摘要:

Proceedinas m - - - ~of the Analytical Division ofThe Chemical SocietyCONTENTS257 Honorary Degrees257 Regional Advisory Editor ofThe A nalyst258 Summaries of Papers: AnnualChemical Congress258 'Developments in Analytical AtomicSpectrometry'291 AD Distinguished Service Award291 Silver Medal292 Fourth SAC Conference292 Euroanalysis 111294 Analytical Division DiaryVolume 13 No 9 Pages 257-294 September 197PADSDZ 13(9)257-294(1976)ISSN 0306-1 396September 1976PROCEEDINGSOF THEANALYTICAL DIVISION OF THE CHEMICAL SOCIETYOfficers of the Analytical Divisionof the Chemical SocietyPresidentD. W. WilsonHun. SecretaryP. G. W. CobbSecretaryMiss P. E. HutchinsonHun. Treasurer Hon. Assistant SecretariesJ. K. Foreman D. I . Coomber, O.B.E.; D.C. M. Squirrel1Editor, ProceedingsP. C. WestonProceedings is published by The Chemical Society.Editorial: The Director of Publications, The Chemical Society, Burlington House, London, W1 V OBN.Telephone 01 -734 9864. Telex 268001.Subscriptions (non-members) : The Chemical Society, Publications Sales Office, Blackhorse Road, Letch-worth, Herts., SG6 1 HN.Non-members can only be supplied with Proceedings as part of a combined subscription with The Analystand Analytical Abstracts.@ The Chemical Society 1976A Meeting of the South East RegiononChemistry in the Packaging IndustryPI RAonOctober 28th, I976will be held atThe following papers will be presented at this whole-day meeting-"The Role of PIRA and the Facilities and Services it Offers to Industry," byF. A. Paine (Director of Packaging Division, PI RA)."Matching the Packaging to the Product," by H. Adcock (Head of AnalyticalDivision, PI RA)."The Determination of Traces of Vinyl Chloride in Various Media," By J. T. Davies(Metal Box Limited)."Some Aspects of Chemistry in Printed Packaging Manufacture," by E. W. Peacock(Printing Services Manager, PI RA).There will be opportunity for participants of this meeting to familiarise themselveswith the facilities and services available at PIRA.Further information can be obtained from Dr. John Warren, Laboratory of theGovernment Chemist, Cornwall House, Stamford Street, London, SEI 9NQ

ISSN:0306-1396    

DOI:10.1039/AD97613FX029

出版商:RSC    

年代:1976

数据来源: RSC

摘要:

292 FOURTH SAC CONFERENCE PYOC. Analyt. Div. Chem. SOC.Analytical Division Diary, continuedOctober, continuedPleizavy Lectuve : “Is Optical Emission Spec-trometry Being Tailored to the Kequire-ments and Budgets of a Large Variety ofUsers ?” by P. W. J . M. Boumans.“Fast Optical Emission Analysis of Metals-Two Points of View,” by K. Slickers.“Improving the Performance and Speed ofAnalysis by Triggered Capacitor Dis-charges,” by N. Kemp.“Some Aspects of Spectroscopy with Induc-tively Coupled Plasmas,” by S. Greenfield.“Some Recent Studies with a MicrowavePlasma Source of Multi-element lktectionSystems,” by G. F. Kirkbright.“Development of Glow Discharge EmissionSpectroscopy for Application in the SteelIndustry,” by A. Butterworth.Cleveland Scientific Institute, CorporationRoad, Middlesbrough.Thursday, 28th, 10.30 a.m.: LeatherheadSouth East Region on “Chemistry in thePackaging Industry.”“The Role of PIRA and the Facilities andServices it Offers to Industry,” by F. A.Paine .“Matching the Packaging to the Product,” byH. Adcock.“The Determination of Traces of VinylChloride in Various Media,” by J. T. Davies.“Some Aspects of Chemistry in PrintedPackaging Manufacture,” by E. W.Peacock.Research Association for the Paper and Board,Printing and Packaging Industries (PIRA),Randalls Road, Leatherhead, SurreySeptembev, 1976 ANALYTICAL DIVISION DIARY 293Analytical Division Diary, continuedSeptember, continued“West European Marketing for LaboratoryAnalytical Instruments,” by J.Drake.“Marketing Research in Process ControlInstrumentation,” by K. Carr-Brion.“Who Needs Automated Polarography ?” byT. Ryan.“The Contribution of P R to the Marketing ofInstrumentation,” by G. Mann.“Who is Wearing the Clinicians Trousers ? ” byJ. S. L. Fowler.“A Small Company Approach to MarketResearch in Automation,” by D. M.Hendry.“The Problems of Training and Education forAutomation,” by D. M. Browning.“The Use of Company Resources in MarketResearch,” by J. Boother.“The Role of the User in Instrument Develop-ment,” by A. W. Hough-Grassby andD. R. Deans.“Instrument Designs for Water QualityMeasurement,” by W. Lancaster.“The Design of Gamma Counters for theRadioimmunoassay Laboratory,” byB. Lumb.The meeting will include an InstrumentExhibition.The University, York.OCTOBERFriday, lst, 7 p.m.: DurhamNorth East Region.Dinner and talk on some aspect of qualitycontrol by R.L. Stephens.Royal County Hotel, Durham.Wednesday, 6th, 2.30 p.m.: LondonAnalytical Division on “Standards.”“Standards and thc Community Bureau ofReference,” by W. Van der Eijk.“Standards-the Aims and Development ofthe Analytical Methods Committee,” byProfessor E. Bishop.“Official Reference Standards for MedicinalSubstances,” by Mrs. S. Richens. TheLinnean Society, Burlington House,Piccadilly, London, W. 1.Tuesday, 12th, 6.30 p.m. : WolverhamptonMidlands Region, jointly with the Birming-ham and West Midlands Section of the CSon “Environmental Protection-TheAtmosphere. ”“A Consideration of the Contribution theMeteorologist May Make to the Protectionof the Atmosphere,” by E.T. Stringer.“Sampling and Analysis for Compliance withPresent and Possible Future Legislation inAir Pollution,” by S. C. Wallin.The Polytechnic, Wolverhampton (entrancein Stafford Street).Friday, 15th, 6.30 p.m.: PrestonPolytechnic Chemical Society.North West Region, jointly with the l’reston“Advances in Forensic Science,” by C. Wood.The Polytechnic, Preston.Educatioiz and Training Group.Discussion on “What Should be Taught inAnalytical Chemistry to Undergraduates ?”to be introduced by W. H. C. Shaw andC. Whalley.Frankland Building, Department of Chem-istry, The University, Birmingham.Wednesday, 20th, 2.15 p.m.: BirminghamThursday, 21st, 2.30 p.m.: LondonBiological Methods Group on “Replacement ofthe Therapeutic Substances Regulations :Compendium on Biological Substances.”Presented by a panel of speakers from theDepartment of Health and Social Securityand the National Institute of BiologicalStandards and Control, headed by J. A.Holgate.Dale Suite, National Institute for BiologicalStandards and Control, Holly Hill, London(nearest tube station Hampstead) .Thursday, 21st, 2 p.m.: LondonJoivlt Pharmaceutical Analysis Group on“New Techniques in Thin-layer Chromato-graphy. ”“The Latroscan THlO-An Advance inQuantitative TLC,” by A. Stafford.“The Application of Mass Spectrometry in theIdentification of Drug Impurities Resolvedby TLC,” by B.Millard.“The Technique and Application of Thin-layerGel Filtration,” by E. A. Hill.The Hall, Pharmaceutical Society of GreatBritain, 1 Lambeth High Street, London,SE1 7JN.Friday, 22nd, 5 p.m.: ExeterWestern Region, jointly with the PeninsulaSection of the CS.“Carbon Furnace Atomic-absorption and-emission Spectrometry,” by J. M. Ottaway.Newman Building, The University, Exeter.Thursday, 28th, 10 a.m. : MiddlesbroughNovth East Region and Atomic SpectroscopyGroup on “Sources for Emission Spectro-scopy. ’ ’Introduction by A. A. Smales.[continued on p. 29An alytica I DivisionSEPTEMBERMonday to Friday, 20th to 24th: SalfordThevmal Methods Gvoup : First EuropeanPlenary lecturers: F. Paulik, J . Zsako andThe University, Salford.Symposium on Thermal Analysis.L.S. Bark.Tuesday to Thursday, 21st to 23rd: SheffieldCS A u t u m n Meetivlg. Symposium organised bythe AD on “Justification of Computers inAnalytical Chemistry.”“Routine Analysis-Computers to Match theJob,” by F. Farren.“Computers in the. Laboratory-A Sympa-thetic Approach,” by I. Telford.“Computer Methods for Data Cleaning andPattern Matching Applied to Mass Spectro-metric Appearance Potential Measure-ments,” by H. F. Tibbals and M. Jones.“Computer Applications to Copolymer Ana-lysis,’’ by R. C. Austin and A. F. Johnson.“The Application of a Minicomputer-basedSystem to Problems of Analytical Signific-ance,” by R. Parsons and R. M. Reeves.“Trends in Computer Technology and TheirInfluence on Analytical Chemistry,” byE.L. Dagless.“Application of Computers in an IntegratedSteel Plant,” by L. Gwilliam.Wednesday, 22nd- Joint Discussion organ-ised by the Education and Training Groupof the AD and the Chemical InformationGroup of the CS.The University, Sheffield.Thursday, 23rd, 10.30 a.m. : LoughboroughPavticle Size Analysis Gvoup, jointly with theIndustrial Sub-committee of the FaradayDivision on “Characterisation of VerySmall Particles. ’’“The Importance of Particle Size in IndustrialApplications,” by Th. F. Tadros.“Sub-micrometre Particles : A Review ofAvailable Techniques,” J . H. Nobbs.“Large Molecules and Small Particles,” byJ . H. Perry.“Flocculation Studies Using an AutomatedLaser Illuminated Particle Counter,” byL.G. Thompson.“Magneto-optical Techniques for StudyingColloidal Particles in Suspension,” byG. H. Meeten.“Sub-micrometre Particle Size Characterisa-tion and Distribution by Mercury Penetra-tion,” by N. G. Stanley-Wood.“Evaluation of the Sartorius SpectralCounter,” by B. Scarlett, R. Buxton andC. R. G. Treasure.Electrical Engineering Department, Univer-sity of Technology, Loughborough.Tuesday, 28th, 10 a.m.: LondonElectvoanalytical Gvoup, jointly with theElectrochemistry Group of the CS on“Students Research Topics in Electro-analysis and Electrochemistry. ”“Polarography of Agrochemicals,” by R. R.Rowe and W. F. Smyth.“Detection Limits of Calcium Ion-selectiveElectrodes in Relation to Ligand-contain-ing Systems,” by B.J . Birch, A. C. Craggs,G. J . Moody and J . D. R. Thomas.“Polarographic Studies of Some ForeignOrganic Compounds in the AqueousEnvironment,” by B. J. Birch, J. P. Hartand W. F. Smyth.“Metal - Metal Oxide pH Electrodes,” byD. Jones.“Two-dimensional Nucleation and Growth-Impedance - Frequency Response,” byT. Dicltinson and Miss A. A. Metcalf.“Electrochemistry of Water in Molten Nit-rates,” by D. Lovering, R. M. Oblath andA. K. Turner.“Galvanostatic Studies of ElectrochemicalNucleation,” by G. K. Hills, G. A.Gunwardena and Mrs. I. Montenegro.Titles of short contributions (about 10 min)should be sent to Dr. B. J. Birch, UnileverResearch Laboratory, Port Sunlight,Cheshire, L62 4XN, from whom further detailscan be obtained.Chelsea College, Manresa Road, London,S.W.3. Please note change of date.Thursday and Friday, September 30th andOctober 1st: YorkAutomatic Methods and RadiochemicalMethods Gvoufis on “Market Research andInstrumentation.”PZenav.y Zectuve: “Supply and Demand ofInstruments in Clinical Chemistry -AProblem of Communications ?” by J .Bierens de Haan.Plenavy Zectuve : “Commercial Developmentand Licensing of New Instruments,” byW. H. Barber.“Role and Concept of Marketing Research atTechnicon lnterna tional Division, ’’ byC. Studievic.[continued inside back coverPrinted by Heffers Printers Ltd Cambridge Englan

ISSN:0306-1396    

DOI:10.1039/AD97613BX031

出版商:RSC    

年代:1976

数据来源: RSC

摘要:

Vol. 13 No. 9 Proceedings September 1976 of the Analytical Division of the Chemical Society 257 Honorary Degrees Honorary DSc degrees have recently been awarded to analytical chemists by British Universities. On July 8th Professor K. Belcher of Birmingham University received an honorary degree at the Queen’s University of Belfast, and on July 9th Professor I. P. Alimarin of the M. V. Lomonosov State University, Moscow, who was elected as an Honorary Member of the former Society for Analytical Chemistry in 1968, received an honorary degree at Birming- ham University.Regional Advisory Editor of The Analyst Dr. J . RtiiiEka has recently been appointed as a Regional Advisory Editor for The Analyst. Biographical notes are given below. Jaromir RdiiEka was born in 1934 in Prague, Czechoslovakia and he received his education there.He studied general and analytical chemistry in the Department of Analytical Chemistry of the Faculty of Sciences a t Charles’ University, in the school of Professor Tomicek and Professor Zyka. In 1957 he graduated from there with a Diploma Thesis on “Polaro- graphy in Glacial Acetic Acid,” part of which was awarded by Professor Heyrovskq, and he was then employed, until 1968, a t the Faculty of Technical and Nuclear Physics at the Tech- nical University of Prague, working mainly in the field of radioanalytical chemistry.During that period he obtained his CSc degree (in 1963) and also the degree of Doctor of Natural Sciences (at the Charles’ University in 1967). He suggested and developed, together with J .Starg, the sub-stoicheiometric method of analysis, for which he was awarded the State Prize in 1965. The method was adopted in a number of radiochemical laboratories and RfiiiCka himself spent one year (in 1965) a t the University of Aston in Birmingham, start- ing there, together with Mo Williams, the research in radiochemistry. After a short period as a lecturer he became a Docent in analytical chemistry at Chemistry Department A of the Technical University of Denmark, where he now teaches basic and instrumental methods of analysis.He has also served as a Secretary and subsequently President of the Danish Analytical Society, as an expert of the UN International Atomic Energy Agency a t the University of Sao Paulo in Brasil, and is presently active in the Danish International Development Agency of the Danish Ministry of Foreign Affairs.His research interest has been mainly in the field of ion-selective electrodes and, together with his co-workers a t Chemistry Department A, he was responsible for the development of the Selectrode and the air-gap electrode. He is currently working on the automation of chemical analyses, and the technique of flow injection analysis, suggested and developed together with E. H. Hansen, now occupies most of his research time. RdiiCka is co- author of a monograph on sub-stoicheiometry and has had published almost 100 papers. RdiiEka, who is married with two daughters, is now a Danish citizen. Dr. J . RDiiEka 257

ISSN:0306-1396    

DOI:10.1039/AD9761300257

出版商:RSC    

年代:1976

数据来源: RSC

摘要:

258 DEVELOPMENTS IN ANALYTICAL ATOMIC SPECTROMETRY ROC. Analyt. Div. Chew. SOC. ANNUAL CHEMICAL CONGRESS The Annual Chemical Congress of The Chemical Society and Royal Institute of Chemistry was held at the University of Glasgow from April 5th to 9th, 1976. Two Symposia were organised by the Analytical Division and summaries of some of the papers given at Symposium A, Develop- ments in Analytical Atomic Spectrometry, are presented below.Summaries of the papers given at Symposium B, Analysis in the Energy Industries, are expected to appear in the October issue. The Fifth Theophilus Redwood Lecture by Professor R. Belcher appeared in full in the June issue of Pyoceedings (p. 153). Developments in Analytical Atomic Spectrometry The following are summaries of nine of the papers presented at the Annual Chemical Congress on April 6th and 7th, 1976.State of Development of Atomic-absorption and Atomic- fluorescence Spectrometry with Furnaces-Plenary Lecture H. Massmann Institut fur Spektrocheunie und Angewandte Spektroskopie, 46 Dortunund, West Gevmany For the measurement of atomic absorption and atomic fluorescence, the sample must be con- verted into atomic vapour as completely as possible.Normally flames are used for this purpose. With such techniques about 70 elements can be directly determined, and analysis can be carried out with high precision and accuracy. Unfortunately, the power of detection of both analytical procedures is not high, especially if compared with some of the methods of emission spectral analysis using arcs.The reason is that solid samples must first be dissolved and, as a consequence, the solid sample is diluted by a factor of nearly 100. If the dissolved sample is nebulised and sprayed into a flame, further dilution occurs as the sample mixes with the gases of the flame. Under such conditions, the concentration of the analyte element in the absorption volume is very small.In order to avoid some disadvantages of the flame techniques in atomic-absorption and atomic-fluorescence spectrometry, atomisers without flames have been used. At present, the most important of such atomisers in atomic-absorption and atomic-fluorescence spectrometry are still furnaces of different kinds. They are suitable for the detection and determination of the same elements that can be determined with flames, but in most instances the latter methods are more sensitive and have a higher power of detection.They are especially important if very small amounts of an element are to be determined. Types of Graphite Furnace Furnace methods have been used in atomic spectroscopy since the first decade of this century, dating back to the well known carbon furnace of Kir~g.l-~ The first application of a graphite furnace to the quantitative determination of element concentrations by atomic-absorption spectrometry (1959) was reported by L’VOV.~-~ L’vov employed arc atomisation in combina- tion with a heated graphite tube.The liquid sample is deposited on the end of a carbon electrode and evaporated to dryness, this electrode is introduced into a hole in the middle of a graphite tube and the tube is then heated by passing an electrical current through it.Simul- taneously, the electrode is heated externally by a d.c. carbon arc in order to volatilise the sample into the hot tube. To prevent combustion of the graphite tube, the furnace device is contained in a chamber filled with an inert gas, e.g., argon. This furnace is very complicated, even after subsequent simplification^.^ A series of simpler types of furnace have now become common in analytical practice, in which their essential parts are related to the graphite-tube furnace described by Massmanns-lo in 1965 and the carbon filament described by West and Williamsll in 1969.A selection of common types of graphite furnace used in atomic-absorption and atomic- fluorescence spectrometry is shown in Fig.1 ; only the main part of each furnace is shown. InH Fig. 1. Types of graphite furnace in atomic-absorption (A,4) and atomic-fluorescence (AF) spectrometry. The optical axis is indicated in each instance. all instances the furnace is heated by resistance heating. Temperatures of 3 000 "C or higher can thus be reached within a few seconds.The shape of the furnace is different, depending on whether absorption or fluorescence is measured for analysis. Furnaces of types (l), (2), (4), (6) and (8) have been developed for atomic-absorption measurements, and (3), (5) and (7) mainly for atomic-fluorescence measure- ments. In order to prevent combustion of the hot graphite, the larger furnaces, types ( l ) , (6), (7) and (8), are normally used in a chamber filled with an inert gas, whereas the smaller furnaces are only shielded by a stream of an inert gas.It is about 1-2 pl for the small furnaces, type (2) or (3).12?13 If, however, a graphite tube of about 4-6 cm in length and 6-8 mm in diameter is used, as in type (1)10314 or type 8,15 or a beaker as in type ( 7 ) , l O ~ ~ ~ the amount of sample may be 50 or even 100 pl.The furnace type (8) is a very effective atomiser for atomic-absorption spectrometry and was first described by Robinson and Wolcott15 in 1974. With this furnace, the sample vapour is not generated immediately in the absorption volume itself but in a separate section of the furnace, from where it is swept by a stream of inert gas into a second section of the furnace, the absorption volume.I n this furnace, the dissociation of molecules is very effective because of the length of time the vapour remains in the hot atmosphere. In this way, interference by molecular absorption is considerably reduced. In addition, chemical interference is reduced. The graphite T-piece can be heated by an electric current that flows through the walls.The stem of the T-piece is connected t o one electrode and each end of the cross-piece is connected to the second electrode. The sample is inserted into the stem and evaporated. The absorption volume is the cross- piece. Similar furnace constructions with a T-piece made from quartz and heated in an electro- magnetic r.f. field had already been used e a r l i e ~ - , l ~ > ~ ~ but the maximum usable temperature was only 1 500 "C.I n 1974, a new graphite-filament atomiser was described by Montaser et aZ.19 The filament consists of a graphite braid 3 cm in length and 1.5-2 mm wide, made of graphite fibres. Much less heating power is needed than with most other graphite atomisers. The power necessary t o achieve a temperature of 2 500 "C is only 350 W.20 The amount of sample normally used depends on the size of the furnace.Essentially, this furnace consists of a T-piece made of graphite. No cooling system is necessary.260 logical samples by atomic-absorption and atomic-fluorescence measurements. cedures show good sensitivity and a high power of detection.20 DEVELOPMENTS IN ANALYTICAL ATOMIC SPECTROMETRY Proc.AnaZyt. Div. Chern. SOC. The graphite-braid atomiser has been used for the determination of trace elements in bio- These pro- Furnace Materials In atomic-absorption and atomic-fluorescence spectrometry, graphite is the most common furnace material. Owing to its high sublimation point, it can even be used at temperatures above 3 000 "C. Contrary to most other materials, graphite shows increasing strength in compression and flexion up to 2 500 "C.Even drastic temperature changes have little effect on the shape and dimensions of graphite pieces. Also of special importance is that it is readily available in a very pure form. A disadvantage is that at high temperatures graphite reacts strongly with some elements to form refractory carbides or other graphite compounds, e.g., intercalation corn pound^,^^-^^ which is why some elements can be determined only with difficulty or not at all.Another disadvantage of the normally used modification of high-purity graphite is its porosity. This graphite allows the sample to soak into the surface. At high temperatures, diffusion of the vapour through the walls of graphite tubes causes a loss of sen~itivity.~~ In order to avoid loss of vapour by diffusion through the walls, pyrolytic g r a ~ h i t e ~ 6 , ~ ~ and glassy carbon2* have been used in furnace devices.Both materials are very dense, hard and non-porous, and are nearly impermeable to gases even at very high temperatures. L ' v o v , ~ ~ Clyburn et aZ.27 and Thompson et aZ.29 have described methods for pyrolysis treat- ment of graphite atomisation systems.The treatment may involve the introduction of a mixture of a hydrocarbon, e.g., methane, and an inert gas if the heated furnace element has a temperature above about 2 000 "C. A significant improvement in performance was observed in determinations of elements such as aluminium, barium, beryllium, silicon, tin, titanium and vanadium.29 This pyrolysis treatment is very important if continuously operating furnace devices are used.30 As the treatment can be repeated many times, the lifetime of the heated element becomes relatively long.Alternatives to graphite are metals with high melting-points, such as tantalum (2 996 "C), molybdenum (2 620 "C) or tungsten (3 380 "C) .31 Using such materials, carbide formation can be avoided, and in addition they are relatively impermeable to gases, which is why some ele- ments show higher sensitivity in tungsten or tantalum than in graphite furnaces.However, these materials are not inert and in many instances they react with the sample during heating, making the determination of some elements difficult or even impossible. McIntyre et aZ.32 investigated the suitability of molybednum and tantalum filaments in determining cobalt, nickel and copper by atomic-absorption spectrometry.Tantalum interacts with these ele- ments and in this instance it is not suitable as the furnace material, especially if cobalt and nickel are to be determined. Molybdenum, on the other hand, shows no noticeable interaction with these elements and gives very reproducible analytical results.In most instances, the determination of aluminium with graphite furnaces has many prob- lems because aluminium reacts with graphite. However, its determination with tantalum, molybdenum or tungsten furnaces is also difficult as aluminium and its compounds also react strongly with these furnace metals.33 In some instances oxides, e.g., beryllium oxide, or carbides, e.g., tantalum carbide, proved satisfactory from the point of view of thermal stability and non-reactivity.Like graphite or metals, these oxides and carbides can act as their own heating elements, but until now they have been of no practical importance in atomic-absorption and atomic-fluorescence spectro- metry with furnaces. Fig. 2 compares the characteristic amounts of some elements that have been determined with carbon-rod and metal-filament furnaces.The term "characteristic amount" is recom- mended by IUPAC34 for the amount of an element that can be determined with 1% absorption. Exact numerical values are meaningless unless the experimental conditions and the analytical problem are accurately defined; for this reason, only ranges are shown.In addition, these values tell us nothing about the attainable detection limits, as they depend not only on the characteristic amounts or the sensitivities but also on the analytical noise levels. If the prevailing analytical noise is due only to the apparatus and the filament atomiser, approximate guide values can be given. Under these conditions, it will be possible Its reducing properties aid in atom formation.September, 1976 DEVELOPMENTS I N ANALYTICAL ATOMIC SPECTROMETRY 261 to reach detection limits that are of the same order of magnitude as or lower than the values indicated for the characteristic amount.However, owing to matrix interferences or to fluctuating blank values, the detection limits are very often much higher.Carbon 'T 1 Metal filament 1 Mg Zn Na Li Rb Mn Cu Cr Ni Pd Pb Bi TI Al Se Ga Cd Be K Ag Ca Fe Sr Co Mo Cs Au Sb Sn As In Hg Comparison of the characteristic amount of some elements that Fig. 2. have been determined with carbon-rod and metal-filament furnaces. Development of Analytical Procedures As in all analytical methods, a very carefully devised procedure must be followed in atomic- absorption and atomic-fluorescence spectrometry with furnaces.For most analytical prob- lems, it is necessary to remove the interfering species of the sample matrix. This removal can be effected by a preceding chemical treatment of the samples, e.g., by chemical enrichment of the trace elements or by separation of the interfering matrix components by chemical methods.Such combinations of chemical pre-treatment of the samples and spectrochemical determina- tion of trace elements are very commonly used. However, if possible one tries to avoid this time-consuming step and to remove the interfering matrix components in the furnace itself. Especially in this instance, the heating programme must be very thoroughly developed. It is relatively easy to develop the correct heating programme for the determination of an element that is difficult to vaporise in a volatile matrix.As an example, Fig. 3 shows the deter- mination of nickel in magnesium sulphate by atomic-absorption spectrometry with a graphite-tube furnace. The most sensitive resonance line of nickel at 232.0 nm lies in a wave- length region in which molecular absorption due to SO, and SO can occur.21 The absorption spectrum of SO, is an electron excitation spectrum that has a fine structure, and this inter- fering background cannot be determined correctly by the usual background measurements. The absorption spectrum of SO is a spectral continuum due to the photodissociation of SO into S and 0.In order to establish the temperature that must be selected in heating step (2) to decompose the sulphate and to remove the interfering components without loss of nickel, the temperature was increased in steps from 900 to 1 800 "C.Experiments A-E were made with a nickel hollow- cathode lamp as the primary source, and therefore we see atomic absorption and background absorption at the wavelength of the nickel line. Experiments A'-E' were made with a deuter- ium lamp as the primary source.In this instance we can measure the background correctly. In heating step (l), a final temperature of 110 "C was selected to evaporate the solvent. The signals are due to background absorption only.262 DEVELOPMENTS IN ANALYTICAL ATOMIC SPECTROMETRY Proc. Analyt. Div. Chem. SOC. Temperature - 900 “C 1100 “ c 1500 “ C 1700 “C 1800 “C Time - Fig.3. Development of a temperature programme for the determination of nickel in magnesium sulphate by atomic-absorption spectrometry with a graphite- tube furnace. The temperature in heating step (2) has been changed between 900 and 1 800 “C. Upper reading with hollow-cathode lamp as the primary source; lower reading with a deuterium lamp. From experiments A and A’, it can be seen that a temperature of 900 “C in heating step (2) is not high enough to remove all interfering species and to avoid background absorption in the atomisation step (3).In experiments C, D and E, a loss of nickel occurs in heating step (2). The atomic-absorption signal in heating step (3) decreases from experiment C to E. The best conditions are obviously those in experiment B.Especially in trace analysis of organic samples, e.g., in biochemical or clinical trace analysis, pre-treatment of the samples is usually necessary. So far, the early hopes of being able to carry out a process equivalent to ashing in the furnace itself have not been fulfilled. Moreover, at the temperatures required to destroy most organic residues, many elements have appreciable vapour pressures and may be lost.Frequently, a suitable heating programme cannot be found. Background Measurements in Atomic-absorption and Atomic-fluorescence Analysis with Furnaces Background measurements in furnace methods are an important aid, without which some analyses would not have been possible. It is particularly necessary to measure the back- ground at low concentrations.This measurement could be carried out with similar samples that contain the analyte element in an undetectable concentration. By such “blank measure- ments,” systematic errors caused by the background can be avoided. If no such blank samples are available, the background can be determined by measuring the spectral background on both sides of the analytical line.This measurement can be carried out sequentially or simultaneously (automatic “background Compensation”). If the background is measured sequentially, random fluctuations of the background from sample to sample cannot be dimin- ished. If the background can be measured simultaneously, then these fluctuations can be very much compensated. Indeed, automatic background compensation is an important technical aid in atomic-absorption and atomic-fluorescence analysis with furnaces.However, the measurement of the background next to the analytical line must be viewed very critically if one is to avoid systematic errors. Particularly in atomic-absorption spectro- metry, accurate background measurements are extremely difficult because the background under the very narrow resonance line, which cannot be resolved by the monochromator, has to be determined.Se$tcmber, 1976 DEVELOPMENTS I N ANALYTICAL ATOMIC SPECTROMETRY 263 In most instances, the background measurement is carried out with a source of continuous radiation.Because of the low resolving power of the monochromators used in atomic- absorption spectrometry, its radiation flux will not be reduced noticeably by atomic absorption, but by broad-band absorption by molecules.This method of background measurement is correct only if the background is a spectral continuum, e.g., a dissociation continuum of molecules. However, the background measure- ment is wrong if the background is due to line-rich electronic excitation spectra of molecules. The actual background then depends on whether or not a rotational line of the molecular spectrum coincides with the atomic-absorption line.Correct background measurements could be carried out with spectral apparatus that has a very high resolving power. As an example, Fig. 4 shows the background interference of the gold absorption line at 267.6 nm due to an InCl absorption band. The two absorption spectra were recorded with a spectrograph that had a very high resolution. The upper spectrum shows the gold absorption line; this line does not coincide with a rotational line of the InCl band.Therefore, measurements with normal atomic-absorption instruments will give too high a value for the background. Au 267.595 nm I 2 1 1 7 - I .I I ! I I I I 1 1 267.2 267.3 267.4 267.5 267.6 267.7 Wavelength/nrn Fig.4. Molecular absorption spectrum of InCl in the region of the gold line, 267.6 nm, recorded with a spectro- graph of high resolution (X = 100 000). T , is the trans- mittance of the photographic plate. A comparison of both spectra shows that the gold line (upper spectrum) does not coincide with a rotational line of the InCl band (lower spectrum).Vnfortunately, there is at present no catalogue of interfering molecular spectra and of the necessary experimental data for atomic-absorption analysis with furnaces. A collection of such data could greatly facilitate the development of reliable analytical procedures. In atomic-fluorescence spectrometry (as in atomic-emission spectrometry), correct back- ground measurement is much simpler because the background and the line can be measured with the same resolving power or spectral band width.In most instances, the background under the line can be interpolated from background readings on bothsides of the analyticalline and the value thus found can be subtracted from the reading a t the peak of the line. 1. 2. 3. 4. 5. 6. 7. S. References King, A. S., Astvophys.J , , 1908, 27, 353. King, A. S., Astrophys. J., 1908, 28, 300. King, A. S., and King, R. B., Astvophys. J., 1935, 82, 377. L’x70v, B. V., Inzh.-Fiz. Zh., 1959, 11, 44. L’x-ov, B. V., Inzh.-Fiz. Zh., 1959, 11, 56. L’vov, B. V., Spectrochiun. Acta, 1961, 17, 761. L’vov, B. V., and Lebedev, G. G., Zh. Prikl. Spektrosk., 1967, 7, 264. Blassmann, H., “Proceedings of XIIth Colloquium Spectroscopicum Internationale,” Hilger and Watts, London, 1965, p.275.264 DEVELOPMENTS IN ANALYTICAL ATOMIC SPECTROMETRY Proc. AnaZyt. Div. Chem. SOC. 9. Massmann, H., “Reinststoffanalytik,” Part 2, Akademie-Verlag, Berlin, 1965, p. 297. 10. Massmann, H., Spectrochim. Acta, 1968, 23B, 215. 11. West, T. S., and Williams, X. K., Analytica Chim. Acta, 1969, 45, 27.12. Amos, M. D., Am. Lab., 1970, August, 33. 13. Amos, M. D., Benett, P. A., Brodie, K. G., Lung, P. W. Y., and Matousek, J. P., Analyt. Chem., 1971, 14. Manning, D. C., and Fernandez, F., Atom. Absorption Newsl., 1970, 9, 65. 15. Robinson, J. W., and Wolcott, D. K., Analytica Chim. Acta, 1975, 74, 43. 16. Martin, T. L., Thesis, University of Stellenbosch, 1973. 17. Robinson, J. W., Slevin, P.J., Hindman, G. D., and Wolcott, D. K., Analytica Chim. Acta, 1972, 61. 18. Robinson, J. W., Wolcott, D. K., Slevin, P. J. and Hindman G. D., Analytica Chim. Acta, 1973, 66, 13. 19. Montaser, A., Goode, S. R., and Crouch, S. R., Analyt. Chem., 1974, 46, 599. 20. Montaser, A., and Crouch, S. R., Analyt. Chem., 1974, 46, 1817. 21. Massmann, H., and Gucer, S., Spectrochim.Acta, 1974, 29B, 283. 22. Rudorf, W., in Emelkus, H. J., and Sharpe, A. G., Editors, “Advances in Inorganic Chemistry,” 23. Croft, R. C., Q. Rev. Chem. SOG., 1960, 14, 1. 24. Emeldus, H. J., and Anderson, J. S., “Ergebnisse und Probleme der Modernen Anorganischen Chemie,” Springer-Verlag, Berlin, Heidelberg, New York, Second Edition, 1954. 25. L‘vov, B. V., and Khartsyzov, A.D., Z h . Analit. Khivn., 1970, 25, 1824. 26. L’vov, B. V., “Atomic Absorption Spectrochemical Analysis,” translated by J. H. Dixon, Adam 27. 28. 29. 30. 31. 32. 33. 34. 43, 211. 431. Volume 1, Academic Press, New York, 1959. Hilger, London, 1970. Clyburn, S. A., Kantor, T., and Veillon, C., Analyt. Chem., 1974, 46, 2213. Kitagawa, K., and Takeuchi, T., Analytica Chim. Acta, 1973, 67, 457.Thompson, K. C., Godden, R. G., and Thomerson, D. R., Analytica Chim. Acta, 1975, 74, 289. Clyburn, S. A., Bartschmid, B. R., and Veillon, C., Analyt. Chem., 1974, 46, 2201. Weast, R. C., Editor, “Handbook of Chemistry and Physics,” Chemical Rubber Publishing Co., McIntyre, N. S., Cook, M. G., and Boase, D. G., Analyt. Chem., 1974, 46, 1983. Brewer, L., and Searcy, A. W., J .Am. Chem. Soc., 1951, 73, 5308. “Nomenclature, Symbols, Units and Their Usage in Spectrochemical Analysis. 111. Cleveland, Fifty-second Edition, 197 1-72. Analytical Flame Spectroscopy and Associated Procedures,” IUPAC Information Bulletin, No. 27, November 1972. Carbon Furnace Atomic-emission Spectrometry F. Shaw, J. M. Ottaway and D. Littlejohn Department of Pure and Applied Chemistry, University of Strathclyde, Cathedral Street, Glasgow, G l 1 X L Most excitation sources in common use in emission spectrometry involve high-temperature flames, plasma or electrical discharges in which the excitation energy is derived from the combustion process or the electrical energy.The use of a carbon furnace atomiser, a purely electrothermal device, provides an emission source with significantly different properties, which may find some applications in analytical chemistry.Our initial studies of this technique, using a commercial atomic-absorption carbon furnace system, were reported last year1 and showed encouragingly low detection limits for several elements. At that time, the expected Boltzmann-type dependence of emission intensity with wavelength appeared to place a limitation on the range of elements available to this technique.The instrument used was the Perkin-Elmer HGA-72 carbon furnace atomiser and optimum conditions found for almost all elements involved atomisation at maximum tube temperature under gas-stop conditions. It is also now clear that useful detection limits are obtained only when the optical system of the spectrometer used with the furnace effectively separates the atomic emission emanating from the centre of the tube from the continuous emission from the walls of the carbon tube.The Boltzmann equation would also predict that the population of atoms in excited energy levels, and hence the emission intensity, would follow an exponential relationship with tem- perature.The attainment of higher temperatures in the carbon furnace should therefore result in significantly improved detection limits for all elements. Higher temperatures have been achieved by reducing the thickness of the central portion of the standard HGA-72 carbon tube, and this procedure has indeed resulted in an improvement in detection limits for many elements by one or two orders of concentration.2 A comparison of detection limits for 17 elements in standard and “modified” tubes was published recently in P~oceedings.~ Useful detection limits have now been obtained for a further eight elements: rubidium, 0.000 19 pg ml-l ; ytterbium, 0.000 36 pg ml-l ; strontium, 0.001 2 pg ml-l ; europium, 0.023 pg ml-l ;Septembev, 1976 DEVELOPMENTS I N ANALYTICAL ATOMIC SPECTROMETRY 265 dysprosium, 0.053 pg ml-l; caesium, 0.018 pg ml-1; holmium, 0.021 pg ml-l; and erbium, 0.042 pg ml-l.The range of elements covered by this technique has therefore been con- siderably extended but it still lacks sensitivity for the volatile elements that are atomised and leave the carbon tube at temperatures that are too low4 for significant excitation of the atoms to occur.The variation with temperature of atomic-absorption and atomic-emission signals of iron and chromium is shown in Fig. 1. As the temperature of the carbon furnace increases, atomic-absorption signals increase up to a maximum when the atoms are released at a rapid rate from the surface of the tube. This effect can be seen clearly for iron but the signal for chromium, which is atomised only at high tube temperatures, has not reached its maximum a t the currently available maximum tube temperatures.A slight levelling off of the chrom- ium signal can be discerned and it would undoubtedly reach a maximum at higher tempera- tures. In contrast, the emission signals for both elements show a distinct exponential or Boltzmann-type dependence on tube temperature.Although emission intensities are related to the vapour-phase temperature in the carbon tube, there will be a close relationship between this and the tube temperature, and the Boltzmann-type response would be expected. Any further increases in tube temperature can therefore be expected to lead to improvements in the detection limits for all elements using carbon furnace atomic-emission spectrometry.2 000 2200 2400 2600 2800 3000 Temperature/ K Effect of temperature on the atomic-absorption signals for (A) 50 pl of 0.1 pg ml-I iron and (B) 50 pl of 0.1 pg ml-I chromium; and effect of temperature on the atomic-emission signals for (C) 50 pl of 0.2 pg ml-l iron and (D) 50 p1 of 0.1 pg ml-I chromium. Peak-height measurement using the modified form of graphite tube.HGA-72 atomiser, gas-stop conditions. Fig. 1. The application of this technique in routine analysis depends on the degree to which inter- ference effects can be controlled but the determination of microgram per gram amounts of lithium in copper has been reported5 and the determination of minor elements in steel is also possible without interference from the matrix.Chemical interferences in carbon furnace atomisation are reduced to a low or negligible level when sample solutions are prepared in oxyanion, e.g., nitrate or sulphate, medium3y6 and this type of interference will be the same in both atomic-absorption and atomic-emission spectrometry. It has been found7 that the degree of ionisation of atoms during carbon furnace atomisation is very low and can be easily controlled, so that ionisation interferences will not be significant either.A number of spectral interferences observed previously in flame emission spectrometry have been studied in carbon furnace atomic-emission spectrometry and in most instances these interferences are also insignificant, particularly when the interfering line is a non-ground-state transition.The background emission from the tube is much less noisy than the flame emission background and varies in a continuous and reproducible manner with wavelength and atomisation time. It can therefore be easily and precisely measured and subtracted or corrected for.266 DEVELOPMENTS IN ANALYTICAL ATOMIC SPECTROMETRY PYOC. AnaZyt. Dh. Chem. Soc. The potential advantages of the carbon furnace as an emission source appear to be con- siderable, although whether, in its finally optimised form, it will compare well or favourably with other existing emission sources remains a question for debate and further research.References 1. 2. 3. 4. 5. 6. 7. Ottaway, J. M., and Shaw, F., Analyst, 1975, 100, 438. Ottaway, J. M., and Shaw, F., AppZ. Spectrosc., 1977, 31, in the press.Ottaway, J. M., Proc. Analyt. Div. Chew,. Soc., 1976, 13, 185. Campbell, W. C., and Ottaway, J. M., Talanta, 1974, 21, 837. Shaw, F., and Ottaway, J. M., Analyt. Lett., 1975, 8, 911. Ottaway, J . M., Proc. Analyt. Div. Chem. Soc., 1975, 12, 176. Ottaway, J. M., and Shaw, I?., Analyst, 1976, 101, 582. Practical Developments in Atomic-f luorescence Spectroscopy-Invited Lecture T.S. West Macaulay Institute j o y Soil Reseavcla, Craigiebuckler, A berdeeiz, A B9 2Q J The basic equation that relates the resonance fluorescence signal for an atomic species to its original concentration in a sample solution can be written1 as where p denotes radiance, 0 the quantum efficiency coefficient for the resonance line, 0: the irradiational cross-section, P the nebulisation coefficient, y the atomisation coefficient, K, the atomic absorptivity coefficient, I the absorbing pathlength and C the gross (analytical) concentration of the analyte species in the sample solution.The subscripts f and a refer to fluorescence and absorption, respectively. The actual signal measured is the amplified response of a transducer, such as a photomulti- plier detector, and has no absolute value, but is used empirically in relation to that produced by a reference amount of the analyte species subjected to the same measurement process with all variable parameters such as excitation/emission geometry, source and atom reservoir character- istics kept as invariable as possible for sample and reference amounts. As always, it is par- ticularly important that the physical and chemical compositions of the two media should be virtually identical.Under radiance from a line or continuum source of conventional intensity in conditions where less than 5% of the incident radiation is absorbed, the atomic species in the reservoir exhibit classical quantum mechanical behaviour so that 0 approximates closely to a constant, as does K,.Consequently, with fixed geometry, and constant source and atomiser conditions, the fluorescence depends linearly on the concentration of the analyte species within the simplifying approximations used in setting up equation (1). Choice of Fluorescence Line The basal resonance line ( i e . , that of lowest frequency) tends to be the most useful analytic- ally but its sensitivity depends on several factors, e.g., its oscillator strength, its intensity of emission by the selected source, the intensity characteristics and number of other source lines that make contributions to the basal resonance frequency by stepwise fluorescence, the relative response of the detector and the transmission characteristics of the monochromator at the basal frequency. Generally, however, the basal resonance frequency is the easiest to excite, has the maximum stepwise contribution, lies closest to the optimum values of detector/ monochromator sensitivity and shows slightly less scatter than resonance lines of higher frequency.Stepwise lines are very useful but they are usually weaker in intensity than basal resonance lines and suffer weakening due to resonance re-emission competition at their own frequency ofSepteinbev, 1976 DEVELOPMENTS I N ANALYTICAL ATOMIC SPECTROMETRY 267 excitation.They are not so subject to scattering as resonance lines and the problem can be minimised by inserting a filter between the atom reservoir and the source, which cuts off the measured frequency. However, direct-line fluorescence is perhaps the most useful alternative to resonance fluorescence.Because the spacings between upper quantum levels are usually less than those between any such level and the ground level, the frequency of direct-line emission is usually much less than that of the basal resonance line. Thus, direct-line fre- quencies tend to fall within the most sensitive ranges of detectors and monochromators.In this instance also the greater separation between the frequencies of excitation and fluorescence emission usually allows a higher throughput of excitation energy when a filter is used to minimise scatter. Thermally assisted fluorescence phenomena and atomic phosphorescence are equally useful in these respects, but are much less common. Choice of Atom Reservoir Diffusion flames, using nitrogen or argon as the nebulisation medium and hydrogen as the fuel gas, have been used but, apart from tin, for which they are useful, they are little used because they suffer from physical and chemical matrix effects and have low atomisation efficiencies.Turbulent flames have a compact and dense format and utilise all of the sample, but they suffer from high and noisy background radiation levels, show poor atomisation efficiency and are surprisingly subject to matrix problems.Laminar-flow hydrocarbon flames using pre-mixed oxidant gases to nebulise the sample solution are undoubtedly the most useful. The intensely radiating primary combustion zone is kept low in the flame so that it produces scarcely any background.Separation of the chemiluminescent outer diffusion mantle of the flame by a sheath of concentrically flowing argon or nitrogen leaves the atomic species in an almost non-radiative zone where fluorescence can be generated and measured easily and also protects atoms against oxidative attack.2 Separation does involve some cooling of the environment of the species being measured, but this scarcely affects fluorescence measurements for most species.Unquestionably, the separated flame is the best form of flame atom reservoir for atomic-fluorescence measurements inasmuch as it provides efficient atomisation for most metal and metalloid species, low emissive/absorptive backgrounds and moderate freedom from matrix effects. Radiofrequency inductively coupled plasmas are excellent sources for atomic-emission spectroscopy and their high temperatures ensure complete gasification of most species.Their high backgrounds in the so-called doughnut region can be avoided by working higher in the plume and possibilities exist of bending the strongly emitting charged plasma away from the stream of neutral species by application of a suitable field, but so far they have been little used for this purpose.The use of an electrothermal atomiser within the flow of plasma support gas to produce discrete sample introduction has merits in avoiding the need for cumbersome desolvating devices. Microwave plasmas offer more compact sources, but their capacity to accommodate water vapour is critically low so that they require extremely vigorous desolvation or must be used with an electrothermal analyser.Sputtering chambers of the hollow-cathode variety have considerable merit for metallurgical specimens and can be adapted to deal with solutions, but in the latter instance they have not been widely used because of the greater facility of flames and electrothermal analysers. Chemical flames are the most common atomisation devices.Electrothermal (Resistively Heated) Atomisers Resistively heated atomisers, such as carbon tubes, cups and rods, metal rods, ribbons and loops, have supplied the principal competition to flames as the most aseful atomisers. For fluorescence work, the rod, ribbon or loop devices are the most useful as they provide a tran- sient free cloud or plume of atoms against a non-radiating background.It must be appreciated, however, that the lifetime of their fluorescent event is frequently shorter than the response time of the amplification/recording system of most spectrometers designed for flame work . Indeed, with such instruments the modulating frequency is often insufficient even to code the signal. The lifetimes are transient because the surface-generated atoms are presented im- mediately to a “cold” environment in which most atomic species condense very rapidly to form polyatomic aggregates.In the presence of preponderant amounts of indifferent aggregat- ing species, e.g., sodium chloride, the rate of nucleation of the analyte species from the gas268 DEVELOPMENTS IN ANALYTICAL ATOMIC SPECTROMETRY R o c . Analyt.Div. Chem. SOC. phase becomes greater than normal. Signals must then be measured more quickly nearer the point of generation. Carbon is the most useful rod material because of its reducing power, but there are instances, such as the use of a tungsten rod, where interferences can be minimised to a much greater extent because the power dissipation in the atomisation environment is considerably greater, thus delaying the onset of n~cleation.~ There are, for example, fewer matrix effects for lead with a tungsten rod than with a carbon rod of similar dimensions.The situation for carbon-rod atomic-fluorescence measurements can be summarised in relation to flame atomisation as follows: (1) higher absolute sensitivity; (2) only small samples required; (3) subjection of the entire sample to atomisation; (4) closer approximation of the fluorescence volume to the desideratum of a point source; (5) different interference pattern ; and (6) necessity for greater operator skill.Excitation Source In fluorescence work, the source should be as intense as possible within limits that may be imposed by self-reversal. It does not need to be a line source, but may be an ultraviolet - visi- ble continuum source as the irradiated atoms act as their own monochromators.High- pressure xenon arc lamps of the modern variety that resemble sealed motor-car headlamps are particularly useful because of their high radiance down to 200 nm. Resolution then depends on the use of suitable monochromation between the fluorescing atoms and the detector - trans- ducer.Such continuum sources may yield detection limits similar to those obtainable by atomic-absorption spectroscopy. Hollow-cathode lamps may yield measurable signals with electrothermal atomisers, but require to be pulsed for flame work. High-intensity lamps (boosted output) of the Sullivan and Walsh or Lowe variety are excellent for someelements, e.g., magnesium, silver, gold, cobalt, nickel and zinc, but are not readily available.Lamps of the Grimm variety have also been used, as have sealed metal vapour lamps of the "street lamp" type, for a limited number of elements. Suitably doped flames and inductively coupled radiofrequency or microwave plasmas have also been employed, but offer no particular advantage. Electrodeless Discharge Lamps At this time, microwave-excited scaled electrodeless discharge lamps are the most useful and generally applicable line sources for the excitation of atomic fluorescence.However, their construction and operation have been points of considerable controversy and the em- pirical approach adopted in the past has caused widespread and diverging views to be published. The heated or thermally stabilised mode of running electrodeless discharge lamps first described by Browner et aZ.4 is a major advance.It separates the heating and excitation requirements of the lamp where previously both were supplied by the microwave power. In some instances, the conditions are antagonistic. For example, alkaline earth metals are easy to excite but their salts are involatile. Recent work5 has shown that such lamps may be heated in $- or Q-wave resonant cavities with an air stream up to 550 "C.In the same study, the critical factors in the manufacture and operation of temperature-controlled lamps were examined. Problems arise chiefly from the presence of impurity species such as oxides or their formation i.tz sit% by overheating of the lamp (usually by operating under uncontrolled conditions or during sealing).The presence of excess of halogen also causes problems owing to quenching. Fig. 1 shows the effect of the air-stream temperature on the intensity at about 228.8 nm of a cadmium electrodeless discharge lamp. At the peak temperature of 310 "C the intensity is about 100-fold greater than that at room temperature and it shows little change with ap- preciable fluctuations in temperature on the peak plateau.Fig. 2 shows a comparison of the performance of a 310 "C stabilised electrodeless discharge lamp against an unstabilised lamp as a function of microwave power. It will be observed that at the optimised temperature, power fluctuations cause little change in intensity. The cadmium 228.8-nm fluorescence peak approximates closely to the intensity peak temperature, but the atomic-absorption peak occurs at about 160-170 "C.This lower temperature for the use of the electrodeless discharge lamp in atomic-absorption measurements is no doubt due to the onset of line broadening and the sharply rising emission curve in Fig. 1 might appear to suggest that absorption signalsSeptember, 1976 DEVELOPMENTS IN ANALYTICAL ATOMIC SPECTROMETRY 269 would lack reproducibility in this range.However, the absorption - temperature profile is in fact flat and easily reproducible. Similar results have been obtained for lamps for arsenic, copper, manganese, phosphorus, sulphur, selenium, thallium, tellurium, zinc, etc., and will be published in due course. 0 100 200 300 Air-stream temperature/"C Fig.1. Variation of lamp intensity a t 228.8 nm for a cadmium electrodeless dis- charge lamp operated in a &-wave resonant cavity as a function of the surrounding air- stream temperature. A studv has also been made of the li:_-----:_ 5 Fig. 2. Variation of lamp intensity and fluorescence response for cadmium at 228.8 nm as a function of applied power using a temperature-controlled electrodeless discharge lamp a t 300 "C: (a) lamp intensity; x d (b) atomic-fluorescence signal for a cadmium solution.A $-wave resonant cavity was used. behaviour of some element pairs such as thallium (chloridej- cadmium, selenium - tellurium and cadmium - zinc.5 Although our results arc limited, a study of intensity - temperature and fluorescence - temperature curves against factors such as the ratio of thallium to cadmium reveals some apparent trends that have worthwhile application.In the pairs studied, the peak emission of the higher temperature element appears to be moved to an appreciably lower temperature for ratios of about unity, while the point of maximum emission of the lower temperature partner is little affected but moves slightly higher.Our conclusions are that temperature-controlled electrodeless discharge lamps are much superior to lamps operating in an uncontrolled environment. They can almost invariably be run at lower powers and show temperature and power plateaus where the intensity scarcely varies. Some lamps do, however, show temperature - intensity peaks rather than plateaus, e.g., thallium chloride, probably owing to halogen quenching, and some extinguish very repro- ducibly at high temperatures, e.g., copper. The temperature-controlled lamp usually exhibits lower noise levels, greater signal to background characteristics in fluorescence and lower susceptibility to tuning changes and position within the resonant cavity than the corresponding uncontrolled lamp.Also, the operational life of such a lamp is, in our experience, much longer. So far, this effect has been scarcely noted for 1 O : l and 1 :10 ratios. A few lamps, e g . , mercury, must be cooled. Laser-excited Fluorescence The excited-state population of atomic species produced by normal sources such as electrode- less discharge lamps is usually a very small fraction of the ground-state population.The radiant power of a tunable dye laser source is such, however, that 50% of the ground-state population can be excited. Under such conditions, maximum atomic-fluorescence and inde-270 DEVELOPMENTS I N ANALYTICAL ATOMIC SPECTROMETRY PYOC. AThdyt. Div. Chevz. sot. pendence of the fluorescence signal from the source intensity can be obtained.Other benefits are the virtual disappearance of self-absorption (because it is balanced by stimulated emission), minimisation of quenching and improved signal to noise ratios. Such lasers are, of course, expensive and have a limited range of application at higher frequencies in their present state of development. It is, however, misleading to consider such lasers simply as very intense thermal sources because the excited populations that they induce may not behave according to the normal laws of quantum mechanics.Thus, a narrow-line single transverse mode laser can behave as a diff raction-limited plane-wave generator, the coherence of whose photon train can radically alter the photon - atom interaction. Groups of atoms in constant phase relation to each other can interact mutually and spontaneously emit coherent radiation.The radiant power of such clusters of co-operating atoms should be orders of magnitude greater than normal fluorescence, but their excited-state lifetime will be greatly decreased. Thus, not only does saturation become more difficult to achieve, but in addition the emitted fluorescence will be anisotropic and have a greater spectral band-width. Non-linear analytical growth curves must also be expected with signals proportional to No2 at low concentrations of analyte solution, where No2 is the number of absorbing atoms per unit path length.Nevertheless, the availability of laser sources presents a challenging new dimension for work in atomic-fluorescence spectrometry. Optics and Multi-element Non-dispersive Atomic Fluorescence Non-dispersive atomic-fluorescence spectrometry is a well recognised technique but the temperature ramp of a resistively heated carbon filament or rod offers another degree of freedom to non-dispersive fluorescence, as the appearance temperature of atomic species varies from one element to another.Mercury, for example, will be atomised completely when the rod temperature reaches about 200 "C (depending on its original form), cadmium by 500 "C, bismuth by about 1000 "C, iron by 2 500 "C and so on.Thus, if the surface of a rod on heating progressively releases these species at successive intervals in time into the surrounding space, fluorescence signals characteristic of each species will appear sequentially if the reservoir is irradiated by the appropriate line sources or a continuum source.A simple apparatus for time-resolved non-dispersive dual-element atomic fluorescence is shown in outline in Fig. 3.6 Reflectance optics of Cassegrainian design are used as they eliminate chromatic aberration, produce a pin-point focus and have high ultraviolet trans- mission. A solar-blind photomultiplier responds simultaneously to all of the resonance Detect0 r n Amdifier u- '+ Power supply - read -ou t A 1 Reflecting objective v a v e c a v i t y Power divider - 1 Microwave power supply r-l Fig.3. Schematic diagram of apparatus used for measuring time- resolved non-dispersive atomic-fluorescence signals using a carbon- filament atomiser. Tirne/s Fig. 4. Oscilloscope trace of time-resolved non-dis- persive atomic-fluorescence signals for 5 ng of cadmium and 5 ng of thallium.Two single-element sources were used.September, 1976 DEVELOPMENTS I N ANALYTICAL ATOMIC SPECTROMETRY 271 stepwise or direct lines for the fluorescing species falling within its response range. Excitation is obtained by two single-element temperature-controlled electrodeless discharge lamps operated from a single microwave generator via a two-port power divider.Fig. 4 shows a typical spectrum produced by such an experiment for 5 ng of cadmium and 5 ng of thallium. Similar experiments yielded good results with mercury - zinc, mercury - cadmium, mercury - thallium, bismuth - cadmium, zinc - cadmium, etc. References 1. 2. 3. 4. 5 . 6. West, T. S., PYOC.SOC. Analyt. Chem., 1974, 11, 198. Kirkbright, G. F., and West, T. S., Appl. Optics, 1968, 7, 1305. Cantle, J . F., and West, T. S., Talanta, 1973, 20, 459. Browner, R. F., Patel, B. M., Glenn, T. H., Kietta, M. E., and Winefordner, J. F., Spectvosc. Lett., 1972, Bartley, D. F., and West, T. S., to be published. King, A. F., and West, T. S., to be published. 5, 331. Determination of Zinc and Cadmium in Biological Samples by Atomic-f luorescence Spectrometry G.S. Fell D. 6. Hough, F. E. R. Hussein and J. M. Ottaway Department of Biochevnistvy, Royal Infivmavy, Glasgow, G4 OSF Depavtment of Puve and Applied Chemistvy, Univevsity of Stvathclyde, Cathedval Stveet, Glasgow, GI I X L Clinical interest in the metabolism of both essential and toxic trace metals has led to a require- ment for the determination of various metals, such as zinc and cadmium, at low levels in biological material obtained from both patients and experimental animals.Although various analytical techniques have the required sensitivity, some, such as neutron- activation analysis, require specialised facilities and are slow in relation to a patient care service.Others, such as electroanalytical methods, require considerable care in the pre- paration of material prior to analysis in order to prevent the introduction of contamination by reagents. Atomic-absorption methods, using carbon furnaces or rods, have the required sensitivity, but variable interference effects due to the high chloride content of biological samples can be a difficulty that is not altogether overcome by existing background correction facilities.Flame atomic-fluorescence spectrometry offers good sensitivity and the conventional flame and nebuliser systems used overcome most of the chemical interference problems. Such systems are well suited to handling liquid samples and are capable of a high analytical work- rate. Unfortunately, there are no commercially available purpose-built atomic-fluorescence instruments.In a previous report,l the adaptation of a standard atomic-absorption spectrometer for atomic-fluorescence measurements was described, using an air - hydrogen flame and an elec- trodeless discharge lamp as the light source. We now report the improvement achieved by thermostating the electrodeless discharge tube,2 by passing heated nitrogen gas into the microwave cavity.This arrangement affords about a 10-fold gain in detection limits, and improves the reproducibility of the system. The instrumental settings for zinc and cadmium are similar to those reported previously except that the electrodeless discharge lamp is thermostated at 220 "C for zinc and 288 "C for cadmium. Application to the Determination of Zinc Although conventional flame atomic-absorption methods offer good sensitivity for the determination of zinc in plasma and urine, it was felt that a semimicro method for zinc in plasma could offer advantages in paediatric cases and in studies on experimental animals.The detection limit for zinc, by the atomic-fluorescence method, is about 0.2 p.p.b., and as the zinc level in plasma is normally about 1 p.p.m., a considerable dilution of plasma can be272 DEVELOPMENTS IN ANALYTICAL ATOMIC SPECTROMETRY PYOC.Analyt. Div. Chem. SOC . employed. Samples of 5 pl of plasma were diluted to 5 ml (1 in 1 000) and the results for zinc assay compared with those from an atomic-absorption procedure3 in which 0.5 ml of plasma was diluted to 2.5 ml.The results were in good agreement with a statistical corre- lation coefficient of 0.95 for 14 samples. The atomic-fluorescence method has also been applied to the determination of zinc in urine and tissue digests, again with good agreement compared with an atomic-absorption method. The method is sufficiently precise jwithin- batch relative standard deviation = 2.7% (20 results)] to be suitable for routine use and its sensitivity offers important economies of sample material and allows an extension of the analysis to studies of the fractions of zinc in plasma such as plasma ultrafiltrates.Application to the Determination of Cadmium The determination of cadmium in blood and urine can form an important part of the clinical investigation of populations thought to be at risk from cadmium toxicity either environ- mentally or as an occupational hazard.The levels found in the blood of non-industrially exposed persons are about 0.005 p.p.m., and most methods of analysis require a pre-concentra- tion procedure before analytical measurement. The atomic-fluorescence method is able to perform satisfactorily using 1 ml of blood diluted to 5 ml with dilute acid prior to aspiration into an air - hydrogen flame.The method using the thermostated electrodeless discharge lamp has a detection limit of about 0.03 p.p.b., of cadmium. Comparisons of blood cad- mium results with those from duplicate analyses performed by Cernik and Sayers4 using a filter-paper punched-disc technique involving a carbon-cup atomic-absorption system4 were acceptable and gave a statistical correlation coefficient of 0.97 for 22 samples.The precision of the method is 3.5% within batch (12 results) and 8.12% between batch (10 results) at the 0.005 p.p.m. level. The method has also been applied to urine analysis for cadmium, but inter-laboratory comparisons are at present unsatisfactory, possibly owing to problems of sample contami- nation and deterioration during storage.a 800 400 i ---- I I I I I I I I I I r-- 0 00 8 I 0 I 4 8 12 16 20 Cdlpg I--’ Fig. 1. Correlation of blood lead and blood cadmium levels in a survey of shipbreakers. Normal maximum levels for each element are indicated by dotted lines. Blood lead values were obtained by the Delves’ cup technique. Each point indicates one sample except for 0, which indicates more than one sample a t that point.The method for blood analysis for cadmium has been applied to various categories of industrial workers and an apparent increase in blood cadmium, together with that in blood lead, has been observed in workers involved in metal cutting operations, such as shipbreakersSe$tembev, 19'76 DEVELOPMENTS I N ANALYTICAL ATOMIC SPECTROMETRY 273 and other demolition workers.sequences for patients undergoing penicillamine treatment for lead t ~ x i c i t y . ~ This result is illustrated in Fig. 1 and has important con- The collaborative testing of analytical methods with A. A. Cernik, EMAS, London, and colleagues in the Royal Infirmary, Glasgow, is gratefully acknowledged. References 1. 2. 3.4. 5. Hough, D. C., and Ottaway, J. M., PYOC. SOC. Analyt. Chewz., 1974, 11, 223. Browner, R. F., and Winefordner, J. D., Spectrochirvz. Acta, 1973, 28B, 263. Peaston, R. T., Med. Lab. Technol., 1973, 30, 249. Cernik, A. A., and Sayers, M. H. P., BY. J . Ind. Med., 1975, 32, 155. Fell, G. S., Hussein, F. E. R., and Ottaway, J. M., to be published. A Kinetic Theory of Atomisation for Atomic-absorption Spectro- metry Using a Graphite Furnace Atomiser Part 111." Matrix Control An Aid to the Selection of Atomisation Parameters for C.W. Fuller Tioxide Iiatevnational Ltd., Central Labovatories, Stockton-on-Tees, Cleveland, T S 1 8 2NQ Several kinetic theories of atomisation have been proposed recentlyl-4 in an attempt to define the atom population existing in electrothermal atoniisers.There are two different analytical models to consider : ( a ) atomisation under increasing temperature, and (b) atomisation under constant temperature (isothermal). An outline of the approaches used for these two models is given below, followed by the application of the isothermal atomisation model to the selection of atomisation parameters for matrix control.Atomisation under Increasing Temperature (i) According to L'vov,~ the rate of change in the number of atoms, N , in an atomiser can be expressed by the equation .. .. * * (1) dN/dt = n,(t) - nz(t) . . .. where nl(t) is the number of atoms entering the system in unit time and n,(t) is the number of atoms leaving the system in unit time. For a constantly increasing atomisation temperature, with A = dT/dt : .... .. * * (2) nl(t) = A t . . .. and 71 .. .. * * (3) [ n,(t)dt = N(0) . . .. J O where T~ is the time to transfer the total number of atoms, N(O), to the system. Therefore, .. .. * - (4) .. * * ( 5 ) n1(t) = ~ N ( O ) ~ / T ~ ~ . . .. Assuming atoms are removed from the system by vapour diffusion, then nz(4 = NIT2 .. .. .. where 7 2 is the average residence time of atoms in the system.Equation (1) can be rewritten to show the rate of change in atom population at time t : dN 2N(O)t N - .. - - (6) .. .. dt 71, 7 2 (ii) According to Torsi and Tessari,2 the rate of release of atoms from a monolayer on the *For details of Parts I, I1 and IV of this series, see reference list, p. 276.274 atomiser surface, nl(t), is given by the equation DEVELOPMENTS IN ANALYTICAL ATOMIC SPECTROMETRY Proc.AnaZyt. DiV. Chem. SOC. where s is the surface area, 8 is the fraction of surface covered at time t, q is the surface con- centration when 8 = 1 and k is the rate constant for the atomisation process. If k follows a normal Arrhenius temperature dependence, then nl(t) = AsqOexp(-AG/RT) . . .. . . * * (8) Assuming that the temperature varies with time as a simple function T = T(0) + at, where T(0) is the initial temperature and a = dT/dt, then equation (8) becomes If the space around the atomiser surface acts as an infinite atom sink, then equation (9) represents a typical peak signal because, as T increases with time, there is a corresponding reduction in the value of 8.(iii) According to Johnson et aZ.,3 if atomisation can occur only when atoms acquire a minimum energy, then the number of atoms with this energy is given by the equation n = Nexp( --EIRT) .. .... . . (10) where N represents the number of atoms in the monolayer. proportional to n, then As the rate of atomisation will be --dN/dt = CNexp( --E/RT) . . .. .. . . (11) Assuming that the analysis volume is directly above the the atomiser surface, then removal of atoms from this volume will occur at the same rate as they enter.If the residence time of atoms in the sample volume is 7, then the number of atoms, A (t), present at any time t is given by the difference between the number of atoms that have entered and left the analytical volume : A(t) = “(0) - N(t)] - “(0) - N(t - T)] = N(t - 7) - N(t) .... . . (12) which can be solved by using equation (11). All of these atomisation theories are applicable to rapid heating atomisers, e.g., graphite rods and cups and metal filaments, and for easily atomised elements in graphite furnaces at high temperatures. Atomisation under Constant Temperature (2) Fuller4 again considered that the rate of change of atom population in the atomiser is nl(t) = k,“(O) - N(t)] .. .. .. . . (13) where N(0) is the initial amount of element introduced to the atomiser and N(t) is the amount of element atomised up to time t. Then, N(t) = N(0)[1 -exp( -kit)] . . .. .. . . (14) given by equation (1). Here which, on substitution back into equation (13), gives n,(t) = k,N(O)exp( -k,t) . . .. .. . . (15) If the rate of loss of atoms from the atomiser is controlled by diffusion and the flow of purge gas through the atomiser, then n2(t) = k,N .. .. . . .. . . (16)Septembev, 1976 DEVELOPMENTS I N ANALYTICAL ATOMIC SPECTROMETRY Hence, 275 dN/dt = k,N(O)exp( -k,t) - k,N .. .. . . (17) This model is applicable to furnace atomisers, particularly when the rate of atomisation is slow, in which event the atomiser temperature will reach equilibrium conditions during atomisation.Applications of Kinetic Models of Atomisation The kinetic models for atomisation can be used to describe various practical analytical parameters, e.g., peak height and integration signal measurement,5 stopped gas flow operati~n,~ interference studies6 and matrix control. The remainder of this paper describes the possible application of the isothermal atomisation model to matrix control.Matrix Control For an analytical determination to be successful, it is necessary for the concentration of the matrix element vaporised during the atomisation period to be below that at which it will interfere in the determination. For example, assume that a matrix concentration up to 100 pgml-l can be tolerated during the atomisation period but that the sample matrix concentration is 10 000 pg ml-l.It is necessary, therefore, to ensure that not more than 1% of the sample matrix is vaporised during the atomisation period. Assuming that the loss of elements from a graphite furnace obeys first-order kinetics, that C represents the concentration of the element remaining in the furnace at time t and that C(0) represents the initial concentration of the element in the furnace at time t = 0, then .... . . (18) -dC/dt = kC . . .. Integrating gives and ln[C(O)/C] = kt . . .. .. .. . . (19a) .. .. .. . . (19b) 1 k t = - ln[C(O)/C] . . Equation (19b) can be used to obtain the time required to remove any amount of element from the furnace at any temperature, provided that the dependence of the rate constant, k, on temperature is known.The loss of elements during the pyrolysisstage of an analytical determination can be obtained by experiment. The procedure entails plotting the peak absorbance value, obtained during the atomisation period, against the pyrolysis time for various pyrolysis temperatures. These results can then be used to obtain the relevant kinetic information.There are two analytical situations that can occur: (a) the matrix (B) is more volatile than the analyte (A) and (b) the analyte (A) is more volatile than the matrix (B). (a) In this situation, 99% of the matrix must be removed, with minimal loss of the analyte (say, less than 5%), prior to atomisation. Using equation (19b) the time, t,,, to remove 99% of the matrix can be obtained: t,, = -ln .... . . (20a) = 4.6/kB .. .. .. .. .. . . (20b) Similarly, for the loss of less than 5% of the analyte, the time, t,, can be obtained from equation (19b) : t, = 1 In[ 100c(0) ] . . . . .. . . (2la) k A (100 -5)C(O) = O.O51/KA . . .. . . .. .. . . (21b)276 DEVELOPMENTS IN ANALYTICAL ATOMIC SPECTROMETRY Proc. AnaZyt.Diu. Chem. SOC. Knowing from experiment how kA and kg vary with temperature, it is possible, mathe- matically or graphically, to obtain the temperature at which equation (20) holds for a typical pyrolysis time of 30 s. If the value of kA, corresponding to this temperature, is substituted into equation (21) the value of t, can beobtained. Provided that this value of t, is greater than or equal to 30 s, then the analytical determination is feasible.( b ) In this situation, most of the analyte (say, 90%) must be atomised before 1% of the matrix. Again, using equation (19b) : t,, = -1n .. .. . . (22a) .. . . .. . . (22b) For the loss of less than 1% of the matrix, the time t, can also be obtained from equation (19b) : t, = I I n [ 100c(0) ] . . . . . .. . (23a) k , (100 -l)C(O) = 0.009 2/kB . . .. .. .. .. . . (23b) For an atomisation time of 2-5 s, ii. is possible to determine the value of kA and hence the temperature required for analysis, The appropriate value of ks, corresponding to this tem- perature, can be obtained together with the resulting value of t,. Provided that this value of t, is greater than or equal to the atomisation time, then the analytical determination is feasible.Although this approach is clearly oversimplified, it does serve to illustrate the possible application of the kinetic models. Also, by using a short computer program, the whole pro- cedure of testing the analytical criteria can be made extremely simple. References 1. 2. 3. 4. 6. 6. L’vov, B. V., “Atomic Absorption Spectrocl-iemical Analysis,” Adam EIilger, London, 1970. Torsi, G., and Tessari, G., Auialyt.Chew., 1973, 45, 1812. Johnson, D. J., Sharp, B. L., West, T. S., and Dagnall, K. M., AYzalyt. Chew., 1975, 47, 1234. Fuller, C. W., Analyst, 1974, 99, 739. Fuller, C. W., Analyst, 1975, 100, 229. Fuller, C. W., Analyst, 1976, 101, 798. NOTE-References 4, 5 and 6 are to Parts I, 11 and IV of this series, respectively.Peak-area Measurement of Transient Signals in Atomic-absorption Spectrometry J. E. Cantle Tnstvumentation Labovatovy ( U K ) Ltd., Technical Services Division, Edgeley Road Tvading Estate, Cheadle Heath, Stockpovt, Cheshive, SK3 OXE Several atomisation techniques that produce transient atom populations are now common- place in routine atomic-absorption spectrometry.Devices such as resistively heated graphite furnaces,1-3 filaments 455 and tantalum ribbons6 have proved useful alternatives to the chemical flame, as they have much greater atomisation efficiencies than conventional pneumatic nebuliser - burner systems. The avoidance of the limiting sample dilution effects that occur in flame cells results directly in much improved detectability for most elements and permits analyses using very small sample volumes, typically 1-100 p1.Micro-sampling techniques that produce transient atom populations and that retain the chemical flame for atomisation include the cup and boat micro-sampljng devices, P.:., that of Delves,' hydride generation methods for the determination of arsenic, selenium, tellurium, etc.8 and discrete-sample nebulisation meth~ds.~-ll The primary characteristic shared by all of these atomisation devices is the transience of the absorption signal generated, when seen inSeptember, 1976 DEVELOPMENTS I N ANALYTICAL ATOMIC SPECTROMETRY 277 contrast to the continuous deflection of the read-out, which is the property of the traditional flame cell.These relatively new types of atomiser will be considered in terms of the measure- ment of the absorption signal produced.When these methods were first used, existing flame- based instrumentation had to be complemented with a chart recorder in order that the new type of display might be recorded and retrieved. The analyst inevitably used a ruler and measured the height of the absorption peak. The concept of integrated absorbance or area measurement was discussed thoroughly by L’VOV~~ and appeared very succinctly in a paper by Aldous et aZ.13 The latter workers observed that almost all experimental work to date had been based on peak absorbance measurements, but these measurements give accurate estimates of analyte concentration only if all of the following conditions are met: (a) the total number of atoms entering the atom cell is constant, requiring a constant degree of atomisation; ( b ) the average residence time in the cell is constant, requiring a stable gas-flow pattern and a reproducible chemical environment in the atom cell; and (c) the atomisation build-up time is constant, requiring a constant rate of sample heating and atom release from the matrix.The first two conditions are likely to be met by most atomisers and many sample matrices.However, the third is more doubtful, since the build-up time is likely to be affected by a number of physical factors, such as the thickness of the nickel cups used for Delves’ cup atomisers, the stability of heating achieved with carbon rod and furnace atomisers and the amount of inert matrix material that can physically impede the atomisation process.With reference to the hydride generation and discrete nebulisation techniques, we can add variables associated with the transfer of discrete volumes of sample molecules into the flame, e g . , viscosity, carrier- gas flow-rate and nebuliser uptake rate. These variables will change the atomisation time, causing large changes in peak absorbance.Integrated absorbance measurements, however, are directly related to an analyte concentra- tion provided that variables (a) and (b) are constant, and these parameters are independent of variations in the rate of release of atoms into the atom cell. Integrated absorbance data should therefore be more precise and interference-free than peak-height data.Aldous et aZ.13 described a computer-linked spectrometer capable of integrating transient absorption signals that was developed to accommodate a Delves’ cup type of atom cell. A considerable improvement in precision was subsequently made available. These types of atomiser and the means by which the signals can best be quantified, are currently being examined in this laboratory.The spectrometers used in these studies are available commercially; the IL 151 and 251 have high-speed digital electronics, specifically designed to cope with transient signals. The spectrometer samples the absorption event at 30-ms intervals andupdates the display at the selected frequency; thus the update time may be &, $, 1 , 4 or 16 s. Hence the peak height can be captured or the absorbance information can be summed to present an integrated area measurement over a period of up to 16 s.The choice of update time is widened considerably when the spectrometer is interfaced to a desk-top calculator. With respect to discrete-sample nebulisation, Manninglo used this technique to measure 11 metals in NBS orchard leaves and Thompson and Goddenll recently measured aluminium, arsenic and tin in steels.In this method, a discrete volume of sample (20-200 p1) is aspirated into a traditional pre-mix chamber - flame system. Sensitivity is similar to that obtained with macro-volumes. The attractions of this very simple procedure are that micro-volumes are employed and solutions with a high dissolved-solids content, such as serum, whole blood and 10% steel solutions, can be handled directly.All work to date has used a chart response with a peak-height measurement. This procedure assumes a constant rate of sample transfer, no viscosity variations and a constant rate of atomisation. Iron, copper and zinc have recently been measured in human control serum in this laboratory by using this method, employing ZOO-$ samples and measuring integrated absorbance over a 4-s time window.The spectrometer was set up for each element in turn and the 4-s manual integration mode selected. In this position, the instru- ment integrated for the period chosen following a 3-s fixed time delay after the integration button was pressed. The blank was obtained by using continuous aspiration of distilled water. The area absorbances of standards (200 pg per 100 ml in each instance) and samples were obtained by passing 200 pl of each through the nebuliser during the 4-s period. The capillary tube was immediately replaced in the distilled-water blank so as to minimise the possibility of The procedure used was as follows.278 clogging of the burner.in Table I. DEVELOPMENTS IN ANALYTICAL ATOMIC SPECTROMETRY PYOC.AnaZyt. Div. Chem. SOC. The results obtained and the precisions of the method are given TABLE I KESULTS FOR DETERMINATION OF METALS I N HUMAN SERUM Element Iielativc staiidard deviation, yo <---*--, Amount Reference Serum Standard found/pg value/pg Zn 1.3 1.4 230 224 I T e 4.4 2.8 216 198 CU 2.1 1.3 189 200 Some preliminary studies indicated that the method can be used to measure lead in whole blood directly.Simultaneous background correction was employed throughout the experi- ments. The integrated absorbance technique introduces considerable versatility in that one can match the sample volume with the counting time, depending on the level expected. In other words one can desensitise by taking a smaller aliquot or lengthening the count time, or both. The Delves’ cup technique for blood-lead determinations generally enables a routine precision of 6-7% to be achieved when recording peak absorbance via a chart recorder. This figure can be reduced by using automatic dilution methods, but significantly a 4-s area measurement has been demonstrated to improve these figures, typically by about 2%. With respect to furnace atomisation, currently available power supplies and programmers are sophisticated and reproducible heating rates can be assumed. Therefore, the measurement of peak height will be an adequate representation of the absorption signal from all simple analy- ses. Very often, however, an interference may be encountered that is manifested in a slowing of the rate of release of analyte atoms into the cell. In these instances, a peak-area measure- ment may enable the interference to be compensated for. It is probably wisest to observe the corresponding peak shapes via a transient recorder before progressing any further.l* In the determination of arsenic and related elements by hydride generation and atomisation in a low-temperature flame, a transient signal is produced as constant masses of arsine, etc., are sequentially atomised. An integrated area absorbance technique is the logical way of handling this type of absorption signal and will compensate for any changes in carrier-gas flow- rate and flame condition that might affect the atomisation rate. Generally, a noisy base-line is a characteristic of this technique and the integration mode described enables a consistent approach to the establishment of a base-line zero on either side of the absorption signal. Conclusion An increasing proportion of routine and published methods of atomic-spectrometric analysis involves atomisers handling discrete sample masses or volumes. The signals generated are thus transient in nature. The observation that Aldous et aZ.13 made in 1973 is still valid, namely that, in the main, these signals are quantified in terms of a peak absorbance. Certain implica- tions of measurement have been considered with reference to four atomisation devices com- monly used and some results have been presented from a currently running project. An integrated area absorbance measurement can now be readily made using currently available spectrometers. Particularly with respect to the very simple devices, discrete nebulisation of micro-volumes, cup sampling techniques and hydride generation, the use of this mode can demonstrably improve the quality of the analysis. 1. 2. 3. 4. 5. 6. 7. 8. 9. References L’vov, B. V., Spectvochim. Acta, 1961, 17, 761. Massmann, H., .$$ectvochim. Acta, 1968, 23B, 215. Cantle, J. E., Lab. Equip. Dig., 1975, 13(3), 97. Cantle, J. E., and West, T. S., Talanta, 1973, 20, 459. Amos, M. D., Bennet, P. &4., Brodie, K. G., Lung, P. W. Y., and MatouSek, J. P., Analyt. Chem., 1971, Hwang, J. Y., Ullucci, P. A., Smith, S. B., Jr., and Malenfant, A. L., Analyt. Chem., 1971, 43, 1319. Delves, H. T., Analyst, 1970, 95, 431. Fernandez, I;. J., Atom. Absovption Newsl., 1973, 12(4), 93. Sebastiani, E., Ohls, K., and Riemer, G., Z. Analyt. Chem., 1973, 264, 106. 43, 211.September, 1976 DEVELOPMENTS I N ANALYTICAL ATOMIC SPECTROMETRY 279 10. 11. 12. 13. 14. Manning, D. C., Atom. Absorption Newsl., 1975, 14, 99. Thompson, D. C., and Godden, R. G., Analyst, 1976, 101, 96. L’vov, B. V., “Atomic Absorption Spectrochemical Analysis,” Adam Hilger, London, 1970. Aldous, K. M., Mitchell, D. G., and Ryan, F. J., Analyt. Chem., 1973, 45, 1990. Warren, J., and Regan, J. G. T., Analyst, 1975, 101, 220.

ISSN:0306-1396    

DOI:10.1039/AD9761300258

出版商:RSC    

年代:1976

数据来源: RSC

摘要:

September, 1976 DEVELOPMENTS I N ANALYTICAL ATOMIC SPECTROMETRY 279 Why Plasma Torches?-Invited Lecture S. Greenfield Albright and Wilson Ltd., Industvial Chemicals Division, P.O. Box 80, Oldbury, Wavley, West Midlands, B69 4LN The immediate answer to the question “Why plasma torches?” or “Why use plasma torches?” is “because they have exceptionally good properties as emission sources in spectroscopy,” and these desirable properties include high sensitivity, ability to excite any element, low detection limits, freedom from matrix effects, wide linear range, stability, convenience in operation and simultaneous multi-element excitation.This paper tries to show that many of the features possessed by the plasma follow naturally from the nature of plasma discharges in general and of the type that we use in particular.Consider first the sensitivity, which is defined as the slope of the calibration graph, so that for a linear calibration the sensitivity is the ratio of the net signal to the concentration that produces it. As the temperature is increased the sensitivity increases, reaches a maximum and then decreases. The temperature at the maximum sensitivity is called the norm tem- perature.The physical explanation of this behaviour is simple qualitatively, although the calculations are complicated. The population of atoms in any energy level depends on the number that have been excited to this level from lower energy states, and this number increases with temperature; so, too, do the losses from this level to higher energy states.The curve of sensitivity veysus temperature is thus the result of the balance between acquisition from lower levels and loss to higher levels. Fig. 1 shows the results of some calculations we have made on the variation of sensitivity with temperature for a few spectral lines. I .g Na 589.0 nm Zn 307.2 nm Ar 763.5 nm -g I (2.1 eV) (8.1 eV) (13.1 eV) 0 5 10 15 20 25 Temperature/K X lo3 Fig.1. Variation of sensitivity with tcrnpera- ture. Table I gives a list of a few lines of interest to us and the norm temperatures that we have It can be seen that even for the most easily excited lines the norm temperatures For anything but the most easily excited elements, a source hotter calculated. are as high as 6 000 K. than this is clearly advantageous.TABLE I SOME LINES OF INTEREST AND NORM TEMPERATURES Element Wavelength/nm Excitation energy/eV Norm temperature/K Na 589.0 2.10 6 000 285.2 4.35 7 500 249.7 4.96 10 500 396.8 9.23 12 000 M8‘ (11) 280.3 12.06 16 000 Eg Ca (TI)280 DEVELOPMENTS IN ANALYTICAL ATOMIC SPECTROMETRY Proc. AnaZyt. Div. Chem. SOC. Inductively coupled plasmas reach temperatures in the region 8 000-12 000 K; the actual temperatures reached will depend on the geometry, the gas flows and the power supplied.In the annular configuration of the plasma that we have always used, the gas stream carrying the sample atoms flows through a central cooler tunnel that is surrounded by the very hot doughnut-shaped region where the applied power is dissipated. When this gas flow emerges from the tunnel, it forms a well defined and narrow tail-flame.Because of its small radius, the emission of the analyte atoms per unit of surface area is high, and this effect contributes to the sensitivity. The sample atoms in the central tunnel are heated by radiation and conduction from the fireball and the system is analogous to a tube furnace. Consider a very crude model (Fig.2), in which the walls of the tunnel are at a constant temperature, To, for its visible length and that beyond this length the temperature decreases exponentially with height. Assume that the sample atoms a t a given height are all a t the same temperature, T , which is the result of heating at a rate proportional to the difference between this temperature and the temperature of the surrounding tunnel.Then, the greater the length of the plasma the greater is the maximum temperature and the closer to the plasma is the position of the maximum. Varia- tion of the plasma temperature, To, at least on this model, simply makes a proportional variation in the sample temperature at all heights. It does not alter the position of the maxi- mum temperature, only its value.An increase in power that causes an increase in the tempera- ture and size of the plasma will result in an increase in the temperature of the sample atoms owing to both effects and a movement of the maximum temperature to a position nearer the discharge. This model is consistent with our experimental observations; a consequence is that in observing how the sensitivity varies with power we cannot ignore the effect of height of observation.To take account of the movement of the maximum temperature, for a given setting of power we scan over a range of heights and choose the maximum value of signal to plot as a function of power. Fig. 3 shows how emission varies with power for sodium (589.0 nm), zinc (307.2 nm) and aluminium (281.6 nm), with excitation energies of 2.1, 8.1 and 17.7 eV, respectively.For the easily excited sodium line, the norm temperature is apparently reached at about 2.5 kW. For all but easily excited lines, the sensitivity increases with power up to high values, and for lines difficult to excite a high power is required in order to obtain an appreciable signal. So far, these correlations of sensitivity with temperature have been indirect, and are really correlations with power.However, we have made some temperature measurements using the It is easy to be misled unless this kind of scanning is used. Top of Height- plasma Fig. 2. Model for sample heating. The longer the plasma, the higher is the temperature attained by the sample and the closer to the top of the plasma does the maximum occur.1 o4 u 0 1 2 3 4 5 6 Power in plasmalkw Fig. 3. Net signal, maximised with respect t o height, as function of power in the plasma: (a) Na 589.0 nm (2.1 eV); (b) Zn 307.2 nm (8.1OeV); and (c) A1 281.6nm (17.7 eV). m, Small generator; and 0, large generator.September, 1976 DEVELOPMENTS I N ANALYTICAL ATOMIC SPECTROMETRY 28 1 zinc lines with wavelengths of 307.2 and 307.6nm.This is achieved by comparing the intensity of emission of the two lines. Fig. 4 shows how the excitation temperature calculated in this way and maximised with respect to height in the tail-flame varies with power. It can be seen that it increases, as expected, and that the dependence is approximately linear in the power range for which a stable plasma is obtainable with our apparatus, namely, greater than about 0.8 kW.I I I 2 4 6 Power/kW Fig. 4. Measured effective excitation temperature as a function of power. 34 I- -. 3 0 5 10 15 20 Tempersture/K x lo3 Fig. 5 . Logarithmic plot of the continuum emission from argon in arbitrary units as a function of temperature. The temperature that we have determined here is a value effectively averaged along the line of sight.While it is possible, by sampling the emission from different lateral positions in the source and subsequently applying an Abel inversion, to recover the true radial distribution of temperature, we have not done so. When a plasma source is used for spectroscopy, what is of ultimate importance is what the spectrometer sees, and this is of necessity a volume extended along the line of sight.On this basis, we can say that the “effective temperature” that we have determined is an appropriate parameter and that its linearity with power is an adequate justification for saying that a correlation of sensitivity with power implies a correlation with temperature. This is defined as a factor x times the noise of the background (taken to be its standard deviation) divided by the sensitivity.What is wanted to give a low limit of detection is a combination of low noise on the background and high sensitivity. The major components of the noise in the background radiation from the source are expected to increase or decrease with the background signal or its square root. Hence the lower the background, the smaller is its expected noise.A high-temperature plasma gives a very high continuum background; indeed, this back- ground is responsible for the dazzling appearance that a high-power plasma possesses. The continuum is due largely to two processes : recombination of ions with free electrons, and the bremsstrahlung due to deceleration of electrons in the neighbourhood of heavy ions. The continuum thus depends strongly on the electron density.Fig. 5 gives the results of an approximate calculation of the continuum for argon ; the dependence on temperature is very strong. Indeed, if the temperature decreases from 10 000 to 8 000 K the continuum is reduced to one hundredth of its value and, if the temperature further decreases to 7 000 K, it is reduced to one thousandth of its value. It is likely that it is this dramatic decrease in emission with temperature that gives the bright part of a plasma its sharp, well defined boundary.The background that would be obtained by using the brilliant fireball as a source would be intolerably high. With an annular plasma, the sample atoms emerge in a narrow tail-flame, which projects some distance beyond the fireball, and it is this tail-flame, in a region some distance away from the fireball, which is used as the spectroscopic source.Here the con- tinuum background is very low, as can be seen in Fig. 6, which is a recorder trace showing Sensitivity is also one of the factors that determine the limit of detection.282 DEVELOPMENTS IN ANALYTICAL ATOMIC SPECTROMETRY Proc. AYzalyt. Div. Chem. Soc.Height- Fig. 6. Tracings from recorder charts 01 gross signal and background as a function of height of observation. signal and background against height in the tail-flame. Because of this effect, we expect the noise on the background to be low, which, coupled with the high sensitivity, means that very low limits of detection are obtainable. Fig. 7 shows the effect of using argon as the coolant compared with nitrogen, which we usually use, The background is further reduced. CN 387.1 Ca 393.4 Al 396.1 Ar/Ar plasma Ar/N2 plasma Fig.7. as coolant. Comparison of spectra obtained with argon and with nitrogen Our reason for using nitrogen in spite of this effect is that extra sensitivity is obtained, al- though the results are more complicated than we once thought.Table I1 shows, for a number of lines in decreasing order of excitation energy and for two different powers, the ratio of the signal to noise ratios obtained with nitrogen to those obtained with argon as the coolant. If this ratio has a value greater than unity, then nitrogen as coolant should give the lower limit of detection; if it is less than unity, argon. Note, incidentally, that for a power of 3 kW, three of the lines listed are not detected.Many more lines have been investigated than appear in Table 11, with excitation energies ranging from 2.9 to 16.3 eV. We have always found that nitrogen is the preferred coolant at 3 kW. At 6 kW, nitrogen is preferred for lines with high excitation energy, but argon for those with low energy. The net signal and hence the sensitivity for all of the lines studied is greater TABLE I1 RATIO OF SIGNAL TO NOISE RATIOS (S/N) FOR NITROGEN TO THOSE FOR ARGON AS COOLANT S/N (nitrogen) SIN (argon) Wavelength/ Excitation r--A-, Elemcnt nm energy/eV 6.0 k W 3.0 kW Mn 259.0 16.3 -* --t Ni 239.5 14.5 1.71 --t Nb 258.4 12.9 1.12 --t V 289.3 11.4 0.82 1.52 cu 324.7 3.8 0.65 1.67 Cr 427.4 2.9 0.70 2.81 * Not observed with argon.t Not observed with argon or nitrogen.Scptcmhev, I976 DEVELOPMENTS I N ANALYTICAL ATOMIC SPECTROMETRY 283 for nitrogen, implying that the tail-flame is hotter. We have seen that both sensitivity and background increase with temperature, but at different rates. It seems that at the positions giving the maximum signal to noise ratios, the higher temperature produced by the use of nitrogen as coolant increases the noise more than it does the signal from elements of low excitation energy, so that for these elements the lower temperature obtained with argon should lead to lower detection limits, even though the sensitivity with nitrogen is higher.For lines with even lower excitation energies, the sensitivity also will be less with nitrogen, for it was easy to exceed the norm temperature for sodium (589.0 nm), with an excitation energy of 2.1 eV.Nevertheless, nitrogen gives better sensitivity for nearly all elements and better detection limits for high-energy lines and we use it for these reasons. The liigh temperature possessed by an inductively coupled plasma has another consequence, namely, freedom from chemical interference; by this is meant the reduction in the intensity due to the formation of refractory compounds, which effectively locks up atoms and prevents them from contributing to the desired emission. It is the kinetic or gas temperature that is important here, as it is by the collision of particles of high kinetic energy that the compounds are broken up. Some plasmas can therefore be efficient emission sources because they have high excitation temperatures but nevertheless still display chemical interferences; if a fair amount of power is dissipated in the plasma, a sufficiently high gas temperature to dissociate the refractory molecules can be expected.We have found no exceptions to this rule in practice. This effect also applies, of course, to radicals such as C, and Fig.8 shows how the well known Swan bands are reduced with power. 3 kW 4 kW 5 kW 6 kW 7 kW Swan bands Fig. ti. Reduction in the intensity of the Swan bands with powcr. All of the desirable features described up to now depend on a high temperature. Some also depend on the particular configuration of the plasma, i.e., the annular fireball and emergent narrow tail-flame.We ascribe this effect to the low background at the low end of the range and to the absence of self-absorption a t the high end. We believe that at heights of interest there is little mixing between the central gas stream carrying the sample and the plasma gas in which the discharge takes place, so that all of the analyte atoms are confined to the tunnel and the tail-flame. In particular, there are no analyte atoms in the cooler regions that surround the tail-flame, so that self-absorption should be low.This explanation is borne out in practice; calibrations that are linear over four orders of magnitude are common. Another factor that can be seen to depend on the configuration of the source is its stability. The stability, of course, depends on other factors also: we obviously need a stable generator and stable gas flows.Its dependence on the annular configuration stems from the fact, already noted, that the sample is confined to the tunnel when passing through the discharge region. In a high-frequency inductively coupled plasma, the power is dissipated largely in the outer layer of the discharge, so that the interior is effectively screened from the applied field.We expect, There are other features that depend very largely on this configuration. The first of these features is the wide linear dynamic range of the calibrations obtained. An illustration is given in Fig. 9.284 DEVELOPMENTS IN ANALYTICAL ATOMIC SPECTROMETRY Proc. Analyt. Din. Chem. SOC. 10 1 v) c, C 3 0 0 000 000 100 10 1 0.0001 0.001 0.01 0.1 I 10 100 Percentage composition Fig.9. Calibration graph obtained for copper in a variety of matrices. Points: 1, BCS 374, phosphor bronze, 89.5% Cu; 2, BCS 264, leaded bronze, 80.5% Cu; 3, BCS 216/2, Duralumin, 4.56% Cu; 4, BCS 181/2, 4% Cu/Al alloy, 3.96% Cu; 5, BCS 300, A1 alloy, 1.28% Cu; 6, BCS 263/1, 5% Mg/Al alloy, 0.09% Cu; 7, BCS 262, 10% Mg/Al alloy, 0.030/, Cu.therefore, the electrical properties and hence the power in an annular plasma to be largely independent of the sample. It is well known that in plasmas of very low power, where a power of, say, only 200 W is supplied, desolvation is necessary in order to prevent the plasma from being extinguished. The reason is simply that there is insufficient energy to maintain the plasma and dissociate the water molecules simultaneously.Even when the generator bas more power, difficulties are sometimes encountered. It is obvious that the dissociation, excitation and perhaps ionisation of the sample must use up some energy and that the amount used up may vary from sample to sample. If the energy required to do this is a negligible fraction of that supplied, the change in plasma properties, such as size and temperature, will also be negligible.The type of stability discussed here means that the plasma will not go out and that the sensitivity will remain constant for a wide variety of sample solutions. This feature really overlaps with another desirable feature, namely, ease of operation, for if this type of stability prevails, then one can start the plasma, load samples and standards on a turntable and leave everything to run automatically.Another aspect of stability depends on the power. This is one reason why a high power is desirable. Application of an Atmospheric-pressure Ion Source to the Measurement of Isotope Ratios at Trace Levels in Solution by Mass Spectrometry R. J. Anderson and A. L.Gray Applied Reseavch Labovatovies Ltd., Wingate Road, Luton, Bedfordshire Atmospheric-pressure ionisation sources1 have previously been reported for the molecular analysis of organic liquids and vapours. For elemental analysis, a new technique has been de~cribed,~,~ plasma sampling mass analysis (PSMA) ,4 permitting direct mass spectrometry of aqueous solutions. Plasma Ion Source and Mass Spectrometer A detailed description of the techniques has been given el~ewhere.~,~ Briefly, a high- temperature d.c.plasma is used as an ionisation source, solution samples being introduced with an ultrasonic nebuliser. The plasma flame is sampled with a small orifice and the extracted ions are directed through a quadrupole mass spectrometer. Transmitted ions are detected by ion- counting techniques and the spectrum produced by scanning the mass analyser is displayed on a chart recorder or fed to a synchronous multi-channel analyser.September, 1976 DEVELOPMENTS I N ANALYTICAL ATOMIC SPECTROMETRY 285 Application The technique has been found to provide sensitivities ranging from lo2 to lo5 or more counts s-l for aqueous solutions of 1 pg ml-l for a wide range of elements, with background levels at a few counts per second.Aqueous solutions of samples are introduced directly a t atmospheric pressure and a high sample throughput is possible. There appears to be considerable potential for trace analysis, and attention is currently being concentrated on exploring the source characteristics. The method is directly suitable for isotope dilution analysis and the determination of isotope ratios with moderate precision. An investigation of the performance on isotope-ratio measurements on some mineral samples has been carried out.Isotope-ratio Measurements A number of samples of lead- and strontium-bearing minerals were obtained in the form of acidic solutions. These solutions were diluted to about 20 pap.m.of the element with de-ionised water. The only other sample treatment was to add about 2% of hydrogen peroxide to the lead samples in the nebuliser sample cell in order to reduce memory effects due to adsorption of lead on the cell and nebuliser walls. The analyser was set to scan a width of 7 a.m.u. across the isotope peaks of interest, from 203-209 a.m.u. for lead and 83-89 a.m.u.for strontium. The maximum scanning rate of the analyser (about 3 x s a.m.u.-l) was used so that the repetition rate of the scans was about 50 per second. The start of each scan was synchronised with the start of the multiscaler sweep and the complete scan was stored in about 200 channels. At the end of the integration period, during which the spectrum was monitored as it accumulated on the cathode-ray tube display, the spectrum was typed out.The channel contents were then added for each peak and corrected for background, using an average background count per channel calculated from the signal accumulated from the blank regions of the spectrum. The isotope ratios were then calculated from the ratios of integrated counts. Lead Samples and Results The first runs were carried out on a J-type lead sample to determine the experimental reproducibility of the method.For each run, the sample cell was loaded with 7 ml of sample solution and a 30-min integration carried out. A series of 10 runs was made and the mean and the experimental standard deviation (oexp.) calculated for each ratio. The standard deviation to be expected from counting statistics (on) was also calculated for each run and its mean value was compared with aexp.These results are also shown in Table 11, as relative abundances with 95% confidence limits ( 2 4 , as are results of measurements on the same material made by Kirchoff .5 The values for a typical J-type (Missouri) lead obtained by Chow et aL6 are shown in both tables. The results of these runs are shown in Table I.TABLE I ISOTOPE-RATIO REPRODUCIBILITY ON A JOPLIN ORE SAMPLE Parameter pb206 Pb20G p 5 aexp. 0.000 70 0.003 9 0.009 7 Pb204 Pb207 Pb208 ~ - Mean ratio 0.046 89 0.733 4 1.861 1 a, (mean) 0.000 57 0.002 5 0.005 0 Missouri leadG 0.045 91 0.722 0 1.871 9 TABLE I1 RELATIVE ABUNDANCES FOR A JOPLIN ORE SAMPLE Relative abundance, yo 7 Method or T----------A---pp-- author Pb204 Pb206 Pb207 Pb2OS PSMA 1.29 f 0.01 27.46 -l 0.06 20.14 f 0.04 51.11 0.07 Kirchoff - 27.58 f 0.28 __ 51.17 f 0.51 Missouri leadG 1.261 27.47 19.84 51.43286 DEVELOPMENTS IN ANALYTICAL ATOMIC SPECTROMETRY PYOC.Artznlyt. Din. Chew. SOC. On the mean ratio calculated from 10 runs, they can be reduced by 410, and this was done in deriving the confidence limits on abundance given in Table 11.Agreement of the mean values with the other determination is well within the quoted limits. The determinations on the Joplin sample each used less than 8 ml of solution, containing about 160 pg of metal. Further measurements were carried out on two NBS standard materials, SRM981 and SRM983, which are common and radiogenic lead, respectively. Mean values of three runs on each are shown in Table 111, together with the true values.' Except for the value for lead-204 in SRM983, where insufficient counts were obtained because of the low abundance, the agreement is reasonably good.With both standards, however, the abundance of the main peak is slightly low. This result is attributed to some loss of counts owing to gain shift in the channel multiplier detector at the peak count rate.A low value for one peak falsifies the abundances of the others and also affects the ratio in which it appears. This effect is particularly serious for SRM983, because the affected peak is that for lead-206, which appears in all of the ratios. If the low value is corrected, the other isotopes and the ratios appear much closer.Corrected values are shown in the last line of Table I11 for each standard. The values for standard deviation of the ratios in Table I relate to a single run. TABLE 111 ISOTOPE RATIOS AND RELATIVE ABUNDANCES FOR NBS STANDARDS Standard Value SRM 981 True7 PSMA PSMA (corr.) PSMA PSMA (corr.) SKM983 True7 Isotope ratios ------- 7 PbZo4 PbZu7 Pb208 PbZo6 pb206 Pb206 0.059 00 0.914 64 2.168 1 0.058 76 0.913 9 2.138 4 0.058 76 0.913 9 2.167 7 0.000 37 0.071 20 0.013 62 0.000 3 0.072 77 0.013 87 0.000 3 0.071 29 0.013 56 7-- Pb204 1.424 5 1.43 1.42 0.034 2 0.03 0.03 Relative abundance, yo Pb206 Pb207 24.144 7 22.082 7 24.32 22.23 24.15 22.07 92,149 7 6.561 1 92.00 6.70 (92.15) 6.57 ----A --I Pb208 52.348 1 52.01 (52.35) 1.255 0 1.27 1.25 Strontium Samples and Results Samples of celestite and strontianite were analysed and adequate peak separation was obtained at a lower resolution than was needed for the lead samples.This result, however, gave higher count rates, which, although desirable for good counting statistics for the low- abundance isotopes, caused counting losses on the main peak. Each run was therefore made twice, the second at a higher resolution, giving a count rate reduced by a factor of approximately 10.Each integration was made for 5 min and the high count rate run used for strontium-84 -86 and -87 values and the lower rate values for strontium-86, -87 and -88. The combination of these two sets of runs enabled adequate statistics to be obtained over all four isotopes. The re- sults are shown in Table IV togerher with estimated 95% confidence limits on the ratios.TABLE IV ISOTOPE RATIOS AND RELATIVE ABUNDANCES FOR STRONTIUM SAMPLES Isotope ratios - ~ . ~ A - . . . - - ~ - 7 Relative abundance, % A Srs4 SrS7 SrS8 I 7 -- Sample SrS6 SrsO Srs6 Srs4 Sr86 Srs7 Srss Celestite 0.058 7 0.695 4 8.089 3 0.596 10.163 7.067 82.208 Strontia- 0.056 9 0.692 4 8.107 2 0.577 10.145 7.025 82.252 - &- 0.000 52 0.006 5 0.07 4 nite & 0.000 56 f 0.007 0.056 No contribution from rubidium was found in these samples.Examination of the back- ground in the position of the rubidium-85 peak enabled the upper limit to the contribution of rubidium-87 to the strontium437 to strontium-86 ratio to be set at 0.000 12.September, 1976 DEVELOPMENTS I N ANALYTICAL ATOMIC SPECTROMETRY 287 Conclusions It has been demonstrated that the new technique of plasma sampling mass analysis can be used to make isotope-ratio measurements with a relative precision of the order of 0.5%. The ease of sample presentation and minimal sample preparation suggest that the technique should have useful application in mineralogical or pollution studies for which large numbers of samples must be handled.Higher precision can be expected with improved ion-detection methods. The interest and assistance of Mr. D. Hagger of Applied Research Laboratories Ltd., Mr. P. J. Moore of the Institute of Geological Sciences, London, and Dr. G. F. Kirkbright of Imperial College, London, who provided samples, is gratefully acknowledged. References 1. 2. 3.4. 5. 6. 7. Caroll, D. I., Dzidic, I., Stillwell, R. N., Horning, M. G., and Horning, E. C., Analyt. Cheun., 1974, 46, Gray, A. L., Analyt. Chem., 1975, 47, 600. Gray, A. L., Analyst, 1975, 100, 289. Applied Research Laboratories Ltd., British Patent No. 1261596, 1969. Kirchoff, H., Spectrochiun. Acta, 1969, 24B, 235. Chow, T. S., Snyder, C. R., and Earl, J. I,., “Isotope Ratios as Pollutant Source and Behaviour Indicat- Catanzaro, E.J., Murphy, T. J., Shields, TV. R., and Garner, E. L., J. Res. Natn. Bur. Staizd., 1968, 706. ors,” IAEA, Vienna, 1974. 72A. 3. Application of Flame and Flameless Atomic-absorption Spectroscopy to Routine Rock Analysis and a Study of Some Matrix Interference Effects in Fla meless Atom ic-a bsor pt ion Spectroscopy John Warren and Michael P.Harrison* Depavtment of Industvy, Labovatovy of the Govevnvnent Chemist, Covnwall House, Stalnfovd Stvcet, London, SE1 9NQ The underlying theme of the presentation was based upon some interference effects recently investigated at the Laboratory of the Government Chemist. The work comprised a study of some specific aspects of both flame and flameless atomic-absorption spectroscopy and can be conveniently arranged into three separate categories, as discussed below.Atomic-absorption Determination of Strontium in Silicate Rocks : A Study of Major Element Interferences in the Nitrous Oxide - Acetylene Flame A fuller account of this work has already been published in The A ~ ~ a l y s t l and should be referred to for more detailed information. A study was made of the major element interferences associated with the atomic-absorption determination of trace amounts of strontium in silicate rock using a nitrous oxide - acetylene flame.Aluminium causes a suppression of the strontium signal, while calcium and magnesium act as partial releasing agents, thus reducing the effect of the aluminium. Exact matching of samples and standards can be avoided by the use of lanthanum, which has been shown to be an effective releasing agent in the nitrous oxide - acetylene flame.Strontium levels in the US Geological Survey rocks AGVI, BCRI and GSPI, in the range 244-670 pg g-l, have been determined with a precision of better than 2%. Spiking experi- ments (100 pg 8-l of strontium) produced recoveries ranging from 98.2 to 101.6%.Strontium can be measured in silicate rock with a limit of determination of 0.6 pg g-l. Determination of Trace Amounts of Gallium in Silicate Rock by Flameless Atomic-absorption Spectroscopy The level of gallium in silicate rock is normally of the order of 10 pg g-l, which is too low for * Present a.ddress : Hutterworth Microanalytical Consultancy Ltd., Teddington, Middlesex.288 DEVELOPMENTS IN ANALYTICAL ATOMIC SPECTROMETRY Proc.Analyt. Div. Chew. SOC. its direct determination by flame atomic-absorption spectroscopy using a typical solution con- taining 1% m/V of dissolved rock. A solvent extraction technique,2 prior to flame atomic absorption, can be usefully employed to effect a concentration increase, but even so, the sensitivity of the method is not really sufficient for the determination of gallium on a routine basis.The higher sensitivity of flameless atomic absorption readily lends itself to the determination of gallium, and the solvent-extraction procedure performs a useful function in "cleaning-up" the sample and thus minimising any matrix interferences. From the analytical scheme shown in Fig. 1 it can be seen that the rock solution is made approximately 6 M with respect to hydrochloric acid and then the gallium is extracted as its chloro complex into isobutyl methyl ketone (4-methylpentan-2-one).Unfortunately, iron is also extracted as iron(II1) chloride, although this effect can be minimised by adding a reducing agent to convert iron(II1) to iron(I1). Ascorbic acid (0.5 g) has proved to be a convenient reducing agent and later work showed that its presence in the ketone extract is also beneficial as it helps to combat matrix interferences.Rock solution Conc. HCI, 20 ml 20 ml I Solvent I Analysis of extract (20 pl) Dry, - 90 "C Atomise, - 1900-1 950 " C HGA Ash, - 440 "C -6 M HCI Reduce Fe3+-.Fe2' extract Fig. 1. Flameless determination of gallium in silicate rock.Level - 10-20p.p.m. of gallium, too low for flame atomic absorption. Extraction as chloro com- plex. A Perkin-Elmer HGA-72 graphite furnace installed in a Model 403 atomic-absorption spectrophotometer fitted with the optical modification was used for the analytical measure- ments. The absorption signals were measured using a conventional chart recorder (0.3 s full-scale deflection) and computed from the peak height.Because of the spreading of the solvent extract in the conventional graphite tube and the deleterious effect that this spreading has on analytical precision, Perkin-Elmer organic, grooved tubes were used throughout the study. A calibration graph was constructed from the measurements obtained from the extracts of a set of synthetic working standards3 containing the equivalent range of 0-50 pg g-l of gallium in rock.The gallium content of an unknown rock extract is determined directly from the calibration graph. Good results were obtained for some US Geological Survey rocks and spiking experiments indicated a 98% recovery at the 10 pg g-l level. The precision of the method using manual injection is 2 4 % (relative standard deviation).Deuterium area background correction was employed.September, 1976 DEVELOPMENTS IN ANALYTICAL ATOMIC SPECTROMETRY Peak-shape Studies in Flameless Atomic-absorption Spectroscopy and the Use of Ascorbic Acid to Combat Some Matrix Interferences 289 Matrix interferences of the inter-element type are known to occur in flameless atomic- absorption spectroscopy, but little is known about the exact physico-chemical processes in- volved.However, at least two main mechanisms can be postulated: (1) a depletion of the atomic vapour owing to the formation of molecular species (volatile or involatile), leading to a reduction in the height of the absorption peak; and (2) a delay in the formation of the atomic vapour, leading to both height reduction and broadening of the absorption peak.It can be seen, therefore, that the shape of the absorption peak may provide an indication of the nature of any interferences encountered. An electronic system for recording the shape of flameless atomic-absorption peaks has been constructed in this laboratory and utilised for both routine and research work. Fig. 2 shows a schematic diagram of the system, which is based on a transient recorder, a device that is capable of catching and storing a transient response and feeding it out at a much slower speed. I I I Atomic-absorption A I spectrometer I-‘- \ Fig.2 . Transient recorder system. The output signal from the atomic-absorption spectrophotometer is simultaneously meas- ured with a conventional chart recorder and a transient recorder. The chart recorder merely displays the peak as a transient “kick” and is used to measure peak height in the normal way.The transient recorder, on the other hand, can be used to output the scan of the stored peak over a period of 200 s and therefore displays the actual shape of the atomic-absorption signal. A cathode-ray oscilloscope, linked to the transient recorder, automatically gives a visual dis- play of the shape of the stored peak, while if a permanent record is also required it can be traced on the chart recorder by means of an optional link, fitted with a switch, to the transient recorder.This means that an analytical signal can first be displayed on the chart recorder as its peak height, and then the same signal can be traced on the chart recorder (via the transient recorder) as its peak profile.Integration of this profile would obviously give a measure of peak area. This device is a useful adjunct to the conventional chart recorder and has proved to be a valuable diagnostic aid in our study of matrix interferences. A study carried out by ourselves on the addition of water-soluble organic materials to aqueous sample solutions4 has indicated an enhancement in the height of the atomisation peak for copper, gallium and lead.This is probably due to the in situ formation of a mixture of sample and finely divided carbon during the thermal destruction stage, and the subsequent improvement in reduction that occurs during the atomisation stage. The addition of 1% m/V of ascorbic acid to a lead solution (0.1 pg ml-1) substantially removes the interference effects of calcium, strontium, magnesium and barium (100 pg ml-l) in both 0.1% nitric and hydro- chloric acid, as shown in Table I.290 DEVELOPMENTS I N ANALI’TICAL ATOMIC SPECTROMETRY PYOC.AfiaZyt. Div. Cherut. S O C . TABLE I COMPARISON OF PEAK HEIGHTS* FROM 0.1 pg ml-1 OF LEAD WITH AND WITHOUT 1% OF ASCORBIC ACID IN SOLUTIONS COXTAINING 100 pg ml-l OF OTHER METALS 0.1 yo Hydrochloric acid 0.1 yo Nitric acid 7 ---- A -_-- ~ ---- L -____ Sample No ascorbic acid 1 yo Ascorbic acid No ascorbic acid 1 yo Ascorbic acid Pb 100 141 100 144 Pb + Ca 64 142 83 142 Pb + Mg 12 231 133 137 Pb -t Sr 77 138 86 139 Pb t Ba 83 143 99 145 * Mean of a t least four injections; coefficient of variation better than 50/,.The usefulness of peak-shape studies is clearly shown in Fig. 3, which illustrates the inter- ference of calcium on lead in both nitric and hydrochloric acid, and the beneficial effect of ascorbic acid. Although the interference effect of calcium is less in nitric acid than in hydro- chloric acid, in terms of peak height, the peak-shape study shows severe peak broadening (peak R) in the nitric acid system, indicative of a delay in the formation of the atomic vapour. Fig. 3. Peak shapes and interferences. Peak shapes obtained from 0.1 pg ml-1 lead solutions, showing the interference effects of 100 pg ml-1 of calcium in both 0.1 yo nitric and hydrochloric acids, and the influence of 1 ”/o m/ V of ascorbic acid on this system : A, 0.1 pg ml-l of Pb in 0.1 yo HCl or HNO,; B, 0.1 pg ml-1 of Pb + 100 pg ml-I of Ca in 0.1% HNO,; C, 0.1 pg ml-I of Pb + 100 pg ml-I of Ca in O . l O / , HC1; D, 0.1 pg ml-l of Pb + 100 pg ml-I of Ca in 0.1 %) HC1 or HNO, plus lo/, yn/V ascorbic acid; and E, 0.1 pg ml-l of Pb in 0.1% HC1 or HNO, plus 1% m/V ascorbic acid. In the hydrochloric acid system the peak symmetry (peak C) is retained, possibly indicating a depletion of the atomic vapour by the formation of lead chloride. Addition of ascorbic acid to both the calcium in nitric and hydrochloric acid systems produces an enhancement of the lead peak height and also restores the peak symmetry (peak D). Furthermore, addition of ascorbic acid to lead alone, in either nitric or hydrochloric acid solution, produces a peak (peak E) of height and shape almost identical with those obtained with the ascorbic acid plus calcium plus lead interferent systems, thus indicating the effectiveness of ascorbic acid in removing the interferences. The lead content of some natural drinking waters has been successfully determined against aqueous lead standards by the addition of 1% m/V ascorbic acid to both samples and stand- ards in order to overcome the matrix interferences associated with the hardness in water. References 1. 2. 3. 4. Carter, D., Regan, J . G. T., and Warren, J., Analyst, 1975, 100, 721. Cresser, M. S., and Torrent-Castellet, J., Talanta, 1972, 19, 1478. Warren, J., and Carter, D., Can. J . Spectrosc., 1975, 20(1), 1. Regan, J. G. T., and Warren, J,, Analyst, 1976, 101, 220.

ISSN:0306-1396    

DOI:10.1039/AD9761300279

出版商:RSC    

年代:1976

数据来源: RSC

摘要:

Septembev, 1976 AD DISTINGUISHED SERVICE AWARD 291 Si her Medal Nominations are invited for the award of the SAC Silver Medal, which is for the encouragement of young scientists working in any field covering the practice and teaching of analytical chem- istry. The award is accompanied by a cash prize and is normally made annually to the candidate who, in the opinion of the Council, has made the greatest contribution and whose work has made the most significant impact in any branch of analytical chemistry.In addition, the future promise of the candidate is taken into consideration. It is hoped to provide an opportunity for the successful candidate to deliver a lecture to the Division on a suitable occasion subsequent to the presentation of the Medal. 1. The award of the Silver Medal will normally be considered annually by the Honours Committee, acting on behalf of the Council of the Division, but an award may not be made if it is considered that the work of no candidate reaches the required standard. 2.Candidates must be British subjects of 38 years of age or under in the year in which the award is considered. Evidence of age will be required.3. The merits of the candidate’s work may be brought to the notice of the Council by any person (being a member of the Analy- tical Division of the Chemical Society) who desires to recommend the candidate by letter addressed to The President, The rules are as follows- 4. 5 . 6. 7. 8. Analytical Division, The Chemical Society, Burlington House, London, WlV OBN. The letter should be accompanied by a short statement on the candidate’s career (date of birth, education and experience, degrees and other qualifications, special awards, etc., with dates, and any other relevant information) and a list of titles of, and references to, papers or other works published by the candidate, independently or jointly. One reprint of each paper (or other work) for which reprints are available should be submitted.Thc award will be made on an over-all assessment of the candidate’s contri- bution, the impact of his/her work and his/her future promise in any field covered by the principles, teaching and practice of the analytical sciences. No restriction is placed as to where the work is conducted. The Committee assessing the applications shall be a t liberty to call any candidate for interview. The successful candidate will receive the sum of L l O O in addition to the medal. Thc decision of the Council shall be final. Any alteration to these Rules shall be subject to the approval of the Council. ~~ Recommendations for the next award should be made to The President, Analytical Division, The Chemical Society, Burlington House, London, W1V OBN, by September 30th, 1976.

ISSN:0306-1396    

DOI:10.1039/AD976130291b

出版商:RSC    

年代:1976

数据来源: RSC

摘要:

292 FOURTH SAC CONFERENCE PYOC. Analyt. Div. Chem. SOC. Eu roanalysis I I I August 20-25, 1978, Dublin The Euroanalysis I11 Conference will be held A travel agent has been appointed through between August 20th and 25th, 1978, in the whom special package tours from most Euro- Science block of University College, Dublin. pean capitals will be available. Further in- The Secretariat is being handled by the Insti- formation can be obtained at present from tute for Industrial Research and Standards Mr. Liam 6 hAlmhain, Institute for Industrial (IIRS) and the first circular will shortly be Research and Standards, Ballymun Road, issued and circulated. Dublin 9, Ireland.

ISSN:0306-1396    

DOI:10.1039/AD976130292b

出版商:RSC    

年代:1976

数据来源: RSC