of the sample with the carrier stream is minimal. Medium dispersion (D = 3-10) corresponds to significant mixing of the sample and carrier stream and is utilised in standard additions experiments and in those applications where analytical reac- tions are necessary. Large dispersion (D > lo), achieved by incorporation of a mixing chamber in the carrier line may be used for on-line sample dilution but more efficient techniques including “electronic” dilutions are favoured for this purpose. A further category of dispersion, which is not illustrated in Fig.1, is required to describe the on-line pre-concentration experiment where D values are <l. In this experiment a relatively large volume of sample (e.g., 10 ml) is passed through, for example, a microcolumn of ion-exchange material and retained analytes are eluted by injection of a small volume ( e . g . , 200 pl) of eluent (giving a nominal pre-concentration factor of 50, assuming quantitative deposi- tion and elution for the analyte). This approach is a powerful method for enhancing sensitivity in ICP spectrometric analysis. Dispersion System D 1-3 1 D 3-10 D 2 10 Classification of FI manifolds. Fig. 1. of Elsevier from reference 9 Response curve i A n Reproduced with permission550 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, SEPTEMBER 1987, VOL.2 Inductively Coupled Plasma Atomic Emission Spectrometry The development of FI-based methodology, as outlined in Table 1, has in part been prompted by certain limitations of the ICP technique. Thus, in the analysis of complex samples, novel FI strategies have been used to minimise physical and spectral interference and to enhance method sensitivity. Guidelines for experimental design have been well documen- ted in the overview papers of Greenfield374 which have done much to stimulate interest in FI - ICP-AES. Flow injection provides an important route for direct sample introduction and early studies3.4,1@13 demonstrated important advantages over conventional procedures including a microsampling capability, high sampling rates and relative freedom from nebuliserhnjector tip blockage.Thus samples of high viscosity or high dissolved solids content such as blood serum which cause nebulisation/transport problems in conven- tional ICP analysis were analysed by FI - ICP-AES. 12713 In parallel with developments in flow injection atomic absorption spectrometry (FI - AAS), novel calibration proce- dures have been utilised in ICP-AES. A standard additions procedure (first described by TysonS for flame atomic absorption spectrometry), whereby the sample served as the carrier stream and into which aliquots (25 pl) of standard solution were injected, was utilised by Greenfield3>4 for the determination of Ca in a cement reference material. A medium dispersion manifold was constructed by insertion of a 100 cm length of tubing between the injection point and the ICP nebuliser.The calibration data are reproduced in Fig. 2 and it is seen that interpolation was used to obtain the unknown sample concentration. Israel and Barnes,14 using a simplified model for inverse dispersion, derived an expression for estimation of the unknown sample concentration in FI - standard additions experiments. The method, which was based on the measurement of two FI signals, was applied to the determination of silica in phosphoric acid. Flow injection systems with a merging zones configuration have been used for implementing internal standardisation11 and the generalised standard additions method. 15 The latter calibration proce- dure, effective in overcoming spectral interferences in ICP emission spectrometry was considerably simplified in the FI mode.Application to synthetic solutions and alloy digests was demonstrated. In general, FI systems have been treated as a novel accessory for standard ICP equipment, with the result that, although improvements in performance and capability of the FI unit have occurred, fewer attempts have been made to modifyhmprove the sample introduction - ICP interface. Fas- sel and co-workers16J7 realised that a principal limitation of FI - ICP experiments was the low analyte transport efficiency associated with pneumatic nebulisation and, consequently, a total injection microconcentric nebuliser, giving essentially 100% efficiency, was designed. The system, applicable to FI or HPLC effluents flowing at 100-200 pl min-1 was found to be comparable to or better than conventional pneumatic nebulisation, in terms of limits of detection, and this perfor- mance contrasts with most FI studies where reduced detection capabilities have been reported.Gustavsson18 has described a new interface for sample introduction in ICP spectrometry and with special reference to continuous flow applications. The system consisted of a conventional concentric nebuliser, a heated chamber for desolvation and a jet separator. Analyte transport efficiencies were 35% with an acceptable solvent loading. Flow injection systems provide a novel approach to on-line sample pre-concentration - matrix removal with traditional chemical separation procedures being scaled down for the purpose.Most studies have utilised microcolumns packed with chelating or ion-exchange materials and these develop- ments have been prompted by an original FI - AAS study by Olsen et al. 19 Activated alumina offers a novel route for analyte pre- concentration as it can function both as an anion and a cation exchanger depending upon solution pH.20 Under acidic and basic conditions it exhibits a high affinity for oxyanions and cations, respectively. Yx- I / J Found 46.3 p.p.m. 40 p.p.m. -5.2 k 0.56 Lu Q 4 p.p.m. -25.6 k 0.4 f. 1.83 I 10 20 30 40 50 60 70 80 90 100 Ca concentration, p.p.rn. Fig. 2. Calibration graph for FI standard additions for NBS Portland cement, 62.09% CaO (= 44.38% Ca); sample mass 0.1031 g in 100 ml (= 45.76 pg ml-1). Reproduced with permission from reference 3 Table 1.Development and status of FI - ICP spectrometry 1964 1975 197% 1981 1981- 1983 1983- 1984- 1986- Elemental analysis by ICP emission spectrometry Concept of FI introduced Combination of FI with atomic absorption spectrometry Scope for FI - ICP experiment Direct sample introduction Calibration techniques Generalised standards addition Direct injection nebuliser On-line pre-treatment chemistry: trace enrichment, matrix removal and speciation Integration of robot and FI systems for automated ICP analysis Acidic Neutral Basic In one study21 a microcolumn of acidic alumina was used to separate and pre-concentrate phosphate from a potentially interfering iron matrix. The FI manifold, illustrated in Fig. 3, consisted of a peristaltic pump, dual injection rotary valve and a microcolumn of acidic alumina.During operation, samples (<2 ml) were injected into the carrier stream and phosphate ions were retained on the column whereas cations were OH- I / \ I Sample Fig. 3. FI manifold (acidic alumina) phosphate. Reproduced with permission reference 21 for trace enrichment of of Pergamon Press fromJOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, SEPTEMBER 1987, VOL. 2 - - 4 W Elute T 55 1 9.5 Sample A * W - * - L I 1.8 Buffer A ml min-' - A - - - Eluate ---.-*w - unretained. An injection of potassium hydroxide solution (200 yl, IM) was then used to elute phosphate into the ICP and obtain an interference-free measurement of P at 213.62 nm. An emission -time response representative of the above sequence is given in Fig.4. The determination of P in a steel reference material was demonstrated. The same approach has been extended to a range of oxyanions21 and a report on the determination of sulphate in waters is given in reference 23. Given the extent and significance of spectral interference in ICP-AESZ4 there is considerable scope for further develop- ment of the analyte enrichment - matrix removal concept. As a recent example, promising results have been obtained for the determination of B in metallurgical samples using a boron- specific chelating resin.25 The development of a rapid speciation scheme for inorganic Cr was based on the FI manifold with acidic alumina.26 As in previous examples the alumina column displayed a high affinity for anionic CrVI in contrast to that for Cr"' and this enabled time-resolved emission for the two valency states to be monitored after injection of the samples.Limits of detection based on a 2-ml sampling volume were 1.4 and 0.2 yg 1-1 for Cr"' and CrVI, respectively. Application of the technique to reference waters revealed that Cr"' was the predominant species. The speciation technique did not offer a pre-concentration capability for the trivalent species and this prompted investigations with the basic form of alumina. Analogous experiments to the previous one indicated quanti- tative retention -elution (with 2 M nitric acid) of Cr"' with a microcolumn of basic alumina and, for a 5-ml sampling volume, a sub-pg 1-1 determination capability was realised.27 The determination of Cr"' in human urine was attempted but co-elution of matrix cations (Ca, Mg) resulted in background interference such that reliable determinations at natural concentration levels were precluded.More recently it has been shown that basic alumina has a high affinity for a wide range of cations28 and studies are currently in progress on the usefulness of this technique for multi-element pre-concentra- tion. Several groups of workers have demonstrated the effective- ness of chelating resins for on-line multi-element pre-concen- tration. Hartenstein et aZ.29 utilised the manifold shown in Fig. 5 to achieve substantially improved limits of detection (1&50 fold) for Ba, Be, Cd, Cr, Cu, Mn, Ni and Pb relative to those for conventional nebulisation. Samples were passed through the microcolumn at a flow-rate of 9.5 ml min-1 and the analytes were subsequently eluted with nitric acid.Use of dual columns and sample loading times of 40-190 s enabled 30 samples per hour to be processed. In a later report the procedure was optimised for ultratrace analysis of waters.3" In the study of Hirata et aZ.31 a microcolumn of Muromac A-1 was used for on-line pre-concentration of Al, Cr, Fe, Ti and V. Load 1 L - J Injector - valve 2 1.8 850 v) c 3 + 2 650 -. c >. v) c .- 2 450 .- C 0 .- $ 250 .- E w 50 H20 0 20 40 60 80 100 t Time's B t A inject inject Fig. 4. Emission (213.62 nm)-time response for injection of (A) phosphate sample (100 pg ml-l of P, 10000 pg ml-l of Fe, 10% HNO,; 200 pl at t = 0 s) and (B) 0.5 M KOH (50 p1 at t = 40 s) Improvements in sensitivity of up to two orders of magnitude relative to conventional nebulisation were reported and with sample loading times of 3&180 s (at 6 ml min-1) a sampling rate of 17 h-l was achieved.The Muromac A-1 resin was deemed resistant to volume changes with changes in pH and this was considered an important advantage over Chelex 100. More recently, the polydithiocarbamate and poly- (phenylurea) - poly(ethy1eneaminodiacetic acid) resins, which have been used extensively in conventional column and batch experiments, were adapted for on-line use in a commercial FI system . 3 2 3 Details were presented on system performance and considerable differences were noted in exchange charac- teristics for many elements relative to the column and batch mode results.33 In the present context reference has not been made to corresponding developments in atomic absorption spectrometry but an evaluation and critical comparison of systems based on 8-hydroxyquinolines,3436 salicylic acid37 and pyridylmethylethylenediamine38 is warranted.It is clear that on-line pre-treatments can considerably enhance the analytical capabilities of ICP-AES and although emphasis, in this discussion, has been given to microcolumn techniques other possibilities including hydride genera- tion3Y140 and solvent extraction41 have met with considerable success. An FI system has been combined with a laboratory robot and ICP-AES for automated sample preparation and analysis of lubricating oils.42 The integrated approach resulted in a more efficient analysis and the analytical precision (over 8 h) for transient signal measurement wzs found to be at least an order of magnitude better for all elements relative to a steady-state measurement.The substantial improvement in precision was related to several factors associated with FI namely: (i) a relatively constant solution flow-rate (irrespec- tive of sample viscosity), (ii) an improved plasma stability and (iii) a reduced build-up of carbon on the torch injector tip. Another recent development, computer-intelligent auto- mated sample dilution, was reported at the 1986 Winter Conference on Plasma Spectrochemistry.43 The FI system performed (if necessary) on-line sample dilution, without operator knowledge or intervention, so that measurements were made at appropriate analyte concentration ranges.This approach has considerable implications in the field of expert analytical systems and an increasing use of FI technology in such systems is projected. Inductively Coupled Plasma Mass Spectrometry Few studies, as yet, have been reported on FI - ICP-MS but given the similarity in source operating requirements many of Injector valve 1 - W PumD 1 Turn valve t t 1 2 Fig. 5 . FI manifold (Chelex 100) for trace enrichment of metal ions. Reproduced with permission of the American Chemical Society from reference 29552 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, SEPTEMBER 1987, VOL. 2 the techniques utilised in emission are transferrable to ICP-MS. Matrix isolation and standard additionshnternal standardisation procedures are of particular relevance given that ICP-MS is prone to matrix Owing to the inherent sensitivity of ICP-MS there may be less of a need for on-line pre-concentration except in situations where measurements of low abundance isotopes or element specia- tion are required.It is well known that continued aspiration of samples of relatively high dissolved sblids may cause deposits not only on the torch injector tip but on the mass spectrometer sampling interface and frequent re-standardisation or instru- ment shut down (in extreme instances) may be necessary. Such problems may be obviated by FI techniques. In a preliminary study, the potential of FI - ICP-MS for the determination of nine trace elements in blood serum has been assessed .47 Few analytical techniques have the necessary sensitivity for such an analysis but the high detection power of ICP-MS46 is an attractive feature when considered in relation to normal concentration levels in serum.48 A detailed presen- tation of procedure and results is not given here but it was shown that major matrix elements (C, C1, N, Na and S) were responsible for both signal enhancement and suppression.Signal suppression was detected for most elements studied and although simulation experiments in a sodium chloride matrix (3000 pg ml-1 Na) revealed similar trends the extent of suppression (30-70%) did not correlate (as might be expected for ionisation interference) with ionisation potential. This matrix effect was minimised using a standard additions procedure and the data for Al, Cd, Mo, Pb and Sb were consistent with literature values.48 Severe spectral interfer- ences due to the formation of polyatomic ions precluded reliable determinations of As (due to 75ArC1+), Cr (due to 5*ArC+), Mn (due to 55ArN+, 55ArNH+) and V (due to 5 1c10+).The collaboration and support of colleagues at Sheffield City Polytechnic, the British Geological Survey and the Royal Armament Research and Development Establishment is gratefully acknowledged. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. References Cresser, M. S., Prog. Anal. At. Spectrosc., 1981, 4, 219. Barnes, R. M., and Fodor, P., Spectrochim. Acta, Part R, 1983, 38, 1191 and references cited therein. Greenfield, S., Ind. Res. Dev., 1981, 21, 140. Greenfield, S . , Spectrochim. Acta, Part B, 1983, 38, 93. Tyson, J. F . , Analyst, 1985, 110, 419.Gallego, M., Luque de Castro, M. D., and Valcarcel, M., At. Spectrosc., 1985, 6, 16. RGiiEka, J., Fresenius 2. Anal. Chem., 1986, 324, 745. Christian, G . D., and RGiitka, J., Spectrochim. Acta, Part B, 1987, 42, 157. RfiiiCka, J., and Hansen, E. H., Anal. Chim. Acta, 1978, 99, 31. Ito, T., Nakagawa, E., Kawaguchi, H., and Mizuke, A . , Bunseki Kagaku, 1980, 29, 332. Jancintho, A. O., Zagatto, E . A. G., Bergamin, F. H., King, F. J., Reis, B.F., Bruns, R . E., and Kowalski, B. R., Anal. Chim. Acta, 1981, 130, 243. Alexander, P. W., Finlayson, R. J . , Smythe, L. E., andThalib, A., Analyst, 1982, 107, 1335. McLeod, C. W., Worsfold, P. J . , and Cox, A. G., Analyst, 1984, 109, 327. Israel, Y., and Barnes, R. M., Anal. Chem., 1984, 56, 1188.15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27, 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40 * 41. 42. 43. 44. 45. 46. 47. 48. Zagatto, E . A. G . , Jacintho, A. O., King, F. J., Resi, B. F., Bruns, R. E., and Araaujo, M. C. U., Anal. Chim. Acta, 1983, 145, 169. Lawrence, K. E., Rice, G . W., and Fassel, V. A., Anal. Chem., 1984, 54, 162. La Freniere, K. E., Rice, G. W., and Fassel, V. A., Spectrochim. Acta, Part B, 1985, 40, 1495. Gustavsson, A . , Spectrochim. Acta, Part B, 1987,42, 111. Olsen, S., Pessenda, L. C. R . , R i i i t k a , J . , and Hansen, E. H., Analyst, 1983, 108, 905. Patterson, R . , “An Introduction to Ion Exchange,” Heyden, London, 1970, p. 17. McLeod, C. W., Cook, I. G., Worsfold, P. J., Davies, J . E., and Queay, J . , Spectrochim.Acta, Part B, 1985,40. 57. Cook, 1. G . , McLeod, C . W., and Worsfold, P. J . , Anal. Proc., 1986, 23, 5. Cox, A. G., McLeod, C. W., Miles, D., and Cook, J., J. Anal. At. Spectrom., 1987, 2, 553. Boumans, P. W. J . M., Fresenius Z. Anal. Chem., 1986, 324, 397 * Anderson, D., and McLeod, C. W., unpublished work. Cox, A. G., Cook, I. G., and McLeod, C. W., Analyst, 1985, 110, 331. Cox, A. G . , and McLeod, C. W., Anal. Chim. Acta, 1986.179, 487. Zhang, Y., Cook, I. G., Cox, A. G., and McLeod, C. W . , in preparation. Hartenstein, S. D., Rfiiitka, J . , and Christian, G . D., Anal. Chem., 1985, 57, 21. Hartenstein, S. D,, Christian, G . D., and RgiiEka, J., Can. J . Spectrosc., 1985, 30, 144. Hirata, S . , Umezaki, Y., and Ikeda, M., Anal. Chem., 1986, 58, 2602. Barnes, R. M., Spectroscopy (Springfield, Oreg.), 1986, I , 25. Wang, X., and Barnes, R. M., presented at the 1987 Winter Conference on Plasma and Laser Spectrochemistry, Lyon, France, 12th-16th January, 1987, paper No. C1. Malamas, F., Bengtsson, M., and Johansson, G., Anal. Chim. Acta, 1984, 160, 1. Marshall, M. A., and Mottola, H . A . , Anal. Chem., 1985, 57, 729. Milosavzevic, E. B., Sohyic, L., Nelson, J . H., and Hendrix, J. L., Mikrochim. Acta, 1985, 111, 353. Fang, Z . , Xu, S., and Zhang, S., Anal. Chim. Acta, 1984,164, 41. Bengtsson, M., Malamas, F . , Torstensson, A . , Reyrell, O., and Johansson, G., Mikrochim. Acta, 1985, 111, 209. Liversage, R. R., Van Loon, J . C., and de Andrade, J. C., Anal. Chim. Acta, 1984, 161, 275. Tioh, N. H . , Israel, Y., and Barnes, R. M., Anal. Chim. Acta, 1986, 184, 205. Kumarmaru, T., Nitta, Y., Nakata, F., Matsuo, H., and Ikeda, M., Anal. Chim. Acta, 1985, 174, 183. Granchi, M. P., Biggerstaff, J. A , , Hillard, L. J., andGrey, P., Spectrochim. Acta, Part B, 1987, 42, 169. Ihrig, P. J., Dobbins, J. T., and Reynolds, R. J . , presented at the 1986 Winter Conference on Plasma Spectrometry, Hawaii, 2nd-8th January, 1986, paper No. 44. McLaren, J. W., Mykytnik, A. P., Willie, S. N . , and Berman, S. S., Anal. Chem., 1985, 57, 2907. McLeod, C . W., Date, A. R., and Cheung, Y. Y., Spectro- chim. Acta, Part B, 1986, 41, 169. Gray, A. L., Spectrochirn. Acta, Part B, 1986,41, 151. McLeod, C. W., Date, A. R., and Cheung, Y. Y., unpublished work. Morrison, G. H., Crit. Rev. Anal. Chem., 1979, November, p. 287. Pup er J 7/58 Received May 5th, 1987

ISSN:0267-9477    

DOI:10.1039/JA9870200549

出版商:RSC    

年代:1987

数据来源: RSC