摘要:

CONFERENCE ANNOUNCE 1995 FIRST CONFE (I In order to promote wlsi mation among analytical scientists of zhc whole Mediterranean Basin t h ~ 1 First Mediterranean Basin Conference m Analytical Chemistry will provide M adequate forum for reportin thoroughly discussing the latest research results in basic and instrumental developments in Analytical Chemistry. Other aims of this Conference are - To promote new oppormniti Sea area (particularly for those in Soar fhern Bank) to attend intmxati ori a I meetings in countries of the region t~ attend training workshop;- on new analytical techniques to attend short cmlaarses on new techniques and trmxki in Analytical Chemistry and to establish new links for research in/or other countries of the region. - To stimulate the progress of Analytical Chemistry as a whole by solving analytical problems affecting the Mediterranean Area.ntists in the ~~~i~~~~ The program has been designed t~ attract participants from industry universities and research centers. The program will comprise invited plenary and key-note lecturers contributed oral papers and posters distributed in several Symposia covering the following topics Education of Analytical Chemistry Environmental Analytical Chemistry Agriculture and Food Analysis Geoanalyticaf. Chemistry and Benefitiafion of Minerals Biomedical Analysis kcheometry and Art Qbjects Preservation Quality Assurance and Harmonization of Procedures. A few Short Courws Special Sessions on "hot" topics and an E,xhibition s f Instrumentation has also been arranged. Invited lecturers who have already confirmed their contribution include M. Val&-cel I.B. Brenner D. Barcel6 S. Carol& A. Laachach. O.X.F. Donnard Khater H. Munfau J. Albaiges B.Y. MeMati P. Quevauvmer etc. CALL FOR PAPERS Titles of submitted oral or poster presentations are solicited with the preli registration card by 30 May 1995. Submision of final Conference abstract^ arc?. requested not later than 30 .June 1995 SOCIAL ACTIVITIES Varied social activities including a i s i t to Granada are For further information and Prof. Alfred0 Sanz-Me Department of Physical and Faculty of Chemistry University sf Qviedo C/ Julim Claverria s/n 33006ovieds lease cantact SPAIN Phone 34-8-5103.180 o 3 4 - 8 - 5 103474 FAX 4 - 8 - 5 103125

ISSN:0267-9477    

DOI:10.1039/JA99510BP003

出版商:RSC    

年代:1995

数据来源: RSC

摘要:

Journal of Analytical Atomic Spectrometry (Including Atomic Spectrometry Updates) JAAS Editorial Board* Chairman B. L. Sharp (Loughborough UK) J. M. Gordon (Cambridge UK) S. J. Haswell (Hull UK) S. J. Hill (Plymouth UK) R. C. Hutton (Winsford UK) D. Littlejohn (Glasgow UK) J. Marshall (Middlesbrough UK) D. L. Miles (Keyworth UK) A. Sanz-Medel (Oviedo Spain) JAAS Advisory Board F. C. Adams (Antwerp Belgium) R. M. Barnes (Amherst MA USA) L. Bezur (Budapest Hungary) M. W. Blades (Vancouver Canada) R. F. Browner (Atlanta GA USA) S. Caroli (Rome Italy) A. J. Curtius (Norianopolis Brazil) J. B. Dawson (Leeds UK) M. T. C. de Loos-Vollebregt (Delft The Nether L. Ebdon (Plymouth UK) M. S. Epstein (Gaithersburg MD USA) Fang Zhao-lun (Shenyang China) W. Frech (Urnei Sweden) A. L. Gray (Egham UK) S.Greenfield (Loughborough UK) G. M. Hieftje (Bloomington IN USA) B. V. L'vov (St. Petersburg Russia) R. K. Marcus (Clemson SC USA) J. M. Mermet (Villeurbanne France) T. Nakahara (Osaka Japan) Ni Zhe-ming (Beijing China) N. Omenetto (lspra Italy) C. J. Park (Taejon Korea) R. E. Sturgeon (Ottawa Canada) V. Sychra (Prague Czech Republic) R. Van Grieken (Antwerp Belgium) A. Walsh ,K. 6. (Victoria Australia) B. Welz (Uberlingen Germany) ids) Atomic Spectrometry Updates Editorial Board Chairman *D. L. Miles (Keyworth UK) J. Armstrong (Edinburgh UK) J. R. Bacon (Aberdeen UK) C. Barnard (Glasgow UK) R. M. Barnes (Amherst MA USA) S. Branch (High Wycombe U K ) R. Bye (Oslo Norway) J. Carroll (Middlesbrough UK) M. R. Cave (Keyworth UK) S. Chenery (Keyworth UK) *J. M.Cook (Keyworth UK) 'M. S. Cresser (Aberdeen UK) H. M. Crews (Norwich UK) J. S. Crighton (Sunbury-on-Thames UK *J. €3. Dawson (Leeds UK) J. R. Dean (Newcastle upon Tyne UK) *A. T. Ellis (Oxford UK) *E. H. Evans (Plymouth UK) J. Fazakas (Budapest Hungary) A. Fisher (Plymouth UK) 'J. M. Gordon (Cambridge UK) D. J. Halls (Glasgow UK) *S. J. Hill (Plymouth UK) K. W. Jackson (Albany NY USA) R. Jowi t t (Middlesbrough UK ) K. Kitagawa (Nagoya Japan) J. Kubova (Bratislava Slovak Republic) *J. Marshall (Middlesbrough UK) H. Matusiewicz (Poznan Poland) A. W. McMahon (Manchester UK) J. M. Mermet (Villeurbanne France) R. G. Michel (Storrs CT USA) T. Nakahara (Osaka Japan) Ni Zhe-ming (Beijing China) P. R. Poole (Hamilton New Zealand) P. J. Potts (Milton Keynes UK) W.J. Price (Budleigh Salterton UK) C. J. Rademeyer (Pretoria South Africa) *M. H. Ramsey (London UK) P. G. Riby (Greenwich UK) A. Sanz-Medel (Oviedo Spain) *B. L. Sharp (Loughborough UK) I. L. Shuttler (Uberlingen Germany) S. T. Sparkes (Plymouth UK) R. Stephens (Halifax Canada) J. Stupar (Ljubljana Slovenia) R. E. Sturgeon (Ottawa Canada) *A. Taylor (Guildford UK) G. C. Turk (Gaithersburg MD USA) J. F. Tyson (Amherst MA USA) P. Watkins (London UK) 6. Welz (Uberlingen Germany) J. Williams (Egham UK) J. 8. Willis (Victoria Australia) *Members of the ASU Executive Committee Editor JAAS Janice M. Gordon The Royal Society of Chemistry Thomas Graham House Science Park Milton Road Cambridge CB4 4WF UK. Telephone + 44 (0) 1223 420066. Fax + 44 (0) 1223 420247. E-mail RSCl @RSC.ORG (Internet) Senior Assistant Editor Brenda Holliday Editorial Secretary Lesley Turney US Associate Editor JAAS Dr.J. M. Harnly US Department of Agriculture Beltsville Human Nutrition Research Center Beltsville MD 20705 USA.. Telephone 301 -504-8569 Assistant Editor Ziva Whitelock Advertisements Advertisement Department The Royal SQciety of Chemistry Burlington House Piccadilly London W1 V OBN UK. Telephone + 44 (0) 171 -287 3091. Fax -1 44 (0) 171 -494 11 34. Information for Authors Full details of how to submit materials for publi- cation in JAAS are 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 communi- cations and letters concerned with the development and analytical application of atomic spectrometric techniques.The journal is pub- lished twelve times a year including comprehen- sive reviews of specific topics of interest to practising atomic spectroscopists and incorpor- ates the literature reviews which were previously published in Annual Reports on Analytical Atomic Spectroscopy (ARMS). Manuscripts intended for publication must describe original work related to atomic spectro- metric analysis. Papers on all aspects of the sub- ject will be accepted including fundamental studies novel instrument developments and prac- tical analytical applications. As well as AAS AES and AFS papers will be welcomed on atomic mass spectrometry X-ray fluorescence/emission spectrometry and secondary emission spec- trometry.Papers describing the measurement of molecular species where these relate to the characterization 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 particularly welcome. Manuscripts on other subjects of direct interest to atomic spectroscopists including sample prep- aration and dissolution and analyte pre-concen- tration procedures as well as the statistical interpretation and use of atomic spectrometric data will also be acceptable for publication. There is no page charge.The following types of papers will be considered. f u l l papers describing original work. 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 receipt. 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 particular facet of analytical 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 elsewhere 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 spacing) should be sent to Janice M. Gordon Editor JAAS or Dr. J. M. Harnly US Associate Editor JAAS. All queries relating to the presentation and sub- mission 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 JAAS Editorial Board (who may be contacted directly or via the Editorial Office) would welcome comments suggestions and advice on general policy matters concerning JAAS. Fifty reprints are supplied free of charge. Journal of Analytical Atomic Spectrometry (JAAS) (ISSN 0267-9477) is published monthly by The Royal Society of Chemistry Thomas Graham House Science Park Milton Road Cambridge CB4 4WF UK.All orders accompanied with payment should be sent directly to The Royal Society of Chemistry Turpin Distribution Services Ltd. Blackhorse Road Letchworth Herts. SG6 IHN UK Tel. +44 (0) 1462 672555; Telex 825372 Turpin G; Fax +44 (0) 1462 480947. Turpin Distribution Services Ltd. is wholly owned by The Royal Society of Chemistry. 1995 Annual subscription rate EEA €51 2.00 USA $941 50 Canada f538.00 (+ GST) Rest of World f538.00. Customers should make payments by cheque in sterling payable on a UK clearing bank or in US dollars payable on a US clearing bank. Air freight and mailing in the USA by Publications Expediting Inc.200 Meacham Avenue Elmont NY 11 003. USA Postmaster send address changes to Journal of Analytical Atomic Spectrometry (JAAS) Publications Expediting Inc. 200 Meacham Avenue Elmont NY 11 003. Postage paid at Jamaica NY 11 431. All other despatches outside the UK by Bulk Airmail within Europe Accelerated Surface Post outside Europe. PRINTED IN THE UK. @The Royal Society of Chemistry 1995. 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.Journal of Analytical Atomic Spectrometry (Including Atomic Spectrometry Updates) JAAS Editorial Board* Chairman B.L. Sharp (Loughborough UK) J. M. Gordon (Cambridge UK) S. J. Haswell (Hull UK) S. J. Hill (Plymouth UK) R. C. Hutton (Winsford UK) D. Littlejohn (Glasgow UK) J. Marshall (Middlesbrougi9 UK) D. L. Miles (Keyworth UK) A. Sanz-Medel (Oviedo Spain) JAAS Advisory Board F. C. Adams (Antwerp Belgium) R. M. Barnes (Amherst MA USA) L. Bezur (Budapest Hungary) M. W. Blades (Vancouver Canada) R. F. Browner (Atlanta GA USA) S. Caroli (Rome Italy) A. J. Curtius (florianopolis Brazil) J. B. Dawson (Leeds UK) M. T. C. de Loos-Vollebregt (Delft The Netherlands) L. Ebdon (Plymouth UK) M. S. Epstein (Gaithersburg MD USA) Fang Zhao-lun (Shenyang China) W. Frech (UmeA Sweden) A. L. Gray (Egham UK) S. Greenfield (Loughborough UK) G. M. Hieftje (Bloornington IN USA) B. V.L'vov (St. Petwsburg Russia) R. K. Marcus (Clemson SC USA) J. M. Mermet (Villeurbanne France) T. Nakahara (Osaka Japan) Ni Zhe-ming (Beijing China) N. Omenetto (Ispra Italy) C. J. Park (Jaejon Korea) R. E. Sturgeon (Ottawa Canada) V. Sychra (Prague Czech Republic) R. Van Grieken (Antwerp Belgium) A. Walsh K. B. (Victoria Australia) B. Welz (Uberlingerr Germany) Atomic Spectrometry Updates Editorial Board Chairman *D. L. Miles (Keyworth UK) J. Armstrong (Edinburgh UK) J. R. Bacon (Aberdeen. UK) C. Barnard (Glasgow UK) R. M. Barnes (Amherst MA USA) S. Branch (High Wycombe UK) R. Bye (Oslo Norway) J. Carroll (Middlesbrough UK) M. R. Cave (Keyworth UK) S. Chenery (Keyworth. U K ) *J. M. Cook (Keyworth UK) *M. S. Cresser (Aberdeen UK) H. M. Crews (Norwich UK) J. S. Crighton (Sunbury-on-Thames UK) *J.B. Dawson (Leeds UK) J. R. Dean (Newcastle upon Tyne UK) *A. T. Ellis (Oxford UK) *E. H. Evans (Plymouth UK) J. Fazakas (Budapest Hungary) A. Fisher (Plymouth. UK) *J. M. Gordon (Cambridge UK) D. J. Halls (Glasgow. UK) *S. J. Hill (Plymouth UK) K. W. Jackson (Albany NY USA) R. Jowitt (Middlesbrough UK ) K. Kitagawa (Nagoya Japan) J. Kubova (Bratislava Slovak Republic) *J. Marshall (Middlesbrough UK) H. Matusiewicz (Poznan Poland) A. W. McMahon (Manchester UK) J. M. Mermet (Villeurbanne France) R. G. Michel (Storrs CT USA) T. Nakahara (Osaka Japan) Ni Zhe-ming (Beijing China) P. R. Poole (Hamilton New Zealand) P. J. Potts (Milton Keynes; UK) W. J. Price (Budleigh Salterton UK) C. J. Rademeyer (Pretoria South Africa) *M. H. Ramsey (London WK) P.G. Riby (Greenwich UK) A. Sanz-Medel (Oviedo Spain) *B. L. Sharp (Loughborough UK) I. L. Shuttler (Uberlingen Germany) S. T. Sparkes (Plymouth UK) R. Stephens (Halifax Canada) J. Stupar (Ljubljana Slovenia) R. E. Sturgeon (Ottawa Canada) *A. Taylor (Guildford UK) G. C. Turk (Gaithersburg MD USA) J. F. Tyson (Amherst MA USA) P. Watkins (London UK) B. Welz (Uberlingen Germany) J. Williams (Egham UK) J. B. Willis (Victoria Australia) *Members of the ASU Executive Committee ~ ~~~~ Editor JAAS Janice M. Gordon The Royal Society of Chemistry Thomas Graham House Science Park Milton Road Cambridge CB4 4WF UK. Telephone + 44 (0) 1223 420066. Fax + 44 (0) 1223 420247. E-mail RSC1 @RSC.ORG (Internet) Senior Assistant Editor Brenda Holliday Editorial Secretary Lesley Turney US Associate Editor JAAS Dr.J. M. Harnly US Department of Agriculture Beltsville Human Nutrition IResearch Center Beltsville MD 20705 USA. Telephone 301 -504-8569 Assistant Editor Ziva Whitelock Advertisements Advertisement Department The Royal Society of Chemistry Burlington House Piccadilly London W1 V OBN UK. Telephone + 44 (0) 171 -287 3091. Fax +44 (0) 171 -494 11 34. Information for Authors Full details of how to submit materials for publi- cation in JAAS are 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 communi- cations and letters concerned with the development and analytical application of atomic spectrometric techniques.The journal is pub- lished twelve times a year including comprehen- sive reviews of specific topics of interest to practising atomic spectroscopists and incorpor- ates the literature reviews which were previously published in Annual Reports on Analytical Atomic Spectroscopy (ARAAS). Manuscripts intended for publication must describe original work related to atomic spectro- metric analysis. Papers on all aspects of the sub- ject will be accepted including fundamental studies novel instrument developments and prac- tical analytical applications. As well as AAS AES and AFS papers will be welcomed on atomic mass spectrometry X-ray fluorescence/emission spectrometry and secondary emission spec- trometry. Papers describing the measurement of molecular species where these relate to the characterization 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 particularly welcome. Manuscripts on other subjects of direct interest to atomic spectroscopists including sample prep- aration and dissolution and analyte pre-concen- tration procedures as well as the statistical interpretation and use of atomic spectrometric data will also be acceptable for publication. There is no page charge. The following types of papers will be considered. Full papers. describing original work.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 receipt. 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 particular facet of analytical 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 elsewhere 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 spacing) should be sent to Janice M. Gordon Editor JAAS or Dr. J. M. Harnly US Associate Editor JAAS. All queries relating to the presentation and sub- mission 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 JAAS Editorial Board (who may be contacted directly or via the Editorial Office) would welcome comments suggestions and advice on general policy matters concerning JAAS. Fifty reprints are supplied free of charge. Journal of Analytical Atomic Spectrometry (JAAS) (ISSN 0267-9477) is published monthly by The Royal Society of Chemistry Thomas Graham House Science Park Milton Road Cambridge CB4 4WF UK.All orders accompanied with payment should be sent directly to The Royal Society of Chemistry Turpin Distribution Services Ltd. Blackhorse Road Letchworth Herts. SG6 1 HN UK Tel. +44 (0) 1462 672555; Telex 825372 Turpin G; Fax +44 (0) 1462 480947. Turpin Distribution Services Ltd. is wholly owned by The Royal Society of Chemistry. 1995 Annual subscription rate EEA €51 2.00 USA $941 50 Canada €538.00 (+ GST) Rest of World €538.100. Customers should make payments by cheque in sterling payable on a UK clearing bank or in US dollars payable on a US clearing bank. Air freight and mailing in the USA by Publications Expediting Inc. 200 Meacham Avenue Elmont NY 11 003. USA Postmaster send address changes to Journal of Analytical Atomic Spectrometry (JAAS) Publications Expediting Inc. 200 Meacham Avenue Elmont NY 11 003. Postage paid at Jamaica NY 11 431. All other despatches outside the UK by Bulk Airmail within Europe Accelerated Surface Post outside Europe. PRINTED IN THE UK. @The Royal Society of Chemistry 1995. 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.

ISSN:0267-9477    

DOI:10.1039/JA99510FX009

出版商:RSC    

年代:1995

数据来源: RSC

摘要:

Journal of Analytical Atomic Spectrometry CONTENTS NEWS PAGES Dbry of CoWwm@s and Courses Future i.wc0 2813 u n PAPERS INVITED LECTURE AMiytIcal Minimalimn Applled to the Detwmlnrtion of Trace Ehments by Atomic Spectmnetry David J. Hall3 INVITED LECTURE Appllclltion of inducthnly Coupled Plasma Atomk Emiwlon and Mam Spectrometry to Forensic Analyslr of Sodium amma H y d m Butymte and Ephodrkre Hydrochloride Karen A. Wolnik Douglas T. Heitkernper John 6. Crowe Barbara 5. Barnes Thomas W. Brueggemeyer INVITED LECTURE Wudhmly Coupled Plasma in FIwmcence Spectrom Sourn and Atom/ ion Rowrwir Stanley Greenfield Effect of Elevated Oar Pressure on Atomhation in Qnphk Furnace conthwun Source Atomic -on Spectmnetry with Linear phatodiode Array Detecth Clare M. M. Smith James M. Hamiy Effect of Fumrce Atomiution Temperature8 on SimuItaneow Multieiment Atomlc Absorption Measurement Ushg a Tm-Hwtsd araphk AtomltM James M.Harnly Bernard Radduk Knowled#e-baad Computer System for tho Detection of Matrix interferences In Atomlc Abmptbn SpcrcbwneMc Methods Wim Pennincka Peter Vankeerberghen D. Luc Massart Johanna Smeyers-Vetbeke Baterminrtion of Selenium in Human Half 8nd Mall by Ei.ctrothemral Atomic Absorption Spectrometry lain Harrison David Littlejohn Gordon S. Fell COmpHbon of Chemlcal ModMufs W the Determination of Gold In Biological Fluids by Ektrothermd Atomic Abrorption Spectrometry Nikdaocl 8. Thomaidis Efrosini A Piperakl Constantinos E. Efstathiou Studies on &bent ExWaction to Daennkrc iodlde Indirectly by Eledmttm~i Atomic Abwption Spectmncvtry Pilar Bermejo-Earrera Manuel Aboal-Somoza Antonio Moreda-PiWiro Adeta BermejO-Barrera DeWnninrtkn and Speciation of Hewy Matllh in Sedlmmts from the CumMan Coast NW England UK Ulhafii Belazi Christine M.Davidson G l l h E. Keating David Laejohn Martin McCartney On-line Reconcentration of Chromium 010 and Specktion of Chromium in Waten by Flame Atomic Abaorptlon Spectmmetry-Benyamin Pesullean Christine M. Davldson DavM Littlejohn Speclation of Amonlc by the Detennimtlon of Total ArrcMic and Ar#nio (Ill) in Marlne Sodiment Sample by Ektrothermai Atomic Absorption Spectrometry P. Bermejo-Banere M. C. Barciela-Alonso M. Ferr6n-Novais A Bem\eiO-Banera Automatic Wavelength Calibration Procedure for lhe with an Optical Spectromebr and Army Detwtor Daran A sadlet David Littlejohn Charles V.perkina Optlmal Accuracy preokion and 8.nvHlvity of inductivelp CoupW plr#nu oplrlcrl Eml#lon Specbmmrtry BkuMIyah of Aluminium Trevor J. Burden J. J. powelf R.P.H. Thompson P. D. Taylor 8hnliltrmour Moa8umment of kotope Ratios In &ilds by Laser Ablation with a Twln Qwdrupde lnducthrety Coupled Plasma M#. -meter Uoyd A. Allen Ho-ming Pang Amold R. Warren R. S. Houk Resonance loniution Mar0 Tumer . Field Snmpllng Technique for the 'Fast Reacthm' Aluminium Fraction In Watem WIna a Flow InJectIon Mlnl-cohnnn systrrm wlth Inductively Coupled Plasma Atomic Emidon Spectmmetric and inducthre& Coupbd Plasma Mass SpedmmWc Dekctkn Ben Fairman; Alfred0 Sanz-Medel Phil Jones Automated Technique for Mercury Betennlnaion at Sub-nanogram per Utre Luvek In Natural Water8 Daniel Cossa Jane Sanjuan Jacques Cloud Peter B.Stockwll Warren T. Corns CUMULATIVE AUTHOR INDEX Ratio M.aruments in i)trontium Udng Twoghoton Two-dour lndral K. Perera Ian C. Lyon Grenviile 168 177 188 187 197 207 21s 221 - 227 241 247 158 259 267 273 281 287 295r 6th Surrey Conference on Plasma Source Spectrometry I1 St. Helier Jersey UK 17-20 September 1995 Invited Lecturers Dr N Walsh (Royal Holloway) Dr C Gregoire (Geological Survey Canada) Professor R Barnes (University of Massachusetts) and Dr A Gray (Imperial College) Call for Papers Papers (oral and poster presentations) on topics associated with all aspects of plasma source mass spectrometry and on ICP-AES and ICP-MS studies in the Earth Sciences. Three copies of abstracts must be submitted before July 2 8 1 9 9 5 Social Programme An informal reception will take place on the Sunday and a conference dinner on Wednesday evening. An accompanying persons' package is available. Registration The residential package covers all meals coffee tea accommodation in single rooms and registration fee. A reduced fee is available for all bona f i d e students Further Details Dr Kym Jarvis NERC ICP-MS Facility CARE Silwood Park Ascot Berks UK SL5 7TE. Tel +44 ( 0 ) 1 3 4 4 294517/6; Fax +44 ( 0 ) 1 3 4 4 873997

ISSN:0267-9477    

DOI:10.1039/JA99510BX011

出版商:RSC    

年代:1995

数据来源: RSC

摘要:

Journal of Analytical Atomic Spectrometry 1995 Fourth International Conference on Progress in Analytical Chemistry in the Steel and Metals Industry May 16-18 Jean Monnet Building Luxembourg Details can be found in J. Anal. At. Spectrom. 1994 9 50N. For details of providing a contribution to the programme or other information contact CEC/CETAS Conference R. Jowitt British Steel plc Technical Teesside Laboratories PO Box 11 Grangetown Middlesbrough Cleveland TS6 6UB. Telephone + 44 642 467144; Fax +44 642 460321. 43rd ASMS Conference on Mass Spectrometry and Allied Topics May 21-25 Atlanta GA USA For further details contact ASMS 1201 Don Diego Avenue Santa Fe NM 87501 USA. Telephone 505 989 4517; Fax 505 989 1073. 5th Annual Flow Injection Atomic Spectroscopy Short Course June 6-8 Amherst Massachusetts USA Details can be found in J.Anal. At. Spectrom. 1994 9 68N. For further information contact Julian F. Tyson Department of Chemistry Lederle GRC Tower University of Massachusetts Box 34510 Amherst MA 01003-4510 USA. Telephone (413) 545 0195; Fax (413) 545 4846. Short Course. High-Performance Liquid Chromatography Loughborough UK Details can be found in J. Anal. At. Spectrom. 1995 10 18N. For further details contact Mrs. S. J. Maddison Department of Chemistry University of Technology Loughboro- ugh Leics. LEll 3TU. Telephone (01509) 222575 222563; Fax (01509) 233 163. July 3-7 SAC 95 July 9-15 Hull UK Details can be found in J. Anal. At. Spectrom. 1995 10 13N. For further information contact The Secretary Analytical Division The Royal Society of Chemistry Burlington House Piccadilly London W1V OBN UK.Vth COMTOX Symposium on Toxicology and Clinical Chemistry of Metals July 10-13 University of British Columbia Vancouver British Columbia Canada Details can be found in J. Anal. At. Spectrom. 1994,9 26N. For further information contact F. William Sunderman Jr. MD Department of Laboratory Medicine University of Conneticut Medical School Room C-2050 263 Farmington CT 06030-2225 USA. Telephone 203-679-2328. 13th Australian Symposium on Analytical Chemistry. In conjunction with 4th Environment Chemistry Conference - Chemistry in Tropical and Temperate Environments July 9-14 Darwin Northern Territory Australia Details can be found in J. Anal. At. Spectrom. 1995 10 19N. For further information contact Dr.Brian Salter-Duke Secretary 13AC/4EC Organizing Committee RACI GPO Box 363 Darwin NT 0801 Australia. 41st International Conference on Analytical Sciences and Spectroscopy August 14-16 Windsor Ontario Canada The 41st International Conference on Analytical Sciences and Spectroscopy sponsored by the Spectroscopy Society of Canada (SCC) and the Canadian Society for Mass Spectrometry (CSMS) will be held from August 14 to 16 1995 at the University of Windsor Windsor Ontario Canada. Contributions from all areas of analytical sciences (encompassing chromatography thermal analysis electrochemistry etc.) and spectroscopy (including atomic molecular electronic vibrational and rotational magnetic mass surface etc.) are invited. A plenary lecture entitled ‘Photochemistry in the Adsorbed State Using Light as a Scalpel and a Crystal as Operating Table’ will be presented by Professor John Polanyi.For more information contact Dr William E. Jones. Telephone (519) 253 4232 ext 2001; Fax (519) 973 7098. The Third Asian Conference on Analytical Sciences ASIANALYSIS 111 August 20-24 Seoul Korea Details can be found in J. Anal. At. Spectrom. 1995 10 18N. For further details contact Prof. Hasuck Kim (Secretariat) ASIANA- LYIS 111 Department of Chemistry College of Natural Sciences Seoul National University Seoul 151-742 Korea. Telephone + 82(2)880- 6638; Fax + 82(2)889-1568; E-mail hausukim- @KRSNUCCl.BITNET Colloquium Spectroscopicum Internationale (CSI) XXIX August 27-September 1 Leipzig Germany Details can be found in J. Anal. At. Spectrom.1993,8 50N. For further details contact Prof. Dr. H. Nickel Forschungszentrum Jiilich GmbH Institut fur Werkstoffe der Energietechnik/RWTH Aachen D- 52425. Telephone (02461) 61 55 65; Fax (02461) 61 36 99. Colloquium Spectroscopicum Internationale (CSI) XXIX Post Symposium ICP-MS September 1-4 Wernigerode/Hartz Germany Details can be found in J. Anal. At. Spectrom. 1994 9 46N. For further details contact Dr. L. Moenke Martin-Luther University Halle-Wittenberg Department of Chemistry Institute of Analytical and Environmental Chemistry Weinbergurg 16 D-06120 Halle Germany. Fax 0049-345-649065. Euroanalysis IX Sep tem ber 1-7 Bologna Italy Details can be found in J. Anal. At. Spectrorn. 1995 10 14N. Journal of Analytical Atomic Spectrometry March 1995 Vol.10 23NFurther information is available from Professor Luigia Sabbatini Euroanalysis IX Dipartimento di Chimica Universita di Bari Via Orabona 4 70126 Bari Italy. 8th International Conference on Coal Science September 10- 15 Instituto Nacional del Carbdn CSIC Apartado 73 33080 Oviedo Spain Details can be found in J. Anal. At. Spectrom. 1994 9 61N. For further details contact Dr. Juan M. D. Tascon 8th ICCS Scientific Programme Chairman Ins titu to Nacional del Carbon CSIC Apartado 73 33080 Oviedo Spain. Telephone + 34.8.528.08.00; Fax + 34.8.529.76.62. Sixth Surrey Conference on Plasma Source Spectrometry September 17-20 Jersey UK Details can be found in J. Anal. At. Spectrom. 1995 10 19N. For further details contact Dr. K. Jarvis NERC ICP-MS Facility Centre for Analytical Res.in the Environment (CARE) Imperial College at Silwood Park Buckhurst Road Ascot Berkshire SL5 7TE UK. Telephone +44(0) 344 294517; Fax +44(0) 344 873997. European Workshop in Chemometrics September 17-22 Bristol. UK The University of Bristol will hold its annual 1995 European Workshop in Chemometrics in Bristol 17-22 September 199 5. For further details contact Mrs. C. Hutcheon School of Chemistry University of Bristol Contock’s Close Bristol BS8 lTS UK. Telephone + 44(0) 117-928 7645 ext. 4221; Fax + 444 0) 1 17-925 1295. Federation of Analytical Chemistry and Spectroscopy Societies Conference October 15-20 Cincinnati Ohio USA Details can be found in J. Anal. At. Spectrom. 1995 10 19N. For further information contact Joseph A. Caruso FACSS National Office 198 Thomas Johnson Dr.Suite S-2 Frederick MD 21702 USA. Telephone (301) 694-8122; Fax (301) 694-6860. First Mediterranean Basin Conference on Analytical Chemistry November 5-10 Cbrdoba Spain For further details contact Prof. Alfred0 Sanz-Medel Department of Physical and Analytical Chemistry Faculty of Chemistry. University of Oviedo C/ Julian Claveria no 8. 3006 Oviedo (Spain). Telephone 34/85/ 103474-103485; Fax .34/85/103480. Biological Applications of Inorganic Mass Spectrometry November 8-9 Norwich UK Details can be found in J. Anal. At. Spectrom. 1995 10 20N. For further information contact Dr. Fred Mellon Institute of Food Research Norwich Laboratory Norwich Research Park Colney Norwich NR4 7UA UK. Telephone +44(0)1603 255 299 (direct line) +44 (0) 1603 255 000 (switchboard/paging); Fax +44 (0)1603 452578 +44 (0)1603 fred.mellon@BBSRC.AC.UK. 507723; E-MAIL International Symposium on Environmental Biomolnitoring and Specimen Banking December 17-22 Honolulu Hawaii USA Details can be found in J. Anal. At. Spectrom. 1994,9 59N. For further information contact K. S. Subramanian Enviroinmental Health Directorate Health Canada Tunney’s Pasture Ottawa Ontario K1A OL2 Canada (phone 613-957-1874; fax 613-941-4545) or G. V. Iyengar Center for Analytical Chemistry Room 235 B 125 National Institute of Standards and Technology Gaithersburg MD 20899 USA (Telephone 301-975-6284; Fax 301-921-9847) or M. Morita Division of Chemistry and Physics National Institute for Environmental Studies Japan Environmental Agency Yatabe-Machi Tsukuba Ibaraki 305 Japan (Telephone 8 1-298-5 1-6 1 1 1 ext.260; Fax 8 1-298-56-4678). 1996 1996 Winter Conference on Plasma Spectrochemistry January 8-13 Fort Lauderdale Florida USA Details can be found in J . Anal. At. Spectrom. 1994,9 53N. For further information contact Dr. R. Barnes ICP Information Newsletter Department of Chemistry Lederle GRC Towers University of Massachusetts Box 34510 Amherst MA 01003-4510 USA. Telephone (413) 545 2294; Telefax (413) 545 4490. International Schools and Conferences on X-Ray Analytical Methods January 18-25 Sydney Australia Details can be found in J. Anal. At. Spectrom. 1994,9,47N. For further information contact AXAA ’96 Secretariat GPO Box 128 Sydney NSW 2001 Australia. Telephone 61 2 262 2277; Fax 61 2 262 2323; Telex AA 176511 TRHOST. Analytica Conference 96 April 23-26 Munich Germany Details can be found in J . Anal. At. Spectrom. 1994,2,69N. For further information contact Messe Miinchen GmbH Messegelande D-80325 Munchen Germany. Telephone +49 89 51 07-0; Telex 5 212 086 ameg d; Fax +49 89 51 07-177. 24 N Journal of Analytical Atomic Spectrometry March 1995 1~01.10

ISSN:0267-9477    

DOI:10.1039/JA995100023N

出版商:RSC    

年代:1995

数据来源: RSC

摘要:

FUTURE ISSUES WILL INCLUDE- Electrothermal Atomic Absorption Spectrometric Determination of Ultratrace Amounts of Tellurium Using Palladium-coated L'vov Platform After Separation-Concentration by Hydride Generation and Liquid Anion Exchanger-Marco Grotti Ambrogio Mazzucotelli Improved Thallium Hydride Generation Using Continuous Flow Methodologies-Les Ebdon Phillip Goodall Steve J. Hill Peter B. Stockwell Clive K. Thompson Determination of Trac:e Elements in Food Contact Polymers by Semi- quantitative Inductively Coupled Plasma Mass Spectrometry and Microwave Digestion Sample Pre- treatment. Performance Evaluation Using Alternative Multi-element Techniques and In-house Polymer Reference Materials-Peter J. Fordham Laurence Castle Helen M. Crews John 24 N Journal of Analytical Atomic Spectrometry March 1995 1~01.10W.Gramshaw Diana Thompson Susan J. Parry Ed McCurdy Direct Determination of Micro Aluminium in High-Purity Tin by Electrothermal Atomic Absorption Spectrometry-Zang Yanfu Zhang Ke Fang Zheng Wang Yunzhou Excitation Temperature and Analytical Parameters for an Ethanol Loaded Inductively Coupled Plasma Atomic Emission Spectrometer-Robert I. McCrindle Cornelius J. Rademeyer Synthesis and Application of an Inert Type of 8-Hydroxyquinoline-Based Chelating Ion Exchanger for Sea Water Analysis Using On-Line Inductively Coupled Plasma Mass Spectrometric Detection-Andreas Seubert G. Petzold J. W. McLaren Determination of Lead in Whole Blood with Inductively Coupled Argon Plasma Mass Spectrometry Using Isotope Dilution-Daniel C. Paschal Kathleen L.Caldwell Bill G. Ting Determination of Fission Products and Actinides in Spent Nuclear Fuels by Isotope Dilution Ion Chromatography Inductively Coupled Plasma Mass Spectrometry-Alonso Garcia Ignacio Jose Fabrizio Sena P. H. Arbore Maria Betti Lothar Koch Glucose as a Chemical Modifier for the Determination of Antimony and Selenium by Electrothermal Atomic Absorption Spectrometry-M. T. Perez- Corona M. B. De La Calle-Guntinas Yolanda Madrid C. Camara Use of Thiourea in the Determination of Arsenic Antimony Bismuth Selenium and Tellurium by Hydride Generation Inductively Coupled Plasma Atomic Emission Spect rome t r y-Hilde Uggerud Walter Luad Rapid Determination of Chromium Electrothermal Atomic Absorption Spectrometry after Microwave Assisted Digestion of Different Samples-Ruma Chakraborty Arabinda K.Das Luisa M. Cervera Miguel De La Guardia In Viuo Sample Uptake and On-line Measurements of Cobalt in Whole Blood by Microwave-assisted Mineralization and Flow Injection Platform Electrothermal Atomic Absorption Spectrometry-J. L. Burguera M. Burguera C. Rondon C. Rivas P. Carrero M. Gallignani M. R. Brunetto COPIES OF CITED ARTICLES The Royal Society of Chemistry Library can usually supply copies of cited articles. For further details contact The Library Royal Society of Chemistry Burlington House Piccadilly London W1V OBN UK. Tel +44 (0) 71-437 8565; fax f44 (0) 71-287 9798; Telecom Gold 84; BUR210; Electronic Mailbox (Internet) LIBRARY @RSC.ORG. If the material is not available from the Society’s Library the staff will be pleased to advise on its availability from other sources.Please note that copies are not available from the RSC at Thomas Graham House Cambridge. Journal of Analytical Atomic Spectrometry March 1995 VoE. 10 25 NRamon M. Barnes Editor Department of Chemistry LGRC Towers University of Massachusetts Am herst MA 01 003-0035 Telephone (41 3) 545-2294 fax 545-4490 Objective The ICP INFORMATION N€WSL€TT€R is a monthly journal published by the Plasma Research Group at the University of Massachusetts and is devoted exdusively to the rapid and impartial dissemination of news and literature information re- lated to the development and applications of plasma sources for spectrochemical analysis. Background ICP stands for inductively coupled plasma discharge which during the past decade has become the leading spectrochemi- cal excitation source for atomic emission spectroscopy.ICP discharges also are applied commercially as an ion source for mass spectrometry and as an atom and ion cell in atomic fluo- rescence spectrometry. The popularity of this source and the need to collect in a single literature reference all of the pertinent data on ICP stimulated the publication of the ICP INFOR- MATION NEWSLETTER in 1975. Other popular piasma sources Le. microwave induced plasmas direct current plasmas and glow discharges also are included in the scope of the ICP IN- FORMATION NEWSLETTER. Scope As the only authoritative monthly journal of its type the ICP INFORMATION NEWSLETTER is read in more than 40 coun- tries by scientists actively applying or planning to use the ICP or other types of plasma spectroscopy.For the novice in the field the tCP INFORMATION NEWSLETTER provides a conase and systematic source of information and background material needed for the selection of instrumentation or the development of methodology. For the experienced scientist it offers a sin- gle-source reference to current developments and literature. Editorial The ICP INFORMATION NEWSLETTER is edited by Or. Ramon M. Barnes Professor of Chemistry University of Mas- sachusetts at Amherst with the assistance of a 20-member Board of National Correspondents composed of leading plasma spectroscopists. The Board members from around the world report news viewpoints and developments. Dr. Barnes has been conducting plasma research on ICP and other dis- charges since 1968.He also serves as chairman of the Winter Conference on Plasma Spectrochemistry sponsored by the I'CP INFORMATION NEWSLETTER. Regular Features oOriginal submitted and invited research articles by ICP and (Complete bibliography of all major ICP publications. oAbstracts of all ICP papers presented at major US and inter- ofirst-hand accounts of world-wide ICP developments. 4pecial reports on dcp microwave glow discharge and other (Calendar and advanced programs of plasma meetings. oTechnica1 translations and reprints of critical foreign-lan- guage ICP papers. Gritical reviews of plasma-related books and software. Conference Activities The ICP INFORMA TION N€WSLETT€R has sponsored seven international meetings on developments in atomic plasma spectrochemical analysis since 1 980 in San Juan Orlando San Diego St.Petersburg and Kailua-Kona. Meeting pro- ceedings have appeared as Developments in Atomic Plasma :3pectrochemkaI Analysis (Wiley) Plasma Spectrochemistry and Plasma Spectrochemistry /I-IV (Pergamon Press) as well as in special issues of Spectrochimica Acta Part B and Journal of Analytical Atomic Spectrometry. The 1994 Winter Confer- ence on Plasma Spectrochemistry will be held in San Diego California January 10 - 15 1994; its proceedings will be published by Fall 1994. Subscription Information !Subscriptions are available for 12 issues on either an annual or volume basis. The first issue of each volume begins in June and the last issue is published in May.For example Volume 18 runsfrom June 1992 through May 1993. Back issues beginning with Volume 1 May 1975 also are available. To begin a subscription complete the form below and submit it with prepayment or purchase information. For additional informa- tion please call (41 3) 545-2294 fax (41 3) 545-4490 or contact the Editor. Credit cards accepted. plasma experts. national meetings. plasma progress. To order complete this section and send it to ICP Information Newsletter %Dr. Ramon M. Barnes Depart- ment of Chemistry Lederle GRC Towers University of Massachusetts Amherst MA 01 003-0035 USA. Start a subscription for the following issue 0 Volurne(s)- (June 19- - May 19- ) or 0 19 (January - December). Enclosed 13 Prepayment 0 Check or money order QVISA D MasterCard Account No. (All 13 or 16 digits) ) or CJ Send invoice. Date Cardholder Name Expiration date Cardholder Signature .Amount Due $ Mail to; Name Organizaticm Address City State/Country- ZI P/Postalcode Telephone Telex/f ax Note For each credjt-card transaction a 4 % service charge will-b added reflecting our bank charges. Current subscription rates are $60 (North America) $85 (Europe South America) or $94 (Africa Asia Indian/Pacific Ocean Areas Middle East and Russia). Back issue rates available on request. All payments should be made with US dollars by draft on a US bank by international money order or by credil card. Foreign bank checks are not accepted. ,13 Purchase order (No.

ISSN:0267-9477    

DOI:10.1039/JA995100024N

出版商:RSC    

年代:1995

数据来源: RSC

摘要:

Analytical Minimalism Applied to the Determination of Trace Elements by Atomic Spectrometry* Journal of Analytical Atomic Spectrometry Invited Lecture DAVID J. HALLS Trace Element Unit Institute of Biochemistry Glasgow Royal Injrmary University NHS Trust Castle Street Glasgow UK G4 OSF In analytical minimalism each stage of the analysis is evaluated to minimize the time cost sample requirement reagent consumption energy requirements and production of waste products. These parameters are often inter-related. If the objective of digestion of biological tissues foodstuffs and environmental samples is taken as the complete dissolution of the trace elements then the time of digestion by conventional heating can be reduced considerably to times comparable with pressure digestion using microwave heating.The development of rapid and simple partial digestion techniques is reviewed. In electrothermal atomic absorption spectrometry by assessing the function and time of each stage in the programme it has been possible to reduce the programme time to about 30 s for a number of determinations. Recent developments in fast furnace technology are reviewed particularly on omission of the ashing stage and drying with hot injection or high temperatures. With reduction in furnace programme time the time taken by the autosampler (30-35 s) becomes dominant. Developments to reduce this time by 10-20 s are discussed. In the evaluation of results minimal time and effort by the analyst is ensured by customized computer programmes. The programmes variants of one or two basic programmes are adapted for each determination to retain the value of standards to correct for blanks and to allow conversion from g I-' to mol I-'.Keywords Partial digestion; electrothermal atomic absorption spectrometry; fast furnace analysis; clinical and biological samples; analytical minimalism The aim of analytical minimalism is to keep analytical processes as simple as possible with the minimum consumption of resources. Parameters that should be minimized are time cost sample requirements reagent consumption and waste pro- duction. Reduction in time is of particular relevance as this increases throughput and reduces costs. As will be apparent many of these parameters are inter-related. For example a reduction in sample volume generally leads to a reduction in reagent consumption and production of waste products.All processes carried out in the laboratory from sample handling to calculation of results and report generation need to be considered in this way. The purpose of this paper is to demonstrate some aspects applied to the determination of trace elements by atomic spectrometry particularly in clinical and biological samples and to review progress in fast furnace analysis in electrothermal atomic absorption spectrometry ( ETAAS). * Presented at the Seventh Biennial National Atomic Spectroscopy Society (BNASS) Hull UK July 20-22 1994. SAMPLE PREPARATION Sample Amount In the clinical field it is becoming more and more relevant to work with small amounts of sample. Commercial multichannel analysers for general clinical chemistry produce a wide range of results from a small volume of sample so that clinicians expect more specialized tests to be produced from a small quantity of blood.This is imperative for samples from neonates and infants from whom the total blood volume sampled must be limited. Athough determination of copper in blood plasma from adults is more economically and rapidly carried out by flame atomic absorption spectrometry ( FAAS) for samples from neonates and infants ETAAS is more appropriate as the sample volume can be limited to 20p1.192 The determination is particularly relevant for premature infants who have limited copper stores because most of the copper stores in the infant are laid down in the third trimester of pregnan~y.~ Failure of the neonate to absorb sufficient copper in its early life can lead to brittle bone formation as a result of a failure to produce sufficient collagen and elastin cross-linking.4 Measurement of iron and copper in the liver is invaluable in confirmatory diagnosis of haemochromatosis and Wilson's disease respectively.Biopsy samples taken by needle may weigh as little as 2-4mg after drying. Care is needed in handling the specimen (particularly to avoid losing it!). Once dissolved in acid determination is straightforward by FAAS for iron or ETAAS for copper (Fig.1). Small scale working was necessary in the development of a method for the determination of arsenic in hair by ETAAS.' Dry sample overnight at 90 "C Weigh sample into acid-washed glass tube (1 0x75 mm) Add 0.4 ml nitric acid Heat at 90 "C for 20 min Add1 ml water Add 1.6 mi water Measure by ETAAS Measure by FAAS Fig.1 Procedure for the determination of Cu or Fe in liver biopsies. For each determination blanks and a reference material BCR No. 165 Bovine Liver are put through the same procedure Journal of Analytical Atomic Spectrometry March 1995 Vol. 10 169As this was intended to replace a previous method using neutron activation analysis the sample requirement of 3-10 mg of hair preferably had to be retained. To obtain adequate sensitivity the sample was dissolved in 100 pl nitric acid taken to dryness and re-dissolved in 100 pl of 1% v/v nitric acid. This was mixed with 100 pl of palladium modifier (1 mg 1-l) for analysis. Digestion of Samples In recent years advances have been made in reduction of the time of preparation of solid samples by slurry ampl ling^,^ and bomb digestion with microwave heating.8 Applications to the analysis of clinical and biological samples have been reviewed in recent Atomic Spectrometry Updates.'-'' In most publi- cations comparison is generally made with classical wet diges- tion techniques which take several hours to complete.Most of the wet digestion techniques date back to the days of colorimetric analysis when it was essential to remove all traces of organic material.' Even recently these objectives have been seen as important and have been studied for a range of digestion procedure^.'^.'^ For most atomic spectrometric analysis a more relevant objective is the complete dissolution of the trace elements to be measured. Many practical analysts have realized that it is possible to achieve this with simpler procedures.Extraction with dilute acid at room temperature is effective at releasing Ca Cd K Mg Mn Na Pb and Zn from biological t i ~ s u e ' ~ ' ~ but not C U ~ ~ ' ' or Fe.16-18 Luterotti et al." determined Cu Mn and Zn in rat liver by homogenizing the material with five times the mass of water adjusting the HC1 concentration to 1 mol l-' shaking for 30 rnin and centrifuging the resulting mixture. The elements were deter- mined in the supernatant by FAAS. Treatment at a higher temperature speeds up extraction." Price" quotes a procedure by Premi and Cornfield22 in which Cu Cr Fe Mn and Zn were quantitatively extracted from plant materials and sewage sludge by boiling with 6 moll-' HC1.Using citrus leaves as a model for plant material Bassen and Bohmer2' found that Ca Cu K Mg Mn and Zn were completely extracted after heating with 3 moll-' HC1 for 15 rnin but Fe was not. Kuennen et aL2 found that moderate pressure produced by digesting samples with 6 moll-' HCl in capped polyethylene bottles at 80 "C for 30min was effective in extracting a range of elements from agricultural crops including Fe. After an evaluation of various hot-plate procedures and room temperature digestion for bovine liver Asp and Lund" recommended a simple procedure based on heating with HN03 for 1-2 h followed by filtration of undissolved material. Accurate results were obtained for Cd Cu Fe Mn Mo and Zn in certified reference materials on determination by inductively coupled plasma atomic emis- sion spectrometry (ICP-AES).minimization of the time required for complete extraction of the elements Cu Fe Mn and Zn from biological tissues was addressed. For Community Bureau of Reference (BCR) Bovine Liver and Pig Kidney reference In a recent materials (RMs) and for fresh bovine liver these elements were completely extracted after heating with HNO at 105°C for only 20min. Digests were centrifuged after dilution and the elements determined by FAAS ETAAS or ICP-AES. Good agreement was shown with results by complete digestion and results on RMs were in good agreement with certified values. This approach has also been used successfully in the determi- nation of the same four elements in a range of foodst~ffs.~~ A similar approach was used by Zima et aLZ6 for the determi- nation of Cd and Pb in animal tissues.After boiling the samples with HNO for 15 min the digests were made up to volume with ethanol to prevent undigested lipids from precipitating. In Table 1 a comparison is made on the basis of the minimization parameters discussed earlier between the partial digestion technique and a rapid microwave bomb digestion procedure described by Van Wyck et al.27 for the determination of iron in tissue. Although the heating time is five minutes longer by partial digestion the cooling time is considerably shorter and the overall time is shorter. Undoubtedly the time by microwave digestion could be shortened but the important point is that the times are comparable.Microwave heating is more efficient in energy consumption than a block heater because pre-heating is unnecessary. In cost of reagents and production of waste products there is very little difference. However the capital cost of a laboratory microwave digestion system with bombs far exceeds the cost of a block heater. If the bomb digestion is carried out in a domestic microwave oven the capital cost of equipment is less than a block heater. Whether this is desirable because of safety considerations is a contentious which will not be pursued here. The important conclusion is that the partial digestion approach offers an alternative to the cost and complexity of a laboratory microwave digestion system. Another obvious inference is that a combination of the partial digestion approach with micro- wave heating would be more energy efficient.An indication of the potential of this is seen in the work of Blust et al.,,' in which brine shrimp samples were digested directly in polypropylene autosampler cups with 100 pl HNO for 5 min with microwave heating. Water (1 ml) was added and the samples analysed by ETAAS. This approach of partial digestion has been applied to the routine determination of Cu and Fe in liver biopsies (Fig. 1) in order to reduce digestion time. It has been found possible to omit the centrifugation step used in the preliminary further simplifying the procedure. DETERMINATION BY ETAAS One of the drawbacks of conventional determination by ETAAS is the prolonged time required to produce a result.In 1983 Bahreyni-Toosi and Dawson,' produced a miniature graphite furnace with Zeeman-effect background correction to allow determinations in a single step. Although the objective of a single step has not been achieved since the work did show Table 1 Comparison of microwave bomb digestion and partial digestion for preparation of tissue samples for determination of iron Procedure Time Energy consumption Waste products Approximate capital cost (1994 prices) Reagent cost and consumption Microwave bomb digestion2' 250 mg sample + 1 ml HNO + 1 ml Hz02 15 min at 25% power 30 rnin cooling Total 45 rnin 0.150 kW h-' Fumes on opening bomb &10000 Very low Partial digestion24 100-200 mg sample + 2 ml HNO 20 rnin at 105 "C 10 min cooling Total 35 rnin 0.167 kW h-' +0.250 k W h-' for preheating Fumes during digestion $650 Very low 5 min centrifugation 170 Journal of Analytical Atomic Spectrometry March 1995 Vol.10the potential time saving of this technique. Practical steps to reduce programme time with conventional furnaces then fol- lowed.32 The purpose of this section is to review progress that has been made over 10 years. Developments are discussed under the four main steps in a furnace programme drying ashing atomization and clean. A further section covers the time taken by the autosampler. Drying Stage In conventional programmes the time allocated to drying can be quite extensive. Three approaches have been tried to reduce the time in the drying stage; minimization of the time in conventional drying injection of the sample onto a preheated tube ('hot injection') and high-temperature drying.Conventional drying Most graphite furnace systems display temperature and allow ramped temperature programming. The temperature displayed is in most cases a function of the power applied to the tube and is not a reading of actual temperature. Because of the thermal mass of the graphite the actual temperature will lag behind the programmed temperature when ramping. Studies32 of the actual temperature attained in a drop of water on the surface of an uncoated graphite tube showed that ramp times of 1-7 s made no difference to the shape of the temperature rise seen. Distinct temperature ramps were seen at programmed ramp times of 10 and 20 s. As the natural temperature response provided ramping of around 100-140 "C the applied tempera- ture ramp was set to the minimum time (1 s) and drying times of 7-12 s were used.Pyrolytic graphite coated graphite surfaces are more hydrophobic and more care is needed in drying. Table 2 shows minimized drying stages suitable for simple matrices for the three types of graphite surface commonly used. Since temperature settings vary between instruments some experimentation with the applied temperature setting may be necessary to obtain satisfactory drying. Although atomization from pyrolytic graphite coated tubes or from platforms may offer higher sensitivity and possibly lower interferences for many applications uncoated graphite tubes provide adequate sensitivity and freedom from interferences and have the advantage of simpler and more rapid drying stages particularly with serum and ~ r i n e .~ ~ ~ Hot injection The ability to dry samples almost instantaneously by injecting samples onto a pre-heated tube is now available on a range of graphite furnace systems. Control of the pipetting speed is important as the rate of sample introduction needs to be slowed down so that the sample is almost immediately dried as it contacts the tube surface which is kept at temperatures of 100-120°C. If the ashing stage is omitted the injection needs to be followed by a purge stage to ensure that water vapour is completely removed before atomization. K n ~ w l e s ~ ~ applied hot injection to the determination of Cr Cu and Ni in water Cd and Cr in urine and Cu in serum thus reducing programme times to < 20 s.Procedures for the determination of Al Cu and Pb in waters and Cu Fe Ni and Pb in biological RMs were developed by Kunwar et a1.35,36 The technique is Table 2 Drying stages for simple matrices ideally suited for the analysis of chelates extracted into isobutyl methyl ketone (IBMK) as Apostoli et al.37 demonstrated for the determination of V in urine by extraction with cupferron. Comparison with conventional drying showed an improvement in precision and a three-fold improvement in sensitivity which was presumed to be due to reduced formation of vanadium carbide. The versatility of the hot injection approach has been shown in a series of application notes from Varian which include the determination of Ag As Bi Cd Pb Sb and Se in stainless Cd and Ni in shellfish tiss~e;~' Cu in urine;42 Pb T1 Bi Cd and Sn in high purity sulfuric acid43 and Se in blood and urine.44 A further approach which has an even longer history is to spray an aerosol of the sample onto a pre-heated tube.The system originally developed by Matousek4' and marketed by Thermo Jarrell Ash as the Fastac system is based on a pneumatic nebulizer of the type used in FAAS. A disadvantage is that not all the sample reaches the atomizer as in FAAS and hence sample volume requirements are greater with this technique. This is overcome by thermospray introduction as Bank et u1.46,47 have shown. In their system a fixed volume (normally 10 pl) of sample was introduced into a flow injection system which pumped the sample through a heated fused silica capillary and deposited the thermospray onto a pre-heated graphite tube or platform.To avoid condensation of water vapour on the windows of the furnace an evacuation system was necessary to remove the water vapour generated. As Cu and Pb in ~eawater;~' Cd in High temperature drying A third approach is to apply indicated temperatures much higher than would normally be thought acceptable for drying. If conditions are chosen correctly drying occurs without sputtering and results with good precision are obtained. For wall atomization on uncoated tubes it was found possible to apply drying temperatures of up to 260°C without loss of precision. Methods for the determination of aluminium and lead in waters were developed without an ashing stage using a drying temperature of 250"C.48 The use of high drying temperatures also helped to expel water vapour rapidly from the tube which was shown to be important to eliminate interference in the determination.For platform atomization Slavin et aL4' heated the platform rapidly by applying an indicated temperature of 700 "C for 1 s and then dried at 400°C. The total drying time was 18 s. The minimum time and temperature for drying of slurries on pyrolytic graphite platforms was studied systematically by Hinds et aLS0 For drying temperatures in the range 200-500 "C the minimum time required for drying was established for volumes of 10 and 20 1-11 of slurry. The time decreased with an increase in temperature and a decrease in sample volume. Temperatures above 400 "C resulted in shoulders and double peaks (probably owing to sputtering).The drying stage chosen for 20 pl aliquots was an applied temperature of 400 "C with a ramp time of 1 s and a hold time of 20 s. These findings were confirmed by Lopez Garcia et aL5' for the analysis of slurries of diatomaceous earths. They reduced the drying hold time to 15s. In all of these cases it seems unlikely that the temperature actually reaches the applied temperature setting. The greater Graphite tube Temperature/"C Ramp time/s Hold time/s Example Ref. Uncoated 140 Pyrolytic graphite coated 130 Platform with pyrolytic graphite coated 160 7 Cu in urine 32 15 Cr in plant digests 33 15 Pb in plant digests 33 Journal of Analytical Atomic Spectrometry March 1995 Vol. 10 171power applied to the furnace causes a more rapid rise in temperature which allows drying of the sample in a very short time probably when the surface of the tube or platform reaches z 100 "C.Ashing or Pyrolysis (Charring) Stage In a conventional heating programme this stage can take a substantial proportion of the total programme time. The functions of an ashing stage are summarized in Table 3. If the ashing stage achieves none of these functions then it is unnecess- ary and can be omitted. For samples with high organic content omission of the ashing stage causes generation of smoke in the atomization stage which obscures and scatters the incident light giving rise to high background absorbance. Biological food or environmental samples which have been digested or blood samples which have been deproteinized have had their organic content largely removed which makes it possible to consider omission of the ashing stage.The ability of an ashing stage to remove compounds that give rise to background absorbance or chemical interferences is often very limited. Except for measurements at very low wavelength (e.g. for As and Se) background absorbance for most biological samples is from sodium and potassium chlorides which cannot be removed until temperatures of > 1OOO"C are achieved. For determinations such as Cu in urine,32 for which the maximum ashing temperature is 900"C ashing makes no difference to the background absorbance or the matrix effect and can be omitted. For others (e.g. Cr in urine52) the background absorbance without an ashing stage is still within the range correctable with a D2-arc system.The use of Zeeman-effect background absorbance offers even greater possibilities for omission of the ashing With respect to the last function of the ashing stage (Table 3) matrix modifiers should be avoided unless really necessary. Slavin et al. have shown that it is possible to get accurate analyses without a modifier and without an ashing stage for a range of digested4' and slurried samples.53 Under these conditions the peak shapes for the same element vary more greatly from matrix to matrix than with conventional programmes with modifiers as Hoenig and Cilissen They concluded that the use of integrated absorbance measurement for volatile elements was necessary under these conditions. Table 4 lists published fast furnace methods showing many examples where the ashing stage could be omitted.When the ashing stage cannot be omitted then the ramp and hold time should be minimized. After 30s very little further change in background absorbance occurs and times in excess of this are rarely necessary. It is only necessary to reduce the background absorbance to a level that the correction system can cope with. For many samples the main function is to remove organic material and the time can be kept short (e.g. Cu in serum2). In many graphite furnace systems the gas flow is automati- cally reduced at a fixed time interval before atomization begins in order to stabilize gas flow. As this normally means zero or very low gas flow products vaporized in this time will not be swept out of the tube before the atomization stage begins and could interfere in the atomization.This needs to be borne in mind when minimizing the ashing time. When the ashing stage is omitted this affects the drying stage. Thus with Perkin- Elmer graphite furnaces (except the HGA 500) programmed Table 3 Functions of a charring or ashing stage 1. To remove organic material. 2. To remove compounds which give rise to background absorption. 3. To remove compounds giving rise to matrix effects. 4. To facilitate reaction with a modifier. for zero gas flow in atomization gas flow is switched off 5 s before the start of the atomization stage and so 5 s needs to be added to the drying stage in Table 2 when the ashing stage is omitted.32 Atomization Stage This measurement step is normally one of the shortest stages in the furnace programme.The time required is evident from the video display of the peak shape and so there is little that needs to be changed from current practice. Clean Stage In procedures developed in this laboratory the clean step has been retained but reduced to a time of 3 s at 2700"C,2,32,48 except when there was evidence of carryover in which case the time was increased to 5 s.52 Lopez Garcia et aLs1 omitted the clean step from their programmes for the analysis of slurries of diatomaceous earths. On testing omission of this stage for the determination of lead they found no loss of sensitivity or precision but a slight increase in background signal over the first 6-7 injections finally reaching a plateau. On the basis of their experience it would seem worthwhile examining for all determinations whether this stage is neces- sary. Certainly for atomization at high temperatures (2500-2700 "C) a further step at a similar temperature may not achieve much more in matrix removal.For refractory elements however it seems unlikely that the clean step could be omitted without introducing carryover. Sampling Time As it became possible to shorten furnace programmes it became obvious that the time taken by the autosampler also needed to be reduced if further reduction in time was to be made. Times taken by some commercial autosamplers are shown in Table 5; separate addition of modifier increases these times. Varian recommended that a volume of blank should be taken up before the sample allowing the blank to wash out the sample.65 Depending on the relative positions of the sample and blank times of between 38 and 42 s were measured.For most samples this is unnecessary and the time then reduces to 33 s. In all these examples the processes of rinsing the tip and injection start only when the furnace programme is complete. Measurement of the temperature decrease of the surface of a L'vov platform in a tube after heating at 2500°C showed33 that the tube had cooled down to under 100°C after about 20s allowing a potential reduction in time of about 10s. A device was built for Perkin-Elmer AS1 and AS40 samplers which allowed them to be controlled by the furnace program- mer so that the sampler could be started before the furnace programme had ended.Details of this have been published,66 showing that time savings of 15 s are possible with an uncoated tube and that using a more conservative 10s saving gave satisfactory results in routine use. Further trials of routine operation with a time saving of 14 s per injection are proving promising. Reduction in sampling time really needs to be considered in the design of the complete ETAAS system as Varian have done in their recently launched Spectraa 600 and 800 series spectrometer systems. Sample and modifier pipetting are car- ried out while the furnace programme is running and injection is made into the tube about 12 s after the end of the heating programme. The smaller mass of graphite in the Varian furnace presumably cools more quickly than the Perkin-Elmer furnace used in the work described above.Using this new system Shrader et aL6' demonstrated that it is now possible to achieve cycle times of 36-47 s in the determination of cadmium in water wine beer and human serum samples. Perkin-Elmer 172 Journal of Analytical Atomic Spectrometry March 1995 Vol. 10Table 4 Applications of fast furnace analysis Element (s) Matrix Food plant water sediment and urine RMs Dialysate fluids Waters Bone soft tissues Waters Ashing stage No Comments Samples (except water) digested before analysis Combined dry/ash stage Hot injection Samples digested with HNO High temperature drying La modifier for Pb Slurries in 5% HNO 0.04% Triton X-100. Problems with Se detection Hot injection onto a transversely heated graphite tube. Solid samples acid digested Slurries in dilute HNO or digests with HN0,-HF. Seawater analysed directly for Cd and Pb matched standards.Samples dissolved in HNO High temperature drying. Matrix- Samples deproteinized Hot injection. Ammonium oxalate NH,H2P04 modifier for Pb modifier for Cd Ref. 49 55 35 56 48 53 57 54 58 32 34 33 52 59 34 51 60 34 2 61 32 36 32 62 63 50 64 Ag As Cd Cr Cu Ni Pb A1 Al Cu and Pb Al Fe Al Pb Yes No No No As Pb Sr T1 Coal fly ash No As Cd Cu Cr Pb Se Ti and V Plant water sediment sewage sludge and urine RMs Soil sediment plant RMs seawater No (except As and Pb in urine As in sediment) As Cd Co Cr Cu Mn Ni Pb V Zn No (except Pb in seawater) Bi Cu Fe Ni Pb Se Te Zn Fine silver No Cd Cd Cr Blood Urine Yes Yes Cr Cu Pb Plant materials No (Cr Cu) Yes (Pb) No Yes Cr Cr Urine Water Cr speciated by extraction with APDC into IBMK.Combined dry/ash stage in slow ramp up to 900°C Hot injection Slurries in 3% HF high temperature Slurries in 3% HF drying no clean stage Cr Cu and Ni Cu Cr and Pb Water Diatomaceous earths Yes No Cu Cr Co Fe Mn and c u c u Cu Cr Fe Pb and Zn Ni Glasses No Serum Serum Sweets chewing gum Yes Yes No Hot injection anti-foam agent used Samples calcined at 400 "C prior to preparation of slurry from residue. High temperature drying c u Cu Fe Ni and Pb Urine Urine milk powder and Blood Blood bovine liver RMs No Yes Pd modifier with H hot injection. Milk Samples deproteinized NH,H,PO,-HN0,-Triton X- 100 High temperature (400°C) drying Teeth digested in HNO and liver digested before analysis modifier Pb Pb No Yes Pb Pb Soils Teeth No No Table5 Times taken by some commercial autosamplers in ETAAS measured from the completion of the heating programme to the start of the programme after injection the determination of blood lead for assessment of lead exposure under the UK Control of Lead at Work Regulations. However the cost of ICP emission spectrometers and mass spectrometers is falling and where there is a need for multielement analyses these techniques offer advantage of time and cost (particularly the cost of analyst's time) when compared with sequential single element analyses by AAS.The cost of an ICP emission spectrometer now approaches that of a complete graphite furnace AA system and recent developments with axial plasma viewing bring detection limits of commercial instruments close to those of ETAAS.68 In the author's laboratory a fast sequential ICP spectrometer with a vertically-orientated plasma has taken over much of the workload formally carried out by ETAAS and FAAS.It was first applied to the determi- nation of aluminium and calcium in tap and purified water used in dialysis machines and replaced sequential determi- nation by fast furnace AAS and FAAS for A1 and Ca respect- ively. Then a method for the determination for A1 in serum was developed which although often working close to the detection limit has proved reliable and robust in the monitor- ing of serum aluminium in patients on dialysis which brings in around 4000 samples per year. These methods69 have given our laboratory good performance in external quality assess- ment schemes.Zinc and copper in serum and magnesium in urine are also now determined by ICP-AES replacing determinations by FAAS. Manufacturer Model Perkin-Elmer AS1 * AS40* Zeeman 3030/AS60 (original software) Zeeman 3030/AS60 (revised software) 1100/AS70 Spectra 30/40 without blank Varian Spectra 30/40 with blank Time/s 29 29 51 41 43 38-42 33 * Early autosamplers with their own controller. Times are therefore independent of the spectrometer they are connected to. have also recently introduced an autosampler the AS-72 which allows preparation of samples while the furnace heating programme is proceeding. Comparison With Other Techniques Most of the AA spectrometers sold are single element instru- ments; multi-element AA spectrometers have still to make an impact.Fortunately there is still a demand for single element analyses. In the author's laboratory most of the requests for trace element analysis require a single element for example Journal of Analytical Atomic Spectrometry March 1995 Vol. 10 173Data stored Data entered Process Printed report Heading I I Standard absorbance Calculate calibration No. of standards and concentration Values values line of best fit Equation of line of best fit Plot calibration graph I I Print calibration graph Element relative atomic mass Fig. 2 Flow chart for computer programs for evaluation nf analytical data customized for particular determinations Reagent blank value Electrothermal AAS is therefore likely to suffer intense competition from these alternative techniques.It seems reason- able to suppose that its future can be predicted from what has happened to FAAS. Flame atomic absorption spectrometers today are relatively simple and compact and have a market as low-cost analysers to laboratories with relatively simple analyt- ical needs (i.e. the analysis of a few elements). To occupy a similar niche for the determination of elements at lower concentrations ETAAS needs to be of lower cost than its competitors relatively simple and fast enough to offer a comparable rate of analysis. Developments in fast furnace analysis have shown that ETAAS does not need to be as complex as conventionally practised. Indeed the work by Slavin and c ~ - w o r k e r s ~ ~ ~ ~ * ~ ~ demonstrating that for a large number of different determinations modifiers and ashing stages are unnecessary leads to considerable simplification of the technique. Further evaluation and development are necessary but the evolution of a simpler set of rules for temperature programming in ETAAS would be a step forward.-- Subtract blank value and calculate concentrations Evaluation of Results This should follow the same concept in reducing the number of steps and time in the evaluation of results to a minimum. Computer software should enable the production of a cali- bration graph from the absorbance measurements of the standards and evaluation of results from the data for the samples. In the field of trace elements in clinical chemistry it is often useful to know results in two sets of units 81-1 and moll-l.Standards and reference material data are frequently in 81-1 units whereas in the UK moll-’ units have been adopted as standard for the reporting of results in clinical chemistry. To obtain results in two sets of units in the same computer program the element has to be specified as the relative atomic mass is required for conversion. From two fundamental programs using linear and quadratic least squares fit respectively for calibration a number of variants have been written in Microsoft Quickbasic specifically for determinations that form a significant part of the workload of our laboratory. This customization allows fixed standard values to be incorpor- ated correction for blank values and results to be evaluated in the two sets of units (Fig. 2). The programme is tailored to follow the pattern of the worksheet so that key entries logically follow in order.For the analysis of solid samples after dissolu- tion e.g. arsenic in hair the masses can be entered to calculate concentrations in the original sample. Results and calibration data are automatically printed out. Print out results in pg r1 and pmol r1 or nmol I-’ CONCLUSIONS Minimisation of the parameters time cost sample volume energy requirements reagent consumption and waste pro- duction leads to analytical processes which are simple rapid and efficient in use of resources. There are further advantages to be gained which are not so immediately apparent. Trouble-shooting is quicker and easier and since the processes are simpler there is less to go wrong. As an example blanks obtained in the partial digestion pro- cedureZ4 were lower than a more complete digestion using HN03-H202 .70 In fast furnace analysis there are incidental advantages of increased tube lifetime and a decrease in time- dependent effects during the course of a batch of analysis such as drift evaporation and contamination of samples.No. of significant figures in result I would like to thank Dr C. Mullins and Mr B. Field of Varian for providing me with details of recent work on fast furnace analysis. Sample absorbance values REFERENCES 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 Evenson M. A. and Warren B. L. Clin. Chem. 1975 21 619. Halls D. J. Anal. Proc. 1988 25 232. Shaw J. C. L. Pediatr. Clin. North. Am. 1973 20 333. Shaw J. C. L. Arch.Dis. Child. 1988 63 448. Halls D. J. in Heavy Metals in the Environment-Proceedings of an International Conference in Edinburgh 1991 ed. Farmer 3. G. CEP Consultants Edinburgh 1991 Vol. 2 p. 123. De Benzo Z. Velosa M. Ceccarelli C. De la Guardia M. and Salvador A. Fresenius’ J. Anal. Chem. 1991 339 235. Miller-Ihli N. J. Anal. Chem. 1992 64 964A. Kingston H. M. and Jassie L. B. Anal. Chem. 1986 58 2534. Branch S. Crews H. M. Halls D. J. and Taylor A. J. Anal. At. Spectrom. 1991 6 69R. Taylor A. Branch S. Crews H. M. and Halls D. J. J . Anal. At. Spectrom. 1992 7 67R. Taylor A. Branch S. Crews H. M. and Halls D. J. J. Anal. At. Spectrom. 1993 8 79R. Taylor A. Branch S. Crews H. M. Halls D. J. and White M. J. Anal. At. Spectrom. 1994 9 87R. Gorsuch T. T. The Destruction of Organic Matter Pergamon Press New York 1970.Krushevska A. Barnes R. M. Amarasinwaradena C . J. Foner H. and Martines L. J. Anal. At. Spectrom. 1992 7 845. Krushevska A. Barnes R. M. Amarasinwaradena C. J. Foner H. and Martines L. J. Anal. At. Spectrom. 1992 7 851. Hinners T. A. Fresenius 2. Anal. Chem. 1975 277 377. De Boer J. L. M. and Maessen F. J. M. J. Spectrochim. Acta Part B 1983 38 739. Asp T. N. and Lund W. Talanta 1992 39 563. Luterotti S. Zanic-Grubisic T. and Dubravka J. Analyst 1992 117 141. 174 Journal of Analytical Atomic Spectrometry March 1995 Vol. 1020 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 Basson W. D. and Bohmer R. G. Analyst 1972 97 482. Price W. J. Spectrochemical analysis by atomic absorption Heyden London 1979.Premi P. R. and Cornfield A. H. Spectrovision 1967 18 2. Kuennen R. W. Wolnik K. A. Fricke F. L. and Caruso J. A. Anal. Chem. 1982 54 2146. Niazi S. B. Littlejohn D. and Halls D. J. Analyst 1993 18 821. Halls D. J. Niazi S. B. and Littlejohn D. unpublished results. Zima S. Sucmanova M. Celechovska O. Rchurkova I. and Sucman E. Proceedings of the 5th Colloquium Atomspek- trometrische Spurenanalytik ed. Welz B. Perkin-Elmer Uberlingen 1989. Van Wyck D. B. Schifman R. B. Stivelman J. C. Rulz J. and Martin D. Clin. Chem. 1988 34 1128. Gilman L. and Grooms W. Anal. Chem. 1988 60 1624. De La Guardia M. Salvador A. Burguera J. L. and Burguera M. J. Flow Znjection Anal. 1988 5 121. Blust R. Van der Linden A. and Decleir W. At. Spectrosc. 1985 6 163. Bahreyni-Toosi M.-H.and Dawson J. B. Analyst 1983,108,225. Halls D. J. Analyst 1984 109 1081. Halls D. J. Mohl C. and Stoeppler M. Analyst 1987 112 185. Knowles M. J. Anal. At. Spectrom. !989 4 257. Kunwar U. K. Littlejohn D. and Halls D. J. J. Anal. At. Spectrom. 1989 4 153. Kunwar U. K. Littlejohn D. and Halls D. J. Talanta 1990 37 555. Apostoli P. Alcssio L. Dal Farra M. and Fabbri P. L. J. Anal. At. Spectrom. 1988 3 471. Beach L. M. Varian Instruments at Work AA-102 Varian Australia Pty Ltd. Mulgrave Australia 1990. Nham T. T. Varian Instruments at Work AA-92 Varian Australia Pty Ltd. Mulgravc Australia 1989. Hoenig M. Vurian Instruments at Work AA-49 Varian Australia Pty Ltd. Mulgrave Australia 1985. Moffat J. H. Varian Instruments at Work AA-112 Varian Australia Pty Ltd.Mulgrave Australia 1993. Hocnig M. and Van Hoeyweghen P. Varian Znstruments at Work AA-66 Varian Australia Pty Ltd. Mulgrave Australia 1986. Beach L. M. Varian Znstruments at Work AA-95 Varian Australia Pty Ltd. Mulgrave Australia 1990. Hoenig M. Varian Znstruments at Work AA-61 Varian Australia Pty Ltd. Mulgrave Australia 1986. 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 Matousek J. P. Talanta 1977 24 315. Bank P. C. de Loos-Vollcbregt M. T. C. and de Galan L. Spectrochim. Acta Part B 1988 43 953. Bank P. C. de Loos-Vollebregt M. T. C. and de Galan L. Spectrochim. Acta Part B 1989 44 571. Halls D. J. J. Anal. At. Spectrorn. 1989 4 149. Slavin W. Manning D. C. and Carnrick G. R. Spectrochim. Acta Part B 1989 44 1237.Hinds M. W. Latimer K. E. and Jackson K. W. J. Anal. At. Spectrom. 1991 6 473. LopCz Garcia I. Arroyo Cortez J. and Hernandez Cordoba M. J. Anal. At. Spectrom. 1993 8 103. Halls D. J. and Fell G. S. J. Anal. At. Spectrom. 1988 3 105. Bradshaw D. and Slavin W. Spectrochim. Acta Part B 1989 44 1245. Hoenig M. and Cilisscn A. Spectrochim. Acta Part B 1993 48 1003. Halls D. J. and Fell G. S. Analyst 1985 110 243. Lian L. Spectrochim. Acta Part B 1992 47 239. Li Z. Carnrick G. R. and Slavin W. Spectrochim. Acta Part B 1993 48 1435. Hinds M. W. J. Anal. At. Spectrom. 1992 7 685. Subramanian K. S. Anal. Chem. 1988 60 11. Bcndicho C. and de Loos-Vollebregt M. T. C. Spectrochim. Acta Part B 1990 45 679. Vinas P. Campillo N. Lopez Garcia I. and Hernandez Cordoba M. Analyst 1994 119 1119. Subramanian K. S. At. Spectrosc. 1987 8 7. Parsons P. J. and Slavin W. Spectrochim. Acta Part B 1993 48 925. Keating A. D. Keating J. L. Halls D. J. and Fell G. S. Analyst 1987,112 1381. Routh R. W. Doidge P. S. Chidzcy J. and Frary B. Am. Lab. 1982 14 80. Halls D. J. J. Anal. At. Spectrom. 1994 9 1177. Shradcr D. Flajnik C. Knowles M. and Vanclay E. Lecture presented at the Pittsburgh Conference 1994. Bedson A. J. Anal. At. Spectrom. 1994 9 37N. Lyon T. D. B. Gibbons J. Keating A. Cunningham C. Halls D. J. and Fell G. S. Ann. Clin. Biochem. accepted for publication. Analytical Methods Committee Analyst 1967 92 403. Paper 4f05.5590 Received September 13 1994 Accepted November 24 1994 Journal of Analytical Atomic Spectrometry March 1995 Vol. 10 175

ISSN:0267-9477    

DOI:10.1039/JA9951000169

出版商:RSC    

年代:1995

数据来源: RSC

摘要:

Application of Inductively Coupled Plasma Atomic Emission and Mass Spectrometry to Forensic Analysis of Sodium Gamma Hydroxy Butyrate and Ephedrine Hydrochloride* Journal of Analytical Atomic Spectrometry Invited Lecture KAREN A. WOLNIK DOUGLAS T. HEITKEMPER JOHN B. CROWE BARBARA S. BARNES AND THOMAS W. BRUEGGEMEYER National Forensic Chemistry Centre US Food and Drug Administration Cincinnati OH 45221 USA The identity and relative amounts of various elements in samples of two compounds sodium gamma hydroxy butyrate (GHB) and ephedrine hydrochloride (ephedrine) have been used to compare items of evidence. GHB unapproved for use in the United States and ephedrine used in the illicit manufacture of methamphetamine are currently of interest to law enforcement authorities. In this paper the analysis of the elemental impurities in GHB by inductively coupled plasma atomic emission spectrometry (ICP-AES) has been used to further the investigation of clandestine manufacturing sources.The following elements were detected in samples of GHB Ba Ca Cd Fe K Mg Ni P Pb Si Sr and Zn. Results were used to demonstrate an association between samples of unknown origin. Analysis of samples of ephedrine of known origin from legitimate manufacturers by ICP-mass spectrometry showed that Al Ba Mn Pt Rb and Sr were the most useful for discriminating between sources and making direct comparisons. Interpretation of results with respect to the intended use of conclusions derived from those results is discussed. Keywords Inductively coupled plasmu atomic emission spectrometry; inductively coupled plusmu muss spectrometry; ephedrine hydrochloride; gamma hydroxy butyrate; sodium oxybate; forensic analysis Historically forensic scientists have used the identity and relative amounts of various elements in samples such as glass and paint to differentiate or link items of evidence. In 1982 seven people in the Chicago area were poisoned by Tylenol capsules purposely tainted with potassium cyanide.In that instance analysis by inductively coupled plasma atomic emis- sion spectrometry (ICP-AES) of trace element contaminants in a relatively pure chemical KCN was used for making comparisons of suspect samples.’ This technique has been extended to include ultratrace analysis by ICP-mass spec- trometry (MS) and has been applied to a number of diverse forensic samples.2 -’ ICP-AES and ICP-MS analysis of forensic drug samples including methamphetamine cocaine heroin and legitimate bulk pharmaceuticals has been described.*-l’ Elemental analy- sis of drug substances has the potential for providing infor- mation regarding geographic origin synthetic route and type of refining treatment used in addition to its usefulness in direct * Presented at the Seventh Biennial National Atomic Spectroscopy Symposium (BNASS) Hull UK July 22-24 1994.sample comparison analyses. Comparative chemical analysis of cocaine samples using gas chromatography for the detection and quantification of manufacturing impurities has been used successfully in criminal prosecution^.'^"^ Elemental analysis by ICP-AES and ICP-MS is applied here to two drugs sodium gamma hydroxy butyrate (GHB) and (-)-ephedrine hydrochloride (ephedrine) currently of interest to law enforcement authorities.The relative amounts and identity of the trace elements present in these substances are evaluated for the ability to discriminate and classify various samples. GHB or sodium oxybate is recognized as a hypnotic an adjunct to ane~thesia;’~ however GHB has not been approved for use in the United States. Consequently the sale and distribution of this drug is illegal. Nevertheless GHB has been promoted as an ergogenic aid (a substance that increases some aspect of muscle performance) and is marketed to athletes bodybuilders and others. It is believed that GHB increases the secretion of natural growth hormone and/or that GHB will improve the quality of sleep and produce a ‘euphoric’ effect.” Use of GHB has caused serious illness in a number of people with symptoms including respiratory problems seizures and coma.16 According to popular literature targeted at users of ergogenic aids sodium oxybate can be easily prepared by the alkaline hydrolysis of gamma-butyrolactone a fairly common industrial s01vent.I~ Ephedrine a stimulant less potent than methamphetamine is often used in the illicit manufacture of methamphetamine a controlled substance.Ephedrine is available in approved prod- ucts over-the-counter in the United States. Unapproved prod- ucts containing ephedrine can be purchased on the black market. In some instances ephedrine is obtained from finished dosage form products.In other cases bulk ephedrine is diverted from legitimate use to serve as the precursor. EX PER1 MENTA L Reagents and Standards The water used was distilled and de-ionized (DDW) (18 MR cm Millipore). The nitric acid used was Baker Instra-Analyzed (Phillipsburg NJ USA) for ICP-AES studies and GFS Chemicals (Columbus OH USA) Double Distilled for ICP-MS studies. Plasma-grade reagent standards (1000 pg ml-’) were used in the preparation of all standard solutions (Spex Industries Edison Park NJ USA and Inorganic Ventures Toms River NJ USA). Journul of Analytical Atomic Spectrometry March 1995 Vo1. 10 177Samples and Preparation Samples of GHB were obtained as investigative samples and were prepared simply by adding 15ml of 2% nitric acid to 0.5 g of powder in an acid-washed 30 ml high density polyethy- lene (HDPE) bottle.The mixture was shaken until all GHB dissolved. Samples of ephedrine were collected by another government agency from two different sources and will be described here as ‘Investigative’ samples 1 and 2. Samples were also obtained from the designated manufacturer’s reserve and are referred to here as ‘Known’ samples 1-4. In addition four samples of ephedrine were obtained from commercial sources for use as comparison samples. These samples are described in this work as ‘Comparison’ samples 1-4. Ephedrine samples were prepared by adding 1.25 g of 50% v/v nitric acid to 0.25 g of ephedrine. This mixture was then allowed to stand for 30min prior to addition of an internal standard and dilution to a final mass of l o g with DDW.Samples were prepared as described above in acid-washed 30 ml HDPE bottles and transferred to acid-washed autosampler tubes immediately preceding ICP-MS analysis. Instrumentation The simultaneous ICP-AES instrument used in this work was a Model 1160 Plasma Atomcomp (Thermo Jarrell Ash Franklin MA USA). The instrument has an updated PC-based data acquisition and readout system. The nebulizer used was a fixed crossflow. The ICP-AES operating conditions are described in Table 1. The ICP-MS instrument was a PlasmaQuad Model PQ2+ (Fisons Winsford Cheshire UK) equipped with a Gilson Model 222 autosampler (Middleton WI USA). A Meinhard concentric nebulizer and Scott-type spray chamber cooled to 5°C were used. The ICP-MS was operated in the scanning data acquisition mode.Operating conditions are listed in Table 2. The following isotopes were used for quantification 27Al ”Mn 85Rb 88Sr 138Ba and lg5Pt. Table 1 ICP-AES operating conditions Forward power/kW Reflected power/W Outer gas flow11 min-I Intermediate gas flow/l min-’ Injector gas flow/l min-’ Solution delivery rate/ml min- Viewing height/mm Background correction/nm 1 0.95 <5 20 0.75 0.75 1.5 15 k 0.03 Table 2 ICP-MS operating conditions Forward power/kW Reflected power/W Outer gas flow/] min-’ Intermediate gas flow/l min-’ Injector gas flow/l min-’ Solution delivery rate/ml min-’ Analyser stage/mbar Intermediate stage/mbar Expansion stage/mbar Scanning parameters Pulse counting dwell/ys Time/p Analogue dwell time Channels per amu No.of integrations Skipped mass regions Internal standard 1.35 <5 14 1 .o 0.80 1 2.1 x 1.5 320 320 20 3 3 12-23 and 28-42 10 ng ml-’ of In < 10-4 RESULTS AND DISCUSSION Sodium Gamma Hydroxy Butyrate Evaluation of the elemental patterns that result from ICP- AES or ICP-MS analysis of a material depends on a number of factors in addition to the accuracy and precision of the analytical technique. The history of the samples being provided the questions being asked about those samples and the intended use of the answers to those questions must all be considered. In the early 1990s the Food and Drug Administration (FDA) was investigating widespread sales of GHB. Samples purchased undercover all across the country were referred to our laboratory for elemental analysis in the hope that the results would focus the investigation and provide information on production and distribution channels.In addition to direct comparison of various suspect samples the investigators wished to know the number of illegal manufactur- ing sources of GHB. The certainty and statistical standards applied to ‘exploratory data analysis’ which is aimed at furthering an investigation are less rigorous than required to provide ‘proof’ in a criminal prosecution. Investigators located two illicit GHB manufacturing sites and samples were collected from each. Results obtained by ICP-AES for the replicate analyses of these ‘known’ samples are presented in Table 3. Samples from one of the manufactur- ing sites (Site A) consisted of two ‘in process’ portions of GHB (A-1 and A-2) and portions taken from 5 filled drums found at the site (A-3 through A-7).The comparison between the ‘in process’ and drum samples was significant since the possibility existed that the drum samples had been manufactured else- where. The sample obtained from the other site (Site B in Table 3) was a single sample of limited quantity <600 mg. The results shown in Table 3 illustrate the similarities between the ‘in process’ samples and the drum samples. Note that concentrations have not been corrected for moisture. Results for drum A-3 are somewhat out of line particularly for calcium magnesium and strontium but are much more closely aligned with Site A than Site B. The obvious differences in the levels of various elements notably cadmium lead calcium mag- nesium and strontium found in the samples from the two sites indicate the potential for discrimination between these two sources based on elemental analysis.Other GHB samples purchased ‘undercover’ were compared with these samples of known origin. Plots of element X versus element Y were useful for facilitating comparisons of large numbers of results. Several suspect samples showed definite similarities to the ‘known’ samples while others tended to group together but were different from the ‘known’ samples. Examples of element versus element plots are shown in Fig. l(a)-(c) for a set of 20 suspect samples (including the Site A drum samples) referred to our laboratory and analysed over a span of several months. Three replicate weighings were analysed for each suspect sample and each sample is therefore represented on the plot by three points.Individual results for the known samples from Site A are symbolized by a star and from Site B by an asterisk. The drum samples and some suspects are symbolized with a circle while other suspects are symbolized with triangles or squares to permit comparison of one plot with another. A plot of calcium versus cadmium [Fig. l(a)] served to separate the set into two groups which coincided with the two known manufacturing sites and corre- sponded to high versus low or non-detectable levels of these two elements. Strontium versus lead [Fig. l(b)] and iron versus barium [Fig. l(c)] further divided the set into three groups. It can be inferred that the suspect samples represented by squares were manufactured at Site B and that the samples represented by circles are associated with Site A.The suspect samples symbolized by triangles appear to be distinct from the other groups. Other element-by-element plots either corroborated 178 Journal of Analytical Atomic Spectrometry March 1995 Vol. 10Table 3 Average concentration (pg g- ’) in GHB samples n = 3 except for B- 1 when n = 2 Site No. A- 1 A-2 A-3 A-4 A-5 A-6 A-7 B-1 Detection limits* Ba 0.165 0.160 0.143 0.195 0.234 0.195 0.222 0.136 Ca 139 148 149 200 154 200 76.6 7.82 Cd d 0.073 d 0.073 d 0.073 d 0.073 d 0.073 d 0.073 d 0.073 1.94 Fe 2.49 2.64 3.14 2.96 3.56 2.15 3.59 8.08 Mg 11.2 11.7 11.4 14.4 11.9 14.3 d 1.1 6.10 P d 1.8 d 1.8 1.9 d 1.8 d 1.8 d 1.8 < 1.8 3.26 Pb d 1.4 < 1.4 d 1.4 < 1.4 d 1.4 d 1.4 < 1.4 3.58 Si 27.5 27.3 24.4 26.6 29.8 27.9 28.8 16.9 Sr 1.56 1.67 0.761 1.55 2.24 1.71 2.23 0.140 Zn 0.262 0.280 0.407 0.9 14 0.508 0.822 0.665 0.937 0.034 0.35 0.073 0.68 1.1 1.8 1.4 1.7 0.037 0.082 * Calculated as 3 times the standard deviation of 9 method blanks multiplied by the dilution factor Ni and K were not detected in these samples.these three groupings or displayed points which were insufficiently resolved to aid in evaluation. In total more than 100 samples of GHB were eventually analysed using ICP-AES. The following elements were detected in various samples Ba Ca Cd Fe K Mg Ni P Pb Si Sr and Zn. Analysis of GHB purchased from legitimate chemical distributors showed much lower levels of contamination with only a few elements above the detection limit.Comparison of elemental results for suspect and known samples and between suspect samples provided information useful to investigators. This information served as a basis for testing knowledge obtained from labels and various documents for corroborating statements made by informants and for linking manufacture and distribution of GHB from coast to coast. However we were unable to provide a definitive answer to the question of the number of sources of manufacture since the cause(s) for the observed variations in elemental results which led to the sample groupings could not be defined. The differences could represent different batches of production different manufactur- ing sites or different histories Le. different sources of contamination.250 200 5 150 cn s 0 100 Y 50 0 I . m m r n l rm (- )-Ephedrine Hydrochloride The results for the ephedrine samples referred to our laboratory were intended for use in a criminal prosecution. The questions asked were well-defined; therefore our approach to the elemental analysis of ephedrine was much more methodical. 3 - 2 - 1 - Analytical method development Preliminary ICP-MS scans were performed on several ephedrine samples to determine which elements might be useful in discriminating between ephedrine manufacturers. No quanti- fication was performed in any of these preliminary experiments. Integrated area counts were compared for all scanned masses. Fourteen elements including Na Al Mn Rb Sr Mo Sb Ba La Ce Gd Dy Pt and Pb were selected for further study.Most of the ephedrine samples were analysed quantitatively on two separate occasions. Seven of the original fourteen elements were found to provide a significant ability to differen- tiate various ephedrine samples (Na Al Mn Rb Sr Ba and Pt); however only six of the seven provided reproducible values from the day-to-day runs. The within-day reproduc- ibility for sodium was acceptable; however the day-to-day reproducibility was poor. It is speculated that this was due to a problem with the cross calibration between the pulse counting and analogue detector modes of the ICP-MS. Higher levels of 8 - 6 - L E W I I I 1 0 0.1 0.2 0.3 0.4 [Baypg ml-’ - Fig.1 Element uersus element plots for GHB samples. + Site A sodium and were found in Some and * Site B A Suspects 0 Suspects o Suspects (see text for further required detection in the analogue mode.The remainder of details) the elements studied utilized only pulse counting detection. Journal of Analytical Atomic Spectrometry March 1995 Vol. 10 179Table 4 Ephedrine results Element/ng g-' Sample Investigative 1 Mean SD (n = 8) Investigative 2 Mean SD (n= 12) Known 1 Mean SD (n=4) Known 2 Mean SD (n=6) Known 3 Mean Known 4 Mean SD (n = 6) SD (n=6) Comparison 1 Mean SD (n = 6) Comparison 2 Mean SD (n=3) Comparison 3 Mean SD (TI = 3) Comparison 4 Mean SD (n=3) A1 622 109 1409 175 729 62* 638 55 1295 832 1278 154 563 60 44 199 848 46574 3955 438 70 Mn 123 16 69.2 22.3 181 19 199 9 129 33 75.5 21.8 74.6 23.0 75.7 4.7 79.3 8.0 131 2 Rb 10.5 1.3 3.8 0.5 15.2 1.4 11.5 1.2 10.7 1 .o 3.8 0.2 1.2 0.2 1.1 0.4 1.2 0.3 4.6 0.2 Ba 264 48 57.4 9.1 333 76 26 1 57 298 38 64.1 8.0 trace trace 8.1 11.2 39.3 1.2 Pt 3.2 0.4 0.88 0.44 2.3 0.3 6.5 1.7 3.3 0.2 2.8 0.9 c DL 19.0 3.5 17.9 1.1 3.0 0.5 Sr 25.9 3.7 24.1 3.0 27.1 3.3 33.5 4.5 27.0 1.6 23.4 3.0 3.3 0.8 2.7 0.8 2.5 0.6 18.4 0.4 * n = 6 for Al.Elemental determinations A summary of ICP-MS results for the six elements selected is shown in Table4. The mean concentrations for the Investigative and Known samples are derived from at least three replicate weighings on each of two separate occasions. Comparison samples 2-4 were analysed in triplicate on only a single occasion. The reproducibility of sample determinations generally ranged from 2-35% with an average of about 15%. As expected the relative standard deviations (RSDs) for con- centrations closer to the detection limit were significantly higher.The RSD for A1 in Known sample 3 was 64%; however the within-day reproducibilities were 8 and 28%. The reason for the difference from day-to-day is not known. Sample detection limits (DLs) were as follows 7.2 ng 8-l Al 3.2 ng g-' Mn 0.42ngg-' Rb 0.42ngg-I Sr 0.54ngg-' Ba and 0.72 ng g-' Pt. Detection limits were calculated as 3 times the standard deviation of 10 measurements of a method blank multiplied by a dilution factor of 40. Comparison of results Comparison of forensic samples for purposes of discriminating or classifying samples can be accomplished in a number of different ways including principal components analysis k- nearest neighbours and linear discriminant analysis.A recent review revealed 205 references dealing with chemical pattern recognition in the past two years.18 However when possible methods which are simply applied and understood are pre- ferred for forensic samples. As criteria for discrimination of glass samples analysed by ICP-AES Hickman3 simply compared the ranges (meanf2SD) of each of the elements determined for two samples. Samples were reported as indis- tinguishable if the ranges overlapped for every element. If the range for at least one element did not overlap the samples were reported as distinguishable. Similarly Zurhaar and 300 v- lo 250 rn 5 .- 200 r 2 4- 150 5 100 50 n V INVl KNW1 KNW2 KNW3 INV2 KNW4 Sample Fig. 2 Average concentration profiles for samples of (-)-ephedrine HCl for Investigative samples (INV); and Known samples (KNW) (A1 =A1 ng g-' -+ 5 ) Mullings' used the mean & 3SD to discriminate glass samples analysed by ICP-MS.The first question asked in the ephedrine investigation was 'Are the two Investigative samples identical?. Fig. 2 shows a comparison of the mean concentrations of Al Mn Rb Ba Pt and Sr for several samples including the two Investigative samples (INV1 and INV2). It should be noted here that the A1 concentration has been divided by a factor of 5 for graphical purposes. The profiles for INVl and INV2 appear to be different for Al Mn Rb Ba and Pt. Using the discrimination criteria of Hickman3 the two samples are distinguishable. The ranges for Al Rb Ba and Pt did not overlap. Significance testing was also used to compare the means for each element." An uncertainty of the difference between the means was calculated using the appropriate Student t value and a pooled SD.This uncertainty is then compared with the measured difference between the means. At the 95% level of confidence the Al Mn Rb Ba and Pt mean concentrations are significantly different. Therefore it was concluded that the two samples are distinguishable. A number of analytical specification documents relating to ephedrine were obtained in conjunction with one of the Investigative samples of ephedrine. The second question regarded our ability to match the Investigative samples to the specification documents. In order to answer this question reserve samples with lot numbers corresponding to those on the documents were obtained directly from the manufacturer.These are the aforementioned Known samples and make it possible to link an Investigative sample with information from the documents by direct comparative analysis. Mean element concentrations for the Investigative samples and Known samples are shown in Fig. 2. Several observations were made on visual inspection of the element profiles. The profile for Investigative sample 1 has the same basic appearance as profiles for Known samples 1-3 which were produced at the same manufacturing plant. In addition the profile for Investigative sample 2 appears very much like the profile for Known sample 4 which was manufactured at a different site than Known samples 1-3. Significance testing was used to compare the mean element concentrations for Investigative sample 1 with those of Known samples 1 2 and 3.For Known sample 3 only the mean A1 concentration showed a significant difference with Investigative sample 1 at the 95% level of confidence. For Known samples 1 and 2 the mean concentrations were significantly different from Investigative sample 1 for three of the six elements. 180 Journal of Analytical Atomic Spectrometry March 1995 Vol. 10INV1 INV2 COM1 COM2 COM3 COM4 Sample Fig. 3 Average concentration profiles for samples of (-)-ephedrine HCl for Investigative samples (INV); and Comparison samples (COM) (Al=Al ng g-' + 500) However the ranges (mean 2SD) for each of the six elements determined in Known samples 1 and 3 overlap with those in Investigative sample 1.Thus Known samples 1 and 3 and Investigative sample 1 are indistinguishable when using the range for di~crimination.~ It would seem highly likely that Investigative sample 1 and Known samples 1-3 are from a common manufacturing source. The differences in mean element concentrations for Investigative sample 2 and Known sample 4 were also tested for significance as described above. At the 95% level of confidence only the mean Pt concentration showed a signifi- cant difference. Nevertheless if the ranges (mean & 2SD) are compared as the measure of discrimination the two samples are indistinguishable i.e. the ranges for Al Mn Rb Ba Pt and Sr overlap. It would seem highly likely that Investigative sample 2 and Known sample 4 are from a common manufacturing source.Four additional samples of ephedrine were obtained from commercial sources for comparison purposes (Comparison samples). Mean element concentrations for the Investigative and Comparison samples are shown in Fig. 3. Note that in this figure the A1 concentration has been divided by a factor of 500 for graphical purposes. A number of observations were made regarding the element profiles. The profiles for Comparison samples 1,2 and 3 are significantly different from the Investigative samples. Comparison sample 1 is distinguish- able from the Investigative samples (mean * 2SD). Comparison samples 2 and 3 are indistinguishable from one another but distinguishable from the Investigative samples. Distributors records which were obtained after testing was completed showed the two samples (Comparison 2 and 3) originated from the same source.In addition Comparison sample 4 shows some similarities to Investigative sample 2. The concen- tration ranges for Rb Pt and Sr overlap in the two samples. Although the samples are distinguishable distributors records showed that Comparison sample 4 and Known sample 4 were manufactured at the same site approximately 1 year apart from each other. The comparisons made using Fig. 3 would indicate that conclusions made earlier are valid and that it may be possible to go beyond simple direct comparison analysis and start to classify samples by manufacturing source based on elemental analysis. This would of course require the analysis of many additional ephedrine lots which were produced over a specified time period.REFERENCES 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 Wolnik K. A. Fricke F. L. Bonnin E. Gaston C. M. and Satzger R. D. Anal. Chem. 1984 56 466A. Koons R. Spectroscopy 1993 8 16. Hickman D. A Forensic Sci. Int. 1983 23 213. Wolnik K. A. Gaston C. M. and Fricke F. L. J. Anal. At. Spectrom. 1989 4 27-31. Zurhaar A. and Mullings L. J. Anal. At. Spectrom. 1990,5,611. Koons R. D. Peters C. A. and Rebbert P. S. J. Anal. At. Spectrom. 1991 6 451. Koons R. D. Peters C. A. and Merrill R. A. J. Forensic Sci. 1993 38 302. Suzuki S. Tsuchihashi H. Nakajima K. Matsushita A. and Nagao T. J. Chromatogr. 1988 437 322. Kishi T. J. Res. Natl. Bur. Stand. 1988 93 469. Violante N. Quaglia M. G. Lopez A. and Caroli S. Microchem. J. 1992 45 79. Sheppard B. S. Gaston C. M. Barnes B. S. and Wolnik K. A. paper presented at the 1994 Winter Conference on Plasma Spectrochemistry Abstract No. ThP26 San Diego CA January 1994. Moore J. M. Meyers R. P. and Jimenez M. D. J. Forensic Sci. 1993,38 1305. Casale J. F. and Waggoner R. W. J. Forensic Sci. 1991,36 1312. The Merck Index 1 1 th Edition ed. Budavari S. Merck and Co. Rahway NJ 1989 p. 1365. Phillips W. N. Anabolic Reference Guide (6th Issue) 1991 Mile High Publishing Golden CO pp. 36-37. FDA Quarterly Actiuities Report First Quarter Fiscal Year 1991 Department of Health and Human Services Public Health Service Food and Drug Administration p. 6. Duchaine D. Underground Steroid Handbook for Men and Women Update Daniel Duchaine USA 1992 pp. 45-48. Brown S. D. Blank T. B. Sum S. T. and Weyer L. G. Anal. Chem. 1994 66 340R. Taylor J. K. Quality Assurance of Chemical Measurements CRC Press Boca Raton FL 1987 pp. 29-30. Paper 4/066928 Received November 2 1994 Accepted December 5 1994 Journal of Analytical Atomic Spectrometry March 1995 VoL 10 181

ISSN:0267-9477    

DOI:10.1039/JA9951000177

出版商:RSC    

年代:1995

数据来源: RSC

摘要:

Inductively Coupled Plasma in Fluorescence Spectrometry Source and Atomhon Reservoir* Journal of Analytical Atomic Spectrometry L I Invited Lecture STANLEY GREENFIELD Department of Chemistry Loughborough University of Technology Loughborough Leicestershire LEI 1 3 TU UK Over the last decade three systems involving the use of inductively coupled plasmas in atomic fluorescence spectrometry have been subjected to similar thorough investigations which have resulted in similar findings. This similarity is not altogether surprising as all the systems employ a low-power plasma as an atomizer and differ only in the source employed namely another low-power plasma a high-power plasma and hollow-cathode lamps. The system employing a high-power plasma is taken as the role model and the investigations and findings on that system are described.Possible ways of improving on the results from all three systems are described and suggestions are made on the manner in which the technique may advance. Keywords Inductively coupled plasma; atomic fluorescence spectrometry In the early 1980s work started independently on three atomic fluorescence spectrometric (AFS) systems incorporating an inductively coupled plasma (ICP). Kosinski Uchida and Winefordner' described a system which they called ICP-ICP- AFS where the radiation from excited species in the tail flame of a low-power plasma was used to excite fluorescence in atomic species in the tail flame of a similar low-power plasma. Greenfield2 described a system incorporating a high-power ICP as the source and a low-power plasma as the atomizer; this system was given the acronym ASIA representing atom- izer source ICPs in AFS.Demers and Allemande3 described a system incorporating an ICP as an atomizer and hollow- cathode lamps (HCLs) as sources. This they called HCL-ICP- AFS and it was the precursor to the commercial instrument manufactured by Baird.4-6 Work on these three systems followed much the same pattern with similar areas of investigation in each instance and with similar results. The first experiments with the ASIA system' were to produce power emission excitation and fluorescence curves of growth. These experiments gave the range and operating parameters of the system and confirmed the fact that an ICP is optically thin is a line source and in addition is an efficient atomizer.There are far fewer spectral interferences in AFS than there are in atomic emission spectrometry (AES) or even atomic absorption spectrometry (AAS) undoubtedly because certain stringent conditions which could lead to interference are not often met with in AFS. First the emission profile of the source must overlap with the absorption profile of the interfering element in the atomizer. Second the population of the interfer- ing element in the correct energy level must be significantly high. Third the amount of energy absorbed by the interfering * Presented at the Seventh Biennial National Atomic Spectrometry Symposium (BNASS) Hull UK July 20-22 1994. element to that emitted as fluorescence radiation must be significant. This freedom from spectral interference was investigated using ASIA.In early experiments the classical work of Larson and Fassel' on background shifts due to radiative recombina- tion continua and interference due to collisional broadening were repeated. Because ASIA employs a.c. coupled electronics and background shifts are essentially d.c. shifts this type of interference was shown to be absent.' In emission there is considerable interference on the emission lines of A1 in the 393-396nm region when 1OOOppm of calcium is added to a 1 ppm solution of aluminium. However there is no such interference in atomic fluorescence with ASIA when a similar amount of calcium is added to a 1 pprn solution of Al. This lack of spectral interference is remarkable. Thus although the major zinc resonance line is only 0.003 nm away from a copper non-resonance line trace amounts of zinc can be determined in pure copper.' Many more instances of spectral interference which occur in emission spectrometry were shown to be absent in A F S .~ ~ Studies were also made" of the non-resonance transitions of lead using filters between the source and atomizer plasmas to isolate the required wavelengths. All the fluorescence mech- anisms resonance Stokes direct line anti-Stokes direct line stepwise line and thermally excited fluorescence were observed and verified. This type of non-resonance fluorescence would not only make spectral interference virtually impossible but would also avoid any problems of scattered light and enable multi-pass optics to be used. Studies of chemical and ionization interferences were made." These included the well known depressive effect of PO4 and A1 on Ca I emission and the enhancement due to alkali metal elements.It was found that when the ASIA operating param- eters had been optimized for the best limit of detection (LOD) by the alternating variable search (AVS) method,12 PO4 had no effect on the fluorescence signal of Ca I up to a PO4 concentration of lo4 ppm the Ca concentration being 1 ppm. On the other hand A1 showed a marked depression of the signal. Na and K gave first an enhancement and then a depression of the signal. The effect of A1 on the Ca was judged to be that of stable compound formation. The effect of the alkali metals was thought to be an ionization effect followed by a quenching effect.On re-optimization for the figure of merit minimum inter- ference the results obtained were different; PO4 had no effect A1 could be tolerated up to a concentration of 900ppm and Na and K up to concentrations of 1000 and 100 ppm respec- tively before a depression could be seen in the fluorescence signal. These experiments confirmed the efficacy of the AVS optimization and the spatial dependence of the reactions in the plasma. Optimization for minimum interference will of necessity degrade the LOD obtained as it does in emission. Journal of Analytical Atomic Spectrometry March 1995 Vol. 10 183When the experiments were repeated with a 1 ppm concen- tration of Ca but with measurement of the Ca I1 line it was found that when optimized for minimum LOD PO4 had no effect but Al Na and K had the expected depressive and enhancement effects.However re-optimization for minimum interference showed no interference from PO and Na and K could be tolerated up to about 300ppm but the depressive effect of A1 could not be reduced. As stated at the beginning similar investigations were carried out by other workers on the other two systems i.e. ICP-ICP- AFS'. l3-I6 and HCL-ICP-AFS.'5-17 The results obtained were in broad agreement with those obtained by ASIA. The selec- tivity was excellent. The linear dynamic range was 5-6 orders of magnitude." Chemical and ionization interferences were no greater than in ICP-AES itself regarded as relatively free from this type of interference. Comparison of the LODs obtained by the three systems is difficult rather like comparing apples with oranges as can be seen from Table 1.All else being equal one might expect ASIA to yield lower LODs than the low-power system ICP-ICP-AFS as the fluor- escence radiation produced in the atomizer is initially proportional to the source radiation. This in turn will be proportional initially to the concentration of excited species in the tail flame which in turn is related to the power in the plasma and the concentrations of the appropriate element in the solution nebulized into the plasma. All of these are much greater in the high-power plasma of ASIA than in the low- power systems. Referring to Table 1 and the non-refractory elements this expectation would appear to be fulfilled. This does not appear to be so when one considers the refractory elements.However it should be noted that the efficiency of transfer of energy from the source to the atomizer is 1-2% in ASIA and 13% in the low-power system owing in the latter instance to superior optics. From experience it can be said that a five-fold improve- ment in LOD would be a modest figure to assume for an increase in transfer efficiency from 1 to 13% source to atomizer in the ASIA system. The energy transfer efficiency from source to atomizer for the Baird instrument is not known. However the transfer efficiency of fluorescence radiation from the atomizer plasma to the detector is likely to be higher than in the other two systems as an interference filter is used to isolate the required wavelength and not a monochromator.It is known that an ultrasonic nebulizer when used in conjunction with the source plasma on ASIA will give an increased signal by a factor of 3-4. Further it is also knownI7 that equal signals can be obtained from two solutions the elemental content of one being 60-70 times lower than the Table 1 LODs (ng m1-l) (3 x sb) obtained by ICP-ICP-AFS,l6 HCL- ICP-AFS'*and ASIA" compared with LODs obtained by ICP-AES19 ICP-ICP-AFS Element ( 13 YO)* Non-refractory elements- Ca 0.6 c o Cr 15.0 c u 0.6 Fe Na 1.5 Zn 3.0 Refractory elements- A1 15.0 B 15.0 Ba 1.4 Mo - Si 10.5 - - ASIA ( 1 Yo)* 0.2 9.5 3 .O 0.4 5.0 0.1 2.0 20.5 28.0 3.5 63.0 54.5 HCL-ICP-AFS (?)* co.1 0.5 0.6 0.1 0.5 <0.1 <0.1 5.0 60.0 25.0 8.0 40.0 ICP-AES 0.15 3.0 3 .O 1.5 1.5 6.0 1.5 6.0 3.0 0.15 7.5 5.0 ~~ ~~ * Efficiency of light transfer source to atomizer.other if the aerosol from the weaker solution is heated and de- solvated before it is passed through the ICP. There are problems associated with the de-solvation of heated aerosols which cannot be discussed here but these problems may not occur with cryogenic membrane de-solvation systems. Adopting the same approach to the atomizer plasma would undoubtedly raise its temperature which would necessitate lowering the power or observing further along the tail flame. If this proved possible then a combination of ultrasonic nebulizer with de-solvation should produce a higher signal. However the greatest advantage to be gained from de- solvation on the atomizer plasma is likely to be the removal of oxygen which should improve the LOD of the refractory elements.The point of this discussion is that an improvement in optical transfer from source to atomizer and from atomizer to the monochromator or filter and detector together with the use of ultrasonic nebulizers and de-solvation should result in LODs at the very least equal to those in ICP-AES for the refractory elements and much better for the non-refractory elements. This is a good place to discuss the future of the ICP in AFS (excluding the laser-ICP systems as being beyond the scope of this paper). The Baird instrument (which has been withdrawn from the market for as far as is known commercial and not scientific reasons) may have been before its time. It was a sophisticated machine somewhat expensive and exceeded the requirements of many laboratories and encouraged a simplistic approach to a technique that requires some understanding.That the technique of ICP-AFS has merit is an established fact but further progress will only come about if the scientific community can be encouraged to follow the example of the pioneers of ICP-AES who were prepared to retrofit an ICP to existing monochromators or polychromators in order to find out for themselves what such a system would do for them. In a similar manner use can be made of the ICP in an existing ICP-AES instrument as a source and as an atomizer in an AFS mode. Also simple instruments along the lines of early AAS machines can be assembled to investigate the use of the technique of AFS. Such an instrumental set-up has been put together at Loughborough University of Technology (LUT) and is shown in Fig.1. Recorder -. -/x I torch Ta i If I a m e' Hollow cathode '0 lamD Fig. 1 Schematic arrangement for BDHCL-ICP-AFS with axial viewing of the ICP 184 Journal of Analytical Atomic Spectrometry March 1995 Vol. 10Basically it consists of a rearrangement of a very old Shandon AAS machine together with a plasma in an end-on configuration. The low-power generator used to drive the plasma was built in-house. The lamp turret and pulsed power supply are retained as is the monochromator and detector which is turned through 90" and placed on its side so that its slit is parallel to the tail flame of the plasma. A lock-in amplifier integrator and recorder complete the set-up. The sources are HCL and boosted discharge hollow cathode lamp (BDHCL) the latter requiring a boosted power supply. The end-on configuration was chosen in order to increase the size of the fluorescence cell.The HCL source can be replaced by the end of a fibre-optic bundle conveying radiation from an ICP in an existing ICP-AES instrument if it is so desired. To facilitate this use of an existing plasma torch the device shown in Fig. 2 was constructed at LUT. The 'top-hat' that is used to prevent 'arc-over on many ICP torches is replaced by a ground silica flange joint. The joint fits into a cylindrical block of machinable ceramic that has a hole bored through the middle to allow the passage of the plasma and tail flame. Another hole is drilled into the block at right-angles to the central hole to allow the insertion of an optical fibre bundle.If the light from the bundle is required to be modulated it is passed through condensing lenses between which is placed a mechanical chopper. Omenetto2' described an ICP-AFS system in which light from an ICP in a commercial polychromator was used via a lens system with an optical chopper to excite fluorescence using a graphite rod atomizer with spectacular LODs being obtained as can be seen in Table 2. A fibre optical bundle Ceramic blocks A m - 1 Collimating beam probes Fig. 2 Plasma radiation transference device and optical chopper Table 2 LODs (ng ml-l) (3 x sb) obtained by ICP-electrothermal atomization ( ETA)-AFSZo compared with LODs obtained by graphite furnace (GF) AASI9 and by ICP-MS19 Element ICP-ETA- AFS GFAAS ICP-MS Ag 0.04 0.05 0.003 Cd 0.01 0.02 0.003 c u 0.02 0.25 0.003 Mg 0.002 0.01 0.007 Zn 0.003 0.3 0.003 could be used to transfer the radiation from the source to the graphite rod more conveniently.Another possible use of the optical bundle light transference device is in the field of molecular fluorescence where light from an existing plasma instrument could be directed into a cuvette containing a solution of a compound of interest which emitted fluorescence when irradiated by light of a suitable wavelength. This fluorescence radiation after passing through a filter or monochromator could be detected in a variety of ways. Tallant" reported about 15 years ago that an ICP has a comparable performance to an arc lamp in molecular fluor- escence.Surprisingly no further work seems to have been reported following this early paper; yet as Tallant pointed out good reasons exist for further work to be done. By introducing a suitable element into an ICP a desired spectral region may be selected as an excitation source. This ability will be of greatest value for regions of the spectrum in which arc sources have relatively weak output. Prominent plasma lines are available through the ultraviolet region with many elements emitting their strongest line below 250 nm and a few below 200nm. Thus for certain applications the ICP may be more desirable as a source for molecular fluorescence than the commonly used arc lamps such as xenon which have relatively weak UV output or like Hg lamps have holes in their intensity - wavelength profiles.All such lamps can show considerable instability in their radiant fluxes. So far the discussion has been of the use of simple instrumen- tation in order to further the use of ICP-AFS and suggestions for the retrofitting of the ICP in existing commercial instru- ments as a source in fluorescence applications. There is also the possibility of replacing the conventional torch in existing instruments with a long-sleeve torch and using it as an atomizer with an HCL or BDHCL as a source. In the latter instance light transfer would be by lens and/or fibre-optic bundle depending on the geometry of the torch box. There is every possibility that the monochromators in the ICP instru- ment could be used to isolate the required wavelength and with the addition of a lock-in amplifier the existing data collection system could be utilized.In conclusion it can be said that the use of the ICP in AFS has been well researched (over 100 papers have been published on this topic in the last 10 years) and the technique has been found to have many of the attributes of AES and AAS and can be virtually free from spectral interference. It has been demonstrated that the LODs obtained by the technique good as they are can be improved by giving attention to optical performance nebulization and desolvation. Although no com- mercial instruments for ICP-AFS are available it has been shown how simple instruments can be built and the plasma in existing ICP-AES instruments can be utilized to perform both atomic and molecular fluorescence and thus further the use of the technique.REFERENCES 1 2 3 4 5 6 7 8 9 10 Kosinski M. A. Uchida H. and Winefordner J. D. Anal. Chem. 1983 55 688. Greenfield S. Anal. Proc. 1984 21 61. Demers D. R. and Allemand C. D. Anal. Chem. 1981,53 1915. Demers D. R. and Allemand C. D. paper presented at the 1981 Pittsburgh Conference Atlantic City NJ paper No. 122. Demers D. R. and Allemand C. D. paper presented at the 1981 Pittsburgh Conference Atlantic City NJ paper No. 123. Demers D. R. Busch D. A. and Allemand C. D. Znt. Lab. 1982 14 40. Greenfield S. and Thomsen M. Spectrochim. Acta Part B 1985 40 1369. Larson G. F. and Fassel V. A. Appl. Spectrosc. 1979 33 592. Greenfield S. and Thomsen M. Anal. Proc. 1987 24 22. Greenfield S. Malcolm F. M. and Thomsen M.J . Anal. At. Spectrom. 1987 2 711. Journal of Analytical Atomic Spectrometry March 1995 Vol. 10 18511 12 13 14 15 16 17 18 Greenfield S. Salmon M. S. and Tyson J. F. Spectrochim. Acta Part B 1988 43 1087. Greenfield S. Salmon S. M. Thomsen M. and Tyson J. F. J. Anal. At. Spectrom. 1989 4 55. Long G. L. and Winefordner J. D. Appl. Spectrosc. 1984,38,563. Long G. L. Voigtman E. G. Kosinski M. A. and Winefordner J. D. Anal. Chem. 1983 55 1432. Walters P. E. Long G. L. and Winefordner J. D. Spectrochim. Acta Part B 1984 39 69. Kruppa R. J. Long G. L. and Winefordner J. D. Spectrochim. Acta Part B 1985 40 1485. Greenfield S. Durrani T. M. and Tyson J. F. ICP I$. Newsl. 1990 16 159. Demers D. R. and Montaser A. in Inductively Coupled Plasmas in Analytical Atomic Spectrometry ed. Montaser A. and Golightly D. W. VCH New York 2nd edn. 1992 p. 553. 19 The Guide to Techniques and Applications of Atomic Spectroscopy Perkin-Elmer Norwalk CT 1993. 20 Omenetto N. paper presented at Analytiktreffen Atmospektrosk Fortschr. Anal. Ammend Haup vortr. meeting held at Karl Marx University Leipzig Germany 1982. Tallant D. R. ICP In$ Newsl. 1979 5 171. 21 Paper 4/04 755 I Received August 2 1994 Accepted August 24 1994 186 Journal of Analytical Atomic Spectrometry March 199.5 Vol. 10

ISSN:0267-9477    

DOI:10.1039/JA9951000183

出版商:RSC    

年代:1995

数据来源: RSC

摘要:

Effect of Elevated Gas Pressure on Atomization in Graphite Furnace Continuum Source Atomic Absorption Spectrometry with Linear Photodiode Array Detection* CLARE M. M. SMITH AND JAMES M. HARNLYT USDA ARS Beltsville Human Nutrition Research Center Food Composition Lab Beltsville MD 20705 USA Absorbances for Al Cd Cr Mn and Pb integrated with respect to wavelength and time were determined at pressures as high as 6 atmospheres (1 atm = 101.325 Pa) in a standard integrated contact cuvette (ICC) (tube ends open) and an end- capped ICC (tube ends restricted) and with the injection port open and plugged. In a standard ICC the maximum increase in signal was limited to a factor of 3 at 6 atmospheres owing to a strong convection component. Plugging the injection port reduced but did not eliminate the convection component for Cd and Pb.Use of an end-capped ICC increased the sensitivity by an average factor of 3.2 but did not reduce the convection component. Plugging the injection port of the end-capped ICC gave sensitivity increases proportional to the pressure increase suggesting that convection had been eliminated. Factors of 4-6 improvement in sensitivity were obtained for atomization at 6 atmospheres. Reasonable agreement was found between the experimental data and a simple model that incorporated diffusion and convection as loss mechanisms. The rate of analyte loss by convection in an ICC at 1 atm is approximately 1/8 that by diffusion. In 3% NaCI Pb showed improved recoveries at higher pressure with the injection port open or plugged.Good recoveries for Cr were obtained only with the injection port plugged; pressure had no effect. With an end-capped ICC the injection port plugged and 6 atmospheres of pressure detection limits of 50 100 30 and 90 fg were obtained for Cd Cr Mn and Pb respectively. Keywords Atomic absorption; graphite furnace; elevated pressure; second surface atomization; halide interferences Since the first development of the graphite furnace there has been considerable interest in atomization at elevated press- ures.'-'' In principle increasing the pressure for a diffusion limited system produces a lower diffusion coefficient a longer analyte residence time and consequently a larger time- integrated signal (peak area). Implicit in this logic is the assumption that the chemical interactions of the analyte are unchanged and that convection is not a significant loss mechan- ism.Thus doubling or trebling the pressure should produce proportional increases in the time-integrated absorbance. Research in this area however has yielded conflicting results.'-'' In the earliest workip3 atomization at increased pressures resulted in a delay in the appearance time an increase in the width of the analytical peak an extension of the linear dynamic range of the calibration curves and an increased freedom from the effect of variation in the line width of the emission source. The analytical sensitivity however was observed to decresse * Presented in part at the Seventh Biennial National Atomic t To whom correspondence should be addressed. Spectroscopy Symposium (BNASS) Hull UK July 20-22 1994.Journal of Analytical Atomic Spectrometry for all the elements studied. Subsequent research by Hoenig et aL4 and Fazakas et aL5-' reported sensitivity enhancements with increased pressure for some elements but not for others. One suggested explanation was that the enhancement was related to the analyte volatility; enhanced sensitivity should be seen only for volatile element^.^-^ Alternatively it was sug- gested that enhanced sensitivity should be seen only for elements with low relative atomic masses.8 Neither of these hypotheses successfully accounted for all the observed results. Fazakas' was the first to suggest that the broadening and shifting of the absorption profile at higher pressures accounted for the reduction in analytical sensitivity. For conventional AAS the profile width of the emitted radiation of the hollow cathode lamp (HCL) is narrower than the absorption profile width by a factor of 3-5 and thus approximates a monochro- matic source.Because the HCL is not truly monochromatic the absorption measurement varies as a function of the width of the two profiles and the relative position of the profile centres. With increasing pressure in the furnace the absorption profile becomes broader and shallower owing to an increase in the collisional width component. This change in shape is also accompanied by a shift in the position of the profile centre; to longer wavelengths in argon and nitrogen and to shorter wavelengths in hydrogen and helium. Chang and O'Haverl' showed that for In the profile centre was shifted by 8 pm relative to the HCL line at an Ar pressure of 4 atmospheres. For Ca at 4 atmospheres the profile centre was shifted by approximately 5 pm.Fazakis' suggested that the only way to effectively measure absorbance at elevated pressures is to use a continuum lamp as the primary radiation source. He pointed out that with an HCL source the change in shape and position of the absorption profile produced a reduction in sensitivity that offset the increased residence time. Thus it was possible to observe the delayed appearance time and the wider analytical peaks expected for a slower diffusion rate,'-3 with no increase or even a decrease in the integrated absorbance. Using a con- tinuum source and an appropriate detection system absorbance can be integrated with respect to wavelength yielding a constant value regardless of the shape or position of the peak.Thus absorbance integrated with respect to wavelength and time should increase proportionally with increasing pressure. Using a continuum source an echelle spectrometer and a quartz refractor plate Chang" found enhancements in sensi- tivity ranging from none to a factor of 3 for a variety of elements at 3 atmospheres of Ar. This lack of a uniform increase in sensitivity made it impossible to reach any con- clusions with respect to the effect of pressure on integrated absorbance. The issue was further confused by the use of non-optimum furnace atomization conditions and analyte concentrations. Journal of Analytical Atomic Spectrometry March 1995 Vol.10 187Recently an instrument has been described which uses a continuum source an echelle spectrometer and a linear photo- diode array (CS-LPDA).''*l3 This instrument offers simul- taneous measurement of intensities over a relatively broad wavelength range (approximately 1 nm) surrounding the absorption profile permitting the computation of wavelength- integrated absorbance. Regardless of the broadening and shift- ing of the absorption profile at elevated pressures accurate wavelength-integrated absorbances are still achieved. With a large spectral bandwidth (500 pm entrance slit-width) the convoluted absorption profile at the focal plane is relatively unaffected by changes due to pressure. This permits the same set of pixels to be used in the absorbance calculation at all pressures. Wavelength integrated absorbance is normalized for the spectral width of each pixel in the same manner that time- integrated absorbance is normalized for the sampling period.Hence the wavelength- and time-integrated absorbance has units of picometre seconds (pm s). An increase in sensitivity proportional to the increasing pressure does not mean a similar increase in the detection limit. Since the increased sensitivity is derived from a longer analyte residence time a longer integration period is needed to acquire the signal. For the shot noise limited case and concentrations close to the detection limit the integrated absorbance noise is proportional to the square root of the integration time. For example if atomization at 10 atmospheres yields an integrated absorbance that is 10 times larger but 10 times broader then the signal-to-noise ratio will only improve by a factor of fl.Thus the improvement in the detection limit will be critically dependent on the integration limits used by the analyst. In addition as stated in the opening paragraph enhancement of the integrated absorbance proportional to increasing press- ure assumes that diffusion is the only mechanism for loss of analyte atoms from the furnace. Frech and L'vovi4 have recently concluded that convection is a small but significant analyte loss mechanism for the transversely heated graphite atomizer (THGA) (Perkin-Elmer Norwalk CT USA). They have described an axial convection flow from the ends of the tube towards the injection port.Simplistically the heated Ar escaping upwards through the injection port creates a slightly negative pressure in the furnace which draws in cooler Ar from the tube ends. The axial convection flow is a function of the gas densities inside and outside the tube and is therefore dependent on the size of the injection port14 and the square root of the tube temperat~re.'~ Since diffusion is roughly dependent on temperature raised to the power of 2 convection is relatively more important for the more volatile elements with lower atomization temperatures. The presence of a convection component will influence the effectiveness with which increased pressure will increase the integrated absorbance. As pressure increases the rate of diffusion decreases linearly.The contribution of the convection component which remains constant with pressure will rapidly replace diffusion as the dominant loss mechanism. If convection controls the loss of analyte then it is unclear what advantage in sensitivity can be achieved by increasing the pressure. Measures taken to reduce the convection component14 will increase the effectiveness with which the increased atom cell pressure enhances sensitivity. In this study the CS-LPDA instrument was used to deter- mine the effect of elevated atomization pressures on the sensitivities of five elements (Al Cd Cr Mn and Pb). An integrated contact cuvettei6 (ICC) was used in the standard mode with the tube ends open and in the end-capped mode with carbon discs inserted in the ends of the tube to restrict the rate of diffusion of the analyte from the furnace.14 Experiments were run with the injection port open and plugged to restrict convection. The effect of pressure on the chloride interference for Pb and Cr was also investigated.MODEL FOR VAPOUR LOSSES The ICC used in this study16 and the THGA used by Frech and L'vovl4 are very similar in their shape and performance characteristics. The ICC is 17 mm long with a 2 mm diameter injection port compared with an 18 mm length and a 1.8 mm diameter injection port for the THGA. Both have an inside diameter of 6mm. In this study the ICC sits in a sealed vacuum cross in the middle of a large gas volume of uniform pressure while the THGA is surrounded by a flow of Ar. Upon heating the ICC the gas inside the ICC expands outwards against the unheated gas in the vacuum cross.Thus during atomization the pressure inside the vacuum cross will increase slightly although the volume of the vacuum cross is sufficiently large that the increase is insignificant. The flow of Ar across the ends of the THGA may contribute to convection but the ICC has a larger injection port. Thus it is reasonable to assume that the convection flow processes described by Frech and L'vov'~ for the THGA apply equally well to the ICC. The integrated absorbance is proportional to the effective analyte residence time zeff which is determined by the resi- dence times associated with diffusion Td and convection z 10ss.l~ ___ 1 1 1 - -+ - Teff 2d Tc The diffusion residence time is inversely dependent on the diffusion coefficient.Thus since where D is the diffusion coefficient at temperature T Do is the diffusion coefficient at temperature T and II is a constant for each element varying from 1.5 to 2.0 the residence time at temperature T can be shown to be (3) In addition the pressure dependence of the diffusion residence time is (4) where Td is the diffusion residence time at pressure P and zdo is the residence time at pressure Po. Frech and L'vovi4 assumed that convection is driven by the temperature gradient between the inside and outside of the furnace. They then assumed that the gas velocity is proportional to the temperature. It can be shown however that the gas velocity is proportional to the square root of the temperature gradient.15 Thus the tempera- ture dependence of the convection residence time is 0.5 Tc=Tco(:) The convection residence time is independent of pressure to a first approximation.Nothing in the gas expansion with tem- perature suggests a pressure dependence as long as the initial pressures (at room temperature) are the same inside and outside the furnace and the gases can expand freely upon heating. To clarify the relationships given in the preceding eqns. (1)-(5) Table 1 provides data for zd z and zeff as a function of temperature and pressure for a standard ICC. These data assume that n = 1.68 (Cd) zd = 1.0 at 1 atmosphere and 1500 "C and z = 8.62 at 1 atmosphere. The last relationship means 188 Journal of Analytical Atomic Spectrometry March 1995 Kol. 10Table 1 Predicted diffusion convection and effective residence times for standard ICC Residence time/s Zeff Temperature/" C 1500 1700 1900 2100 2300 2500 1 atm 3 atm 6 atm 0.90 2.2 3.5 0.76 1.9 3.1 0.65 1.7 2.7 0.57 1.5 2.5 0.50 1.3 2.2 0.44 1.2 2.0 zd 1 atm 3 atm 1 .o 3 .O 0.84 2.5 0.71 2.1 0.61 1.8 0.53 1.6 0.47 1.4 6 atm 6.0 5.0 4.2 3.7 3.2 2.8 Zc 1 atm 3 atm 6 atm 8.6 8.6 8.6 8.2 8.2 8.2 7.8 7.8 7.8 7.4 7.4 7.4 7.1 7.1 7.1 6.9 6.9 6.9 that loss of analyte by convection is 8.6 times slower than loss by diffusion.This value was taken from data from Frech and L'VOV.'~ They showed that the experimentally sensitivity loss for the THGA as compared with the HGA (a factor of 2.9) was greater than theoretically predicted (a factor of 2.6). Using eqn. (l) it can be shown that the discrepancies in the sensitivit- ies can be accounted for by a convection component 8.6 time weaker than diffusion or a convection residence time 8.6 times larger than the diffusion residence time.This value serves as an initial approximation of the convection component. If diffusion and convection are both present in the ratios assumed above then the integrated absorbances should be proportional to the values for zeff. If only diffusion is present then the absorbances should be proportional to the values for z d . In the latter case increases in pressure should produce proportional increases in integrated absorbance. Table 2 provides similar data for an end-capped ICC.14 For the end-capped ICC it was assumed that zd=5.0 at 1500°C and 1 atmosphere of Ar. This value was chosen because it provides a factor of 3.2 increase in zeff which is in reasonable agreement with the experimental results that will be presented in this study.It was assumed that z = 8.6 the same as for the standard ICC. Fig. 1 provides a graphical display of the computed values for zeff from Tables 1 and 2. Only the extreme temperatures (1500 and 2500 "C) have been shown in Fig. 1 to keep the plots as simple as possible. In Table 2 and Fig. l(a) and (b) it can be seen that the end-capped ICC provides higher effective residence times and that the curvature induced by the convection is more pronounced. This is because convec- tion for the end-capped ICC is relatively larger compared with diffusion (1/8.6 versus 1/5) than it is for the standard ICC (1/8.6 versus l/l).Thus at 1500"C an increase in pressure from 1 to 6 atmospheres predicts an increase in zeff or integrated absorbance by a factor of 3 for the standard ICC and a factor of 2 for the end-capped ICC. EXPERIMENTAL Instrumentation The CS-LPDA instrument has been described previously. l2 The instrument consists of a 300 W xenon arc lamp (Cermax ul . E .- c Q C U 0' 1 I 1 I 1 0 1 2 3 4 5 6 Pressure/atmosDhere Fig. 1 Effective residence time as a function of pressure (1 atmos- phere = 101 325 Pa) temperature and the contributing loss mechan- isms for (a) standard ICC and (b) end-capped ICC A 1500°C (diffusion); B 1500 "C (diffusion and convection); C 2500 "C (diffusion); and D 2500 "C (diffusion and convection) LX3OOUV lamp and PS300-1 power supply ILC Technology Sunnyvale USA) operated at a current of 20 A and an Cchelle spectrometer (Spectraspan V Fisons Instruments Sunland CA USA) fitted with a 256 pixel linear photodiode array (S3904-2564 Hamamatsu Bridgewater NJ USA).Entrance slit dimensions of 500 pm were used. The system is controlled from a PDPll/23 + microcomputer (Digital Equipment Bedford MA USA). Table 2 Predicted diffusion convection and effective residence times for end-capped ICC Residence time/s Zeff TemperaturerC 1 atm 3 atm 6 atm 1 atm 1500 3.2 5.5 6.7 6.0 1700 2.8 4.9 6.2 4.2 1900 2.4 4.5 5.7 3.6 2100 2.2 4.1 5.3 3.1 2300 1.9 3.8 4.9 2.7 2500 1.8 3.5 4.6 2.4 3 atm 6 atm 15 30 12 25 11 21 9.2 18 8.0 16.0 7.1 14.2 1 atm 3 atm 6 atm 8.6 8.6 8.6 8.2 8.2 8.2 7.8 7.8 7.8 7.4 7.4 7.4 7.1 7.1 7.1 6.9 6.9 6.9 Journal of Analytical Atomic Spectrometry March 1995 Vol.10 189A side heated ICC identical with that described by Lundberg et a1.16 and powered by an HGA-500 power supply (Perkin- Elmer Norwalk CT USA) was used throughout this work. The ICC was 17 mm long and had an inner diameter of 6 mm. The ICC was enclosed in a vacuum housing which consists of a six-way stainless-steel cross with flanges on each port. The windows in the flanges are reinforced from both sides to allow operation of the furnace under both vacuum conditions and at elevated pressures of up to 10 atmospheres. Two types of ICCs were used in this work. Standard ICCs with the tube open at both ends and end-capped14 ICCs with carbon discs in the open ends of the tube. The discs were 1.5 mm thick and had an aperture of 3.2 mm.The discs were inserted in the factory prior to the pyrolitic graphite coating process. Neither type of ICC were supplied with platforms. Platforms from a slotted style HGA furnace were laid on the wall of the standard ICCs. This approach was not entirely satisfactory since the platforms never became attached to the wall and occasionally shifted position introducing an additional source of uncertainty into the signal measurement. No plat- forms were used with the end-capped ICC since it was imposs- ible to get the platforms through the end-caps into the furnace. The heating rates of both the standard and end-capped ICC were less than that of the HGA or the THGA since the impedance match of the ICCs and the HGA-500 power supply was less than optimum.With maximum power heating using the sensor diode the ICC was heated from room temperature (296 K) to 2770 K in approximately 2 s (approximately 1200 K s-'). Absorbance Calculations Wavelength integrated absorbance is computed for each scan of the array by summing the absorbance for each of the pixels over the absorption pr~file.'~.'~ Time integrated absorbances are computed in the normal manner by summing the wave- length integrated absorbances over the atomization cycle. This value is then multiplied by the spectral bandwidth (SBW) of a 25 pm pixel (2.2-3.8 pm for the elements determined in this study) and 0.02 s (50 Hz absorbance computation frequency for all elements) to normalize for time and dispersion. All integrated absorbances are reported in units of pm.Standard Solutions and Reagents Standard solutions were prepared from 1000 pg ml-' stock solutions of A1 (as chloride) and Cd and Pb (as nitrate) supplied by J. T. Baker (Phillipsburg NJ USA) and of Cr and Mn (as nitrate) supplied by Inorganic Ventures. Standards were pre- pared in nitric acid supplied by Seastar Chemicals (Seattle WA USA) so as to allow a final concentration of 2% m/v HNO,. A 200 pg ml-' Pd(N03)2 solution was prepared from a 5000 pg ml-' Pd(N03)2 solution supplied by Inorganic Ventures. Solutions of 3% m/v NaCl were prepared from the solid supplied by J. T. Baker. Methodology Sample solutions were deposited onto a platform in the standard ICCs and onto the furnace wall in the end-capped ICCs. The chamber was evacuated to 100 m Torr (1 mTorr = 0.133 Pa) to dry the sample and to remove the air present.The chamber was then filled with Ar to the desired pressure and the sample atomized under maximum power using the con- ditions listed in Table 3. Neither a drying or a pyrolysis step were used. The chamber was then re-evacuated and the furnace fired again as a clean-out step. Finally the chamber was brought back up to atmospheric pressure with Ar the lid removed and the next solution deposited in the ICC. Table 3 Atomization temperatures and integration times Element A1 Cd Cr Mn Pb TemperaturePC 2500 2000 2500 2400 2000 Time/s 8-18 5-12 8-20 8-15 5-12 RESULTS AND DISCUSSION Standard ICC The integrated absorbances shown in Table 4 were measured for the atomization of Al Cd Cr Mn and Pb using a standard ICC with a platform at applied pressures of 1 2 3 and 6 atmospheres of Ar.Each absorbance represents the average of multiple determinations (n > 3). Figs. 2 and 3 make it easier to see the pattern of the increase in sensitivity as a function of pressure. In each figure the integrated absorbances have been normalized by division of the values obtained at 1 atmosphere. In Fig. 2 the plots for Al Cr and Mn have similar shapes with A1 showing a slightly greater increase in the signal with pressure. These plots for integrated absorbance are not entirely consistent with those of the predicted effective residence times in Fig. 1. Although the overall increase in sensitivity at 6 atmospheres is reasonable (a factor of 2.5-3) indicating that 8.6 is a reasonable estimate of the convection residence time Table 4 Wavelength and time-integrated absorbances (pm s) versus pressure for a standard ICC Integrated absorbance/pm s A1 Cr Mn Pb Cd Concentration*/ng ml- 20 10 10 40 5 Spectral bandwidtht/pm Pressure/atmosphere Open Injection Port:- 1 2 3 6 Plugged Injection port 1 2 3 6 3.2 0.21 1 0.426 0.666 0.744 0.93 1 1.82 2.97 3.19 - 3.8 0.198 0.380 0.513 0.578 0.749 1.03 1.84 1.98 2.8 0.473 0.946 1.20 1.35 0.563 1.26 1.61 1.75 2.8 0.358 0.549 0.510 0.378 0.378 0.725 0.848 1.08 2.2 0.636 1.17 1.38 1.23 0.660 1.31 1.82 2.51 * 10 pl injections used.t In pm for a 25 pm pixel. a 4.0 I u - C (0 3.5 2 3.0 2 (0 2 2.5 +- (0 f Cr Pb k 0.5 z 0 1 2 3 4 5 6 Pressure/at mosp here Fig.2 Normalized absorbance for a standard ICC as a function of Ar pressure.Normalization by absorbance at 1 atmosphere 190 Journal of Analytical Atomic Spectrometry March 1995 Vol. 10w 4.0 C ; 3.5 0 (0 3.0 2.5 w 5 2.0 3 .w .' 1.5 -0 - 1.0 0.5 .- - - - - - - - 0 1 2 3 4 5 6 Pressure/atmosphere Fig. 3 Normalized absorbance for the standard ICC with the injection port plugged as a function of Ar pressure. Normalization by absorbance at 1 atmosphere (compared with a diffusion residence time of l) the shapes are different. Experimentally the increase in sensitivity between 1 and 3 atmospheres is too great and the increase in signal above 3 atmospheres is too little. The integrated absorbances in Fig. 2 for Cd and Pb the most volatile elements reach a maximum at 3 and 2 atmos- pheres of Ar respectively and then decrease.At 6 atmospheres of pressure Pb has returned to the signal level at 1 atmosphere. These decreases with pressure were found consistently and were significant compared with the experimental precision. These experimental results are in obvious disagreement with the predicted results in Table 1 and Fig. l(u). The model cannot produce a reversal in sensitivity for the standard ICC under any circumstances. The only possible means of inducing rever- sal is to assume that the convection component increases with pressure. Such an increase in convection with pressure could also explain the unexpectedly small increase in sensitivity for Al Cr and Mn above 3 atmospheres. The degree of increase in convection need not be much. A 20% increase in convection with a 100% increase in pressure would provide sufficient reversal in the model to agree with the experimental results for Cd.Using the assumptions of the model [eqns. (2)-( 5 ) and that z is constant with respect to pressure] eqn. (1) can be used to estimate the values of z for each of the elements in Fig. 2. At 6 atmospheres of pressure z d = 6 and the ratio of the integrated absorbances at 6 and 1 atmospheres is an approxi- mation of zeE. Solving for z yields the values shown in Table 5. The convection component appears to be strongest (z is smallest) for the most volatile elements and grows less severe as the elements become less volatile. This suggests that axial convection is strongest early in the atomization cycle. Axial convection should not be confused with convection due to gas expansion at the start of the atomization cycle.'' Calculations using the equations of Gilmutdinov and Fishman19 have shown that with a platform the expansion convection component is not significant for even the most volatile elements.The com- puted values for z for Cd and Pb are roughly a factor of 4; low compared with the value of 8.6 derived from the work of Frech and L'vovI4 and used in the model. In order to eliminate loss of analyte by convection the injection port of the furnace was plugged using a small piece of graphite rod. The effect of pressure on integrated absorbance was then measured for the same five elements. Integrated absorbances as a function of pressure are shown in Table4 and plots of the normalized integrated absorbances as a function of pressure are shown in Fig.3. Each absorbance again represents the average of multiple determinations (n = 3 or more). Plugging the injection port eliminates the reversal for Cd and Pb but produces little change in the plots of Al Cr and Mn. None of the elements show the linear increase in integrated absorbance that would be expected if convection had been eliminated. The Cd and Pb results suggest a dimin- ishment but not an elimination of the convection component while the results for Al Cr and Mn suggest no change in any of the loss mechanisms as a result of plugging the injection port. These observations are supported by the estimates of z in Table 5. The computed values for z for Cd and Pb are now in better agreement with those for Al Cr and Mn which are essentially unchanged.Plugging the injection port also introduces a second surface effect.20 The carbon plug tends to heat more slowly than the tube wall and initially serves as a site for condensation of the analyte. As the plug heats further the analyte is volatilized a second time from this second surface. In this study the second surface phenomenon was observed as a broadening of the analyte peak i.e. the atomization peaks from the platform and plug were unresolved. The primary effect of the second surface phenomenon is to delay the atomization time of the analyte. This effect is more pronounced for the more volatile elements Cd and Pb. Another perspective of the data in Table 3 is presented in Fig. 4 where the absorbance ratios injection port plugged/ injection port open are plotted as a function of Ar pressure for all five elements. These results show two different trends.The first trend is that the ratios for the most volatile elements Mn 0 1 2 3 4 5 6 Pressu re/atmosp here Fig. 4 Absorbance ratio (injection port plugged/injection port open) as a function of pressure Table 5 Computed convection time constants (z,~) relative to diffusion (1.0) Standard ICC End-capped ICC Element Open port Plugged port Open port Plugged port A1 8.5 Cr 5.7 Mn 5.4 Pb 1.3 Cd 2.9 8.0 4.7 6.5 5.5 10.0 - 3.2 3.8 NC* 1.4 - 6.0 14.0 30.0 23.0 * NC = not computable; result is negative indicating absorbance is less at 6 atmospheres than at 1 atmosphere. Journal of Analytical Atomic Spectrometry March 1995 Vol. 10 I 9 1Cd and Pb show a steady increase with pressure.These data suggest that convection is reduced by plugging the injection port. At 1 atmosphere where diffusion is dominant there is little difference between zeff and zd and elimination of z will have little influence on zeff (Table 1). As pressure increases and zeff diverges increasingly from zd elimination or reduction of the convection component can result in a progressive increase in zeff. Consequently the results for Cd and Pb in Fig. 4 suggest a reduction in the convection component and are consistent with the data in Table 5. The second trend observed in Fig. 4 is that the less volatile elements Al Cr and Mn show no increase or a slight decrease in the absorbance ratio with increasing pressure. For Mn the absorbance ratio at 1 atmosphere is slightly greater than unity and then increases slightly and decreases slightly as the pressure increases to 6 atmospheres.This fairly constant absorbance ratio suggests that there is little change in the loss mechanism with the injection port open or closed. This is consistent with the data (for Al Cr and Mn) in Table 5 and is directly opposed to the conclusion reached for Cd and Pb. A1 and Cr show the same trend observed for Mn in Fig. 4; the absorbance ratios decrease slightly with increasing pressure. For A1 and Cr however the absorbance ratios at 1 atmosphere are close to 4. For a diffusion controlled system estimates of analyte lost through the injection port range from 17 to 34%.3,21,23 An increase in sensitivity by a factor of 4 suggests that convection is dominant in the standard ICC and that it is eliminated by plugging the injection port.However the fact that the absorbance ratio remains constant with increasing pressure suggests that there is little change in the loss mechan- ism with the injection port open or closed. Also the increase in integrated absorbances for A1 and Cr as a function of pressure (Figs. 2 and 3) did not change when the injection port was plugged. Thus it seems clear that convection was relatively unchanged by plugging the injection port. The large enhancement of A1 and Cr remains unexplained. End-capped ICC The effect of increasing gas pressure on the sensitivity of Cd Pb Mn and Cr was further investigated using end-capped ICCs. These modified furnaces have been used by Frech and L‘vov to increase the gas-phase temperature near the ends of the THGA and to reduce matrix trapping.It was not possible to place platforms in the end-capped ICCs used in this work and so the samples were deposited on the furnace wall. Table 6 shows the integrated absorbances obtained for the end-capped ICCs as a function of pressure. Figs. 5-8 show the normalized integrated absorbances for Cd Pb Mn and Cr as a function of pressure for the standard and the end-capped ICCs with the injection port open and plugged. The integrated absorbances were normalized using the absorbances obtained at 1 atmosphere of pressure with an open injection port and a standard ICC. Thus the plots reflect sensitivity enhancements due to increasing pressure plugging of the injection port and use of end-capped tubes.With the injection port open the end-capped ICC provides an average increase in sensitivity of a factor of 3.2 over that of the standard ICC (injection port open) for the four elements studied here. A strong convection component is also still in evidence. Ignoring the increase in sensitivity the shape of the plot of integrated absorbance as a function of pressure is very similar for the end-capped ICC and the standard ICC. This is also shown in Table 5 where the estimates for z for the two types of cuvettes with the injection port open are very similar. The flattening of the data above 3 atmospheres pressure is again worse than predicted by the model. This flattening can only be accounted for if the original estimate of z is decreased Table 6 Wavelength and time-integrated absorbances (pm s) versus pressure for an end-capped ICC Pressure/atmosphere Open Injection Port:- 1 2 3 6 1 2 Plugged Injection Port:- Integrated absorbance/pm s Cr Mn Pb Cd Concentration*/ng ml- 10 10 40 5 Spectral bandwid t h”fpm 3.8 2.8 2.8 2.2 0.642 1.55 1.28 2.02 1.18 2.79 1.19 3.30 1.38 3.11 0.543 2.88 1.34 3.61 0.308 2.31 1.03 3.08 1.50 2.73 1.44 5.46 2.88 5.83 2.58 6.94 3.98 7.52 2.57 12.9 7.50 13.0 *10 p1 injections used.7 In pm for a 25 pm pixel. to approximately 4.0 (a factor of 2 increase in the convection component) or if z = 8.6 but decreases slightly with pressure. The one major deviant from the general agreement of the behaviour for the standard and endcapped ICCs with the injection port open is Pb.A strong decrease in the integrated absorbance is observed with increasing pressure that far exceeds that observed for Pb or Cd in the standard tube. These data yielded a negative value for zc since only an increasing convection component could produce a decreasing sensitivity. With the injection port plugged each element with the exception of Cr shows a nearly linear increase in signal with pressure. In addition the data in Table 5 show a dramatic increase in 7,. Thus it appears that plugging the injection port of the end-capped ICC drastically reduces the convection component and leaves diffusion as the dominant loss mechanism. Plugging the injection port of the end-capped ICC produces an average increase in signal of 1.5 for the four elements (Fig. 5-8).The dramatic difference in enhancement between elements observed upon plugging the injection port of the standard ICC (Fig. 4) was not observed with the end-capped ICC. Table 2 predicts a slight increase for the more volatile elements (an increase from 1.3 at 2500°C as compared with 1.5 at 1500°C). Experimentally the enhancement for Cr was 1.6 and that for Cd and Pb was 1.4 and 1.2 respectively. At 6 0 1 2 3 4 5 6 Pressu re/atmosp here Fig.5 Normalized absorbance for Cd as a function of Ar pressure standard ICC with A injection port open and B injection port plugged and end-capped ICC with C injection port open and D injection port plugged. Normalization by absorbance at 1 atmos- phere standard ICC injection port open 192 Journal of Analytical Atomic Spectrometry March 1995 VoZ.100 18 - & 16 - v) I) 73 m 14 h 0) '0 c 8 - - 12 - 4- - .w .- c I /-/ C / A CI I I I 1 I + 0 1 2 3 4 5 6 Pressure/atmosphere Fig. 6 Normalized absorbance for Pb as a function of Ar pressure standard ICC with A injection port open and B injection port plugged; and end-capped ICC with C injection port open and D injection port plugged. Normalization by absorbance at 1 atmos- phere standard ICC injection port open m 60 I 1 0 1 2 3 4 5 6 Pressure/atmosphere Fig. 7 Normalized absorbance for Mn as a function of Ar pressure standard ICC with A injection port open and B injection port plugged; and end-capped ICC with C injection port open and D injection port plugged. Normalization by absorbance at 1 atmos- phere standard ICC injection port open 14 1 I 1 I I 1 I 1 I 0 1 2 3 4 5 6 Pressure/atmosphere Fig.8 Normalized absorbance for Cr as a function of Ar pressure standard ICC with A injection port open and B injection port plugged; and end-capped ICC with C injection port open and D injection port plugged.Normalization by absorbance at 1 atmos- phere standard ICC injection port open atmospheres the model (Table 2) predicts enhancement factors of 3.1 at 2500 "C and 4.5 at 1500 "C for plugging of the injection port. Experimentally values of 1.9 3.5 and 5.6 were observed for Cr Mn and Cd. The low results for Pb with the injection port open made it impossible to compute a result. The only element that failed to show a linear increase in sensitivity with increasing pressure was Cr. Every plot for Cr in Fig.8 is linear to 3 atmospheres (an increase of 2.5 in sensitivity) and then shows no change with an increase in pressure to 6 atmospheres. A possible explanation for this flattening is a spectral interference in the wings of the analyte absorption profile. Absorbance by a polyatomic species in this region will reduce the reference intensity and reduce the analyte absorbance. It is well documented that a CN band exists close to the Cr line. It is possible that as pressure increases the temporal overlap of the interfering band increases. Initial studies have not yielded any obvious interferences but more studies are planned. Without a platform in the end-capped ICC the second surface effect of the plug is readily observed for Cd and Pb. For both two fully resolved atomization peaks one from the wall and one from the plug are observed.For the other elements the resolution of the double peaks decreases with decreasing volatility of the element and with increasing press- ure. For Cd and Pb the second peak occurs significantly later than the first peak (approximately 0.75 s). Thus the plug delays volatilization until higher temperatures are reached. This offers one possible explanation of the Pb results. The extreme decrease in integrated absorbance observed for Pb with increas- ing pressure was clearly reversed when the injection port was plugged. An explanation of the Pb data based on physical phenomena is difficult since Cd a slightly more volatile element does not show the same pattern. It is possible that without a plug Pb is subject to a chemical or spectral interference.With the plug and higher atomization tempera- tures the interference is significantly reduced. Further research is necessary to investigate this interference. Chloride Interferences This study has focused primarily on explaining the data on the basis of physical factors. It is reasonable to assume however that the increased pressure will also have a significant effect on chemical reactions in the gas and condensed phases. These chemical changes can lead to significant changes in the population of the analyte. In an initial attempt to investigate changes in the chemistry the classic interference of chloride on Pb and Cr was examined as a function of pressure. Fig. 9 presents the recovery of Pb in a 3% m/v NaCl solution as a function of increasing gas pressure using the standard ICC with the injection port open and plugged.These results show that as pressure is increased the recovery of Pb increases significantly. Plugging the injection port has very little effect on the recovery of Pb from the chloride matrix. It has been suggested that the mechanism of chloride interference is the formation of volatile chloride species that are expelled by the rapid evolution of matrix gases.24 Atomization at elevated pressure offers two advantages. The increased resi- 110 1 100 90 80 2 $ 70 60 50 0 [r 40 0 1 2 3 4 5 6 7 Pressure/atmosphere Fig. 9 and B injection port plugged Recovery of Pb in 3% m/v NaCl with A injection port open Journal of Analytical Atomic Spectrometry March 1995 Vol.10 193Table 7 Comparison of sensitivities (pg) Pressure/ Instrument Furnace atmosphere 5000 (ref. 25) HGA 1 4100ZL (ref. 26) THGA 1 CS-LPDA (ref. 13) HGA 1 CS-LPDA (this work) ICC-Std 1 CS-LPDA (this work) ICC-Std 6 CS-LPDA (this work) ICC-EC 6 Injectioin port A1 Cd Open 10.0 0.35 Open 31.0 1.3 Open - 0.60 Open 13.0 0.80 Plugged 0.87 0.20 Plugged - 0.04 Cr Mn Pb 3.3 2.2 12.0 7.0 6.3 30.0 - 2.3 11.0 8.5 2.6 14.0 0.85 0.70 4.6 0.65 0.10 0.64 Table 8 Comparison of detection limits (pg) Pressure/ Instrument Furnace atmosphere 5000 (ref. 25) HGA 1 4100ZL (ref. 26) THGA 1 CS-LPDA (ref. 13) HGA 1 CS-LPDA (this work) ICC-Std 1 CS-LPDA (this work) ICC-Std 6 CS-LPDA (this work) ICC-EC 6 Injection port Open Open Open Open Plugged Plugged Detection (pg) A1 Cd 1 .o 0.3 5.0 0.3 0.4 1 .o 0.5 0.1 0.2 - 0.05 - Cr Mn Pb 1 .o 1 .o 5.0 2.0 2.0 3.0 - 0.5 0.9 0.8 0.8 1 .o 0.1 0.3 0.5 0.1 0.03 0.09 0 1 2 3 4 5 6 7 Pressure/atmosphere Fig.10 Recovery of Cr in 3% m/v NaCl with A injection port open and B injection port plugged dence time of the analyte in the furnace will increase the probability of the Pb chloride species being reduced to Pb atoms. In addition the delayed appearance time due to pressure or second surface atomization will result in higher furnace temperatures when the Pb is vaporized. Fig. 10 presents the recovery of Cr in a 3% m/v NaCl solution as a function of increasing gas pressure using a standard ICC with the injection port open and plugged. In this case at all pressures the recovery of chromium is signifi- cantly higher with the injection port plugged than with it open.Plugging of the injection port which appears to dramatically increases the residence time and sensitivity of Cr (Fig. 4) also favours the decomposition of any analyte chloride species (Fig. 10). Thus the physical effect of the longer residence time appears to have a favourable influence on the chemistry for both Cr and Pb. Figures of Merit Tables 7 and 8 provide a comparison of sensitivities and detection limits obtained for Al Cd Cr Mn and Pb for a progression of furnaces and instrumental systems. Table 7 shows that compared with a standard ICC with open injection port at 1 atmosphere of pressure an end-capped TCC with plugged injection port at a pressure of 6 atmospheres provides characteristic mass improvement factors of 20 13 14 and 22 for Cd Cr Mn and Pb respectively.Although all absorbance data in this study have been reported in units of pms characteristic mass (pg s-') has been used instead of intrinsic m a d 3 (pg pm -' s - I). Fortunately characteristic and intrinsic mass are very similar for this echelle and a 25 pm pixel width. The new sensitivity improvements shown in Table 7 did not result in comparable improvements in the detection limits owing to the necessary increases in the integration time. The detection limits reported for the CS-LPDA with the HGA13 were based on a 5 s integration time for all elements. The range of integration times used for the CS-LPDA with the ICC between 1 and 6 atmospheres are listed in Table 3. The baseline absorbance noise levels were unaffected by the press- ure but as shown in Table 3 longer integration times were necessary as the pressure increased.Still significant improve- ment has been noted for the detection limits for all elements. The improvement is approximately an order of magnitude compared with the commercially available instruments. REFERENCES 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 L'vov B. V. Spectrochemical Analysis by Atomic Absorption Spectrometry Adam Hilger London 1970. Sturgeon R. L. Chakrabarti C. L. and Bertels P. C. Spectrochim. Acta Part B 1977 32 257. Sturgeon R. L. and Chakrabarti C. L. Prog. Anal. Atom. Spectrosc. 1978 1 5. Hoenig M. Vanderstappen R. and van Hoeyweghen P. Analysis 1978 6 433. Fazakas J. Spectrochim. Acta 1982 37B 921. Fazakas J. Spectrochim. Acta 1983 38B 455. Fazakas J. and Zugravescu P. G. Spectrochim. Acta 1988 43B 897. Fazakas J. and Hoenig M. Talanta 1988 35 403. Fazakas J. and Zugravescu P. G. Appl. Spectrosc. 1977 42 521. Chang J-C and O'Haver T. C. J. Anal. At. Spectrom. 1990,5663. Chang J-C PhD Thesis University of Maryland 1990. Harnly J. M. J. Anal. At. Spectrom. 1993 8 317. Smith C. M. M. and Harnly J. M. Spectrochim. Acta 1994 49B 387. Frech W. and L'vov B. V. Spectrochim. Acta 1993 48B 1371. Gilmutdinov A. private communication. Lundberg E. Frech W. and Harnly J. J. Anal. At. Spectrom. 1988 3 1115. L'vov B. V. and Frech W. Spectrochim. Acta 1993 48B 425. Holcombe J. A. Spectrochim. Acta 1983 38B 609. Gilmutdinov A. Kh. and Fishman I. S. Spectrochim. Acta 1984 39B 171. Holcombe J. A. and Sheehan M. Y. Appl. Spectrosc. 1982 26 631. 194 Journal of Analytical Atomic Spectrometry March 1995 Vol. 1021 22 Baxter D. C. and Frech W. Spectrochim. Acta 1987 42B 459. Frech W. L‘vov B. V. and Romanov N. P. Spectrochim. Acta 1992,47B 1461. 26 Perkin-Elmer Corporation 1991 ‘Recommended Conditions for THGA Furnaces’ Part Number B050-6158 Publication B3110.06. 23 Guell 0. A. and Holcombe J. A. Spectrochim. Acta 1988 43B 459. 24 Welz B. Akman S. and Schlemmer G. J . Anal. At. Spectrom. 1987 2 793. 25 Perkin-Elmer Corporation 1988 ‘Recommended Conditions for Zeeman Background Correction (HGA Furnaces)’ Part Number 0993-8199 Rev B. Paper 41051 78E Received August 28 1994 Accepted October 24 1994 Journal of Analytical Atomic Spectrometry March 1995 Vol. 10 195

ISSN:0267-9477    

DOI:10.1039/JA9951000187

出版商:RSC    

年代:1995

数据来源: RSC

摘要:

Effect of Furnace Atomization Temperatures on Simultaneous Multielement Atomic Absorption Measurement Using a Transversely= Heated Graphite Atomizer* Journal of Analytical Atomic Spectrometry JAMES M. HARNLY USDA ARS Beltsville Human Nutrition Research Center Nutrient Composition Laboratory Building 161 BARC-East Beltsville MD 20705 USA BERNARD RADZIUK Bodenseewerk Perkin-Elmer Corporation GmbH Postfach 101 761 0-88647 Uberlingen Germany The sensitivities and signal-to-noise ratios (SNRs) of five elements (Cd Pb Cu Cr and V) were characterized for atomization by a commercially available transversely heated graphite atomizer (THGA) as a function of the atomization temperature with and without a Pd-Mg(N03)2 matrix modifier and with a standard furnace and an "end-capped" furnace to restrict diffusional loss of the analyte.These five elements were selected to provide a representative range of atomization temperatures. In general the sensitivities of Cd Pb Cu and Cr decreased with increasing atomization temperature and the sensitivity of the least volatile element V increased with temperature. The most suitable temperature for simultaneous determination of these elements was 2500 "C as dictated by V. The loss in sensitivity for the atomization of Cd and Pb at 2500°C was 25 to 35% considerably less than predicted by mass diffusion. With use of the modifier and 'end- capped' THGAs the temperature dependence of the sensitivity approached that predicted by mass diffusion. With photon shot noise and optimization of the integration interval the best SNRs for all elements in the simultaneous multielement mode or in the single element mode are found at approximately 2500 "C and no compromise is necessary. Keywords Electrothermal atomic absorption spectrometry; multielement; transversely heated graphite atomizer A potential hindrance to simultaneous multielement determi- nations by electrothermal atomic absorption spectrometry (ETAAS) is the necessity of using compromise furnace tempera- tures for multiple analytes.In the traditional single-element mode of operation the furnace temperatures for the drying pyrolysis and atomization steps are optimized for the element to be analysed and the sample matrix. In the drying and pyrolysis steps temperatures gas types and flows and chemical additives (matrix modifiers) are selected to provide reproduc- ible drying and to remove or minimize the effect of matrix components which provide chemical and/or spectroscopic interferences. In the atomization step the temperature and gas flow are selected which optimize the analytical sensitivity.Thus selection of temperature is critical to the optimization of sensitivity and the minimization of interferences. In contrast simultaneous multielement determinations demand compromise furnace temperatures. The degree of compromise is a function of the suite of elements to be * Presented at the Seventh Biennial National Atomic Spectroscopy Symposium Hull UK 20-22 July 1994. determined. The more extreme the least and most volatile elements are the less room there is for compromise.Thus Cd and Pb dictate that the pyrolysis temperature be less than 450°C (without a modifier) and B and Ti dictate that the atomization temperature be close to 2500 "C. If these elements are to be included in the determination there is little room for change in these temperatures. Consequently depending on the suite of elements to be determined the compromise parameters may be dramatically different than individually optimized single element conditions. The questions of interest are (i) how severely will the sensitivity and the detection limit of the volatile elements be degraded using an atomization tempera- ture suitable for the least volatile element and (ii) how severely will the chemical and spectroscopic interferences be increased using a pyrolysis temperature suitable for the most volatile element? This study will concentrate primarily on the first question. The first detailed investigation into the use of compromise multielement furnace temperatures by Harnly and Kane,' examined the variation of peak height and peak area (inte- grated) sensitivities as a function of the atomization tempera- ture for a conventional longitudinally-heated graphite furnace (Model HGA-2100 Perkin-Elmer Corp.Norwalk CT USA). It was determined that with or without a platform a maximum pyrolysis temperature of 500 "C and a minimum atomization temperature of 2700°C were necessary for a suite of nine elements (Co Cr Cu Fe Mn Mo Ni V and Zn). The maximum pyrolysis temperature (500 "C) was determined by Zn and the minimum atomization temperature (2700°C) was determined by Mo and V.With these parameters the peak area sensitivity of Zn was reduced by 50% (as compared to the maximum sensitivity at an atomization temperature of 1500 "C) for platform atomization and by less than 5% for atomization from the wall. These parameters were subsequently used for a prototype multielement continuum source instru- ment for the determinations of trace elements in a wide variety of biological and food The limitation of a low pyrolysis temperature and restricted ability to remove trouble- some background was offset by the inherent background correction capability of the prototype continuum source instrument. These early studies demonstrated that the longitudinally- heated furnace is not ideal for multielement determinations.Although platform atomization offered an approximation of temporal isothermality these tubes are far from spatially isothermal. A temperature drop of 1200 "C between the middle and the ends of the tube has been documented for atomization at 2700 "C6 As a result elevated atomization temperatures are Journal of Analytical Atomic Spectrometry March 1995 Vol. 10 197necessary to determine even the most volatile elements. Still the cooler ends of the tube are sites of condensation for the less volatile elements and less volatile components of the matrix. This condensation can result in elemental memory between atomizations for all but the most volatile elements and can give rise to increased matrix interferences. For multiele- ment determinations where a large dynamic range is desired to eliminate the need for multiple dilutions memory between atomizations is a critical problem.' A recent investigation by Berglund et aL7 examined the use of a commercially available transversely heated graphite atom- izer (THGA) with palladium as a matrix modifier for the simultaneous determination of 8 elements (Ag Cd Co Cr Mn Mo Pb and V) using a single set of atomization param- eters.The use of palladium allowed a pyrolysis temperature of 900°C and the use of a transversely heated furnace made possible an atomization temperature of 2400 "C and dramati- cally reduced memory between atomizations (compared with a longitudinally heated furnace). Experimentally determined characteristic masses for atomization at 2400 "C were com- pared with calculated values for the same temperature and with calculated values for optimal single element temperatures.They showed that the loss in sensitivity ranged from 10% to 20% for Cd Co Cr Pb and V and was negligible for Ag Mn and Mo. They concluded that the transversely heated furnace offers better prospects for multielement determinations than the longitudinally heated furnace. The necessity of using a matrix modifier for multielement ETAAS is not clear cut. A modifier often produces a more heat stable form of the analyte that allows higher pyrolysis temperatures and like platform atomization delays volatiliz- ation of the analyte until the furnace has reached a higher stabilized temperature. Both of these features may significantly reduce interferences and background attenuation.Atomization at a higher temperature however will reduce the analyte integrated sensitivity due to the higher diffusion coefficient and the shorter residence time. In some cases the modifier may also delay volatilization of the matrix and lead to co-volatilization of the matrix with the analyte. This can lead to interferences more serious than those observed without the modifier. In addition recent evidence suggests that the analyte may be trapped or co-condense with the modifier in the cooler gases at the ends of the tube even with a THGA.8-9 Thus the use of a matrix modifier may not be entirely beneficial. One means of increasing the sensitivity of the THGA and minimizing matrix trapping is to restrict the size of the openings at the ends of the tube.'*'' This is done by inserting carbon discs with small apertures (approximately 3.2 mm) in each end of the tube prior to the pyrolytic graphite coating process.The 'end-caps' reduce the open area thus reducing the rate of diffusional loss and increasing the peak area sensitivity of the THGA by factors of 1.7 to 2.5. This restores the sensitivities of the THGA to levels comparable to the HGA. Without end- caps the standard THGA is less sensitive than the HGA because of the shorter tube length (18 mm for the THGA 28 mm for the HGA) dictated by the use of a longitudinal magnetic field for Zeeman-effect background correction. The end-caps also serve to raise the gas phase temperature at the ends of the tube minimizing condensation in this region. In this study the suitability of one set of furnace parameters for the simultaneous determination of five elements (Cd Cr Cu Pb and V) was evaluated with respect to sensitivity and signal-to-noise ratios (SNRs) using a commercially available AAS with hollow cathode lamps as sources a THGA and longitudinal Zeeman-effect background correction.Studies were conducted with and without a Pd-Mg(NO,) matrix modifier and with standard and end-capped THGAs. The temperatures experienced by the analytes were examined by 198 Journal of Analvtical Atomic Svectrometrv. March 1995. comparing the peak shape with the temperature-time plots of the wall and platform. Sensitivity losses at higher temperatures due to diffusion and condensation with the modifier were quantified. Signal-to-noise ratios were evaluated as a function of temperature and integration time. EXPERIMENTAL Instrumentation In this study a Model 4100ZL atomic absorption spectrometer (Perkin-Elmer Corp.Norwalk CT USA) equipped with a THGA and a longitudinal Zeeman-effect background correc- tion system was used. Two types of THGAs were used; the conventional design with the tube open at the ends and an 'end-capped' design' with carbon disks restricting the aperture at both ends to 3.2 mm. Methods Sensitivity measurements All sensitivity measurements were made in the 'Test Mode' using the furnace programme shown in Table 1. The Test Mode permits the specified temperature to be increased system- atically. Initially the atomization temperature was increased from the lowest to the highest temperature in increments of 100 "C.In later experiments the temperature steps were increased to 200 "C. Ten repetitions were made at each tempera- ture. The lowest temperature varied according to the element. Integration times of 10 or 15 s were adjusted as necessary at low temperatures. In all cases 20 p1 sample sizes were used. The same Test Mode was used to determine the maximum allowable pyrolysis temperature for each element with and without a modifier. It was determined that 400°C was the maximum pyrolysis temperature for multielement determi- nations without a modifier and 700°C was the maximum pyrolysis temperature with a modifier. Peak time measurements All time measurements for appearance peak maximum and half height front and back slope were made from traces of the atomization signal.A time resolution of 0.02 s was obtain- able which was comparable to the instrument resolution with a 54 Hz absorbance computation frequency. Consequently all times are listed to the nearest 0.02 s. The full width at half height (At+) was computed as the difference between the front and back slope half height times. Pyrometric temperature measurements Continuous temperature measurements were made with a Model 1130 automatic optical pyrometer (Ircon Inc. Niles IL USA) with a P-4 objective lens assembly. Three temperature modules were used which provided linearization over the ranges 900-1600 "C 1100-2000 "C and 1500-3000 "C. The Table 1 Furnace programme for standard and end-capped THGA* Step Ramp time/s Hold time/s TemperaturerC Dry (1) 20 10 140 Dry (11) 10 10 200 P y r o 1 y s i s 5 15 400/700t Cool down 1 9 20 Atomization 0 5 Varied Clean out 1 3 2500 * 20 1 sample sizes used in all cases. No modifier/With modifier.Matrix modifier was 5 pg Pd and 3 pg Mg(NO,),. VOl. 10continuous signal from the pyrometer was sampled at approxi- mately 54Hz using a microprocessor equipped with an ana- logue-to-digital converter. Data acquisition was initiated by a pulse from the 4100ZL 2 s prior to the atomization cycle and lasted for 8 s. All platform and wall temperatures were acquired using emissivities of 0.93 and 0.77 respectively. Measured temperatures recorded with the automatic pyrometer and these emissivities compared well with those obtained using a van- ishing filament pyrometer.Platform temperatures were meas- ured with the pyrometer focused on the platform through the dosing hole. Wall temperature measurements were made with the tube mounted in an upside down position; the pyrometer was focused on the outside wall of the tube exactly opposite the position where the platform temperatures were measured. The platform temperatures were not corrected for radiation reflected from the wall. The overshoot of the measured platform temperatures most noticeable between 1300 and 1900 "C strongly suggested that radiation from the tube wall was being reflected from the platform and that the overshoot of the platform temperature was simply a reflection of the overshoot of the tube wall temperature. Application of Falk's correction formula6 reduced the apparent temperatures at all times but failed to change the shape of the profiles i.e.the overshoot persisted. Additional temperature measurements were made with the top third of the furnace removed. This eliminated the source of the reflected radiation. In this configuration the overshoot of both the wall and the platform increased. If it is considered that the THGA platform is concentric to the tube wall is less than 1 mm from the wall and covers 40% of the wall surface then the possibility of the temperature overshoot of the platform is not unreasonable. For these studies the overshoot had no effect on the analytical results. Consequently the uncorrected temperature profiles have been shown. The question of the validity of the overshoot of the platform temperature will be left for another study.It is generally accepted that the optical pyrometer has an absolute accuracy of k5O"C. Variation of the automatic pyrometer can exceed this level unless the emissivity setting is calibrated using a vanishing filament pyrometer. In addition a slight variation of the recorded temperature was observed depending on the temperature range used (upper half of the 1100-2000°C range or the lower third of the 1500-3000°C Table 2 Atomization temperature^'^ TemperaturerC Element 1300 -4% 1400 Cd 1500 Ag K Na Pb 1600 TI 1700 Bi Cs 1800 Au Rb Te Zn 1900 2000 As 2100 Fe In 2200 Li Pd Pt Sn 2300 2400 2500 2600 Lu Cu Mg Mn Sb Se Al Ba Be Cr Ga Ge Ni Yb Co Ir Mo Rh Ru Sr V B Ca Dy Ho La P Pr Ti Table 3 Parameters for five selected elements14 range).In all the best estimate of the accuracy of the traces presented in this study is L 100 "C. Mass diflusion The dependence of the rate of mass diffusion on the gas phase temperature is given by D = Do ( T/To )" for a diffusion controlled system where D is the diffusion coefficient at temperature T Do is the diffusion coefficient at the reference temperature To and n= 1.68 1.88 1.91 1.84 and 1.93 for Cd Cr Cu Pb and V re~pective1y.l~ Since the residence time is inversely proportional to the rate of diffusion eqn. (1) can be expressed as where td is the residence time at temperature T and tdo is the residence time at the reference time T,. It is expected that the integrated absorbance is proportional to t,. Reagents The elemental standards employed for this study were 1 ngml-' of Cd 10ngml-I of Cr 25 ngml-' of Cu 25 ng ml-' of Pb and 50 ng ml-' of V in 0.1% HNO,.These solutions were prepared fresh routinely from stock solutions of 2 mg ml-' 2 mg ml-' 5 mg ml-' 25 mg ml-' and 50mgml-' respectively in 0.5% HNO,. Each of the stock solutions were prepared by appropriate dilution of 1000 mg ml-' commercial standards (Merck GmbH Darmstadt Germany) and concentrated HNO,. Elemental standards in the matrix modifier were made by dilution of the stock solutions as described in 0.1% HNO 250 pg ml-' of Pd and 150 pg ml-I of Mg(NO,),. Thus the pipetting of each aliquot of standard solution into the furnace placed 5 pg of Pd and 3 pg of Mg(NO,) onto the platform. RESULTS AND DISCUSSION Sensitivity Table 2 shows that the optimum atomization temperature for elements commonly determined by ETAAS covers a range of 1300-2600°C for the THGA.I4 From this list five elements (Cd Cr Cu Pb and V) were chosen to provide a representative temperature distribution ( 1400-2400 "C) consisting of the most frequently determined of the elements.The selected elements and their recommended matrix modifiers pyrolysis tempera- tures and characteristic masses are shown in Table 3.14 For this study the furnace program shown in Table 1 was used for all determinations and a matrix modifier consisting of 5 pg of Pd and 3 pg of Mg(NO,) for each 20 pl aliquot was chosen. With no modifier a pyrolysis temperature of 400°C was used while with the Pd-Mg(NO,) modifier a pyrolysis temperature of 700°C was used.Higher pyrolysis temperatures resulted in losses of Cd. Figs. 1 and 2 show the relative sensitivity of the five elements studied as a function of the programmed atomization tempera- ture without and with the Pd-Mg(NO,) modifier. The relative sensitivity is expressed as the ratio of the characteristic mass specified by the Perkin-Elmer Corporation for the THGA Element Atomization temperature/"C Recommended modifier Pyrolysis temperature/"C mdpg Cd 1400 50 I% PO 3 Clg M!mOJ) 700 1.3 Pb 1500 50 pg pg Mg(N03)2 8 50 30.0 c u 1900 !% Pd @g Mg(N03)2 1200 17.0 Cr 2300 15 Pi3 Mg"O,) 1500 7.0 V 2400 none 1200 42.0 Journal of Analytical Atomic Spectrometry March 1995 VoZ. 10 1991.5 1.3 1.1 €- f" 0.9 0.7 0.5 500 1500 2500 Programmed atomization temperaturePC Fig. 1 Relative sensitivity for A Cd Pb 0 Cu + Cr and + V as a function of the programmed atomization temperature for the standard THGA with no matrix modifier.Dashed line is the predicted loss in sensitivity for the diffusion limited case 1.5 1.3 1.1 i- €? 0.9 0.7 0.5 \ \ \ \ \ 500 1500 2500 Programmed atomization temperaturePC Fig.2 Relative sensitivity for A Cd 0 Pb 0 Cu 0 Cr and + V as a function of the programmed atomization temperature for the standard THGA with 5 pg of Pd and 3 pg of Mg(NO,) as a matrix modifier. Dashed line is the predicted loss in sensitivity for the diffusion limited case (Table 3) and the experimentally determined characteristic mass at each temperature. Thus ratios greater and less than unity reflect sensitivities greater and less than the manufac- turers specified characteristic masses.Since the specified characteristic masses are based on results for a large number of THGAs sensitivities for individual THGAs can be expected to show a random distribution around this value. The dashed line represents the decrease in sensitivity to be expected if the loss of atoms from the furnace is determined by mass diffusion at the programmed temperature. The value of n= 1.68 used for this plot is appropriate for Cd but is low for the other four elements (see Experimental section). For Pb and Cu the slope of the dashed line in Figs. 1 and 2 should be even steeper. Comparison of the experimental data to the predicted mass diffusion controlled curve must be made by shifting the data vertically i.e. comparison must be made at the same temperature since the slope is temperature dependent.In Figs. 1 and 2 the plots of sensitivity as a function of the programmed atomization temperature with the exception of V all exhibit the same pattern; an initial increase is followed by a decrease and then a plateau. The initial increase is a result of increasing atomization efficiency. At the lowest tem- peratures integration intervals as long as 15 to 20 s were not sufficient to cover the whole peak. As the programmed atomiz- ation temperature increased the analyte was more efficiently volatilized from the carbon surface and losses from the furnace in the molecular form were reduced. For Cd atomization temperatures less than 900°C were not possible in the maxi- mum power mode because the temperature sensing diode could not be calibrated.For V the temperature of maximum atomization efficiency is most likely not reached. The following decrease in sensitivity is to a first approxi- mation a result of increasing mass diffusion with increasing temperature. The decrease in sensitivity should be predictable from the temperature. In Fig. 1 it can be seen that only Cu (from 1900 to 2400°C) shows reasonable agreement with the predicted diffusion loss curve. Failure of the plots for Cd and Pb to exhibit the predicted slope serves to emphasize the over simplification of this model. It is possible that mass diffusion is the dominant loss mechanism but the temperatures determining the diffusion rate are lower than the programmed final stabilized temperature of the furnace.Most likely the loss in sensitivity can best be predicted by a convolution of the time dependent diffusion and analyte signal functions. The loss in signal is made more complex by the fact that loss due to the convective force of gas expansion with heating while generally negligible for an optimal atomization temperature becomes a significant component when atomization tempera- tures are employed which exceed the optimal temperature by 500 to 1000°C. In addition changes in the absorption coefficient and thermal diffusion as a function of temperature may also be contributing factors. A quantitative evaluation of the integrated signal as a function of the programmed atomiz- ation temperature is beyond the scope of this paper. Use of the Pd-Mg(N03)2 modifier (Fig.2) shifts the tem- perature of maximum sensitivity to a higher temperature for Cd Pb and Cr and reduces the maximum sensitivity values. It can be seen that Pb and Cu both show reasonable agreement with the plot based on loss by mass diffusion. The slope for Cd appears unchanged compared with atomization without a modifier (Fig. 1). The general loss in sensitivity between Figs. 1 and 2 is predictable from incomplete volatilization of the analyte (analyte is partially occluded by the modifier) at lower temperatures and higher diffusion coefficients resulting from volatilization at higher temperatures in the presence of the modifier. Frech and L'vov~,~ have also described a reduction in sensitivity due to matrix trapping of the analyte with the Pd modifier at the cooler ends of the furnace.It is not possible to distinguish between the two loss mechanisms without specific knowledge of the temperatures experienced by the analyte vapour. Ultimately in Figs. 1 and 2 the volatile elements show a lack of response at the highest temperatures. This plateau in sensitivity is a result of the analyte experiencing identical thermal conditions despite the difference in the final tempera- ture. This can happen only if the analyte is lost from the furnace before the maximum temperature is reached. Cadmium Pb and even Cu show constant sensitivities at higher tempera- tures. Thus the atomization for all three elements is completed before the final furnace temperature is reached. Figs. 1 and 2 show clearly that inclusion of V among the five elements to be determined dictates that 2500°C is the lowest atomization temperature that can be used.However the loss in sensitivity for Cd Pb Cu and Cr atomized at 2500"C is not as great as predicted for loss of the analyte from the furnace by diffusion. Without a modifier (Fig. 1) the losses are only 24% 17% 28% and 9% respectively. With the Pd-Mg(N03)2 modifier the losses were 34% 23% 34% and 4% respectively. In both cases the shapes of the plots (Figs. 1 and 2) were very reproducible. Relative precisions for 200 Journal of Analytical Atomic Spectrometry March 1995 Vol. 10Programmed atomization temperature/"C No Modifier 1300 1500 1700 1900 2100 2300 2500 With Modifier 1300 1500 1700 1900 2100 2300 2500 2500 Y 2 2100 E -. *-' 1700 e 1300 tl12 1.02 1.04 0.98 1.06 1.02 1.06 1.06 2.12 1.60 1.50 1.52 1.38 1.34 1.36 - - - - Cd t At112 1.14 0.32 1.10 0.24 1.06 0.24 1.14 0.26 1.10 0.24 1.14 0.24 1.10 0.22 2.56 0.94 1.88 0.52 1.68 0.36 1.64 0.28 1.46 0.22 1.44 0.22 1.46 0.24 t112 2.04 1.32 1.22 1.32 1.18 1.18 1.22 - - 2.50 2.00 1.70 1.74 1.70 Pb t p At112 2.32 0.56 1.58 0.54 1.44 0.42 1.46 0.32 1.34 0.28 1.38 0.32 1.36 0.26 - - - - 3.32 1.68 2.44 0.88 1.98 0.56 1.92 0.38 1.86 0.34 Table 4 Times for half height (tliz) peak maximum (tp) and half widths (Atll2); standard THGA - ten atomizations at each temperature were approximately 1 % for all of the elements.The loss in sensitivity shown in Figs. 1 and 2 would appear to be a very tolerable sacrifice in exchange for mu1 tielemen t capability. Peak Temperatures To discriminate between atom loss due to faster diffusion at higher temperatures and matrix trapping an attempt was made to characterize the temperature experienced by the analyte.This was done by comparing the time of the peak maximum (Table 4) with the temperature-time profiles of the furnace wall and platform (Fig. 3) obtained using an automatic pyrometer. To eliminate temperature variations between THGAs the peak time and temperature measurements were made using the same THGA used for the sensitivity measure- ments in the previous section. The furnace temperature profiles in Fig. 3 are similar to those previously reported by Hagdu and Frech." In Table 4 the time of the + height of the rising edge (t+) the time of the peak maximum ( t J and the full width at half height (At+) serve to give a rough approximation of the peak shape.Data are not listed when the first two parameters could not be observed within a 5 s atomization cycle. In every case the peaks had a pronounced asymmetry at lower temperatures (the front slope is steeper than the back slope) and became increasingly symmetric (equal slopes front and back) with increasing atomization temperature. As the programmed temperature was increased the peaks shifted to earlier times and At$ of the peaks decreased. For all the elements the peak times approached a minimum value as the temperature increased. n 2900 I hG= P 2500 ' W 2500 / 1 - - w 2100 - - - 2100' W 1700 P 1700' W 1300 1 6 P 1300 1 I I I I I I 1 1 2 3 4 5 Time/s Fig.3 Wall (W) and platform (P) temperatures measured by auto- matic optical pyrometer as a function of time for a standard THGA for programmed atomization temperatures of 1300 1700 2100 and 2500 "C 900 ' 0 c u Cr - - - - - - - - 1.96 2.42 1.32 - 1.76 2.02 0.74 2.80 3.57 - 1.58 1.74 0.42 1.84 2.09 0.39 1.58 1.74 0.32 1.85 2.04 0.24 1.56 1.70 0.30 1.75 1.87 0.19 - - - - - - 2.66 3.82 - - - - 2.02 2.52 1.24 3.22 4.46 - 1.68 1.90 1.50 1.96 2.28 0.46 1.60 1.76 0.34 1.88 2.06 0.26 1.64 1.78 0.34 1.84 1.96 0.22 V - - - 1.92 2.20 0.86 1.94 2.12 0.64 - - - 1.98 2.14 0.66 1.92 2.14 0.62 The temperatures at the time of the peak maximum deter- mined from the times in Table4 and the temperature-time plots in Fig.3 are shown in Table 5. It was assumed that the temperature experienced by the analyte was equal to the average of the wall and platform temperature.It can be seen that without the modifier the average peak temperatures for Cd and Pb quickly reach a maximum (1685 and 1999"C respectively) as the programmed atomization temperature increases while those for Cu Cry and V systematically increase. The regions of constant temperature for Cd and Pb are consistent with the previously discussed regions of constant absorbance (Fig. 1). The effect of the Pd-Mg(NO,) modifier on the peak temperatures appears to be directly related to the volatility of the element. For V the modifier appears to have no effect on either the time of the peak maximum (Table 4) or the tempera- ture (Table 5). For Cu and Cr the peak times are greatest at the lower temperatures and approach the values obtained without the modifier as the programmed temperature increases.At the lower temperatures the delay in the time of the peak maximum has little effect on the peak temperature since the stabilized temperature has been reached. At higher tempera- tures the rate of change of the temperature is greater but the times of the peak maximum are in much closer agreement. As Table 5 Temperature ("C) experienced by analyte at peak maximum standard THGA Programmed atomization temperature/"C No Modifier 1300 1500 1700 1900 2100 2300 2500 With Modifier 1300 1500 1700 1900 2100 2300 2500 Temperature at time of peak maximum* Cd Pb c u Cr V 1333 1214 1555 1503 - - - 1680 1684 1616 - - 1685 1850 1798 1864 - 1677 1999 2069 2087 - 1723 1985 2268 2352 2382 1727 1988 2305 2478 2638 - - - - - - - 1212 1462 1647 - 1659 - - 1822 1825 1834 1861 - 2053 2076 2080 2105 - 2072 2314 2280 2376 2382 2075 2452 2370 2557 2656 - - - - * Temperatures determined from Fig.3 using the peak maximum times from Table 4. Journal of Analytical Atomic Spectrometry March 1995 VoZ. 10 201a result the modifier has little effect on the peak temperatures experienced by Cu and Cr. For Cd and Pb the peak maxima are delayed by a minimum of 0.3 and 0.5 s respectively and by as much as 1 or 2 s at lower temperatures. At the lower temperatures the time delay has no effect on the peak tempera- tures since the stabilized temperature has been reached. At higher temperatures the temperature maximum observed with- out the modifier no longer exists. With the modifier the peak temperatures for Cd and Pb increase monotonically with the programmed temperature.Thus the modifier is successful in delaying the atomization of Cd and Pb until the furnace is close to the stabilized temperature. Matrix Trapping The data in Table4 and Figs. 1 and 2 show that with the addition of the modifier the analyte peaks for Cd and Pb are shifted to higher temperatures and all elements have reduced areas. For Cd and Pb the decrease in area can arise from two different sources. At lower temperatures the atomization efficiency will be reduced. As long as the bulk of the modifier remains on the platform there is the possibility of occluded analyte remaining in the modifier. At higher programmed atomization temperatures the higher peak temperatures in the presence of the modifier result in higher diffusion rates.The peak temperatures for Cu Cr and V however are roughly the same and yet the sensitivity has decreased significantly. One explanation for this decrease in sensitivity is matrix Table 6 presents the predicted and experimentally measured recoveries (absorbance with modifier/absorbance without modifier x 100%) for all five elements in the presence of the Pd-Mg(N03)2 matrix modifier with a programmed atomization temperature of 2500 "C. The predicted recoveries were calculated assuming diffusional loss at the listed peak temperatures. The predicted recoveries are then used to com- pute temperature adjusted recoveries from the experimental recoveries. Deviations from 100% recovery after adjustment for the peak temperatures is assumed to be the result of matrix trapping.Thus matrix trapping results in a 9-32% decrease in the analytical sensitivities. Table 7 shows the experimental and temperature adjusted recovery for each element in the presence of the Pd-Mg(N03)z matrix modifier as a function of the programmed atomization temperature. In general after adjusting for temperature Cd Pb and Cr show a trend of increasing recovery as the programmed temperature increases. Only Cu shows an opposite trend. It should be noted that the temperature adjusted recoveries are based on the experimental peak times in Table 4 the temperature-time plots in Fig. 3 and the absorbance data in Figs. 1 and 2. While it is not possible to rigorously evaluate the precision of the temperature adjusted recoveries a value of 10 to 20% seems reasonable.Thus only general trends should be considered. Frech and L'vov' found recoveries of 80-87% for the atomization of Co at 2400°C in the presence of pg-masses of Au Cu and Pd. They also observed that recoveries decreased from 100 to 80% for the atomization of Co in the presence of 5 pg of Pd and 15 pg of Mg(NO,) as the programmed atomization temperature increased from 1800 to 2400 "C. Similar observations were made for Au in the presence of 20pg of Ag; recoveries decreased from 85 to 76% as the programmed atomization temperature increased from 1500 to 2200 "C. These recoveries were not adjusted for the peak temperatures. It can be seen in Table 7 that in this work the experimental recoveries for Cd Pb and Cu decrease and those for Cr show a slight increase as the programmed atomization temperature increases.When the recoveries are adjusted for temperature there is an improvement with increasing atomiz- ation temperature for Cd Pb and Cr. Of course the pattern is complicated by the lack of efficient atomization of the Cd and Pb at lower temperatures. Signal-to-noise Ratio The data in Figs. 1 and 2 support the point made in the introduction that with extremely volatile or non-volatile Table 6 Recovery of analyte in presence of Pd-Mg(N03)* modifier with a programmed atomization temperature of 2500°C Standard THGA Peak temperature* Temperature Element No modifier With modifier recoveryj- (YO) predicted Cd Pb c u Cr V 1727 1988 2305 2478 2638 2075 2452 2370 2557 2656 73 71 95 93 99 Experimental recovery1 (Yo) 64 55 76 81 84 Temperature adjusted experimental recovery9 (YO) 88 77 80 87 85 * From Table 5.Assuming only diffusional loss and the peak temperatures in columns 1 and 2. 1 Absorbance with modifier (Fig. 2)/absorbance without modifier (Fig. 1) x 100%. tj Experimental recovery/predicted recovery x 100%. Table 7 Percentage recovery of analyte in the presence of Pd-Mg(N03):! modifier as a function of temperature standard THGA Programmed atomization temperat ure/"C 1300 1500 1700 1900 2100 2300 2500 Cd Pb Exp* Adjt 80 74 78 71 78 76 74 83 67 89 64 86 64 88 Exp* Adjj- - - 70 68 61 66 58 74 55 78 Exp* - - 83 77 75 77 76 c u Cr Adj t Exp* Adjt - 87 80 76 77 76 - - 76 77 79 80 81 86 * Experimental recoveries from Figs. 1 and 2.t Adjusted recoveries assuming diffusion and based on temperatures in Table 5. 202 Journal of Analytical Atomic Spectrometry March 1995 Vol. 10elements in the suite to be determined there is little room for compromise. The data for V in Figs. 1 and 2 dictate that an atomization temperature of 2500 or 2600°C be used. At these temperatures the sensitivity loss for Cd Pb and Cu ranged from 17 to 28% without the use of the Pd-Mg(NO& modifier and from 23 to 34% with the modifier. The greater loss with the modifier comes from both matrix trapping and increased mass diffusion (see Table 6). Based on the sensitivity data the simultaneous determination of the suite of five elements con- sidered in this study requires a 2500°C atomization tempera- ture and is accompanied by a 17-34% loss in sensitivity for the volatile elements. The SNR however not sensitivity is the most informative figure of merit.Consequently the limiting noise source must also be considered. The limiting noise for furnace atomization signals is dependent on the analyte concentration. At higher concentrations analyte fluctuation noise arising from the deposition volatilization and atomization of the analyte is limiting. In this study as reported earlier the relative precisions of signals with absorbances greater than 0.1 were constant at approximately 1.0%. At these signal levels there is no compro- mise. If all the analytes are present in large quantities i.e. all signals are greater than 0.1 then the SNRs for all elements will always be approximately 100.At lower analyte concentrations near the detection limit photon shot noise is limiting and the precision for integrated signals is dependent on the square root of the integration time. Thus the integrated absorbance noise will be minimized if the integration interval is kept as short as possible. This is achieved if the integration interval is matched to the duration of the analytical ~ i g n a l ' ~ ? ' ~ and high atomization temperatures are used (Table 4). The analytical signals however also decrease with increasing temperature (Figs. 1 and 2) so the exact relationship of the SNRs to atomization temperature is not obvious. If it is assumed that the analytical peaks are roughly triangular in shape then At3 in Table 4 is proportional to the time required to integrate the whole peak and the square root of At3 will be proportional to the integrated absorbance noise.Fig. 4 presents the relative SNRs computed by dividing the sensitivities of Figs. 1 and 2 by the square roots of the values for At+ in Table 4. Based on photon shot noise an atomization temperature of 2500°C is optimum with and without the modifier. In both cases however the atomization temperature is not dictated by the desire to include V but because 2500°C offers the best SNRs for all five elements and no compromise + 0- + I 1200 1900 2600 1200 1900 2600 Temper at u re/" C Fig. 4 S/N for A Cd 0 Pb 0 Cu 0 Cr and + V as a function of the programmed atomization temperature for the standard THGA (a) without and (b) with matrix modifier is required.It can also be seen that the SNRs are better for all elements without a modifier. Of course only standards were used in this study. In the presence of sample matrices the utility of the modifier for higher temperature pre-treatment may be essential. End-Capped THGAs-Sensitivity A series of studies identical to those just described for standard THGAs were performed using end-capped THGAs identical to those described by Frech and L'vov.' Figs. 5 and 6 show the relative sensitivity of the end-capped THGAs as a function of the programmed atomization temperature without and with modifier respectively. The plots in Figs. 5 and 6 have the same general shapes as those in Figs. 1 and 2 but the relative sensitivities for the end-capped THGA are enhanced by factors of 1.4 to 1.7 with no modifier and by factors of 1.7 to 3.1 with the modifier.The enhanced sensitivities are undoubtedly due to the reduced apertures of the end-capped THGA which increase the analyte residence times and consequently the \ 1000 2000 3000 Programmed atomization temperature/"C Fig. 5 Relative sensitivity for A Cd Pb 0 Cu + Cr and + V as a function of the programmed atomization temperature for the end- capped THGA with no matrix modifier. Dashed line is the predicted loss in sensitivity for the diffusion limited case 3 2 E F" ( 1 1 2000 3000 Programmed atomization temperaturePC Fig.6 Relative sensitivity for A Cd 0 Pb 0 Cu 0 Cr and + V as a function of the programmed atomization temperature for the end- capped THGA with 5 pg of Pd and 3 pg of Mg(N0,)2 as a matrix modifier.Dashed line is the predicted loss in sensitivity for the diffusion limited case Journal of Analytical Atomic Spectrometry March 1995 Vol. 10 203peak areas. Frech and L ' V O V ~ ~ ~ have also shown that this enhancement can arise from a reduction in matrix trapping. It can also be observed that the end-capped THGAs (Figs. 5 and 6) shift the temperatures at which the peak areas are maximum and provide temperature dependent sensitivities which are close to that predicted for the diffusion limited case (dashed line n = 1.68). With respect to simultaneous multielement determinations V again dictates that the minimum suitable atomization tem- perature is 2500°C. With no modifier atomization at 2500°C means a loss of 3 to 20% in sensitivity for the other four elements (Fig.5) as compared to their maximum sensitivity at lower temperatures. With the Pd-Mg(N03)2 modifier atom- ization at 2500°C is accompanied by an 8 to 30% loss in sensitivity for the other four elements (Fig. 6). The losses in maximum sensitivity were slightly less with end-capped THGAs as compared to the standard THGAs but the differ- ence between the temperature of maximum area and 2500°C was also less. The decrease in sensitivity resulting from the Pd-Mg(NO,) modifier was significantly less for the end-capped THGA than the standard THGA. With a programmed atomization tem- perature of 2500 "C sensitivities for the end-capped THGA were reduced by 0 to 25% using the modifier. For the standard THGA a reduction of 16 to 41% was observed with use of the modifier.This suggests that matrix trapping has been reduced but it is necessary to adjust the absorbances for changes in the peak temperatures before making a final judgment . The peak widths in Tables 4 and 8 also suggest that there is a difference in the interaction between the analyte and the matrix modifier in the standard and end-capped furnaces. In a standard furnace the matrix modifier results in a larger half- width of the peaks. This effect is more pronounced at the lower temperatures. In an end-capped furnace the peak width for Pb still increase with the use of the modifier but Cd Cu and Cr show significant decreases in width. Tables 5 and 9 show that there is not a significant difference in the temperature at the time of the peak maximum.The peaks in the end-capped tubes occur at a later time but the reduced overshoot of the wall temperature results in lower average (of the wall and platform) temperatures. Thus it is possible that the broader peaks of the standard furnace with the modifier (as compared to the end-capped tube with the modifier) are due to co-condensation of the analyte with the modifier near the ends of the furnace. 2500 W 2500,P 2500 - Y 2100 ?! .F & 1700 Q E ' 1300 f ! ! P 1300 . . . /A I I 1 I I 0 1 2 3 4 5 900 ' Ti mels Fig. 7 Wall (W) and platform (P) temperatures measured by auto- matic optical pyrometer as a function of time for an end-capped THGA for programmed atomization temperatures of 1300 1700,2100 and 2500°C End-Capped THGAs-Peak Temperatures Fig. 7 shows the temperature-time profiles for the wall and platform of an end-capped THGA.Compared to the standard THGA in Fig. 3 it can be seen that the overshoot of the wall temperature is practically non-existent and the differences between the stabilized temperatures of the wall and platform are very small (within experimental error) at higher atomization temperatures (2100-2500 "C) and less than 100 "C at lower temperatures. These results are in agreement with those pre- viously reported by Hagdu and Frech.15 In addition the heating rates of the end-capped THGAs are significantly lower compared with the standard THGA. In Fig. 3 the platform of the standard THGA reaches 2200 "C in slightly less than 2 s while it takes the end-capped THGA about 2.7 s. This suggests that a significant fraction of the current through the tube is conducted through the end caps.The more complete enclosure formed by the end-capped furnace also brings the stabilized temperatures of the wall and platform into better agreement. Table 8 presents the times measured for the analyte peaks using the end-capped THGAs. As was observed with the standard THGA the symmetry of the analytical peaks improves and all the peak times approach a minimum value as the programmed atomization temperature increases. The use of the Pd-Mg(NO,) modifier delays the peak appearance (larger values for t3 and tP) for all elements except V. The length of the delay appears to be roughly proportional to the volatility of the element. Compared with the standard THGA the peaks for the end-capped THGA with and without modi- Table 8 Times for half height (front slope tljz) peak maximum (tp) and half widths (Atll2); end-capped THGA Programmed Cd Pb cu Cr V atomization temperature/"C No Modifier 1300 1500 1700 1900 2100 2300 2500 With Modifier 1300 1500 1700 1900 2100 2300 2500 t112 1.32 1.34 1.26 1.26 1.26 1.20 1.26 2.58 2.08 1.88 1.84 1.74 1.72 1.76 t P 1.50 1.46 1.40 1.38 1.36 1.34 1.40 3.14 2.38 2.14 1.98 1.84 1.82 1.90 4 2 t l l 2 0.82 2.12 0.54 1.80 0.46 1.62 0.44 1.68 0.44 1.58 0.44 1.54 0.44 1.56 1.08 - 0.62 3.28 0.50 3.10 0.38 2.58 0.28 2.28 0.26 2.22 0.28 2.26 tP 2.68 2.14 1.90 1.90 1.74 1.72 1.72 - 4.18 3.64 3.22 2.68 2.54 2.56 At112 h i 2 0.90 - 0.62 - 0.52 2.57 0.46 2.22 0.34 2.04 0.42 2.10 0.44 2.08 - - - 3.28 1.52 2.72 1.18 2.40 0.70 2.28 0.60 2.30 tP - - 3.06 2.68 2.30 2.32 2.32 -.- 4.52 3.28 2.72 2.48 2.48 At, t1/2 - - 1.76 - 1.16 2.66 0.66 2.36 0.54 2.42 0.48 2.32 - - 1.46 - 0.58 2.94 0.36 2.74 0.34 2.56 t P - - - 3.32 2.70 2.66 2.54 - - - - 3.08 2.84 2.62 - - 2.12 - 1.02 - 0.64 2.64 0.56 2.66 - - - - - - - - 0.58 - 0.30 2.80 0.20 2.76 204 Journal of Analytical Atomic Spectrometry March 1995 Vol. 10fier are all delayed. This reflects the slower heating rates observed for the end-capped THGAs in Fig. 7. The behaviour of the width of the peaks (At$) appears to be element dependent. With a standard THGA the use of the modifier produced almost no change in At+ at the highest atomization temperatures (2300 and 2500 "C). With the end- capped THGA the use of the modifier produces no change in width for V but the widths for Cd Cu and Cr are significantly narrower and the width of Pb is significantly broader with the modifier at the highest temperatures. A direct comparison of At+ for the standard and end-capped THGA without the modifier shows that the end-capped THGA produces wider peaks for all elements Using the modifier the end-capped THGA provides comparable widths for Cd Cu and Cr and broader peaks for Pb and V.Table 9 gives the average of the wall and platform tempera- tures at the time of the peak maximum based on the peak times in Table8 and temperature-time traces in Fig. 7. The average peak temperatures for the end-capped THGA in Table 8 are consistently lower than those for the standard THGA in Table 5. This difference arises from the higher wall temperatures of the standard THGA as compared with the end-capped THGA.Not surprisingly the platform tempera- tures at the time of the peak maximum (data not shown) are approximately the same for both types of THGA. The volatiliz- ation and atomization processes should occur at the same temperature for the same elemental standards and similar carbon surfaces. However the wall temperatures of the stan- dard THGA at the times of the peak maximum range from Table 9 Temperature ("C) experienced by analyte at peak maximum end-capped THGA Programmed atomization temperaturePC No Modifier 1300 1500 1700 1900 2100 2300 2500 With Modifier 1300 1500 1700 1900 2100 2300 2500 Temperature at time of peak maximum* Cd Pb Cu Cr v 1216 1210 I - - 1411 1396 ~- - - 1508 1524 1593 - - 1525 1658 1710 1772 - 1570 1785 1942 2012 - 1581 1781 2083 2177 2274 1590 1784 2139 2274 2476 1216 - - - - 1401 1480 - - - 1522 1620 1658 - - 1658 1756 1775 - - 1837 2012 2016 2054 - 1854 2139 2130 2224 2262 1876 2285 2234 2330 2486 * Temperature determined from Fig.7 using the peak maximum times from Table 8. 300 to 400°C hotter than those of the end-capped THGA. Thus at all but the lowest programmed atomization tempera- tures the peak temperatures of the end-capped THGA are 100-200 "C lower than those for the standard THGA. End-Capped THGAs-Matrix Trapping Recovery of all five elements in the Pd-Mg(NO,) matrix modifier was computed for a programmed atomization tem- perature of 2500°C. The recoveries were adjusted for the experimentally measured peak temperatures as previously dis- cussed for matrix trapping with the standard THGA.Table 10 shows that the recoveries ranged from 93 to 108%. As men- tioned previously the experimental uncertainty for the recover- ies is in excess of & 10%. Thus within experimental error the end-capped THGA eliminates matrix trapping for the Pd-Mg( NO,) modifier. End-Capped THGAs-Signal-to-noise Ratio As discussed previously for the standard THGAs the SNRs for the end-capped THGAs are dependent on the limiting noise source. At higher concentrations where analyte fluctu- ation noise is limiting the SNRs of all elements will be 100 (with a 1.0% analyte fluctuation) as long as the analyte signals remain above 0.1. Near the detection limit photon shot noise is limiting and relative SNRs in Fig.8 are determined by dividing the sensitivities by the square root of At+. It can be seen that the conclusions reached for the standard THGAs (b) + 1200 1900 2600 1200 1900 2600 Tern pe ra t u re/"C Fig. 8 SjN for A Cd 0 Pb 0 Cu 0 Cr and + V as a function of the programmed atomization temperature for the end-capped THGA (a) without and (b) with matrix modifier Table 10 Recovery of analyte in the presence of Pd-Mg(NO,) modifier with a programmed atomization temperature of 2500 "C end- capped THGA Peak temperature* Element No modifier With modifier Cd Pb c u Cr v 1590 1784 2139 2274 2476 1876 228 5 2234 2330 2486 Temperature predicted recovery? (YO) 79 67 93 96 99 Temperature adjusted Experimental experimental recovery$ (Yo) recovery4 (YO) 77 73 98 89 94 98 108 105 93 95 * Temperatures from Table 5.t Assuming only diffusional loss and the peak temperatures in columns 1 and 2. $ Absorbance with modifier (Fig. 2)/absorbance without modifier (Fig. 1) x 100%. 4 Experimental recoveryjpredicted recovery x 100%. Journal of Analytical Atomic Spectrometry March 199.5 Vol. 10 205also apply to the end-capped THGAs. When photon shot noise is limiting an atomization temperature of 2500 “C should be used not because this is the minimum temperature for efficient atomization of V but because this temperature with and without the modifier offers the best SNRs for all five elements without compromise. With the end-capped THGA the effect of the Pd-Mg(NO& modifier on the SNR varies with each element. With the modifier the SNRs for Cu and Cr improve Pb is worse and Cd and V are unchanged. The SNRs for the end-capped THGA are better than those for the standard THGA for all elements with or without the modifier. This is a result of the enhanced sensitivities of the end-capped THGA which are achieved without any loss of transmitted intensity. REFERENCES 1 2 3 4 Harnly J. M. and Kane J. S. Anal. Chem. 1984 56 48. Harnly J. M. Miller-Ihli N. J. and O’Haver T. C. Spectrochim. Acta Part B 1984 39 305. Lewis S. A. OHaver T. C. and Harnly J. M. Anal. Chem. 1984 56 1651. Lewis S. A. O’Haver T. C. and Harnly J. M. Anal. Chem. 1985 57 2. 5 6 7 8 9 10 11 12 13 14 15 16 17 Harnly J. M. and Garland D. G. in Methods in Enzymology eds. Riordan J. F. and Vallee B. L. Academic Press Inc San Diego CA 1988 vol. 158 p. 145. Welz B Sperling M. Schlemmer G. Wenzel N. and Marowsky G. Spectrochim. Acta Part B 1988. 43 1187. Berglund M. Frech W. and Baxter D. C. Spectrochim. Acta Part B 1991 46 1767. Frech W. L‘vov B. and Romanov N. P. Spectrochim. Acta Part B 1992 47 1471. Frech W. and L‘vov B. Spectrochim. Acta Part B 1993,48 1371. Frech W. Cedegren A. Lundberg E. and Siemer D. D. Spectrochim. Acta Part B 1983 38 1435. Siemer D. D. and Frech W. Spectrochim. Acta Part B 1984 39 261. L‘vov B. and Frech W. Spectrochim. Acta Part B 1993 48,425. L‘vov B. V. Spectrochim. Acta Part B 1990 45 633. Perkin-Elmer Corporation 1991 “Recommended Conditions for THGA Furnaces,” Part Number B050-6 158 Publication B3 1 10.06. Hagdu N. and Frech W. Spectrochim. Acta Part B 1994,49,445. Harnly J. M. J. Anal. At. Spectrom. 1988 3 43. Voigtman E. Appl. Spectrosc. 1991 45 237. Paper 4/05 182C Received August 24,1994 Accepted September 30 I994 206 Journal of Analytical Atomic Spectrometry March 1995 Vol. 10

ISSN:0267-9477    

DOI:10.1039/JA9951000197

出版商:RSC    

年代:1995

数据来源: RSC