ISSN:0144-557X    

DOI:10.1039/AP98825FX025

出版商:RSC    

年代:1988

数据来源: RSC

摘要:

ANPRDI 25(7) 209-260 (1988) Proceedings of the Analytical Division of The Royal Society of Chemistry CONTENTS 209 Editorial 21 0 The New President 21 1 Obituary 21 2 SAC Silver Medal 213 Retiring President's Address 213 'Analytical Chemistry: a Question of Decisions' by D. C. M. Squirrel1 21 6 Chair for Analytical Chemist 21 7 SUMMARIES OF PAPERS 248 250 21 7 Professor Ottaway Memorial Meeting 21 7 220 222 224 226 229 230 232 233 235 237 238 240 244 246 'Graphite Furnace Atomic Emission Spectrometry-the Rediscovery of a Technique' by D. Littlejohn 'Evaluation of a Mathematical Model for Peak Interpretation in Graphite Furnace Atomic Absorption Spectrometry Based on Free Analyte Atom Redeposition on Carbon Surfaces' by B.Welz, B. Rudziuk and G. Schlemmer 'Graphite Furnace Atomic Absorption Spectrometry on the Way to Absolute Analysis' by Boris V. L'vov 'Comparison of Bead-making Techniques for X-ray Fluorescence Spectrometry' by Richard Steventon and Andrew Cunningham 'Novel Instrumentation for Inductively Coupled Plasma Atomic Emission Spec- 'Determination of Cobalt in Plasma and Urine by Electrothermal Atomisation Atomic 'Laser-excited Atomic and Molecular Fluorescence-Preliminary Investigations for 'Faster Analysis of Biological Samples by Electrothermal Atomisation Atomic Absorp- 'Determination of Boron in Plants by Graphite Furnace Atomic Absorption Spec- trometry' by s. P. Corr, D. H. Hall, D. Littlejohn and C. V. Perkins Absorption Spectrometry Using Palladium Matrix Modification' by Barry Sampson Thallium' by L.M. Garden, D. Littlejohn, K. Dittrich and H.-J. Stark tion Spectrometry' by D. J. Halls trometry with Slurry Atomisation Using Matrix Modification and Totally Pyrolytic Graphite Tubes' by Neil W. Barnett, Les Ebdon, E. Hywel Evans and Pierrick Ollivier 'Furnace Atomisation for Multi-element Atomic Absorption Spectrometry' by James M. Harnly 'Application of Atomic Spectroscopy in Clinical Chemistry' by G. S. Fell 'The ICP-Is it The Real Thing?' by J. Marshall 'In-situ Pre-concentration in Flame Atomic Spectrometry' by T. S. West 'Probes and Furnaces from Molecular Emission Cavity Analysis' by N. Pourreza, Alan Townshend and Paul S. Turner 'Radioanalytical Studies of Electrothermal Atomisation Atomic Absorption Spec- trometry' by John E.Whitley, Richard Hannah and David Littlejohn 'Flame and Furnace; Emission and Absorption-a Historical Dialogue' by A. M. Ure 'Some Results of Joint Research Activities on the FANES Technique Between the Academy of Sciences, Berlin, and Strathclyde University, Glasgow' by Heinz Falk Tvnnset and minted bv Black Bear Press Limited, Cambridge, England Analytical Proceedings Continued inside back cover July 1988 ... ANALYTICAL PROCEEDINGS, JULY 1988. VOL 25 111 25 1 'The Teaching of Analytical Science in the 21st Century' by John F. Alder 'Early Optics and Spectroscopy-The Scottish Dimension' by D. Thorburn Burns 253 253 'Ramblings Through the Landscape of Analytical Atomic Spectroscopy' by Thomas C.O'Haver 256 Equipment News 259 Publications Received 259 Ronald Belcher Memorial Lectureship (Rules) 260 Analytical Division Distinguished Service Award (Rules) 260 Course 260 Conference 260 New Edition of Orange Book Continued from back cover ~ Analytical Applications of Spectroscopy Edited by C.S.Creaser, University of East Anglia and A.M.C. Davies, Institute ofFuod Research, Norwich This new book provides a 'State-of-the-Art' review of the applications of the major spectroscopic techniques and will prove invaluable to researchers involved in this form of analysis The book provides wide-ranging coverage of recent developments in analytical spectroscopy, and in particular the common themes of chromatography - spectroscopy combinations, Fourier transform methods and data handling techniques.Each section includes a review of key areas of current research, written by spectroscopists who have made major contributions in their respective disciplines, as well as short reports of new developments in these fields. These common themes have played an increasingly important part in recent advances in spectroscopic techniques and emphasise the multidisciplinary approach of present research. , 502 pages ISBN 0 85186 383 3 Price E47.50 ($99.00) lnformat ion Services To order or for further information, please write to: Royal Society of Chemistry, Distribution Centre, Blackhorse Road, Letchworth, Herts SG6 IHN, UK. or telephone (0462) 672555 quoting your credit card details. We now accept Access/Visa/MasterCard/EuoCard. RSC Members are entitled to a discount on most RSC publications and should write to: The Membership Manager, Royal society of Chemistry, 30 Russell Square, London WCl B SDT, UK.

ISSN:0144-557X    

DOI:10.1039/AP98825BX027

出版商:RSC    

年代:1988

数据来源: RSC

摘要:

ANALYTICAL PROCEEDINGS, JULY 1988, VOL 25 209 Editorial Analytical Chemistry Trust Introduction The Analytical Division of the Royal Society of Chemistry is unique amongst the RSC Divisions in that it has at its disposal a substantial Trust Fund, which is used to augment the income that it receives from the RSC to meet its running expenses. The Trust Fund was established about the time of the amalgamation of the Society for Analytical Chemistry with the Chemical Society and by maintaining a policy of capital growth and by ploughing back some of the income, the value of the Trust Fund has steadily appreciated over the years to buffer the effect of inflation. The Fund now provides a substantial income, the use of which is central to the present method of operation of the Divi- sion.The accounts for the Analytical Chemistry Trust are published annually in Analytical Proceedings and show how the income is used. As part of the continuing and regular review of Analytical Division activities, the use of the Trust Fund was discussed at a special meeting of the Trust Advisory Committee in Autumn 1987, to which the Treasurers of all Groups and Regions were invited and subsequently, by the Trustees. To aid these discussions a paper was produced, summarising the role played by the Trust Fund in Divi- sional activities, and Trustees felt that members of the Division might like to have a summary of that paper, and of the discussion that ensued, published in Analytical Proceedings, in view of the importance of the Trust Fund to Divi- sional activities. Current Situation For convenience, the expenditure of the Division can be divided into two cat- egories, running costs and additional expenditure.Each of these is considered in turn. Running Costs The RSC allocates an annual grant to the individual Divisions, which is determined primarily by the number of members in the Division. The Analytical Division, because of the additional sums available from the Trust, has more latitude and has established a comprehensive range of activities to benefit its members, which are heavily dependent upon support from the Trust Fund for them to be able to continue at the current level. During the financial year 1987, the Trust Fund will have contributed f24 000 to running costs of the Division, between two and three times the value of the RSC annual grant of fll000.Trust Fund income is conse- quently clearly central to the continuing operation of the Division in its present form. As would be expected, a major source of the expenditure results from Council and committee expenses. The Analytical Division has a large Council with a total membership of about forty, which meets four times a year and reflects the Divi- sion’s considerable national and technical coverage. This compares with some other Divisions which have smaller Councils that meet three times a year. Divisional activities are supported by a number of committees, which are directly account- able to Council and are responsible for maintaining the comprehensive activities of the Division. These committees include Group Liaison and Policy, Programmes, Analytical Methods, Honours, Education and Training, Trust Advisory and Honor- ary Officers, as well as the Secretaries’ Conference and any working parties set up by Council.Some committees need to meet several times a year (for example the Group Liaison and Policy Committee), others, such as the Secretaries’ Confer- ence, are convened annually and some, for example, the special meeting of Treasurers, meet only when this is deemed to be needed by Council. The Division also takes on the responsibility for the circulation of Divisional, Regional and Group information to all members five times a year. This should be con- trasted with methods other Divisions have at their disposal to communicate with their members. The combined costs of Committee expenses, of postage and of distribution account for about half the operating costs of the Division, the major component of the other half being the cost of organising Divisional scientific meet- ings, which, of course, is largely covered by income from those meetings.Additional Expenditure The major regular contribution made by the Trust to the activities of the Division, outside the subsidy to the running costs, is the support for research, normally through studentships. The current rules allow for a commitment to support up to 12 man years of work at any one time and the Trust normally budgets to spend between f20 000 and f30 000 per annum. The Trust also provides funds for awards such as the Boyle, Gold and Silver Medals, Distinguished Service and Belcher awards and for the Redwood Lecturer.It also pays the expenses asso- ciated with the implementation of these awards. The cost of these studentships and awards during 1987, excluding Gold and Silver medals, amounted to another f30 000. In addition to the regular items, the Trust Fund supports new initiatives to promote and support analytical science when these have been endorsed by Trustees. Thus, Schools Lectures were recently introduced, while a current major expenditure is the publicity video, which is intended to strengthen the image of analytical chemistry in schools and other organisations that have an interest in learning more about modern analytical chemistry. A short time ago, the Trust made a large interest-free loan to the RSC to assist with computerisation of Ana- lytical Abstracts. Additional, smaller grants are made from time to time when agreed by the Trustees.Future Options In the past the policy of the Trust Fund can be considered to have had four main aims: ( a ) , to strengthen the activity of the Division by contributing to running costs; (b), to support analytical research through studentships or fellowships; ( c ) , to re-invest a proportion of income to ensure a well-supported Fund in future years; and ( d ) , to have a surplus available which can be used to support new ana- lytical initiatives if Trustees agree. This policy is in the spirit of the aims laid down in the Trust document itself. The contribution from the Trust Fund to the running costs of the Division calls for one of the largest contributions from the Trust Fund.The estimated subsidy during the current year is again likely to be between two and three times the grant received from the RSC and recognises the comprehensive nature of the Division’s activities. It clearly underpins the activity at both Regional and Group level, enables national and technical coverage to be comprehensive and leads to a programme of scientific meetings whose number probably matches that of all other Divisions of the RSC combined. Circula- tion of information also provides a com- prehensive service to members so that, by subsidising running costs, all members of the Division gain direct benefit from the Trust. The Trust Advisory Committee and Trustees therefore considered it to be reasonable to continue to subside the operating costs of the Division from the Trust Fund in order to maintain the current breadth of coverage but, at the same time, emphasised the need to keep a careful watch on costs so that they remain firmly under control.Running costs of the210 ANALYTICAL PROCEEDINGS, JULY 1988, VOL 25 Division must be expected to rise some- what in the future for various reasons, for example, a new committee of Council has been set up (the Education and Training Committee), whose expenditure will not be balanced by income, and the allocation to Groups and Regions has been raised from f300 to f400 per annurn. The cost of management of the Trust Fund itself will also show a sharp increase from 1987 onwards as a result of the requirement after the “big bang” that Fund Managers and Market Makers needed to be sep- arated. As a consequence, in the future the Trust Fund will be charged an appreci- able fee for Fund Management.Support given to the students is clearly an important use of the Division’s Trust Fund as it advances analytical knowledge and fulfils the scientific and technical aims of the Trust, as well as raising the profile of the Division and of analytical science, in scientific and technical circles; this use of the Trust income was endorsed by Trustees. The financing of the various awards supported by the Division was also considered to be a good use of the Trust Fund because it gives the Division the opportunity to recognise distinguished achievements in analytical science. There was also a firm view that the Trustees should ensure a financially strong Trust Fund in the future, so that the members of the Division will be in a position to continue to benefit from the extra income that the Fund offers and to sponsor analytical activitites. However, Trustees agreed that there remains scope for sup- port of some novel initiatives by the Fund although these will need to reflect the breadth of the membership and of the interests of the Division. A number of the new initiatives have been proposed by the Trust Advisory Committee and are cur- rently being discussed by Trustees to determine how they can be implemented. T. B. PIERCE

ISSN:0144-557X    

DOI:10.1039/AP9882500209

出版商:RSC    

年代:1988

数据来源: RSC

摘要:

210 ANALYTICAL PROCEEDINGS, JULY 1988, VOL 25 The New President Duncan Thorburn Burns was born in Wolverhampton in 1934. After an early itinerant education in Rugby, Manches- ter, Wolverhampton and Newton Stewart he arrived in Gomersal, where he almost became a Yorkshireman but, spending his summers in Scotland, failed to develop the requisite skills or devotion to cricket. His initial chemical education owed much to the heuristic approach of his chemistry master at Whitcliffe Mount Grammar School, Cleckheaton, William J. Quin, and to the six years of summer vacation work, mainly analytical, at British Belting and Asbestos Ltd., working directly for Dr. Charles G. Addingley, the Chief Chemist. Both Mr. Quin and Dr. Add- ingley were Leeds graduates, so it was natural that after leaving Whitcliffe Mount he went to Leeds and undertook the accelerated version of the course known at the time as “Special Studies in Chemistry.” That he remained at Leeds following the death of his father at the end of his first year was due to Professor (now Lord) Frederick Dainton’s advice and arrangements for practical support at what was a difficult period.He graduated in 1955 and stayed on to do research in physical chemistry under Dr. G. F. Smith with the aid of a West Riding Research Scholarship, obtaining a PhD in 1959. His first appointment was to Medway College of Technology, Chatham, as an Assistant Lecturer in Physical Chemistry in 1958, becoming a Lecturer and an ARIC in 1959. It was whilst at Medway he first came in contact with Professor Ronald Belcher via the Midlands Associa- tion for Qualitative Analysis.Over the following years, Ron’s encouragement, support and timely advice was greatly appreciated and is now very much missed. In 1963 he was appointed to Woolwich Polytechnic as the Senior Lecturer in Analytical Chemistry. This post was not achieved without some difficulty since the first attempt failed, as the appointing panel lacked the necessary members of the Governing body as the result of a serious London fog. Dr. A. I. Vogel, the then Head of Department, was a hard but fair taskmaster, well read, with an amaz- ingly broad knowledge of chemistry. He well remembers in the middle of his first term the Chief Steward, Vernon Kyte, A.I.V.’s usual emissary, asking what had been published since arrival.As luck had it, the reply was “I’m correcting proofs!” From that time on a good relationship existed between himself and Vogel. Membership of the SAC (under Harold Brookes’ guidance) and a return to the Midlands were both effected in 1966 upon appointment to Loughboroug h Univer- sity of Technology as Senior Lecturer and Head of the Analytical Chemistry Sec- tion. Professor Reginald F. Phillips gave very positive encouragement to the estab- lishment of Short Courses in Analytical Chemistry, an MSc course and an ana- lytical research grouping. Transfer to a Readership occurred in 1971. He was awarded the James Taylor Prize by the Sheffield Metallurgical and Engineering Society in 1970 and the first substantive DSc degree by the University in 1972. Whilst at Loughborough he was active in the Microchemical Methods Group, the establishment of the Education and Training Group and the Midlands Region, becoming Chairman of that Region just before appointment, in 1975, to the oldest UK Chair of Analytical Chemistry, that is, that at The Queen’s University of Belfast.Upon arrival in Belfast the first task, with the enthusiastic support of his academic and technical staff, was to estab-ANALYTICAL PROCEEDINGS. JULY 1988, VOL 25 21 1 lish an SERC and DEN1 recognised MSc course in Analytical Chemistry, followed by a Northern Ireland Region of the Analytical Division. Via the good offices of the Scottish Region a Northern Ireland Sub-committee was quickly set up and in 1980 became a full Region, of which he is the current Chairman.Other Divisional activities included Chairman of the Mi- crochemical Methods Group, Member- ship of Programmes, Group Liaison and Policy. Analytical Methods, Honours and Trust Advisory Committees. Within the KSC he has served as Chairman of the Northern Ireland Section, a member of the Working Party for the Indicative Register for Analytical Chemists and on the Open Tech Project. He is the present Chairman of the Local Committee for the 1990 Annual Congress, which is to be held in Belfast. Within Queen’s University he has been active as a Member of Council and, as Chairman of the University Safety Com- mittee 1979-1985, designed and deve- loped a structure to cope with the onset of the Health and Safety at Work Order (NI). He is a Professorial Curator in his second term of office and was Chairman of the Department of Chemistry 1982- 1986.In 1982 he was the Theophilus Red- wood Lecturer and also received the RSC Award and Medal (sponsored by BDH) for “Analytical Reactions and Analytical Reagents.” In addition to studies in reac- tion chemistry and solvent extraction - spectrophotometry, work has been car- ried out in instrumental methods includ- ing chromatography, luminescence, elec- troanalysis, atomic spectrocopy and elec- tion spin resonance spectroscopy. His- torical topics studied include the relation of British and Irish Analytical Chemistry in a European context, which produced a series on Robert Boyle and on the early analysis of solutions. Recent and current historical research concerns the prepara- tion of a definitive history of optical spectroscopy.Much of this latter has been presented as ilivited lectures both at home and abroad and in invited publications. Overall, he is author and co-author of some 230 papers and reviews and has co-authored 4 books and contributed to 3 others. The year 1984 was also of significance as he was elected to the Fellowship of the Royal Society of Edinburgh and also to Membership of the Royal Irish Academy. Interest in the affairs of the Federation of European Chemical Societies came in 1980 when he attended the FECS Work- ing Party for Analytical Chemistry (WPAC) meeting in Graz as an observer, introduced by Professor R. Belcher. Since that date he has attended every meeting as either an observer, UK alternate or Irish delegate, and has attended all of the Euroanalysis Conferences, being heavily involved in the organisation of Euro- analysis 111, held in Dublin, as well as giving the Opening Lecture.It is thus with great pride and pleasure that at Euro- analysis VI, Paris, he heard that Euro- analysis VIII would come to the UK, to Edinburgh in particular, under his Presidency. In the International Union of Pure and Applied Chemistry he is currently Chair- man of the oldest commission V/1 “Ana- lytical Reactions, Reagents and Separa- tions,” having previously served as its secretary. He was a member of the Royal Society delegations to the IUPAC Coun- cils in Lyon (1985) and in Boston (1987). He is currently a member of the British National Committee for Chemistry, Chairman of its Analytical Sub-Commit- tee and a Member of the Royal Society - RSC Joint Committee on Chemical Nomenclature.He is the current Chair- man of the Northern Ireland Branch of the British Association for the Advance- ment of Science and was on the local committee for the 1987 Annual Meeting held in Belfast. Professor Burns also serves the Province with the Department of Education NI Advisory Board on Post Graduate awards and Membership of the Poisons Board. By way of relaxation, like most recent Presidents, he enjoys fine food, wine and good company. He also spends a deal of time in antique and book shops in search of material in support of his historical studies. So far none of Professor Burns’ three children has followed him into chemistry, Mary Jane is a professional caterer, Susan a second year BEd geography student and James, the last hope for the subject, has three years to wait for before his 1l-t examinations. The incoming President is conscious of the merits of his predecessors and that as the first representative from Ireland, North or South, since Sir Charles Carn- eron 1893-1894, he has much to live up to. He is looking forward to the challenge and also to the pleasures in store, includ- ing SAC 1989.

ISSN:0144-557X    

DOI:10.1039/AP9882500210

出版商:RSC    

年代:1988

数据来源: RSC

摘要:

ANALYTICAL PROCEEDINGS. JULY 1988, VOL 25 21 1 Obituary Dr. C. A. Johnson CBE C. A. Johnson (“Johnny”) died suddenly at his home in Sutton, Surrey, on Mon- day, May 2nd, 1988. Although he had officially retired three months earlier as Secretary and Scientific Director of the British Pharmacopoeia Commission, he was still involved with many residual tasks: his death occurred just after his return from Strasbourg, and before an intended visit to Stockholm. Johnny’s initial training was in phar- macy, and his career started in phar- maceutical analysis; first with consulting analysts, then briefly with the Phar- maceutical Society of Great Britain, fol- lowed by nearly ten years with Boots in Nottingham, where he became head of their Analytical Development Group. Just 25 years ago he joined the staff of the British Pharmacopoeia Commission, where he became Scientific Director and, in 1976, added the post of Secretary.As well as directing the progress of the British Pharmacopoeia, he played a major part in the development of its European counterpart, serving as Chair- man of many expert groups, and, in 1977-80, as Chairman of the European Pharmacopoeia Commission. During that time he also fostered liaison with the US Pharmacopeia. By great good fortune Johnny’s Head of Department at Boots was the late Dr. D. C. Garratt, who encouraged him, with many others of his staff, to take an active interest in the Society for Analytical Chemistry. Johnny’s interests in anything were never superficial and he contributed with vigour and effect to the affairs of the SAC, subsequently the Analytical Divi- sion.A brief account of his achievements was published in Analytical Proceedings (1983, 20, 196). Johnny considered that his main contribution lay in developing the SAC Conferences, leading up to the Society’s Centenary Celebrations. He was Honorary Secretary of the first and second SAC Conferences in 1965 and 1968 (Nottingham) and the third in 1971 (Durham), and was Chairman of the Executive Committee of the 1974 Cen- tenary. He was justifiably proud of the organ- isation of the Centenary, and ascribed its success to the individual work of Commit- tee Members. Those involved know that it was founded on Johnny’s enormous energy and meticulous attention to detail which guided the near-impossible task of holding a major international conference in London in July to a well-nigh perfect conclusion. Those who attended the Opening Ceremony in the Royal Institu- tion may remember the background music (chosen by Johnny) and his announcement of the 27 participating societies, each foreign one first in its own language-an exercise which, with char-212 acteristic thoroughness, involved prior trips with a tape-recorder to various embassies and travel agencies.Johnny received awards from many bodies, which he accepted with pride and modesty. In 1967 he was awarded the Harrison Memorial Medal of the Phar- maceutical Society of Great Britain. In 1983 he was awarded the Distinguished Service Award of the Analytical Division. In 1985 he was awarded an Honorary DSc by Bradford University, and in the same year he was made an Honorary Member of the Nobile Collegio Chimico Far- maceutico in Rome.His appointment as a CBE was announced in the last New Year’s Honours List, and he was to become an Honorary Member of the ANALYTICAL PROCEEDINGS, JULY 1988, VOL 25 Hungarian Pharmaceutical Society in October, 1988. Johnny’s reputation as a speaker on topics ranging from the specialist scien- tific to sheer entertainment attracted invi- tations to a host of countries from Canada to Australia and from the Far East to the Far West. The foregoing account describes some of the activities of a man of exceptional energy and clarity of mind. To Johnny’s many friends at home and abroad it is only part of the story. He was well educated in language, in food, in drink and in music (he was known, on rare occasions, to enliven the streets of foreign cities late at night with delightfully sung arias).He had the ability to relax with friends over a well chosen meal preceded, accompanied and followed by equally well-chosen drinks, and enlivened by leisurely, varied and stimulating conversation. His travels with friends to north-east England, to France, Belgium, Ireland and Hungary, some- times on unplanned car-trips, engendered a wealth of anecdotes which he had no difficulty in remembering. Johnny’s work for many organisations, both scientific and social, and his friend- ship with countless individuals world- wide, will be sadly missed and long remembered. He leaves Theresa, his widow, three daughters and a son, and five grand- children.D. W. WILSON SAC Silver Medal On the recommendation of its Honours Committee, the Council of the Analytical Division, at its meeting on May l l t h , awarded the Sixteenth Society for Analytical Chemistry Silver Medal to Dr. R. D. Snook of the Department of Instrumentation and Analytical Science of the University of Manchester Institute of Science and Technology. ALAN DATE MEMORIAL AWARD In recognition of the considerable contribution made by Dr. Alan Date to the field of inductively coupled plasma source mass spectrometry, an annual commemorative award has been created to encourage talented young scientists to broaden their overseas scientific experience. Candidates should be working in the field of atmospheric pressure plasma source mass spectrometry (inorganic) and will be required to submit a written resume of their work in this field. It is intended that this will form the basis of a publication. An amount of up to S1,OOO will be available to the successful candidate(s) Further details of the award may be obtained from: Dr. Robert Hutton, VG Elemental, Ion Path, Road Three, Winsford, Cheshire CW7 3BX.

ISSN:0144-557X    

DOI:10.1039/AP9882500211

出版商:RSC    

年代:1988

数据来源: RSC

摘要:

ANALYTICAL PROCEEDINGS, JULY 1988, VOL 25 213 Retiring President‘s Address Analytical Chemistry: a Question of Decisions D. C. M. Squirrel1 9 Graysfield, Welwyn Garden City, Hertfordshire AL7 4BL It will be clear from the title that this address is to be somewhat philosophical in nature. When one has been away from front-line practical research for more than a couple of years the philosophical line is probably the right one to take, for one has had, at least in theory, the opportunity to reflect a little on past experience, present day and longer term problems and on changes likely to occur in the future, particularly in one’s own area of the profession. Analysis, as we all know, is all about questions, answers and decisions, and the gift given to all good analysts, the gift of analytical thinking.My generation has been very fortunate in that our working experience has covered the 30-40 year period linking the best of classical analytical procedures, through the wide use of commercially produced instrumentation, which increased rapidly in sophistication in the 1960s, to the computer revolution and the ultra-modern technology we have today in the 1980s. A history book of progress has been encompassed within our workspan and so we should perhaps be better equipped than many to comment on the questions and decisions that the analyst has faced and is preparing to face in the future. I would like to consider the following aspects of our profession and discuss, as appropriate, where change may be required. I acknowledge that some of the points I shall cover arise from discussions with members of the Division and from opinions expressed by speakers at our scientific meetings.My grateful thanks for those discussions. In addition, since preparing these notes, some of the points discussed have also been touched on in the recent literature. Events have overtaken me. Let us consider the analyst and his profession; the job and the job-holder; the person, his education, training and role; and the tools of the trade. The job is, of course, different depending on whether we are in: manufacturing industry, heavy or light; manufacturing industry, batch or continuous; public service; clinical biochem- istry; education, secondary; education, tertiary; research and development; production and quality control; or quality assurance.In all instances, however, we are concerned with data and effective communications. The person or “personality” portrayed by the job holder may also be different depending on whether we are concerned with industry (analytical research, technique specialism, quality control analysis and quality assurance, standardisation - legislation and analytical management) or education (course structuring - organisation, teaching, examining, research - development supervision, generation of ideas and consul- tancy). All the above “personalities” will, of course, have or should have an interest in the role of their professional society and its publications and will have realised long ago that their profession is now one that requires the knowledge and skills not only of chemistry but also of other disciplines.Let us now look at the job and also at what should not have changed over the years, the basic structure and requirements that will remain common to all the variants of the job. 1. A PROBLEM, request, question and raises Our work starts with 2. A QUESTION regarding the background to the problem leading to 3. A DECISION-concerning priority and need followed by another 4. DECISION-on the analytical action-WHO?- Then follows 5. THE EXPERIMENTAL-any development required. WHAT?-WHERE?-WHEN? -the sampling, preparation, separation and measurement. Followed by 6. VALIDATION of the data produced 7. INTERPRETATION of the data 8. DECISIONS on CONCLUSIONS and REPORT 9. DECISIONS on RECOMMENDATIONS for ACTION The problem has been that too many in the profession have confined their interest to stages 5-7 and to a lesser extent stages 4 and 8 in the above list.The analyst should be making an appropriate input at all stages. In industry, the analyst must be concerned to understand well the manufacturing processes used in his sector; to realise that his Company does not sell analysis but does sell quality, analysis being the means to an end, an overhead; that efficiency and cost effectiveness are of paramount importance and that particular and frequent attention must be paid to the why? when? where? and how? of the analytical requirements. With obvious modification to detail the above principles hold also in the public service and in the clinical - biochemical areas. Within some industries, one of the principal distinctions in the job lies between the research and the routine process/ quality control functions.Although different in content, there is a common objective of improved service and efficiency. Analytical research is carried out to provide for the needs of the industry in terms of support for new product development, process problem solving and method, combined with instrumentation systems, development. Many of these needs will be communicated via the production analyst who, whilst having a prime objective of keeping production and quality demands fully satisfied, must always be looking for points where improvements can be made. This requires an awareness of research advances and a willingness to try out new principles in harmony with research colleagues.In some instances too, a sharing of development and market research costs may be appropriate. The essence is co-operation, communication, concern and co-ordination. Where a large amount of analytical process control is involved, particularly in continuous and large batch processes, much joint work with engineering and computer services departments is also most essential. With increasing demands more than keeping pace with the improved analytical capability brought about by modern technology, analysis must be well managed at every level. In some instances this will mean changes, changes not only in equipment, technique and methodology, all with built-in quality assurance protocols, but also in the better use of all available information and analytical data to produce more meaningful and effective conclusions, recommendations and decisions.This may or will, particularly in larger companies, require ownership of computerised laboratory information214 ANALYTICAL PROCEEDINGS, JULY 1988, VOL 25 management systems (LIMS) and appropriate understanding of the installed system is required by all levels of staff from technician to senior manager.Experience in this and other high technology analysis areas is expensive to obtain but is clearly ready to be shared by those who have it. Contributions to our own Analytical Division meetings, in particular the Automatic Methods Group, have clearly demonstrated this and such opportunities to learn from others should not have been missed and certainly should be taken up in the future.Education and Training This brings me conveniently to education, which in our profession more than any other is a continuing exercise. More than 99.9% of analytical development is carried out outside the particular laboratory in which we work and so to keep up to date we have a big learning problem, only part of which is to be solved by reading our journals, excellent though they are. Attendance at scientific meetings, personal contact and discus- sions are most important if we are to improve performance. What are the education, training and experience require- ments for analysts to fill posts at all levels in the hierarchy of our profession , from analytical technician through analytical scientist to analytical manager and beyond? At the start it is pleasing to see that much more is being done by the chemical industry and by the Industrial, Education and Analytical Divisions of the RSC, and others, to revive, at secondary education level, the interest and enthusiasm for chemistry that we all knew in our schooldays, for it is there that real interest in our science begins.One thing is clear, more good and enthusiastic teachers, properly qualified in chemistry, are an essential for the future. The question is, what can be done to ensure that they are recruited? This must be a subject for a separate discussion. The training of analytical technicians is well catered for by the BTEC courses and those leading to HND, HNC and LRIC qualifications. These will also be helped by the Analytical Chemistry by Open Learning (ACOL) system.At all levels of tertiary education, the need for high quality teaching is clearly accepted although the pressures on some teaching staff appear to be increasing to an unsatisfactory level. I believe that the big question for the future will be-What to teach at first degree Honours level? There is clearly a limit to the amount of information that can be imparted, understood and practised in the laboratory in a 3-year full-time course. Opinions differ widely as to what is the ideal tertiary training for a potential analytical chemist. The word potential is important here, as potential is not always fulfilled. Some industrialists prefer to recruit a person with a good Honours Degree in Chemistry with the usual ancillary supporting subjects and then impart the necessary additional training in advanced analytical techniques, electronics, com- puting, management systems, control systems, etc., by post- graduate courses and in-house training over several years, tailored to the demands of the industry, the job and the career aspirations and capability of the job holder.This, they feel, gives the right background and balance to enable the complete senior analyst to develop or evolve. Some companies will require a research degree as the starting point, but this may not necessarily be in an analytical subject; the idea being to start with someone with a broader based experience in chemistry. Other, usually smaller, companies may go to the other extreme and would prefer a person with a degree essentially in analytical science, covering the relevant components of chemistry, physics, instrumentation, electronics techniques, computing, control systems, statistics, etc.They want to be able to employ someone with a multi-discipline graduate training such as will enable them to cope with all eventualities. Of course, even with such training, considerable experience will still be required to ensure competence and further, it will be necessary to maintain the chemistry content of the course at a level which would still permit entry as a Graduate Member of the RSC; that is, if the graduate expects eventual registration as a Chartered Chemist, and this could be important in future career moves. Analytical research within our universities and polytechnics must continue at a realistic level, both to maintain our excellence in new science developments and to provide stimulation to the teaching staff.Industry can devote only limited personnel and resources to speculative ventures and they and instrument manufacturers must rely heavily on “front-line” activity within academia. The emphasis should, however, be on relevant front-line activities. The answer to both funding and “relevance” problems must lie in the recognition of the importance of analytical research to science and the community and in more joint ventures between the research schools, industry, medical departments and instru- ment manufacturers, in addition to appropriate funding from research councils. Much has been done in the encouragement of collaborative efforts and these must continue to be facilitated.Role Having produced our potential analysts, what will be their role within industry or public service? I have already made the point that whatever position they may be in, if they want to succeed in terms of reaching the more senior positions, they must take a wide interest in and make significant inputs into the whole cycle of analytical influence. This may not always be easy and it should be realised that not all analysts will reach senior positions. Qualification itself is not a passport to fame and, as in other professions, room at the top is limited and will in many instances only be open to those who have the aforementioned flair, wider interests and experience. Within the analytical experience cycle, analytical research and method development for new productsiapplications is important, both as a service to the employer and also in the provision of excellent training.This, coupled with application work in new techniques and sometimes actual instrument development, will often bring out and demonstrate an analy- tical flair in some particular area of the science which may usefully lead on to a longer period spent on instrumental specialisation and/or the development of a centre of expertise. However, the analyst remaining in a “narrow” field for a long time may have to consider a career development move in the light of the supply to demand ratio of the expertise that he offers, unless, of course, he is dedicated only to a particular area of analytical science and the value of his expertise and reputation, both internally and externally, is fully recognised by his employers. What then of the super specialist in a new technique, conceived and developed in a university and judged to be of great commercial potential by an instrument company with a significant application potential in industry.Such expertise may well be in high demand, particularly during the launch years of the technique and its associated variants, and the young researcher will keep an eye on the limitations of the job market in a narrow field and will judge when to expand his horizons. Expand he will certainly have to, eventually. Some questions that must be asked are: do we suffer from a too narrow view of our over-all potential?; do too many take the attitude “Once an analyst: always an analyst! (and I won’t be very good at anything else)” to the extreme of isolationism?; should we not more often consider where else, within the activities of our employers, our skills can be effectively applied? Quality Assurance Good quality results is something that all responsible analysts have always tried to ensure by careful standardisation, calibration, the use of internal standards, replications, check methods, “round robins” and the use of certified referenceANALYTICAL PROCEEDINGS, JULY 1988, VOL 25 215 materials, etc.Therefore, to most of us Analytical Quality Assurance and the concepts of Good Laboratory Practice are not new. The difference now is that, quite rightly, we have to assure others that our results are valid and are of the substance and quality demanded.This means better use of statistics and validation programmes, documentation of quality assurance protocols, documentation of all quality assurance actions and the use of control charts, etc. I do not intend to dwell on the detail but do wish to make the point that, depending on circumstances, up to 20% of analytical effort may have to be allocated to quality assurance. Also, the degree of automation in the analysis systems and methods in use can markedly influence the manner of the quality assurance. The “black boxes” must produce results of the guaranteed quality deman- ded and must be shown to do so. A high grade of supervisory staff and training is thus still required in the modern labora- tory. In general, although the need for conventional manipula- tive analytical skills may have diminished for routine work in some areas, they are still an essential for many non-routine operations and in “general” laboratories. I suggest, therefore, that for a person wishing to be regarded as an analytical chemist, training and perfection in such skills remains neces- sary.The Tools of The Trade So what are the tools that the analyst has at his disposal at the close of this decade? Most important, assuming a sound knowledge of chemistry and other analytical sciences, are the human attributes of integrity. high intelligence and the concept of analytical thought. Memory and speed of recall can, of course, be assisted by computer technology and in some areas short-term relief can be obtained from expert systems.The human sensors and senses continue to be important, not only in observing and recording data with real-time interpreta- tion and reaction, but also in selective communication. Our efficiency is, however, greatly enhanced by the aids to our hearing, sight, touch, smell and memory provided by modern analytical instrumentation, now capable of once unimaginable resolution and sensitivity. It is not my intention to review the field of modern instrumentation: our Analytical Division subject groups keep us well informed in this area. I simply wish to say that although the headlines often capture only the highly sophisticated and expensive instrumental techniques, large increases in human efficiency can be obtained from the innovative use of low and medium priced machines, particu- larly when they can be left to work, even for quite short periods, unattended.I say this in encouragement to some smaller laboratories who have to work on a very limited budget or have a limited sample throughput. Some re-thinking of methodology will, however, almost certainly be required. In some larger industries, the service industries and in the clinical field, where very large numbers of similar samples are submitted for examination for a variety of specific analytes, the pros and cons of the central laboratory equipped with high cost automated equipment, the smaller, dispersed, analytical facil- ity, on-line or off-line analysis and remote. unmanned analyser systems, will continue to be evaluated and many new systems and protocols tried.It is an exciting period in this area and one in which the multi-disciplinary nature of analytical science will be most evident. Automation is no threat to the professional analyst but it may be a challenge to his flexibility and powers of lateral thinking. Requests for more attention to be paid to the solution of sampling and sample presentation problems. particularly in relation to automated and on-line systems, has understandably not produced much published information on the subject, other than for a small number of generally applicable units. Each commercial process presents its own problems and security of the process may well inhibit publication of details. Sampling thus presents, and will continue to present, analysts with a challenge requiring lateral thinking and a chance to show their engineering as well as chemical prowess.Finally, on the subject of tools of the trade, can I repeat some of the information from a slide I used in a lecture on automated analysis, given in the early 1970s. Table 1 sets out the differing requirements for an instrument or analyser system to be used for research as compared with routine process analysis. Table 1. Different requirements for research and process control anal ysers Requirements-Research Requirements-Process control Important lmportan t Flexibility Speed and reliability Maximum information High reproducibility High accuracy Adequate sensitivity High sensitivity Simplicity of operation Reliability Ease of calibration Less Important Less Important Simplicity of operation Flexibility Ease of calibration Operator attention Minimum of maintenance High absolute accuracy With the addition of entries about data transfer and quality assurance these listings still apply.It is, however, sometimes more difficult these days to buy simplicity for process analysis applications. This can present maintenance problems if the superfluous capability is mechanical and make understanding by junior staff more difficult if the control and data processing options are over-complex to select. Concerns In all the foregoing, I have simply tried to record background experience in order better to consider the alarm bells which I hear ringing in some quarters in relation to the future of analytical chemistry and those practising, teaching and manag- ing in our profession.Let us consider some examples of these SOS messages: (1) S 0 S Sampling Optimisation Standard materials S (2) S 0 Standard methods Obsolescence Statistics ( 3 ) S 0 S Scholarship Outlook Status sos- I Sampling. The need for continued attention to the develop- ment of new sampling ideas and the avoidance of sampling errors has already been mentioned. One action here is to publish our ideas, even simple ones, whenever this is possible. Optimisation. Optimisation of analysis conditions is import- ant but it must be realised that over-optimisation for a single analyte or function may render the procedure more liable to interferences and go against the desire for robust methodology. particularly for process control. Standard Reference Materials. These are essential for both manual and automated test procedures, particularly for pri- mary standardisation, method validation and quality assurance checks. There is clearly a need for an extension in the range of standard materials commercially available with full informa- tion on correct storage conditions and an indication of shelf life,216 ANALYTICAL PROCEEDINGS, JULY 1988, VOL 25 sos-2 Standard methods.Standard methods of analysis that have been well validated are, of course, recognised as extremely valuable and indeed essential, particularly when they form part of a specification defining the quality or performance of a product. The arguments that the standardisation can slow up improvements in methodology do have some foundation, however, psrticularly when new equipment is involved or when historical comparison with past results is being made against final product performance.Change may not always be readily accepted unless sufficient work is done to correlate the results obtained by the improved method with that of the old. There seem to be sound arguments too for avoiding unnecessary mandatory detail in legislative methods. Where minor changes, controlled by performance checks, can be tolerated to allow for local circumstances and the availability of equipment, then these should be clearly stated. Obsolescence. Obsolescence is a worry which has been voiced in terms of instrumentation, methodology and indeed, people. Rapid technological progress, almost by definition, can lead to rapid obsolescence of the knowledge we have and the equipment we buy.This we know to be true of domestic appliances as well as scientific equipment. I am sure that instrument manufacturers will do their best to keep our older instrumentation supplied with spares and service but there will clearly be tighter limitations which may well worry the user about to spend part of a controlled budget on an expensive piece of hardware. Obsolescence in methodology is, in part, also linked to speed of instrumental development and the validation of new techniques but, as indicated under standardi- sation, interim improvements need not be inhibited in an initially robust procedure. Statistics. The value of correctly applied statistics is fully appreciated. The fears that are expressed about statistics and now chemometrics, I think, largely stem from a lack of understanding of the influence that these tools will have.There is a communications or education job to do and, as always, there is a need for the simplest treatment possible, capable of providing the required information, to be recommended. SOS-3 Scholarship. 1 have already said something about education and training. No-one disputes the need for well educated and trained, chemistry based, analytical scientists. The main worry stems from the question whether analytical science will receive its correct share of the training resources that are available and in how these resources will be used. More enquiry and collaboration with and between, for example, industrial, medical, public, and local and central government users of analytical services, is clearly necessary.Outlook. The outlook of some seems a little despondent. This is not surprising in areas where serious cut-backs have occurred and where the employment of analysts has been affected. If an over-all study was made, however, I would be surprised if within industry and public service, analysts had not fared better than other scientists and engineers. Current employment prospects are also equally good for good candi- dates. Status. Worries on this question I find particularly difficult to deal with. Some analysts genuinely feel that their profession is considered, even by other chemists, as the poor relation, more vulnerable than others to the effects of organisational changes, rationalisation and other influences.The experience of others does not support this view and some note a marked improve- ment in recognition in recent years, with analysis receiving a larger slice of the available cake. There is clearly a public relations or promotional job to be done in some areas that is not necessary and could even be counter productive in others. Status is a factor which thus requires very careful considera- tion, perhaps on an individual basis, taking into account many of the factors and influences that I have previously outlined. Certainly, however, we must seek proper representation with our organic, inorganic and physical chemistry colleagues on committees and executive bodies responsible for resource allocations. For the future, it must be remembered that analytical science does not operate in isolation and individual performance will alway outway the corporate image, which I am sure is itself on an upward trend. Finally, I must emphasise that in the discussion of “concerns” I have, of necessity, had to concentrate on some negative aspects in our profession. There are, of course, overwhelmingly more positive and exciting facets which provide great job satisfac- tion, and with these all analysts are familiar. I am of the opinion that some of the apparent problems are beginning to be understood and that provided we accept the need and inevitability of further changes (and we have already seen many) then analytical science and those who are engaged in it will meet all the challenges placed before it and remain in a position of key importance, where chemistry interacts with all other scientific disciplines to serve both science and the community. Chair for Analytical Chemist We are pleased to announce that the first holder of the endowed Philips Chair in Analytical Chemistry in the Department of Pure and Applied Chemistry of the University of Strathclyde will be Dr. D. Littlejohn.

ISSN:0144-557X    

DOI:10.1039/AP9882500213

出版商:RSC    

年代:1988

数据来源: RSC

摘要:

ANALYTICAL PROCEEDINGS, JULY 1988, VOL 25 217 Professor Ottaway Memorial Meeting The following are summaries of twenty of the papers presented at a Joint Meeting of the Analytical Division and the Scottish Region hela on November 5th and 6th, 1987, in the Royal Society of Edinburgh, George Street, Edinburgh. A further paper, by Professor M. L. Hitchman, appeared in the June issue of The Analyst, p. 875. Graphite Furnace Atomic Emission Spectrometry-the Rediscovery of a Technique D. Littlejohn Department of Pure and Applied Chemistry, University of Strathclyde, Cathedral Street, Glasgow G 7 7XL Although I had known John Ottaway since 1971, I did not become fully involved with his research group until 1975, when I started my PhD studies. By this time, John and his students had initiated a number of projects devoted to the study of atomisation and interference processes that occur in elec- trothermal atomisation atomic absorption spectrometry (ETA- AAS) .When conducting chemical interference experiments, one of John’s senior students, Bill Rowston (sadly also no longer with us), noticed that flashes of radiation were produced from the vapour in the centre of the tube during atomisation. This encouraged another of John’s students, Frank Shaw, to make further measurements, this time with the spectrometer in the emission mode, and a number of sensitive “graphite furnace atomic emission spectrometry” measurements were obtained for a variety of elements. A brief literature survey quickly indicated that the measure- ment of atomic emission from an electrically heated tubular furnace was not novel.John and his students had, indeed, re-discovered a procedure used 70 years earlier for fundamen- tal spectrophysical studies of atoms and molecules. Those interested in the history of the subject are probably well aware of the general supposition that electrothermal atomisers, as used in ETA-AAS, derived initially from the work of King in the early 1900s.l.2 However, it may be of interest to note that King ascribed the development of electrically heated tube atomisers to Liveing and Dewar,-? as is evident from the following extract from one of King’s earliest “furnace emission” publications. 1 “Attempts to produce spectra in tubes tested in furnaces were made by several early investigators, but the first really effective method seems to have been that used by Liveing and Dewar in the course of their experiments on the reversibility of lines .. .” It was true, however, that no one had made use of this phenomenon for analytical purposes, and in this respect the Strathclyde group’s work was innovative. The work of Ottaway and Shaw on atomic emission measurements with a modern graphite furnace atomiser was first published in 1975,4 and in this preliminary appraisal, the authors concluded that: “It appears, for a number of elements at least, that the technique of carbon furnace atomic emission spectrometry warrants further detailed examination if extremely low detection limits are required.” This statement was a vague reference to some basic limitations of AAS instrumentation used in ETA-AES.In particular, John was concerned about the need to make a second “blank tube” measurement after every analytical signal to compensate for the continuum background emission pro- duced by the hot atomiser tube during the atomisation stage. Also, he was aware that the peak atomic emission intensity did not occur until after the peak atom concentration had been achieved and so improvements in excitation conditions were clearly required in order to make best use of the analyte vapour produced by a sample. Another thing that intrigued John, right from the start of his research in graphite furnace (now, electrothermal atomisation) AES, was the mechanism of atom excitation. A number of papers produced in the mid-1970s by other authors had suggested that conditions of local thermal equilibrium (LTE) were not achieved in a graphite furnace during the atomisation stage.5.6 John and I, in particular, could not understand why this should be so.It was not surprising, therefore, that research conducted at the University of Strathclyde on the subject of ETA-AES covered: firstly, fundamental studies; secondly, the develop- ment of automatic background correction routines and other spectrometer studies; and thirdly, atomiser studies designed to improve excitation conditions. Fundamental Studies The principal aim of work conducted in this area was to answer two questions: firstly, is excitation in ETA-AES thermal or non-thermal, and secondly, what mechanisms contribute to atom excitation in a graphite tube atomiser? If LTE exists, then the same value of temperature describes the energy distribu- tions associated with the kinetic energy of electrons and macro-species (atoms and molecules), rotational and vibra- tional excitation in molecules, atom excitation and ionisation and the spectral distribution of the continuum tube wall radiation, in accordance with expressions derived by Maxwell, Boltzmann, Saha, and Planck.It was not possible to make measurements to investigate the energy status for all of these phenomena. However, temperatures calculated for vibrational excitation (from CN band emission), atom excitation (from Ni and Fe atom emission) and ionisation (from atom and ion emission intensities of Ba, Ca, Eu, Sr and Yb) were sufficently close to one another, and to the radiation temperature derived from the spectral intensity distribution of the tube wall radiation, that LTE was considered to exist under the “gas stop” conditions that applied during ETA-AES measure- ments.7-9 This conclusion was supported by independent AAS measurementslO published about the same time as the excita- tion study results.With regard to excitation mechanisms, the existence of LTE conditions makes it difficult to comment with any certainty upon the relative contributions of different processes. However, it is probable that molecular collisions are a major source of excitation and de-excitation in a graphite furnace atomiser. Even when argon is used as the furnace purge gas8 there is a sufficient concentration of impurity molecules, such218 ANALYTICAL PROCEEDINGS, JULY 1988, VOL 25 Table 1.Effect of instrument development on ETA-AES detection limits of twenty-two elements Number of elements in detection limit rangeing ml- <0. 1-1 1-10 1u-100 100-1000 1977 Wall atomisationino auto 4 6 10 2 1982 Wall atomisationiauto 11 5 4 2 1982i5 Platform or probe atomisationi 14 5 3 0 background correction background correction auto background correction as O2 and N2, in the furnace environment to establish LTE conditions.8." There is, however, another source of excitation available in a graphite tube atomiser that does not normally feature to any great significance in many other emission sources (e.g., combustion flame, arc, microwave induced plasma). In a graphite tube, the sample vapour is surrounded by a cylinder of intense continuum radiation and, at the temperatures used for atomisation and excitation, photon absorption can make a significant contribution to atom excitation.B.11 This conclusion was supported by measure- ments made at a reduced pressure of 2 torr in a sealed electrothermal atomiser, which indicated that under similar analyte atom concentrations and tube temperatures the atomic emission intensity for Cr was similar to that achieved at atmospheric pressure.11 When an electrothermal atomiser tube is heated to high temperatures, electrons are produced by thermionic emission and concentrations in the order of 101"-1012 electrons cm-3 are produced. Calculations have indicated that collisions with electrons are a possible means of excitation in ETA-AES, especially at temperatures in excess of 2500 "C, but this process is less important than molecular collisions and photon absorp- tion.8-11 Automatic Background Correction in ETA-AES In order to introduce some form of automatic background correction in ETA-AES measurements, conventional atomic absorption spectrometers were abandoned in favour of spec- trometers modified for wavelength modulation by insertion of either a vibrating quartz refractor plate12913 or a rotating quartz chopperI4,ls immediately after the entrance slit or before the exit slit. The procedure of wavelength modulation allowed a small spectral interval around an analyte line to be scanned repetitively so that off-peak measurements of the background could be used to correct the signal at the analyte wavelength.Displacement of the wavelength at the exit slit was achieved by vibrating a quartz plate of fixed thickness through a 15" anglel2.13 or by inserting quartz plates of different thicknesses into the optical beam at the same angle, in the form of a quartz chopper or sectored wheel. 14315 With modulation frequencies of 20-40 Hz, giving corresponding detection frequencies of 4&80 Hz, the speed of background correction was fast enough to cope with atomic emission signals produced when modern electrothermal atomisers were used as an emission source. Up until the introduction of wavelength modulation background correction, the detection limit in ETA-AES was limited by the amount of scattered black-body radiation that entered the monochromator along with the analyte emission.16.17 The deployment of wavelength modulation had a significant influence on the detection limits achieved by ETA-AES," as indicated in Table 1.The application of the technique to the analysis of complex matrices was also enhanced by this instrumental development and a number of methods were developed for the determination of trace metals such as Cr, Cu, Mn, Pb and others in clinical samples.1X-22 The procedures generally involved minimal sample preparation (e.g. , dilution with water) and detection limits for the elements were around 1 ug 1-1 or lower. Recent work on the computer control of wavelength modulation has produced a system that is capable of perform- ing ETA-AES measurements and continuum source atomic absorption spectrometry (CSAAS) with a single instrument system.2"24 In many respects, this arrangement is a modern equivalent of the flame atomic absorption - flame emission duality of commercial AAS - AES instruments.Atomiser Studies to Improve Excitation Although the ability to compensate for unwanted background radiation was an important breakthrough in the development of ETA-AES, the need to improve excitation conditions, for volatile elements in particular, also had to be addressed. When samples are vaporised from the wall of a graphite tube, atoms of volatile elements are produced at comparatively low temperatures and are lost from the tube by diffusion before the maximum temperature can be achieved. In order to obtain volatilisation into a higher temperature environment, a num- ber of procedures were investigated, as itemised in Table 2.Examples of modified atomiser tubes that are used to give improved excitation are given in Fig. 1. In tubes C and D, the Fig. 1. Modified Perkin-Elmer HGA72 graphite tubes for graphite furnace atomic emission spectrometry: A , the "high-temperature" tube; B, the tapered tube; C , the volatile elements tube; and D, the cup tube end sections were of thinner wall thickness and heated faster and to a higher temperature than the central section, where the sample was deposited either on the wall (tube C) or in a cup (tube D). As the end sections heated the vapour prior to atom formation, volatile analytes were vaporised into a higher temperature environment than would have been possible with atomisation from the wall of an unmodified tube.25 Delayed atomisation to achieve volatilisation into a high temperature environment was also achieved by use of platform atomisa- tion26.27 and probe atomisation.28.29 A comparison of ETA- AES detection limits for some elements by platform and probe atomisation are given in Table 3. The detection system was an echelle spectrometer equipped with the quartz plate back- ground correction system.For the medium volatile elements ~~ Table 2. Atomiser studies to improve excitation 1. Tube designs for wall atomisation 2. Platform atomisation 3. Probe atomisation 4. Constant temperature 2-phase atomiser 5 . FANES atomiserANALYTICAL PROCEEDINGS, JULY 1988, VOL 25 3.71 3.36 3.01 2.66 2.31 ,- 1.96 N 0 219 ~ -- ~ - - - - - Table 3.ETA-AES detection limits* (ng ml- I ) Wavelength1 Echelle spectrometer system Element nm Platform Probe Cd 326.11 50 3 Cr 425.43 0.023 0.012 c u 324.75 0.034 0.01 Mn 403.08 0.029 0.008 Pb 405.78 3.4 Zn 307.59 - 0.45 - 23.5 * 50 pl injection volumes. i- Not detected at 10 pg ml- I . (Cr, Cu and Mn) probe atomisation gave only a minor advantage over the platform. However, for the volatile elements (Cd, Pb and Zn), probe atomisation yielded signifi- cantly better detection limits. The information presented in Table 1 shows that the use of platform or probe atomisation gave a general improvement in detection limit values when operated in conjunction with automatic background correction by wavelength modulation. Under these conditions, two thirds of the 22 elements under consideration gave detection limits less than 1 ng ml-1.At the present time, ETA-AES detection limits for 42 elements have been published; 19 elements have detection limits in the range <0.01-1 ng ml-1, 19 detection limits are in the range 1-100 ng ml-1 and four values are >lo0 ng ml-1. Recent Developments Further improvements in thermal excitation have been realised by the development of a two-phase atomiser which completely separates the processes of vaporisation and excitation in ETA-AES.3” However, it is the thermal nature of excitation in a graphite tube atomiser which provides a fundamental limitation to the improvement in detection limits that can be achieved by atomiser design. Detection limits at the ng ml-1 level are only likely to be achieved for elements with an excitation potential less than 4.5 eV.In order to achieve useful detection limits for elements with higher excitation potentials, some form of non-thermal excitation is required. In this respect 0.55 t 0.20 0.15 0.50 0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.90 1.00 1.10 Time x lO-”s Fig. 2. Furnace atomic non-thermal excitation spectrometry signal for chlorine: He discharge at 25 torr, 60 mA; ClI837.6 nm line used; 20 p1 of 20 pg ml-’ C1 as KCl. Peak height, 351.5 counts; peak area, 55.2 counts s-*; peak position, 2.86s; half width, 0.278s; background height, 1683.7 counts; background area, 11 085.2 counts 5-1; background position. 2.84 s Table 4. Comparison of detection limits for FANES and ETA-AES obtained using an echelle spectrometer and wavelength modulation Detection Limit1pg Wavelength1 Element nm FANES* Reference ETA-AES Reference 320.1 4 2.6 14 249.8 80 NMi 447.8 23 000 NP$ 228.8 326.1 150 28 439.0 6800 NP 479.5 5100 NP 425.4 0.6 28 302.1 0.7 26 372.0 9 403.3 2 26 206.2 4800 NP 285.2 3 15 26 214.9 210 NP 253.6 280 NP 283.3 405.8 23 28 196.0 800 34 NP VU) 437.9 184 100 13 213.9 4 34 1200 28 307.6 AgV) B(I) Br(I) W I ) CJ(I) (11) CrO) 357.9 Feu) Ga(1) 287.4 0.6 I(I) M d I ) 313.3 32 250 13 Mg(I) P(I) (1) P W ) s e u > Zn(I> 1 34 4 12 34 * FANES detection limits without reference are new work. i NM = not measured.: NP = not possible to measure.220 ANALYTICAL PROCEEDINGS, JULY 1988, VOL 25 atomiser by omitting the evacuation - discharge stage prior to atomisation. With this device therefore, it is possible to select the most appropriate mode of operation in order to achieve the best excitation conditions for any particular analyte or set of analyte elements.Concluding Comments Although the research in ETA-AES initiated by John Ottaway in 1974 was more a venture of rediscovery than invention, much more has been achieved than could have been predicted following the initial “preliminary experiments” and applica- tions. The current status of the technique owes much to the development of platform and probe atomisation and the provision of an efficient means of automatic background correction by wavelength modulation. The advent of the FANES source has clearly opened a new chapter to the ETA-AES story and, if commercial development of the technique is to take place, it will be in the form of a dual-excitation source that can be operated with or without auxiliary excitation.Research and development in ETA-AES and FANES continues at the University of Strathclyde so there is every reason to expect that there will be many more publications from John’s group on atomic emission techniques with a graphite furnace. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. References King, A. S . , Astrophys. J., 1905, 21, 236. King, A. S . , Astrophys. J., 1908, 27, 353. Liveing and Dewar, Proc. R. Soc., 1882, 34, 119. Ottaway, J . M., and Shaw, F., Analyst, 1975, 100, 438. Adams, M. J . , and Kirkbright, G. F., Anal. Chim. Acta, 1976, 84, 79. Sturgeon, R. E., and Chakrabarti, C. L., Spectrochim. Acta, Part B , 1977, 32, 231.Littlejohn, D., and Ottaway, J . M., Analyst, 1978, 103, 595. Littlejohn, D., and Ottaway, J . M., Analyst, 1979, 104, 208. Littlejohn, D., and Ottaway, J . M., Anal. Chim. Acta, 1978, 98, 279. Van den Broek, W. M. G. T., de Galan, L., Matousek, J . P., and Czobik, E. J . , Anal. Chim. Acta, 1978, 100, 121. Littlejohn, D., and Carroll, J., Spectrochim. Acta, submitted for publication. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21 1 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32, 33. 34. Epstein, M. S . , Rains, T. C., and O’Haver, T. C., Appl. Spectrosc., 1976, 30, 324. Baxter, D. C., Duncan, I. S . , Littlejohn, D., Marshall, J . , and Ottaway, J . M., J. Anal. At. Spectrom., 1986, 1, 29. Ottaway, J . M., Bezur, L . , and Marshall, J . , Analyst, 1980, 105, 1130.Bezur, L., Marshall, J . , and Ottaway, J . M., Spectrochim. Acta, Part B , 1984, 39, 787. Littlejohn, D., and Ottaway, J. M., Analyst, 1977, 102, 553. Littlejohn, D., Hutton, R . C., and Ottaway, J. M., “Proceed- ings of Symposium on Electrothermal Atomisation,” 20th CSI, Chlum a Trebone, Czechoslovakia, 1977, p. 191. Ottaway, J. M., Bezur, L., Fakhrul-Aldeen, R., Frech, W., and Marshall, J., “Proceedings of 1980 International Work- shop on Trace Element Analytical Chemistry in Medicine and Biology,” Neuherberg, Munich, FRG, 1980, p. 575. Ottaway, J . M., in Facchetti, S . , Editor, “Analytical Tech- niques for Heavy Metals in Biological Fluids,” Amsterdam, 1982, p. 171. Marshall, J . , and Ottaway, J. M., Talanta, 1983, 30, 571. Harnly, J . M., Patterson, K. Y., Veillon, C., Wolf, W.R . , Marshall, J . , Littlejohn, D., Ottaway, J . M., Miller-Ihli, N. J., and O’Haver, T. C., Anal. Chem., 1983, 55, 1417. Frech, W., Ottaway, J . M., Bezur, L., and Marshall, J . , Can. J. Spectrosc., 1985, 30, 7. Marshall, J., Carroll, J . , Littlejohn, D., Ottaway, J . M., O’Haver, T. C . , and Harnly, J . M., Anal. Proc., 1985, 22,67. O’Haver, T. C . , Harnly, J. M., Marshall, J . , Carroll, J . , Littlejohn, D., and Ottaway, J . M., Analyst, 1985, 110, 451. Littlejohn, D., and Ottaway, J . M., Analyst, 1979, 104, 1138. Bezur, L., Marshall, J . , Ottaway, J . M., and Fakhrul-Aldeen, R., Analyst, 1983, 108, 553. Marshall, J . , Bezur, L., Fakrul-Aldeen, R., and Ottaway, J . M., Anal. Proc., 1981, 18, 10. Giri, S. K., Littlejohn, D., and Ottaway, J.M., Analyst, 1982, 107, 1095. Marshall, J., Giri, S. K., Littlejohn, D . , and Ottaway, J. M., Anal. Chim. Acta, 1983, 147, 173. Baxter, D. C . , Frech, W., and Lundberg, E., Analyst, 1985, 110, 475. Falk, H., Hoffmann, E., and Ludke, C., Spectrochim. Acta, Part B , 1981, 36, 767. Falk, H., Hoffman, E., and Ludke, C., FreJenius Z. Anal. Chem., 1981, 307, 362. Littlejohn, D., Carroll, J . , Quinn, A. M., Ottaway, J . M.. and Falk, H . , Fresenius Z . And. Chem., 1986, 323, 762. Falk, H., Hoffmann, E., Ludke, C., Ottaway, J. M., and Giri, S. K., Analyst, 1983, 108, 1459. Evaluation of a Mathematical Model for Peak Interpretation in Graphite Furnace Atomic Absorption Spectrometry Based on Free Analyte Atom Redeposition on Carbon Surfaces B. Welz, B. Radziuk and G.Schlemmer Department of Applied Research, Bodenseewerk Perkin-Elmer & Co. GmbH, 0-7770 Uberlingen, FRG The processes occurring in an electrothermal atomiser are complex and dependent on many factors. This is reflected by the great variety of absorbance peak shapes recorded during atomisation. In the years since the introduction of the technique by L’vov,~ various mathematical approaches to the physical interpretation of peak forms have been taken. These range from the simple kinetic model of Fuller2 to the Monte Carlo techniques of Black et a1.3 Although it has been shown4 that the shape of absorption signals is different, particularly as regards the degree of tailing, for different tube materials such as polycrystalline electro- graphite, pyrolytic graphite and glassy carbon, and that this is not due to differences in the heating characteristics of the atomisers, there have been few attempts to account for this factor. Smetss mentioned the possibility of atom redeposition on the graphite surface, and Holcombe and co-workers6.7 suggested that a series of adsorption - revaporisation reactions may take place along the tube during the atomisation step, and that the form of the signal profile may therefore be largely determined by the degree of “stickiness” of the analyte atoms to the tube material.Musil and RubeSka8 extended the models of Fuller2 and van den Broek and de Galan9 to include rate constants for redeposition and re-evaporation and were able to solve the resulting system of equations to arrive at an analytical expression for the number of atoms in the gas phase as a function of time.In this work, the applicability of such a model to the interpretation of the decay of signals generated in a tube-type graphite furnace under a wide range of conditions wasANALYTICAL PROCEEDINGS, JULY 1988, VOL 25 22 1 investigated by fitting computer-generated curves to experimental data. A numeric method was used, making available information about the distribution of atoms in the gas phase and on the surface under the conditions assumed for this model and permitting the relatively simple extension of the model to account for both temporal and axial non-isothermal conditions in the furnace. Plots of atom density versus time were generated, using detailed temperature data characteristic of the tube materials. The elements Ag, Co, Pd, Ir, V and Cr, representing extremes of volatility and reactivity with carbon, were studied using combinations of tubes and platforms made from pyrolytic graphite (PG), glassy carbon (GC), polycrystalline electro- graphite (EG) and electrographite with a pyrolytic graphite coating (PGC).Experimental Instrumentation and Reagents The analytical data were generated using a Perkin-Elmer HGA-500 graphite furnace equipped with an AS-40 auto- sampler and installed in a Perkin-Elmer Model 5000 atomic absorption spectrometer. Temperature measurements were made using an Ircon Modline I1 infrared pyrometer with P3 optics. Data were collected and manipulated using an Epson PC+ microcomputer. Computation An extended model was developed in order to overcome the main drawback of the model of Musil and RubeSka, that is the assumption of isothermal conditions.The tube was divided into ten axial segments, five on each side of the tube centre, for each of which dynamic temperature measurements were made. Thus, the assumption is reduced to uniform temperature for the length of a segment, rather than constant and uniform temperature over the whole tube. Starting with an atom population located at a single point at the centre of the tube wall or platform, numeric expressions accounting for the processes depicted in Fig. 1 were used for I I - LX -4- Lx- 1 4 1 I I t i t i I A I I I I I I I I I t I .. t I t I TGx(f), Twx(t) 1 TGX+ 1 (d, Twx+ I ( t ) , Tpx+ 1 Fig. 1. Schematic representation of the re-deposition, re-vaporisa- tion and diffusion processes for two segments, one including the end of a platform. Here x is the segment number (in the figure x = 3) and L , T , D and k are the appropriately subscripted segment lengths, temperatures, diffusion coefficients and “rate constants”; N is the number of atoms in gas phase stepwise calculation of the population distribution among the original sample, gas and redeposited phases.The calculation was made for atoms transported by diffusion to each tube segment. The absorbance versus time curve was generated by summing the gas phase population of all the segments for each interval. Results and Discussion Differences in rates of decay could, to some extent, be quantified using the simple model of Musil and RubeSka, which assumes isothermal conditions. A reasonable consistency in the calculated degree of reactivity, as represented by a redeposition coefficient, could be obtained for the materials and elements tested.However, assumptions made are not rigorous. The results obtained were dependent on the starting point of the fit. Further, there were instances in which no acceptable fit could be obtained. Hence the model in this form is of very limited use. In the application of the extended model an attempt was made to treat the parameters in a physically realistic fashion. The numerical approach made it possible to include the actual temperature at the actual point in the tube in each individual calculation and thus to account for a temperature dependence of the reaction rate through an activation energy term. Gas-phase diffusion was assumed to be the only factor responsible for axial transport of atoms and the “stickiness” of a material for a given analyte was expressed in terms of the energy required for revaporisation of atoms from the surface. 0 2 4 Time/s Fig.2. Absorbance profiles for chromium, calculated using the extended model (dotted lines), fitted to measured peaks (solid lines). (a) PGC tube, PG platform, 2300°C; ( b ) GC tube, PG platform, 2300 “C In Fig, 2, numerically generated curves are compared with the absorbance profiles for chromium initially atomised from a PG platform in tubes made of PGC and GC, the two most different materials available. The same activation energy for reatomisation from the platform was used in each instance and only the value for the tube wall was changed.As might be ‘w-1 I 0 2 4 Ti meis Fig. 3. Calculated distribution of atoms as a function of time in a PGC tube with atomisation from a PG platform. NR - 1 and NRw-5, atoms redeposited on the wall in segments 1 (tube endyand 5 (tube centre); N,, atoms in the gas phase222 ANALYTICAL PROCEEDINGS, JULY 1988, VOL 25 expected, the value of E2 obtained for the “stickier” material, GC, was higher. The numeric nature of the curve generation process made available data on the atom distributions between the gas phase and the surface in each tube segment at each calculation interval. Hence the contributions of atoms from various parts of the tube to the analytical signal could be evaluated.For example, for chromium atomised from a PG platform in a PGC tube (Fig. 3), it can be seen that the number of atoms deposited on the wall in the outermost segment is still increasing even after chromium has essentially been removed from the central segment. These atoms will, for the most part, not be re-atomised under steady-state conditions, but rather con- tribute to a memory effect in a subsequent atomisation as the ends of the tube are then, as has also been reported by Falk et af. , l o initially heated to a higher temperature. Conclusions The decay of the absorbance peak, in particular for an element such as chromium, which interacts extensively with graphite, cannot be described in terms of simple diffusion of atoms out of the tube. Curves generated using the model of Musil and RubeSka can much better describe the latter portions of absorbance profiles than is possible without accounting for interactions between atoms and the atomiser surface and the difference in reactivity between PG, EG and GC can therefore be expressed quantita- tively.However, as the processes involved in both atom formation and surface interaction are not independent of temperature, the above model describes the leading edge of absorbance signals not at all and the trailing edge only approximately. A better description of the whole event can be obtained if the temperature dependence of the reaction rates is included in the model. The “stickiness” of a material for a given analyte atom may be expressed in terms of an activation energy for revaporisa- tion.This energy is in general different from that calculated for initial atomisation of the sample. Much more experimental work is required to substantiate the applicability of this approach and to investigate the degree to which simplifying assumptions are justified. References 1. L’vov, B. V., Eng. Phys. J . USSR, 1959, 2(2), 44. 2. Fuller, C. W., Analyst, 1974, 99, 739. 3. Black, S. S . , Riddle, M. R., and Holcombe, J. A., Appl. Spectrosc., 1986, 40, 925. 4. Schlemmer, G., and Welz, B., Fresenius Z. Anal. Chem., 1986, 323, 703. 5. Smets, B., Spectrochim. Acta, Part B , 1980, 35, 33. 6. Holcombe, J. A., and Rayson, G. D., Prog. Anal. A t . Spectrosc., 1983, 6 , 225. 7. Holcombe, J. A., Rayson, G. D., and Akerland, N., Spectro- chim.Acta, Part B , 1982, 37, 319. 8. Musil, J., and RubeSka, I., Analyst, 1982, 107, 588. 9. van den Broek, W. M. G. T., and de Galan, L., Anal. Chem., 1977,49, 2176. 10. Falk, H., Glismann, A., Bergann, L., Minkwitz, G., Schubert, M., and Skole, J., Spectrochim. Acta, Part B , 1985, 40, 533. Graphite Furnace Atomic Absorption Spectrometry on the Way to Absolute Analysis Boris V. L’vov Department of Analytical Chemistry, Polytechnical Institute, Leningrad 195251, USSR In trying to form an over-all picture of the recent decade in the development of graphite furnace atomic absorption spec- trometry (GFAAS), one can only be amazed at the tremen- dous changes that have occurred in the instrumentation, methodology and analytical characteristics of the method. For example, the injection and analysis of liquid samples have been completely automated.As a result, the random error of determination had dropped from 3-8% to 1-2%. The sensitiv- ity of determination for many elements has improved by a factor of up to 10 through the use of pyrolytically coated tubes and fast furnace heating. 1.2 The analytically useful lifetime of graphite tubes has increased from 50-200 to 500-1000 cycles.3 The use of improved background correction procedures, in particular the Zeeman effect, has permitted accurate correc- tion of up to 1.7-2 absorbance. In addition, it may be that a universal matrix modifier, palladium, has now been found.4.5 Perhaps the most significant achievement has been the commercial development of the platform atomisation tech- nique.There is no doubt that methodology based on the “stabilized temperature platform furnace” (STPF) concept has made it more likely that sample solutions can be analysed with calibration provided by analyte reference solutions rather than standard samples. The STPF concept has been subjected to detailed investiga- tion at the Department of Analytical Chemistry, Leningrad Polytechnical Institute. The most significant results obtained recently are summarised below. 1. The presence of gaseous carbon at enhanced concentrations during the heating of a graphite furnace has been experimentally verified.6 This finding has been used to interpret some unusual effects of furnace material, sample vaporisation techniques and sheath gas on the analytical signal. 2. The molecular spectra in the range 200-300 nm observed during the vaporisation of Al, Ga, In and TI in graphite furnaces have been shown to originate from A12C2, GaC2, InC2 and In2C2, T1C2 and T12C2 molecules.7 3.The effect on the absorption signal of sample redistribution along the tube axis during thermal pre-treatment has been studied and the need to carry out atomisation under gas-stop conditions has been validated.8 4. Characteristic masses for 40 elements have been theoretic- ally calculated and compared with experimental values .9 The corresponding coefficients of atomic diffusion in argon for 60 elements have been calculated.10 5. A method has been developed to determine alkaline earth and rare-earth elements using a modified version of the STPF procedure.11 Owing to the use of a Ta platform and to lining the graphite tubes with Ta foil, the number of elements which can be determined under STPF conditions has risen above 50.11 This seemingly slight modification of the furnace has made it possible to determine all alkaline earth and rare-earth elements using STPF conditions, with a sensitivity similar to that for other elements (Table 1).The furnace exhibits almost no memory effects and there is little tailing of signals for the elements most difficult to atomise, e.g., Ce, La, Pr and Y (Fig. 1). 6. An automated technique for curvature correction of calib- ration graphs has been developed. 12 Information on the shape of two calibration pulses, corresponding to two analyte masses, is recorded by the computer. The analyteANALYTICAL PROCEEDINGS, JULY 1988.VOL 25 223 Table 1. Measurement conditions and sensitivity of determination for alkaline earth and rare earth elements: STPE conditions with Ta platform and Ta-lined tube Atomisation Characteristic massipg Wavelength/ temperature/ Element nm "C Hold time/s Peak height Peak area Ba . . . . . . Ce . . . . . . Dy . . . . . . Er . . . . . . E u . . . . . . G d . . . . . . H o . . . . . . L a . . . . . . Nd . . . . . . Pr . . . . . . s c . . . . . . S m . . . . . . S r . . . . . . Tb . . . . . . Tm . . . . Y . . . . . . Yb . . . . . . 553.6 567.0 421.2 400.8 459.4 407.9 410.4 550.1 492.5 494.0* 391.2 429.7 460.7 432.6 371.8 410.2 398.8 2550 2500 2550 2550 2400 2600 2600 2400 2550 2550 2550 2500 2500 2600 2550 2660 2400 4 3 4 4 4 5 4 4 4 3 4 4 3 4 4 5 4 2.5 2.8 6.6 1.7 7.0 1000 200 250 46 330 3.6 30 0.3 80 1.1 25 0.75 4.3 6.5 4.2 3200 17 210 18 550 140 1300 100 90 38 8.7 0.8 3.1 1.7 * Superposition of two lines at 493.974 and 494.030 nm with lower levels at 2874 and 1377 cm-1 and intensity ratio 1 : 2.mass for the smaller signal is chosen such that the peak height lies within the linear part of the calibration graph. The larger signal is selected to cover the whole range of absorbance values up to the absorbance plateau or the roll-over level. Care must be taken to ensure that the two signals are of similar shape. By comparing the continuous values of A = f(t) and A2 = f(t) for the two signals, one can readily construct a normalised calibration graph, A = f(A,), where A. is the absorbance assumed to exist with ideal linear calibration.One has to find for it the non-linear regression coefficients, a and b, for the general algorithm which approximates this curve. In an analysis, the coeffi- cients a and b are used in the computer to correct the Ao = A/(1 + UA + bA2) 0.5 0 0.5 0 T 0.5 0 1 0 ( b ) I I I 0 2 4 Timeis Fig. 1. Absorbance signals for (a) 90 ng of Ce, ( b ) 20 ng of La, (c) 50 ng of Pr and ( d ) 5 ng of Y, obtained under measurement conditions specified in Table 1 running values of A to obtain A. and subsequently to calculate in the usual way the integrated absorbance, SA,(t)dt, and to determine the analyte mass (m) by the equation rn = rn0~Ao(t)dt/0.0044. These practical achievements have been accompanied by some progress in the study of the physics and chemistry of the processes occurring in graphite atomisers. At the same time, it should be admitted that the level of our theoretical knowledge is inadequate for the logical solution of many practical problems. Moreover, sometimes whatever knowledge we have hinders rather than promotes the development of the tech- nique.How else could one explain the belated solution of the problem of alkaline earth and rare-earth element determina- tion with the STPF procedure or the slow research on the optimum matrix modifier. Therefore, the expansion of fundamental research into the process of atomisation is now more urgent than ever before. Of the various methods employed at present in such experimental studies, the most informative, in our opinion, are scanning electron microscopy (to investigate the furnace material3 and the physical state of sampless), quadrupole mass spec- trometryl3.14 and optical molecular spectroscopy.6.' These techniques offer a unique possibility of determining the gas-phase composition, including unstable molecules (C,, MC,, CN, MCN) directly in the course of sample atomisation.One of the primary methodological goals on the way to absolute analysis is to reveal and remove the major causes of calibration instability and the difference between the charac- teristic masses obtained even for instruments of the same type. This variability can reach a factor of two and may originate from (i) differences in the measurement conditions chosen by the analyst (lamp type and current, spectral slit width, atomisation temperature and integration time), (ii) possible uncontrollable differences in the instrument performance (tube and platform geometry, quality of pyrolytic graphite coating, auto-sampler pipette calibration, Zeeman-effect cor- rector adjustment and magnetic field strength, calibration of readings in absorbance) or (iii) external factors (natural variations of atmospheric pressure, instability of dilute refer- ence solutions, changes of analyte concentrations in the sampling cup because of water evaporation).If one takes into account and controls properly all these factors, one will undoubtedly be able to reduce the variability of instrument calibration down to the level of random errors. Naturally, the impact of these factors on calibration and the possibilities of controlling them are essentially different.The224 ANALYTICAL PROCEEDINGS, JULY 1988, VOL 25 largest difficulties are apparently met at present in attempts to monitor the quality of pyrocoatings during experiments and the performance of the Zeeman-effect background corrector. 15 In this connection it would possibly be preferable to compare the characteristic masses measured with new tubes and different instruments, but without the Zeeman-effect corrector. An additional advantage of using the characteristic masses measured without the Zeeman-effect corrector lies in the possibility of comparing them with the results of theoretical calculations.9 When one considers the various advances achieved in graphite furnace AAS, and the potential of the method is taken into account, one is directed to the conclusion that the road to absolute analysis does not appear to be as long as it seemed a decade ago.However, additional research of fundamental processes that occur in the graphite furnace is still required. The author expresses his sincere thanks to Dr. David Little- john for his help in the preparation of this paper. References 1. “Analytical Methods Using the HGA Graphite Furnace, Part No. 009358,” Perkin-Elmer, Uberlingen, 1977. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. “Techniques in Graphite Furnace Atomic Absorption Spec- trometry, Part No. 6993-8150,” Perkin-Elmer, Ridgefield, CT, 1985. Welz, B., Schlemmer, G., and Ortner, H. M., Spectrochim. Acta, Part B, 1986, 41, 567. Schlemmer, G., and Welz, B., Spectrochim.Acta, Part B , 1986, 41, 1157. Voth-Beach, L. M., and Shrader. D. E., J. Anal. At. Spectrom., 1987, 2, 45. L’vov, B. V . , J . Anal. A t . Spectrom., 1987, 2, 95. L’vov, B. V., Norman, E. A . , and Polzik, L. K . , Zh. Prikl. Spectrosk., 1987, 47, 711. L’vov, B. V., Polzik, L. K., and Yatsenko, L. F . , Talanta, 1987, 34, 141. L’vov, B. V., Nikolaev, V. G., Norman, E. A . , Polzik, L. K., and Mojica, M., Spectrochim. Acta, Part B , 1986, 41, 1043. L’vov, B. V . , Nikolaev, V. G . , Zh. Prikl. Spektrosk., 1987,46, 7. L’vov, B. V., Nikolaev, V. G . , and Norman, E. A , , Z h . Anal. Khim., 1988, 43, 46. Bayunov, P. A . , and L’vov, B. V., Zh. Anal. Khim., 1987, 42, 621. Styris, D. L., Anal. Chem., 1984, 56, 1070. Styris, D. L., Fresenius Z .Anal. Chem., 1986, 323, 710. Slavin, W., Carnrick, G . R., and Manning, D. C., Pittsburgh Conference on Analytical Chemistry and Applied Spectro- scopy, 9-13 March, 1987, paper No. 271. Comparison of Bead-making Techniques for X-ray Fluorescence Spectrometry Richard Steventon and Andrew Cunningham British Steel Corporation, Strip Products Group, Ravenscraig Works, Motherwell, Lanarkshire ML 1 1 S W The fusion of a sample with a flux such as lithium tetraborate to produce a glass bead is a well established technique to reduce secondary absorption and particle size effects in the analysis of material by X-ray fluorescence spectrometry (XRFS). The technique itself suffers from difficulties caused by, for example, inhomogeneity of the mixture, insolubility of the sample material in the flux and recrystallisation of the glass melt on cooling.These effects can be minimised by careful control of the condition of fusion and of casting, therefore, chemists have attempted to automate the process in order to standardise these conditions. Over the years several solutions to this problem have been designed and marketed in the form of different instruments and the aim of this work was to compare the techniques and apparatus now available for the production of glass beads in XRFS. The following different methods studied were chosen to cover the various improvements realised over the years: (i) muffle furnace technique; (ii) Claisse stirrer; (iii) Perl’x 2; (iv) Labor-Schoeps AAG 50B; and (v) Leco FX 100 fluxer. For each technique ten beads were prepared using samples of iron ore sinter which were analysed by XRFS using our standard conditions for iron ore analysis.In each test all the beads were analysed consecutively and a randomly selected bead from each series was analysed ten times. From these results were calculated (i) the over-all standard deviation of the instrument plus sample preparation (So) and (ii) the standard deviation of the instrument (SI). In order to make meaningful comparisons between the techniques, the standard deviation due to sample preparation (S,) was calculated using the expression Table 1. Comparison of relative standard deviations (RSD) Iron ore sinter Method Fe Si02 A1203 TiOz CaO MgO P Mn Perl’x2 . . . . . . 0.099 0.48 4.6 - 0.62 1.47 0.62 - Schoeps . . . .0.23 0.52 5.52 3.28 0.95 1.00 0.94 0.4 FX 100 . . . . . . 0.26 0.14 0.80 - 0.35 2.65 - - Claisse . . . , . . 0.24 1.09 2.91 - 0.36 1.92 - Furnace . . . . 0.59 1.90 5.25 1.24 1.29 - 0.77 Approximate composition, Yo . . 58 6 2 0.2 9.6 2.5 0.07 0.11 - - Slag CaO MgO SO2 A1203 Ti02 MnO Fe Perl’x2 . . . . . . 0.34 0.22 0.16 1.93 1.21 0.34 1.84 Schoeps . . . . 0.72 - 0.21 2.33 0.58 2.23 FX 100 . . . . . . 0.47 - 0.32 1.75 2.31 0.25 0.45 Approximate composition, ‘/o . . 0.5 12.0 34.0 9.2 0.4 0.8 0.7 -ANALYTICAL PROCEEDINGS. JULY 1988, VOL 25 225 Table 2. Advantages and disadvantages of the devices Method Advantages Perl'x 2 . . . . . . Very safe; furnace is enclosed. Idiot-proof; very comprehensive safety circuits prohibit malpractice. Easy temperature regulation; manual shows graph of induction current vs.temperature. Very reproducible. Pre-oxidation cycle is available. Self-diagnostic microprocessor control. AAG50B . FX 100 Fuses more than one sample at a time. Pre-oxidation cycle is available. Very reproducible. . Fuses more than one sample at a time. Pre-oxidation cycle is available. Facility for casting into solution. Excellent crucible swirling mechanism. Excellent crucible - casting dish - beaker Self-diagnostic microprocessor control. Very reproducible. retaining arrangement. Claisse stirrer . . . . Very reproducible. More than one fusion at a time. Muffle furnace . . . . Simple and robust. More than one fusion at a time. Disadvantages Only fuses one sample at a time. No facility for casting into solution. * Poor crucible-retaining mechanism.No facility for casting into solution. * Fusion temperature is more difficult to Possible to cast flux in absence of casting Requires modification to take standard Fusion temperature is more difficult to Gas-flow regulators are very coarse. regulate using flames. dish. size casting dishes. regulate using flames. Successful operation is very much Crucible - casting dish retainers are very Dangerous casting technique. Fusion temperature is more difficult to No forced air cooling. Labour intensive. No flux agitation. Least reproducible samples. Slow. No forced air cooling. operator-dependent . poor. regulate using flames. * The manufacturers of both the Perl'x 2 and AAG 50B market devices capable of casting either into casting dishes or into solution.S, gives a more realistic comparison because the contributions from instrumental variations, e.g., tube power, detector power supply and sample cups, have been removed and the results are therefore more representative of the errors arising from the bead making procedure. In each instance the beads were prepared by fusing 1 g of sample with 9 g of lithium tetraborate at 1200°C and casting into a 40 mm diameter platinum -gold casting dish. The bottom surface of the bead was used for the analysis. The exercise was repeated using a blast furnace slag sample for the Labor- Schoeps, Leco Fluxer and Perl'x instruments. Fe S 0 0 d m LL L I D 41 Ti CaMg PMn Fe Si A Schoeps u Ti CaMg PMn t I 1 n %4LcL.l FeSi AlTiCaMgPMn Fig. 1. sinter Relative standard deviation profile for each method.Iron ore Table 1 shows the relative standard deviations (RSD) obtained from this work. Fig. 1 shows the profiles obtained for each method. The RSD is defined as one standard deviation expressed as a percentage of the mean concentration in the sample. Methods Method (i): Muffle Furnace In the furnace technique the flux and sample mixture, which has been thoroughly mixed, is transferred into a 70-ml capacity platinum - gold crucible and placed in a muffle furnace at 1200°C for 15 min. At the end of this time the mixture is swirled and cast into a platinum - gold alloy casting dish which has been pre-heated to a cherry red heat. The bead is then allowed to cool to room temperature for 13 min. Method (ii): Claisse Stirrer The Claisse stirrer was designed to swirl the melt continuously whilst the fusion process was carried out. It consists of six gas burners arranged in a circle around a central tube, which is held near its base by a universal joint.A motor at the base causes the assembly to tilt about 20" from the vertical and rotate while in operation, thus swirling the melt during fusion. Above each burner is a set of two clips which hold a crucible and casting dish, which allows the casting dish to be heated during the fusion of a sample. These clips are connected by a gear mechanism to a nut on top of the central tube. On rotating this nut, the casting dish passes underneath the crucible, the crucible tilts and the melt is cast into the casting dish within the flame of the burner. The rotation of this nut causes all six226 crucibles and the casting dish to rotate simultaneously. When the beads have cooled to room temperature the can be analysed by XRF.Method (iii): Perl’x 2 The Perl’x 2 is a microprocessor-controlled furnace which contains a radiofrequency induction generator to heat the fusion crucible and casting dish, The crucible and casting dish are enclosed by a hinged furnace cover to protect personnel from electrical and thermal risks during operation. The microprocessor allows six different melting programs to be stored. Manual controls are used to pre-set the operating parameters and initiate operation. The melt is agitated during the fusion process and the bed is cast automatically and cooled by a jet of air. Method (iv): Labor-Schoeps AAG 50B The Schoeps AAG 50B has two pairs of oxygen - propane burners, each pair heating a crucible and a casting dish.The various steps in the fusion method are controlled by electrical timers. The crucibles are held in position above the burners by two bars which are operated pneumatically and the casting dish is held by two retractable arms. The melt is agitated during fusion and when the fusion is complete the mixture is cast into a pre-heated casting dish which is then cooled by an air jet. Method (v): Leco FX 100 Fluxer The FX 100 is a microprocessor-controlled unit which uses three oxygen - propane burners for fusion. The instrument can ANALYTICAL PROCEEDINGS, JULY 1988, VOL 25 either cast into platinum moulds to make beads or cast into stirred acid solutions for ICP, AAS or other methods of chemical analysis.At the conclusion of the fluxing procedure, the device casts the molten flux into the casting dish and the bead is cooled by a jet of air. For solution preparation, the casting dishes are simply removed and beakers containing an appropriate acid solution are placed on the stirring motor base. In common with the Perl’x and Schoeps, the FX 100 can be programmed as follows: (a) low or high heat settings with or without agitation (for pre-oxidation, etc.); (b) casting time; (c) forced air cooling time for beads; (d) stirring time for solution; and (e) number of burners used. Advantages and disadvantages of the different devices are given in Table 2. Conclusion The results that we obtained showed that the furnace method gives results which, in general, are less reproducible than the automatic methods by factors of between one and four.This is probably due to the lack of agitation during fusion. The automatic methods give better results with all elements and very good results, approaching classical values, for some elements. No one instrument seems to be better overall than any other therefore other considerations such as cost, ease of use and the particular analytical application should be borne in mind when deciding which technique to use. Novel Instrumentation for Inductively Coupled Plasma Atomic Emission Spectrometry S. P. Corr, D. H. Hall and D. Littlejohn Department of Pure and Applied Chemistry, University of Strathclyde, Cathedral Street, Glasgow G 7 7x1 and C.V. Perkins Philips Scientific, York Street, Cambridge CB7 2PX Although the inductively coupled plasma (ICP) is now extensively used in atomic emission spectrometry (AES), its high purchase and operational costs have to some extent limited its use to higher budget laboratories. To reduce these instrument costs and expand the use of the ICP, alternative generator designs have been suggested and lower running costs have been achieved by using N2,1.2 air3.4 or water”6 coolants to reduce argon gas consumption. This paper describes the characteristics of a micro-torch ICP designed to operate at 80 MHz with reduced power input and lower argon flow-rates compared with conventional macro- torch operation. For the evaluation, the micro-ICP was used with a novel echelle spectrometer that has some advantages over other spectrometers of equivalent focal length.Micro-torch ICP and High-frequency RF Generators Torch miniaturisation has been proposed as a means of reducing the argon consumption and power requirements of an ICP. A micro-ICP, if operated at the same power density as its larger counterpart, should, in theory, exhibit a similar analy- tical performance. Allemand et al.7 obtained a reduction in the coolant argon flow-rate from 15 to 8.5 1 min-1 for a 9 mm i.d. torch at power levels of 700-1000 W. Weiss et a1.8 also developed a 9 mm i.d. torch which could sustain a discharge at 500 W and 7 1 min-1 of argon. Although this discharge demonstrated excellent sensitivity and calibration linearity, it proved to be more susceptible to chemical interference effects than the standard 18 mm i.d.torch. The reason for this inferior performance was believed to arise from the interaction that occurred between the sample aerosol and the radiofrequency (RF) coupling directed into the small plasma. In all RF discharges, power is introduced near the boundary (or skin) of the plasma with RF coupling decreasing exponentially towards the plasma centre. This feature is characterised by the “skin depth,” which by definition, is the distance from the plasma boundary where the power has fallen to l/e of its surface value. Weiss et a1.* recognised that it was necessary to reduce the power dissipated into the central channel in order to minimise the influence of the analyte on the power density in the plasma.Table 1. Operating conditions for the 50- and 80-MHz ICPs 50MHz 80MHz Power inputikW . . . , . . . . . . 1 .O 0.85 Intermediate argon flow-ratell min-I . . 0.2 0.13 Carrier argon flow-ratell min-1 . . . . 0.65 0.65 Observationheightimm . . . . . . 15 16 Outer argon flow-rateil min-1 . . . . 15 10 They concluded that an increase in frequency beyond the standard 27.12 MHz was required, as an increase in frequency results in a decrease in the skin depth and would therefore reduce power dissipation into the central channel. An increase in frequency should allow easier sample introduction using aANALYTICAL PROCEEDINGS, JULY 1988, VOL 25 227 lower carrier gas velocity at the same flow-rate. Previous work9 has shown that a lower carrier gas velocity aids the suppression of chemical interferences in the ICP.Although few comparative studies on different operating frequencies exist, there is evidence to suggest that such an increase in operating frequency may also be beneficial in improving signal to background ratios (SBR).IO A more recent study11 on a 148-MHz ICP suggests that an upper frequency limit might exist beyond which a deterioration in analytical performance is observed. In this work, a study was made of a micro-ICP system operating at 80 MHz. The micro-ICP was operated using a 9 mm i.d. demountable quartz torch. Spectrometer In ICP-AES the use of a spectrometer of adequate resolving power is of critical importance. Spectrometer performance can affect the speed, accuracy and precision of the analytical measurements made.A major requirement for any spec- trometer used in ICP-AES is that it should offer a high enough resolution to allow the isolation of analyte lines of interest from inherently line-rich emission spectra. Many publications have demonstrated that spectral interferences can be significant in ICP-AES owing to the broadening and richness of the spectra.12 In addition to high resolution, the design of any spectrometer for ICP-AES must also include several other desirable features. In the presence of high concentrations of an intense emission element, e.g., Ca, stray light may be observed. This results in an increase in the measured back- ground radiation at analyte wavelengths, and if uncorrected can introduce large errors in the measurement of trace metal concentrations. The spectrometer must therefore be designed to contribute low stray light levels.The light throughput efficiency of the spectrometer is important from a consider- ation of noise, particularly at low wavelengths where detection limits are limited by background shot noise. Other factors, such as spectral range, stability, background correction and speed of analysis, must also be considered to achieve a compromise in cost, size and performance. Most conventional ICP systems provide reasonably high resolution spectrometers but normally the highest resolution can only be achieved by some compromise in performance or design. For example, a convenient way to increase resolution in a conventional grating spectrometer is to use narrower slits. However, this places a limitation on light throughput and wavelength stability if the slits are reduced below a certain minimum.Table 2. Spectral characteristics of mini-Cchelle and SOPRA echelle spectrometers Mini-Cchelle SOPRA echelle Linear Linear Wavelength/ Resolving dispersion1 Resolving dispersion/ nm power nrnmm-1 power nmrnm-1 200 710 000 0.1446 1800000 0.031 300 480000 0.2131 1 200 000 0.042 400 354 000 0.2892 910000 0.062 500 278 000 0.3680 720000 0.085 An alternative to the normal diffraction grating, which does not require a long focal length spectrometer, a finely ruled grating, or the use of very narrow slits to achieve high resolution is the echelle grating. 13.14 The main difference between an echelle and a normal diffraction grating is that the echelle is illuminated at high angles of incidence (ca.63") on the short side of the grooves. Its groove spacing is also much coarser and typically 79 or 316 lines mm-1 is the number employed. As a result of this geometry, the diffracted light is produced over a narrow angle as a larger number of spectral orders, each spectral order overlapping the adjacent order. Owing to this order overlap, the echelle grating cannot be used on its own but requires a secondary dispersive system to separate "stacked" orders. A prototype mini-echelle spectrometer, which combines the benefits of high performance, compact design and low cost, has been constructed at Pye Unicam, Cambridge. The instrument is a sequential scanning spectrometer and is operated under software control. In this study.the performance characteristics of the mini-kchelle spectrometer were evaluated in operation with 50 MHz macro-ICP and 80 MHz micro-ICP torches. Experimental The prototype mini-echelle spectrometer was initially evalu- ated with the Philips 50-MHz PV8490 ICP. This evaluation also contributed to a comparison of the analytical characteristics of a conventional 50-MHz ICP (18 mm i.d. torch) and an 80-MHz ICP (9 mm i.d. torch). A schematic diagram of the micro-torch is shown in Fig. 1. The torch was a demountable type with the three quartz tubes held in an accurately machined brass block (provided by Prof. Dr. P. W. J. M. Boumans, Philips Research Laboratories. Eindhoven). Intermediate tube (7.5 rnm ad.)' (6.5 mrn i.d.1 ( 6 mm 0.d.) (4 mm i.d.1 \ / ,P,uzlt"o:.r (9 mm i.d.) / Inner tube (2.7 mm 0.d.) (1.1 mm i.d.) Coolant (plasma) g a s inlet Intermediate (auxiliary) gas inlet \ Machined brass block Sample (carrier) aerosol inlet Fig.1. Demountable micro-torch The performance of the mini-echelle spectrometer was compared with that of the SOPRA echelle as reported by Boumans and Vrakking.15 Features of special importance for both monochromators are presented in Table 2. Results and Discussion Background Equivalent Concentration Table 3 presents the background equivalent concentrations (BEC) obtained for the 50- and 80-MHz ICPs at several wavelengths. BEC values are a measure of the analyte sensitivity for a specific wavelength and allow a fairly rapid characterisation of the instrument response without the need to perform any statistical evaluations, as is normally required for detection limit comparisons.BEC values are dependent o n several instrumental parameters such as sample uptake rate.228 ANALYTICAL PROCEEDINGS, JULY 1988, VOL 25 argon flow-rate, input power and observation height. The greater the sensitivity of the ICP, the lower is the BEC value. Table 3. Comparison of aqueous solution background equivalent concentrations (BEC) for the 50-MHz (macro) and 80-MHz (micro) ICPs with the mini-echelle spectrometer BEC/ng ml- Spectral lineinm B I 208.459 G e I 209.426 B I 249.678 B I 249.773 G a I 294.364 PbII 220.3.53 In11 230.606 MnII 257.610 V I I 309.311 Y I I 371.030 BaII 493.409 50 MHz 106 490 149 64 610 625 1121 10 45 20 19 80 MHz 48 33 44 24 161 78 1 869 15 1.53 4.8 1 .5 Considering the BECs obtained for the two plasmas, it is apparent that for the atom lines considered, the high-frequency micro-ICP demonstrated increased sensitivity.For the ion lines the position is less clear but generally the micro-ICP exhibits a similar performance to the conventionally sized torch. Previ- ous comparative studies on low- and high-frequency ICP generators have indicated that as the frequency was increased the excitation temperature and electron density number decreased.10.16 As a consequence of this, decreases in the line and background continuum intensities were obtained. Capelle et al.1" reported that in their study of generators operating between 5 and 56 MHz, the intensity of the continuum decreased more rapidly with increasing frequency than the net analyte intensity.This resulted in an increase in the signal to background ratios. In contrast, however, a more recent study comparing the performance of a 148-MHz ICP and a 27.12- MIlz ICP11 did not lead to a similar conclusion and no improvement in SBR was obtained at the higher frequency. As 148 MHz is higher than any other operating frequency previously considered, it may be that an upper frequency limit exists, above which the analytical performance of the ICP deteriorates. With respect to the BEC values obtained in this study for the 50- and 80-MHz ICPs, the results generally agree with the observations made by Capelle et al.10 No measurements of excitation temperature or electron number density have been made as yet with the 80-MHz micro-ICP, but the improvement in sensitivity obtained for the atom lines, compared with the 50-MHz results, suggests that the 80-MHz micro-torch plasma is cooler than its larger counterpart.If this is correct the degree of ionisation in the 80-MHz plasma should be lower than for the 50-MHz ICP, which is confirmed by the slightly poorer sensitivity exhibited by some ion lines in the higher frequency plasma. Interferences Previous studies7.8 on micro-torch systems operated at 27.12 MHz have indicated that chemical interferences are worse than in conventional plasmas. To study the effect of operating the micro-torch at a higher frequency, the response of several atom and ion lines in the presence of 1000 pg ml-1 sodium and 1000 pg ml-1 phosphate was observed. It was found that the level of interference from both sodium and phosphate remained low, typically in the range 04% for the elements considered.This suggests that operating the ICP at 80 MHz improves the coupling efficiency between the oscillator and the plasma when using a micro-torch and therefore allows easier sample introduction and so reduces the interaction between the sample aerosol and the RF coupling. Detection Limits Table 4 lists the aqueous detection limits obtained for several elements using the 50- and 80-MHz ICPs and mini-echelle spectrometer. The detection limits found for the micro-ICP compare favourably with those of the 50-MHz macro-ICP. However, unlike the 80-MHz BEC values, no obvious improvements in atom line detection limits were found compared with the 50-MHz ICP and the magnitude of any improvement found for BEC values (Table 3) was not reflected in the measured detection limits.This may be related to the greater instability of the micro-ICP, which exhibited a slightly larger flicker noise component than the macro-ICP. Table 4. Comparison of aqueous solution detection limits for the 50-MHz (macro) and 80-MHz (micro) ICPs with the mini-kchelle spectrometer Detection limiting ml- 1 Spectral lineinm B I 208.959 G e I 209.426 B I 249.678 B I 249.773 G a I 294.364 Pb I1 220.353 In11 230.606 C o I I 238.892 MnII 257.610 V I I 309.311 Y I1 371.030 B a I I 493.409 50 MHz 1.8 7.6 2.5 1.1 10.3 11 17 1.2 0.1.5 0.85 0.28 0.47 80 MHz 2.3 7.5 1.6 0.9 3.5 29 28 4.2 0.4 3.7 0.25 0.13 A comparison was made of the aqueous detection limits found with the mini-echelle and those obtained by Boumans and Vrakkingls using the 1.5 m focal length SOPRA echelle spectrometer with pre-disperser. Both systems employed the same 50-MHz ICP and had accurately derived background relative standard deviations for the detection limit analysis. Detection limits can be used as a guide to the overall instrument performance as they are dependent not only on the RF generator and plasma variables but also on the light throughput, stability and resolution offered by the spec- trometer. The detection limits for both kchelle systems (Table 5) did not differ by more than a factor of 2-3, which is the uncertainty factor in detection limit determinations.The limits Table 5. Comparison of aqueous solution detection limits for the SO-MHz ICP with SOPRA and mini-ichelle spectrometers Detection limiting ml-1 Spectral lineinm As I1 193.696 As I 197.197 MoII 202.030 G e I 209.426 PbII 220.353 In11 230.606 CoII 238.892 R h I I 249.077 MnII 257.610 MgII 279.5.53 C a I I 294.364 V I I 309.311 Ti11 334.941 Y I1 371.030 NdII 401.225 SmII 442.434 B a I I 493.409 ~ SOPRA kchellel5 11 21 11 11 4.0 7.7 1 .0 8.8 0.08 0.01 4.4 0.34 0.17 0.07 1.3 2.1 0.09 ~ ~~ Mini-kchelle 8.2 1.4 7.6 18 11 17 11 1.2 0.15 0.01 5.9 0.55 0.30 0.21 2.6 3.5 0.17ANALYTICAL PROCEEDINGS. JULY 1988, VOL 25 229 of detection achieved by the mini-echelle are extremely good owing to its excellent light throughput and relative freedom from shot noise.The flicker noise contribution to the relative standard deviation for the background signal was about 0.6%.Conclusions A 9 mm i.d. ICP torch has been successfully operated at a generator frequency of 80 MHz. Detection limits and the level of interference compare favourably with those obtained for an 18 mm i.d. torch. The micro-plasma was sustained at 820 W and with a total argon flow-rate of 11 1 min-1. Further improvements in torch and coil design will be necessary to permit the use of lower power and argon gas flow-rates. The noise characteristics and light throughput of the mini-echelle spectrometer are excellent. However, one pos- sible design improvement would the the availability of a variable slit width which could be of advantage in the analysis of complex sample matrices that give line-rich spectra. The authors gratefully acknowledge the provision of a SERC CASE grant with Tioxide International (D.H.H.) and a postgraduate studentship from Pye Unicam Ltd.(S.P.C.) and thank Dr. W. F. Knippenberg and Prof. Dr. P. W. J. M. Boumans of Philips Research Laboratories, Eindhoven, for the loan of the 80-MHz ICP. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. References Fassel, V. A , , Appl. Spectrosc., 1981, 35, 292. Montaser, A . , and Mortazavi, J., Anal. Chem., 1980,52, 255. Ripson, P. A. M., de Galan, L., and de Ruiter, J . W., Spectrochim. Acta, Part B , 1982, 37, 733. Van der Plas, P. S. C., de Waaij, A. C . , and de Galan, L., Spectrochim. Acta, Part B , 1985. 40, 1457. Kawaguchi, H . , Ito, T., Rubi, S . , and Mizuikc, A . , Anal. Chem., 1980, 52, 2440.Kornblum, G . R . , Van der Waal, W., and de Galan, L., Anal. Chem., 1979, 51. 2378. Allemand, G. R., Van der Waal, W., and de Galan, L., Anal. Chem., 1979, 51,2392. Weiss, A. D., Savage, R. N., and Heiftje, G . M., Anal. Chim. Acta, 1981, 124, 245. Kornblum, G. R., and de Galan, L., Spectrochim. Acta, Part B , 1977, 32, 455. Capelle, B., Mermet, J.-M., and Robin, J . , Appl. Spectrosc., 1986, 36, 102. Webb, B. D., and Denton, M. F., Spectrochim. Acta, Part B , 1986, 41, 361. Dahlquist, R. L., and Knoll, J. W., Appl. Spectrosc., 1978, 32, 1. Richardson, D . , Spectrochim. Acta, 1953, 6, 61. Keliher, P. N., and Wohlers, C. C., Anal. Chem. 1976, 48, 333A. Boumans, P. W. J . M., and Vrakking, J. J. A. M., Spectro- chim. Acta, Part B , 1987, 42, 553. Gunter, W. H., Visser, K., and Zeeman, P.B., Spectrochim. Acta, Part B , 1983, 38, 949. Determination of Cobalt in Plasma and Urine by Electrothermal Atomisation Atomic Absorption Spectrometry Using Palladium Matrix Modification* Barry Sampson Department of Chemical Pathology, Charing Cross Hospital, Fulham Palace Road, London W6 8RF Although cobalt has been used for the treatment of anaemia in ureaemic patients,’ its use has been associated with toxic side effects, including fatal cardiomyopathy,2.3 and consequently its use has now largely been discontinued. Marginally increased Co concentrations have been reported in the blood and plasma of uraemic patients who have not received Co supplement^.^,^.^ Increased concentrations have also been measured in tissues from uraemic subjects.6 High Co concentrations in the plasma of patients with chronic renal failure (CRF) have been related to a decrease in cardiac function.’ The ranges of whole blood and plasma Co concentrations reported range from less than 0.1 to 1000 pg 1-1; the currently accepted consensus is that normal values are less than 1 pg 1- l .These low concentrations present a difficult analytical prob- lem, being close to the detection limit of graphite furnace techniques. Previously, direct assays have used ammonium phosphate, magnesium nitrate or nitric acid as the matrix modifier.8.9 This paper examines an alternative strategy for matrix modification of Co in a direct assay. Palladium has recently been suggested as approaching the utility of a “universal” matrix modifier for atomic spectrometry.1 c 1 2 The biological applications pub- lished so far include assays for Te,l3 Se,10.12 As14 and Au.15 It has also been reported that Pd will stabilise Co to an ashing temperature of 1200 *C,lz whereas the maximum temperature that can be attained using ammonium phosphate as modifier is only 1000°C. * Presented at the 2nd Nordic Symposium on Trace Elements in Human Health and Disease, Odense, Denmark, 17-21 August, 1987. Experimental All experiments were performed with a Perkin-Elmer 3030 atomic absorption spectrometer with D2 background correc- tion. An HGA-500 furnace and AS-40 autosampler were used. Pyrolytic coated graphite tubes were used throughout. Argon was the purge gas, compressed air was the alternate gas. A commercial Co standard solution (1000 mg 1-1) was used to prepare a stock standard. Working standards were prepared from this solution as required.The ammonium phosphate matrix modifier was prepared from Specpure grade NH4H2P04 (Johnson Matthey Chemicals, Royston, UK). Palladium solutions were prepared from PdCI2 (Johnson Matthey Chemicals). Triton X-100 (BDH scintillation grade) and Dow-Corning antifoam, DBllOA, were also used in the preparation of matrix modifiers as described below. For method development studies, freeze-dried bovine serum was used to avoid undue exposure to potentially infectious human samples. In later experiments this material was used, spiked with increasing amounts of Co, as an internal control. The endogenous concentration of Co found in these specimens was 0.5-2.0 pg 1-1.The final method tested involves the addition of 25 p1 of modifier to 400 pl of sample in a polystyrene test-tube while mixing the whole on a “whirlimixer.” The modifier solution contains 9 g 1-1 of Pd, 1% V/V Triton X-100 and 1% V/V antifoam. The unknown samples are then analysed using the spiked samples as standards. The concentration of Co in the spiked serum is first determined by standard additions. The furnace programme used is shown in Table 1.230 ANALYTICAL PROCEEDINGS, JULY 1988. VOL 25 Results and Discussion Matrix modification and furnace conditions were initially investigated using aqueous standards and later with spiked urine and plasma samples. The maximum temperature for a plasma matrix was found to be 1400°C, held for 5-10 s.In practice it was found that a lower background absorbance was obtained using a temperature of 1250-13OO0C, held for 15-20 s. Under these conditions, the background absorbance from 20 p1 of urine is less than 0.1. Table 1. HGA-500 programme for the determination of cobalt in plasma Ramp Hold Step Temperaturei’C timeis time/s Internal gas” 1 90 5 10 2 120 60 10 3 600 30 30 Air (300 ml min- 1 ) 4 600 1 10 5 1300 20 10 6: 2500 0$ 2 Argon(0mlmin 1 ) 7 2650 1 5 * Argon (300 ml min-1) unless stated otherwise. i. Read command. $ Maximum power heating. The sensitivity required of the assay is very high if it is to measure Co concentrations at normal levels, so it is essential to use the largest possible sample volume. The maximum sample volume that could be used was found to be 40-60 1-11.Above this volume the background absorption shows a pronounced increase. The sample volume used for all results reported here was 40 pl. The detection limit of the assay was calculated to be 0.1-0.15 pg 1-1. A satisfactory recovery was obtained up to a concentration of Co of 10 pg 1-1. Within- and between-batch precision was assessed by the assay of two control samples, one with an “endogenous” concentration and the other spiked to a higher concentration. At the normal level of Co, a precision of greater than 15% is acceptable, whereas at elevated concentrations the precision attained is nearer to 5%. For studies of normal Co concentra- tions, however, the extra sensitivity given by the solvent extraction methods may be preferable. A factor in the poor precision found at low concentrations may be the error involved in assigning an endogenous concentration to the control serum samples used as standards. There are few reference materials available for Co.The urine Seronorm 108 (Nycomed, Norway) has a provisional recommended Co concentration of 11 pg 1-1,16 with a range of 9-14 pg 1-1. The concentration of Co found by this method was The very low concentrations of Co in blood make it essential to take stringent precautions to minimise contamination at all stages. The Co contamination found in various items used to collect routine blood samples was therefore compared. The amount of additional Co was low in all instances-approxi- 9 pg 1-1. mately the same as the expected normal concentration of Co in plasma. In view of this it was decided that the samples collected without any special precautions being taken would be ade- quate, as we only needed to identify raiskd concentrations of Co. All results on control and patient samples must therefore be interpreted as containing an obligatory blank of up to 0.3 pg 1-1, which cannot be accurately subtracted or reduced further. The normal range for Co found to be 0.35 k 0.15 pg 1-1 (n = 35). Samples were taken from 29 CRF patients on maintenance haemodialysis (HD); the mean concentration of Co found was 1.46 k 0.57 pg 1-1. Only one patient had a plasma Co concentration within the range found in the normal subjects (<0.65 pg 1-1); the maximum concentration found was 2.85 yg 1-1. We have been unable to demonstrate any relationship between the plasma Co concentration and cardiac function in these patients. 17 The use of Pd as a matrix modifier for the assay of Co raises the prospect that in addition to its reported use in the assay of the more volatile elements, discussed earlier, Pd may also find application in the thermal stabilisation of other first row transition elements. 1. 2. 3. 4. 5 . 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. References Lancet, 1976, 11, 26. Curtis, J. R., Goode, C. C., Herrington, J., and Urdanetta, L. E., Clin. Hephrol., 1976, 5 , 61. Pehrsson, K., and Lins, L. E . , Lancet, 1978, 11, 51. Astrug, A., Kuleva, V., Kuleff, I., Kiriakov, Z . . Tomov. A , , and Djingova, R., Trace Elem. Med., 1984, 1, 65. Lins, L. E., and Pehrsson, S. K., Trace Elem. Med., 1984, 1, 172. Pehrsson, S . K., and Lins, L. E., Nephron, 1983, 34, 93. Clynne, N., Lins, L. E . , and Pehrsson, S . K., Trace Elem. Med., 1984, 2, 44. Angerer, J., and Heinrich, R., Fresenius Z . Anal. Chem., 1984, 318, 37. Delves, H . T., Mensikov, R., and Hinks, L., in Bratter. P., and Schramel, P., Editors, “Trace Element Analytical Chemistry in Medicine and Biology,” Volume 2, de Gruyter, Berlin and New York, 1983, p. 1123. Schlemmer, G., and Welz, B., Spectrochim. Acta, Part B , 1986,41, 1157. Voth-Beach. L. M., and Schrader, D. E., J . A n d . At. Spectrom., 1987, 2, 45. Voth-Beach, L. M., and Schrader. D . E . , Spectroscopy, 1986, 1, 49. Weibust, G . , Langmyhr, F. J . , and Thomassen, Y., Anal. Chim. Acta, 1981, 128, 23. Xiao-Quan, S., Zhe-Ming, N., and Li, Z.. Anal. Chim. Acta, 1983, 151, 179. Xiao-Quan, S . , Egila, J . , Littlejohn, D . , and Ottaway, J . M.. J . Anal. At. Spectrom., 1987, 2, 299. Molin Christensen, J., Inhat, M., Stoeppler, M., Thomassen, Y., Veillon, C., and Wolynetz, M., Fresenius Z. Anal. Chem., 1987, 326, 639. Maher, E. R., Sampson, B., and Curtis, J. R., Clin. Sci., 1987, 73, Suppl. 17, 42P.

ISSN:0144-557X    

DOI:10.1039/AP9882500217

出版商:RSC    

年代:1988

数据来源: RSC

摘要:

230 ANALYTICAL PROCEEDINGS, JULY 1988. VOL 25 Laser-excited Atomic and Molecular Fluorescence-Preliminary Investigations for Thallium L. M. Garden and D. Littlejohn Department of Pure and Applied Chemistry, University of Strathclyde, Glasgow G 1 IXL K. Dittrich and H.-J. Stark Sektion Chemie, Karl-Marx-Universitat, Talstr. 35, DDR-70 10 Leipzig, GDR Laser-excited atomic fluorescence spectrometry (LAFS) with electrothermal atomisation has been shown to be a very sensitive technique for the determination of trace metals, especially when vaporisation in a graphite tube is used to produce atoms prior to excitation by a dye laser.l.2 This mode of atom production is preferred as higher vapour phase atomANALYTICAL PROCEEDINGS. JULY 1988. VOL 25 23 1 concentrations can be obtained compared with filament or rod atomisation.When the graphite tube conditions are chosen to promote the formation of molecules of the analyte, it is possible to carry out measurements by laser-excited molecular fluorescence spectrometry (LAMOFS). The molecular flu- orescence of molecules of indium and aluminium has been investigated and described in a recent publication .2 In this work the investigation of LAFS and LAMOFS has been extended to the determination of thallium. Of particular interest were the interference effects, caused by halide salts and acids, on the atomic fluorescence of thallium, and the optimisation of conditions for the production of thallium halide molecular fluorescence. The instrumentation used for the LAFS and LAMOFS measurements has been described previously.l.2 Details regarding the laser - spectrometer system used are given in Table 1. The electrothermal atomiser was similar in dimensions to a Perkin-Elmer HGA 500, but had slots cut in the side of the tube to allow passage of the laser radiation. 1 Table 1. Instrumental parameters Component Parameter Nitrogen laser . . Laser supply voltage Power Pulse duration Pulse frequency Pulse energy Operating pressure Wavelength Range of dye Frequency doubling crystal Optimum working range Entrance slit height Entrance slit width Dyelaser . . . . Dye max. Monochromator Prism Parameter value 15 kV 250 k W 4 ns 6 Hz 2 mJ 7.9 kPa 337.13 nm CZ462 525-594 nm 560 nm NaH2P04 NaCl 210-360 nm 2 mm 3 mm Thallium Atomic Fluorescence It was considered important to evaluate the interference caused by various concentrations of hydrohalic acids on the thallium direct line fluorescence at 352.943 nm.The wavelength of the dye laser beam was set at 553.574 nm and excitation of fluorescence was carried out at the 276.787-nm wavelength using the frequency-doubled dye laser beam. The mass of thallium injected for all experiments was 2 pg per 10 pl of solution. Initially, concentrations of hydrochloric, hydro- bromic, hydriodic and hydrofluoric acids of up to 1 M were added to the thallium solutions and the LAFS signals recorded. In a second series of measurements, the thallium solutions contained 1 M nitric acid and various concentrations of each hydrohalic acid up to 1 M. The atomisation programme used to generate thallium atoms was: a drying stage at 400 "C for 15 s, ashing at 600 "C for 15 s and atomisation at 1500 "C for 5 s.The results obtained when various concentrations of hydrochloric. hydrobromic, hydriodic and hydrofluoric acids were added to the thallium solutions are given in Table 2 in the form of percentage recovery of signal. In general, addition of an,acid to the thallium solutions caused a depression of the thallium atomic fluorescence signal, the effect becoming more pronoun- ced at higher concentrations of the acid. This trend may be attributed to the formation of thallium halide molecules and also to the quenching effect of residual levels of hydrohalide molecules remaining in the graphite tube during atomisation. At increased concentrations of acid, the effect of both these factors would become more pronounced, thus decreasing the fluorescence intensity of thallium. The trend observed for the effect of hydrofluoric, hydro- chloric, hydrobromic and hydriodic acids was not as expected.Thallium fluorescence was found to be suppressed most by the presence of hydrobromic and hydriodic acids and least by hydrofluoric acid. At any concentration of acid, the effect on the thallium fluorescence intensity was in the order HBr = HI > HCl > HF. The dissociation energies of thallium fluoride. chloride, bromide and iodide are 4.77, 3.82, 3.37 and 2.95 eV. respectively. From these dissociation energies, it was expected that the effect of the acids on the thallium intensity would have been the reverse of that actually observed.Thallium fluoride has the greatest dissociation energy and so if the dominant factor during atomisation was the formation of thallium halide molecules, then the thallium atomic fluorescence intensity should have been decreased most in the presence of hydro- fluoric acid and least in the presence of hydriodic acid. Table 2. Effect of HF, HCI, HBr and HI on recovery of thallium LAFS signals. Thallium concentration. 2 pg per 10 $-injected volume Recovery of thallium intensity. '% Concentrationh HF HC'I HB r HI 0 100 100 100 100 0.01 34 13 17 0.1 109 32 0 0 1 .o 61 9 - - - The most likely reason for the observed results is the quenching effect of residuai volumes of acid vapours that remained in the tube during the atomisation stage. All the dilute solutions of the hydrohalic acids have (constant) boiling- points of less than 130°C and the boiling-point tends to decrease as water is evaporated from the solution.3 It is likely, therefore, that most of the acid molecules would have been removed from the atomiser during the drying stage. However.it is possible that some of the acid could have impregnated the graphite tube to form salts with trace metal components of the graphite, or could exist as trapped acid molecules. The amount of halide remaining in the tube would probably increase at higher acid concentrations. As the size of the halide ions increases in the order iodide > bromide > chloride > fluoride. it would be expected that the quenching effects of the iodide and bromide ions would be greatest and so cause a more severe reduction in the thallium atomic fluoresence intensity.27 1.40 271.60 27 1.80 Wavelength nm Fig. 1. fluoride at 271.59 nm obtained by LAMOFS Point-by-point molecular fluorescence spectrum for thallium When 1 M nitric acid was added to the thallium - hydrohalic acid solutions, a slightly greater suppression of the thallium atomic fluorescence signal was observed compared with measurements obtained for solutions to which no nitric acid had been added. This may be attributed to the increased quenching effects, caused by the presence of NO, species during the atomisation stage. In general, however, no substan-232 tial differences to the trends given in Table 2 were observed. Hydriodic and hydrobromic acids caused the greatest suppres- sion of the thallium signal and hydrofluoric acid the least.Thallium Molecular Fluorescence An attempt was made to observe thallium molecular fluorescence for TlBr at 291.0 nm, TIC1 at 281.9 nm and T1F at 271.37, 271.59 and 272.49 nm. The atomiser programme described previously was used, which included an atomisation temperature of 1500 “C. No molecular fluorescence was observed for TlBr or TlC1. When the molar ratios of T1: X (X = halide) used in the experiments were considered, it was found that the ratios were 1 :0.26 and 1:0.58 for bromide and chloride, respectively. In order to ensure complete production of thallium halide molecules, an excess of the halide should be present in the solutions injected into the atomiser. It is likely, therefore, that insufficient thallium bromide or chloride molecules were produced to give satisfactory sensitivity for molecular fluorescence measurements.Resonance fluorescence was observed for thallium flu- oride; in this instance the molar ratio of thallium to fluoride in the injected solution was 1 : 2.15, i . e . , an excess of fluoride was present. The molecular fluorescence signals obtained for TlF at 271.59 nm are shown in Fig. 1. Similar spectra were obtained for T1F fluorescence at 271.37 and 272.49 nm. In addition to the presence of an excess of fluoride in the thallium solution, it is ANALYTICAL PROCEEDINGS, JULY 1988, VOL 25 prob-- 5le that measurement of the T1F fluorescence signals was enh .ed by the higher dissociation energy of thallium fluoride (4.7. V) compared with those of the other thallium halide mole, x.Conclusions The interference effects caused by hydrofluoric, hydrochloric, hydrobromic and hydriodic acids on the atomic fluorescence of thallium suggest that the dominant factor is quenching of the excited atoms by molecular collisions, rather than a vapour phase chemical effect. In order to achieve molecular flu- orescence of a thallium halide, it is necessary to have a molar excess of the halide salt in the thallium solution (as demon- strated in this work for thallium fluoride) and to operate the atomiser at a temperature that minimises thallium atom formation. References 1. 2. 3. Dittrich. K., and Stark, H.-J., J. Anal. A t . Spectrom., 1986, 1 , 237. Dittrich, K., and Stark, H.-J., 1. A n d A t . Spectrom., 1987, 2, 63. Bock, R., “A Handbook of Decomposition Methods in Ana- lytical Chemistry,” International Textbook Co., Edinburgh, 1979, p.56. Faster Analysis of Biological Samples by Electrothermal Atomisation Atomic Absorption Spectrometry D. J. Halls Trace Metals Unit, Department of Biochemistry, Glasgow Royal Infirmary, Glasgow G4 OSF The success of electrothermal atomisation atomic absorption spectrometry (ETA-AAS) in determining trace elements in biological specimens has led to an increased demand for analyses using this method, and this has highlighted the slowness of the technique. A programme of research, carried out to evaluate the changes that can be made to furnace programmes in ETA-AAS, has resulted in savings in analysis time. Principles Applied ramp rates are rarely necessary in drying stages as the furnace itself requires a finite time (about 10 s) to reach its final temperature.Hence the furnace provides its own ramping. For simple matrices (e.g., waters, digested samples), the time for the drying steps can be reduced to about 8 s for uncoated tubes1 and about 16 s for pyrolytically coated graphite surfaces (wall or platform atomisation).2 For serum and urine, dilution with a 0.25% V/V Triton X-100 solution helps to equalise the spreading of standards and samples and allows faster drying. In some determinations, the ashing stage plays no important role and can be omitted (Table 1). For methods using a L'vov platform, the commonly used cooling stage seems unnecessary and can also be omitted.2 ~~ ~ Table 1. Determinations in which an ashing stage was found to be unnecessary Cu in urine' Cd in urine4 Pb in bloodl Cr in urine7 Cr in plant digests2 Cu in plant digests2 Pb in teeth digests' (after deproteinisation) By using these principles, the programme time can be reduced to about 30 s which, with an autosampling time of 30 s, gives cycle times of about 60 s.Examples of Applications Reported applications of these principles include the determi- nation of aluminium in water1 and dialysate fluids,3 cadmium in bloodl and urine,4 copper in urine1 and plant digests,2 chromium in plant digests2 and lead in blood,' plant digests2 and teeth digests.5 Tables 2 and 3 show furnace programmes, Table 2. Furnace programme for the determination of Cu in serum. Samples and standards are diluted 1 + 10 with 0.25% ViVTriton X-100 and 0.01 M nitric acid and then determined off the wall of an uncoated graphite tube.Total cycle time, 58 s Temperature/ Ramp time/ Hold time/ Step "C S S 1. Dry . . . . 110 1 8 2. Ash . . . , 900 2 5 3. Atomise . . 2700 2 10* 4. Clean . . . . 2700 1 2 * Internal gas flow-rate, 10 ml min-l; -2 s record and -2 s autozero. which have been developed for the determination of copper in serum, based on the method described by Evenson and Warren,6 and for chromium in urine, intended for use in monitoring occupational exposure. Both programmes were designed for use with a Perkin-Elmer 2280 spectrometer in conjunction with an HGA 500 furnace, peak height quantifica- tion being obtained from the recorder output. The programme for copper in serum, shown in Table 2, contains the conven- tional four stages, but these have been reduced to a minimumANALYTICAL PROCEEDINGS, JULY 1988, VOL 25 233 Table 3.Furnace programme for the determination of chromium in urine for assessment of occupational exposure. Samples and standards are diluted 1 + 1 with 0.25% V/VTriton X-100 and 1% V/V nitric acid and then determined off the wall of an uncoated graphite tube. Total cycle time, 48 s Temperature/ Ramp time/ Hold time/ Step "C S S 1. Dry . . . . 140 1 7 2. Atomise . . 2400 0 5" 3. Clean . . . . 2700 1 4 * Internal gas flow-rate, SO ml min-I; -2 s record and -2 s autozero. Conclusions By looking critically at the function and timing of the steps in a furnace programme, major economies in analysis time can be made.In many of the determinations studied the ashing stage played no important part and a re-evaluation of the role of this stage seems necessary. Programme times can be reduced to about 30 s, but the long times (30 s or more) taken by present furnace autosamplers make further reduction of the pro- gramme time pointless; instead, faster methods of sample introduction are required. that is consistent with good reproducibility and accuracy. Because of the organic content of serum, the ashing stage was not omitted. Urine contains little organic material, however. and for the determination of chromium in this fluid (Table 3), the ashing stage could be omitted with no effect on accuracy, but with an increase in background absorbance. Background absorbances were still small and within the range for which a deuterium-arc background system is capable of correcting.' In the determination of cadmium4 and copper' in urine, it was found that the omission of the ashing stage had no effect on the background absorbance.References 1. 2. 3. 4. 5 . 6 . 7. Halls, D. J., Analyst, 1984, 109, 1081. Halls, D. J . , Mohl, C., and Stoeppler, M., Analyst. 1987, 112. 185. Halls, D. J., and Fell, G. S . , Analyst, 1985, 110, 243. Halls, D. J . , Black, M. M., Fell, G. S . , and Ottaway, J. M., J . Anal. A t . Spectrom., 1987, 2, 305. Keating, A. D., Keating, J . L., Halls, D. J . , and Fell, G. S., Analyst, 1987, 112, 1381. Evenson, M. A , , and Warren, B. L., Clin. Chem., 1975,21,619. Halls, D. J., and Fell, G. S . , 1. Anal. A t . Spectrom., 1988, 3, 105.Determination of Boron in Plants by Graphite Furnace Atomic Absorption Spectrometry with Slurry Atomisation Using Matrix Modification and Totally Pyrolytic Graphite Tubes Neil W. Barnett, Les Ebdon," E. Hywel Evans and Pierrick Ollivier Department of Environmental Sciences, Plymouth Polytechnic, Drake Circus, Plymouth PL4 8AA The determination of boron by electrothermal atomisation atomic absorption spectrometry (ETA-AAS) has been ac- knowledged to be particularly difficult,l-S owing to the involatile nature of this element and its carbides, which may be formed at the graphite tube wall. Manning et al. 1 overcame this problem to some extent by adding calcium as a matrix modifier, postulating that the formation of the more volatile calcium boride species occurred preferentially to that of boron carbide. Similarly, van der Geugten2 used calcium and magnesium as modifiers, and this approach was further extended by Szydlowski3 who used the modifiers barium, calcium, chromium, magnesium and strontium.The best sensitivity was obtained using Ba(OH)2 as the modifier, even when compared with that obtained using BaC12. Although previous workers have postulated that atomisation is aided by the formation of more volatile species, it is possible that the preferential formation of carbides with the modifiers also helps, by competing with the formation of boron carbides. This paper reports the investigation of matrix modifiers :^or boron and their application to the direct analysis of plant materials using slurry atomisation and totally pyrolytic graph- ite furnace tubes.Experimental Apparatus All analyses were performed with a Pye Unicam PU SP9 atomic absorption spectrometer (Pye Unicam, Cambridge, UK) equipped with a computer-controlled graphite furnace (Pye Unicam, PU 9095). The spectrometer operating para- meters are shown in Table 1 and the furnace heating conditions used for boron determination are given in Table 2. Injections into the furnace were made manually using an adjustable micropipette. Furnace tubes were commercially available pyrolytically coated, or totally pyrolytic, graphite tubes (Pye Unicam). Table 1. Spectrometer operating parameters Wavelength . . . . . . 249.8 nm Spectralbandpass . , . . 0.2nm Lamp current . . . . . . 10.0 mA Background correction . .Continuum source Integration time . . . . . . 6.0 s Reagents and Standards All chemicals were of analytical-reagent grade (BDH, Poole, Dorset, UK). Modifier stock solutions were prepared by dissolving the appropriate amounts of their salts in dilute (2 M) nitric acid. A boron stock solution was prepared by dissolving the appropriate amount of H3B03 in de-ionised, doubly distilled water. Standard solutions were prepared by serial dilution as required. Table 2. Furnace heating conditions for boron determination Stage Temperature/"C Timeis Ramp ratePC s- l 1 160 20 100 2 200 15 5 3 1400 10 50 4 1400 10 5 3000 6 2000 6 - - - - Determination of Boron by Slurry Atomisation * To whom correspondence should be addressed. The following certified reference materials were analysed:233 ANALYTICAL PROCEEDINGS, JULY 1988.VOL 25 NBS SRM 1571 (orchard leaves) (National Bureau of Stan- dards, Washington, DC, USA), NBS SRM 1573 (tomato leaves) (National Bureau of Standards) and BCR CRM 281 (rye grass) (Bureau of Community Reference, Brussels, Belgium). Portions of the reference materials were dried and 0.5-0.7 g were accurately weighed into acid-washed crucibles. The materials were charred at 250°C for 6 h, then quantitatively transferred into acid-washed polypropylene bottles together with approximately 10 g of zirconia spheres and 10-20 ml of Ni(N03)2 (1000 pg ml-1) were accurately added. The bottles were agitated on a flask shaker for 1 h to reduce the particle size as described by Sparkes and Ebdon,6 after which the bottles were shaken vigorously before each sampling.Results and Discussion Preliminary Investigations Preliminary investigations using a 1000 ,ug ml-1 Ca - Mg mixture and pyrolytically coated tubes had shown that the boron absorbance signal was greatly increased compared with signals obtained for unmodified boron; however, a consider- able memory signal still remained. Two cleaning solutions, namely NaF (1% m/V) and CH30H - H2S04 (4 + 1 V/V), were investigated, the assumption being that they would form volatile species such as BF3 and B(CH3)3 during a separate cleaning cycle between analytical measurements. The furnace heating conditions used during the cleaning cycle are shown in Table 3. ~~~ Table 3. Furnace heating conditions for cleaning cycle Stage Temperature/"C Time/s Ramp ratePC s-1 1 200 10 50 2 1500 5 50 3 3000 5 2000 4 5 6 - - - - - - - - - Both solutions considerably reduced the memory signal, although the CH30H - H2S04 solution had a more detrimen- tal effect on tube lifetimes.Eventually, the combination of totally pyrolytic tubes, which had longer lifetimes and were less prone to memory effects (Fig. l ) , and the NaF cleaning solution was adopted for all further investigations. Fig. 1. Absorbance signal for boron with Ni modifier (10 pg) using a totally pyrolytic graphite tube. A , Boron (20 ng); and B, memory after cleaning with NaF (1 YO mW) Matrix Modification The greatest signal enhancements were achieved using the matrix modifiers Ni, Mg and La, which were therefore investigated further. The effect of varying the modifier concentration on the boron absorption is shown in Fig.2. The three modifiers behaved similarly in that a plateau was not reached until a much larger amount of modifier than of boron was present. This suggests that the atomisation of boron is aided by the preferential formation of chemical species between the modifiers and components, either on the tube wall or in the matrix, which would otherwise form less volatile species with boron. The formation of more volatile species between boron and the modifiers did not seem to occur, as any such reactions would be expected to be complete when the modifiers were within an order of magnitude of the equimolar concentrations with boron. 1.2 a, C m + g 0.8 n Q 0.4 0 0.8 W c m + g 0.4 n Q 0 Modifier, nmol I 1 I 2 50 500 750 Modifier nmol Fig.2. Absorbance signal [(a) peak area and ( b ) peak height] for boron (4.5 nmol) in the presence of varying amounts of modifiers. A, Ni; B, La; and C, Mg The assumption that modification was taking place in the condensed phase on the surface of the graphite tube wall was supported by further experiments. Aliquots of boron solution (50 ng) and modifier (10 pg) were placed either together (i.e., from a mixture) or separately (i.e., as spatially separated drops) on the tube wall, and the boron absorption measured in each instance. An absorbance signal was obtained only when modifier and boron were at the same location, thus ruling out the possibility of a gas phase reaction leading to the formation of more volatile metal borides. Finally, a qualitative scan of the surfaces of the tubes employed was undertaken using a scanning electron micro- scope (SEM).Boron could not be determined by energy dispersive X-ray fluorescence (XRF) spectrometry owing to its low relative atomic mass; however, there was no evidence of any modifier remaining on the pyrolytically coated or totally pyrolytic graphite tubes. Hence, re-addition of the modifier was necessary each time an analytical measurement was made.ANALYTICAL PROCEEDINGS, JULY 1988, VOL 25 235 Calibration and Analysis of Plant Materials Calibration was achieved using either a Mg or Ni modifier, both at a concentration of 1000 pg ml-1, mixed with the standard boron solutions. The resulting calibration graphs are shown in Fig. 3. 0.6 3.4 V c m + :: n 6 0.2 0 P ? I 1 I I 1 4 8 12 16 20 Boron ng Fig.3. Mg (1000 pg ml-1 each) modifiers Calibration graphs (peak height) for boron with A , Ni; and B, Boron was determined in various certified reference materials having both certified and indicative values for the boron concentration. The results, shown in Table 4, support the view that this rapid, sensitive method offers encouraging prospects for the determination of trace amounts of boron in environmental samples. Table 4. Results of boron determination by graphite furnace atomic absorption spectrometry with slurry atomisation. Values in paren- theses are indicative values only Boron concentrationipg g- I Reference material Certified value Found NBS 1571 orchard leaves . . 33 k 3 48 k 5 NBS 1573 tomato leaves .. (30) 42 k 2 BCR281 ryegrass . . . . (5-15) 10 k 2 References 1. 2. 3. 4. 5 . Manning, D . C., Fernandez, J . J . , and Peterson, G. E.. Third Annual FACSS Meeting, Philadelphia, PA, 1976. Van der Geugten, R. P., Fresenius Z . Anal. Chem., 1981, 306, 13. Szydlowski, F. J., Anal. Chim. Acta, 1979, 106, 121. Holding, S. T., and Rowson. J. J . , Proc. Inst. Pet., 1982, 227. Tang, L. C., and Troyer, M. A . , “Direct Determination of Boron in Archival and Library Materials by Flameless Atomic Absorption Spectrometry,” The Library of Congress, Washing- ton, DC, USA. Sparkes, S.. and Ebdon, L., Anal. Proc.. 1986, 23, 410. 6 . Furnace Atomisation for Multi-element Atomic Absorption Spectrometry James M. Hardy USDA, Beltsville Human Nutrition Research Center, Nutrient Composition Laboratory, Beltsville, MD 20705, USA The topic of this lecture is appropriate for the John Ottaway Memorial meeting because John was an enthusiastic advocate of multi-element atomic absorption spectrometry and was very interested in the development of new atomisation - excitation sources.I was fortunate to have been able to collaborate with John and his group in several areas of mutual interest. In 1977, a simultaneous multi-element atomic absorption continuum source (SIMAAC) spectrometer was developed through the collaboration of USDA and Professor Tom O'Haver at the University of Maryland.' In retrospect, SIMAAC is primarily a sophisticated detection system. As only commercial atomisers were used initially, the innovative aspects of SIMAAC lay in high speed data acquisition, high spectral resolution with wavelength modulation and the versatility afforded by a powerful, dedicated computer system.By modification of the software, SIMAAC can be used for absorption or emission measurements, with graphite furnace or flame atomisation, and for a variety of signal processing regimes (i. e., background corrected absorbance and emission intensities, extended analytical ranges, temporal background corrections and/or Fourier filtering). Following the development of SIMAAC the research emphasis shifted to the development of improved atomisation - excitation sources. Commercially available graphite furnaces are not very suitable for multi-element determinations, mainly because they offer poor compromise atomisation conditions and are subject to memory effects (material left in the furnace from a previous atomisation) at higher analytical concentra- tions.In addition, commercial furnaces exhibit high molecular (continuum) backgrounds (which produce poor signal to noise ratios) and severe gas-phase interferences. These defects in the performance of graphite atomisers can be traced to the heating characteristics, specially to the lack of spatial and temporal isothermality. The development of a furnace with spatial and temporal isothermality would provide a more ideal atomiser for both multi-element and single element atomic absorption spectrometry (AAS). The search for an improved atomiser has led to collaboration with a number of researchers. In each instance, the prototype atomisers examined offered capabilities not available with commercial systems. To evaluate these prototype atomisers, the commercially available graphite furnace with a platform, [ Perkin-Elmer (Ridgefield, CT)] and the stabilised tempera- ture platform in furnace (STPF) philosophy were used as a reference .2 l?v'ith atomisation from a platform, a conventional graphite furnace approaches temporal isothermality by retarding atomi- sation until the furnace and interior atmosphere are close to the maximum temperature.Spatially, however, a large tempera- ture gradient exists between the hottest furnace temperatures found at the centre of the tube and the much lower tempera- tures at its end.-' The cooler ends of the tube are where236 condensation of the analyte atoms and molecules occurs, giving rise to molecular absorption, gas-phase interferences and analyte memory between atomisations.Compromise condi- tions for multi-element determinations are decided by the volatility of the elements. At higher atomisation temperatures, the peak-area signals for the volatile elements decrease. The volatile elements are efficiently atomised at lower tempera- tures, but the residence time decreases as the atomisation temperature and the diffusion coefficient increase. Conse- quently, at 2700"C, the peak areas for Mo and V are still increasing, whereas those for Cd, Pb and Zn are 2-4 times smaller than at the lower, optimum atomisation temperatures. Elements of intermediate volatility, such as Co, Cr, Cu, Fe and Ni, show maximum peak areas at atomisation temperatures of about 2500 "C.In the early 1980s SIMAAC was used as a detector for multi-element furnace emission in collaboration with Professor John Ottaway and his group.4.5 As conventional furnaces were used, no change was anticipated in the isothermal characteris- tics. Platform atomisation was also advantageous for emission measurements; it prevented atomisation from occurring until the furnace was sufficiently hot to cause excitation of the analyte. However, the cooler ends of the tube were still conducive to the existence of large populations of ground-state atoms and this resulted in severe self-reversal of most resonance transitions, even at very low concentrations. Conse- quently, as had been shown previously, non-resonance lines provided the best signal to noise ratios.Only a few simul- taneous determinations were made because SIMAAC was not set-up for the non-resonance lines. SIMAAC has been used with the aerosol deposition method developed by Instrumentation Laboratory6 as a means of improving compromise analytical conditions .7 It was found that all elements gave a constant signal (peak area or height) after a threshold temperature had been exceeded. Hence, signals for the volatile elements were not compromised by the higher temperatures necessary for atomisation of the non- volatile elements. The temperature response of the elements was not due to a difference in the isothermal characteristics of the furnace, but arose because of the comparatively slow heating rate of the furnace compared with the analyte residence time.A laboratory-built tungsten wire probe, modelled after that reported by Manning et d . , 8 was employed with the aim of further improving temporal isothermality. The approach was easy to implement and the results for all but the non-volatile elements were excellent .9 At an atomisation temperature of 27OO0C, non-analytical signals were observed for Mo; it was suspected that the Mo was forming an amalgam with the tungsten wire. After 5-10 firings at this temperature, it was observed that the sharp edges of the tungsten wire had disappeared, suggesting a slight melting of the wire and further supporting the amalgamation hypothesis. The tungsten probe was, therefore, unsuitable for multi-element determinations for the elements of interest.In a second collaboration with Professor Ottaway and his group the carbon probe was examined.1" Again, improvements in the temporal isothermality were anticipated, whereas the spatial isothermality would be expected to be unchanged. The volatile elements showed the expected decrease in peak area with increasing atomisation temperature. 1 1 Elements of inter- mediate volatility, however, showed a shift of the maximum signals to lower temperatures. The optimum temperatures for Cu and Cr were 2100 and 2300 "C, respectively. The shift of the maximum signal to lower temperatures suggests a shorter residence time. It seems probable that the diffusion of the analyte out of the tube through the slot cut for the probe was significant. Gas-phase interferences from chlorides were not significantly different for probe and platform atomisation.Peak signals, however, occurred 5-10 times earlier with probe atomisation, suggesting that temporal overlap of the analyte ANALYTICAL PROCEEDINGS, JULY 1988, VOL 25 and the interferent was more significant than temporal isothermality . In collaboration with Professor James Holcombe, the second surface atomiserl2 was used with SIMAAC. Preliminary results showed that transfer of most of the volatile elements (Cd, Pb and Zn), and elements of intermediate volatility (Co, Cr, Cu, Fe and Mn), from the furnace wall to the second (tantalum) surface could be accomplished at a temperature of 2000 "C.13 Hence, with atomisation at 2700 "C, only the non-volatile elements (Mo and V) would be atomised from the carbon surface.No further results could be obtained as atmospheric oxygen penetrated the enclosure of the Perkin- Elmer furnace and routinely destroyed the tantalum surface after 4-6 atomisations. More recently, the double furnace, developed by Dr. Eric Lundberg and Dr. Wolfgang Frech, was used with SIMAAC. The upper furnace employs integrated contacts14 to provide spatial isothermality and the sample is volatilised in one of two ways: from a platform (one-step operation) or from a carbon cup with its own power supply (two-step operation). The two-step operation provides better temporal isothermality and allows more versatility in obtaining compromise atomisation conditions. For a one-step operation, and a two-step operation with a single atomisation step, the elements of intermediate volatility showed maximum peak areas at temperatures between 1900 and 2200 T .1 5 The shorter residence times were due to the use of shorter tubes (19 vs. 27 mm) and to the spatial isothermality. Using the two-step operation with a double atomisation step (ramp to 1500°C and then step to 25OO0C), nearly optimum peak areas were obtained for both the volatile (Cd, Pb and Zn) and non-volatile (A1 and Cr) elements. Further experiments showed that memory, molecular absorp- tion and gas-phase interferences were reduced using both one- and two-step operation compared with the conventional, commercial, graphite furnace. For the gas-phase interferences (CuC12 as interferent), the improvement in recovery was greatest for the volatile elements.It was also noteworthy that the additional use of the second furnace and power supply for the two-step operation, and the lack of an automatic sampler, made these determinations much less routine. In conclusion, the ideal atomiser for multi-element determi- nations should provide spatial and temporal isothermality. Separation of the volatilisation and atomisation processes, as accomplished in the two-step operation of the double furnace, provides the best compromise for multi-element operation. In addition, the ideal furnace should be sealed to the atmosphere and yet still retain the convenience of automatic sampling. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. References Harnly, J. M., O'Haver, T. C., Golden, B., and Wolf, W. R., Anal. Chem., 1979, 51, 2007. Slavin, W., Manning, D.C., and Carnrick, G. R., At. Spectrosc., 1981, 2, 137. Sturgeon, R. E., and Berman, S. S., Anal. Chem., 1983. 55, 190. Marshall, J., Littlejohn, D., Ottaway, J. M., Harnly, J. M., Miller-Ihli, N. J . , and O'Haver, T. C., Analyst, 1983, 108, 178. Marshall, J . , Littlejohn, D., Ottaway, J . M., Harnly, J. M., Miller-Ihli, N. J., and O'Haver, T. C., Spectrochim. Acta, Part B , 1984, 39, 321. Sotera, J. J., Cristiano, L. C., Conley, M. K., and Kahn, H. L., Anal. Chem., 1983, 55, 204. Harnly, J. M., J. Anal. At. Spectrom., 1986, 1 , 287. Manning, D. C., Slavin, W., and Myers, S . , Anal. Chem., 1979, 51, 2375. Harnly, J . M., unpublished results. Littlejohn, D., Marshall, J . , Carroll, J., Cormack, W., and Ottaway, J. M, Analyst, 1983, 108, 893.Carroll, J . , Miller-Ihli, N. J., Harnly, J. M., Littlejohn, D., Ottaway, J. M., and O'Haver, T. C., Anulyst,. 1985.110,1153.ANALYTICAL PROCEEDINGS, JULY 1988, VOL 25 237 12. Rettberg, T. M., and Holcombe, J. A . , Spectrochim. Acta, Part B , 1984, 39, 248. 13. Rettberg, T. M.. Holcombe, J. M., and Harnly, J. M.. unpublished results. 14. 15. Lundberg, E., Baxter, D. C., and Frech, W., J . Anal. At. Spectrorn., 1986, 1, 105. Lundberg, E., Frech, W.. and Harnly, J. M., in preparation. Application of Atomic Spectroscopy in Clinical Chemistry G. S. Fell Trace Element Unit, Department of Pathological Biochemistry, University of Glasgow, Royal Infirmary, Glasgow G4 OSF John Ottaway had a keen interest in applying advanced analytical techniques to “real samples” and, over a period of many years, a mutually beneficial friendship developed between my laboratory staff at Glasgow Royal Infirmary and the analytical chemists in Strathclyde University.The use of AAS in clinical analysis has recently been thoroughly reviewed by Delves.’ A hospital is able to provide a range of biological samples-blood, plasma, urine, tissue- which require trace element analyses for an ever increasing number of clinical applications (Table 1). The number of elements of clinical interest is considerable and their concen- trations can range from hundreds of p.p.m. to below 1 p.p.b. Analytical techniques have to combine extreme sensitivity with accuracy, yet cover a wide range of concentrations and have a reasonable rate of throughput. Table 1. Applications of AAS in clinical chemistry 1.Nutritional biochemistry 2. Inborn errors of essential element metabolism 3. Occupational and environmental toxicology 4. Medical uses of inorganic compounds as drugs 5. Toxicology of metal-contaminated pharmaceuticals 6. Disorders of trace element metabolism-secondary to organic disease, infection and trauma ~ ~~ The availability of a number of commercial quality assurance products and other biological materials which have been subjected to inter-laboratory comparisons provided a useful means for an emerging generation of Strathclyde chemists to validate new atomic spectroscopic techniques. The list of jointly supervised theses (Table 2) gives an indication of the main lines of enquiry which we have pursued. Much effort was expended in development of flame atomic fluorescence (AFS) and a prototype instrument was used to determine Cd in blood, urine and tissue.2 Although no commercial exploitation of the AFS system with automatic scatter correction was forthcoming, development of this technique continues.3 Thanks to the “Strathclyde connection” we were able to apply electrothermal atomisation (ETAAS) methods very early on to a number of clinical problems.Thus, we were able to detect high levels of aluminium in blood serum, domestic water supplies and tissue samples obtained during investigation of a fatal condition (dialysis dementia) affecting home renal dialysis patients in the West of Scotland.4.5 The speciation of essential trace elements in human blood plasma was explored by chromatographic separation and ETAAS used to detect metals in the fractions.6 This line of enquiry continues7 and is of great importance to the under- standing of trace element metabolism.The problem of interferences in ETAAS analysis has been addressed by a variation of means and was reviewed by John Ottaway.8 Trace metal contamination of pharmaceutical products is known to occur during their manufacture. We have used AAS methods to show that commercially produced human albumin and other “blood products,” which are widely used clinically, contain relatively high amounts of aluminium, chromium, manganese, iron and other metals which are potentially toxic.9 Table 2. Jointly supervised higher degrees Year Student Title 1977 Martin Hall Atomic Fluorescence Spectrometry as a Technique for the Analysis of Cadmium and Zinc in Biological Fluid Fluorescence Spectrometry to the Determination of Cadmium and Zinc in Water and Biological Materials and Delves Cup Atomic Absorption Spectrometry to Trace Metal Analysis in Biological Fluids Spectrometry Coupled with Biochemical Separation Methods to Study the Blood Transport of Essential and Toxic Metals in Human Health and Disease 1980 Martin Hall Evaluation and Application of a Purpose Built Atomic Fluorescence Spectrometer for the Analysis of Cadmium and Zinc in Biological Materials 1980 Joseph Sneddon Developments in the Application of Atomic Fluorescence Spectrometry to Clinical Analysis Fluorescence Spectrometric Methods for Clinical Analysis Essential and Therapeutic Metal Species 1982 Alistair Brown Atomic Spectrometric Determina- tion of Selenium in Biological Materials-Applications in Hospital Nutrition Fluorescence for Lead and Cadmium Analysis 1983 James Hendry Matrix Effects and the Use of Platform Atomisation in Blood Lead and Cadmium Analysis Biological Materials by Atomic Spectrometric Methods Estimation of Toxic Metal Species Generation for the Determination of Arsenic, Lead and Selenium by Atomic Absorption Spectrometry 1977 Fareed Hussein Applications of Atomic 1977 Sohrab Sabet The Application of Electrothermal 1979 Philip Gardiner Application of Atomic Absorption 1981 Edet Ekanem The Development of Atomic 1982 Jelas Haron Biochemical Separations of 1982 Prem Sthaphit Development of Atomic 1984 Douglas Baxter The Determination of Chromium in 1984 Murdo Black Biochemical Separation and 1986 Hugo Rincon Continuous Flow Hydride Degree MSc PhD PhD PhD PhD PhD PhD PhD PhD PhD PhD MSc PhD MSc238 ANALYTICAL PROCEEDINGS, JULY 1988, VOL 25 The prolonged use of plasma exchange transfusion with aluminium-contaminated albumin has been shown to cause bone disease in patients who also have impaired renal function.10 Pharmaceuticals used in large volume and given intravenously should be “screened” for the presence of unsuspected metal contamination which could be potentially toxic.This can be achieved by a simultaneous multi-element technique such as ICP - MS. Although this procedure suffers from significant analytical interferences, these can be reduced and applications to the multi-element analysis of biological samples are likely to be numerous.1’ Collaboration between the GRI Trace Element Unit and Strathclyde University will continue and provide a future generation of analytical chemists with interesting and difficult problems-as John Ottaway would surely have wished.References 1. Delves, H. T., Ann. Clin. Biochem., 1987, 24, 529. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. Scott, R., Mills, E . A., Fell, G. S . . Hussein, F. E. R., Ottaway. J. M., Fitzgerald, F., Lamont. A . , and Roxburgh. S . . Luncet, 1976, ii, 396. Ekanem, E. J., Barnard, C. L. R . , Ottaway, J . M., and Fell, G. S., Talanta, 13, 1986, 55. Elliot, H. L . , Dryburgh, F., Fell, G. S . , Sabet. S . . and Macdougall, A. I., Br. Med. J . , 1978, 1101. Gardiner, P. E., Ottaway, J. M., Fell, G.S.. and Halls, D. J . , Anal. Chim. Acta, 1981, 128, 57. Gardiner, P. E., Ottaway, J. M., Fell, G. S . , and Burns, R. R.. Anal. Chim. Acta, 1981, 124, 281. Gardiner, P. E., Top. Curr. Chem., 1987, 141, 146. Ottaway, J. M., Anal. Proc., 1984, 21, 55. Fell, G. S . , and Maharaj, D . , Lancet, 1986, ii. 467. Maharaj, D., Fell, G. S . , Boyce, B. F., Ng, J. P., Smith, G. D., Boulton-Jones, J . M., Cumming, R. L. C . , and Davidson, J . F . , Br. Med. J . , 1987, 295, 693. Lyon, T. D. B., Fell, G. S . , Hutton, R. C.. and Eaton, A. N . , J . Anal. At. Spectrom., 1988, 3, 265. The ICP-Is it The Real Thing? J. Marshall IC1 plc, Wilton Materials Research Centre, Middlesbrough, Cleveland TS6 8JE The purpose of this paper is to reflect on the contribution of the Strathclyde group, under the direction of the late Professor J.M. Ottaway, to the development and application of the inductively coupled plasma (ICP) as a spectroscopic source. The title of the lecture is taken, in part, from a play by Tom Stoppard, whose work John greatly admired. However, it also indicates something of John’s approach to analytical chemistry in that the practical relevance of the science was always stressed. Work on inductively coupled plasmas began at Strathclyde in 1981 as part of a growing collaboration on instrument development with Pye Unicam and Philips.’ Although the contributions of the research group to developments in atomic absorption spectrometry are perhaps the best known ,* the work in the ICP area was informed by considerable experience of the utilisation of the graphite furnace as an emission source over the previous decade.3 John Ottaway pointed out, for example, that while electrothermal atomisation atomic absorp- tion spectrometry (ETA-AAS) was criticised for being prone to interferences, the technique often out-performed induc- tively coupled plasma atomic emission spectrometry (ICP- AES) in this respect if the analyte to matrix concentration ratio was used as a criterion. John also advocated the view that the ICP should be viewed from the top rather than from the side in order to minimise the contribution to the total signal from the plasma continuum, in the same manner as in ETA-AES.4 Recent reports have shown that this idea is analytically viable.5 Wavelength modulation is a background correction tech- nique that has been employed with considerable success in flame6 and furnace7 emission spectrometry and in continuum source atomic absorption spectrometry.8 As a result of several collaborative projects between the Strathclyde group, Profes- sor T.C. O’Haver of the University of Maryland and Dr J. M. Harnly of the US Department of Agriculture, a modular computer-controlled background correction system was devel- oped for application in emission or absorption spectrometry, for the measurement of steady-state (e.g., flame or plasma) or transient (e.g., furnace or flow injection) signals.9 Essentially, the same hardware could be utilised with appropriate software to provide an extremely flexible multi-source, multi-technique instrument system. The details of the ICP implementation have been described by Hall et al. 10 A quartz refractor plate, stepper motor driven under the control of an Apple TIe microcom- puter, is placed in the optical path inside the monochromator and is used to scan in the immediate wavelength region of the emission line.Spectral positions for background correction can then be selected by the operator using cursor controls. This procedure offers a number of advantages for ICP [and indeed for direct current plasma (DCP)] work. Firstly, the system provides correction on a manual or automatic basis for the background continuum, which may be affected by components of the sample matrix. Secondly, because the process is interactive, judicious selection of correction positions may be made to overcome partial line overlap.Once these are selected, the system can alternate rapidly and repetitively between background and analyte peak positions, thus provid- ing automatic correction and discrimination against low frequency noise. Finally, the system provides a relatively inexpensive method of obtaining the performance of slew- scanning monochromators in instruments with manual wavelength control (e.g., the Spectrametrics echelle mono- chromator used in this work). The renewal of interest in emission spectrometry, resulting primarily from the commercial introduction of the ICP source, has led to a re-examination of the resolution requirements of spectrometers used for this purpose. Echelle spectrometers have increasingly been used in ICP-AES because these systems provide resolving power that approaches the physical line width, and consequently, they offer high spectral selectivity.Further, systems of this type are compact and are attractive from the point of view of designing high performance, low cost, benchtop instrumentation for plasma spectrometry. Although ICP systems are now widely employed for elemental analysis, the expense of such instruments, in terms of hardware, software and consumables, undoubtedly limits the potential size of the market. Indeed, there is a need for relatively simple low cost ICP equipment, which can be used to perform sequential determinations of perhaps 6 1 2 elements, to com- plement the range of automated AAS systems now available. Although it is clear that echelle spectrometers can now be designed and manufactured fairly cheaply to satisfy optical performance requirements, a reduction in hardware cost may also be possible by reconsidering the specification of the ICP itself.Most commercial ICP generators operate at 27 or 40 MHz, with output power in the range 1-2.5 kW. Argon gas consumption using the standard 18 mm i.d. torch is usuallyANALYTICAL PROCEEDINGS, JULY 1988, VOL 25 239 between 12 and 20 1 min-*. It is possible to reduce gas consumption by miniaturising the torch.” This approach also allows a reduction in the input power necessary to sustain a stable discharge, and has consequent cost implications for generator design. Decreasing the torch diameter, however, increases the technical complexity of its manufacture, and reduces the tolerances needed to provide acceptable plasma characteristics.For a given operating frequency, the minimum torch size is limited by the plasma skin depth. It is necessary to avoid dissipation of power in the central analyte channel of the plasma as this may increase interference effects. However, the plasma skin depth may be reduced by operating at higher frequency and this may also offer advantages in terms of signal to background ratio. 12 A study was made of the performance of an 80 MHz micro-ICP system (provided by Dr. P. W. J. M. Boumans) as part of a collaborative project between the Strathclyde group, Philips Nat. Lab., Eindhoven, and Pye Unicam. The micro- ICP consisted of an exactly scaled-down (9 mm i.d.) demount- able version of the standard Philips 18 mm i.d.torch. Analytical investigations of this plasma were carried out using a commercially available Spectrametrics echelle spectrometer and a laboratory-constructed, computer-controlled, mini- echelle spectrometer designed by Dr. C . V. Perkins of Pye Unicam. Comparisons were made with the performance of a Philips PV8490 50 MHz conventional ICP source on the same spectrometers. It was found that the 80 MHz micro-ICP provided better background equivalent concentrations (BECs) than the 50 MHz system fitted with a “standard” 18 mm i.d. torch.13 There was some indication that atom lines benefitted more than ion lines as a result of a reduction in the temperature of the micro-ICP, caused by the increase in operating frequency. It is likely that this improvement in sensitivity arises from the fact that the background intensity has a greater dependence on temperature than does the analyte intensity.It is also possible that the use of a micro-torch results in an increase in the atom density in the plasma. However, it was found that this improvement in BEC did not translate directly into an improvement in detection limits because the micro- plasma was less stable than the 50 MHz system. Nevertheless, detection limits in aqueous solutions, obtained using the micro-ICP, compared favourably with those for the conven- tional ICP. Interference studies indicated that the micro-ICP provided marginally poorer performance with respect to the standard system when operated under compromise plasma conditions. It was found that, although it was possible to maintain a stable plasma at power levels of around 600 W while aspirating the aqueous solution, the optimum setting for a typical “hard” line’? [such as Mn(I1) at 257.610 nm] was 0.82 kW.Similarly, although the micro-ICP could operate using a coolant argon flow-rate of 7 1 min-1, for compromise analytical performance, a flow-rate of 11 1 min-1 was required. Since these studies were carried out, there have been improve- ments in the design of standard 18 mm i.d. torches that have allowed a reduction of 50% or more in argon consumption. As the micro-ICP was a scaled-down version of the standard system, the design could clearly now be modified to provide still lower gas and power consumption. The selection of compromise conditions for ICP analysis is a complex matter, as the mechanism for interference effects is not yet clearly established.However, the concept of “hard” and “soft” lines14 appears to offer a pragmatic approach to this problem. A study was made of the effect of plasma parameters on the behaviour of a number of “hard” and “soft” lines in the presence of interferents.15-16 It was established that the adoption of “hard” line conditions provided a general method of minimising interferences in multi-element analysis at the cost of a minor compromise on sensitivity for some “soft” lines. An effective demonstration of this conclusion was provided in the determination of copper in serum by ICP-AES.17 It was found that easily ionised elements in the serum matrix caused an interference in the measurement of copper using plasma conditions that had been optimised for maximum analyte sensitivity.Although this problem could be removed by matrix-matching of samples and standard solutions, it was more convenient to alter the ICP conditions to parameters optimised for “hard” lines. Excellent agreement was obtained using this approach, which is based on the concept of energy profiles in plasma rather than on empirical investigations. At ICI, there is a need to analyse a wide variety of sample types on a routine basis. The application of informed compro- mise conditions has played a significant role in the recent adoption of ICP-AES as the principal analytical tool in our laboratory. The philosophy of performing direct analysis whenever possible, as championed by John Ottaway, remains the target.Myers - Tracy signal compensation has been used in conjunction with a Babington-type nebuliser to improve measurement precision in solutions containing high levels of dissolved solids.18 Using modern auto-tuning generators, it is now possible to introduce solvents such as methanol and acetone into the ICP directly via the normal sampling system and this is obviously an important step forward in the analysis of petrochemicals and plastic materials. However, despite these technical improvements, ICP-AES is undoubtedly limited in its sensitivity for many elements, and in asking the question in the title, consideration should also be given to mass spectrometric detection. In addition to the high sensitivity offered by inductively coupled plasma mass spectrometry (ICP-MS), the relative simplicity of the spectra obtained (which are clearly amenable to analysis by expert systems) and the speed of data acquisition (e.g., in semi-quantitative analysis) are significant advantages over ICP-AES.The latter technique, by virtue of the time constant of the measurement, has the potential for direct multi-element analysis of solids via laser ablation and electrothermal vaporisation-such a system is presently under construction in our laboratory. Conclusion However interesting a technique, or an aspect of its develop- ment, may be, it is unlikely that it will be the most appropriate tool to examine all potential problems in a given area. Advances in research are self-perpetuating-the technically impossible will always remain. The ICP is no more the real thing than its predecessors or successors. What matters most is the endeavour of those involved in overcoming the challenges posed to analytical science. John Ottaway was always keen to promote collaboration on the local, national and international scale, as evidenced by much of the work described here. Those people represent John’s contribution and legacy as much as the science itself. To them, John Ottaway was The Real Thing. The author would like to acknowledge the contribution of members of the Strathclyde group, both past and present, whose work in the ICP area since 1981 has been briefly summarised here. The provision of equipment, and the technical support and advice of Dr. P. W. J. M. Boumans and J. J. A. M. Vrakking of Philips National Laboratory, Eind- hoven, and Dr. C. Perkins of Pye Unicam, is particularly acknowledged. Thanks are due to the Pye Foundation and the SERC for support of the work at Strathclyde. The author would like to thank ICI plc for permission to publish this paper. References 1. 2. 3 . 4. 5. 6. Marshall, J., Littlejohn, D., and Ottaway, J . M., Strathclyde Rev., 1983, 7. Ottaway, J. M., At. Spectrosc., 1982, 3, 89. Ottaway, J. M., Bezur, L., and Marshall, J . , Analyst, 1980, 105, 1130. Littlejohn. D., and Ottaway, J . M.. Anafyst, 1977, 102, 553. Davies, J., Dean, J . R., and Snook, R. D.,Analyst, 1985,110, 535. Epstein, M. S . , and O’Haver, T. C., Spectrochim. Acta, Part B , 1975, 30, 135.240 ANALYTICAL PROCEEDINGS, JULY 1988, VOL 25 7. 8. 9. 10. 11. 12. Bezur, L., Marshall, J., and Ottaway, J . M., Spectrochim. Acta, Part B , 1984,39, 787. Marshall, J., Ottaway, B. J . , Ottaway, J. M., and Littlejohn, D., Anal. Chim. Acta, 1986, 180, 357. O’Haver, T. C., Harnly, J. M., Marshall, J., Carroll, J., Littlejohn, D., and Ottaway, J. M., Analyst, 1985, 110, 451. Hall, D. H., Littlejohn, D., Ottaway, J. M., and O’Haver, T. C., Anal. Proc., 1986, 23, 18. Weiss, A. D., Savage, R. N., and Hieftje, G . M., Anal. Chim. Acta, 1981, 124, 245. Capelle, B., Mermet, J . M., and Robin, J., Appl. Spectrosc., 1982, 36, 102. 13. 14. 15. 16. 17. 18. Marshall, J . , Bamiro, F. O., Corr, S . P., Littlejohn, D., and Ottaway, J. M., to be published. Boumans, P. W. J. M., and Lux-Steiner, M.Ch., Spectrochim. Acta, Part B., 1982, 37, 97. Bamiro, F. O., Littlejohn, D., Marshall, J., and Ottaway, J. M., Anal. Proc., 1983, 20, 602. Bamiro, F. O., PhD Thesis, University of Strathclyde, 1984. Bamiro, F. O., Littlejohn, D., and Marshall, J., J . Anal. At. Spectrom., 1988, 3, 279. Marshall, J . , Rodgers, G., and Campbell, W. C . , J. Anal. At. Spectrom., 1988, 3, 241.

ISSN:0144-557X    

DOI:10.1039/AP9882500230

出版商:RSC    

年代:1988

数据来源: RSC

摘要:

240 ANALYTICAL PROCEEDINGS, JULY 1988, VOL 25 ln Situ Pre-concentration in Flame Atomic Spectrometry T. S. West Macaula y Land Use Research Institute, Craigiebuckler, Aberdeen A89 2QJ The author of this contribution to the Professor Ottaway Memorial Meeting was privileged to act as the external examiner for his PhD submission at the University of Exeter and had close contact with his work continuously since that time. John Ottaway was always appreciative of novelty and simplicity of concept in analytical techniques; it is hoped that this paper at his Memorial Meeting satisfies both these criteria. The Concept of In Situ Pre-concentration There are numerous instances when the analyte to be determined is present in the test sample at a concentration below the capability of the available analytical techniques.In such instances it is usually separated from the sample and, in the process, concentrated in a smaller volume of solution or some other matrix. Extraction into a water-immiscible solvent is frequently used, or ion exchange or co-precipitation. The pre-concentrated sample is then submitted for analysis or further manipulated as necessary, e.g. , elution from an ion-exchange column, dissolution of the matrix in co-precipita- tion or back-extraction into an aqueous phase in solvent extraction. The use of such pre-concentration procedures necessarily involves additional equipment , such as container vessels, and additional chemicals, such as acids, resins and solvents. All of these constitute sources of contamination by the test ion itself and possibly interferents and may be a serious source of error, particularly in situations where the analyte may be present at similar levels in the original sample, Conversely, the test ion may be partially lost by adsorption on the walls of additional equipment, e.g., container vessels. The electroanalytical chemist has as a resource a nearly ideal in situ pre-concentration technique for metal cations which uses no additional chemicals or apparatus, viz. , cathodic electrodeposition on a hanging mercury drop from the test solution and subsequent measurement as the deposited ion is subjected to anodic dissolution, i.e., the techniques of the various forms of anodic-stripping voltammetry. Until fairly recently the closest the atomic spectroscopist has come to this desideratum of not using added chemicals or additional equipment has been in the technique of electrother- mal atomisation.It is possible repeatedly to dry out solution samples in graphite furnace tubes or on rods before heating fully for atomisation. However, for a variety of reasons, the chief of which is, for most samples, vastly increxed back- ground or matrix effects, this is frequently not a practical proposition. In flame-based atomic absorption spectrometry the sensitiv- ity of measurement is considerably lower than th'at of the electrothermal atomisation techniques as the atomic species from the sample are diluted extensively in the rapidly expanding flame gases and are distributed throughout the body of the flame. In contrast, the dilution effect in electrothermal atomisation is considerably less and much of the sample atoms coming, as they do, from a point source can be located at the time of measurement within the beam of analysing radiation from the hollow-cathode lamp.The electrothermal techniques give a transient signal and are subject to background absorption effects from matrices; the flame techniques give a steady signal during aspiration of the sample solution and are generally considerably less prone to background absorption phenomena. Generally, therefore, flame techniques are easier to use. Long-pathway Atomic Absorption Spectrometry Various experiments have been made previously to endow the more user-friendly characteristics of flame spectrometry with higher sensitivity by restricting the flowing flame gases into a Table 1.Instrumental parameters for the atom trapping AAS determination of 12 elements Working conditions Ag As Au Bi Cd Cu Mn Hollow-cathode lamp current/mA 1 4 4 3 2 3 5 Bandwidthinm . . . . . . . . 0.2 1.00 0.2 0.2 0.5 0.5 0.2 Wavelengthinm . . . . . . . . 328.1 193.6 242.8 223.1 228.8 324.8 279.5 Distance between collector tube Obscuration of optical path by Flamemixturei.. . . . . . . A-A A-B A-A A-A A-A A-A A-A Flamecondition$ . . . . . . R 0 0 R 0 0 S Backgroundcorrection . . . . No Yes No Yes Yes No No (base) and the burner headimm 5 21 7 5 7 5 9 tube, YO * . . . . . . . . 50 35-40 25 10-15 50 50 35-40 * Obscuration figures are approximate. i. A-A = air - acetylene; A-B air - butane. j: R = reducing; 0 = oxidising; s = stoicheiometric.Pb 3 1 .0 217.0 5 50 A-A R Yes Sb 4 0.2 217.6 5 35-40 A-A R Yes Se 7 1 .o 196.0 10 25 A-A 0 Yes T1 6 0.5 276.8 6 3 5-40 A-A R No Zn 2 0.5 213.9 12 50 A-A 0 YesANALYTICAL PROCEEDINGS, JULY 1988. VOL 25 24 1 narrow contained path. Two of these are well worthy of mention. Fuwa and Valleel passed the flame from a so-called total consumption burner into the end of an open horizontally mounted 1 m long silica tube and aligned the beam of the hollow-cathode lamp along its longitudinal axis, thus obtaining an approximately 1 m long absorption path. This gave increased sensitivity, but the technique suffered from strong background absorption by the flame gases, including those of the primary zone which were also directed along the tube. Hingle et a1.2 avoided much of this background problem by using a separated flame.The interconal flame gases above the primary combustion region were pulled along a silica tube with end-windows by gentle suction, the remainder of the gases being burned at an exit vent on top of the tube. The beam of analysing radiation passed through the silica end-windows along the axis of the tube, which contained the atomic species of the analyte generated in the primary combustion region. Again, this device of the long-path flame increased the sensitivity more or less as predicted by the Lambert-Beer absorption law, but there were problems from memory effects in the tube, obscuration of the end-windows, etc. The gain in sensitivity in both these long-path flame devices was interesting rather than useful for practical analyses.Atom Trapping In Situ The idea of trapping atomic species or their pre- or post-cursors within the body of a flame was first demonstrated by Lau et al.3 A 4 mm water-cooled silica tube was mounted in the flame gases of a conventional approximately 10 cm long air - acetylene atomic absorption burner immediately above the burner head slot and just below the light path of the beam from the hollow-cathode lamp. Coolant water (Fig. 1) was passed through the silica tube whilst the analyte solution was aspirated into the flame. The cold outer surface of the tube condensed and progressively collected metallic species generated as free atoms in the lower parts of the flame together with oxides, carbonates or other compounds of the analyte species aspirated in the flame.After collection or trapping in this way for a given II A[ 1 Drain Fig. 1. Apparatus. Block diagram in plan. Collector tubc A vertically situated ca. 0.5 cm above slot of burner head B. Air and water passed through tube via tap, C period of time, aspiration of the sample solution was disconti- nued and distilled water was aspirated in its place. The coolant water stream in the silica tube was replaced by blowing a stream of air through it. Within a very short period, usually ca. 2-5 s depending on the vapour pressure of the condensed analyte metal or metal compound, the temperature of the surface monolayer of material on the tube rose sufficiently to release the atomic species into the analqtsing beam of radiation, thus producing a transient atomic absorption signal the size of which was proportional to the amount of the trapped analyte species.The flame gases remove the analyte species from the surface of the silica tube rapidly and the absorption signal returns to the zero or background level (Fig. 2). At this point the coolant water may be switched back through the hot tube (Fig. 1) without cracking it because the coefficient of thermal expan- sion of silica is very small. The next sample may then be aspirated, collected and analysed, etc. Thus, Lau et al.3 were able to achieve in situ pre-concentration within the body of the flame without resorting to the addition of further chemicals and with the introduction of only one physical surface, in much the same way as the hanging mercury drop in anodic-stripping voltammetry and which, like that device, is self-cleaning in the analytical A Fig.2. Signal for Cu2+ (1 p.p.m.) in 0.1 M HCl after 1 min collection.4 Aspiration begins for 1 min at A . Collection ceases at B when the blank solution is aspirated thereafter. Coolant water blown from tube at C. Signal completed within 6 s and coolant water switched back into tube at D The absorption signals obtained for copper were about ten times more sensitive than that obtained using the spectrometer in the normal way. According to Lau et al.3 elements with shorter wavelength resonance lines, e.g., cadmium, lead and zinc, gave poorer detection limits owing, it was thought, to larger background signals when the trapping tube was heated. The elements of higher melting-point were trapped, but not re-atomised so readily.The alkali and alkaline earth elements were also readily trapped, but etched the silica tube and could not be determined. Development of Atom Trapping Atomic Absorption Spec- trometry The atom trapping technique was re-examined by Khalighie.4 Variables such as the position of the tube with respect to the burner head and to the beam of radiation, the diameter of the tube, the effect of ageing its surface, the collection time and other variables for elements such as cadmium, copper and lead, the non-release of trapped calcium and the effects of some ions thought likely to interfere were studied. It was found that whereas low concentrations of aluminium decreased the sensitivity for copper, for example, as expected, large concen- trations, e.g., 1000 p.p.m., increased the sensitivity by a factor of 50-100%.The white deposit of alumina on the tube could easily be removed by brushing. It was further shown that a silica collector tube pre-coated with alumina, by pre-aspiration of a 1000 p.p.m. aluminium solution, behaved with equal efficiency and it was concluded that the increased sensitivity was due to the larger specific surface of the collecting interface produced by the presence of finely divided alumina. Vanadium behaved in the same way as aluminium, but the yellow deposit of vanadium oxide sintered into the surface of the silica and could not be removed by brushing. The existence of maxima on some absorption signals was attributed to the deposition of242 ANALYTICAL PROCEEDINGS, JULY 1988, VOL 25 more than one chemical species of the analyte on the tube and which released the atomic form of the analyte at different temperatures.Subsequently these phenomena were studied in greater detail and optimum conditions were established for the determination of a much wider range of elements.5Jj Positions of Trapping In Situ and Measurement The position of the 4 mm tube was not particularly critical, but was found to be such that the flame should be able to unite above it. It was expected that the population of atoms released from the surface of the silica tube would be greatest imme- diately above it. The largest signals were observed with approximately 25-30% obscuration of the hollow-cathode lamp by the upper surface of the tube, i.e., at grazing incidence.Under these conditions,7 Cd, Pb, Se and Zn in an air - acetylene flame and As in an air - propane flame gave sensitivities that were respectively 18-, 48-, 3-, 80- and 60-fold greater than those obtained using the same spectrometer in the conventional way. The imposition of a pair of horizontal 2-mm slits in the light path on either side of the flame reduced light by-passing the densest cross-section of atoms in the flame and further increased the signals for Cd, Cu, Pb, Se and Zn by factors of 4.5, 2.5, 2.5 and 5.5, respectively. Use of Collector Tubes Other than Silica Metal tubes were examined in place of silica to obtain colder surfaces on which to collect the more volatile elements, such as cadmium and selenium.A nickel tube was three times more sensitive than a silica tube for cadmium, but it tended to retain selenium during the release cycle and although it trapped copper it would not release it, presumably owing to alloy formation. A copper tube, on the other hand, was between three and four times more sensitive than the nickel tube for both cadmium and selenium.8 Further examination6 showed that the use of metal tubes, including titanium and stainless steel, was fraught with practical difficulties because of physical distortion after a few heating - cooling cycles. In general, it was found easier and better to incorporate ca. 1000 p.p.m. of Cu2+ in the cadmium and selenium test solutions rather than to pre-metallise the silica tube surface by spraying the 1000 p.p.m.solution. Alumina-coated Silica Collector Tubes Following the earlier observation4 that whereas small amounts (ca. 100 p.p.m.) of aluminium decreased copper signals, 1000 p.p.m. had the opposite effect, aluminium and other refractory oxide-forming elements such as vanadium and iron were examined as oxide-coating materials for silica tubes, the coating being simply effected by pre-spraying a strong (1000 p.p.m.) solution of the metal salt before in situ collection of the analyte i o n 9 Significantly, it was found that not only was the signal for some analyte species increased, but also problems due to the interference, by attack on silica, of the alkali metals and alkaline earths were eliminated and the interference of other oxide-forming co-elements such as cobalt, iron, nickel and manganese were reduced.6.10 The effect of iron, a common contaminant, was eliminated by pre-coating the alumina- coated surface with a 1000p.p.m.iron solution from the flame or by using an iron oxide-coated tube. The iron oxide-on-alumina is preferable, however, because the conglomerate can easily be brushed off the silica tube subsequently. With high concentra- tions of nickel or cobalt pre-coating was not successful, so that high concentrations of these elements must be reduced or eliminated prior to analysis for other analyte species. The physico-chemical nature of the protective oxide coatings of alumina and iron deposited by the flame on the silica tube were examined by X-ray crystallography, infrared absorption spectrometry and electron microscopy and their purity by spark-source mass spectrometry.6.10 The alumina was found to be highly pure, mainly &alumina with a small amount of the 8 form and with a d 1 ym spheroid format. The iron oxide, which could not be removed by brushing, consisted of more densely packed haematite platelets containing a small amount of metallic iron. Whereas these coatings yielded higher signals than an uncoated silica tube for cadmium, lead and zinc, the signals for gold and manganese were slightly lower, but were free from interference from the alkali metals.The signals measured in all of these experiments were those of the peak heights. It was observed that with the alumina- coated tubes copper was released more quickly (6 s) than from an untreated silica tube (10 s) but, in general, irrespective of the appearance time, the peak shapes from vanadium oxide, alumina or iron - alumina coated tubes were broader owing to stronger retention on the oxide surface, presumably because of the interactions on, or in, the oxide substrate.These observa- tions suggest that the integration of peak areas would be more Table 2. Comparisons of characteristic concentrations and detection limits of 12 elements determined by conventional flame AAS and atom trapping AAS Ag As Au Bi Cd Cu Mn Pb Sb Se TI Zn Conventional flame A A S : Characteristic concentration*/ pgml-* . . . . . . . . 0.036 0.92 0.22 0.25 0.012 0.088 0.037 0.12 0.44 0.38 0.28 0.01 Detectionlimitt/pgml-1 . . 0.025 0.85 0.018 0.05 0.004 0.044 0.011 0.04 0.65 0.4 0.15 0.005 Relative standard deviation ( n = 12), .. . . . . 1.5 7 4 5 3 5 2.5 3.5 5.5 8 4 5 A t o m trapping A A S (silica atom trap): Characteristic concentration*/ pg ml-' 30-scollection . . . . . . 0.003 0.087 0.049 0.022 0.0013 0.015 0.010 0.010 0.038 0.050 0.043 0.002 1-mincollection . . . . . . 0.0016 0.045 0.037 0.012 0.0007 0.0075 0.005 0.0057 0.021 0.028 0.022 0,001 2-mincollection , . . . . . 0.0009 0.024 0.028 0.006 0.00046 0.004 0.003 0.0029 0.01 1 0.015 0.015 0.0005 3-mincollection . . . . . . 0.0006 0.019 0.022 0.004 0.00037 0.0028 0.002 0.0020 0.007 0.0088 0.012 0.00036 col1ection)iygml-' . . . . 0.0005 0.05 0.020 0.010 0.0001 0.004 0.002 0.005 0.036 0.030 0.009 0.001 ( n = 12),O/O . . . . . . 2.9 10 8.7 8 7.4 5 4.2 6.8 8.3 14 5 6.3 Detection limit+ (2 min Relative standard deviation * Sometimes called sensitivity (pg ml-1 required for 1% absorption).t Signal to noise ratio 2 : 1.ANALYTICAL PROCEEDINGS, JULY 1988, VOL 25 243 beneficial than the peak-height measurements used in this work. Nature of Species Trapped by In Situ Concentration Physical examination of the deposits after spraying strong solutions of Ag, Au, Cd, Co, Cu, Pb, Se and Zn showed that they were almost entirely metallic, and this was confirmed by X-ray crystallographic and electron microscopy examination. Ca, Cr, K, Li, Mg, Mn and Na, on the other hand, are trapped almost entirely as silicates or oxides, as are A1 and V.6.11 For some elements, the presence of multiple peaks, e.g., with Cu, indicates that the trapped species may be a mixture of mainly metal with some oxide.Again, physical examination reveals the presence of oxide films at the edges of the tube where the partial pressure of oxygen will be at its greatest owing to the diffusion or entrainment of atmospheric oxygen in the swiftly flowing flame gases. The use of a slightly fuel-rich flame removed the smaller peak for copper, indicating that the deposition was entirely as free metal in the richer flame. It is an interesting corollary of this observation that because virtually all metal and metalloid species in the flame will be deposited on the cold tube, they will be released progressively (the non-volatile ones first) at different periods of time as the interface heats up, thus making possible, for example, time- resolved atomic fluorescence signals as in carbon filament non-dispersive atomic fluorescence spectrometry.12 A second and perhaps more significant possibility suggests itself, namely that, depending on the flame temperatures- particularly cooler flames such as argon-diluted pre-mixed air flames or hydrogen-diffusion flames employing aspiration of the solution on nitrogen-it may be possible to examine speciation for some metals or metalloids using this technique. Mechanism of Atomisation of In Situ Deposited Species The temperatures of the tube were measured in the flame with and without coolant water flowing through it.j.6.11 Thermo- chromic crayon marks made on the cold tube showed that its surface temperature ranged from 75 to 100°C from top to bottom in the air - acetylene flame. This agrees well with the obervation that water vapour was only rarely seen to condense from the burning flame gases and aspirated water and even then only at the outer edges of the flame.Optical pyrometry and thermochromic crayons showed the temperature of the hot tube during the release cycle to be around 6704350°C within about 20 s of switching off the coolant water and thereafter not to increase much more. It is clear, therefore, that the tube is sufficiently cool to condense almost any metal or metalloid from the flame gases except possibly mercury and to a lesser extent selenium and arsenic (for which a cooler air - propane or air - butane flame is satisfactory). On the other hand, these data appear to suggest that in the release cycle the tube will not be sufficiently hot to boil off or even melt some of the species that give satisfactory signals in practice.However, the temper- ature measurements by optical pyrometry correspond to that of the bulk of the tube and not that of its surface. There is every reason to suppose, however, at least to a first approximation, that the temperature of the interface between the tube and the flame gases will be fairly close to that of the flame, which is the heat-generating phase. There is a good correlation between the appearance times of the various metals and their melting- points, but not their boiling-points, e.g., for cadmium, copper, gold, lead, selenium and silver5.11 (Fig. 3). The normal temperature-related vapour pressure of the trapped metals on the surface will become appreciable after softening or melting and, at that stage, additional atomisation due to sputtering by energetic flame species will become appreciable.In essence, therefore, it can be concluded that the atomisation from the species trapped in the monolayer at the silica or alumina- coated silica surface is likely to be little different from that which occurs at the surface of the particles of solid formed in normal flame atomic absorption spectrometers by evaporation of the aspirated droplets of analyte solution. l o 7 2 A:Na A L i A Mn I I I I I I I I 400 600 800 1000 1200 1400 1600 1800 2000 2200 Fig. 3. Correlation between appearance time and melting-point of metal. The silicate-forming alkali metals which etch the tube appear above the curve. Cr and Mn which deposit largely as oxides lie below it.The curvature is analogous to the shape of the temperature - time curve found by optical pyrometry and thermochromic crayon response M.p./K Analytical Parameters for In Situ Atom Trapping Atomic Absorption Spectrometry and Analytical Applications The behaviour and optimum conditions for application of the in situ pre-concentration technique of atomic absorption spectrometry are summarised for twelve elements in Table 1. Table 2 compares the characteristics of the technique using a silica tube with those of conventional atomic absorption spectrometry and the behaviour of alumina and iron oxide coated tubes is given in Table 3. From these data it can be seen that the technique is more sensitive by one or two orders of magnitude in the characteristic concentration following in situ collection for 2-3 min for most metals (Table 2) and that the relative standard deviation is only slightly less favourable even at these low levels.Because most samples presented for analysis usually contain appreciable concentrations of sodium, alumina- or sometimes iron oxide/alumina-coated collector tubes have been used in most practical applications. Selenium has been determined in plant tissues13 following combustion in an oxygen flask and absorption in acetic acid, after which the sample solution was aspirated into an air-acetylene flame and collected on an alumina-coated tube. Various alternatives to the silica tube were examined and a range of potential interfering ions. The results compared well with those obtained by the standard piazselenol fluorimetric method.Lead and cadmium, total and extractable, were determined in soils14 using aqua regia digests for the total content and EDTA and acetic acid extracts for the exchangeable content. Whereas the conventional flame technique was just sufficiently sensitive for the total lead content of the surface soils, it was not sufficiently so for cadmium. The acetic acid extracts could only be examined by the atom trapping technique because of the smaller amounts removed by acetic acid. An alumina or iron oxide - alumina coated silica tube was used to pre-concentrate the lead and cadmium. The enhancement in sensitivity under these practical conditions following a 2-min collection was 30-fold for both elements.Calcium chloride solution, ca. 0.05 M, is frequently used as an extractant in soil analysis because it gives a better correlation with plant uptake than many other extractants used to simulzite the plant availability of trace elements such as cadmium. The concentration of cadmium in such extractants is generally too low to be determined by conventional flame atomic absorption and additionally the high calcium content causes background problems. Graphite furnace atomic absorp- tion measurements are more sensitive, but suffer badly from volatility and re-combination effects of the calcium chloride.244 ANALYTICAL PROCEEDINGS, JULY 1988, VOL 25 Table 3. Characteristic concentrations and detection limits of the 12 elements determined by atom trapping AAS using alumina and iron oxide coated tubes TI Zn Ag As Au Bi Cd Cu Mn Pb Sb Se Alumina coated tube (2-min collection): Characteristic concentration/ pgml-1 .. . . . . . . 0.0015 0.072 0.039 0.010 0.0005 0.008 0.008 0.0026 0.044 0.023 0.024 0.0005 Detectionlimit/pgml-1 . . . . 0,0008 0.15 0.035 0.030 0.0003 0.019 0.010 0.0011 0.048 0.05 0.011 0.0017 Relative standard deviation ( n = 12), O/O . . . . . . 4 10 9 10 8 7 7 7.5 12 12 6 4 Iron oxide coated tube (2-min collection): Characteristic concentration/ pgml-1 . . . . . . . . 0.0022 0.095 - * 0.012 0.0002 -* - * 0.003 0.090 0.022 0.027 0.0002 Detectionlimit/pgml-* . . . . 0.001 0.19 - * 0.040 0.0001 -* - * 0.015 0.10 0.045 0.012 0.0005 Relative standard deviation * Not applicable. (n = 12),"/0 . . . . . . 6 11 - * 10 8 - * - " 7 11 13 5 5 The in situ collection technique on an alumina-coated silica tube, however, yielded excellent results,15 comparable to those obtained by the complicated and time-consuming procedure involving chloroform extraction of the dithizonate with back- extraction into hydrochloric acid and conventional atomic absorption spectrometry.Another example of the sensitive nature of the in situ pre-concentration technique being used to good effect is the determination of the arsenic content of top soils with alumina and iron oxide - alumina coated tubes with 4-min collection times. A standard additions technique and matrix-matched calibration graphs gave good agreement.10 In the determina- tion of zinc in sea water the latter was diluted 10-fold to reduce sodium matrix problems and the zinc was collected on an alumina-coated silica tube for 1 min before atomisation.Standard additions and matrix-matched calibration graphs were again used.10 Similarly, cadmium, copper, lead, man- ganese, silver, thallium and zinc have been determined in drinking water using an alumina-coated silica collector tube and an in situ pre-concentration for 4 or 5 min as necessary. With zinc the accuracy was capable of being checked by conventional atomic absorption spectrometry, and graphite furnace atomisation was possible for lead and managanese.10 Spark-source mass spectrometry was used to check the results obtained by collection on an iron oxide/alumina-coated silica tube for 2 min of bismuth, cadmium, lead and silver in nitric acid - perhydrol digests of bovine liver samples.With cadmium, the collection was only for 30 s. Conclusions The direct analyses of the 12 elements discussed in this paper in very diverse samples and at levels that were difficult or impossible without extensive time-consuming and potentially contaminating handling procedures using other techniques of atomic absorption spectrometry or spectrophotometry illus- trates the viability of atom trapping atomic absorption spec- trometry using in situ pre-concentration directly within air - acetylene or other pre-mixed flames. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. References Fuwa, K. W., and Vallee, B. L., Anal. Chem., 1963, 35, 942. Hingle, D. N., Kirkbright, G. F., and West, T. S., Talanta, 1968, 15, 199. Lau, C., Held, A., and Stephens, R ., Can. J. Spectrosc., 1976, 21, 100. Khalighie, J . , DZC Dissertation, Imperial College, London, 1975. Khalighie, J . , PhD Thesis, University of Aberdeen, 1979. Lau, C. M., PhD Thesis, University of Aberdeen, 1982. Khalighie, J., Ure, A. M . , and West, T. S . , Anal. Chim. Acta. 1981, 131, 27. Khalighie, J., Ure, A. M., and West, T. S . , Anal. Chim. Acta, 1982, 134, 271. Khalighie, J . , Ure, A. M., and West, T. S . , Anal. Chim. Acta, 1979, 107, 191. Lau, C. M., Ure, A . M . , and West, T. S . , Bull. Chem. SOC. Jpn, 1988, 61, 79. Khalighie, J . , Ure, A. M., and West, T. S . , Anal. Chim. Acta. 1980, 61, 79. West, T. S . , Pure Appl. Chem., 1978, 50, 837. Lau, C. M., Ure, A. M., and West, T. S . , Anal. Chim. Acta, 1982, 141, 213.Lau, C. M., Ure, A. M., and West, T. S . , Anal. Chim. Acta, 1983, 146, 171. Fraser, S. M., Ure, A. M., Mitchell, M. C., and West, T. S . , J. Anal. At. Spectrom., 1986, 1 , 19. Probes and Furnaces for Molecular Emission Cavity Analysis N. Pourreza, Alan Townshend and Paul S. Turner School of Chemistry, University of Hull, Hull HU6 7RX Molecular emission cavity analysis (MECA) is a technique in which emissions from simple molecules (e.g., S2, HPO) are generated in a small cavity placed in a hydrogen-based flame.' The sample may be a solid or liquid, placed in the cavity on the end of a probe before it is inserted into the flame, or it may be a gas or vaporised species introduced into the cavity already placed in the flame.2 In the first approach, the position of the cavity in the flame and its rate of introduction therein must be very reproducible, especially as the emissions from more volatile species occur very rapidly (within 1 s).Such reproduci- bility can be achieved by completely automating the cavity introduction process and by microcomputer control of the parameters governing the rate and duration of insertion, residence time in the flame and subsequent cavity cooling time. For example, the relative standard deviation (RSD) for the determination of 0.25 pg of phosphorus as phosphate, on the basis of its emission as HPO, was decreased from 4.5% for the manual method to 0.9% for the automated procedure (n =ANALYTICAL PROCEEDINGS, JULY 1988, VOL 25 L 245 10) .3 Similar improvements were achieved in the determina- tion of sulphur compounds.4 It was thought that it would be advantageous if the sample volatilisation process took place away from the flame, and if the volatile products were carried into a cavity already placed in the flame.Such an approach, involving chemical conversion into volatile species, is already common in many atomic spectroscopic techniques (e.g. , hydride generation) and has been extensively applied in MECA.2 Pourrezas used an electrically-heated tantalum filament to volatilise samples, and the volatile products, which included, for example, sulphur- containing species, were transported to a water-cooled cavity by a stream of nitrogen. The calibration graphs shown in Fig. 1 were obtained for 2-yl sample solutions of three inorganic sulphur compounds; similar responses were obtained for organic sulphur compounds.The detection limits were 0.25, 4.0, 2.0 and 2.5 ng of S for Na2S03, Na2S04, methionine and sodium lauryl sulphate, respectively. The RSD was normally 45%. Phosphates couId be determined with similar and selenium with less sensitivity. The procedure is very simple, the sample solution being injected on to the filament, through a septum, followed by controlled electrical heating of the filament. 40 5 20 .- v) .- E w 10 I I 0 100 200 Amount of sulphur!ng Fig. 1. Calibration graphs for (A) Na,SO,; (B) K2S208; and (C) Na2S04 in 2 4 samples, obtained by tantalum filament volatilisation Attempts were also made to apply this approach to the direct determination of sulphur in steel samples. However, the metal samples caused very rapid deterioration of the tantalum filament.To overcome this problem a modified Massmann furnace was used to heat the steel samples in a current of hydrogen.6 Gas inlet and outlet ducts were attached to the ends of the furnace, through which hydrogen entered and left the interior of the graphite tube. Nitrogen was retained as the shielding gas for the other parts of the furnace. Graphite tubes without holes were employed for sample injection and hydrogen was used as the fuel gas for the flame which burnt inside the cavity.' By heating 9.5-mg samples of cast iron containing 0.05% of sulphur to 1600°C in this type of furnace, the hydrogen sulphide produced gave intense transient Sz signals when transported to the cavity by the stream of hydrogen, but the reproducibility between samples was poor.When thiourea was used in place of the iron sample so as to provide a similar amount of sulphur to that present in the iron, signals of similar intensity were obtained, but the RSD improved to 5% (n = 6). The detection limit was 4 ng of S , which is very similar to that Concentration of sulphur in steel, O/O Fig. 2. Calibration graphs for sulphur in ferrous metals obtained by direct Massmann furnace volatilisation in a stream of hydrogen. (0) High purity iron; (V) cast iron; (0) plain carbon steel (Mn <1%, Cu <0.5%); ( A ) plain carbon and low alloy steel ( 2 7 0 , Mn >1%); ( X ) highly alloyed steel, with respect to Mn or Cu (9.5-mg sample); and (m, A, a) highly alloyed steel, with respect to Cr and Ni (9.5-mg sample) attained with the tantalum filament vaporiser.Therefore, the irreproducibility seemed to arise from the sequence of metallic samples analysed. Another problem that is likely to arise with metallic samples is the effect of metal - sulphur bonding on the release of H2S. Comparison of metal - sulphur bond energies showed that, of the metals commonly present in steels, only manganese forms a stronger bond with sulphur (205 kJ mol-1) than does iron(II1) (96 kJ mol-1) .s Cobalt(II), chromium( 111), nickel(I1) and copper(I1) all form weaker bonds with sulphur. Hence, only manganese would be expected to hinder sulphur release more than the main matrix element, and possibly depress the emission. Interestingly, when such metals were added to thiourea in the furnace, and the temperature was raised to 1600 "C, all the metal ions tested depressed the peak height by the following amounts: MnII (94%) > Fe"' (83%) > CU" Log (concentration of sulphur x 1000, O/O) Fig.3. Logarithmic plot of the results shown in Fig. 2 after correction for the effects of (0) Mn and Cu; (0) Mn alone; and ( A ) not corrected246 (31%) > Ni" (29%) > Cr"J (24%) > CoJI (12%) = MoV1 This trend provides some confirmation of the predicted effect of these metals. Part of the irreproducibility arose because of the accumula- tion of the metal samples in the furnace as more samples were analysed. This, and other sources of error, could be accommo- dated by bracketing the unknown sample with two standard metal samples of similar composition.When this was done, reproducible results could be achieved for 5-20 mg-samples of powdered metals and calibration graphs could be obtained over the range 0.002-0.06% of S in various iron and steel samples (Fig. 2). It is evident from Fig. 2 that different types of steel and iron samples fall on different calibration graphs and that the points for high manganese or copper steels are greatly displaced to lower sensitivity compared with the other calibra- tion points. A statistical survey of the results was carried out in order to identify any significant trend caused by any of the metals present in the samples in addition to iron. This survey showed that only two metals [ v i z . , manganese and copper (when manganese was also present)] influenced the S2 intensity and that the magnitude of the effect could be quantitatively related to the metal concentration.It was therefore possible to calculate correction factors to accommodate the effects of these two metals. When these corrections were made the ANALYTICAL PROCEEDINGS, JULY 1988, VOL 25 calibration graph shown in Fig. 3 was obtained. Hence, this procedure allows the direct determination of sulphur in powdered ferrous metals. As with the tantalum filament, phosphorus and selenium as well as tellurium, arsenic and antimony could be determined in a similar way. The authors thank the Ministry of Defence for a grant to P. S. T. and the Bureau of Analytical Samples for providing the analysed, powdered samples. 1. 2. 3. 4. 5 . 6. 7. 8. References Bogdanski, S.L., Burguera, M., and Townshend, A . , CRC Crit. Rev. Anal. Chem., 1981, 10, 185. Henden, E., Pourreza, N., and Townshend, A . , Prog. Anal. At. Spectrom., 1979, 2, 355. El-Hag, I. H . , and Townshend, A . , J . Anal. At. Spectrom., 1986, 1 , 383. Evmiridis, N., and Townshend, A., J. Anal. At. Spectrom.. 1987, 2, 339. Pourreza, N., PhD Thesis, University of Birmingham, 1981. Turner, P. S., PhD Thesis, University of Hull, 1986. Bogdanski, S. L., Henden, E . , and Townshend, A., Anal. Chim. Acta, 1980, 116, 93. "Selected Values of Chemical Thermodynamic Properties," National Bureau of Standards, Washington, DC, 1952. Radioanalytical Studies of Electrothermal Atomisation Atomic Absorption Spectrometry John E. Whitley Scottish Universities Research and Reactor Centre, East Kilbride G75 OQU Richard Hannah and David Littlejohn Department of Pure and Applied Chemistry, University of Strathclyde, Cathedral Street, Glasgow GI IXL The application of radioanalytical techniques for diagnostic purposes is extremely useful in many areas of analytical chemistry.In electrothermal atomisation atomic absorption spectrometry (ETA- AAS) , the insignificant activation of graphite simplifies analyte distribution studies by neutron activation analysis (NAA), although in some instances detec- tion limits can be limited by activation of other constituents. When suitable radioactive tracers with half-lives of less than a few days can be produced, long-term contamination of components is avoided and detection limits better than those obtained with NAA can often be achieved.In ETA-AAS, radioactive tracers have been used to investigate pre-atomis- ation losses, atomisation mechanisms and transport phe- nomena.132 The limits of detection attainable by NAA or with tracers produced at the SURRC (where irradiation for up to 6 h at 3.6 X 1012 n cm-2 s-1 is available) are comparable. Commercially Table 1. Limits of detection (LD) for radioanalytical tracer studies (ng) High specific activity NAAISURRC tracer tracer Isotope LD Isotope LD 2.3 min 28A1 3.8 min -52V 2.6 h 56Mn 10 min 60C~m 14 h 69Znm 120 d 75Se 26 h 76 As 65 h 197 Hg 2.7 d ~YXAU 7 0.5 0.04 5 40 70 130 0.3 0.01 Not available 16 d 4RV 10-4 312 d 54Mn 10-4 5 y6Wo 10-4 244 d 6sZn 10-2 120 d 7sSe 10-3 18 d 74As 10-4 47 d203Hg 0.3 2.7 d 198Au 10-3 available tracers of high specific activity improve the limit of detection, but their half-lives are usually longer and introduce possible contamination of components, which hinders replicate studies.The detection limits for selected elements are given in Table 1. A preliminary study of the use of radioanalytical techniques in ETA-AAS has been conducted by the SURRC and the Analytical Chemistry Research Group, University of Strathclyde. Non-destructive NAA has been applied to the semi-quantitative parametric analysis of graphite strips used in probe ETA-AAS3 and the 76As isotope (t: = 26 h) has been used for tracer studies with a hydride generation - ETA-AAS method35 Semi-quantitative Parametric Analysis of Graphite Probes To establish background levels of trace elements in different types of graphite probes, specimens were irradiated in a flux of 2.3 X 1012 n cm-2 s-1 in the pneumatic transfer system, transferred into clean containers and counted on a 130 cm3 Ge(Li) detector.The amounts of the elements detected were calculated by taking into account the neutron flux and the efficiency of the detector. This parametric approach is appro- priate in instances where elementary standards are difficult to prepare, but is not as accurate as the comparative method. The amounts of elements detected in short radiations of three different probes are shown in Table 2. The microporous glassy carbon (MPGC) probes contained higher levels of Al, Ba, Ti and W than the electrographite material. This may not be of great significance, however, as probe atomisation is not recommended for the determination of these elements by ETA-AAS.In contrast, the individual levels of Mn and V were very similar for all three probes irradiated and the amount of247 ANALYTICAL PROCEEDINGS, JULY 1988, VOL 25 Na in the MPGC probe was smaller than for the electrograph- ite probe. Table 2. Elements observed on neutron activation of graphite probes (ng) Element Isotope 1* 2; 3: 4 min irrudiation- A1 2.3 min26A1 110 24 NDli Ti . . S.8min5lTi 15 46 ND V . . 3.7min'W 8 0.5 0.7 30 rnin irradiation- Na 15h24Na 1s 26 78 Ar 1.8h4'Ar ND 4 22 Mn 2.6h"Mn 0.2 0.5 0.3 Ba 83 min *3yBa 310 1050 ND W 24h187W 42 31 2 * Pyrolytically coated microporous graphite. t As 1 but tip not coated. $ Polycrystalline electrographite.li ND = not detected. A comparison of the results obtained for MPGC probes 1 and 2 suggests that the deposition of a pyrolytic graphite layer on the probe causes an increase in the A1 and V levels, but a decrease in those of Ba and Ti, which was unexpected. Further studies of this type would be useful in assessing the purity of different graphite materials intended for use in ETA-AAS and would permit an evaluation of the contamination of tubes and probes, caused by pyrolytic graphite deposition. One problem in this work that has to be taken into account is the presence of the 41Ar isotope (ti = 2.4 h) (from activation of occluded air) which may hinder the determination of elements with compar- able half-lives. Tracer Studies of Hydride Generation - Graphite Furnace Deposition for the Determination of Arsenic by ETA-AAS The 76As isotope (ti = 26 h) was used to investigate transfer processes in a variation of the method described by Sturgeon et ~ 1 .~ for the determination of As by hydride generation with in situ furnace concentration. Arsine (ASH,) was generated by introducing 5% mlV NaBH4 - 1% m/V NaOH into solutions of 76As tracer in a Perkin-Elmer MHS-10 instrument and the vapour was transferred into a Perkin-Elmer HGA-74 atomiser situated in a Pye Unicam SP-2900 AA spectrometer. The hydride was passed into the atomiser via the internal purge gas supply lines. Entry of the vapour was, therefore, from the ends of the tube. For deposition, the atomiser was set to give a tube temperature of 800-900 "C; for atomisation, the maximum temperature was fixed (about 2600 "C).The radioactive tracer levels were determined in various components of the system, such as the hydride generation reaction cell for the HGA-74 atomiser tube, by counting them in or on a 3 x 3 in well type NaI detector. The levels were related to the activities observed in similar components containing 100 ng of labelled As. The counting of tracer solutions in the generation cell before and after hydride generation showed that the residual activity after generation was 1.4 +0.3.% of the initial activity (n = 6). The efficiency of the collection of As by an uncoated electrographite tube was found to be 24 k 2%, which is much lower than the 100% efficiency reported by Sturgeon etal.5 at a similar deposition temperature.The retention by intermediate components was 0.5%. The distribution of As in a tube was investigated by counting sections of the tube after generation. Most of the As was found to be deposited at the extremities [Fig. l(a)]. A similar distribution of indigenous Na, K and Au was found by NAA. A more even distribution was obtained with a shaped tube [Fig. l(b)], although the tube used for the distribution experiments, illustrated in Fig. l ( b ) , was compara- tively new and the total collection efficiency was low. When an older, shaped, tube was used, the collection returned to 25% and better precision and AAS sensitivity were obtained compared with the results for a standard tube. Tracer studies also indicated that incomplete atomisation of As occurs after decomposition and deposition.This is primarily a limitation of the maximum heating rate and temperature attainable with the HGA-74 atomiser. Conclusions Background levels of tracer elements in graphite probes determined by NAA can probably be explained by the method of manufacture and are indicative of the levels of detection expected for future atomisation studies. The use of a 76As tracer (ti = 26 h) had indicated that a hydride generation - ETA-AAS method exhibits efficient generation, but with incomplete deposition and atomisation of As. Modification of the tube profile improves the distribution of As in the tube, but further work is required to obtained efficient atomisation and to achieve a total mass balance. Fig. 1. 1 2 3 4 5 Segment N o . Distribution of 7hAs in (a) a normal HGA-74 tube and (b) 1 2 3 4 5 Segment No.a shaped HGA-74 tube, after 76AsH3 decomposition and deposition of 'GAS248 ANALYTICAL PROCEEDINGS, JULY 1988, VOL 25 References 3. Littlejohn, D., Lab. Pract., 1987, 36(10), 126. 4. 5 . Sturgeon, R. E., Willie, S . N., and Berman, S. S . , J. Anal. At. Spectrum., 1986, 1, 115. Sturgeon, R. E., Willie, S. N., Sproule, G. I., and Berman, S. S . , J. Anal. At. Spectrom., 1987, 2, 719. 1. 2. Schmid, W., and Krivan, V., Anal. Chem., 1985, 57, 30. Veillon, C., Guthrie, B. E., and Wolfe, W. R., Anal. Chem., 1980, 52, 457. Flame and Furnace; Emission and Absorption-a Historical Dialogue A. M. Ure Department of Pure and Applied Chemistry, University of Strathclyde, 295 Cathedral Street, Glasgow GI 1XL Although the pioneering concept of the spectral line was almost demonstrated by the Glaswegian Thomas Melvill as early as 1756, it was not until the description of the absorption lines in the sun’s spectrum and of the sodium emission lines in a candle flame, given by W.H. Wollaston in 1802 (and named after the later Fraunhofer), that both atomic absorption and emission spectrometry made their appearance on the scientific stage, The work of many other scientists, including a number of Scots, as described by Professor D. Thorburn Burns elsewhere in these proceedings, culminated in the firm assignment of spectral lines to elements by Kirchoff and Bunsen in 1860.1 Thereafter, the development of qualitative and semi-quantita- tive atomic emission spectrography was rapid, as evinced, for example, by the analysis of wines and rocks by Simmler in 1861.2 Nebulisation of the sample solution into the flame did not become an efficient process until the demonstration of a concentric pneumatic nebuliser by Gouy3 in 1879, and this formed the basis for the first quantitative use of flame atomic emission spectrography by Lundegardh4 in 1928.This made use of the air - acetylene flame with pneumatic nebulisation in a technique which was the forerunner of much of the subsequent science of analysis by atomic emission spectrometry (AES) and AAS 1912 1930 I 1939 ,”i” 1960 1970 Woods: Experimental AAS AFS Hg Na Muller, Muller and Pringsheim: Hg AAS Woodson: Hg AAS Walsh: Alkemade and Milatz: Analytical FAAS Well developed FAAS atomic absorption spectrometry (AAS) .Lundegardh also introduced the use of microdensitometry for the measurement of spectral line blackness as well as a method for the correction of flame background emission.5.6 The sensitivity of the technique for five elements is given in Table 1; a precision of around +5% was achieved. This quantitative multi-element emission technique survived in routine use at the Macaulay Institute until about 1950 and in its heyday was widely applied to the agricultural analysis of solutions for the presence of manganese and the alkali and alkaline earth elements. Table 1. Lundegdrdh emission spectrography (air - C1H2) Approximate detection limits/ mg 1 - 1 Cu, 0.3 Ba, 100 Pb, 100 Ag, 0.5 K, 5.0 Lundegardh’s approach formed the basis for the develop- ment of simple flame photometers, particularly for potassium , AES T5 lq49 ’9”’ 1950 1 1960 I 1965 Baum: Total consumption burner Pratt and Larson: interference filters Merton: Replica gratings Ramsay e t a / .: Pt loops in flame Well developed FAES Willis: N20 - C2H2 flame Fig. 1. Evolution of atomic absorption and emission spectrometryANALYTICAL PROCEEDINGS, JULY 1988, VOL 25 249 using photocells and coloured filters. This was the state of the art of AES by 1940. At about this time, Ramage,7 using solid powdered samples rolled in a spill of filter paper and inserted into the flame, introduced what would now be called probe atomisation. Further extension of AES at this time was limited by the technology available. During this period, two develop- ments in AES using non-flame atomisation occurred which were significant pointers to the electrothermal atomiser.These were the seminal work with the electrothermal furnace atomiser by King8.9 in emission spectroscopy and the less well known work of Gatterer,’” who, in the process of purifying carbon rods for arc and spark spectrography by passing high currents through them, realised that the flame produced in this process radiated the impurity spectra which he was able to measure. Retracing our steps to consider the rise of the second partner in the dialogue between AES and AAS, we can see in Fig. 1 the steps in the evolution of analytical atomic absorption spec- trometry beginning with the experimental work of Woods in 1912 in AAS-and also in atomic fluorescence spectroscopy (AFS)-with the elements Hg and Na. This soon led to practical mercury meters”-13 in the 1930s.Here, however, the progress ceased and in the years immediately before the Second World War the analytical spectroscopy was dominated, apart from mercury determination, by Lundegardh’s flame emission spectrography and simple filter flame photometry. Post-War, many of the limitations to the advance of AES were removed, as shown in Fig. 1. New burners appeared, such as the total consumption burner introduced by Baum and, later, new high-temperature flames, such as the revolutionary, reducing, dinitrogen oxide - acetylene flame. This increased the number of elements accessible by AES to elements with higher excitation potentials and with refractory oxides. Improvements in monochromation by means of interference filters and cheap and effective spectrometers based on Mer- ton’s work with replica gratings, together with the revolution in electronics, and in particular the availability of photomultiplier tubes, led to the evolution of successful multi-element flame photometers with detection limits for a wide range of elements in the p.p.m.and sub-p.p.m. range. However, the pendulum swung away from emission back to atomic absorption with the inspirational work of Sir Alan Walsh and of Alkemade and Milatz in 1955. Walsh’s sub- sequent development of practical flame atomic absorption spectrometry (FAAS) revolutionised inorganic analysis. In the 1960s and 1970s the two approaches FAES and FAAS were engaged not only in dialogue but even in acrimonious contention.It was somewhat ironic that the post-war advances in monochromation, which were essential for FAES, accom- modated well the less stringent needs of FAAS. While the advantages of multi-element analysis could be advanced by FAES, the benefits of reduced spectral interferences were cited in favour of FAAS. In terms of sensitivity the arguments were about equally balanced as Table 2, derived from Pickett and Koirtyohann,l4 indicates. Table 2. Comparison of performance of FAES and FAAS. [Adapted from Pickett and Koirtyohann (1969)“q No. of elements More sensitive by Equally sensitive by More sensitive by FAES FAES and FAAS FAAS 24 17 21 Here we must turn to another dichotomy and another competitive dialogue, viz., that of flame versus furnace methods.The history of the development of the electrothermal atomiser, which stemmed from the work of King and was developed by many other scientists including L ’ v o v , ~ ~ Mass- mann16 and West and Williams,l7 is shown in Fig. 2. Again, by about 1970, ETAAS, usually in the form of graphite furnace atomic absorption spectrometry (GFAAS) had emerged as an extremely sensitive analytical technique which required only very small samples. However, because the interference effects were large and not well understood and atomisation mechan- isms were often obscure, it could, at this period, be argued- and it was-that ETAAS lacked the element specificity of GFAAS and was often too inaccurate for the practical analyst without extensive method development. With the pioneering work of L ’ V O V , ~ ~ who pointed out the importance of isothermal atomisation and introduced the L’vov platform, the validity of ETAAS for analysis was established.It was in this era that the late John Ottaway and his group at Strathclyde University contributed a good deal to the fundamental understanding of the process of atomisation.19 From this, practical methods for minimising interference effects evolved, among which the use of probe atomisation in ETAAS by his group was a most important development. In more recent years the acceptance of ETAAS has no longer been in doubt and the development 1905 1908 I I 1959 I 1967 I 1968 I 1969 I I I I 1970 / 1975 1978 King: ETAES L‘vov: Furnace AAS Ulfvarson: Pt Fit. Atomiser Massman: GFAAS West and Williams: Carbon rod atorniser AAS Well developed ETA-AAS L’vov: platform AAS 1975 Ottaway and Shaw: Molnar e t a / .: Furnace AES 1980 Ottaway et a/.: Platform furnace AES Fig. 2. History of the development of the electrothermal atomiser250 story could, therefore, have terminated there. However, a new contender to ETAAS appeared on the horizon to give a new twist to the AES - AAS dialogue. Professor Ottaway’s group,20 and Molnar et ~ 1 . 2 1 in the USA, introduced the concept of ETAES in the form of graphite furnace atomic emission spectrometry. The four techniques evolved so far can be compared in terms of detection limit, as shown in Table 3, which has been adapted from Parsons ef a1.22 and Adams and Kirk b r igh t .23 Table 3. Average detection limits [adapted from Parsons et al.(1983)22]. All values given in ng ml-l FAAS ETAAS FAES* ETAESt 103 1.8 94 86 * N 2 0 - C2H2 flame. t Estimated from Adams and Kirkbright (1976).23 From this comparison it can be concluded that ETAES is not superior to FAES in terms of sensitivity-as might have been predicted by consideration of their maximum excitation temperatures. Once more the late John Ottaway’s g r o ~ p 2 ~ demonstrated, however, that when effective background correction methods, which they devised, were applied to furnace emission, dramatic improvements in detection limit were achieved, as shown elsewhere in these proceedings by Dr. D. Littlejohn. Throughout this discussion I have excluded (from considera- tions of time and space available) the techniques of arc and samples. Nor have I considered the new challenge to these flame and furnace methods arising from the work of Greenfield and Fassel, who have used the inductively coupled plasma as an ANALYTICAL PROCEEDINGS, JULY 1988, VOL 25 emission source. I have, therefore, no doubt that new chapters covering this historical dialogue are still being written.1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. References Kirchoff, G., and Bunsen, R . , Ann. Phys. (Poggendorj? Ann.), 1860, 110, 161. Simmler, R. T., Jahresber. Naturforsch. Ges. Grauhundens Gouy, A . , Ann. Chim. Phys., 1879, 18, 5 . Lundegirdh, H., Ark. Kemi, Mineral. Geol., 1928,10A, No. 1. Lundegirdh, H., “Die Quantitative Spektralanalyse der Elemente,” Gustav Fischer, Jena, 1929.Lundegardh, H . , Z. Phys., 1930, 66, 109. Ramage, H., Nature (London), 1929, 123, 601. King, A. S . , Astrophys. J., 1905, 21, 236. King, A . S . , Astrophys. J., 1908, 27, 353. Gatterer, A., Spectrochirn. Acta, 1944, 2, 252. Muller, IS., Z. Phys., 1930, 65, 739. Muller, K., and Pringsheim, P., Naturwzssenschaften, 1930, 18, 364. Woodson, T. T., Rev. Sci. Instrum., 1939, 10, 308. Pickett, E. E . , and Koirtyohann, S. R., Anal. Chem., 1969,41, 28A. L’vov, B. V., J . Eng. Phys. (USSR), 1959, 2, 44. Massmann, H., Spectrochim. Acta, Part B , 1968, 23, 215. West, T. S . , and Williams, X. K., Anal. Chim. Acta, 1969, 45, 27. L’vov, B. V., Spectrochim. Acta, Part B , 1978, 33, 153. Littlejohn, D., and Ottaway, J. M., Analyst, 1979, 104, 208. Ottaway, J. M., and Shaw, F., Analyst, 1975, 100, 438.Molnar, C. J., Chuang, F. S., and Winefordner, J. D., Spectrochim. Acta, Part B , 1975, 30, 183. Parsons, M. L., Major, S . , and Foster, A. R . , Appl. Spectrosc., 1983, 37, 41 1. Adams, M. J . , and Kirkbright, G. F., Anal. Chirn. Acta, 1976, 84, 79. Littlejohn, D . , and Ottaway, J . M., Analyst, 1977, 102, 553. (1859-60), 1861, N.F. 6 , 194. Some Results of Joint Research Activities on the FANES Technique Between the Academy of Sciences, Berlin, and Strathclyde University, Glasgow Heinz Falk Central Institute for Optics and Spectroscopy, Academy of Sciences of the GDR, Rudower Chaussee 5, 1199 Berlin-Adlershof, GDR Aims of Cooperation Graphite furnace atomic absorption spectrometry (GFAAS), introduced by L’vov in 1959,’ has occupied a prominent position in analytical laboratories undertaking extreme trace element analyses since it became available for routine labora- tories after the work of Massmann in 1976.2 The reason for this is the superiority of the detection limits attainable with GFAAS for most elements compared with those available using other established atomic spectroscopic methods.As a typical monoelement method, GFAAS has its limitations when several elements have to be analysed in one sample. As a consequence, studies have been started in several laboratories on the use of the very efficient graphite furnace atomiser as an atom re$ervoir for emission spectrometry.”4 It was the purpose of these investigations to develop an analytical method with the detection power of GFAAS but additionally with simultaneous multi-element capability and higher dynamic range.Whereas John Ottaway’s group used a thermal excitation of sample vapour generated inside the graphite furnace,s the concept of non-thermal excitation was studied in our labora- tories in Berlin.6 Collaboration between our groups was established in 1980 to exchange instrumental and analytical experience with our methods and to find their limitations.* Thermal versus Non-thermal Excitation The first experiments with graphite furnace atomic emission spectrometry (GFAES) and furnace atomisation non-thermal excitation spectrometry (FANES) were carried out using spectrometers without special means for background correc- tion. The differences between GFAES and FANES sources was essentially the vacuum-tight FANES workhead with an anode to drive a glow discharge at low pressure of about 1 kPa using the atomiser tube as a hollow cathode.The first experiments showed that absolute detection limits in the picogram range were attainable by both GFAES and FANES. However, a striking difference between the two methods has also been found, namely, a degradation of the detection power of GFAES with increasing excitation potential * A programme for joint research activities was settled within the Cultural Exchange Agreement between the Royal Society in the UK and the Academy of Sciences of the GDR.ANALYTICAL PROCEEDINGS, JULY 1988, VOL 25 25 1 of the atomic level used for the analysis."5 In contrast, there was no systematic dependence of the detection limit on the excitation potential for FANES.6 These effects could be understood by considering the excitation mechanisms: GFAES is governed by the thermal equilibrium between the atomiser tube and all components present inside the tube,' whereas in FANES the electron temperature responsible for excitation is much higher than the temperature of the tube wall.' As a consequence, the population of excited levels in GFAES follows a Boltzmann distribution corresponding to the tube wall temperature, which is limited to about 3000 K.Experimental Investigations of the Spectral Background As the ultimate detection limit attainable by AES methods depends on the fluctuations of the spectral background of the excitation source, this problem was studied jointly by our laboratories.For this purpose a high-resolution echelle monochromator incorporating automatic background correc- tion, as developed in Glasgow, was used. According to the thermal nature of the excitation in GFAES there exists a smooth continuous background following Planck's law, which can be efficiently corrected for by using wavelength modula- tion.8 In contrast to GFAES, the FANES background emission is structured as shown by measurements in Glasgow. Moreover, the spectral features of the background change when the tube temperature is increased beyond roughly 1700 K.9 Results of a background study proved that there is no improvement achieved by using wavelength modulation instead of intensity modulation for background correction. As shown by later experiments, wavelength modulation may lead to an improve- ment in the detection limit in FANES if there exists a correlation between the background intensities at the analy- tical wavelength and the spectral position that is used for background correction.10 The requirement for a successful application of wavelength modulation for background correc- tion is a proper choice of the reference wavelength based on a knowledge of the nature of the background spectrum at and near the analytical wavelength. Instrumentally. it must be possible to set each reference wavelength near to the analytical wavelength individually. Peculiarities of FANES Compared with GFAES Essentially, all steps of sample preparation, including drying and ashing, are identical in GFAES and FANES. The atomisation in FANES takes place at low gas pressures, resulting in a shorter residence time of the analyte within the furnace compared with GFAES, which is operated at atmo- spheric pressure.The shorter residence time of sample atoms in FANES implies a corresponding decrease of the atomiser efficiency, typically in the per cent. range.11 As a consequence of the shorter residence time of sample species in FANES, the separation of the analyte vapour from matrix constituents may be more efficient. This fact was successfully exploited when we determined Cd in blood by FANES.12 Another interesting feature offered by the non-thermal plasma in FANES is the higher degree of dissociation of molecular species of the sample present in the gas phase of the atomiser. Molecules with high bond dissociation energies are only incompletely dissociated in a graphite furnace because of its maximum temperature of less than 3300 K.We observed a striking difference between GFAES and FANES with respect to dissociation when we analysed Ga in GaC12. The FANES signal appeared at 800 K whereas GFAAS signals could be observed only above 2000 K. In this instance Ga could be analysed using an atomisation temperature of 1700 K with the detection limit (30) below 1 pg. A full understanding of non-thermal dissociation in FANES and the analytical implica- tions need further investigations. Conclusions The collaboration between physicists in our institute and the chemists at Strathclyde University brought about a better insight into the processes governing new spectroscopic analy- tical methods such as FANES.Moreover, the analytical performance could be improved. Another important reason for continuation of this cooperation is the use of the latest instrumental developments available in our respective labora- tories. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. References L'vov. B. V., Inzh, Fiz. Z h . , 1959, 2. 44. Massmann, H.. Proc. Anal. Div. Chem. Soc., 1976, 13, 258. Falk, H.. Spectrochim. Acta, Part B, 1977, 32, 417. Ottaway, J . M., and Shaw, F.. Appl. Spectrosc., 1977, 31. 12. Littlejohn. D., and Ottaway, J. M.. Analyst, 1979, 104. 208. Falk, H., Hoffmann, E., and Luedke, Ch., Spectrochim. Acta, Part B, 1981, 36, 767. Falk, H., Hoffmann, E., and Luedke, Ch., Spectrochim. Acta, Part B, 1984. 39. 283. Bezur, L., Marshall, J ., and Ottaway, J . M., Spectrochim. Acta, Part B, 1984. 39, 787. Falk, H., Hoffmann, E . , Luedke. Ch., Ottaway, J . M., and Giri. S. K., Analyst, 1983, 108, 1459. Falk, H., Becker-Ross, H., Florek, S . , Hoffmann, E., Luedke, Ch.. and Tischendorf, R . , Proc. Colloq. Atomspektrom. Spurenanalytik, Constanz, 1987, in the press. Falk, H., and Tilch, J . , J . Anal. Ar. Spectrom. 1987, 2, 527. Falk, H., Hoffmann, E., Luedke, Ch., Ottaway. J . M., and Littlejohn, D., Analyst, 1986, 111, 285. The Teaching of Analytical Science into the 21st Century John F. Alder Department of Instrumentation and Analytical Science, UMIST, P. 0. Box 88, Manchester M60 1 QD As well as undertaking both fundamental and applied analy- tical research throughout his life, John Ottaway was a teacher of our scientific discipline.As external examiner of the DIAS MSc course in Instrumentation and Analytical Science at UMIST. we had the opportunity to discuss both his and our plans for the development of analytical science teaching into the future. We benefitted then from his foresight and will no doubt continue to do so. In considering the teaching of analytical science into the future. it is appropriate to look at our recent history and identify the trends that may indicate future directions. Analy- tical chemistry was taught extensively in the period up to the early 1960s as part of inorganic chemistry. The now classical techniques of wet analytical chemistry were an ideal way of teaching chemical properties in the context of contemporary inorganic chemistry.They also provided excellent training in precise and accurate chemical manipulation measurements. With the rapid scientific developments of the 1950s and 1960s, and particularly with the growth in the electronics, petrochem- ical and polymer industries, came new requirements and a new252 ANALYTICAL PROCEEDINGS, JULY 1988, VOL 25 emphasis. The advent of new concepts in the understanding of inorganic chemistry through complexes, ligands, chelates, etc., changed the way inorganic and analytical chemistry was taught. Parallel advances in organic synthesis and analysis completely changed the way this topic was taught too. Equally great and exciting developments in biochemistry and physical chemistry, and the growth of the international electronics and instrument industries, sent a whirlwind through the whole of science in the 1960s and 1970s, leaving us in the 1980s still in a state of high flux.The requirements of modern chemistry teaching are now broader. The demands of modern industry and the new technologies of semiconductors, biotechnology, clinical and environmental monitoring, automatic plant operation, etc., are greater and more exacting than ever before. The teaching of modern analytical chemistry is at a super- ficial level in many, but I must emphasise by no means all, university chemistry courses. The number of university and polytechnic Chairs of Analytical Chemistry in the British Isles is as low now as it was a decade ago, and it was considered to be too low then! In my view, this is not due to sectarian discrimination, but is a reflection of the perceived r61e of the analytical chemist by academics and to some extent by scientists and engineers in manufacturing industry. It is the conflict between the image of the analyst - technician in the “lab” and the analytical chemist in the research laboratory or design and planning team.There is a role and a need for analytical chemists in both of these areas and it is up to us as a profession to ensure that through our teaching, both of these images are projected correctly into the minds of both our students and colleagues. Chemistry is analytical chemistry. Measurement is a prere- quisite of understanding and an essential part of control. All chemical systems need to be understood and controlled, from the most trivial synthesis to the total ecosystem.Never before has the chemical science community had such a need for good chemical analysis. It is equally true to say that engineering is measurement and indeed that science is analytical science; it was always thus. So why do we find ourselves in our present situation both as regards our representation in chemistry departments and therefore in our analytical chemistry teach- ing? The answer is not trivial-indeed the contributory factors are complex. Like many aspects of our lifestyle over the post-war period, we have been overtaken by events. The development of inorganic chemistry undoubtedly weakened the position of analytical chemistry in the pure chemistry university curriculum. Simultaneously with this was the rapid growth of instrumental methods of measurement across a wide front, not all of which in the British Isles originated from the analytical chemistry community.Much of the development in this field was in the USA and became dominated by large multi-national instrument manufacturers such as Perkin- Elmer, Varian, Beckman and Philips. Developments in NMR, MS, X-ray, electron microscopy, chromatography, etc., occurred very quickly, supported by the instrument industry and aimed primarily at the industrial manufacturers and public service users. Quite a lot of this development occurred in parallel to the efforts of analytical chemists in UK universities who could hardly compete with the research teams in the instrument companies and US universities, owing to a lack of financial and personnel resources.Indeed, in the eyes of some of our colleagues, the academic analytical chemist was more or less redundant. This process has, to some extent, slowed down over the last fivc years. The emphasis in instrument development is more towards data handling and microprocessor control of instrumentation. The modern analytical scientist has at his disposal a very wide range of techniques, and equipment of remarkable quality and sophistication, with which and from which to build new analytical methods. There has been a semantic shift since the 1960s. The analytical chemist, as indeed he was, has become the analytical scientist. He has more tools available to do his work now than ever before. He has had to learn more physics, electronics, computing, mathematics and biochemistry than he ever needed to before-but he still has his roots, and indeed his foundations, in chemistry. Be they biochemical, environmental or metallurgical problems he is faced with, the success of the analysis that he carries out will depend fundamentally on his understanding of the chemical properties of the system under analysis and on the physical - chemical principles of the analytical method employed. What is equally important is that the analytical scientist has much more to offer the chemical science and engineering community than ever before.The trend over the last decade is clear and one is readily able to identify a curriculum for the teaching of our science. It must be based firmly on modern chemistry, of that there is no doubt. The teaching has to incorporate physics, mathematics and the elements of biochemistry to be relevant to modern and future needs. Electronics and the design of instruments is now essential reading for the analytical scientist of the 21st century. Data manipulation, statistics, chemometrics and control using computers are important skills to be taught to the modern analyst. Of most importance is the need to cement the structure together to teach the philosophy of analytical science. The student must be taught a systematic approach to analysis, based on a firm understanding of all aspects of the chemistry of the system, knowledge of the principles of measurement available to him, the transformation of chemical activity and properties into electrical and optical signals and how these signals can be interpreted, calibrated and verified. With this training he will be able to undertake the analysis of any problem and work towards an accurate quantification of the important system parameters. He will also play an important r d e in system design and management. Gone are the days of the analyst in a service r61e. But where should this training take place? In a land of plenty, the high schools, departments or institutes of analytical science, of which we dream, could, and indeed may, become concrete. In the reality of the late 1980s, a period of contraction and diminishing resources, we are unlikely to see major resource commitment to this end by either central government or industry. We must continue to operate within our existing frontiers: the secondary schools, polytechnics, colleges and universities. It is important to identify the requirements of both the academic and industrial chemistry community in particular and to ensure that our teaching and research is relevant to both their needs and our’s. This must be a two-way process, particularly at the level of research and teaching. New initiatives in molecular recognition, surface phenomena, sen- sors, catalysis, superconductors, etc., are ideal pastures for fruitful collaboration. Initiatives in other areas, such as process automation, simulation and control, biotechnology and med- ical engineering, offer opportunities to input our philosophy into other fields and benefit from them too. By continuing to operate on this broad basis, the analytical scientists of the 21st century will be able to offer an excellent environment for the training of many generations of world leaders in our most important scientific discipline. John Ottaway believed in this; he was not alone. With his example, the legacy of his contribution to modern analytical science and the inspiration he gave his colleagues, we will continue towards the goal he, and those who have gone before, have set us.

ISSN:0144-557X    

DOI:10.1039/AP9882500240

出版商:RSC    

年代:1988

数据来源: RSC

摘要:

ANALYTICAL PROCEEDINGS, JULY 1988, VOL 25 253 Early Optics and Spectroscopy-The Scottish Dimension* D. Thorburn Burns Department of Analytical and Inorganic Chemistry, The Queen’s University of Belfast, Belfast BT9 5AG It is appropriate in a Symposium to commemorate the life and work of John Michael Ottaway to review the Scottish aspects of particular role of the Royal Society of Edinburgh5 in its early development. The Society’s role was via its Fellows and Honorary Fellows6 and its publishing activities. The period Thomas Melvill in 1752 to William Dittmar’s acquisition of a spectroscope for Anderson’s University (now the University of Strathclyde) which remains there to this day. Key figures and his major research interest, atomic spectr~sopy,l-~ and the 2. 3. reviewed covers the first account of a sharp spectral line by 4.their major contributions are summarised in Table 1. 5. troscopy, 1874-1974,” Proc. Anal. Div. Chem. SOC., 1975, 12, 155. West, T. S., “Some Early British Contributions to Atomic Spectroscopy,” Proc. Anal. Div. Chem. SOC., 1977, 14, 177. Thorburn Burns, D., “Towards a Definitive History of Optical Spectroscopy. Part I. Simple Prismatic Spectra: Newton and His Predecessors,” J . Anal. At. Spectrosc., 1987, 2, 343. Thorburn Burns, D., “Towards a Definitive History of Optical Spectroscopy, Part 11. Introduction of Slits and Collimator Lens Spectroscopes Prior to and Just Post Kirchoff and Bunsen’s Studies,” J . Anal, At. Spectrosc., 1988, 3, 285. Campbell, N., and Smellie, R. M. S . , “The Royal Society of Edinburgh (1783-1983) ,” Royal Society of Edinburgh, Edin- burgh, 1983.Biographies of all those listed in Table I, except for Blair, Ritchie and Swan, may be found in Gillespie, C. C., Editor, “Dictionary of Scientific Biography,” Volumes I-XVI, C. Scribner, New York, 1970-1980. Those for Blair, Ritchie and Swan are in Stephen, L., Editor (Volumes 1-21), Stephen, L., and Lee, S., Editors (Volumes 22-26), and Lee, S., Editor (Volumes 27-43), “Dictionary of National Biography,” Smith and Elder, London, 1885. 6. References 1. Thorburn Burns, D., “One Hundred Years of Atomic Spec- * Full paper will be “Towards a Definitive History of Optical Spectroscopy, Part I11.”3,4 Table 1. Contributors and their major contributions to optical spectroscopy Election to the Royal Society of Name Dates Edinburgh London Contributions to optics and or spectroscopy ThomasMelvill .. . . . . 1726-1753 c. 1750* Prism experiments with flames. Description of sharp sodium yellow Systematic studies of dispersive powers of glasses and liquids. Liquid “Nicol prism. ” Preparation of thin mineral sections for transmission Flame spectra. Solar spectroscopy. Molecular absorption. Photometry. “Ritchie wedge” line aplanatic lens microscopy Polarisation. Optical toys-kaleidoscope and lenticular stereoscope Robert Blair . . . . . .c. 1748-1 828 1786 WilliamNicol . . . 1768-1 85 1 1838 DavidBrewster . . . . . . 1781-1 868 1808 1815 WilliamRitchie . . . . John Frederick William Herschel . . . . William Henry Fox Talbot 1790-1 837 1828 1792-1 871 1800-1877 Hon.1832 Hon. 1858 1813 1831 Flame spectra. Theory of absorption of light First statement of the analytical potential of spectral studies. Polarisation of heat. Concept of a continuous spectrum Reflection prism mount. Detailed flame spectral studies. Detection limit for sodium. Photometry Position of bright lines independent of state of combination, nature of flame or temperature. Reflecting scale. Astronomical and solar spectroscopy. Use of very pure salts for spectroscopic studies Photography JamesDavidForbes . . . . WilliamSwan . . . . . . 1809-1 868 181 8-1 894 1831 1848 1832 - Robert Wilhelm Bunsen 1811-1899 Hon. 1864 (For.) 1875 (For.) 1857, resigned 1874 1862 1882 - Gustav Robert Kirchoff . . 1824-1 887 Hon. 1867 CharlesPiazziSmyth . . . . 1 8 19- 1 900 1846 Detailed high-dispersion solar spectra and spectra of gases BalfourStewart .. . . . . PeterGuthrieTait . . . . William Dittmar . . . . 1828-1 897 1831-1901 1833-1892 Hon. 1878 1861 1863 Theoretical account of adsorption and emission of light Polarisation spectroscope of great dispersion Acquisition of spectroscope for Andersonian University * Philosophical Society of Edinburgh, precursor of Royal Society of Edinburgh. Ramblings Through the Landscape of Analytical Atomic Spectroscopy Thomas C. O'Haver Department of Chemistry and Biochemistry, University of Maryland, College Park, MD 20742, USA The life and career of Professor John M. Ottaway was distinguished by his great ability to bring people together and to catalyse their interaction both scientifically and socially.Professor Ottaway's laboratories in Glasgow were continu-254 ANALYTICAL PROCEEDINGS, JULY 1988. VOL 25 ously host to a succession of visiting scientists from all over the world. As I try in this paper to give an overview of some of the projects of John's group in which I came to be involved, the international scope of the cross-fertilisation becomes obvious. This is a clear illustration of the old adage that we are able to reach new heights only because we are standing on the shoulders of others. John's work was also characterised by an ability to see things in a fresh perspective and a willingness to question the conventional wisdom. Quite a number of new ideas in analytical chemistry and spectroscopy were investigated first by John's group or with his encouragement.Not every new path becomes a well-travelled road, but certainly the trail- blazers give all of us a clearer vision of the landscape. I first encountered John Ottaway's work in the literature in 1975, when Mike Epstein, who was then a graduate student in my group, began to look at the possibility of measuring atomic emission in electrothermal atomisers. We thought we might be able to do alkali metals and possibly a few other favourable metals by emission rather than the standard absorption measurement. We soon found that this had already been demonstrated by Ottaway and Shaw,lJ so we had to make some last-minute changes in strategy, emphasising our use of a wavelength modulation technique to provide a continuous correction for the time-dependent background emission sig- nal.3 Wavelength modulation had been used previously by several investigators for the purpose of obtaining derivative spectra, normally of molecular spectra.There are various ways of achieving modulation of the wavelength of a dispersive spectrometer, most requiring modification and careful realign- ment of the optical path of the spectrometer. One of the simplest was the oscillating refractor plate technique first applied to atomic emission spectrometry by W. Snelleman4 from the University of Utrecht in the Netherlands, while he was a Visiting Scientist at the National Bureau of Standards working with Ted Rains and Oscar Menis. (The original spectrometer and refractor plate built by Snelleman is still being used at NBS.) Mike Epstein and I had subsequently published a modification and simplification of that technique ,5 which we then applied to the carbon furnace emission measurements.The result was a substantial improvement in detection limits for most elements. Several years later the technique was used to good effect by the Ottaway group and is still used routinely in carbon furnace emission measurements. I first met John Ottaway in person at the 1976 FACSSICSI meeting in Philadelphia, where we both gave papers on our work in graphite furnace emission spectrometry. One of John's students, Bob Hutton, stayed in the States for a few weeks after the meeting to work on the wavelength modulation technique at the National Bureau of Standards with Ted Rains and Mike Epstein, who by this time had graduated and was working at the Bureau.At this stage in our work we thought of the carbon furnace emission technique as primarily a method for alkali metals or as a back-up method for the odd element for which a hollow- cathode lamp was not available. It never occurred to us to exploit the technique more generally. We would never have thought it would be possible to achieve the great improvements in detection limits and expansion of the scope of elemental coverage which were accomplished by the Ottaway group in later years. Our attention had already turned to the idea of continuum source atomic absorption, which we felt would have better detection limits for most of the elements than emission measurements in the same atomisers. The reason is that in atomic absorption it was possible to utilise a primary light source that has a much higher black-body temperature (e.g., a 5000°C xenon'arc lamp) than the temperatures achievable in carbon furnaces or chemical flames (3000 "C).We therefore expected a substantial advantage in detection capability, especially at short wavelengths. In the context of the modern practice of atomic absorption spectrometry using conventional commercial instrumentation, the notion of continuum source atomic absorption is novel and somewhat contrary to established ideas. Historically, however, the first use of atomic absorption as a measurement technique, made in the 19th century by J. N. Lockyer, was based on continuum sources.6 This interesting fact was uncovered by John Ottaway7 years after continuum source atomic absorption had been rediscovered.This rediscovery, initiated by Gibson, Grossman and Cooke in 1962,* occurred only after Walsh's line-source method had been adopted as the standard tech- nique. In the 1960s and 1970s, a number of groups flirted briefly with the continuum-source idea, but none of these systems proved practical. Two of the key developments introduced separately during this period were the use of wavelength modulation by Snellemang and the use of a high-dispersion Echelle spectrometer by Keli her and Wohl- ers.1" We were attracted to both these ideas and sought to combine them in one system. In about 1975, we began a collaboration with Peter Keliher at Villanova University, which involved my student Andy Zander driving up to Villanova once or twice a week.The result of this work was to establish that the combination of an intense continuum source, a high dispersion spectrometer and a wavelength modulation detection system could achieve detection limits which were substantially superior to previous continuum source AA systems and which began to approach the performance of conventional line source instruments. * I Our interest soon began to turn to the potential of the continuum source system to perform simultaneous multi-ele- ment measurements. Line source AA has developed exclus- ively as a single-element method, and although various emission methods could operate effectively in simultaneous multi-element mode, AA was still widely favoured for its operational simplicity and its ability to utilise the highly sensitive electrothermal atomisers which were becoming popular at that time.Earlier pioneering attempts to develop multi-element atomic absorption systems, based on multiple line sources12 or on continuum sources,13 had not resulted in practical systems. Wayne Wolf, of the USDA Nutrient Composition Laboratory, became interested in the possibility of utilising a multi-element continuum-source AA system for the analysis of nutritionally significant trace elements in foods and foodstuffs. Through him we were able to obtain funding from USDA to develop the first prototype of what came to be known as SIMAAC (Simultaneous Multielement Atomic Absorption with a Continuum source). The construction was begun in my laboratories and the system was then moved to the nearby Nutrient Composition Laboratory where it could be interfaced to a laboratory minicomputer.14 The development of this system and its application to important practical analytical problems became the thesis projects of Jim Harnly, Nancy Miller-Ihli and Sue Lewis. One of the fundamental problems that had to be solved in developing a practical multi-element AA system is the problem of dynamic range. Atomic absorption, and indeed all absorp- tion spectrometric methods, are traditionally considered to be restricted to a dynamic range of only about three decades. The performance at high concentrations, where the absorbance ex- ceeds one or two absorbance units, is limited by the failure of Beer's Law. In simultaneous multi-element applications this limitation is likely to be significant, because of the large differences between the concentrations of different elements in the same sample.For example, much biological material contains large concentations of Na, K, Ca, and Mg, but only traces of Cr, V, Mo and Ni. A sample preparation and dilution procedure that yields acceptably high concentrations for the trace constituents is likely to be too concentrated in the major elements. Initially, we sought to minimise this problem by using a less sensitive line for some elements (e.g., the 404 nrn potassium line) and by committing two spectrometer channels to some elements, each tuned to a line of different sensitivity.ANALYTICAL PROCEEDINGS. JULY 1988. VOL 75 However, this approach reduces the number of different elements which can be measured simultaneously and is not a general-purpose solution for all elements and for all sample types.Some elements simply do not have useful less-sensitive lines (e.g., zinc), and for some elements (e.g., magnesium) the secondary lines fall at inconvenient wavelengths. A more general solution to this problem was to use the spectral information which is uniquely provided by a continuum-source measurement. By taking intensity readings while the wavelength modulator is tuned to the sides of the atomic absorption line, we could effectively adjust the sensitivity of measurement to suit the concentration. The computerised wavelength system allowed this to be done on a simultaneous multi-element basis, without sacrificing the double-beam dynamic background correction capabilities, without suffering a significant signal-to-noise degradation at low concentrations, and without knowing the approximate analyte concentration in each sample beforehand.15 In 1980 we began a long and fruitful direct collaboration with John Ottaway’s group at Strathclyde.Through a three-way co-operation between Strathclyde University, the USDA Nutrient Composition Laboratory and the University of Maryland, John sent several of his graduate students to the States to work with us for a period of several months. The first of these was John Marshall, followed by John Carroll and David Halls. In these projects we combined the Strathclyde group’s experience with novel atomisation schemes with our computerised data acquisition technology.The SIMAAC instrumentation made it practical to investigate the simul- taneous multi-element potential of carbon furnace atomic emission16 and probe atomisation. 17 The probe atomisation technique, based on Boris L’vov’s concept of the ideal conditions for electrothermal atomisation ,18 had been de- veloped by the Strathclyde group19 in the form of a sample insertion probe attachment for conventional carbon furnace atomisers. The concept of the probe was to delay the vaporisation of the sample until the surrounding tube wall and fill gas had reached a constant temperature. Numerous studies had shown that probe atomisation could reduce interferences in many instances. The original SIMAAC instrumentation employed a complex and rather expensive laboratory minicomputer to acquire and reduce multi-element analytical data, whereas previous single- channel systems had used analog electronics. The use of a computerised data acquisition system provided many useful benefits in addition to the ability to handle multi-channel data.By the early 1980s it had become common practice to employ low-cost 8-bit and 16-bit personal computers for simple laboratory data acquisition applications. In a series of visits to Strathclyde in 1984-1986, Jim Harnly and I developed a personal computer data station for wavelength modulation atomic emission and absorption spectrometry, based on the inexpensive Apple 11 microcomputer.2(&22 Subsequently, a more advanced data station was developed based on the IBM-PC AT,21 with software written in PASCAL.These systems provided a convenient way to study aspects of background correction and peak width measurement in electrothermal atomisation and have been applied to carbon furnace atomic emission and absorption and to the FANES (Furnace Atomic Non-thermal Emission Spectrometry) technique originated by Heinz Falk of the GDR Academy of Sciences. The ability of the continuum-source method to measure the absorption spectra of atoms and molecules in the analytical volume has lead to some interesting studies. One of the projects in which I got involved in the summer of 1986 was an investigation of the background absorption caused by the addition of a high concentration of Mg(N03)2 as a matrix modifier in carbon furnace atomic absorption.24 Measure- ments of the wavelength dependence of the Mg(N03)2 background absorption showed that the major spectral feature was a broad band of roughly Lorentzian shape centred exactly 255 at the Mg resonance wavelength of 285.2 nm.For the atomisation of 20 1-11 of 1% rn/VMg(NO&, this band had an apparent FWHM of about 15 nm, so it would be a major source of background absorption for the determination of Mn (279.5 nm), Pb (283.3 nm), Sn (286.3 nm) and Ga (287.4 nm). Our hypothesis was that this band was caused by the line wings of the Mg resonance line. Model calculations showed that the observed data could easily be explained by a theoretical line profile model based on the assumption of complete atomisa- tion of Mg and an approximate collisional line width of 0.002 nm. The general implication of this study is that atomic absorption, as well as molecular absorption and light scatter- ing, contributes to background absorption. Strong resonance lines of major constituents may have an influence many nanometres away from their line centres.Of course, such background absorption is at least approximately corrected by conventional deuterium or Zeeman systems. Nevertheless, excessive background absorption leads to errors and loss of precision, so it is common practice to seek analytical conditions which minimise background absorption. Thus, it is especially important to know whether the background absorption arises from atoms or from molecules or unvaporised particles; atomisation conditions that give more complete atomisation, although generally desirable in reducing matrix effects and the background from molecules and particles, will not, of course, reduce the atomic absorption from the matrix constituents.1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. References Ottaway, J. M., and Shaw, F., Analyst, 1975, 100, 438. Ottaway, J. M., and Shaw, F., Anal. Lett., 1975, 8, 911. Epstein, M . S . , Rains, T. C., and O’Haver, T. C., Appl. Spectrosc., 1976, 30, 324. Snelleman, W., Rains, T. C., Lee, K. W., Cook, H. D., and Menis, O., Anal. Chem., 1970, 42, 394. Epstein, M. S . , and O’Haver, T. C., Spectrochim. Acta, Part B, 1975, 30, 135. Lockyer, J. N., “Studies in Spectrum Analysis,” D. Appleton & Co., London and New York, 1878. Marshall, J., Ottaway, B. J., Ottaway, J.M., and Littlejohn, D., Anal. Chim. Acta, 1986, 180, 357. Gibson, J. H., Grossman, W. E., and Cooke, W. D . , “Proceedings of the Feigl Anniversary Symposium,” Elsevier, Amsterdam, 1962. Snelleman, W., Spectrochim. Acta, 1968, 23, 403. Keliher, P. N., and Wohlers, C. C., Anal. Chem., 1974, 46, 682. Zander, A. T., and O’Haver, T. C., and Keliher, P. N., Anal. Chem., 1976, 48, 1166. Mavrodineanu, R., and Hughes, R. C., Appl. Optics, 1968,7, 108. Furuta, N., Haraguchi, H., and Fuwa, K., Anal. Chem., 1977, 49, 1263. Harnly, J. M., O’Haver, T. C., Golden, B., and Wolf, W. R., Anal. Chem., 1979, 51, 2007. Harnly, J. M., and O’Haver, T. C., Anal. Chem., 1981, 53, 1298. Marshall, J., Littlejohn, D., Ottaway, J. M . , Harnly, J. M., Miller-Ihli, N. J., and O’Haver, T. C., Analyst, 1983, 108, 179. Carroll, J., Miller-Ihli, N. J., Harnly, J. M., Littlejohn, D., Ottaway, J . M., and O’Haver, T. C., Analyst, 1985,110, 1153. L’vov, B. V., Spectrochim. Acta, Part B, 1978, 33, 153. Giri, S. K., Littlejohn, D., and Ottaway, J. M.,Analyst, 1982, 107, 1095. Marshall, J., Carroll, J., Littlejohn, D . , Ottaway, J. M., Harnly, J . M., and O’Haver, T. C . , Anal. Proc., 1985,22,67. O’Haver, T. C., Harnly, J. M., Marshall, J., Carroll, J . , Littlejohn, D., and Ottaway, J. M . , Analyst, 1985, 110, 451. O’Haver, T. C . , and Ottaway, J. M., “Personal Computer Data Station for Atomic Spectrometry with Wavelength Modulation,” Paper #611, 14th FACSS Meeting, Detroit, MI, USA, 1987. O’Haver, T. C., Carroll, J., Nichol, R., and Littlejohn, D., J. Anal. At. Spectrom., 1988, 3 , 155. Macdonald, L. R., O’Haver, T. C., Ottaway, B. J., and Ottaway, J. M., J. Anal. At. Spectrom., 1986, 1, 495.

ISSN:0144-557X    

DOI:10.1039/AP9882500253

出版商:RSC    

年代:1988

数据来源: RSC