ISSN:0003-2654    

DOI:10.1039/AN97904FX041

出版商:RSC    

年代:1979

数据来源: RSC

摘要:

ANALAO 104 (1 244) 993-1 104 (1 979)ISSN 0003-2654November 1979993101710301037105010551062107010751083108710911094109711001102THE ANALYSTTHE ANALYTICAL JOURNAI- OF THE CHEMICAL SOCIETYCONTENTSREVIEW. Direct Analysis o f Solids by Atomic-absorption Spectrophoto-metry-F. J. LangmyhrDetermination o f Lead, Bismuth, Zinc, Silver and Antimony in Steel andNickel-base Alloys by Atomic-absorption Spectrometry Using DirectAtomisation o f Solid Samples in a Graphite Furnace-Svenerik Backmanand Rune W. KarlssonExtraction o f Nanogram Amounts of Cadmium and Other Metals from AqueousSolution Using H examet h y lenaammon i u m H examet h y lened i t h iocarbamateas the Chelating Agent-A. Dornemann and H. KleistDetermination of Microgram Arnounts o f Precious Metals Using X-rayFluorescence Spectrometry-Paul R.Oumo and Evert NieboerX-ray Fluorescence Determination of Platinum and Palladium i n PlatinumConcentrates Using a Solution Technique-Z. Cruickshank and H. C. MunroDetermination o f Iron(l1) in Rock, !Soil and Clay-L. Th. BegheijnDetermination o f Mercaptoacetic Acid in Hair Waving and Depilatory Products-N. Goetz, P. Gataud and P. BoreREPORTS BY THE ANALYTICAL METHODS COMMITTEEDetermination of Small Amounts o f Nickel in Organic Matter by Atomic-Microbiological Assay of Avoparcirr i n Animal Feeds and Pre-mixesabsorption SpectrometrySHORT PAPERSFast and Simple Polarographic Method for the Determination of Free andTotal. Sulphur Dioxide i n Wines and Other Common Beverages-P.Bruno,M. Caselli, A. Di Fano and A. TrainliDirect Differential-pulse Polarographic Determination of Mixtures o f FoodColouring Matters, Chocolate Brown HT, Tartrazine and Green S-A. G.Fogg and K. S. Yo03-(2'-Thiazolylazo)-2,6-Diaminotoluene as a Selective and Sensitive Reagentfor the Spectrophotometric Determination of Palladium-F. GarciaMontelongo, V. Gonzslez Diaz and C. R. Tallo GonzilezRapid Spectrophotometric Method for the Determination of Arsenic(1ll) inBorate Glasses-Saryoo Prasad Isingh, Ram Pyare, Gur Prasad and P. NathCO M M U N ICATIO N SAn Improvement in the Determination of Available Lysine in Carbohydrate-Ca I i brat i o n of Pro posed Wet - co ml bust i o n Pro ced u re with D ry- co m bust i o nrich Samples-R. J. Hall and K. HendersonMethod for the Determination of Total Carbon in Soils-R. C. DalalBook ReviewsSummaries of Papers in this Issue-Pages iv, vi, vii, x, xii, xivPrinted by Heffers Printers Ltd Cambridge EnglandEntered as Second Class at New York, USA, Post Offic

ISSN:0003-2654    

DOI:10.1039/AN97904BX043

出版商:RSC    

年代:1979

数据来源: RSC

摘要:

iv SUMMARIES OF PAPERS I N THIS ISSUE November, 1979Summaries of Papers in this IssueDirect Analysis of Solids by Atomic -absorption SpectrophotometryA ReviewSummary of ContentsIntroductionPrevious literatureInstrumentsAtomisation methodsAtomisation in flame cellsAtomisation of undispersed powdersAtomisation of powders suspended in solid dispersing agentsAtomisation of powders suspended in liquid dispersing agentsAtomisation of powders sampled by arcing or sparking and suspendedAtomisation in tube or crucible cells, and in cells in the form of a T, invertedAtomisation from rod, strip or braid cellsAtomisation from an electrically heated rod or a closed tube placed in aAtomisation in d.c. or a.c. arcsAtomisation by cathodic sputteringAtomisation by laser or discharge lampsOther atomisation methodsConclusionin gaseous dispersing agentsT or +flameAtomisation of air- and water-suspended solidsReactions in atomisation cellsThermochemistryKineticsInterferencesSamples, sampling, sampling errors anti sample preparationStandards, standardisation and techniques of measurementAccuracy and precisionTime of analysisContamination controlApplicationsAppendixConclusionReferencesKeywovds : Review ; solids analysis ; atomic-absovption spectvophotometryF. J.LANGMYHRDepartment of Chemistry, University of Oslo, Oslo 3, Norway.Analyst, 1979, 104, 993-1016Vi SUMMARIES OF PAPERS I N THIS ISSUEDetermination of Lead, Bismuth, Zinc, Silver and Antimony inSteel and Nickel-base Alloys by Atomic-absorption SpectrometryUsing Direct Atomisation of Solid Samples in a Graphite FurnaceNovember, 1979A fast and simple method of determining lead, bismuth, zinc, silver andantimony in steel and nickel-base alloys has been developed using unmodifiedcommercial atomic-absorption equipment. The method is based on thedirect atomisation of solid metal samples in a graphite furnace.The samplescan weigh between 1 and 20 mg, but test results are influenced by the shapeof the samples. Calibration graphs havebeen drawn using steel samples with known contents. Practical contentranges, e.g., lead 0.03-150 pgg-l and bismuth 0.03-6 pgg-l, and lowdetection limits, down to 0.02 pg 8-1, have been obtained by selecting suitablelines of analysis. The relative standard deviation (1s) is approximately 6%of the content of all elements investigated throughout the stated range ofcontent.The time required for analysis is short, being about 6min for aduplicate determination.The matrix effects are very slight.Keywords ; Direct analysis of solid samples ; lead, bismuth, zinc, silver andantimony detevmination ; steel and nickel-base analysis ; atomic-absorptionspectrometry ; graphite furnaceSVENERIK BACKMAN and RUNE W. KARLSSONSandvik AB, 45-TMK, Fack, S-811 00 Sandviken, Sweden.Analyst, 1979, 104, 1017-1029.Extraction of Nanogram Amounts of Cadmium and Other Metals fromAqueous Solution Using HexamethyleneammoniumHexamethylenedithiocarbamate as the Chelating AgentA group of metals can be extracted from aqueous solution by using hexa-methyleneammonium hexamethylenedi thiocarbamate as the chelating agentand a mixture of 2,4-dimethylpentan-3-one and xylene as the organic phase.A description is given of a procedure fclr the determination of microgram andnanogram amounts of the nine metals silver, bismuth, cadmium, copper, nickel,lead, thallium and zinc in aqueous solution.The influence of a high content of iron and copper on the extraction isdescribed.Keywords A tomic-absorption spectvometry ; metal extraction ; nanogvamamounts of metals ; dithiocarbamate ch,elationA.DORNEMANN and H. KLEISTBayer AG, TVerk Gerdingen, AC-F Untersuchungslaboratorium, D 4150 Krefeld 11,FRG.Analyst, 1979, 104, 1030-1036November, 1979 SUMMARIES OF PAPERS I N THIS ISSUEDetermination of Microgram Amounts of Precious Metals UsingX-ray Fluorescence SpectrometryMicrogram amounts of noble metals were localised with ammonium sulphideon filter absorbent pads and in cellulose pellets for spectrometer counting.The Ka, lines of ruthenium, rhodium and palladium (tungsten tube) and theLcc, lines of osmium, iridium, platinum and gold (molybdenum tube) wereemployed in conjunction with a lithium fluoride (200) analysing crystal.Atthe 95% confidence level, detection limits of 1.0 pg (ruthenium, rhodium andpalladium) and 0.6 p g (osmium, iridium, platinum and gold) were observed for thepellet technique, with values of 0.6 and 0.2 pg, respectively, for the absorbent-pad method.The average coefficient of variation for the determination of10 p g of the seven metals studied was 6.5% for both sample presentations.No inter-elemental matrix interferences were observed among the noblemetals themselves. However, the presence of more than 200 pg of nickel orcopper reduced the slopes of the calibration graph by a constant factor of10% for the lighter metals, and amounts of more than 400 pg of the basemetals reduced the slopes by 20% for the heavier members. Good agree-ment was found between the X-ray fluorescence procedures and standardatomic-absorption methods in analysis of ore concentrates.Keywords : X-ray fluorescence spectrometvy ; precious metals ; absorbent-padtechnique ; cellulose-pellet techniquePAUL R. OUMO and EVERT NIEBOERDepartment of Chemistry, Laurentian University, Sudbury, Ontario, P3E 2C6,Canada.Analyst, 1979, 104, 1037-1049.viiX-ray Fluorescence Determination of Platinum and Palladiumin Platinum Concentrates Using a Solution TechniqueAn X-ray fluorescence solution technique for the determination of platinumand palladium in platinum-bearing material is described.Ruthenium hasto be removed prior to the measurement of the platinum and palladium.Mercury and thorium are used as internal standards. The method is preciseand is more rapid than the gravimetric method normally used.Keywords : X-ray fluorescence spectrometry ; platinum determination ;palladium determination ; mevcury and thorium internal standards ; platinumconcentrates2. CRUICKSHANK and H.C. MUNROJohannesburg Consolidated Investment Company Limited, Minerals ProcessingResearch Laboratory, P.O. Box 13017, Knights, Transvaal 1413, South Africa.Analyst, 1979, 104, 1060-1054.Determination of Iron(I1) in Rock, Soil and ClayA rapid and direct method for the determination of iron(I1) in silicates isdescribed. Redox processes frequently occurring during decomposition aresuppressed satisfactorily by limiting the reaction time to 10 s while main-taining the temperature a t 60-65 "C. Reproducible decomposition tempera-tures are achieved by mixing concentrated sulphuric and hydrofluoric acids(1 + 3 V / V ) in the reaction vessel. The coloured iron (11) - 1,lO-phenanthrolinecomplex is used in the spectrophotometric determination of the twovalency states of iron, iron(I1) directly and iron(II1) by difference aftersubsequent reduction by hydroquinone. Mean results of duplicates of theUSGS geological standards G-2, AGV-1 and BCR-1 are within 0.1% andthose for DTS-1 and PCC-1 within 0.3% of the quoted average valuesfor iron(I1) oxide.Keywords Iron(II) detevmination ; silicates ; hydrojZuoric acid decomposition ;spectrophotometryL. Th. BEGHEIJNDepartment of Soil Science and Geology, Agricultural University T47ageningen,P.O. Box 37, 6700 AA Wageningen, The Netherlands.Analyst, 1979, 104, 1055-1061

ISSN:0003-2654    

DOI:10.1039/AN97904FP097

出版商:RSC    

年代:1979

数据来源: RSC

摘要:

November, 1979 THE ANALYST ixindustrial svstem - m - ~- - ~- ~For 14 years DlGlCO has installed advanced scientific and industrial computers across allfive continents - yes even in the U.S.A.! Installations vary from 'micros' to large'megaminis' all absolutely program and data compatible with a full range of peripheralsand instrument interfaces.Single and multi terminal systemsProcess control 0 Networking0 Medical instrumentationData loggingMathematical simulationGeophysics instrumentationFront end processing0 Multi terminal Basic systems0 Remote terminal emulationIntelligent terminals0 Teaching systemsHigh level or low level languages, interpreters or compilers, fully commissioned ornqked' systems, number crunchers, data loggers, factory controllers or just dataprocessors - test our products against any competition and measure the result forourself.Furthermore they are designed and manufactured right here in Britain - so that you cantalk to the people who actually make the hardware and solve the problems face to face.start small Tick an of the above applications,forwardlit to us and we'll send you an indepth installation dossier free of charge.Digico LimitedWedgwood Way, Stevenage, Herts , England SG1 4PYTelephone Stevenage (0438) 4381 Telex 825508Northern Area OfficeNorth Eastern Chambers, Station Parade, HarrogateTelephone (0423) 501447 Telex 57648 X SUMMARIES OF PAPERS I N THIS ISSUEDetermination of Mercaptoacetic Acid in Hair Waving andDepilatory ProductsNovember, 1979Identification of mercaptoacetic acid in hair waving and depilatory productscontaining other mercapto acids is possible by using thin-layer chromato-However, gas-chromatographic methods are to be preferred for accurate,quantitative determination.Four gas-chromatographic methods are pro-posed ; the first and the second involve derivatisation of the mercaptoaceticacid with diazomethane to the methyl ester, methyl thioether derivative.In the third method, the mercaptoacetic acid is converted into its methyl estergraphy.by use of fi-toluenesulphonic-acid.chromatography of underivatised mercaptoacetic acid.The last method involves the- directKeywords : Mercaptoacetic acid ; hai;v waving products ; depilatory products ;gas - liquid chromatographyN. GOETZ, P.GATAUD and P. BOR:EL'Oreal Research Laboratories, 1 Avenue cle Saint-Germain , 93601 Aulnay-sous-Bois,France.Analyst, 1979, 104, 1062-1069.Determination of Small Amounts of Nickel in Organic Matter byAtomic-absorptnon SpectrometryReport prepared by the Metallic Impurities in Organic MatterSub-committeeThe method for the determination of nickel in organic matter is based onthe atomic-absorption procedure already shown to be applicable to the deter-mination of lead; both nickel and lead can therefore be determined on thesame analytical sample. Organic mattter is destroyed by wet oxidation inthe presence of sulphuric acid. After dilution, the nickel (and lead) arerecovered from the aqueous acid solution by extraction of the complexesformed with ammonium tetramethylenedithiocarbamate complexes into 4-methylpentan-2-one. The nickel content is determined by aspirating theextract into an air - acetylene flame,, measuring the atomic absorption at232.0 nm and comparing the resulting signal with a calibration graph preparedby submitting standard nickel solutions to an identical procedure.It isstressed that practice in sample prepa,ration must be thorough before resultscan be reliable.Keywords : Nickel determination ; wet oxidation ; ammonium tetramethylene-dithiocarbamate ; atomic-absorption spectrometryANALYTICAL METHODS COMMITTEEThe Chemical Society, Burlington House, London, WlV OBN.Analyst, 1979, 104, 1070-1074November, 1979 THE ANALYST xi~~ ~ I TREATISE ON ANALYTICAL CHEMISTRY voi.2 Part 1:__Theory and Practice 2nd Ed.I edited by I.M. Kolthoff, University of Minnesotaand P.J. Elving, University of MichiganA complete and definitive source of information for all analytical chemists, designed tostimulate fundamental research in pure and a plied analytical chemistr . Coverage includesaspects of classical and modern analytical clemistry, and the scientiic and instrumentalfundamentals of analytical methods.0471 05510 7 approx. 770 pages In Press approx. $62.45/f 28.65FOUNDATIONS OF CHEMICAL ANALYSISby 0. Budevsky, Faculty of Pharmacy, Academy of Medicine, Sofia, BulgariaThis concise yet comprehensive introduction makes plain the theory of chemical analysis bymeans of a new and unusual teaching approach.By reinforcing chemical explanations withnumerous raphs, Dr. Budevsky presents a visual overview of equilibria, and particularly intitrimetry, %ereby showing the behaviour of chemical species i n solution, as seen from theevidence of the equilibrium constant as a measure of reaction. The treatment of acid-baseequilibria i s highly original and detailed, and has special significance for acid-base titrimetryin non-aqueous media, and facilitates calculation of titration curves.These new and simple methods in teaching will help the reader’s understanding o f topicspresented. Numerous worked examples in practical analysis are an additional bonus towardscomprehension, and valuable chemical data are tabulated i n Appendixes I-Vli.(EllisHorwood Series in Analytical Chemistry; Editors: Dr. R.A. Chalmers, and Dr. Mary Masson,University of Aberdeen)085312 113 3 372 pages September 1979 $45.25/f18.50Published by Ellis Horwood itd., Chichester, and distributed by lohn Wiley & Sons Ltd.INTRODUCTION TO MODERN LIQUIDCHROMATOGRAPHY 2nd Ed.by L.R. Snyder, Technicon lnstrurnents Corporation;and J.J. Kirkland, €.I. du Pont de Nemours & Co.A completely rewritten edition that incorporates the latest developments in the practicalapplication of liquid chromatograpy. It covers the basics of LC, the six methods and theirapplications, and various specialized areas, and provides material for in-depthcomprehension of how HPLC i s performed, what the necessary materials are, and possibleapplications.0471 03822 9 approx. 848 pages In Press approx.$39.25/f18.00ORGANICS ANALYSIS USING GASCHROMATOGRAPHY/MASS SPECTRQMETRYby W.L. Budde and J.W. Eichelberger, Environmental Monitoring and Support Laboratory,US Environmental Protection Agency, CincinattiA practical guide for scientists, managers, and technicians who perform or contract toperform analyses of organic pollutants in water, air, sediment, or fatty tissue samples. Itemphasizes quality control and the versatility of the powerful gas chromatography/massspectrometry technique.Published by Ann Arbor Science Inc., and distributed by john Wiley & Sons Ltd.0250 40318 8 270 pages October 1979 $22.00/€11~ 3xii SUMMARIES OF PAPERS IN THIS ISSUEMicrobiological Assay of Avoparcin in AnimalFeeds and Pre-mixesNovember, 1979Report prepared by the Antibiotics in Animal Feedingstuff sSub-Committee.A microbiological method for the determination of avoparcin in animal feeds(10-20 mg kg-l; lowest limit of determination 2 mg kg-l) and pre-mixes(20 g kg-l) is described.The method has been submitted to collaborativestudy by a UK Committee and six other EEC Member States, and gave meanrecoveries of 92.8, 104.2 and 101.8%, with standard deviations 17.6, 14.0and 14.6% absolute from chick mash, pig meal and pre-mix, respectively. Inthe method, the sample is extracted with a mixture of acetone, water andhydrochloric acid, and the antibiotic activity of the clarified extract is deter-mined by measuring the diffusion of avoparcin in an agar medium inoculatedwith Bacillus subtilis.Keywords ; Avoparcin assay ; antibiotics ; animal feeds ; microbiological assayANALYTICAL METHODS COMMITTEEThe Chemical Society, Burlington House, London, W1V OBN.Analyst, 1979, 104, 1075-1082.Fast and Simple Polarographic Method for the Determination of Freeand Total Sulphur Dioxide in Wines and Other Common BeveragesShovt PaperKeywords : Direct-pulse polarography ; sulphur dioxide determination ; wineanalysisP.BRUNO, M. CASELLI, A. DI FANO and A. TRAIN1Istituto di Chimica Analitica dell'Universit2, Via Amendola 173, 70126, Bari, Italy.Analyst, 1979, 104, 1083-1087.Direct Differential-pulse Polarographic Determination of Mixturesof Food Colouring Matters, Chocolate Brown HT,Tartrazine and Green SShort PaperKeywords : Food colouring matters ; diffevential-pulse Polarography ; ChocolateBrown H T ; tartrazine; Green SA.G. FOGG and K. S . YO0Chemistry Department, Loughborough IJniversity of Technology, Loughborough,Leicestershire, LEll 3TU.Analyst, 1979, 104, 1087-1090November, 1979 THE ANALYST xiii* Same day despatch on all orders receivedbefore 3 p.m. * 7-day repair service. * Hire service for seminars etc.SEND NOW FOR BROCHUREPRESTONS TIMER DIVISIONCapitol House, Churchgate, Bolton, Lancs BL1 lLYReg in England No 541349 Tel 27035 Telex 63221F l e a s e send me your literature on timers I for industry and sport, without obligation.I Name__p-p- ~ ~ - ~ - -~I PositionI Company--------w--II 1II TA/Hb , ' AddressI- ___-I ~-; PRESTONS TIMER DIVISION ;I POST TO:- PRESTONS TIMER DIVISION, .! Capitol House, Churchgate, Bolton, Lancs.BL11LY. ,'IIIII Address mI II ILimm-mmmmm=m-m=m= Jr i i i i = w m i m = m = i = m =II I enclose f 1 NameTo: May & Baker Ltd (Wail-planner Offer)Essex House, Station Road, Upminster, Essex RM14 2JTPlease send wall-planners (1980/81) at f l eachpostal orderlcheque payable t o May& Baker Ltd I(please print clearly) LA160xiv SUMMARIES OF PAPERS I N THIS ISSUE3- (2'- Thiazolylazo)-2,6-Diaminotoluene as a Selective and SensitiveReagent for the Spectrophotometric Determination of PalladiumNovember, 1979Short PapevKeywovds 3- (2 '-Thiazolylazo) -2,6-diaminotoluene veagent ; palladium deter-mination ; spectvophotometvyF.GARCIA MONTELONGO, V. GONZALEZ DIAZ and C. R. TALL0Department of Analytical Chemistry, University of La Laguna, Tenerife, CanaryIslands.Analyst, 1979, 104, 1091-1094.GONZ~LEZRapid Spectrophotornetric Method for the Determinationof Arsenic(II1) in Borate GlassesShovt PapevKeywovds : Arsenic( 111) detevmination ; spectvophotometvy ; borate glassesSARYOO PRASAD SINGH, R,4M PYARE, GUR PRASAD and P. NATI-IDepa.rtment of Ceramic Engineering, Institute of Technology, Banaras HinduUniversity, Varanasj-22 1005, India.Analyst, 1979, 104, 1094-1097.An Improvement in the Determination of Available Lysinein Carbohydrate-rich SamplesCommunicationKeywords : Lysine detevmination ; cavbohydrate-rich samples ; 2,4,6-tvini-tvobenzenesulphonic acidR. J. HALL and K. HENDERSONMinistry of Agriculture, Fisheries and Food, Agricultural Development and AdvisoryService, Analytical Chemistry Department, Government Buildings, Kenton Bar,Newcastle upon Tyne, NE1 2YA.Analyst, 1979, 104, 1097-1 100.Calibration of Proposed Wet- combustion Procedure withDry- combustion Method for the Determination of TotalCarbon in SoilsComunuTaicationKeywords : Chromic acid digestion; soil ovganic cavbon determination ; dvycombustionR. C. DALALDepartment of Soil Science, Waite Agricultural Research Institute, Glen Osmond,South Australia 5064.Analyst, 1979, 184, 1100-1101

ISSN:0003-2654    

DOI:10.1039/AN97904BP103

出版商:RSC    

年代:1979

数据来源: RSC

摘要:

November 1979 Vol. 104 No. 1244 The Analyst Direct Analysis of Solids by Atomic-absorption Spectrophotometry A Review F. J. Langmyhr Department of Chemistry University of Oslo Oslo 3 Norway Summary of Contents Introduction Previous literature Instruments Atomisation methods Atomisation in flame cells Atomisation of undispersed powders Atomisation of powders suspended in solid dispersing agents Atomisation of powders suspended in liquid dispersing agents Atomisation of powders sampled by arcing or sparking and suspended Atomisation in tube or crucible cells and in cells in the form of a T inverted Tor + Atomisation from rod strip or braid cells Atomisation from an electrically heated rod or a closed tube placed in a flame Atomisation in d.c. or a.c.arcs Atomisation by cathodic sputtering Atomisation by laser or discharge lamps Other atomisation methods Conclusion in gaseous dispersing agents Atomisation of air- and water-suspended solids Reactions in atomisation cells Thermochemistry Kinetics Interferences Samples sampling sampling errors and sample preparation Standards standardisation and techniques of measurement Accuracy and precision Time of analysis Contamination control Applications Appendix Conclusion References Keywords Review ; solids analysis ; atomic-absorption spectrophotometry Introduction Since its introduction in 1955 by the Australian physicist Walsh,l atomic-absorption spectro-photometry (AAS) has developed into one of the most versatile techniques in analytical chemistry.The method permits the simple rapid and inexpensive determination of nearly 70 elements in materials ranging from geology mining and metallurgy to biology pharmacy and medicine. Today AAS is indispensible in those fields where large numbers of samples have to be analysed on a routine basis such as clinical chemistry and geochemical assay. 99 994 LANGMYHR DIRECT ANALYSIS OF SOLIDS Analyst Vol. 104 The high sensitivities and the low limits of determination make AAS particularly useful in trace analysis and pollution control. The accuracy and precision of AAS with flame atomisation are in general sufficiently high to allow the determination not only of the minor and trace constituents but also of the main components of many materials. AAS based on atomisation in flames is now a well established and universally employed technique.However during the past decade much work has been devoted to the problem of replacing the conventional method of flame atomisation by more sensitive and convenient methods. Among the new techniques atomisation in non-flame cells has been found to be particularly promising. Compared with flame atomisation non-flame cells have the advan-tages of requiring smaller amounts of sample of having lower limits of determination and of being better suited for the direct analysis of solid materials. The principle of analysing solids by atomising the components directly from the solid state offers some definite advantages the time-consuming decomposition step can be omitted and the analysis can be carried out without addition of reagents and without any separation and/or concentration steps; the risks of introducing contaminants and of losing the element to be determined are thus considerably reduced.The disadvantages of the direct AAS analysis of solids are that the method is destructive, that normally only one element can be determined at a time that the use of small samples may introduce sampling errors and that interferences may give systematic errors during the measurement of absorption. In this paper a survey is given of the equipment and the procedures that have been employed for the direct AAS analysis of solids; a bibliography on the applications of the technique is also included. The basic theory of AAS is not included; for comprehensive treatises on this subject reference is made to the many advanced textbooks in the field.Previous Literature Some previous survey the whole or parts of the present field. General informa-tion on electrothermal atomisation in AAS can be found in a recent monograph.6 Instruments A selection of instruments and atomisation cells for AAS is commercially available and annual surveys of the equipment are published.' The flame technique is still the most widely used method for atomisation and consequently all instruments have this basic equipment. Most manufacturers also offer other atomisation cells some of which are suitable for the direct analysis of solids. The discussion of the AAS instruments will be confined to pointing out the demands that the analyses of interest here make on the instrumentation.In many such analyses solid or liquid matrix particles and matrix molecules cause considerable background absorption which may introduce serious systematic positive errors. It is therefore imperative that the instrument is equipped with a device that compensates for this interference. Today most AAS instruments are or may be equipped with background correctors the correction for non-specific absorption of radiation being based on consecutive measurement of atomic absorption plus background and of the background alone. It should be noted that the arc-source deuterium lamps normally employed for background correction only cover the wavelength range 180-350 nm. Some manufacturers offer instru-ments equipped with two background correctors the second being an inexpensive tungsten lamp.This lamp compensates for non-specific absorption in the wavelength range 300-800 nm and it is recommended that this lamp be used down to about 300 nm; in this way the lifetime of the much more expensive deuterium lamp may be prolonged. The background correctors are only capable of compensating for a certain amount of non-specific absorption. Specifications from instrument manufacturers indicate maximum correction capacities varying from 0.4 to 1.7 absorbance units. The capacity of background correctors may be measured by inserting sieves of varying mesh size in the light path of the instrument November 1979 BY ATOMIC-ABSORPTION SPECTROPHOTOMETRY 995 In the direct AAS analysis of solids the amplitude and profile of the signal depend on the rates of free atom formation and loss.The rate of free atom formation depends on the temperature and rate of heating on the reactions occurring in the cell and on the composition of the matrix. The factors governing the loss of free atoms are the expulsion and diffusion of atoms out of the absorbing zone and if a purging gas is led through the cell of the flow-rate of the gas stream. The amplitude of the signal is attained when the rates of atom production and atom loss are identical. As the rate of free atom loss is difficult to control in atomisation cells the observed signal will depend mainly on the rate of formation of free atoms. If a number of free atoms are produced rapidly they give a large signal of short duration whereas if the same number is produced slowly the signal will be lower and of longer duration.An important problem is to define whether the analysis should be based upon measuring peak heights or peak areas. Traditionally the peak-height signal has been used and instru-ments of earlier design were not equipped with devices for measuring the areas under the peaks. Today most instruments measure both amplitudes and peak areas ; the amplitudes are either displayed digitally or may be read on an absorbance scale. The measurement of the peak heights of the fast transient signals obtained during the direct analysis of solids requires a fast electronic handling system. This system should be able to record signals of a duration of about 0.1 s. The chart recorders normally employed have a response time of 0.3-0.5 s for full-scale deflection.This may be sufficient for recording small transient signals but is not adequate for registering large signals of short duration. Instead of a laboratory recorder a transient recorder a storage oscilloscope or as in recently introduced instruments a video screen may be used for signal registration. The main objection to the use of peak-height measurements is that even for the more volatile elements and for high rates of heating the rate of free atom formation and conse-quently the profile and the amplitude of the peak may vary considerably with the compo-sition of the matrix. It is therefore strongly recommended to base the direct analysis of solids upon the measurement of peak areas. A measuring system that records both the absorption peak and the area under the peak is very useful.In addition to avoiding the errors of varying rates of atomisation the measurement of peak areas has the advantages of often giving a better precision of giving a calibration graph that is linear over a larger range and of permitting the use of lower atomisation temperatures; the last factor increases the life of the atomisation cell and reduces the risks of sputtering and of producing non-atomic species in such amounts that the compensating capacity of the background corrector is exceeded. The atomisation cell should preferably be connected to a power supply that permits both step and ramp heating (ramp heating allows gradual heating of the cell). A low rate of heating between the drying ashing and atomisation steps reduces the risks of violent reactions and sputtering.It is hoped that future developments in AAS may lead to the introduction of spectro-photometers and atomisation cells designed to meet the special requirements of the solid-sampling technique such as cells of a suitable size faster signal handling systems and improved background correction facilities. The use of the Zeeman effect may solve some of the problems adhering to the equipment employed today. The Zeeman instruments have a single radiation source so that the beam alignment is no longer a problem they have a simpler optical system the background measurements are made closer to the analyte line and they can compensate for a larger non-specific absorption of radiation.To the author’s knowledge only two instruments of this type are commeicially available [the Hitachi (Japan) and the Erdmann & Griin (West Germany) atomic-absorption spectrophotometers] . In addition to the Zeeman effect the Faraday effect i.e. a system utilising beam polarisa-tion has potentiality as a method for correcting the background. Atomisation in Flame Cells Atomisation of undispersed powders powdered solids directly into flames have not been described in the literature. Atomisation Methods To the author’s knowledge AAS methods based on the introduction of undispersed However 996 LANGMYHR DIRECT ANALYSIS OF SOLIDS Analyst Vol. 104 powdered samples have been placed into small platinum or graphite crucibles which were then heated in a flame.8 The low temperatures obtained in these atomisation cells allow the analysis of only the most volatile elements.Atomisation of powders suspended in solid dispersing agents An attempt at the AAS analysis of a powdered sample mixed with a dispersing agent was made9 with a miniature Archimedian screw for introducing the sample into a gas stream feeding a pre-mixed flame. The samples were diluted 100-fold with calcium carbonate of constant particle-size distribution. The equipment was used for determining palladium in cat a1 ysts . Another p r o c e d ~ r e l ~ - ~ ~ is based on mixing the sample with sodium chloride or sodium carbonate depositing the mixture between the threads of a steel screw and atomising the analyte by moving the screw into an air - acetylene flame.The technique was applied to the determination of lithium rubidium caesium and lead in silicate minerals and rocks. The principle and application of a flame cell based on the use of a solid fuel has been described.15-18 The powdered solid sample is mixed with a solid propellant the mixture is pressed into a tablet and one or more tablets are burned just below the radiation beam of the instrument. The method was used for the determination of various metals in their ores. Atomisation of powders suspended in liquid dispersing agents At an early stage of the development of AAS the method found widespread application in the analysis of wear metals in used lubricating and hydraulic oils and greases. The only pre-treatment necessary was to dilute the sample with an organic solvent in order to over-come the viscosity problems during aspiration.In these samples the metals are largely present as suspended particles of metals and/or alloys the efficiency of aspiration and atomisation in flames depend on the particle size. Most of the workers who have made contributions in this field agree that large particles are not completely vaporised and quantitative results may therefore not be obtained. Refer-ences 19-34 pertain to the use of this technique. The analysis of suspensions of wear metals in oils and greases by flame atomisation is now replaced by atomisation in furnaces; by the latter method complete vaporisation of the suspended particles is secured. A device was con~tructed~~ for keeping an aluminium oxide catalyst suspended in water and in various organic solvents; among the dispersing agents tested methanol gave the highest readings for cobalt and molybdenum.The samples were run against closely matched solid standards. Dispersions of metals and alloys can be prepared by operating a spark immersed in ~ a t e r ~ 6 the dispersion being aspirated into a flame. The method has been applied successfully to the determination of minor and trace elements in steel and aluminium alloys. Other workers3' determined tin in suspensions of tin(1V) oxide tin(1V) sulphide and tin(I1) oxide in water without or with agents added to prevent settling. Good results were obtained when the samples and standards were closely matched. Fundamental studies on the efficiency of atomisation of suspended particles have been made.32938 In these investigations metal pa,rticles (of copper iron and silver) and geo-logical samples (sulphide ores zinc concentrate stream sediments and silicate rocks) were suspended in water or in mixtures of organic solvents and the suspensions were aspirated into flames.The results clearly demonstrated the influence of particle size on the atomisation efficiency, which was found to be well below loo% even for particles of diameter 1 ,urn. The determination of trace elements in titanium(1V) oxide pigments has been described.39 It has been shown that the particle size of this pigment is easily reduced by hand-grinding to below 1 pm and the atomisation efficiency can therefore be expected to be relatively high. The only advantages of the methods surveyed in this section are that the sample treat-ment in some methods such as those based upon dispersing the powdered sample in a liquid, is simple and rapid and that many determinations can be made on the same suspension.However a number of disadvantages some of which are serious are inherent in techniques based on atomising undispersed or dispersed solids in flames. The main objection is the incomplete atomisation of the suspended particles. As is apparent from the papers surveye November 1979 BY ATOMIC-ABSORPTION SPECTROPHOTOMETRY 997 above and also other theoretica140 and experimental s t ~ d i e s ~ l - ~ ~ complete vaporisation of particles in a flame is possible only when the particle size is reduced to below 1 pm. As most analysts know the reduction of the particle size of many samples to below 1 pm is virtually impossible.Other disadvantages are the difficulties connected with the preparation of homogeneous mixtures of the sample and a solid dispersing agent the risk that the solid mixture may sputter during heating and that particles settle in liquid dispersing agents. The techniques discussed in this section are not to be recommended in those instances where a high degree of accuracy and precision is required. The methods may however, be useful in such fields as geochemical prospecting. Atomisation of 9owders sampled by arcing OY sparking and suspended in gaseous dispersing agents Solid materials may be transformed into an aerosol of fine particles by sampling with a d.c. arc or a high-voltage spark. A gas (air or argon) fed through the arc or spark chamber, transports the particles into a flame for atomisation and analysis.The method allows convenient sampling of metals and alloys ; non-conducting solid powders may be sampled either from a conventional graphite supporting electrode44 or be mixed with copper powder and pressed into discs of a suitable form.45 By the present technique the mass nebulised is of the order of a few milligrams; these amounts will normally be sufficient to ensure representative sampling. From the data reported in the literature it seems that the accuracy and precision are satisfactory. The method is claimed by some workers46 to be rapid and simple while other workers45 state that the time of analysis is lengthy mainly because of the long pre-spark time necessary in order to reach a steady state.The mechanism of the formation of sample particles is not sufficiently well known. For all techniques based on the introduction of solid particles into flames it is of para-mount importance that the size of the particles is so small that complete vaporisation is secured. The results of some workers46 indicate that this may represent a problem. Unfortunately the effect of varying the parameters upon the particle size distribution has not been studied. Until the results of further work are available no definite conclusions can be d6awn as to the general applicability of the technique to the direct analysis of solids. The matrix and mutual interference effects should also be investigated. Atomisation in Tube or Crucible Cells and in Cells in the Form of a T Inverted T or + The atomisation cells preferably employed today for general non-flame work and for the direct analysis of solids are those in the form of a simple tube or crucible.To some extent more complicated constructions in the form of a T inverted T or + are also used. All of these cells are heated electrically by one or more resistance or induction circuits. Carbon tube furnaces for the study of absorption spectra are d e s ~ r i b e d ~ ~ ~ * ~ in the literature from the early part of this century. Pioneering work on the construction and application of atomisation cells of the present type was started by L’vov49 in 1959; in a series of papers he and his co-workers set a founda-tion for later developments in this field.General surveys6JO on the construction of electrothermal atomisers and special surveys2v3 on those employed in direct analysis of solids have been published. The cells are made of graphite or metal (mostly tantalum or tungsten); the preferred material today is graphite. Graphite can be heated rapidly to above 3300 K its reducing property is of advantage in many atomisations the material is pure or can easily be purified by heating and it is inexpensive and easy to form. The disadvantages inherent in cells made of graphite are that they are combustible that some require a considerable amount of electric power (up to 15 kW) and that they may exhibit memory effects when the element to be determined is incompletely atomised and accumulates in the cell; the latter problem is likely to occur with elements forming stable carbides in these instances higher atomisation temperatures and/or longer atomisation times are required.About 30 papers describe cells belonging to the present group 998 LANGMYHR DIRECT ANALYSIS OF SOLIDS Analyst Vol. 104 The graphite atomisation cells are normally made of the high-purity material used for emission spectrography. To reduce the porosity of the graphite (and consequently the loss of free atoms by diffusion) to minimise carbide formation for certain elements and to prolong the life of the cell the whole or interior surface may be coated with a layer of pyrolytic graphite. As the coating is progressively removed during use the cell has either to be re-coated or better be continuously coated by adding a hydrocarbon e.g.methane or propane to the purging gas.51 Similar beneficial effects are obtained by soaking the graphite cell in certain metal salt solutions; by heating protective coatings of carbides of boron, hafnium molybdenum niobium tantalum titanium tungsten silicon vanadium or zirconium are formed.52 It is also possible to make cells of glassy carbon which is denser and mechanically stronger than the normal or pyrolytically coated graphite. The cells offered by the manufacturers are of the tube or crucible type; to the author's knowledge cells of the T inverted T or + form are not commercially available. The tube cells should preferably have an inner diameter greater than 6 mm; in tubes of smaller diameter it is difficult to introduce solid samples and the sample is also likely to interfere with the beams from the radiation sources.In comparison with the tube cells cells of the crucible type have a shorter absorbing zone and are therefore likely to exhibit a poorer sensitivity. In some types of crucible cells the evaporated species also condense more rapidly than in tube cells. The crucible cells are normally heated by resistance heating the cell being clamped between two graphite rods. In order to obtain reproducible temperatures it is important that the parts forming the contacts are closely matched and that they are in good contact. It is not recommended to remove the crucible for weighing and then re-instal it for atomisation. The crucible cells have the advantage of requiring less power than most tube cells; as the sample is placed in the bottom of the crucible it does not interfere with the light.The cells of the T,53 inverted T54 and + 5 5 9 5 6 forms are of a more complicated construction; they are bigger than the tube or crucible cells and require more power and separate heating circuits may be necessary for the various parts. The maximum temperature obtained in most of these cells is 2800 K. Some of the cells of the + form55956 allow rapid and successive analyses of various materials. The long absorption tube of these cells gives a high sensitivity, and the well below the absorption tube keeps the samples away from the beams of the radiation sources. Among the various atomisation cells described in the literature the dual chamber cell should be mentioned5'; this cell has a closed cylindrical graphite chamber around a con-ventional graphite tube; its construction is similar to a type described in a previous paper.58 The sample is placed in the outer chamber and on heating the vapours diffuse at different rates into the inner tube where the absorption is measured.The cell can accommodate larger samples than most tube or crucible cells and may be useful for separating analytes from matrices producing large amounts of smoke. However the low temperature of the cell allows only elements of high volatility to be determined. Other devices for separating analytes from non-absorbing species have been suggested. One of the recent constructions is a modification of a T-type of furnace,59 in which the sample crucible is closed with a porous graphite cover that filters out and/or decomposes the smoke particles.However analytes forming stable carbides may be trapped in the cover. The cells constructed for the separation of analytes from non-absorbing species have a complicated design and can only be used for the more volatile elements. The versatility of the atomisation cells discussed in this section is demonstrated by their applications; inorganic and organic materials in the form of solid powders drillings cuttings of sheet or fibres samples of soft and hard tissue etc. have been analysed successfully. As is apparent from the above discussion of the analysis of wear metals in used lubricating and hydraulic oils and greases, the atomisation in flames of metals in suspension cannot be considered as quantitative; however by atomising these samples in the present type of cells accurate results can be expected for a number of metals normally encountered in used lubricating and hydraulic oils and grea~es.60-6~ The samples may also be added to the present types of cells as slurries in water to which either sodium hexametapho~phate~~ or a thixotropic thickening agent is added.66 The samples may also be added as suspensions in liquids November 1979 BY ATOMIC-ABSORPTION SPECTROPHOTOMETRY 999 Atomisation from Rod Strip or Braid Cells Horizontal graphite rods with a hole in or a slit along the upper surface for the sample, strips made of metal (tantalum tungsten molybdenum or platinum) and graphite braid have been described and some are also commercially available.In comparison with the graphite tube or crucible cells the present group of atomisation devices have certain attractive advantages; they are small and of a simple construction, thus requiring less power they are heated and cooled more rapidly and those made of metal do not form carbides. The main disadvantage of using rod strip or braid cells is that no additional energy is available in the space above the cell to prevent relatively rapid condensation of the evapora-ted species and consequently the effective lifetime of the absorbing atoms is considerably shorter than that when tube cells are used. The braid cells do not seem to have been applied to the direct analysis of solids; the construction makes this type less adaptable to the analysis of powders.Metal cells become brittle as a result of reactions with the samples and with air the metals employed are rather expensive and some such as tungsten are often not sufficiently ductile to permit the production of more complicated forms. The tendency during recent years has been to replace the present types of cells with tube cells made of graphite. Some of the present types of cells have a small sample capacity. Atomisation from an Electrically Heated Rod or a Closed Tube Placed in a Flame Various w0rkers~~-70 have described atomisation cells that consist of an electrically heated rod or a closed tube placed horizontally in a flame; the pulverised solid samples are placed in either an open slit or a crater in the rod or in an axial hole in the rod the hole being closed with graphite powder or a graphite washer.The flame transports the evaporated species through the radiation beam and prevents rapid condensation. From the data published on the use of the closed graphite tube about 30 elements can be determined including such low-volatility elements as silicon titanium molybdenum and vanadium. Inorganic materials such as minerals rocks ores ceramics slags oxides and salts have been analysed successfully. A special type of atomisation cell consists of two graphite tubes one fitted coaxially into the other58; the cell is heated electrically and during heating it is enveloped in a reducing flame. The solid samples are placed in the cavity between the tubes and on heating the vapours diffuse through the porous walls of the inner tube; to prevent diffusion through the outer tube and the end parts these are either made of glassy carbon or covered with pyrolytic graphite.The different rates of diffusion make it possible to make measurements that are otherwise influenced by a large non-specific absorption of radiation. The cell can be heated only to 2900 K. Unfortunately there are few data to demonstrate the feasibility of this cell. Some of the constructions belonging to the present group have the advantage of having a large sample capacity (from tens to hundreds of milligrams) and the risk of introducing sampling errors is thus reduced. t o offer any definite advantages. The disadvantages are that facilities for both electric and gas heating are required.In comparison with the simple tube cells the present types of equipment cannot be seen The equipment surveyed in this section is not commercially available. Atomisation in D.c. or A.c. Arcs It is not unexpected that the well established excitation methods of emission spectro-graphy also have been employed for atomisation purposes. Various ~orkers~l-73 have reported the direct analysis of solids by atomisation in d.c. or a.c. arcs the atomic absorption being measured in the gap between the electrodes. However as the temperatures in the arcs are considerably higher than in the flames and furnaces normally employed in AAS the ratio of neutral atoms to excited and ionised atoms is less favourable, in particular for the alkaline and alkaline earth elements. The d.c.and a.c. arcs are powerful devices for vaporising solid materials 1000 LANGMYHR DIRECT ANALYSIS OF SOLIDS Analyst Vol. 1U4 The higher sensitivity obtained for a number of heavy metals when measured by the absorption instead of the conventional emission method may be utilised in laboratories equipped with these instruments. The above instruments for the direct analysis of solids by atomisation in arcs are not equipped with lamps for background correction and non-specific absorption has to be compensated for by repeating the measurements at the wavelength of a non-absorbing line. In comparison with atomisation of solids in graphite tube cells the present technique cannot be seen to offer any advantages. From a survey of the literature on the subject it appears that there is little interest in the method.Atomisation by Cathodic Sputtering In various paper^,^^-^^ in particular by Australian workers cathodic sputtering has been shown to be an efficient technique for producing atomic vapour from metals and alloys. The method is also applicable to the analysis of non-conductive materials such as rocks and ores but it is then necessary to mix the sample with metal powder e.g. copper in order to make it conductive. The particles released from a previously cleaned cathode are usually ground-state neutral atoms that leave the cathode surface with high energy; thermal equilibrium of the sputtered atoms with the filler gas is rapidly established and in the vicinity of the cathode the sputtering and the diffusion rates are such that the concentration of atoms reaches a steady-state value.The equipment and procedures for carrying out analyses by cathodic sputtering are relatively simple ; when the surface has been properly processed the amounts sputtered are sufficiently high to prevent any serious sampling errors. The time required for processing the surface of the material to be sputtered may vary considerably, the cleanliness of the cathode surface and the purity of the filler gas are critical and the complex mechanism by which particles are released from the cathode is not known. These problems and the conflicting results reported in the literature have made analysts and manufacturers sceptical about the future of cathodic sputtering. The small number of papers that have been published on this subject in recent years reflects the present lack of interest in the technique.Further work is obviously required to ascertain its potentiality as a method for analysis by absorption. However it should be mentioned that the method is being increasingly used for solid sample analysis by emission. Atomisation by Laser or Discharge Lamps When a pulse of laser light is focused on a solid sample a cloud of vaporised material is formed that contains neutral and excited atoms ions and other species. The laser-produced plasma has a bright emission; however the free atoms outlast the emission period thus allowing AAS measurements to be made. Direct analysis of solids by laser atomisation has been described by a number of worker^.^^-^^ In the method in one of these papers,82 the metal sputtered from the sample is collected on the inside of a small graphite cylinder which is subsequently heated for atomisation and analysis.The samples analysed have mainly been solid metals or alloys and powders of inorganic materials pressed into pellets; however one papers7 also describes the analysis of various biological materials such as plants liver blood and muscle. The workers in this field have employed widely different equipment instrumental arrangements and measuring techniques; while some workers have obtained satisfactory analytical results, others have obtained only semi-quantitative data. The advantages of the technique are that it is rapid it is applicable to solid materials in almost any form it allows the determination of a large number of elements and it requires a minimum of sample pre-treatment.The disadvantages are that the technique cannot match the detection limits achieved by conventional atomisation in tube or crucible cells it requires careful standardisation for each matrix system the small amounts of sample vapourised may cause sampling errors and the high energy of the laser pulse causes eruption of liquid or solid particles from the samples. In addition to the use of lasers discharge lamps have also been employed for atomisation There are however a number of problems inherent in the present technique. Conclusions on the potentiality of the present technique are difficult to draw November 1979 BY ATOMIC-ABSORPTION SPECTROPHOTOMETRY 1001 of solids. However as the sensitivity of analysis with sources of continuous spectra is normally one or two orders of magnitude worse than that obtained with the line sources normally used the discharge lamps are at present of less importance as the primary light source.A special applications1 of a discharge lamp as the atomisation cell for solid samples should be mentioned. The external surface of a lamp was coated with a thin film of an oxide a flash of the lamp evaporated the film and the atomic lines of the spectrum were used to determine the content of impurities. Other Atomisation Methods The plasma formed by an electric discharge in a dielectric channel has a high temperature and pressure and may be used both as an atomisation cell for solid samples and as a source of a continuous spectrum.It has been demonstrateds2 that elements that are difficult to atomise such as silicon aluminium and titanium are easily evaporated from quartz glass. As the work carried out on the technique and its applications is limited it is difficult to evaluate its potentiality. It seems that the rapid evaporation of the sample and the turbulent ejection of the vapours may cause sputtering of solid or melted particles. Conclusion Today electrothermal atomisation is the preferred method for solid sample analysis and the discussion in the following sections will therefore be limited to the use of this type of atomisation cell. Atomisation of Air- and Water-suspended Solids Within the field of environmental analytical chemistry the analysis of suspended particles in the atmosphere and in water constitutes an important branch.The direct analysis of leadv3 and cadmiums4 in air particulates is possible by introducing air directly into a specially designed atomisation cell. However most analyses of air particulates are made by separating the solid particles by filtration either through a filter made of cellulose glass or a plastic material,v5-104 or through a porous graphite cup or ~ y l i n d e r l ~ ~ - ~ ~ * ; in the former method the determinations are made by transferring the whole or part of the filter into a suitable atomisation cell whereas in the latter technique the graphite filter also serves as the atomisation cell. A patentlm describes equipment that collects electrostatically solid particles in gases in a graphite tube and then transfers the tube into a heating circuit for atomisation.Equipment has also been developed for sampling air particulates in the field. Particulate matter in water may be analysed directly after f i l t r a t i ~ n . ~ ~ l l ~ The direct analysis of solids suspended in air and water can be made rapidly and requires only small amounts of sample. Reactions in Atomisation Cells A knowledge of the reactions that occur in atomisation cells is highly desirable in order to choose the most favourable conditions and to control and minimise the effects of interferences. Theoretical and experimental studies are complicated by the non-equilibrium conditions that often prevail during the reactions the high temperatures and the correspondingly high rates of reaction and the lack of reliable thermochemical data.The ensuing discussion will be limited to a short survey of the reactions that are likely to occur in atomisation cells made of carbon or metal. More comprehensive treatments have been published.6*111J12 Atomisation cells are normally equipped with power supplies that are controlled auto-matically e.g. by a microprocessor and allow the sample to be heated in at least four stages, viz. the drying ashing (or pyrolysis) atomisation and cleaning steps. In addition the analyst can select the time of heating the interval and rate of heating between the steps and the flow-rate of the purging gas. During the drying stage at about 370 K the main reaction will be the removal of non-essential water. However some highly volatile elements and metal-containing compounds, such as mercury and the organomercurials may be lost during this operation 1002 LANGMYHR DIRECT ANALYSIS OF SOLIDS Analyst Vol.104 The ashing stage involves one or more heating steps in the range from 620 K to above 1800 K. The main purposes of this operation are to destroy organic matter to remove essential water and interfering elements and compounds by evaporation or sublimation to transform carbonates sulphates nitrates hydroxides etc. into the corresponding oxides, or to transform the analytes into compounds that are less volatile and more convenient for the atomisation processes. In the analysis of certain inorganic materials such as metals and alloys the ashing step may be omitted. The volatility of the analyte or its compounds sets the upper limit of the ashing step.When volatile metals such as cadmium zinc and bismuth are to be determined the ashing temperature should not exceed 620 K. During low-temperature ashing arsenic selenium and tellurium may be completely or partly volatilised either in elemental form or as volatile compounds. Losses of these elements during ashing can be avoided by adding a metal solution e.g. of copper nickel, silver or molybdenum to the sample the metals forming thermally stable compounds with the ana1~tes.l~~ Certain metals such as molybdenum and vanadium allow ashing temperatures above 1800 K. The ashing temperatures recommended in the literature vary considerably with the type of cell employed and the material being analysed and it may therefore be necessary to establish the upper limit of the ashing temperature.During the atomisation step the cell is heated to temperatures in the range 1100-3300 K, the analyte is vaporised and the absorption of the free atoms is measured. The amplitude and shape of the peak are determined by the rate of the atomisation process. If the analysis is based on the measurement of peak heights the rate of free atom formation should be high for both sample and standard; this is normally obtained by maintaining a high rate of heating. However when peak areas are measured instead of peak heights the kinetics of the reactions lose their significance and assuming that the analyte is fully atomised and that all free atoms spend the same time in the light beam the shape of the signal may vary without influencing the analytical results.As recommended above direct AAS analysis of solids should preferably be based on the measurement of peak areas. The purpose of the fourth and final heating step is to clean the cell and is carried out by heating the cell for a few seconds at the maximum temperature. Any residue remaining in the cell should be removed by blowing or brushing; as the latter operation usually affects the interior surface of graphite cells it should be followed by a cleaning step. Thermochemistry Despite the possibility that equilibrium may not be attained in atomisation cells thermo-chemical calculations may still be of value for explaining the reactions in graphite and metal cells. These calculations are complicated by the large number of processes taking place in the solid and gaseous state and the complexity of the transport mechanisms.On the other hand the high temperatures limit the number of compounds in the gaseous phase to the simple thermally stable diatomic molecules. The tendency of some elements such as selenium tellurium bismuth and potassium to form polyatomic species has also to be taken into account. Various ~ ~ r k e r ~ ~ ~ ~ ~ ~ ~ ~ have related the experimentally established appearance tempera-tures of elements atomised in graphite cells with the temperatures at which the free energies of the reduction of the oxides with carbon to free gaseous metals were zero. The agreement between the temperatures was satisfactory for some elements but it was poor for other systems.In those instances where the appearance temperatures were higher than the temperatures at which the free energies were zero the discrepancies can be explained by assuming that at the latter temperatures the rates of reaction or the vapour pressure of the analytes were too low. L’vovl12 discussed the formation of thermally stable carbides pointing out that these reactions may play a more important role than had previously been assumed. On the basis of the heat of carbide formation and its relationship with the appearance temperature the experimentally determined and estimated appearance temperatures for 35 elements were compared. For many analytes the agreement was satisfactory; however for a number of systems in particular the alkali and alkaline earth elements there were considerabl November 1979 BY ATOMIC-ABSORPTION SPECTROPHOTOMETRY 1003 discrepancies.A possible reason for these differences is the formation of intercalation compounds i.e. the tendency of some elements in particular those which are easily ionised, to enter the interlamellar space in graphite and to form compounds with carbon. The complete decomposition of intercalation compounds requires prolonged heating at temperatures above 2300 K ; these compounds are less likely to be formed in cells made of glassy carbon or in graphite cells covered with a metal carbide or with pyrolytic graphite. Calculations of the free-energy changes occurring in the formation of stable carbide com-pounds show that metal oxides frequently react to form stable carbides a t temperatures below that at which the reduction of the metal oxide by carbon to gaseous metal atom occurs.Some metal oxides e.g. antimony(II1) oxide have relatively high vapour pressures at the appearance temperatures of the free gaseous metal atoms and some oxides may there-fore be lost by evaporation. In calculations of the thermal dissociation of oxides in a graphite environment the equilibria between carbon and oxygen also have to be taken into account. In atomisation cells made of metal of which the tantalum strip cell is a well known type, the most likely path to the formation of gaseous atoms is the thermal dissociation of metal oxides or other compounds. The reduction of a number of metal oxides with tantalum is thermodynamically feasible ; however in comparison with the dissociations the rates of reaction for the reduction of oxides in contact with the dense metal are probably low.When nitrogen is used as a purging gas for graphite atomisation cells CN absorption bands may be observed; at temperatures above 2300K the absorption of these bands increases exponentially with increasing temperature. In graphite cells purged with nitrogen some elements react with CN molecules to form cyanides,l12 and the formation of these compounds reduces the sensitivity of the determina-tions. By replacing nitrogen with argon the difficulties in the use of the former gas are avoided. Today argon is the preferred purging gas. From the above survey it appears that thennochemical calculations may explain many of the reactions that occur in carbon or metal atomisation cells; however for a number of systems the processes are not yet fully understood and further work is required.Kinetics In a studyll6 of the behaviour of copper in the range 1720-2220 K the atomisation processes were explained by assuming a slow first-order reaction involving reduction of copper oxide by carbon followed by rapid volatilisation of copper. In a series of papers117-11s other workers described the supply of the analyte by an Arrhenius-type rate constant; they relate the removal of the analyte to convection and diffusion. The studies were made with the use of a home-made open graphite rod cell and the data are therefore not directly applicable to the commercial graphite tube cells. The removal of analyte vapour from a graphite tube furnace has been found120 to occur by a loss (about 20%) through the aperture for the sample injection by diffusion through the tube walls (about 20%) and by diffusion to the cooler parts of the cell where condensation takes place.In a recent paper,121 previous contributions in the field are critically surveyed and a more advanced theory is introduced. The paper describes a study of the supply and removal of analyte vapours in graphite crucible and tube atomisation cells. In both cells the release of the atoms from the graphite surface was found to be determined by the temperature of the surface and was explained by an Arrhenius-type rate constant. The removal of the atoms from the cells varied with the type of cell employed. In the crucible cell diffusion was found to be the dominating process while in the larger tube cell operated under gas flow conditions convection is the dominating process.When the latter cell was operated under gas stop diffusion and to some extent expansion are the main factors in the removal process. From theoretical considerations and experiments it was demonstrated that under gas stop conditions only 25% of the sample is contained in the tube cell; under flow con-ditions this amount is reduced to less than 10%. It should therefore be possible to increase considerably the sensitivity of these determinations. The cited paper121 gives in a con-densed form an excellent review of the kinetics of the reactions in graphite atomisation cells. Papers on the kinetics of atomisation processes are limited.The atomisations were made in a graphite tube furnace 1004 LANGMYHR DIRECT ANALYSIS OF SOLIDS Analyst Vol. 104 Interferences The present technique is subject to various interferences that may introduce serious negative or positive systematic errors; a survey of commonly occurring interferences is given below. The main source of interferences is the matrix. During atomisation the matrix may produce atom molecule and radical interfering species and solid or liquid particles formed by sputtering from the sample by destruction products of inorganic or organic matter or by condensation of vaporised atoms or compounds in the cooler part of the atomisation cell. If the matrix produces atoms that absorb close to the absorption line of the analyte direct spectral interference occurs.A small number of such interferences have been reported122; however in most instances this interference can be avoided by making the measurements at another wavelength. Another type of spectral interference is encountered when the matrix gives off atoms with absorption lines that are not close enough to cause direct spectral interference but sufficiently close to be included in the spectral range of the slit (approximately 0.5 nm); these lines will absorb radiation from the background corrector lamp but not from the analyte light source, and consequently a negative correction (error) will be applied to the absorption signal of the analyte. The latter situation was recently documented123 in the determination of selenium in the presence of large amounts of iron.Iron has a number of resonance wave-lengths close to the most sensitive selenium line at 196.0 nm and these iron lines absorb emission from the deuterium background corrector but not from the selenium emission source. Again the interference can be avoided by selecting another selenium line. It was recently demonstrated12* that the interference of iron on the 196.0-nm selenium line can be avoided by carrying out the measurement with an instrument based on the Zeeman principle for background correction. The matrix may also produce molecules or radicals that absorb radiation; absorption spectra of molecules of commonly appearing salts oxides and hydroxides have been recorded by various workers.126 These molecules exhibit one or more broad maxima in the spectral range used for AAS measurements; in addition to the spectral continua they may also contain line-rich electronic spectra.While the background corrector compensates for the former interference the corrector does not correct for the fine structure of the molecular background. The broad-band spectra of the molecules of salts etc. will normally be compensated for by the background corrector system. However when large amounts of absorbing species are produced rapidly the consecutive measurement of the atomic absorption and of the background is not operating sufficiently rapidly and a negative signal is observed. This disturbance can be avoided by introducing a smaller sample and employing a lower rate of he at ing . The matrix will of course also affect the rate of the atomisation process and thus the amplitude and the shape of the peak.The rate of reaction may seriously affect analyses based on the measurement of peak heights whereas it does not affect data obtained from peak-area measurements. The scattering of radiant energy by very small particles (with a diameter less than one tenth of the wavelength of the radiation measured) can be expected to occur according to the Rayleigh theory Le. the scattering (s) is proportional to the fourth power of the inverse of the wavelength (A) (s = l/X4) and the effect of scattering is therefore much greater in the ultraviolet than in the visible range. Scattering by large particles can be expected to occur according to the Mie theory and in these instances only a slight wavelength dependency would be expected.Within the range of slit widths of most AAS instruments the absorption by scattering can be considered to be constant. Non-specific losses of radiation by scattering are normally compensated for by employing a background corrector. However the rapid formation of large amounts of scattering particles may exceed the compensating capacity of the corrector. In the analysis of biological and other materials the analyst is often faced with the problems associated with the presence of large amounts of halides in particular sodium chloride. The halides affect the measurements in various ways; in addition to the effect on the rate of atomisation and the non-specific absorption of molecules or particles the formation o November 1979 BY ATOMIC-ABSORPTION SPECTROPHOTOMETRY 1005 volatile gaseous monohalides of the analytes represents a serious source of interference.The problems associated with the formation of gaseous monohalides were discussed in a recent review.l12 The interference from chloride (and presumably other halides) can be removed or be compensated for by the following means (a) addition of sulphuric orthophosphoric nitric or ascorbic acid and removal of chloride by evaporation; (b) addition of ammonium nitrate solution and removal of chloride as the ammonium saltll3; (c) removal of sodium chloride by evaporation at about 1500 K ; ( d ) binding chloride in a compound that is more stable than the monochloride of the analyte112; (e) carrying out the measurements according to the method of standard additions; and (f) adding hydrogen to the purging gas and removing chloride as hydrogen chloride gas.The last method has been used successfully in the analysis of lead in steels,126 the separation being effected at 900-1OOO K. Samples Sampling Sampling Errors and Sample Preparation As is apparent from the bibliography on the use of the present technique (see below) a variety of solid materials has been analysed by the techniques surveyed under Atomisation Methods. In addition to the samples originally present as solids materials in the liquid state e.g. biological fluids can be analysed by transforming them into solids by drying, dry ashing plasma ashing or lyophilisation. These operations serve the purpose of con-centrating analytes present in amounts near or below the lower limit of determination.After being transformed into the solid state the material should be homogenised by grinding. It should be noted that certain elements and compounds such as mercury may be lost during these operations. Losses of volatile elements during the above operations can be avoided by adding a stabilising agent.l13 In most methods for the direct AAS analysis of solids the amounts of sample range from less than 1 mg and up to about 20 mg; some atomisation cells allow amounts of sample up to several hundred milligrams to be atomised. Whereas the determination of minor or trace elements in a sample weighing say 500 mg does not normally introduce any serious sampling errors the determination of the trace elements in samples with a mass of the order of 1 mg raises problems relating to the sampling errors and the risks of introducing contaminants.All sampling operations should of course, be made with the utmost care in order to obtain a representative and uncontaminated sample and to reduce the sampling errors to a minimum. Brittle inorganic substances such as minerals rocks ores salts and certain elements and alloys should not be crushed or pulverised in conventional equipment made of alloyed steel, but should be crushed or ground in agate corundum or carbide mortars and pestles. The particles of these samples normally can be readily ground manually to pass a 270-mesh (63-pm) sieve; further grinding to pass a 400-mesh (37-pm) sieve may take a considerable time.However with the use of automatically operated grinding equipment the particle size of inorganic brittle materials can be reduced to a considerably finer state of subdivision. In the author's laboratory,127 about 500 mg of the granite GH (a reference sample issued by Centre de Recherches Pdtrographiques et Geochimiques France) was pulverised for 3 h in an automatic agate mortar and pestle. The particle-size distribution (established with a Coulter Counter) is shown in Table I. From Table I it appears that about 86% of the particles were smaller than 8 pm. About 1% of the particles had a size in the range 3240 pm and this fraction may consist of biotite, the particle size of which is difficult to reduce. However the whole sample passed a 400-mesh sieve i.e.the maximum particle size was 37 pm. It should be noted that prolonged grinding may result in losses of volatile elements e.g., mercury from a silicate or sulphide matrix. Samples that have been in contact with metal sieves should not be used in the subsequent analysis and the portions of the sample taken for the sieve tests should be discarded. If the sample has to be sieved before analysis nylon-meshed plastic sieves should be employed. Brittle inorganic materials are usually analysed in the form of powders. The errors introduced by the sampling of powders for trace analysis vary considerably with the distri-bution pattern of the analyte. The trace element may either be evenly distributed in one or more of the main components of the matrix or be present as discrete particles disseminate 1006 LANGMYHR DIRECT ANALYSIS OF SOLIDS Analyst Vol.104 TABLE I PARTICLE SIZE DISTRIBUTION OF THE GRANITE GH Particle size/pm Content % <1.6 2.6 1.6-2.0 4.7 2.0-2.6 9.0 2.6-3.0 12.6 3.0-4.0 16.4 4.0-6.0 16.2 6.0-6.0 14.8 6.0-8.0 10.9 8.0-9.0 7.6 9.0-12.7 3.6 12.7-1 6.0 1.3 16.0-20.0 1.8 20.0-24.0 0.3 24.0-32.0 0.0 32.0-40.3 1.1 40.3-60.8 0.0 Sum 101.6 throughout the matrix; as would be expected the former instance gives a smaller sampling error. To demonstrate the order of magnitude of the sampling error under the conditions of the less favourable situation a method described in the literature1% was used to estimate the error for a distribution of mercury(I1) sulphide in a matrix of iron(I1) sulphide.l29 In Table I1 the relative standard deviation of the content of mercury is given for 5-mg samples con-taining 1 3 10 and 30 p.p.m.of mercury and for various particle sizes. As is apparent from Table 11 the sampling error may be so large under combinations of unfavourable conditions that reliable analytical data are unobtainable. TABLE I1 SAMPLING ERROR FOR A DISTRIBUTION OF MERCURY(II) SULPHIDE IN A MATRIX OF IRON(II) SULPHIDE Particle ASTM diameter/ mesh CCm number 108 140 63 270 37 400 20 10 --Approximate number of particles per 6 mg of sample 1.6 x lo9 2.4 x lo6 1.9 x 106 1.3 x 104 3.8 x 104 Relative standard deviation of the Hg content % for samples containing : 1 p.p.m. 3 p.p,m. 10p.p.m. 30p.p.m. I A I 232 134 73 43 81 47 26 16 48 27 16 9 19 11 6 3.6 6.7 3.9 2.1 1.2 However when the samples are ground to a very fine state of subdivision the sampling error under the conditions of an unfavourable distribution pattern may be reduced to such low levels that the sample size can be reduced to below 1 mg.This is exemplified by the direct AAS determination of copper in granite GH.12' The particle size distribution of this sample after grinding is shown in Table I. It was assumed that copper was present in the granite as discrete particles of chalcopyrite (CuFeS,) that all particles were present as spheres having an average diameter of 5 p m and that the specific gravity of the sample was 2.7. The number of particles per milligram was calculated to be 5.7 x lo6 and a 0.2-mg sample therefore contains about 106 particles.Using the above method for estimating the sampling error and a sample size of 0.2 mg the relative standard deviation of the content of copper was calculated to be 20%. This is a large error; however as the number of particles is probably considerably higher than lo6 the error obtained in actual analysis may prove to be smaller. The content of copper in granite GH was then determined by direct AAS; from 1 + 1 mixtures of granite and graphite powder six portions corresponding to amounts of granite ranging from 0.16 to 0.34 mg were transferred into a graphite tube furnace and copper wa November 1979 BY ATOMIC-ABSORPTION SPECTROPHOTOMETRY 1007 determined by measurements against the National Bureau of Standards (NBS) Standard Reference Material No.614 (Trace Elements in Glass). The average content of copper was found to be 13 p.p.m. with a relative standard deviation of 16% ; the latter is substantially lower than the above estimated value. The recommended value for copper in granite GH is 14 p.p.m. It can be concluded that the error associated with the sampling of finely powdered inorganic materials is acceptable for amounts of sample down to fractions of 1 mg. Ductile metals and alloys are normally sampled by drilling. The sampling error may again be serious when the element to be determined is present as or in irregularly distributed particles. However as is apparent from the data from applications (see below) direct AAS analysis has been used successfully in the determination of a number of trace constituents in irons and steels; in these analyses the amounts of sample ranged from 1 to 12 mg.Recent studieslS0 of the distribution of lead in mild steel stainless steel a nickel-base alloy and ferromolybdenum and of antimony in mild steel demonstrated that the two trace metals are evenly distributed and that amounts of sample of 2 mg can be taken without introducing any serious sampling errors. In the sampling and treatment of solid materials of human animal and plant origin the diversity of the samples and the widely differing properties of the analytes make it difficult to give any general procedure. The amount of material available is often small and the elements to be determined may be inhomogeneously distributed and consequently sampling errors may occur.Many samples of biological origin can be homogenised by grinding in the wet state; the grinding equipment should be made of materials that do not contaminate the sample. Samples of suspensions and tissues can be homogenised by solubilisation in quaternary ammonium bases; it should be noted that these reagents may contain appreciable amounts of contaminants. Standards Standardisation and Techniques of Measurement The direct AAS analysis of solids is normally based on measurements against suitable standards; in principle it is possible to carry out “absolute” analyses but this possibility does not seem to have been applied to the AAS analysis of solids. The following standardisation methods have been employed measurements against solid standard samples of natural materials or industrial products; measurements against synthetic solid standards; measurements according to the standard additions technique ; measurements against standard solutions.When reliable solid standards are available with certified values for the elements to be determined and with matrices corresponding to those of the samples measurements against such materials are to be preferred. The selection of such standard samples in particular for the analysis of trace components is limited with respect to both materials and elements; however the selection of such materials is steadily increasing. Lists of suppliers of reference and standard materials have been publi~hed.~$l~~ In the certificates of analysis such as those issued by the NBS the certified values are based on a sample size of at least 260 mg.When these samples are employed as standards in the direct AAS analysis of solids the amount taken for atomisation will often be of the order of 1 mg and the sampling error has to be taken into consideration. However as is apparent from the precision obtained from the use of various NBS standard reference materials the sampling error for a number of analytes does not seem to be serious for amounts of sample of the order of 1 mg. When no suitable solid standard is available it is possible to use synthetic solid standards. Such standards are produced commercially with a graphite or silica matrix; they can of course also be prepared from pure substances and with a matrix resembling that of the sample.However the preparation of these samples is tedious and time consuming. It is also possible to apply the standard-additions technique by adding standard solutions to a series of accurately weighed solid samples of approximately the same mass. In these analyses it is recommended to add constant volumes of the standard solutions to let the standard solution soak into the sample and to establish the position of the standard-additions graph by the method of least squares 1008 LANGMYHR DIRECT ANALYSIS OF SOLIDS Analyst VoE. 104 In a few instances it has been found possible to analyse solids by measurements against calibration graphs established by atomising aqueous standard solutions. The standard solutions should be prepared from high-purity metals or compounds (preferably dissolved in nitric or sulphuric acid).In these instances the sample and the standards have a different matrix composition the atomisation mechanisms will be different and the absorption signals are also likely to differ. The measurements should again be based on the addition of equal volumes of the standard solutions and on the measurement of peak areas instead of peak heights. The use of some of the above standardisation techniques is illustrated by the data listed in Table 111 obtained from electrothermal atomisation and AAS determination of cadmium, manganese and lead in hydroxyapatite and two samples of animal bone.182 TABLE I11 RESULTS FOR CADMIUM MANGANESE AND LEAD IN HYDROXYAPATITE AND TWO SAMPLES OF ANIMAL BONE Sample Method* Hydroxyapatitet .A B Animal bone IAEA . . Values recommended by IAEA C A D Animal bone Weider C A D %t p.p.m. 0.33 0.34 Cd 1 Srr 18 18 % Mn - 2 S r l P.P*m* % 0.94 3 0.84 6 Pb - 0 Srt P.P.m. % 2.6 12 2.5 12 Not given 0.072 - ---0.036 -12 --22 32 16§ 33 12 32 11 6.4 20 0.0 22 - -- -6.8 16 6.9 15 0.8 18 9.8 14 9.7 10 - -- -* Method A (standard addition) standard solution is added to the sample solution with atomisation Method B (standard addition) standard solution is added to the solid Method C (standard addition) standard solution Method D solid sample is atomised in the graphite furnace. sample with atomisation in the graphite furnace. is added to the sample solution with atomisation in the flame.in the graphite furnace with solid hydroxyapatite as standard. t Average result with relative standard deviation (sc). 1 The sample employed as the solid standard. tj Relative standard deviation of the mean value. The content of an element in a reference material that has a suitable matrix composition can of course also be established by analysis. Trace element contents can often be deter-mined by decomposing the sample and carrying out the analysis by a conventional flame AAS method by atomising the sample solution in a furnace or by any other suitable method. These important determinations should be made by at least two reliable methods. The techniques of measurement may vary from the plotting of a full calibration graph before atomising the sample to the alternating measurement of standards and samples.As the sensitivity may vary during the measurements the latter approach is to be preferred. By carrying out at least five determinations of the analyte statistical methods e.g. for the rejection of data can be applied. Accuracy and Precision A number of systematic errors may affect the accuracy of the present technique and no estimate of the errors can be given. The most serious sources of systematic errors are the losses of the analyte by evaporation or sputtering during the drying and/or ashing steps incomplete atomisation chemical and spectral interferences the inability of the background corrector to compensate for large non-specific absorptions different rates of atomisation for the sample and standard (when peak heights are measured) varying times of residence of the atoms in the cell incorrec November 1979 BY ATOMIC-ABSORPTION SPECTROPHOTOMETRY 1009 adjustment of the sources of radiation swelling of the sample during drying ashing or atomisation the use of slow signal measuring systems for rapid signals and incorrect values assigned to the standard samples.The accuracy should of course always be established by analysing a reliable standard with the same or a similar matrix composition; the number of analyses should be sufficient to apply statistical methods such as the t-test to decide whether or not systematic errors are present. The precision is established by performing a series of analyses and assuming that the errors are random and have a normal distribution calculating the standard deviation.These calculations are normally made without establishing the actual distribution ; it should be noted that analytical results for trace elements may exhibit other distributions such as a log-normal distribution. In the direct analysis of solids the precision is mainly affected by the errors associated with the sampling and the absorption measurements. Relative standard deviations of 5-10% are frequently obtained for elements present at the 1 p.p.m. level; similarly at the 1 p.p.b. (parts per log) level values in the range l0-30% have to be considered as normal. Time of Analysis Assuming that the instrument is ready for measurement and that the sample has been prepared for analysis approximately ten solid samples can be analysed per hour.Contamination Control The direct AAS analysis of solids has the important advantage of not requiring any decomposition separation and/or concentration steps; compared with many other analytical methods the risks of introducing contaminants are considerably reduced. However the sample may take up impurities during the sampling and sample preparation steps such as crushing grinding sieving and homogenisation from the ambient atmosphere and from contact with glass plastic materials metal knives and spatulas. To avoid or reduce the risks of introducing metallic contaminants it is recommended to employ whenever possible, thoroughly cleaned non-metal equipment such as agate mortars and pestles nylon-meshed sieves plastic sample bottles and spatulas and silica knives.The impurities from the ambient air may be reduced by working in a glove-box fitted with filters or in a laboratory kept under a constant positive pressure of filtered air. For more comprehensive treatments of the contamination problems in trace analysis, reference is made to a monograph133 and a recent review paper.lM Applications In Table IV a bibliography is given of the applications of the direct AAS analysis of solids; in this list of applications all atomisation methods have been considered. The applications are classified according to the composition of the matrix the first group consisting of materials with an inorganic matrix and the second listing substances with an organic matrix. TABLE IV APPLICATIONS Refer-Materials analysed Elements determined ences Materials analySea Elements determined A .Mat& wifh a- inorganic matrix Acids (see Salts) Airsuspendedmatter Pb Pb Be Pb - Air suspended matter (cont.) Cd Pb 93 Ph _. 10s 106 9s cd Cd Cu Mn Pb Pb Pb 97 Alumina (see Aluminium Pb 96 oxide) -Ag Be Cd Hg Pb Se 107 Aluminium Cr Pb 108 Ca Cd 98 Zn Pb 100 Cu Mg Zn Cd Cu Pb Zn 101 T1 Be 104 cu Refer-ences 09 102 94 103 136 138 137 138 139 140 4 1010 LANGMYHR DIRECT ANALYSIS OF SOLIDS Analyst Vol. 104 TABLE IV-continued Refer-ences Refer-ences 138 77 Materials analysed Aluminium(cont.) . . . . Aluminium-base alloys . . Elements determined Materials analysed Iron cont). . . . . . . Elements determined Zn Cr Cu Fe Mn Ni Si Ti cu 141 77 85 Bi c cu Mn Mo Ni Si V BI 162 163 164 Zn Zn Cu Mg Mn Cu Mn Cu Ni c u 138 141 36 9 0 7R 80 67 69 35 137 9 142 i% Fe cu In Iron ores .. . . . . Iron oxide . . . . . . Laterite . . . . . . Lead . . . . . . Lead-base aliok . . . . Leadores . . . . . . Lithium compounds (see Magnesite brick : :: Magnesiumoxide . . . . Manganeseore . . . . Mercury ores Metal oxide films (see Tin ’ ‘ oxide films) Molecular sieve . . . Molybdenum . . . . Molybdenum oxide . . . . Nickel-base alloys . . . . Salts) 16 67 137 136 161 138 16 Aluminium oxide . . . . Ag Cd Cu Sb Zn Bi Bi ~~ c u Cu Fe Mg Co. Mo Pb G i Pd Zn -Ti cu cu -2 67 80 69 2 16 Antimony Apatite (see Hydroxyapatitk and Phosphate rock amcentrates) .. . . Ash. * . ~ m u c a r G l ; a t e . . . . Bases (seesalts) . . . . Bauxite . . . . Born nitride’ . . . . Brass . . . . . . co Ti Hg -Pb c u Ga In Cd Cu Ag Be Cd Fe Mn Ni, Fe Ni Cu Ni Cu Ni A1 Pb Zn T1 c u -Pb Sb --143 67 137 144 77 90 78 78 --Pb -9 166 69 88 130 166 166 82a 78 166 167 69 168 83 68 69 9% 169 170 80 60 171 --71 172 173 174 176 8 176 Al Zn Ca Cd Cu Eu Mg Zn Fe Pb Bronze . . . . . . Bi Pb SC Sn Te T1 Bi Pb Zn Ni 82 82a 146 80 Cadmium . . . . . . Calcium oxide Carbon (see Graphite) : Cast iron (see Iron) .. catalysts . . . . . . Chromiumoxide ._ Clay . . . . . . . . Coal . . . . . . . . Nickel compounds . . . . Niobium . . . . . . Niobium oxide . . Phosphate rock concentrai& Quartz . . . . . . Al Zn Si Al Cu Fe Ag Bi Cd In Pb T1 Zn cu --9 36 84 67 137 146 145 148 69 149 76 160 44 151 77 152 16 44 88 --a 69 67 153 153 69 84 -Pd Co Mo c u c u Ga In Cd Be Ca Fe Na Ni Pb Cd Cu Pb Zn Cu NI V Mn Zn P Fe Ni Si Zn Cd Pb Zn c u Fe K Na Na Ti co 2 --Au Au Quartz amorphous Quartz glass . . Rare earth oxides . . Rocks (see Silicates) Salts acids bases . . Eu Ti Ca Cu Fe K Mg Mn, Na ‘Li ‘Li c u Al Co Cu Fe Na Ni, Pb Zn -2; 2 sn TI copper .. . . . . Sediments (see also Silicates) . . silica (see Silicon &.ihe) Silicates (see also Sediments) Copper ores . . . . Copper slags (see Slags) . . corundum . . . . . . Crayons (see Paints) . . Dolomite . . . . . . Perrites . . . . . . Ferromanganese . . . . Ferrosilicon . . . . . . Fly ash (see Ash) . . . . Gadoliniumoxide . . . . Glass (see also Quartz) . . 3 Cr Cu Ni F% Cd cd CU Cd Pb Cd Pb 10 11 177 84 12 178 83 179 13 69 14 180 66 181 127 143 69 69 80 66 66 69 2 127 69 78 182 183 184 186 -Ca cd Cu Eu Mg Zn Ag Cu Mn Zn Cu Mn Zn Cu Ni Pb Au Au Cs Rb c u Ag Cd W T1 Zn Pb 164 127 143 Goldores . . . . . . Gold ore tailings .. . . Graphite . . . . . . 68 17 68 166 79 71 co Cs Li Rb cd Cr Cu Cs Li Rb c u Au Al. Zn Ag Ca Cu Ag’ Bi’cd Cr c u Mu &o. bb. )Sb.’Sn.kl. i n Pb Fe Fe c u Pb c u Feihin Pb ’ ’ . 144 9 Silicon carbide . . Silicon nitride . . Silicon oxide. . . . Pd As Cd Cu Fe Hg Pb, Se. Zn 156 86 167 168 68 69 169 132 160 69 161 Mn Ag Au Bi Cd H g Mu, hi $b gb i n .ri 27 elements Ag Bi Cd Cr Cu Fe, Mn Pb Cd Mn Pb Al Pb Bi As: ~~ cd c u P. v Slag . . . . . . Sludge . . . . CU Cd Cu Pb cd Cu Pb Cr Ni Zn Cd Cr Cu Ni, Al Bi Ca Cd, - Fe Mg fi, Pb Zn Pb Zn a cu Hpdroxyapatite . . . . Iodineoxide . . . . Inm . . . . . . . . Soil . . . . . . Steel . . . . .. 70 186 81 Cr Gr. Mn Ni S November 1979 BY ATOMIC-ABSORPTION SPECTROPHOTOMETRY 101 1 TABLE IV-continued Refer- Refer-Materials analysed Elements determined ences Materials analysed Elements determined e n c e Steel (cont.) . . . . . . Cr Mn Ni Bi 46 161 89 141 138 78 162 90 187 188 163 189 164 130 190 Leaves (see also Plants) 215 204 01s 219 146 59 164 220 415 146 204 221 Al Al Sb Sn Al Sb Sn Cr Cu Fe Ni Bi Al Cr Mo V Ag Bi Pb Zn co Pb ~~ Cd Cd Cu Pb Zn Be Al Cu Fe Mn Pb Sr, H g Cd H g Ag Cu Mn Pb Se Pb Zn Be Cd Cu Pb Zn cd Zn Liver . Al3 Cr Cu Mn Ni Pb Pb Ag Bi Cd Zn Pb Zn Ag Bi Cd I n Bi Cd Cu In Pb Sb, Te TI In Fe c u Al Zn Cr Cu Fe Mn Ni Si Fe Sn Sb Al Pb Cu Fe Mn Pb Cu Ni Al Zn Cd Li Co Cr Cu Mn Ni Cd Mg Mn Cd 1 B b I J 1%3U rrqj arru 113 81 154 59 82a 191 137 Sulphide ores .. . . 216 57 169 205 220 217 216 222 217 223 87 224 225 226 227 228 229 2 3.0 231 232 233 54 234 235 236 Pb Pb Cd Cu Mn Pb Cu Fe Mn Pb Zn Pb Cd Pb Ph Co Cr Cu Pb Zn Ag Bi Co Cu Fe Mn, Al Cr Cu Pb Zn c u c u Cu Zn Pb Pb Cd Pb Pb Cu Fe Mn Si Cd Cu Mn Ph Cu Pb Cu Fe Ag cd Co Cr Cu Fe, Co Cr Cu Fe Mn Ni Ni Pb Zn Mg Mn Pb Sb 192 137 69 80 155 193 69 37 91 82 194 65 127 155 195 196 197 197a 136 Sulphide ore concentrates . . Sulphur . . . . . . Synthetic mixtures.. . . Tantalum . . . . . . Lung . . . . . . Tin ores Tinoxide fi& 1 :: Titanium . . . . Titaniumoxide . . . . Molluscs . . . . Muscle . . . . Mussels . . . . Nails . . . . . . Tungsten . . . . . . Uranium . . . . . . Uranium oxide . . . . Zinc . . . . . . . Paint . . . . Paper . . . . 198 161 138 _ _ cd Cd Cd Cr Cu Fe Mn Ni Pb. Si 77 198 138 67 199 144 69 Zinc-base alloys . . . . Zinc oxide . . . . . . Zirconium . . . . . . Zirconium oxide . . . . Cd Cd c u La Y Bi Cd Pb Sb Sn Sb Plankton . . Plants (see also Grass, Leaves Tomatoes and Wheat) . . . . 87 70 70a 237 232 68 54 238 239 240 241 242 232 233 239 243 244 213 220 220 245 246 160 203 207 140 72 110 207 ---B.Materials with an organic m Algal cells . . . . . . Amoeba Blood-lyophiiised ashed ' . (seealsoSerum) . . . . Bone . . . . . . . . Dental material (see Teeth) Fibres (see Silk) . . . . Fish . . . . . . . . Fish meal . . cattix Mn Mn Polymers . . . 200 201 202 160 132 --203 204 205 204 206 207 208 209 210 211 -Au Al Cu Fe Au Cu Pb Cr Cu Fe Ti Si V Ag Zn Cd Mn P b Pb co P Co. Mn. Zn Pulp . . . . . . Resin (see Polymers) Rubbex . . . . Seaweed Serum ashed . : Silk . . . . . . Silkworm . . . . Silkworm egg . . Skin . . . . . . Teeth . . . . Textiles (see Silk) . . Tissue . . . . Tomatoes (see also Plant Urea . . . . Water suspended matter Wax crayons (see Paint) Wheat (see also Plants) Cu Fe Mn Si Cd Cu Mn Pb -Pb Si Sn Pb Grain (see Wheat) .. . . Grass (see also Plants) . . Hair . . . . . . . . -c u Pb co Ag Cu Fe Pb Cu Fe Mn Zn Cu Fe Mn Zn Cu Pb Ag AI Co Cr Cu Fe, Mn Xi Si Pb Ag '41 Bi Co Cr Cu, Fe Mg Mn Ni Pb, Si Zn Ag Co Cu Fe Ni Pb, Co Cu Fe Mn Pb Cr Cd Pb Ag Mn Zn 212 Si 213 214 132 215 216 21 7 Co Mn Zn c u TI Ag Cd Mn Pb Pb Ivory . . . . . . Kidney . . . . . . Hg Cd Pb -c u The appendix gives references to some studies that were found to give Iess reliable or only semi-quantitative results or which proved to be unsuccessful. Appendix The literature contains descriptions of atomisation cells which according to the a ~ t h o r s ~ ~ ~ ~ ~ are applicable to the direct analysis of solids; however in these papers no data relating to applications are given 1012 LANGMYHR DIRECT ANALYSIS OF SOLIDS Analyst Vol.104 AAS was usedlM to examine various grades of spectroscopic carbon in the form of rods, and arsenic cadmium copper iron mercury lead selenium and zinc were detected. The impurities were found to be distributed throughout the matrix as well as on the surface. By heating the rods at 1570 K for varying periods of time the content of impurities could be reduced; however cleaned rods were rapidly re-contaminated by absorption of metal impurities from the ambient atmosphere. For the atomisation of volatile elements microcrucibles made of platinum or graphite have been employed.8 The equipment was tested by atomising cadmium from a sample of syenite ; however analytical results were not given.Calibration graphs were reported for silver and gold in mixtures with graphite but no data were given for applications to inorganic and/or organic materials. Atternpt~2~~ at the direct AAS determination of cobalt in feed grains and forages were unsuccessful the reasons being stated to be a large non-specific absorption of radiation and other interference problems. Solid samples of polymers and filter-paper discs were atomised in an inverted T-type of furnace.24e The elements detected or determined in these samples were not reported. The direct determination of aluminium and vanadium in lyophilised and ground powders of bovine liver and of cat and rabbit brain has been attempted2S0; however the accuracy and precision obtained were unsatisfactory .A carbon dioxide laser has been employedse to atomise silver from copper-base alloys but the method was found to give only semi-quantitative results. The content of chromium in brewer’s yeast has been determined by the solid sampling technique261; however the results were much lower than those from samples wet ashed in a closed vessel. The former analysis included ashing at temperatures up to 1620 K and it seems highly probable that chromium was lost at this high temperature. An a.c. arc has been used73 as an atomiser for AAS. Conclusion From the applications listed in Table IV it appears that 40 elements have been determined by the present technique the concentrations of the analytes ranging from more than 100 p.p.m.to about 0.1 p.p.b. (lo9). All types of solid materials can be analysed the only limitation being that the sampling error should be acceptable for amounts of sample of the order to 1-10 mg. Direct analyses of solids have been made with a wide variety of equipment; a number of atomisation cells were home-made and of special designs and the instrumental arrangements have also varied considerably. The most useful and extensively employed atomisation cell for solids is the simple cylindrical graphite tube. The direct AAS analysis of solids has considerable potentiality as a general method for trace analysis and will be of particular value in those fields such as medicine biology and environmental chemistry where only small amounts of sample are available.1. 2. 3. 4. 6. 6. 7. 8. 9. 10. 11. 12. 13. 14. References Walsh A. Spectrochim. Ada 1955 7 108. L‘vov B. V. Talanta 1976 23 109. Razumov V. A. Zh. Prikl. Spektrosk. 1976 24 1117. Wanninen E. Editor. “Analytical Chemistry. Essays in Memory of Anders R.ingbom,” Pergamon Langmyhr F. J. Talanta 1977 24 277. Fuller C. W. “Electrothermal Atomization for Atomic Absorption Spectrometry,” Analytical “Annual Reports on Analytical Atomic Spectroscopy,” The Chemical Society London. Prudnikov E. D. Zh. Prikl. Spektrosk. 1972 17 352. Coudert M. A. and Vergnaud J. M C.R. Hebd. Skanc. Acad. Sci. Paris 1969 268 1225; Analyt. Govindaraju K Mevelle. G and Chouard C. Chem. Geol. 1971 8 131.Govindaraju K. Mevelle G. and Chouard C. Bull. SOC. Fr. Ckram. 1972 96 47. Govindaraju K. Herman R. Mevelle G. and Chouard C. Atom. Absorption Newst. 1973 12 73. Govindaraju K. Mevelle G. and Chouard C. Analyt. Chem. 1974. 46 1672. Govindaraju K. Analusis 1976 3 164. Press Oxford 1977 p. 461. Sciences Monographs No. 4 The Chemical Society London 1977. Chem. 1970 42 1303 November 1979 BY ATOMIC-ABSORPTION SPECTROPHOTOMETRY 1013 16. 16. 17. 18. 19. 20. 21. 22. 23. 24. 26. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 61. 52. 53. 54. 66. 66. 67. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70. 70a.71. 72. 73. 74. 76. 76. 77. 78. Venghiattis A. A. Atom. Absorption Newsl. 1967 6 19. Venghiattis A. A. and Whitlock L. Atom. Absorption Newsl. 1967 6 136. Venghiattis A. A. Spectrochim. Acta 1967. 23B 67. Miaud P. and Robin J. Bull. Soc. Chim. Fr. 1968 No. 2 864. Sprague S. and Slavin W. Atom. Absorption Newsl. 1963 2 20. Sprague S. and Slavin W. Atom. Absorption Newsl. 1966 4 367. Burrows J. A. Heerdt J . C. and Willis J . B. Analyt. Chem. 1966 37 579. Means E. A. and Ratcliff D. Atom. Absorption Newsl. 1966 4 174. Slavin S. and Slavin W. Atom. Absorption Newsl. 1966 5 106. Moore E. J. Milner 0. I. and Glass J. R. Microchem. J. 1966 10 148. Mostyn R. A. and Cunningham A. F. J. Inst. Petrol. 1967 53 101. Schallis J. E. and Kahn H. L. Atom.Absorption Newsl. 1968 7 84. Bartels T. T. and Slater M. P. Atom. Absorption Newsl. 1970 9 76. Kriss R. H. and Bartels T. T. Atom. Absorption Newsl. 1970 9 78. Kahn H. L. Peterson G. E. and Manning D. C. Atom. Absorption Newsl. 1970 9 79. Jackson D. R. Salama C. and Dunn R. Can. Spectrosc. 1970 15 17. Taylor J. H. Bartels T. T. and Crump. N. L. Analyt. Chem. 1971 43 1780. Lacour A. Teinturier C. and Isabelle D. B. Me’thodes Phys. Anal. 1971 7 49. Barnett W. B. Kahn H. L. and Peterson G. E. Atom. Absorption Newsl. 1971 10 106. Jackson K. W. Aldous K. M. and Mitchell D. G. Appl. Spectrosc. 1974 28 669. Kashiki M. and Oshima S. Analytica Chim. Ada 1970 51 387. Ghiglione. M. Eljuri E. and Cuevas C. APPZ. Spectrosc. 1976 30 320. Harrison W. W. and Juliano P. O. Analyt.Chem. 1971 43 248. Willis J. B. Analyt. Chem. 1976 47 1752. Fuller C. W. Analyst 1976 101 961. L’vov B. V. “Atomic Absorption Spectral Analysis,” Second Edition Adam Hilger Bristol 1970. Lebedev V. I. Zh. Analit. Khim. 1969 24 337. Bastiaans G. J. and Hieftje G. M. Analyt. Chem. 1974 46 901. Panichev N. A. Prudnikov E. D. Tatti A. V. and Turkin Y. I. Zh. Prikl. Spektrosk. 1975 22, Kantor T. and Pungor E. Paper presented at the 17th Colloquium Spectroscopicum Internationale, Human H. G. C. Scott R. H. Oakes A. R. and West C. D. Analyst 1976 101 265. Winge R. K. Fassel V. A. and Kniseley R. N. Appl. Spectrosc. 1971 25 636. King A. S. Astrophys. J. 1908 27 363. Maller L. A. Annln Phys. 1927 82 39. L‘vov B. V. J. Eng. Phys. 1969 2 44. Kirkbright G. F. Analyst 1971 96 609.Szydlowski F. J. Peck E. and Bax B. Appl. Spectrosc 1978 32 402. Fritzsche H. Wegscheider W. Knapp G. and Ortner H. M. Talanta 1979 26 219. Woodriff R. Stone R. W. and Held A. M. Appl. Spectrosc. 1968 22 408. Robinson J . W. Wolcott D. K. and Rhodes L. Analytica Chim. Acta 1976 78 286. Talmi Y. and Morrison G. H. Analyt. Chem. 1972 44 1466. Headridge J. B. and Smith D. R. Talanta 1972 19 833. Hadeishi T. and McLaughlin R. D. Analyt. Chem. 1976 48 1009. Katskov D. A. Kruglikova L. P. L’vov B. V. and Polzik L. K . Zh. Prikl. Spektrosk. 1974 20, Nichols J. A. Jones R. D. and Woodriff R. Andyt. Chem. 1979 50 2071. Brodie K. G. and Matougek J. P. Analyt. Chem. 1971 43 1667. Omang S. Analytica Cham. Ada 1971 56 470. Alder J. F. and West T. S Analyfica Chim.Acta 1972 58 331. Reeves R. D. Molnar C. J. and Winefordner J. D. Analyt. Chem. 1972 44 1913. Reeves R. D. Molnar C. J. Glenn M. T. Ahlstrom J. R. and Winefordner J . D. Analyt. Chem., See ref. 39. Fuller C. W. and Thompson I. Analyst 1977 102 141. Borzov V. P. L’vov B. V. and Plyushch G. V. Zh. Prikl. Spektrosk. 1969 11 217. Men’shikov V. I. Malykh V. D. and Shestakova T. D. Zh. Andit. Khim. 1974 29 2132. Katskov D. A. Kruglikova L. P. and L’vov B. V. Zh. Analit. Khim. 1976 30 238. Razumov V. A. Zh. Analit. Khim. 1977 32 383. Jackson K. W Ebdon L. and Webb D. C. Paper presented a t the 21st Colloquium Spectro-Belyaev Y. I. Ivantsov L. M. Karyakin. A. V. Pham Hung Phi and Shemet V. V. Zh. Analit. Belyaev Y. I. and Gordeev V. V. Oceanology 1972 12 766.Malykh V. D. Men’shikov V. I. Morozov V. N. and Shipitsyn S. A. Zh. Prikl. Spekfrosh. 1972, Russel B. J. and Walsh A. Specfrochim. Acfa 1969 15. 883. Gatehouse B. M. and Walsh A. Spectrochim. Acta 1960 16 602. Walsh A. Appl. Spectrosc. 1973 27 336. Gough D. S Analyf. Chem. 1976 48 1926. McDonald D. C. A d y f . Cbm. 1977 49 1336. 391. Florence 1973. 739. 1972 44 2206. scopicum and 8th International Conference on Atomic Spectroscopy Cambridge 1979. Khim. 1968 23 508. 16 12 1014 LANGMYHR DIRECT ANALYSIS OF SOLIDS Analyst Vol. 104 79. 80. 81. 82. 82a. 83. 84. 86. 86. 87. 88. 89. 90. 91. 92. 93. 94. 96. 96. 97. 98. 99. 100. 101. 102. 103. 104. 106. 106. 107. 108. 109. 110. 111. 112. 113.114. 116. 116. 117. 118. 119. 120. 121. 122. 123. 124. 126. 126. 127. 128. 129. 130. 131. 132. 133. 134. 135. 136. 137. 138. 139. Mossotti V. G. Laqua K. and Hagenah W.-D. Spectrochim. Acta 1967 23B 197. Karyakin A. V. and Kaigorodov V. A. Zh. Analit. Khim. 1968 23 930. Krivchikova 8. P. and Demin V. S. Zh. Prikl. Spektrosk. 1971 14 592. Nikolaev G. I. and Podgornaya V. I. Zh. Prikl. Spektrosk. 1972 16 911. Thomerson D. R. and Godden R. G. Baird-Atomic Ltd. Braintree Essex personal communication. Karyakin A. V. Pchelintsev A. M. Shidlovskii A. I. Vul’fson E. K. and Tsingarelli M. N. Zh. Vul’fson E. K. Karyakin A. V. and Shidlovskii A. I. Zh. Analit. Khim. 1973 28 1253. Osten D. E. and Piepmeier E. H. Appl. Spectrosc.1973 27 165. Vul’fson E. K. Karyakin A. V. and Shidlovskii A. I. Zav. Lab. 1974 40 945. Vul’fson E. K. Gribovskaya I. F. Karyakin A. V. and Yanushkevich A. F. Zh. Analit. Khim., KQntor T. Pdos L. Fodor P. and Pungor E. Talanta 1976 23 585. MatonSek J. P. and Orr B. J. Spectrochim. Acta 1976 31B 475. Ishizuka T. Uwamino Y. and Sunahara H. Analyt. Chem. 1977 49 1339. Kaigorodov V. A. Zh. Prikl. Spektrosk. 1975 23 389. Kaigorodov V. A. and Grishko G. A. Zh. Analit. Khim. 1977 32 205. Loftin H. P. Christian C. M. and Robinson J . W. Spectrosc. Lett. 1970 3 161. Christian C. M. 11 and Robinson J . W. Analytica Chim. Acta 1971 56 466. Matousek J. P. and Brodie K. G. Analyt. Chem. 1973 45 1606. Amos M. D. Brodie K. G. and Matoulek J. P. “Proceedings of the Pittsburgh Conference on Roques Y.and Mathieu J. Analusis 1973 2 481. Brodie K. G. and Matougek J. P. Analytica Chim. Acta 1974 69 200. Brodie K. G. and Matoulek J . P. Boll. Lab. Chim. Prov. (Bologna) 1975 1 279. Noller B. N. and Bloom H. Atmos. Environ. 1975 9 505. Noller B. N. Bloom H. and Parker C. R. “Proceedings of the 5th International Conferenceon Atomic Spectroscopy Melbourne Australia 1976,” Australian Academy of Science Dominion Press Victoria 1976. Noller B. N. and Bloom H. Analyt. Chem. 1977 49 346. Thomassen Y. Solberg R. and Hanssen J. E. Analytica Chim. Acta 1977 90 279. Havezov I. and Tamnev B. 2. Analyt. Chem. 1978 291 127. Woodriff R. and Lech J. F. Analyt. Chem. 1972 44 1323. Siemer D. Lech J. F. and Woodriff R. Spectrochim. Ada 1973 28B 469. Siemer D.D. and Woodriff R. Spectrochim. Ada 1974 29B. 269. Lech J. F. Siemer D. and Woodriff R. Environ. Sci. Technol. 1974 8 840. Kriese M. Kuhl J. Neumann S. and Torge R. German (BRD) Patent 2435091 1976. Favretto Gabrielli L. Pertoldi Marletta G. and Favretto L. Atom. Absorption Newsl. 1977 16 4. Sturgeon R. E. and Chakrabarti C. L. Prog. Analyt. Atom. Spectrosc. 1978 1 5. L‘vov B. V. Sfiectrochim. Acta 1978 33B. 153. Ediger R. D. Atom. Absorption Newsl. 1975 14 127. Aggett J. and Sprott A. J. Analytica Chim. Acta 1974 72 49. Campbell W. C. and Ottaway J. M. Talanta 1974 21 837. Fuller C. W. Analyst 1974 99 739. Paveri-Fontana S. L. Tessari G. and Torsi G. Analyt. Chem. 1974 46 1032. Torsi G. and Tessari G. Analyt. Chem. 1975 47 839. Tessari G. and Torsi G.Analyt. Chem. 1975 47 842. Sturgeon R. E. and Chakrabarti C. L. Analyt. Chem. 1977 49 1100. van den Broek W. M. G. T. and de Galan L. Analyt. Chem. 1977 49 2176. West T. S. Pure APPL. Chem. 1978 50 837. Manning D. C. Atom. Absorption Newsl 1978 17 107. Fernandez F. J. Myers S. M. and Slavin W. Perkin-Elmer Corp. Norwalk Conn. personal Massmann H. El Gohary Z. and Giicer S. Spectrochim. Acta 1976 31B 399. Frech W. and Cedergren A. Analytica Chim. Acta 1976 82 93. Wibetoe G. Thesis Department of Chemistry University of Oslo Norway 1978. Wilson A. D. Analyst 1964 89 18. Solberg. R. Thesis Department of Chemistry University of Oslo Norway 1973. Lundberg E. and Frech W. Analytica Chim. Acta 1979 104 75. Koch 0. G. Pure Appl. Chem. 1978 50 1551. Langmyhr F. J.and Kjuus I. Analytica Chim. Ada 1978 100 139. Zief M. and Mitchell J. W. “Contamination Control in Trace Element Analysis Chemical Analysis, Volume 47,” John Wiley New York 1976. Mizuike A. and Pinta M. Pure Appl. Chem. 1978 50 1519. Siemer D. D. Koteel P. Wei H.-Y. Emmel R. and Sotera J. in Ewing G. W. Editor “Environ-mental Analysis,” Academic Press New York 1977 p. 175. Massmann H. Spectrochim. Acta 1970 25B 393. Langmyhr F. J. and Rasmussen S. Analytica Chim. Acta 1974 72 79. Headridge J . B. Lab. Pract. 1974 23 5 . Quentmeier A. Hagenah W.-D. and Laqua K. “Proceedings of the 5th International Conference on Atomic Spectroscopy Melbourne Australia 1975,” Australian Academy of Science, Dominion Press Victoria 1976. Prikl. Spektrosk. 1973 18 610. 1975 30 1913.Analytical Chemistry and Applied Spectroscopy Cleveland Ohio 1973.” communication 140. 141. 142. 143. 144. 145. 146. 147. 148. 149. 150. 151. 152. 153. 154. 166. 156. 157. 158. 159. 160. 161. 162. 163. 164. 165. 166. 167. 108. 169. 170. 171. 172. 173. 174. 175. 176. 177. 178. 179. 180. 181. 182. 183. 184. 185. 186. 187. 188. 189. 190. 191. 192. 193. 194. 195. 196. 197. 197a. 198. November 1979 BY ATOMIC-ABSORPTION SPECTROPHOTOMETRY Headridge J. B. and Smith D. R. Talanta 1971 18 247. 1015 Gries W. H. and Norval E. Analytica Chim. Acta 1975 75 289. Ashy M. A. Headridge J . B. and Sowerbutts A. Talanta 1974 21 649. Nikolaev G. I. Zh. Analit. Khim.1964 19 6s. Siemer D. D. and Wei H.-Y. Analyt. Chem. 1978 50 147. Katskov D. A. and L‘vov B. V. Zh. Prikl. Spektrosk. 1969 10 382. Norval E. and Gries W. H. Analytica Chim. Acta 1976 83 393. Talmi Y. Analyt. Chem. 1974 46 1005. Gladney E. S. Atom. Absorption Newsl. 1977 16 42. O’Reilly J. E. and Hale M. A. Analyt. Lett. 1977 10 1095. Langmyhr F. J. and Aadalen U. Department of Chemistry University of Oslo Norway, Pemsler J . F. and Rapperport E. J. Analytica Chim. Acta 1972 58 15. Sullivan J. V. personal communication cited in Walsh A. APPZ. Spectrosc. 1973 27 335. KAntor T. Fodor P. and Pungor E. Analytica Chim. Acta 1978 102 15. Langmyhr F. J. Thomassen Y. and Massoumi A. Analytica Chim. Acta 1973 67 460. Owens J. W. and Gladney E. S. Atom. Absorption Newsl.1975 14 76. Nikolaev G. I. Zh. Analit. Khim. 1965 20 445. Robinson J. W. and Hindman G. D. Spectrosc. Lett. 1972 5 169. Belyaev Y. I. Koveshnikova T. A. and Kostin B. I. Zh. Analit. Khim. 1973 28 2111. Belyaev Y. I. and Koveshnikova T. A. Zh. Analit. Khim. 1974 29. 659. Grushko L. F. Krasil’shchik V. Z. Lifshits M. G. and Chupakhin M. S. Zh. Analit. Khim. 1977, Langmyhr F. J. Lind T. and Jonsen J Analytica Chim. Acta 1975 80 297. Headridge J. B. and Smith D. R. Talanta 1972 19 833. Andrews D. G. and Headridge J . B. Analyst 1977 102 436. Aziz-Alrahman A. M. and Headridge J . B. Talanta 1978 25 413. Andrews D. G. Aziz-Alrahman A. M. and Headridge J. B. Analyst 1978 103 909. Marks J . Y. Welcher G. G. and Spellman R. J. Appl. Spectrosc. 1977 31 9. Headridge J .B. and Thompson R. Analytica Chim. Acta 1978 102 33. Timofeev €3. F. Ivanova M. I. Elyutin A. V. Maiorov I. A. Nedler V. V. and Filimonov L. N., Langmyhr F. J. Solberg R. and Thomassen Y. Analytica Chim. Acta 1977 92 105. Koizumi H. and Yasuda K. Analyt. Chem. 1976 48 1178. Goleb J . A. and Yokoyama Y. Analytica Chim. Acta 1964 30 213. Brady D. V. Montalvo J . G. Jr. Glowacki G. and Pisciotta A. Analytica Chim. Ada 1974, 70 448. Katskov D. A. and L’vov B. V. “The Analytical Chemistry of Rare Metals and Semiconductor Materials.” (in Russian) Moskovski Dom Nauchno-tekhnicheskoi Propagandi Imeni F. E. Dzerhinsky Moscow 1970. Belyaev Y. I. Pchelintsev A. M. and Zvereva N. F. Zh. Analit. Khim. 1971 26 1295. Luecke W. Eschermann F. Lennartz U. and Papastamataki A.J. Neues Jb. Miner. Abh. 1974, Belyaev Y. I. Pchelintsev A. M. Zvereva N. F. and Kostin B. J. Zh. Analit. Khim. 1971 26, Goleb J. A. Appl. Spectrosc. 1971 25 522. Data from University of Gottingen Department of Mineralogy Federal Republic of Germany, Langmyhr F. J. and Thomassen Y. 2. Analyt. Chem. 1973 264 122. Langmyhr F. J. Stubergh J. R. Thomassen Y. Hanssen J . E. and DoleZal J. Analytzca Chim. Gong H. and Suhr N. H. Analytica Chim. Acta 1976 81 297. Bouvier J.-L. and Abbey S. Analusis 1977 5 332. Mitchell D. G. Mills W. N. Ward A. F. and Aldous K. M. Analytica Chim. Acta 1977 90 275. Lester J. N. Harrison R. M. and Perry R. Sci. Total Environ. 1977 8 163. Stoveland S. Astruc M. Perry R. and Lester J . N. Sci. Total Environ. 1978 9 263. Carrondo M. J.T. Perry R. and Lester J. N. Analytica Chim. Acta. 1979 106 309. Ghiglione M. and Eljuri E. Acta Cient. Venez. 1970 21 123. Backman S. Analytical Chemistry Section Sandvik AB 81 1 01 Sandviken Sweden personal Tsujii K. and Kuga K. Bunseki Kagaku 1978 27 415. Eljuri E. Lunar J. and Ghiglione M. Appl. Spectrosc. 1979 33 170. Lundberg E. and Frech W. Analytica Chim. Acta 1979 108 75. Langmyhr F. J. Solberg R. and Wold L. T. Analytica Chim. Acta 1974 69 267. Karmanova N. G. and Pogrebnyak Y. F. Zh. Analit. Khim. 1979 34 213. Ishibashi W. Sat6 M. and Hashimoto M. Bunseki Kagaku 1974 23 597. Kanda M. Hori Y. and Matsumoto I. Bunseki Kagaku 1975 24 299. Goleb J . A. Analytica Chim. Acta 1966 34 135. Goleb J. A. Analyt. Chem. 1963 35 1978. Page A. G. Godbole S. V. Deshkar.S. B and Joshi B. D. Analyt. Lett. 1978 All 619. Page A. G. Godbole S. V. Kulkarni M. J. Shelar S. S. and Joshi B. D. Z. Analyt. Chem. 1979, unpublished data. 32 218. Zh. Analit. Khim. 1977 32 507. 120 178. 492. reported by Welz B. CZ-Chem.-Techn. 1972 1 460. Acta 1974 71 35. communication. 296. 40 1016 LANGMYHR 199. Krivchikova E. P. and Vasileva N. M. Zh. Analit. Khim. 1973 28 928. 200. Lorch D. W. Atom. Absorption Newsl. 1978 17 95. 201. Data from Carlsberg Laboratory Copenhagen Denmark reported by Welz B. CZ-Chem.-Techn., 202. Langmyhr F. J. Busch S. G. and Knutsen T. Department of Chemistry University of Oslo, 203. Lundgren G. and Johansson G. Talanta 1974 21 257. 204. Hadeishi T. Appl. Phys. Lett. 1972 21 438. 205. Langmyhr F. J and Aamodt J.Analytica Chim. Ada 1976 87 483. 206. Langmyhr F. J. and Orre S. Department of Chemistry University of Oslo to be published. 207. Grobenski Z. Melcher M. and Welz B. 2. Analyt. Chem. 1978 290 144. 208. Renshaw G. D. Pounds C. A. and Pearson E. F. Nature Lond. 1972 238 162. 209. Renshaw G. D. Pounds C. A. and Pearson E. F. J . Forens. Sci. 1973 18 143. 210. Alder J . F. Alger D. Samuel A. J. and West T. S. Analytica Chim. Acta 1976. 87 301. 211. Graef V. J. Clin. Chem. Clan. Biochem. 1976 14 181. 212. Alder J F. Samuel A. J. and West T. S. Analytica Chim. Acta 1976 87 313. 213. Alder J . F. Pankhurst C. A. Samuel A. J. and West T. S. Analytica Chim. Acta 1977 91 407. 214. Alder J . F. Samuel A. J. and West T. S. Analytica Chim. Acta 1977 94 187. 216. Church D.A. Hadeishi T. Leong L. McLaugNin R. D. and Zak B. D. Analyt. Chem. 1974, 216. Jackson K. W. and Mitchell D. G. Analytica Chim. Acta 1975 80 39. 217. Jackson K. W. Marczak E. and Mitchell D. G. Analytica Chim. Acta 1978. 97 37. 218. Kahn H. L. Fernandez F. J. and Slavin S. Atom. Absorption Newsl. 1972 11 42. 219. Brady D. V. Montalvo J. G. Jr. Jung J. and Curran R. A. Atom. Absorption Newsl. 1974 13, 220. Fujiwara K. Umezawa Y. Numata Y. Fuwa K. and Fujiwara S. Bunseki Kagaku 1977 26, 221. Pickford C. J. and Rossi G. Atom. Absorption Newsl. 1976 14 78. 222. Stegavik K. Mikalsen G. Ophus E. M. and Mylius E. A. Bull. Environ. Contam. Toxicol. 1976, 223 Lord D. A. Breck W. G. and Wheeler R. C. ASTM Spec. Techn. Pubn 1976 573 95. 224. Lord D. A. McLaren J. W. and Wheeler R.C. Analyt. Chem. 1977 49 257. 226. Barnett W. B. and Kahn H. L. Clin. Chem. 1972 18 923. 226. Van Stekelenburg G. J. Van de Laar A. J. B. and Van der Laag Clinica Chim. Ada 1976 59, 227. Sohler A. Wolcott P. and Pfeiffer C. C. Clinica Chim. Acta 1976 70 391. 228. Henn E. L. Atom. Absorption Newsl. 1973 12 109. 229. Mitchell D. G. Aldous K. M. and Ward A. F. Atom. Absorption Newsl. 1974 13 121. 230. Lau 0. W. and Li K. L. Analyst 1975 100 430. 231. Castellani F. Riccioni R. Gusteri M. Bartocci V. and Cescon P. Atom. Absorption Newsl., 232. Kerber J . D. Koch A. and Peterson G. E. Atom. Absorption Newsl. 1973 12 104. 233. Langmyhr F. J. Thomassen Y. and Massoumi A. Analytica Chim. Acta 1974 68 305. 234. Sychra V. Kolihova D. and Dudova N. Chemicki Listy 1976 70 737.236. Simon P. J Giessen B. C. and Copeland T. R. Analyt. Chem. 1977 49 2285. 236. Machiroux R. and Dupont J . C. Analytica Chim. Acta 1976 85 231. 237. Kerber J. D. Atom. Absorptiort Newsl. 1971 10 104. 238. Henn E. L. Analytica Chim. Acta 1974 73 273. 239. Hwang J. Y . Instrumentation Laboratory Inc. Wilmington Massachusetts USA personal 240. Girgis-Takla P. and Chroneos I. Analyst 1978 103 122. 241. Hein H. and Schrader W. “Applied Atomic Absorption Spectroscopy,” Pamphlet No. 3 1976, 242. Ediger R. D. Knott A. R. and Peterson G. E. Perkin-Elmer Corp. Norwalk Conn. USA. 243. Favretto L. Atom. Absorption Newsl. 1976 15 98. 244. Muzzarelli R. A. A. and Rocchetti R. Talanta 1976 22 683. 245. LidCn S. and Lundberg E. J. Invest. Derm. 1979 72 42. 246. Langmyhr F. J. Sundli A. and Jonsen J. AnaZytica Chim. Actu 1974 73 81. 247. Morrison G. H. and Talmi Y. Analyt. Chem. 1970 42 809. 248. Hageman L. Torma L. and Ginther €3. E. J. Ass. Ofi. Analyt. Chem. 1976 58 990. 249. Robinson J. W. and Wolcott D. K. Analytica Chim. Acta 1975 74 43. 260. Krishnan S. S. Quittkat S. and Crapper D. R. Can. J. Spectrosc. 1976 21 25. 261. Behne D. Bratter P. Gessner H. Hube. G. Mertz W. and Rosick. U. 2. Analyt. Chem. 1976, Received Awil 30th 1979 Accepted June 26th 1979 1972 1 460. Norway unpublished data. 46 1352. 118. 736. 15 734. 233. 1977 16 57. communication. Bodenseewerk Perkin-Elmer p. 48. 278 269

ISSN:0003-2654    

DOI:10.1039/AN9790400993

出版商:RSC    

年代:1979

数据来源: RSC

摘要:

Analyst, November, 1979, Vol. 104, p p . 101 7-1029 1017 Determination of Lead, Bismuth, Zinc, Silver and Antimony in Steel and Nickel-base Alloys by Atomic-absorption Spectrometry Using Direct Atomisation of Solid Samples in a Graphite Furnace Svenerik Backman and Rune W. Karlsson Sandvik AB, 45-TMK, Fack, S-811 00 Sandviken, Sweden A fast and simple method of determining lead, bismuth, zinc, silver and antimony in steel and nickel-base alloys has been developed using unmodified commercial atomic-absorption equipment. The method is based on the direct atomisation of solid metal samples in a graphite furnace. The samples can weigh between 1 and 20 mg, but test results are influenced by the shape of the samples. Calibration graphs have been drawn using steel samples with known contents.Practical content ranges, e.g., lead 0.03-150 p g g-l and bismuth 0.03-6 pg g-l, and low detection limits, down to 0.02 p g g-l, have been obtained by selecting suitable lines of analysis. The relative standard deviation (1s) is approximately 6% of the content of all elements investigated throughout the stated range of content. The time required for analysis is short, being about 6 min for a duplicate determination. The matrix effects are very slight. Keywords : Direct analysis of solid samples ; bad, bismuth, zinc, silver and antimony determination ; steel and nickel-base analysis ; atomic-absorption spectrometry ; graphite furnace It is well known that even small concentrations of certain trace elements, e.g., lead, bismuth, zinc and tin, will reduce hot ductility, creep strength and weldability and impair certain of the mechanical and physical properties of various steel products.High-alloy steels and superalloys are particularly susceptible in this respect. Many of these trace elements are characterised by high volatility at elevated temperatures compared with the main elements of a steel or a nickel-base alloy (see Table I). The effects of trace elements in steels and nickel-base alloys have been described by Mayer and Clark1 and by Ljungstrom.2 The maximum limits that have been set for impurities in certain types of material are very low. The German DIN 1701 specification for refined nickel, for example, stipulates a limit of 1 pg g-l for lead, zinc, silver and antimony and 0.2 pg g-l for bismuth.So far it has not been possible to determine these low contents without time-consuming separation^,^^^ the purchase of expensive a p p a r a t ~ s ~ . ~ or the construction of frequently complex home-built TABLE I ELEMENTS WITH BOILING-POINTS BELOW 2 250 "C AND BOILING-POINTS OF MAIN ELEMENTS Boiling-point/ Element "C Sb* 1750 As* 613 (subl.) At 337 Ba 1640 Pb* 1740 cs 678 P (white) 280 Cd* 766 Ca* 1484 In 2 080 K 774 * Investigated by us. Element Mg* Mn Na Ag* Se* T1 Te Sn* Bi* Zn* Boiling- point/"C 1090 1962 883 2212 685 1460 990 2 270 1560 907 Main element Fe Si Cr Ni Mo w Ti Nb Ta co cu Boiling- point/"C 2 750 2 355 2 672 2 732 4612 5 660 3 287 4 742 5 420 2 870 2 6671018 BACKMAN AND KARLSSON: DETERMINATION OF Pb, Bi, Zn, Analyst, Vol. 104 equipment .‘j--ll More recently developed techniques for the routine determination of trace elements in steel have included atomic-absorption spectrometry with electrothermal atomisation.12 Normally, however, this technique has demanded the dissolution of the metal samples in mineral acids as a preparatory treatment, and it has involved considerable matrix effects.Although the graphite furnace technique has been known13J4 for amost 20 years and has been commercially available for about 10 years, it has not been widely used in the analysis of solid metal samples. The authors of most of the methods referred to in the literature have used graphite furnaces of their own design. This was done, for example, by Langmyhr and c ~ - w o r k e r s ~ ~ ~ ~ J ~ in determining cadmium, lead, silver, thallium and zinc in minerals, silicon, f errosilicon and ferromanganese, while Headridge and co-~orkers*~~7~18 also used a furnace of their own design for the determination of, among other substances, bismuth in steel (limit of detection = 0.004 pg g-l).On the other hand, Lundberg and Frech19 and Marks et aL20 used commercial furnaces and compared various makes when determining lead, bismuth, selenium, tellurium, thallium and tin. Most commercial furnaces are designed for analysing solutions, and for this reason there are usually difficulties involved in using them for the direct analysis of solid samples. Instrumentation Laboratory Inc., however, has augmented its graphite furnace with standard equipment employing a test boat technique that has proved very suitable for the direct analysis of solid samples.During our search for faster and more sensitive methods of trace-element determination, we considered atomic-absorption spectrometry with electro- thermal atomisation combined with dissolution in mineral acids to be an interesting alternative,l2 but after a few months’ experimentation we encountered the following dis- advantages, which have also been described in the literature : (a) much depends on the way in which a sample is injected into the graphite furnace and (b) the method has pronounced matrix effects. These experiments had very promising results, and the technique has been in use for the routine determination of certain trace elements since February 1976. Instead, our main concern has been applicability to concentration ranges at practically useful levels and to obtain measurements that are sufficiently accurate and precise and that have sufficiently low limits of detection to facilitate a quantitative study of the influence of the trace elements on material properties. The sensitivity of this method is nonetheless adequate, and there are few techniques that can offer the same low limit of detection as the direct atomisation of solid samples.Direct atomisation of solid metal specimens was tested a t the beginning of 1976. No great efforts have been made to obtain extremely low limits of detection. Experimental Materials Steel and nickel-base standards Commercial standards from British Chemical Standards (BCS) , the National Bureau of Standards (NBS) and Jernkontoret (JK, sold by the Instituet for Metallforskning) were used, together with BCO standards from AB Bofors St%l, Bofors, Sweden, trace alloys from the USA and internal standard materials.Graphite tubes and boats RWO 332) were used. Gas Pyrolysed square section tubes (IL 441 18) and unpyrolysed graphite boats (Ringsdorff Spectrally pure argon. Apparatus A tomic-absorption spectrometry An Instrumentation Laboratory, Model 251, double-beam instrument with facilities that include background correction, peak-height measurement, peak-area measurement and A-, 8-, 4- or 16-s integration time was used.November, 1979 Graphite furnace The furnace was an Instrumentation Laboratory, Model 455, with facility for gas stop while pressurised, six heating steps [two for drying (steps 1 and 2), two for ashing (steps 3 and 4) and two for atomisation (steps 5 and 6) (see Table 11)] plus a “Clean” step giving 80% of maximum power for 5 s.The time for each step is pre-set by means of a thumb-wheel. Each “click” represents 5 s, except for the first atomisation step (step 5), where each click corresponds to 1 s. The furnace temperature is measured by means of a tungsten-wire resistance thermometer. The maximum temperature is approximately 3 500 oC,21 Le., the tungsten melts. Ag AND Sb IN STEEL AND NICKEL-BASE ALLOYS BY AAS 1019 TABLE I1 ANALYTICAL PARAMETERS Pb Bi Equipment Parameter - Zn Ag Sb Spectrometer . . . . Content rangelwg g-l 0.03-15 1-150 0.03-3 0.2-6 3-140 0.02-2 6-200 Lamp current/mA 5 2 3.5 3.5 5 5 8 Wavelength/nm 283.3 363.9 306.8 306.8 307.6 338.2 259.8 Band width/nm 0.3 0.3 0.5 0.5 0.5 0.3 0.3 Integration time/s 16 4 4 16 16 16 16 Mode* SBA SBA SBA SBA SBA SBA SBA Detector voltage/V 620 700 700 700 620 700 620 Sample masslmg 1-4 1-15 3-7 3-7 3-8 1-4 18-22 - kurnace .. . . Position 5 6 mmmmmm Power, % 75 60 90 80 75 80 70 80 50 65 60 60 70 72 Integration start X X X X X X X Time/s 8 1 0 7 5 6 5 6 1 0 8 1 0 7 1 5 6 1 0 “Clean” on Yes No No No No No YeS Gas Ar Ar Ar Ar Ar Ar Ar Gas flow, s.c.f.h. 20 20 20 20 20 20 20 * SBA = Single-beam atomisation. Balance 10 3 0.01, 100 & 0.1 and 1000 & 1 mg) was used. A Cahn DTL electronic semimicro balance with three measuring ranges (maximum Calculator Calculations were made on an Expanded Texas SR60 with alphanumeric display, 1920 programming steps and 100 storage memories.The calculator is simply an aid to the operator for instructions, calculations, etc., during routine analytical operations and is not connected to the balance or the instrument. Recorder A Houston Instruments Omniscribe, Model 51 11-2, single-channel recorder was employed. Investigations Line selection Most of the analytical wavelengths indicated in the manuals published by manufacturers of atomic-absorption spectrometers22 are too sensitive for a practically serviceable range of contents (see Table 11) to be obtained by analysing a solid metal sample in a graphite furnace. When determining trace elements via dissolution, there is always the possibility of diluting until one reaches an optimum range of content. Solid samples can be assayed in smaller masses, but then weighing errors are liable to be a major source of error affecting the results.Wavelengths that are less sensitive by up to several powers of ten can therefore be used. These wavelengths must be of high intensity and should preferably be over 250 nm, so as to make background absorption as low as possible. Table I1 lists the wavelengths that have been used and that satisfy this condition. All of these wavelengths, except for lead 283.3nm, give linear calibration graphs within the content ranges specified in Table 11. Other wavelengths for lead are 261.4 and 363.9 nm; experiments indicated an interference from zinc on the former and the latter is not sensitive enough (non-resonance line).1020 BACKMAN AND KARLSSON : DETERMINATION OF Pb, Bi, Zn, Analyst, Vol.104 Selection of parameters for the graphite furnace Obviously, there is no need for drying or pre-ashing when determining trace elements directly in solid samples, and so steps 1 4 for controlling the furnace heating are not used (see Table I1 and Apparatus). The first atomisation step (step 5 ) is used to set the power in such a way that the furnace is heated exactly to the point where the trace element begins to vaporise. This applies particularly to bismuth (0.03-3 pg g-l) and lead (1-150 pg g-1). The power for the second atomisation step (step 6) is then set high enough for the trace element to be completely evaporated within the, integration time selected, but low enough to avoid background radiation and background absorption.“Clean” can be used as an extra atomisation step to raise temperatures during the final seconds of integration, as well as for cleaning purposes. Fig. 1 shows examples of evaporation processes for bismuth with the second atomisation step set to various power ratings. If the power is too low (A, 20% of maximum output), bismuth will not be evaporated within the integration time selected. If, on the other hand, the final temperature is too high (C, lOOyo of maximum power), back- ground absorbance will increase, thus impairing the peak to background ratio. 2 0) c ; 1 2 51 0 J 1 1 I I t I , I D E F D E F D E F Time ___+ Fig. 1. Distillation curves for 5.0ng of bismuth a t various power ratings on last atomisation step: A = 20% of maximum effect; B = 70% of maximum effect; C = 100% of maximum effect; D = last atomisation step and integration starts; E = last atomisation stops; and F = integration stops.Wavelength, 306.8 nm. Shape and mass of samples The Instrumentation Laboratory 251 instrument can be used to measure absorbance as either peak height or peak area. When the atomisation process was recorded via a recorder, an irregular shape was obtained for certain elements. Lead in particular tended to have two peaks, the relative sizes of which varied according to the heating rate. This would lead to poor reproducibility if peak-height measurement were used. Preliminary experi- ments showed that peak-height measurement gave a deviation roughly three times greater than that given by peak-area measurement.L ~ n d b e r g ~ ~ and Marks et al.20 reported the same finding. Peak-area measurement was therefore employed in all of the experiments described below. However, the simplified type of peak-area measurement performed by this instrument gives not the area beneath the graph but the mean absorbance, irrespective of integrationNovember, 1979 1021 time. In practice, this means that a signal received within 4 s, given a 4-s integration time, will give four times the absorbance for the same measurement over a 16-s period. Let us consider how this type of integration works if a trace element is evaporated rapidly, i.e., gives high, narrow absorbance peaks, and consider whether one would then obtain a higher mean absorbance than one would if the same trace element were evaporated more slowly within an integration period of the same duration.To investigate this point, chips of two kinds were prepared from the same samples. Type 1 took the form of balls, with a small area to mass ratio, obtained from an end mill with a high feed rate, while type 2, which had a large area to mass ratio, took the form of thin flakes obtained from a cylinder cutter with a slow feed. Given identical masses and otherwise identical analytical conditions, the absorbance obtained for type 2 was about 20% higher than for type 1 samples. Comparing this result with that for drilled chips, which are the type most commonly occurring at Sandvik AB, we found the absorbance of type 2 to be about 15% higher. Fig. 2 shows -absorbance as a function of mass and type of chip in lead analysis with a wavelength of 283.3nm.No such pronounced difference was observed with the other trace elements, deviations in these instances being within the range of normal errors of measurement. Ag and Sb IN STEEL AND NICKEL-BASE ALLOYS BY AAS 0.3 0, m : 0.2 2 a 2 0.1 0 I 1 5 10 Amount of leadlng I Fig. 2. Absorbance as a function of the amount of lead and various types of drillings. 1 ( O ) , Thin flakes; 2 (A), varying volumes of lead nitrate solution; 3 (a), drilled chips; and 4 ( O), balls. Resonance line 283.3 nm. As the results were influenced by the shape of the chips, it was suspected that the mass of the samples might also affect the determination of a trace element. To test this hypothesis, absorbance was measured as a function of the absolute amount of trace elements. Several standards with various contents of trace elements were used in order to obtain various masses.Figs. 3 and 4 show that absorbance is a linear function of the absolute amount of the element measured and that it is independent of the mass of the sample in the area examined. Matrix efects The elements that may be presumed to disrupt atomisation are those which are vaporised during atomisation or those which bind the trace elements strongly in chemical compounds. Chemical compounds of this kind with high boiling-points include sulphides and oxides.1022 BACKMAN AND KARLSSON: DETERMINATION OF Pb, Bi, Zn, Analyst, VoZ. 104 0.4 al C $ n" 0.2 4 0 Amount of trace element/pg Fig. 3. Relationship between absorbance and sample size: a, 5mg; A, 10mg; and 0, 15mg.A = Pb 363.9nm; B = Zn 307.6nm; and C = Sb 269.8 nm. Table I11 shows the melting-points of sulphides and oxides of lead, bismuth, antimony, zinc and silver. I t can be seen that lead forms a relatively non-volatile sulphide, but non- sulphur alloyed steel contains little sulphur, which means low sulphide formation. On the other hand, when steel contains more than 0.1% of sulphur, as in sulphur-alloyed free-cutting steels, a 40% reduction in the lead signal is obtained at 283.3 nm. This interference may be due to sulphide formation, because it is halved by the addition of about 10 mg of copper to the graphite furnace. The interference is only perceptible at lead contents exceeding 7 pg gl. The oxides formed with all trace elements except zinc are readily volatile.Zinc oxide does not melt until 1975 "C, but it then decomposes readily at about 2700 "C. Zinc has not been observed to be prone to oxide interference. 0.4 W C -e 51 n 0.2 0 5 10 15 Amount of trace element/pg Fig. 4. Relationship between absorbance and sample size: @, 1 mg; A, 2 mg; and 0, 3 mg. A = Ag 338.3 nm; B = Pb 283.3 nm; and C = Bi 306.8 nm. Manganese is the only common steel alloying element that can be vaporised easily. Figs. 5 and 6 show recorder traces for BCS 336, showing the point at which manganese vapour is formed in relation to lead and bismuth. It can be seen that the manganese vapour is formed before either lead or bismuth is vaporised, but it does not affect the signal for the trace element.November, 1979 Ag and Sb IN STEEL AND NICKEL-BASE ALLOYS BY AAS TABLE I11 MELTING-POINTS OF SULPHIDES AND OXIDES OF LEAD, BISMUTH, ZINC, ANTIMONY AND SILVER r 1 Element Sulphide Oxide Lead .... .. 1114 888 Bismuth .. . . 685 (d) 860 Antimony . . . . 550 666 Silver . . .. .. 825 300 (d) Zinc . . .. .. 1185 (s) 1975 *d = Disintegrates; s = sublimes. 1023 This has also been confirmed by analysis of four 18Cr/8Ni/2.5Mo doped steels with manganese contents varying between 0.006 and 1.76%. Lead was determined using the method described here and also using three other wet-chemical methods at various labora- tories. As can be seen from Table IV, no significant difference was obtained by this method compared with the other methods. Andrews and Headridges have also concluded that manganese does not cause interference.Time ___* Fig. 5. Recorder trace for BCS 336, showing that the manganese vapour has almost reached its maximum when lead starts to vaporise. Settings according to Table 11. Mn = 279.5 nm. Time ___* Fig. 6. Recorder trace for BCS 336, showing that the manganese vapour has almost reached its maximum when bismuth starts to vaporise. Settings according to Table 11. Mn = 279.6 nm. For the investigation of other matrix effects, calibration graphs were plotted for low-alloy and alloyed steel and for nickel-base alloys with or without molybdenum, manganese, copper, etc. In Fig. 7, absorbance per milligram at a sample amount of 2 mg is plotted as a function of the lead content of a BOCr/lONi steel, a 20Cr/70Ni nickel-base alloy and a low-alloy steel.The points coincide well along a straight line, which means that wide variations in the concentration of the main elements of the matrix have little or no effect on the result of the analysis. Calibration elements by direct atomisation of solid samples. the graphite boat, followed by drying, ashing and atomisation. against the absolute amount of trace elements. Three methods can be used in plotting a calibration graph for the determination of trace (a) The addition of a particular amount of trace elements from an aqueous solution in Absorbance is plotted The advantage of this method is that it1024 BACKMAN AND KARLSSON : DETERMINATION OF Pb, Bi, Zn, Analyst, VoZ. 104 TABLE IV EFFECT OF MANGANESE ON LEAD DETERMINATION Lead contentlpg g-1 r A -7 Method Sample Mn, yo described A* Bt C$ 426-8 .. . . 0.006 53 57 53 63 426-9 . . . . 0.307 90 93 92 91 426-10 . . . . 0.735 59 64 60 60 426-11 . . . . 1.762 60 61 67 67 * Method as in reference 12. t Extraction of PbI with isobutyl methyl ketone and atomic- absorption spectrometric determination in organic phase. Direct flame atomic-absorption spectrometry. gives a calibration that is independent of standards and, therefore, of other analytical methods, but the atomisation process is not the same as for solid samples. Because atomi- sation proceeds more rapidly, the instrument's mode of peak-area measurement causes it to show an absorbance roughly 10% greater than would be obtained with the same absolute amount of trace elements atomised from a solid drilled sample (see Fig.2). Andrews and Headridge8 also obtained dissimilar curves for the atomisation of solid samples and aqueous solution. (b) The second method is to employ a standard with a known trace-element content. Here too, absorbance is plotted against the absolute amount of trace elements. Atomisation takes place in the same manner as for a sample. The drawback is that not many standards have analytical values for trace elements that are reliable. The third method is to use various amounts of several standards and to plot absorbance against mass of each standard taken. From the graphs obtained are derived a second series of calibrations correlating absorbance against concentration of element sought, expressed in micrograms per gram. Each graph in this second series corresponds to a pre-selected sample mass (see Fig.8). The advantage of this method is that calibration is based on several standards and that one can expect a certain equalisation of errors, together with the exposure of errors of any considerable magnitude. (c) k 8 5 0.1 - e 8 a n 0 5 15 Lead concentrationlpg g-' Fig. 7. Calibration graphs for various types of alloys: 0, 20CrllONi; 0, 20Cr/ 70Ni; 'and A, low-alloy steel. The third method was therefore used in plotting all graphs. For each element we used five to ten standards analysed by ourselves and at other laboratories (see Table V). In the daily calibration, only one standard needs to be analysed to check and correct for sensitivity changes (determination of sensitivity factor, KO).November, I979 Ag AND Sb IN STEEL AND NICKEL-BASE ALLOYS BY AAS 1025 g 0.2 f n Q 0.1 S 0.2 c e 5( a 0.1 0 1 2 3 4 5 6 7 0 1 2 3 4 5 6 7 Mass of samplehg Concentration of bismuth/pg g-’ Fig.8. contents. from (a) by plotting graphs for a constant mass. (a) Absorbance V ~ Y S U S mass calibration graphs for samples with various bismuth A, 5.4 pg g-l; B 3.2 pg g-l; C, 1.5 pg g-l; and D, 0.47 pgg-l. (b) Constructed A, 1 mg; B, 3 mg; and C, 5 mg. Calculation of results The trace-element content of the sample was calculated by means of the equation (A - B)Ko K1m + K , Amount of trace elements (pgg-l) = where A = absorption, B = background absorption, KO = relative sensitivity constant, m = mass of sample in milligrams, K , = the slope of the l-mg curve [see Fig. 8 ( b ) ] and K , = the intercept.Procedwe Weigh the cali- bration standard or the sample within the mass interval given for each element in Table I1 to within 0.05mg. Transfer the chip or chips from the scale into the graphite boat and then place the boat in the graphite furnace with the aid of a non-magnetic boat holder. Push the starter button. Feed the mass of the sample into the calculator together with the absorbance obtained after about 30 s, so as to obtain the content. To avoid incorrect results due, for example, to contamination or heterogeneous sample material, determina- tions must be made at least in duplicate. The stability of the instrument is sufficient for at least four duplicate determinations of silver, antimony and lead to be performed between Set the parameters for the element to be measured as in Table 11.TABLE V STANDARDS USED WHEN PLOTTING CALIBRATION GRAPHS25 Standard JK8E .. .. JK8F .. .. BCS 336 .. .. BCS 337 . . .. BCS 336 .. .. SDN 02* . . .. SDN 03* . . .. BCO 31H .. .. Hfs SS24* . . .. BCS 326 .. .. BCS 329 .. .. NBS 362 .. .. Hfs 36Q* .. .. Grade 17Cr/llNi/3Mo 17Cr/llNi/3Mo 18Cr/9Ni 18Crll ONi 18Cr/9Ni 17Cr/13Ni/3Mo 17Cr/13Ni/3Mo lOCr/ lNi/lMo 20Cr/26Ni/4Mo 17Cr/l2Ni/3Mo 0.4Cu Mild steel Mild steel Pb Pb 283.3 nm 363.9 nm 2.2 2.2 4.3 4.3 7.2 7.2 10.8 - 14.6 14.6 - 47 - 106 - 129 - - Bi 306.8 nm 0.12 0.47 3.2 - - 1.8 0.81 Zn 307.6 nm 11 43 64 112 30 - - Sb 269.8 nm 19 6 28 - - 40 62 180 130 - * Internal standards.1026 BACKMAN AND KARLSSON: DETERMINATION OF Pb, Bi, Zn, AnaZyst, VoZ. 104 calibrations.The graphite boat is changed for each re-calibration except with antimony, for which a new boat is used for every determination. Nine duplicate determinations are performed for bismuth and zinc. Results The detection limits, stated as concentration at a signal equal to three times the back- ground, are given in Table VI, together with the average relative standard deviation (2s) for the data shown in Tables VII-XI. Many standards were analysed in order to determine the lead, bismuth, antimony, zinc and silver contents. Tables VII-XI show the mean values obtained, the number of deter- minations (n) , relative standard deviations (2s) and comparative values. TABLE V I STANDARD DEVIATIONS AND LIMITS OF DETECTION Wavelength/ Element nm Lead .. .. .. 283.3 363.9 Bismuth .. .. 306.8 Antimony . . .. 269.8 Zinc . . .. .. 307.6 Silver .. .. 383.3 Limit of detection/ Relative standard deviation (24, yo 0.02 20 12 0.03 12 6 12 1 10 0.01 10 CLg f2-1 - Discussion Although the calibration graphs are plotted on the basis of several standards, only one is normally used for the daily calibration check. This check is carried out by weighing chips within the mass range indicated in Table 11. The mass and the absorbance obtained are fed into the calculator, which then adjusts the calibration graphs (factor KO). BCS 336 proved to be very useful in calibration checks for lead (7.2 pg g-l), bismuth (3.2 pg gl), silver (0.70 pg g-l) and zinc (64 pg gl). BCS 329 is used for calibration checks on antimony (0.018%). The contents were determined by ourselves and also on a joint basis under the aegis of the Swedish Ironmasters' Association.2s TABLE VII RESULTS OF LEAD DETERMINATIONS Sample JKBE.... JK8F .. BCS 331 . . 333 . . 334 . . 336 , . 336 . . 337 .. BCO 30H .. 33H . . 34H .. 36H . . 40H . . 41H . . 42H . . 43H . . D6-911 . . D6-912 . . D6-913 .. NBS 361 . . .. .. .. .. .. .. .. . . .. .. .. .. .. .. .. .. .. .. .. .. Mean result/ 1.4 4.3 6.4 6.0 11.3 16.3 7.4 11.2 CLg 8-1 113 119 60 7 26 47 144 132 11 3.4 4.0 0.24 n 10 12 10 12 10 18 12 10 8 6 6 6 6 6 R 6 6 6 6 2 Relative standard deviation (24, yo 0.2 0.6 0.6 0.6 1.2 2.0 0.7 1.1 19 24 7 1 2 6 9 10 1.0 0.3 0.9 - Comparative valueslpg g-l 2' 4.4* 6.3. 6.8' 11t 15t 7 t 12t 1211 103* 63' 26* 42* 133. 1171 7.4* 10.7* 2.21 4.0* 0.2# * Collaborative results obtained by at least two laboratories according to Burke.' t Certificate value.Reference 26.November. 1979 Ag and Sb IN STEEL AND NICKEL-BASE ALLOYS BY AAS 1027 Sample JK8E . . J K8F BCS 320 325 330 33 1 332 333 334 335 336 337 NBS 361 363 365 BCO 32H 44H D5-911 D6-912 D6-913 BCS 342 45 1 .. .. .. .. .. .. . . .. .. .. .. .. .. .. . . .. .. .. .. .. .. .. TABLE VIII RESULTS OF BISMUTH DETERMINATION .. .. .. .. .. .. .. .. .. .. . . .. . . .. . . . . .. .. . . . . .. .. Mean result/ Elg g-l n 0.12 6 0.47 12 0.52 2 0.37 2 2.9 2 0.05 6 0.09 6 0.07 7 0.06 8 0.06 8 3.3 10 0.05 8 4.8 8 5.2 7 < 0.03 2 0.12 6 0.15 6 0.64 4 1.21 4 0.33 4 0.74 2 0.26 2 Relative standard deviation (2s). yo 0.03 0.04 - - - 0.01 0.02 0.01 0.01 0.01 0.04 0.01 0.6 0.7 0.02 0.02 0.08 0.12 0.10 - - - Comparative valueslpg g1 .. 0.56* 0.36* 3.0* 0.055* 0.088* 0.060* 0.047* 0.062* 3.4* 0.036* 4 t 6.41 < 0.053 - - 0.58t 1.18t 0.31t 0.6* 0.20* * Values obtained by Andrews and Headridge.6 t Certificate value . # Reference 26 . The advantages of direct atomisation of solid samples are as follows: (a) (b) The sample does not have to be dissolved . This shortens the time required for Drying and pre-ashing are not necessary. which means a further reduction in the analysis and eliminates the risk of contamination from chemicals . time needed for analysis . TABLE IX RESULTS OF ZINC DETERMINATION Sample JK8E .. JK8F .. BCS 331 . . 332 .. 333 .. 334 .. 335 .. 336 . . 337 . . BCO 30H . . 31H .. 32H . . 33H . . 3433 . . 35H . . 40H . . 41H .. 42H ..43H .. 4433 .. .. .. .. .. .. .. .. .. .. .. .. . . .. .. .. .. .. .. .. .. Mean result/ Pg g-' n 12 10 45 6 52 8 96 8 108 8 116 14 112 14 64 20 56 14 7 6 6 8 10 6 8 6 8 6 7 8 7 7 10 6 20 8 14 8 22 8 Relative standard deviation (2s). % 4 4 4 10 8 11 11 6 8 1 3 1 2 3 2 1 2 5 3 3 Comparative valueslpg g-l 12* 45* 64* 98* 115* 121* 110' 64* 61* 6.6*; 6.5t 5.4*; 5.2t 10.8*; 10.6t 8.0*; 7.2t 7.6*: 8.0t 5.8*; 7.4t 5.9* 11.2*; 12.0t 20.4*; 20.7t 12.3*; 13.9t 21.4*; 22.5t * Direct determination by flame atomic-absorption spectrometry . t Extraction of Zn . SCN complex with isobutyl methyl ketone followed by atomic- absorption spectrometry .1028 Analyst, vol. 104 (c) The procedure for inserting the sample in the graphite furnace is simple, which means that operators can be trained very quickly.( d ) Very few matrix effects occur in the materials investigated. (e) High sensitivity. (f) Relatively uncritical atomisation temperature. BACKMAN AND KARLSSON: DETERMINATION OF Pb, Bi, Zn, Sample J K8E .. JK8F .. BCS 326 . . 329 . . 331 .. 332 .. 333 .. 334 .. 336 .. 336 .. 337 .. TABLE X RESULTS OF ANTIMONY DETERMINATION .. .. .. .. .. 0 . .. .. .. .. .. Mean result/ Clg K' n 21 4 4 4 61 4 176 8 46 4 36 4 43 4 22 4 36 4 34 4 31 4 Relative standard deviation (24, yo 4 1 8 20 6 5 6 3 6 6 6 Comparative valueslpg g1 20. 6t 60* 150t 42. 37* 37* 21* 31* 30* 30* * Photometric determination using Rhodamine B (collaborative). t Certificate value. Of course, the method also has some drawbacks: (a) Small sample amounts, so that the results are liable to be unrepresentative if the ( b ) The method is a relative one, which means that any errors in the standards used for (c) The dynamic range is small. The relative precision of the method, given a 95% confidence level, is 10-12%. material is not sufficiently homogeneous.calibration are fed into the results of the analysis. Long serial analyses have not displayed any tendency to impair precision, because calibration checks are carried out frequently and the graphite boat is replaced at short intervals. When, however, the sample amounts involved are as small as is the case here, there is always a risk of the result being made incorrect by heterogeneity. To reduce this risk, each element TABLE XI RESULTS OF SILVER DETERMINATION Sample BCS 331 .. 332 . . 333 .. 334 .. 336 . . 336 .. 337 .. BCO 30H . . 34H .. 35H .. 40H .. 42H . . NBS 348 . . NBS 65e . . 168 .. 349 . . 363 .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. Mean result/ 0.26 0.31 0.21 0.34 0.47 0.70 0.28 1.19 0.65 0.26 1.30 0.80 1.01 0.23 0.90 0.18 0.03 Pg g-l n 4 6 6 6 6 6 6 6 4 4 6 4 6 6 6 6 2 Relative standard deviation (24, yo 0.04 0.04 0.04 0.06 0.06 0.04 0.03 0.11 0.06 0.04 0.12 0.08 0.09 0.02 0.06 0.02 - Comparative valueslpg g-1 0.26* 0.29* 0.26* 0.32* 0.44* 0.70* 0.29* -- - 0.23* 1.37* 0.8t 0.3t 0.9t 0.2t 0.02# - * Determined jointly according to reference 4. t Reference 4. $ Reference 26.November, 1979 Ag AND Sb IN STEEL AND NICKEL-BASE ALLOYS BY AAS 1029 is determined at least twice. During the time the technique has been in routine use, clear non-homogeneity has been established in only about ten out of a total of about some 8000 samples.27 This non-homogeneity would probably not have been noticed if a different analytical procedure involving the weighing in of larger amounts had been used.It follows that this method can also be used for studies of heterogeneity. Data referring to the correctness of the method are given in Tables VII-XI. The results agree, within the range of measuring accuracy, with the results reported in the literature and with those obtained by other methods. With the furnaces that are now commercially available17 it is possible to obtain final temperatures of up to more than 3000 “C. Given this final temperature, elements with boiling-points below 2250 “C can be vaporised with sufficient rapidity.For example, silver, which has a boiling-point of 2212 “C, can be analysed with no memory effect, whereas tin, with a boiling-point of 2270°C, cannot be analysed without a memory effect. Table I lists elements with boiling-points below or around 2250 “C. Many of these elements cannot be analysed, owing to nonspecific absorbances, atmospheric impurities, analytical wave- length with excessive or insufficient sensitivity, etc. Standards or reference methods of sufficient sensitivity are lacking for many of the elements in Table I that are of interest to the steel industry, e.g. , cadmium, indium, selenium, tellurium and thallium. Apart from the materials mentioned above, minerals, ores, ferro-alloys, misch metal, etc., can also be analysed using this technique. However, the problem is that these types of material are not sufficiently homogeneous for representative samples to be obtainable by simple means. The work involved in determining trace elements in these materials is there- fore mainly connected with sampling and the preparation of samples. The authors extend their thanks to Sandvik AB for permission to publish this paper and to their colleagues for suggestions and comments. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. References Mayer, G., and Clark, C. A., Metallurgist Muter. Technol., 1974, 6, 11. Ljundstrom, L. G., Scand. J. Metall., 1977, 6, 676. Bohnstedt, U., Dt. Edelstahl Werke, Technische Berichte 11, 1971, 11, 101. Burke, K. E., Talanta, 1974, 21, 417. Danielsson, L., and Englund, B., Jernkontorets Forskning, Serie D 110, Forskningsuppgift 458/70, Thornton, K., Analyst, 1969, 94, 958. Langmyhr, F. J., and Rasmussen, S., Analytica Chim. Acta, 1974, 72, 79. Andrews, D. G., and Headridge, J . B., Analyst, 1977, 102, 436. Aziz-Alrahman, A. M., and Headridge, J . B., Talanta, 1978, 25, 413. Andrews, D. G., Aziz-Alrahman, A. M., and Headridge, J . B., Analyst, 1978, 103, 909. Headridge, J. B., and Thompson, R., Analytica Chim. Acta, 1978, 102, 33. Frech, W., Analytica Chim. Acta, 1975, 77, 43. L’Vov, B. V., Spectrochim. Acta, 1961, 17, 761. Massmann, H., Spectrochim. Acta, 1968, 23B, 215. Langmyhr, F. J . , and Thomassen, Y., 2. Analyt. Chem., 1973, 264, 122. Langmyhr, F. J . , Stubergh, J. R., Thomassen, Y., Hanssen, J . E., and Doleial, J., Analyti~a Chim. Headridge, J . B., and Smith, P. R., Talanta, 1971, 18, 247. Headridge, J. B., and Smith, P. R., Talanta, 1972, 19, 833. Lundberg. E., and Frech, W., Analytica Chim. Acta, 1979, 104, 75. Marks, J. Y., Welcher, G. G., and Spellman, R. J., Appl. Spectrosc., 1977, 31, 9. L’Vov, B. V., Spectrochim. Acta, 1978, 33B, 153. “Atomic Absorption Methods Manual,” Volumes 1 and 2, Instrumentation Laboratory, Wilmington, Lundberg, E., Appl. Spectrosc., 1978, 32, 276. Weast, R. C., Editor, “Handbook of Chemistry and Physics,” Fiftieth Edition, Chemical Rubber Palvannne, V., “Framtagning och Analysering av Referensprov,” Delrapport for Forskningsuppgift, Dulski, T. R., and Bixler, R. R., Analytica Chim. Acta, 1977, 91, 199. Lundberg, E., and Frech, W., Analytica Chim. Ada, 1979, 104, 67. J erkontoret, Stockholm, 1975. Acta, 1974, 71, 35 Mass., 1975 and 1976. Co., Cleveland, Ohio, 1969. JK 426/76, Jerkontoret, Stockholm, 1979. Received March 27th, 1979 Accepted May 24th, 1979

ISSN:0003-2654    

DOI:10.1039/AN9790401017

出版商:RSC    

年代:1979

数据来源: RSC

摘要:

1030 Analyst, November, 1979, Vol. 104,pp. 1030-1036 Extraction of Nanogram Amounts of Cadmium and Other Metals from Aqueous Solution Using Hexamethyleneammonium Hexamethylenedithio- carbamate as the Chelating Agent A. Dornemann and H. Kleist Bayer AG, Werk Uerdingen, A C-F Untersuchungslaboratorium, D 4150 Krefeld 11, FHG A group of metals can be extracted from aqueous solution by using hexa- methyleneammonium hexamethylenedithiocarbamate as the chelating agent and a mixture of 2,4-dimethylpentan-3-one and xylene as the organic phase. A description is given of a procedure for the determination of microgram and nanogram amounts of the nine metals silver, bismuth, cadmium, copper, nickel, lead, thallium and zinc in aqueous solution. The influence of a high content of iron and copper on the extraction is described.Keywords : A tomic-absorption spectrometry ; metal extraction ; nanogram avnozGnts of metals ; dithiocarbarnate chelation Liquid - liquid extraction is the preferred process for the enrichment of trace amounts of heavy metals from aqueous solution for subsequent determination by means of atomic- absorption spectrometry. In view of the high selectivity of atomic absorption the main emphasis is on group extraction processes. The organic extract should be suitable for both electrothermal atomisation and atomisation in a flame. With an extract of this kind, in which the individual concentrations of metals generally vary a great deal, it is possible to take the better suited atomisation technique for the determination of the particular met a1 under consideration.Firstly, a group of nine metals, silver, bismuth, cadmium, cobalt, copper, nickel, lead, thallium and zinc, were to be extracted simultaneously, and secondly, cadmium contents in the nanogram per gram range were to be determined in the presence of a large excess of iron. It was found that the extraction processes described in the literature as a preliminary stage to atomic-absorption spectrometry1,2 were not suitable for the two purposes referred to above. The most widely used method involves the use of ammonium tetramethylene dithiocarbamate (also known as ammonium pyrrolidine dithiocarbamate, APDC) as the complexing agent, and 4-methylpentan-2-one (methyl isobutyl ketone, MIBK) as the organic phase. Although MIBK has good combustion properties in the flame, it can lead to problems in electrothermal atomisation. A further disadvantage of MIBK is its high solubility in water.Moreover, the cadmium - APDC complex is insufficiently stable in #an acidic medium. This paper describes some investigations that led to the use of hexamethyleneammonium hexamethylenedithiocarbamate (HMA IIMDC) as the chelating agent and of a mixture of xylene and 2,4-dimethylpentan-3-one (diisopropyl ketone, DIPK) as the organic phase. Further experiments clarified the behaviour of the recommended extraction system towards an excess of iron and copper. This study of the extractability of heavy metals had a dual purpose. Experimental Reagents distilled before use. The inorganic reagents used were of the highest available purity; organic solvents were 4-MethyJfientan-2-one (methyl isobutyl ketone, MIBK) . This compound boils at 117-1 18 "C.2,4-Dimethylpentan-3-one (diiso@ro$yl ketone, DIPK) . This compound boils at 124-125 "C. Xylene. Hexamethyleneammonium hexamethylenedithiocarbamate (HMA HMDC) . To a solution This solvent is a mixture of the various isomers.DORNEMANN AND KLEIST 1031 of 224 ml of distilled hexamethyleneimine (boiling-point 136-138 "C) in 300 ml of xylene, which is being cooled in an ice-bath, add, within 30 min and with constant stirring and cooling, 60 ml of distilled carbon disulphide (boiling-point 46.2 "C). Collect the white crystalline precipitate on a funnel, wash it three times with diethyl ether and then dry it between filter-papers. Caution-Hexamethyleneimine is a severe poison and appropriate precautions should be taken.The preparation of each special reagent used in the procedure is described in the appro- priate section. Apparatus burner for the air - acetylene mixture and a deuterium background compensator. Perkin-Elmer atomic-absorption spectrometer. Perkin-Elmer graphite furnace, Model HGA500. The Model 420 is used, with a three-slot Results A series of experiments was performed to find the best conditions for the pre-concentration procedure, investigating the chelating agent, the organic solvent and organic metal standard solutions, which should enable a check of the completeness of extraction. The influence of higher concentrations of total extractable metals was also investigated. The results of these experiments are reported below.Chelating Agent The chelating agent should permit complete extraction over the widest possible ranges of pH and concentration. Attempts to extract cadmium at the 100 pg 1-1 level, and at a pH of 2.8, with APDC - MIBK gave poor results. The conditions of the extraction are: sample volume, 100 ml, containing 10 pg of cadmium; addition of 2.5 ml of APDC solution (1 g of APDC in 100 ml of water); pH adjusted to 2.8 after the addition of APDC; addition of 10 ml of MIBK before extraction. As the mixing time of the organic and aqueous phases was increased, the extraction yield diminished.3 After shaking for 1 min, about 75% of the cadmium was recovered; after shaking for 10 min, only about 8% was recovered. Complete extraction is possible only for concentrations of cadmium below 10 pg l-1.4 The rate of decomposition of the initially formed cadmium - APDC chelate is influenced by the pH, the total salt content of the aqueous phase, the material of the vessel and the exposure to light.Complete extraction of cadmium at the 100 pg 1-1 level is possible at pH 4, but these higher pH values restrict the possibilities of selective isolation of cadmium. Another derivative of dithiocarbamic acid, hexamethyleneammonium hexamethylene- dithiocarbamate (HMA HMDC),"s proved to be more suitable for the extraction of cadmium from an acidic medium. Extraction at pH 2, including shaking for up to lOmin, resulted in a complete cadmium yield. Twenty micrograms of cadmium can be completely extracted from 100 ml of an aqueous solution of sodium chloride (10 g 1-1) with 5 mg of HMA HMDC and 10ml of MIBK in the pH range 1-10, the pH in the aqueous phase being measured after the addition of the HMA HMDC solution.Extraction from acidic solution is of particular importance. Organic Phase Volume stability In order to be able to check the process of enrichment, a clearly defined enrichment factor is desirable. This requires that the added volume of the organic phase remains stable during the course of the extraction. In this respect MIBK as the organic phase has a number of dis- advantages. Its solubility in water at room temperature is about 2% V / V , which is relatively high.9 When 500 ml of water were shaken with 10 ml of MIBK, the organic phase disappeared almost completely.The solubility is greatly dependent on the total salt content of the aqueous solution (Table I). As a result, additional variables are introduced into the analytical procedure, and this must be taken into consideration when the enrichment factor exceeds 5. 4-MethylPentan-2-one.1032 DORNEMANN AND KLEIST : DITHIOCARBAMATE EXTRACTION OF NANOGRAM Analyst, VoZ. 104 TABLE I VOLUME CHANGE OF ORGANIC PHASE Aqueous phase 100m1, pH 7 . . . . .. 100 ml HC1, pH 2 .. .. 600 ml H,O, pH 7 . . . . 100 ml H,O, pH 7 . . .. 100 ml NaCl solution, 4 g NaCl . . 100 ml NaCl solution, 10 g NaCl 100 ml HCl, pH 2 100 ml NaCl solution, 4 g 'NaCl ' 100 ml NaCl solution, 10 g NaCl 100 ml H,O, pH 7 . . .. 600 ml H,O, pH 7 . . .. Organic solvent taken MIBK MIBK MIBK MIBK MIBK DIPK DIPK DIPK DIPK 70% DIPK- 30% xylene 30% xylene 70% DIPK - Volume of organic phase added/ml 10 10 10 10 10 10 10 10 10 10 10 Volume of organic phase after equilibration/ ml 8.0 8.0 8.6 8.8 t0.5 9.2 9.2 9.4 9.6 >9.8 >9 Decrease in volume of organic phase, % 20 20 14 12 > 95 8 8 6 4 <2 < 10 2,4-DimethyZpentan-3-one.In comparison with MIBK, DIPK represents a considerable improvement. Its solubility in water at room temperature is only 0.5% V/V.99 Vari- ations in the total salt content of the solution have only a slight effect on solubility (Table I). A further improvement in volume stability can be achieved with a mixture of 70% V/V of DIPK and 30% V/V of xylene. When 100ml of water were shaken with 10ml of this solvent mixture no change in the volume of the organic phase could be detected with the usual measuring vessels used for analytical work.When 500ml of water were shaken with 10ml of this mixture the reduction in volume of the organic phase was less than 1 ml, this volume decrease being partially caused by the fact that small droplets of the organic phase adhered to the walls of the vessel, or remained suspended in the aqueous phase. Thus, the solvent mixture DIPK - xylene permits enrichment factors of up to 50; no additional steps are necessary, and data on the salt content of the aqueous phase are not required. Mixture of 70% 2,4-dimethylpentany3-one and 30% xylene (V/V). Sensitivity and behaviouur during atomisation Flame atomisation of a given concentration of metal in the three organic phases investi- gated resulted in the same absorbance.This absorbance is about 3.5 times greater than the absorbance of aqueous metal solutions with the same nebuliser - burner assembly. In electrothermal atomisation, the solvent mixture DIPK - xylene evaporates, leaving no residue. MIBK extracts, however, tend towards resinification and coke formation upon heating; this effect impairs the reproducibility of the atomisation process, particularly for elements, such as cadmium, which have low atomisation temperatures. Metal Standard Solutions in the Solvent Mixture Pureearation Standard solutions of metals made up in the extraction solvent mixture are necessary in order to check the yield of extract from synthetic solutions and to establish calibration graphs. Oil-soluble standards are available commercially for only some of the metals covered by the group-extraction procedure.ll Tests carried out with a commercial oil- soluble standard for cadmium (cyclohexanylbutyric acid cadmium salt) were unsatisfactory. The following process proved to be suitable for use with all metals covered by the group extraction procedure.The starting solution used is a standard solution in nitric acid of the metal, to which some citric acid and about a ten-fold volume of formic acid is added. To this mixture is added about the same volume of DIPK. The result is a clear, homo- geneous solution. Part of this solution is then diluted by a factor of about 1000 with the extraction solvent mixture containing HMA HMDC. In this way organic standard solu- tions are obtained, the compositions of which largely correspond to those of the solutionsNovember, 1979 AMOUNTS OF CADMIUM AND OTHER METALS FROM AQUEOUS SOLUTION obtained from the extraction procedure.Working procedure. 1033 Details of their preparation are given under Stability If kept in a laboratory without protection against the light, organic metal standard solutions will remain stable for at least 1 d. If kept in a refrigerator at 278 K, they are stable for several weeks. The metal-containing extracts obtained by means of the solvent mixture DIPK - xylene exhibit the same stability as the organic metal standard solutions when the extracts are pipetted off from the aqueous layer after phase separation. Influence of Extractable Metals Inplzence of iron In addition to the metals mentioned in the group extraction process, the following metals can also be wholly or partially extracted into the organic solvent mixture: arsenic, iron, gallium, germanium, indium, manganese, antimony and tin.The alkali metals, the alkaline-earth metals, aluminium, selenium, yttrium, the lanthanides, titanium, zirconium and hafnium remain in the aqueous phase. When examining samples with greater contents of extractable metals, an important question is up to what total content of extractable species can this method be applied? Details of the influence of iron on the group extraction are discussed elsewhere.12 The iron content of the aqueous solution can be as high as 25 mg 1-1 without having any detri- mental effect on the group extraction of the nine elements silver, bismuth, cadmium, cobalt, copper, nickel, lead, thallium and zinc.Interferences caused by a greater content of iron (up to 250mg1-l) can be excluded by adding ammonium fluoride (5 g1-l) or by pre- extraction of the iron from 2 N hydrochloric acid solution with N-nitroso-N-phenylhydroxyl- amine. I@uence of copper Copper was investigated as an example of a metal that forms particularly stable com- pounds with dithiocarbamates.13-15 Table I1 shows the results of experiments on the extraction of the elements covered in the group extraction process with aqueous solutions of increasing copper content. The extraction carried out under the conditions set out ,in the working procedure was repeated up to four times in order that the recovery of the metals should be as complete as possible, even at higher copper concentrations.Copper contents of up to 25 mg 1-1 had only a slightly detrimental effect on the extraction of the other metals. With copper contents of 50mg1-1 the first extract contained, in addition to part of the original copper, only some of the silver, cobalt and nickel used (see Table 11); the other metals were found in the subsequent extracts. This result is somewhat surprising in view of the fact that, for the stoicheiometric conversion of the metals with the chelating agent, the capacity of the extraction solution was not exhausted in the first extrac- tion. When the starting solution had a higher copper content, a constant amount of metal (about 40mg) was extracted per extraction step from the aqueous starting solution.The highest copper concentration of 150mg1-1 led to the deposition of a greater amount of precipitate, which adhered to the wall of the vessel and was not completely dissolved by the subsequent extractions. For the purposes of practical analysis, it follows from these experiments that the total concentration of metals to be extracted must not exceed 25 mg 1-1 if the metals are to be completely recovered in a single extraction. Procedure for the Determination of Nine Metals in Water7 Method The weakly acidic aqueous solution (acid concentration 0.5 M) is brought into a pH range of between 2 and 3 by adding a formate buffer solution. After chelating with HMA HMDC, the elements to be determined are extracted with a DIPK - xylene mixture.The metal content of the extract is measured by atomic-absorption spectrometry with an1034 DORNEMANN AND KLEIST : DITHIOCARBAMATE EXTRACTION OF NANOGRAM Analyst, vo,!. 104 TABLE I1 INFLUENCE OF COPPER ON THE EXTRACTION Metal content of initial aqueous solution, 25 pg 1-1 of each metal; each extraction performed according to the standard procedure. Copper concentration added/mg 1-1 0 5 12.5 25 37.5 60 100 150 Extraction run 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3 4 Recovery of heavy metals, yo I A 8 Ag Bi >95 >95 >95 >95 - - - - - - - - >95 >96 - - - - > 96 91 6 88 86 11 9 63 29 95 6 44 - 47 79 7 11 23 62 - 12 25 55 - - - - - - - - - co > 95 - - > 96 - - > 95 - - >96 >95 - - - - 61 35 30 61 5 20 20 20 31 - Ni > 95 - - > 95 - - > 95 - A > 95 > 95 - - - - 45 50 29 47 16 10 15 41 - - cu - - - > 96 1 > 96 1 >95 1 > 95 1 76 12 1 39 38 1 26 26 26 1 - - - - air - acetylene flame or an electrothermal atomiser.Metal standard solutions, prepared from an aqueous stock solution, in the DIPK - xylene extraction mixture make it possible to check the completeness of extraction from aqueous standard solutions. Reagents This solution contains 1000 mg 1-1 of silver, bismuth, cadmium, cobalt, copper, nickel, lead, thallium and zinc in water. Dissolve 1000 mg of each of the metals to be determined in nitric acid, sp. gr. 1.41, using gentle heat in order to effect com- plete dissolution. Transfer the solution into a 1-1 calibrated flask, and dilute to the mark with water.The final acid concentration of stock solution I should be about 0.1-0.5 M. Metal nitrates can also be used to prepare this stock solution, provided that their metal content is established by an independent means. If stock solution I contains silver, it must not contain chloride ions; solutions containing lead must be free from chloride and sulphate ions. Prepare aqueous standard solutions by dilution of aqueous stock solution I with 0.1 M nitric acid. This solution contains 50 mg 1-1 of silver, bismuth, cadmium, cobalt, copper, nickel, lead, thallium and zinc in organic solvent. Pipette 5 ml of aqueous stock solution I into a dry, 100-ml calibrated flask. Add 50 ml of formic acid and 0.25 g of citric acid monohydrate. Prepare organic metal standard solutions by dilution of organic stock solution I1 with the extraction solution.The use of dry, 25-ml calibrated flasks and plastic disposable pipettes (0.1-1.0 ml) is recommended. Organic metal standard solutions should be kept in a cool, dark place. Formate bufler solution. Dissolve 268 g of formic acid and 14 g of citric acid monohydrate in about 350 ml of water. Add slowly, with constant cooling and stirring, 243 g of sodium hydroxide. To this mixture add 50 mg of m-cresol purple and dilute the solution to 1 1 with water. Wash this solution twice with 50 ml of extraction solution in order to remove trace amounts of extractable metals. Aqueous stock solution I . Aqueous metal standard solutions. Organic stock solution I I . Adjust to the mark with 2,4-dimethylpentan-3-one.Organic metal standard solutions.November, 1979 AMOUNTS OF CADMIUM AND OTHER METALS FROM AQUEOUS SOLUTION 1.7 g of HMA HMDC in 75 ml of xylene, heating gently if necessary. to the mark with 2,4-dimethylpentan-3-one and keep it in a cool, dark place. HMA HMDC solution, 0.2 M in methanol. of HMA HMDC in methanol, heating gently if necessary. ture and adjust it to the mark with methanol. 1035 Dissolve, in a dry, 250-ml calibrated flask, Adjust the solution In a dry, 100-ml calibrated flask, dissolve 5.5 g Cool the solution to room tempera- Extraction solution, 0.025 M HMA HMDC. VeriJication of complete extraction Before analysing the actual samples, prepare at least three aqueous metal standard solutions, the concentrations of which correspond to the expected concentration ranges of the metals to be determined. Run these aqueous standard solutions through the method described under Working procedure.Then prepare organic standard solutions, the con- centrations of which correspond to the aqueous standard solutions prepared. Extraction is complete if the organic extracts of the aqueous standard solutions give the same absorbance values as the corresponding organic metal standard solutions. Working procedure When complete extraction has been verified, produce calibration graphs by measuring the absorbances of suitable organic standard solutions. In the procedure a 400-ml sample of water is treated with 20 ml of extraction solution, corresponding to an enrichment factor of 20. Other volumes, up to a ratio of the volume of the aqueous phase to the volume of the organic phase of 50: 1, can be used.Measure 400 ml of the water sample in a graduated cylinder and transfer into a 500-ml calibrated flask. Add 20ml of formate buffer solution; the colour of the indicator should be a pure yellow. If a red colour appears, add an additional 20 ml of formate buffer solution. Next add 2 ml of HMA HMDC solution in methanol and shake the flask vigorously. Wait for about 5 min, then add 20 ml of extraction solution and again shake the flask vigorously for at least 3 min. Wait for about 10 min in order to allow the layers to separate. Then carefully add water until the organic layer is completely in the neck of the flask. Adjust the absorbance reading of the atomic-absorption spectrometer to zero while aspirating the extraction solution.Next, aspirate the organic layers of each prepared sample, the organic layer of a blank treated in the same way as the samples and at least three organic standard solutions per element. In addition, aspirate mixed solvent (30% V/V xylene and 70% V/V DIPK) between each sample and between each standard in order to prevent clogging of the nebuliser. Finally, subtract the blank reading from the observed absorbance of each sample to obtain the true absorbance value of the sample. Calculation linear graph paper, of the absorbances of standards of the respective metal. Determine the concentration of each metal in the extract of each sample from plots, on Calculate the TABLE I11 EXTRACTION OF CADMIUM Aqueous solution I 1 Concentration/ Volume/ P.lg I-' ml 0.1 20 0.2 20 0.5 800 1 800 2.5 400 3 800 6 800 10 400 Extraction solution volume/ 1 1 16 16 20 16 16 20 ml Ratio of phases 20: 1 20: 1 50: 1 50: 1 20: 1 50: 1 50: 1 20: 1 Recovery, oh > 95 >95 > 95 > 95 > 95 > 95 > 95 > 95 Measurement technique* Graphite furnace Graphite furnace Flame Flame Flame Flame Flame Flame *Measurement : Perkin-Elmer, Model 420, with either graphite furnace HGASOO accessory, or an air - acetylene flame.1036 DORNEMANN AND KLEIST metal concentration of the samples by dividing the concentration values of the sample extracts by the enrichment factor of the extraction.Application Of the numerous check analyses that have been carried out, only those dealing with the extraction of small amounts of cadmium are reported.They are tabulated in Table 111. In all instances, recoveries were better than 95%. An inter-laboratory comparative test organised by DIN (German Institute for Standardisation) involved the determination of five metals in a sample of water. The starting solution was drinking water, which had been spiked with five heavy metals; ten laboratories participated in this test. The results are given in Table IV. Application of the Dixon testis for a 5% significance level led to TABLE IV INTER-LABORATORY COMPARATIVE TEST, DIN, OCTOBER 1978 Element Cd c o c u Ni Pb Number of laboratories . . . . 10 10 10 10 10 Outliers .. .. .. .. 1 Arithmetic mean of concentrations - 1 - - found/pg 1-1 . . * . .. . . 2.8 16 126 20 28 Absolutelpg 1-1 .. .. . . 0.19 1.6 10 1.7 3.5 Relative (variance), yo . . . . 6.7 9.9 7.9 8.6 12.4 Standard deviation the rejection as outliers of one cadmium value and one nickel value. these outliers the variance for lead was 12.4% and for the other metals below 10%. After elimination of 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. References Kirkbright, G. F., and Sargent, M., “Atomic Absorption and Fluorescence Spectroscopy,” Academic Dean, J. A., and Rains, T. C., Editors, “Flame Emission and Atomic Absorption Spectrometry,” Dornemann, A., Kleist, H., and Goergens, W.. 2. Analyt. Chem., 1977, 284, 97. Department of the Environment, “Cadmium in Potable Waters by Atomic Absorption Spectro- Thorn, G. D., and Ludwig, R. A., “The Dithiocarbamates and Related Compounds,” Elsevier Busev, A. I., Byrko, V. M., Tereschtschenko, A. P., Novikova, N. N., Naidina, V. P., and Terentev, Dornemann, A., and Kleist, H., 2. Analyt. Chem., 1978, 291, 349. Zolotov, Y . A., Analyst, 1978, 103, 56. Ginnings, P. A., Plonk, D., and Carter, E., J. Am. Chem. SOC., 1940, 62, 1923. Saylor, J. H., Baxt, V. J., and Gross, P. M., J . Am. Chem. SOG., 1942, 64, 2742. Dean, J . A., and Rains, T. C., in Dean, J. A., and Rains, T. C., Editors, “Flame Emission and Dornemann, A., and Kleist, H . , 2. Analyt. Cham., 1979, 295, 116. Wickbold, R., 2. Analyt. Chem., 1956, 152, 259. Eckert, G., 2. Analyt. Chem., 1957, 155, 23. Bode, H., and Tusche, K . J., 2. Analyt. Chem., 1957, 157, 414. Dixon, W. J., Ann. Math. Statist., 1951, 22, 68. Press, London, 1974, pp. 491-497. Volume 3, Marcel Dekker, New York, 1975, pp. 635-638. photometry,” Tentative method, HM Stationery Office, London, 1976. Publishing Company, Amsterdam, 1962. P. B., J . Analyt. Chem. USSR, 1970, 25, 665. Atomic Absorption Spectrometry,” Volume 2, Marcel Dekker, New York, 1971, pp. 327-339. Received March 2nd, 1979 Accepted June 4th. 1979

ISSN:0003-2654    

DOI:10.1039/AN9790401030

出版商:RSC    

年代:1979

数据来源: RSC

摘要:

Analyst, November, 1979, Vol. 104, pp. 1037-1049 1037 Determination of Microgram Amounts of Precious Metals Using X-ray Fluorescence Spectrometry Paul R. Oumo and Evert Nieboer Department of Chemistry, Laurentian University, Sudbury, Ontario, P3E 2C6, Canada Microgram amounts of noble metals were localised with ammonium sulphide on filter absorbent pads and in cellulose pellets for spectrometer counting. The Ka, lines of ruthenium, rhodium and palladium (tungsten tube) and the La, lines of osmium, iridium, platinum and gold (molybdenum tube) were employed in conjunction with a lithium fluoride (ZOO) analysing crystal. At the 95% confidence level, detection limits of 1.0 p g (ruthenium, rhodium and palladium) and 0.6 p g (osmium, iridium, platinumand gold) were observed for the pellet technique, with values of 0.6 and 0.2 p g , respectively, for the absorbent- pad method.The average coefficient of variation for the determination of 10 p g of the seven metals studied was 6.5% for both sample presentations. No inter-elemental matrix interferences were observed among the noble metals themselves. However, the presence of more than 200 p g of nickel or copper reduced the slopes of the calibration graph by a constant factor of 10% for the lighter metals, and amounts of more than 400 pg of the base metals reduced the slopes by 20% for the heavier members. Good agree- ment was found between the X-ray fluorescence procedures and standard atomic-absorption methods in analysis of ore concentrates. Keywords : X-ray fluorescence spectrometry ; precious metals ; absorbent-pad technique; cellulose-pellet technique The precious metals include silver and gold and metals of the platinum group (platinum, palladium, rhodium, ruthenium, iridium and osmium).Almost all of the platinum-group metals are associated with ultramafic rocks, and nickel and copper are the chief metals extracted from these rocks1 Consequently, substantial amounts of the noble metals are recovered as by-products of nickel and copper smelting procedures. The scarcity and ever growing industrial use of the precious metals have necessitated accurate and precise methods of analysis. The following analytical procedures are generally employed : spectrophoto- metry,2 atomic-absorption ~pectrophotometry,~~~ gra~imetry,~ spectrochemical methods273 and X-ray fluorescence ~pectrometry.~?~ Of these procedures, X-ray fluorescence (XRF) spectrometry appears to offer some special attractive features, such as simple sample preparation and presentation, as well as the capability of analysing a single sample for a multiplicity of elements.However, published reports3s5 on the application of XRF spectro- metry to the determination of microgram amounts of precious metals lack detail and often only consider individual members of the group. This study was launched in order to devise a simple spectrometer presentation method that permits the determination of all of the precious metals in a single sample aliquot, to evaluate the lowest limits of detection and to study matrix and inter-elemental interference effects.Experimental Procedures Materials and Chemicals All the ore concentrate samples analysed were of a nickel - copper sulphide polymetallic type and were supplied by INCO Metals Company, Sudbury, Ontario, Canada. All chemicals used were of analytical-reagent grade. Preparation of Standard Materials and Samples Gold, platinum and palladium standard stock solutions were prepared by dissolving the individual metal sponges in aqua regia. Iridium, rhodium, ruthenium and osmium standard stock solutions were prepared from ammonium hexachloroiridate( IV) , ammonium hexa- chlororhodate(II1) , ammonium hexachlororuthenate( IV) and ammonium hexachloro- osmate(IV), respectively.1038 OUMO AND NIEBOER: DETERMINATION OF MICROGRAM AMOUNTS AnabySt, vd. 104 Test samples for calibration graphs, matrix and other studies were prepared as described below.Exactly 0.50ml of the appropriate standard solution was added to a 2.2 cm diameter absorbent pad, previously impregnated with 0.30 ml of 23.7% ammonium sulphide solution. The absorbent-pad discs used were cut to 2.2 cm diameter from 4.7-cm filter absorbent pads (Cat. No. AP1003700) made by Millipore Corp., Redford, Mass., USA. During the sample preparation, the cut discs were positioned in grooves cut in a plastic sheet of 1.5-cm thickness. After the addition of the standard solution, the discs were dried for 30 min in the oven at 105 "C, after which they could be presented directly to the sample holder of the X-ray spectrometer for counting. The second procedure for presenting samples to the spectrometer involved the preparation of cellulose pellets.Samples (0.50 g) of microcrystalline cellulose were weighed into medium- sized porcelain crucibles, then 0.5 ml of 23.7% ammonium sulphide solution was added to each, followed by 1.00-ml aliquots of the appropriate sample solution. The sample mixtures were then dried in the oven at 105 "C for 45 min, allowed to cool, mixed thoroughly using a mortar and pestle, and subsequently pressed into pellets at an applied pressure of 15 ton in-2 for a duration of 15 s with a semi-automatic press. Spectrometer Details X-ray spectrometer. summarised in Table I. The X-ray fluorescence measurements were carried out with a Philips, Model PW 1220, Conditions under which the measurements were carried out are TABLE I SPECTROMETER DETAILS Parameter value for individual analvsis Parameter ' Pd Rh Ru All Pt Ir 0 s X-ray tube Voltage/kV .. CurrentlmA . . Crystal .. Collimator . . Counter* .. Analytical line Peak "20 .. Backkround,. "20 Background,, "20 Counting timels .. .. .. w W .. .. .. 95 95 .. .. .. 20 20 .. .. .. LiF (200) LiF (200) .. .. .. Fine Fine . . . . . . gfp + scint gfp + scint .. .. .. Kal KUl . . . . . . 16.76 17.51 .. .. .. 15.25 15.25 . . . . . . 18.76 18.76 .. .. .. 100 100 W 95 20 LiF (200) Fine gfp + scint Kal 18.41 15.26 18.76 100 Mo 95 20 LiF (200) Fine gfp + scint 36.97 36.23 La, - 100 Mo 96 20 LiF Fine (200) gfp + scint La1 38.01 36.23 - 100 Mo 95 20 LiF (200) Fine gfp + scint 39.20 36.23 Lal - 100 Mo 96 20 LiF (200) Fine gfp + scint 40.41 36.23 LEI - 100 * gfp, Gas-flow proportional counter; scint, scintillation counter.Spectrometer Calibration The standard samples described above were used to establish calibration graphs for each metal covering the concentration ranges 0-30 pg for the absorbent-pad technique and 0-80 pg for the cellulose-pellet method. A master standard pellet was prepared from cellulose containing 8Opg of each precious metal. In this approach to calibration, the ratios of the corrected intensities of standard samples relative to the corrected intensity of the master sample were plotted against the concentration of the precious metals in the standard samples. Unknown concentrations were evaluated by referring the corrected count ratios of the unknown pellets to this calibration graph. Matrix Effect Studies Studies of the matrix effect were carried out to investigate possible absorption and enhance- ment of the characteristic analytical lines of individual noble metals by other noble metals and the base metals copper and nickel.Matrix effects among the precious metals themselves were studied using the absorbent-pad sample preparation technique, whereas the effects of copper and nickel were studied by the pellet method. Precision Studies palladium or of iridium, platinum and gold, or 10 pg of all six noble metals. Replicate samples were prepared containing either 10 pg each of ruthenium, rhodium andNovember, 1979 OF PRECIOUS METALS USING X-RAY FLUORESCENCE SPECTROMETRY 1039 Lowest Limits of Detection The lowest limits of detection were calculated for the confidence limits of 99.7y0, 95.4% and 68.3y0, which correspond to 30b, 20b and 0b, respectively, with Ob representing the standard deviation associated with the evaluation of the background6 : where Nb is the number of counts, Rb the counting rate and Tb the counting time in seconds ( N b = &,Tb).The values of Ob used in the calculation of the lowest limits of detection were expressed in micrograms by dividing them by m, where m is the slope of the appropriate calibration graph with units of counts per second per microgram. Analysis of Ore Concentrates Ore concentrate samples were dissolved and pre-treated in order to isolate and concentrate the precious metals using one of the following procedures: (a) separation of the precious metals by leaching with 12 M hydrochloric acid followed by coprecipitation of the precious metals with tellurium'; (b) nickel sulphide collection of precious metals by fire assay8-10; (c) fusion with sodium peroxidell; and ( d ) digestion with aqua regia.The identification of the ore samples analysed and the corresponding pre-treatment employed in each instance is considered in the Discussion. In the hydrochloric acid leaching process, 30 g of finely ground ore sample were digested for 3 h on a hot-plate with 600 ml of 12 M hydrochloric acid. When the dissolution of the base metals was completed, 200 ml of hot distilled water were added, followed by 10 ml of tellurium(1V) chloride solution (2.5 g 1-l). The mixture was then boiled for 5 min, after which 30 ml of tin(I1) chloride solution (500 g 1-1 in 2.5% V/V hydrochloric acid) were added.After boiling for 10 min, the mixture was filtered and the residue was washed, on the filter, with 50% V/V hydrochloric acid. The filter and contents were digested in 250 ml of aqua regia for 1 h. This digestion mixture was then filtered and the residue was discarded after thorough washing with 50% V/V hydrochloric acid on the filter. Evaporation, nearly to dryness, of the filtrate was then carried out, and the residue was taken up in a suitable volume of 25% V/V hydrochloric acid - 1% V/V nitric acid and was diluted to volume. Additional dilutions of this stock solution were effected with 8% V/V hydrochloric acid. The sulphide collection procedure is especially suited to samples containing substantial amounts of elemental copper.Ore samples (10 g) were mixed with 60 g of disodium tetra- borate, 30 g of sodium carbonate, 15 g of silica, 16 g of nickel powder and 8 g of elemental sulphur. The mixture was fused in a furnace at 1000 "C for 2 h, and the melt was then quickly poured from the clay crucible into a steel mould. The nickel button obtained was weighed, ground, re-weighed and subjected to the 12 M hydrochloric acid leaching procedure described above. It involved fusing (in a flame) 250 mg of ore with 4 g of sodium peroxide in a zirconium crucible until the melt was clear (this usually required 10-15 min). The fusion mixture was leached with 25 ml of distilled water and was subsequently acidified with 40 ml of 12 M hydrochloric acid, and then diluted to a suitable volume.Additional dilutions of this stock solution were effected with 8% V/V hydrochloric acid. Aqua regia digestion was achieved by heating, for a suitable period of time, 3-g samples with 25 ml of aqua regia on a hot-plate. The supernatant was isolated by filtration using a fine sintered-glass filter, and was then evaporated nearly to dryness. The residue derived from this filtrate was dissolved in 5 ml of 25% V/V hydrochloric acid - 1% V/V nitric acid and diluted to volume (50 ml) with distilled water. Cellulose pellets and absorbent discs were prepared from stock sample solutions in the usual manner using 1 .OO- and 0.50-ml aliquots, respectively. The ore concentrate samples employed (see Discussion) were also analysed independently by INCO personnel using conventional flame atomic-absorption spectrophotometry after pre-treating the various samples in the manner described above, Peroxide fusion was employed for samples with low levels of nickel and copper.Ru Rh Pd Ir Pt Au Ru Rh 0 s TABLE I1 CALCULATION OF CORRECTED SPECTROMETER COUNTS For peak 28-values used see Table I ; C is the count rate, with the subscript denoting the angle relative to the peak 28-value; and CF is the correction factor.Met a1 Technique Corrected sample counts/s-l .. .. .. .. .. .. .. .. .. Absorbent pad Absorbent pad and pellet Absorbent pad and pellet Absorbent pad and pellet Absorbent pad and pellet Pellet C2e - G e - 1.51 + C2e + 2.tdCF12 C2e - (C2e - 2.97)CF C2e - (C2e - 1.dCF C2e - (C2e - 0.74lCF C2e - (C2e + o.&F Pellet Absorbent pad and pellet C2e - (C2e - 4.18)CF Definition and value of CF 2Cte / ( G e - 2.28 + G e + 1.26) 2 ~ 0 / ( G e - 1.51 + Cie + 2.00) G e I G e - 2.97 G e ICZe - 1.78 G e I G e - 0.74 G e ICte + 0.36 CF = 0.97 -j= 0.01 CF = 1.05 & 0.01 CF = 0.90 -+ 0.01 CF = 1.06 -j= 0.01 CF = 0.96 & 0.01 CF = 1.02 -J= 0.01 G e /C* 2e + 1.26 Cie K,", - 418 CF = 1.09 & 0.01 CF = 0.92 & 0.01 * Indicates that the count rates correspond to the blank sample identified in the last column.Blank Absorbent pad treated with 0.3 ml of (NH4),S solution and 0.5 ml of 8% V/V HCl As for Ru As for Ru As for Ru As for Ru As for Ru Pellet made with 0.53 g of cellulose treated with 0.5 ml of (NH4),S solution + 0.5 ml of 8% V/V HC1 As for Ru As for RuNovember, 1979 OF PRECIOUS METALS USING X-RAY FLUORESCENCE SPECTROMETRY 1041 Results Spectrometer Details and Evaluation of Backgrounds The appropriate spectrometer settings for each individual analysis are summarised in Table I.Similarly, the manner in which corrected counts were evaluated is provided in Table 11. The correction factor, CF, takes into account sloping backgrounds and tube- target contaminations and was evaluated with a metal-free blank (see the last column of Table 11). Table I11 gives a comparison of the backgrounds evaluated as directed in Table I1 with those obtained by a least-squares method applied to graphs of uncorrected counts vemas amount of precious metal. The agreement between the two sets of values is good, and in most instances is within the range covered by one standard deviation based on replicate determinations by the correction-factor method.TABLE I11 BACKGROUND EVALUATION FOR ABSORBENT-PAD TECHNIQUE See Tables I and I1 for appropriate peak 26 values and correction-factor definition ; samples used contained equal amounts of Ru, Rh, Pd, Os, Ir, Pt and Au. Background count rate at peak 2Cvalue Metal Pd . . .. .. Rh .. .. .. Ru . . .. .. Au . . .. .. Pt . . .. .. Ir . . .. .. 0 s . . .. .. w Correction-f actor method* 469 f 9 430 f 9 395 f 8 210 f 7 187 f 6 173 f 6 173 f 6 - Least-squares methodt 45 7 428 393 193 191 175 171 * Values & standard deviation; corresponds to replicate t Evaluated by extrapolating graphs of uncorrected samples. count rates versus amount of precious metal. Calibration Graphs In Table IV, a summary is provided of the regressional parameters pertaining to standard calibration graphs plotted as corrected count rates versus the amount of a single precious TABLE IV REGRESSIONAL PARAMETERS FOR CALIBRATION GRAPHS Graphs of corrected count rates versus amount of metal on absorbent pad.A least-squares method that minimised the sum of the squares only of the y-residuals was used. Samples used contained only the metal indicated. Metal Au . . Pt . . Ir . . Pd . . Rh .. Ru . . 0 s .. Slope (ratel . . 30.4 . . 12.6 . . 11.3 . . 11.4 . . 4.8 . . 8.6 . . 10.7 Correlation Intercept coefficient* at - 7.7 0.999 10.3 -8.7 1.000 2.2 1.8 1.000 3.5 2.4 0.997 9.8 -0.6 0.998 2.8 -3.1 0.998 6.3 0.9 0.999 3.9 * Correspond to a level of significance of < O .l % ( p < 0.001). t o = 1/C(y-residuals)Z/N, where N is the number of data, points.1042 OUMO AND NIEBOER DETERMINATION OF MICROGRAM AMOUNTS Analyst, vat?. 104 metal on the absorbent pad. The values of near unity of the correlation coefficients emphasise the high degree of linearity observed. It is worth noting that the highest sensitivity (largest slope) was observed for gold, and the lowest for rhodium. The standard calibration graphs for all seven metals studied for the cellulose-pellet sample presentation method gave straight line graphs through the origin, up to the maximum concentration tested of 80 pg per pellet. These consist of graphs of the ratio of corrected sample counts to the corrected master standard counts veysus the amount of metal in the pellet.The standard samples contained ruthenium, rhodium, palladium, osmium, iridium, platinum and gold at the same concentra- tion. Typical correlation coefficients (Y) were 1.000 for osmium, gold and platinum, 0.994 for iridium, 0.996 for rhodium, 0.999 for ruthenium and 0.998 for palladium; the levels of significance (9) were less than 0.001 in all instances. Interestingly, no drastic differences in sensitivities for the various metals were observed for the pellets. This uniform response is demonstrated by the nearly constant slopes for ruthenium, rhodium and palladium (approximately 1.2 x 10-2) and for iridium, platinum and gold (approximately 1 x low2) given in Table VI (samples with no copper or nickel added). Good linearity was again found over the experimental concentration range.Matrix Effects The data in Table V for the absorbent-pad technique show no strong trend in corrected count differences for sample mixtures containing 20 pg each of palladium, rhodium, ruthenium, platinum, gold, iridium and osmium compared with the data for samples with the same amount of a single metal. The observed slopes of the calibration graphs for rhodium and ruthenium were not altered significantly by 30 pg of palladium or platinum or by 30pg of both of these metals (absorbent-pad presentation, data not reported). Similarly, it is seen in the cellulose-pellet sample presentation method in Table VII (columns 5 and 8) that the same intensities were observed for samples containing either 10 pg of three or six noble metals, respectively.In Table VI, the effects of nickel and copper on the intensities of the analytical lines of the various noble metals are summarised. TABLE V COMPARISON OF CORRECTED COUNTING RATES OF EACH PRECIOUS METAL IN A MIXTURE AND WHEN PRESENT ALONE (ABSORBHNT-PAD TECHNIQUE) Amount of Corrected count rates each metal/ A > Sample te Pd Rh Ru Au Pt Ir 0 s Mixture of Pd, Rh, Ru, Pt, Au, Ir and 0 s . . . . .. . . 20.0 227 94 177 625 246 231 200 Individual precious metal . . . . . . 20.0 230 93 154 603 250 225 224 The presence of 50 pg of nickel generally increased the observed intensity, corresponding to 6-15% increases in the calibration graph slopes ; ruthenium, rhodium and palladium being most affected. In contrast, 100 pg of nickel had no measurable effect on the intensities of the six noble metals studied.However, line intensities were reduced by about 10% for ruthenium, rhodium and palladium by 200 pg of nickel, but those for iridium, platinum and gold were not. When substantial amounts of copper were present, namely 400 and 800 pg, slopes were also smaller by approximately 10% for the lighter members and up to 20% for the heavier members. Pellets containing up to 20 pg of each precious metal studied, effectively exhibited no changes in intensity when copper was added in amounts between 40 and 600 pg. This apparent contradiction must be qualified by the observation that the standard deviation for replicate measurements was of the same magnitude as the expected 10-20% modifications in intensity due to the copper (see the next section).In separate studies, copper in the amounts of 50400 pg had no measurable bearing on the magnitude of the correction factors for ruthenium, rhodium, palladium, iridium, platinum and gold (pellet method).November, 1979 OF PRECIOUS METALS USING X-RAY FLUORESCENCE SPECTROMETRY 1043 TABLE VI SUMMARY OF REGRESSIONAL PARAMETERS FOR CALIBRATION GRAPHS IN THE PRESENCE OF COPPER AND NICKEL (CELLULOSE-PELLET TECHNIQUE) In contrast to the information in Table 11, two angles were employed for Ru and Rh to evaluate the background. Samples contained a mixture of all six precious metals. Metal Ru Rh Pd Ir Pt Au Pellet content of matrix element No Ni or Cu added Ni content = 50 pg per pellet Ni content = 100 pg per pellet Ni content = 200 p g per pellet Cu content = 400 p g per pellet Cu content = 800 pg per pellet No Ni or Cu added Ni content = 50 pg per pellet Ni content = 100 pg per pellet Ni content = 200 pg per pellet Cu content = 400 pg per pellet Cu content = 800 pg per pellet No Ni or Cu added Ni content = 50 pg per pellet Ni content = 100 pg per pellet Ni content = 200 pg per pellet Cu content = 400 fig per pellet Slope 1.23 f 0.08 1.37 1.18 1.14 1.13 1.13 1.2 f 0.1 1.31 1.12 1.08 1.09 1.07 1.3 f 0.15 1.49 1.31 1.20 1.18 ( x 102) Cu content = 800 bg i)er pellet 1.15 No Ni or Cu added 1.01 f 0.03 Ni content = 25 pg per pellet 1.01 Ni content = 50 pg per pellet 1.02 Ni content = 100 pg per pellet 0.98 Ni content = 200 pg per pellet 1.00 Cu content = 400 pg per pellet 0.89 Cu content = 800 pg per pellet 0.90 No Ni or Cu added 1.02 f 0.02 Ni content = 25 pg per pellet 1.03 Ni content = 50 pg per pellet 1.13 Ni content = 100 pg per pellet 1.02 Ni content = 200 pg per pellet 1.04 Cu content = 400 pg per pellet 0.85 Cu content = 800 pg per pellet 0.88 No Ni or Cu added 1.05 f 0.01 Ni content = 25 pg per pellet 1.02 Ni content = 50 pg per pellet 1.11 Ni content = 100 pg per pellet 1.03 Ni content --- 200 p g per pellet 1.08 Cu content = 400 pg per pellet 0.85 Cu content = 800 pg per pellet 0.87 Intercept ( x 102) -1.2 f 0.4 - 5.66 0.13 1.73 0.20 - 0.88 -1.7 f 1.3 -2.38 3.00 4.72 0.62 1.81 -0.3 5 2 -5.15 -0.62 3.99 0.93 3.57 1.37 0.55 1.2 f 0.8 -0.33 - 1.44 1.08 1.22 1.4 f 0.5 -0.31 -1.12 -0.77 - 1.99 2.13 0.55 1.3 f 0.2 1.00 0.21 0.58 - 1.15 1.01 -0.26 Correlation coefficient * 0.991 0.998 0.997 0.997 1.000 0.988 0.999 0.997 0.997 0.997 0.988 0.998 0.997 0.998 0.995 1.000 1.000 0.999 0.993 0.997 1 .ooo 0.995 0.999 0.999 0.993 0.997 1.000 0.997 0.999 0.999 0.995 0.996 1 .ooo - - - - - - U t ( x 102) - 3.93 1.66 2.24 1.93 0.64 4.38 1.15 2.11 2.15 1.98 4.98 1.98 2.34 1.58 2.60 0.45 0.30 1.16 2.44 1.78 0.40 1.29 0.48 0.99 2.70 1.49 0.44 1.09 0.55 0.85 2.28 1.75 0.54 - - - - - Number of data points, N 6 6 6 6 6 5 6 6 6 6 6 5 6 6 6 6 6 5 6 5 6 6 6 6 5 6 5 6 6 6 6 5 6 5 6 6 6 6 5 * Level of significance is <O.l% ( p < 0.001).t a = 2/C(y-residuals)2/N, corresponding to graphs of count ratios (sample to master standard) veisus amount of the specific precious metal in the cellulose pellet. Precision Studies The reproducibility observed (Table VII) for replicate samples was comparable for the two sample presentation methods studied.For 10-pg amounts of the precious metals, the standard deviations were in the range 0.3-1.2 pg, with means of 0.7 and 0.6 pg for the absorbent-pad and cellulose-pellet presentation procedures, respectively. The standard deviation values in columns 4, 7 and 10 of Table VII when multiplied by the factor 10 are converted into the coefficient of variation: (a in micrograms per microgram of metal) x 100 yo. Finally, the reproducibility of calibration graph slopes for replicate sets of calibra- tion standards was generally good. For gold, platinum and iridium, average deviations of 1-3% were observed in the magnitudes of the slopes, while the reproducibility was not as good (average deviations in the slopes of 7-12%) for ruthenium, rhodium and palladium (see the “No Ni or Cu added” entries in Table VI).1044 OUMO AND NIEBOER: DETERMINATION OF MICROGRAM AMOUNTS Analyst, VoZ.104 TABLE VII PRECISION STUDIES Count ratios refer to the ratio of corrected counts of sample to master standard sample, multiplied by 100; multiplication of the quantities in columns 4, 7 and 10 by the factor 10 yields coefficients of variation; SD is the standard deviation. Absorbent-pad technique c Average count Metal ratio* Ru . . . . 20.3 Rh .. . . 17.4 Pd . . . . 29.2 Ir . . . . 19.8 Pt . . . . 23.9 Au . . . . 43.8 SD of ratio SD/pg 1.22 0.60 1.63 0.94 1.65 0.57 0.86 0.43 1.50 0.63 4.90 1.12 r Average count ratio 15.9t 17.3t 15.5t 11.4t l0.6$ 11 .O$ Cellulose-pellet technique SD of count SD of ratio SD/pg ratios ratio 0.86 0.54 16.2 1.52 1.61 0.93 17.9 0.62 1.00 0.64 15.6 1.88 0.52 0.46 11.7 0.57 0.36 0.33 10.9 0.42 0.56 0.53 10.3 0.69 A - Average SD/H 0.94 0.35 1.20 0.49 0.39 0.57 * Corresponds to 8 replicate samples; samples contained 10.0 pg of all six metals.t Corresponds to 4 replicate samples; samples contained 10.0 pg of Ru, Rh and Pd. $ Corresponds to 5 replicate samples; samples contained 10.0 pg of Ir, Pt and Au. Corresponds to 4 replicate samples; samples contained 10.0 pg of all six metals. Detection Limits The lowest limits of detection based on background count rates measured for a counting time of 100 s are summarised in Tables VIII and IX. It is evident that the detection limits calculated are better by a factor of about two for the absorbent-pad technique (Table IX) relative to the pellet method (Table VIII).The detection limits for the heavier elements osmium, iridium, platinum and gold are twice as good or better than those for ruthenium, rhodium and palladium. TABLE VIII LOWEST LIMITS OF DETECTION OF VARIOUS PRECIOUS METALS Results found using the cellulose-pellet technique with a counting time, TI,, of 100s. The lowest limit of detection was calculated as 3 4 m , 2a1,lm and ob/m for the confidence intervals of 99.7%, 95.4% and 68.3%, respectively. The slope factor m corresponds to the slope of calibration graphs consisting of plots of corrected counting rates zlevsus concentration ; U b is defined in equation (I), Atomic number 44 46 46 76 77 78 79 Metal Ru Rh Pd 0 s Ir Pt Au Slope factor, m/ counts s-l pg-l 6.4 5.6 4.7 5.0 5.3 5.5 5.6 Lowest limit of detection (pg g-l) at the specified confidence limit Qb/ I A counts s-l 99.7% 95.4% 68.3%’ 2.67 1.26 0.84 0.42 2.74 1.48 0.99 0.49 2.87 1.83 1.22 0.61 1.50 0.91 0.61 0.30 1.59 0.90 0.60 0.30 1.71 0.94 0.63 0.31 1.68 0.89 0.59 0.30 Analysis of Precious Metal Concentrates In Fig.1, the results of analysing the ore concentrates by XRF spectrometry are com- pared with the accepted values determined by flame at omic-absorpt ion spect ropho t omet ry . The calibration graphs used corresponded to those for mixtures of precious metals using the absorbent-pad technique and, except for one ore concentrate, to those for noble metal mixtures in the presence of copper for samples analysed in the pellet form (see Table VI). The exception was the sample designated MY (see Discussion), for which the calibration graphs in Table VI, without nickel or copper added, were used.Except for the platinum level in one sample, the agreement between the two methods was good, as indicated byNovember, 1979 OF PRECIOUS METALS USING X-RAY FLUORESCENCE SPECTROMETRY 1045 the regressional parameters reported in the legend of Fig. 1. Of the two sample presenta- tions used, the cellulose-pellet procedure showed the best accord with the atomic-absorption (AA) spectrophotometric values (slope of fitted line, rn = 1.0). The absorbent-pad results were generally 5% lower than the AA or pellet XRF values (1% = 0.95). 80 60 20 I / 0 10 20 30 40 50 60 70 80 0 10 20 30 40 50 60 70 80 Accepted value (AA) Accepted value (AA) Fig.1. Comparison of the analysis of ore concentrates by X-ray fluorescence spectro- metry and accepted value by flame atomic-absorption spectrophotometry. (a), Absorbent- pad presentation: slope, m = 0.95; intercept, b = 0.00; correlation coefficient, r = 0.996; and level of significance, p t0.001. (b), Cellulose-pellet presentation: m = 1.02; b = -0.10; r = 0.997; p <0.001. Perfect agreement is depicted by the dotted line, and the units of concentration are either % m/m or ounces per ton. The anomalous Pt values shown (indicated by question marks) were excluded from the regressional analyses. , Ruthenium; 0, palladium; A, gold; 0, indium; 0, rhodium; and A, platinum. Discussion Sample Presentation The ammonium sulphide impregnation technique for localising metal ions was devised after preliminary localisation studies with nickel(I1). Initially, spotted absorbent pads were exposed to anhydrous hydrogen sulphide in a closed atmosphere.It was found that the black nickel(I1) sulphide precipitate preferentially accumulated at the edges, and thus centrifugal migration of the spotted nickel(I1) occurred. The pre-treatment of the absorbent TABLE IX LOWEST LIMITS OF DETECTION OF VARIOUS PRECIOUS METALS Results found using the absorbent-pad technique with a counting time. Tb, of 100 s. The lowest limit of detection was calculated as 3ub/m, 2Ublm and ub/m for the con- fidence intervals of 99.7y0, 95.4% and 68.3%, respectively. The parameter m is the slope defined and given in Table IV.Atomic number 44 46 46 76 77 78 79 Lowest limit of detection (pg) at the specified confidence limit r I A Element 99.7% 95.4% 68.3% Ru 0.69 0.46 0.23 Rh 1.29 0.86 0.43 Pd 0.57 0.38 0.19 0 s 0.36 0.24 0.12 I r 0.36 0.24 0.12 Pt 0.33 0.22 0.11 Au 0.16 0.10 0.061046 OUMO AND NIEBOER: DETERMINATION OF MICROGRAM AMOUNTS Analyst, vd. 104 pads with ammonium sulphide circumvented this migration. As the precious metal samples were dissolved in 8% V/V hydrochloric acid, the formation of complex ammonium chloro derivatives and various sulphides presumably occurred on contact between the sample aliquot and the impregnated absorbent pad. To prevent losses due to liquid adsorption on to container walls, ammonium sulphide was also used as the precipitant in the cellulose-pellet sample presentation method.It appears that the two sample presentation techniques developed offer some definite advantages over other known micro-sampling methods. As implied in the previous para- graph, evaporation procedures are often accompanied by centrifugal migration of ions, resulting in a non-uniform localisation on the filter discs. MacNevin and Hakkila5 did not mention this problem although they may have overcome it by adding the liquid aliquot to the centre of an oblong of paper placed concave upwards and resting on its four corners. Another approach is to collect a precipitate on a membrane filter.3J2 For example, Pietzner and Werner3313 determined gold by XRF spectrometry after tellurium collection, and trapping the precipitate on a membrane.They, and others using filter membranes,12 point out that lack of adhesion was encountered with voluminous precipitates. Cracking and peeling of the precipitates occurred. In addition, the adsorption properties of some colloidal precipi- tates make handling difficult prior to and during the filtration step.12 In this work total amounts of 210 pg of noble metals could be handled in the absorbent-pad procedure, without difficulty, although matrix studies involving large amounts of nickel and copper (400- 800 pg) did result in poorly adhering precipitates. None of these problems were encountered using the pellet method. Finally, the anion-exchange collection used by Taylor and Beamish3314 to determine small amounts of ruthenium was reported to give a uniform distribution on the exchange paper, and appears to compare favourably with the absorbent- pad and pellet techniques.Background Evaluation and Resolution An examination of the X-ray fluorescence emission lines for other elements15J6 revealed that possible interferences by overlap with the chosen analytical lines (Table I) was minimal. Potentially interfering lines were either very weak, of high excitation energy, or belonged to elements normally absent in samples containing the precious metals (e.g., lanthanides and actinides). The only observed exception was the overlap of the copper K/3 line with the osmium Lal, preventing the determination of osmium in samples containing appreciable amounts of copper. In studies on concentrated solutions of copper - nickel mattes in hydrochloric acid, the K/i? line of copper also affected the determination of p l a t i n ~ m .~ ~ ~ 7 ~ 1 The interference by overlap reported by MacNevin and Hakkila5 of palladium Ka with rhodium Ka, and platinum La, with iridium La,, was not evident in our work. Tube-target contamination was not serious, as indicated by the values of the correction factor, CF, of 1.0 The near constancy of the background intensities for palladium, rhodium and ruthenium (tungsten tube) and for gold, platinum, iridium and osmium (molybdenum tube) reaffirms the good base-line linearity. Tube- target contamination precludes the use of the tungsten tube for gold, platinum, iridium and osmium. The considerably higher background count rates for the lighter members (palladium, rhodium and ruthenium) account for their poorer detection limits.0.1 reported in Table 11. Sensitivity, Detection Limits and Concentration Ranges A comparison of slopes expressed in count rate per niicrogram of metal on the absorbent pad (see Table IV) gave the relative sensitivity order gold > osmium, platinum, iridium, palladium > ruthenium > rhodium. Consequently, the lowest detection limits reported in Table IX should and do follow the same trend. In contrast, there was not much spread in the corresponding slopes for the pellet method (see Table VIII), or in the detection limits (Ta.ble VIII). The reduced sensitivity for the cellulose pellet is not surprising, as it corre- sponds to a thicker and more dense and compact matrix.(X-ray fluorescence intensity obeys Beer’s law, A = ppx, where A is the absorbance, p the mass absorption coefficient, p the density and x the thicknes6) The observed detection limits for the absorbent-pad technique between 0.10 and 0.50pg (excluding rhodium) a t the 95% confidence interval, asNovember, 1979 OF PRECIOUS METALS USING X-RAY FLUORESCENCE SPECTROMETRY 1047 well as the corresponding range of 0.60-1.0 pg (excluding palladium) for the cellulose pellets, are comparable or better when compared with the values reported for individual precious metals3 However, they are an improvement by several orders of magnitude over those reported in the only detailed XRF spectrometric analysis of microgram mixtures of noble met a k 5 The concentration range for which linear calibration graphs might be expected appears to be very wide indeed according to reports by other w0rke1-s.~~~ Linearity up to milligram levels has been reported for iridium, platinum, palladium and rhodium for an evaporation sample-presentation te~hnique,~ amounts up to 150 pg for ruthenium by the anion-exchange paper method and osmium and gold up to 100 pg for precipitate collection on filter-paper.3 Good linearity of calibration graphs for percentage-level concentration ranges are also known : 0.2-2y0 for ruthenium in silica pellets (and 0.08-0.20 g 1-1 in solution),19 0.08-0.16~0 for rhodium, l.2-l.8y0 for platinum, 1.5-3.0% for silver and 2-6% for palladium in pellets of industrial platinum concentrates, copper - nickel slimes and products of their conversion,20 and 04% for platinum in flattened silver beads obtained by cupellation in the classical fire- assay process.21 Similar ranges have been cited by Beamish et aL3 for less recent studies.Consequently, the limit of 80 pg of each precious metal corresponding to a total of 560 pg chosen as the arbitrary upper limit in the cellulose-pellet studies should not constitute the actual experimental limiting level. In contrast, and as mentioned above, the maximum amount that can be handled by the absorbent-pad procedure would be limited by the adhesive properties of the precipitate. Nevertheless, total amounts of noble metals up to 200 pg pose no problem. Precision The observed values of the coefficient of variation corresponding to 1Opg of precious metal between 3 and 12% with an average of 6.5% (see Table VII) are comparable to, or smaller in magnitude than, those reported for analogous3 and larger concentration^^^^^ of individual precious metals.Comparable estimates were associated with the reproducibility of calibration graphs. Even though the observed precision for some of the metals studied was poorer than that of existing analytical norms, it is felt that it is adequate for the simple, but relatively fast, sample preparation techniques employed. It is reasonable to assume that some streamlining of the procedure and the familiarity accompanying repeated use should improve the precision. Inter-elemental Effects AS not much information is available on inter-elemental effects, the results obtained in this work are discussed in some detail.The lack of any obvious inter-elemental effect in the data in Table V, in which a comparison of counts of 20 pg of individual noble metals is compared with the same amount present in mixtures, was reinforced by the count ratios summarised in Table VII. In this table, no differences in intensities are recorded for samples containing three or six noble metals. This result is not surprising as the absorption coefficients for the absorption of the La, lines of osmium, iridium, platinum and gold by these same metals and by ruthenium, rhodium and palladium are all in the range of 110-160 (for compilations of absorption coefficients, p, consult MullerZ2 and Jenkins and de Vriess), and thus mixtures of these metals provide a nearly constant matrix.The value of the mass absorption coefficients for the absorption by sulphur of these La, lines is considerably lower (about 60). Similarly, there is very little self-absorption by palladium, ruthenium and rhodium of their own and each others Kcc lines (p rn 15). Absorption of these lines by osmium, iridium, platinum and gold is also low (p w70), and by sulphur is negligible (p -5-7). Consequently, favourable absorption coefficients combine with the low relative atomic mass cellulose matrix in the pellet and the thin layer of precipitate spread over a relatively large area on the absorbent pad, to generate matrices relatively free of inter-noble metal matrix absorption and enhancement effects. A similar conclusion was reached by Coombes et aZ.2l as they found that small amounts of palladium, rhodium, gold and iridium did not affect the intensity corresponding to 1% of platinum (Lcc line) in silver cupellation beads. Platinum metals did not interfere appreciably with the determination of gold (LBl line) in hydrochloric acid solution^.^,^^1048 Analyst, “ol.104 The observed influences of nickel and copper may have been expected on the basis of their known emission and absorption properties. Firstly, the interference of the copper KP line in the determination of osmium has already been mentioned. Secondly, copper and nickel do not strongly absorb the Ka lines of ruthenium, rhodium and palladium (p WOO), but they do more significantly affect the La lines of iridium, platinum and gold (p = 220- 280; the values for osmium La line absorption are 40 for copper and 280 for nickel).As the matrix studies reported in Table VI correspond to mixtures of all seven noble metals, the presence of 50 pg of nickel in the pellet should have rendered the matrix lighter for the Kcc radiation of ruthenium, rhodium and palladium, but not for the La lines of osmium, iridium, platinum and gold (see values of p given earlier). The 6-15% increases in calibration graph slopes due to the 50 pg of nickel for ruthenium, rhodium and palladium plus the lack of any significant effect on the slopes for the heavier metals are therefore reasonable. The presence of larger amounts of copper or nickel might be expected to reduce the total amount of radiation reaching the surface, and thus to reduce the detection sensitivity for all the precious metals.This effect was observed, as reductions in calibration graph slopes of up to 20% were recorded (Table VI). Presumably, the levelling off of this absorption effect for additions of more than 200 pg of copper and nickel for the lighter members and more than 400 pg for the heavier members corresponds to the condition of adding a large amount of a weak to moderate absorber as a diluent to dominate the matrix absorption.6 For small amounts of precious metals (q., 10 pg) the reductions in intensity were not statistically significant. OUMO AND NIEBOER: DETERMINATION OF MICROGRAM AMOUNTS Analysis of Ore Concentrates The ore concentrates examined and the pre-treatment employed (given in parentheses) were as follows: MY, a copper - nickel concentrate (hydiochloric acid leaching) ; RR, concentrate MY treated to remove most of the nickel as the carbonyl derivative (nickel sulphide collection) ; ISRM, INCO standard reference material, which is a Bessemer con- verter matte (peroxide fusion); SC, sulphur cake, concentrate RR with most of the copper removed (aqua regia digestion) ; and a precious metal concentrate (peroxide fusion).Each sample was analysed for ruthenium, rhodium, palladium, iridium, platinum and gold. It was not possible to analyse for osmium because of overlap of its La emission line with the K/3 line of copper, as X-ray fluorescence spectrometer scans showed the presence of substantial amounts of residual copper in the XRF samples (except in the instance of the MY concentrate).The under-estimation of the metal levels by the absorbent-pad method relative to the cellulose-pellet procedure (as well as atomic absorption) may possibly be due to the fact that the calibration graphs used were determined in the absence of copper in the standard samples (see Results). Apart from the platinum level in a single sample, the agreement between the XRF methods developed and the AA procedure is that which one might expect from an analytical pro- cedure having average coefficients of variation of 6.5%. The anomalous results obtained for platinum corresponded to the SC ore concentrate. As no separation of base metals was carried out for this sample, it is conceivable that some unidentified element interfered. The remaining deviations from the accepted values were random.This observation indicates that the XRF spectrometric procedure is not dependent on the nature of the ore concentrates examined, or on the separation and dissolution pre-treatments employed. The implication of this lack of gross matrix dependence is that ore samples can be examined by XRF without any pre-treatment provided that the calibration standards are of approximately the same composition and that high enough concentrations of the precious metals occur in the ore samples. have indeed determined silver, palladium, rhodium, ruthenium, gold and platinum by direct measurements on pellets prepared from industrial platinum concentrates, copper - nickel slimes and products of their conversion. This is in contrast with those used for pellet samples.As already mentioned, Shestakov et The authors thank J. Bozic and S. Maggs, Central Analytical Service Laboratory, INCO Metals Co., Ontario Division, Copper Cliff, Ontario, for their assistance and advice. D. Guest of the Geology Department, Laurentian University, also provided technical help. Financial support €or P. R. Oumo by the Canadian International Development Agency is gratefully acknowledged, as is the research funding received from Laurentian University.November, 1979 OF PRECIOUS METALS USING X-RAY FLUORESCENCE SPECTROMETRY 1049 References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. Youngquist, W. L., “Investing in Natural Resources : Today’s Guide to Tomorrow’s Needs,’’ Dow The National Research Council (USA), “Platinum-Group Metals,” National Academy of Sciences, Beamish, F.E., Lewis, C. L., and Van Loon, J . C., Talanta, 1969, 16, 1. Gilchrist, R., and Wichers, E., J . Am. Chem. Soc., 1935, 57, 2565. MacNevin, W. M., and Hakkila, E. A., Analyt. Chem., 1957, 29, 1019. Jenkins, R., and de Vries, J. L., Palmer, I., and Streichert, G., “The Coprecipitation of Nobel Metals with Tellurium. Jones - Irwin, Homewood, Ill., 1975, p. 165. Washington, D.C., 1977, pp. 66-78. Practical X-ray Spectrometry,” Second Edition, Springer-Verlag, New York, 1969. I. Platinum, Palladium, Rhodium and Gold,” National Institute for Metallurgy, Johannesburg, Report No. 1273, 1971. RobCrt, R. V. D., van Wyk, E., and Palmer, R., “Concentration of the Noble Metals by a Fire- assay Technique Using Nickel Sulphide as the Collector,” National Institute for Metallurgy, Johannesburg, Report No. 1371, 1971. RobCrt, R. V. D., and van Wyk, E., “The Effects of Various Matrix Elements on the Efficiency of the Fire-assay Procedure Using Nickel Sulphide as the Collector,” National Institute for Metal- lurgy, Johannesburg, Report No. 1705, 1975. Dixon, K., Jones, E. A., Rasmussen, S., and Robkrt, R. V. D., “The Efficiency of the Fire-assay Procedure with Nickel Sulphide as the Collector in the Determination of Platinum, Silver, Gold and Iridium,” National Institute for Metallurgy, Johannesburg, Report No. 1714, 1975. Jeffery, P. G., “Chemical Methods of Rock Analysis,” Pergamon Press, Oxford, 1970, p. 29. Watanabe, H., Berman, S., and Russell, D. S . , Talanta, 1972, 19, 1363. Pietzner, H., and Werner, H., 2. Analyt. Chem., 1966, 221, 186. Taylor, H., and Beamish, F. E., Talanta, 1968, 15, 497. White, E. W., and Johnson, G. G., “X-ray Emission and Absorption Wavelengths and Two-Theta Tables,” ASTM Data Series DS 37A, Second Edition, American Society for Testing and Materials, Philadelphia, Pa., 1970. Vandorpe, B., and Durr, J.] Analusis, 1977, 5, 38. Strasheim, A., and Wybenga, F. T., Appl. Spectrosc., 1964, 18, 16. Wybenga, F. T., and Strasheim, A., Appl. Spectrosc., 1966, 20, 247. Leoni, L., Braca, G., Sbrana, G., and Giannetti, E., Analytica Chim. Acta, 1975, 80, 176. Shestakov, V. A., Arkhipov, N. A., Makarov, D. F., and Kukushkin, Yu. N., Zh. Prikl. Khim., Coombes, R. J., Chow, A., and Flint, R. W., Analytica Chim. Acta, 1977, 91, 273. Miiller, R. O., “Spectrochemical Analysis by X-ray Fluorescence,” Plenum, New York, 1972. Chow, A., and Beamish, F. E., Talanta, 1966, 13, 539. Leningr., 1974, 47, 1035. Received December 13th, 1978 Accepted April 12th, 1979

ISSN:0003-2654    

DOI:10.1039/AN9790401037

出版商:RSC    

年代:1979

数据来源: RSC

摘要:

1050 Analyst, November, 1979, Vol. 104, pp. 1050-1054 X-ray Fluorescence Determination of Platinum and Palladium in Platinum Concentrates Using a Solution Technique Z. Cruickshank and H. C. Munro Johannesburg Consolidated Investment Company Limited, Minerals Processing Research Laboratory, P.O. Box 13017, Knights, Transvaal 1413, South Africa An X-ray fluorescence solution technique for the determination of platinum and palladium in platinum-bearing material is described. Ruthenium has to be removed prior to the measurement of the platinum and palladium. Mercury and thorium are used as internal standards. The method is precise and is more rapid than the gravimetric method normally used. Keywords : X-ray fluorescence spectrometry ; platinum determination ; palladium determination ; mercury and thorium internal standards ; platinum concentrates An accurate X-ray fluorescence spectrometric technique was required for the determination of platinum and palladium in a variety of platinum-bearing materials. A literature survey yielded few instrumental methods.lS2 Owing to the complexity and heterogeneity of these materials, it was decided to use a solution technique to obviate possible mineralogical and particle size effects.Absorption effects would be countered by the use of internal standards. It is possible by correct planning and scheduling to analyse concurrently large numbers of samples for platinum and palladium, thus halving the analysis time normally required for the determination of these elements by the gravimetric procedure.This method does not require pre-analysed samples to be used as standards, but uses pure platinum and palladium solutions covering the calibration range required. Experimental Dissolution The silica content of the samples varied widely, in some instances being as high as 20%, which made treatment with hydrofluoric acid advisable. This was carried out in PTFE beakers either on a steam-bath or on a low-temperature hot-plate. The hydrofluoric acid treatment was followed by leaching of the residue with aqua regia and subsequent filtration through a Millipore filter. The resultant residue was scanned on an X-ray fluorescence spectrometer and found to contain appreciable amounts of platinum group metals, which made the dissolution of the residue necessary. It was also found that the mass of residue varied considerably from sample to sample.The residue was fused with sodium peroxide, leached and added to the aqua regia solution. Choice of Internal Standards The above solution was scanned on an X-ray fluorescence spectrometer so that suitable internal standards could be chosen. The main criteria for choosing an internal standard are that it must not be present in the sample, it must have an X-ray fluorescence line whose energy is similar to that of the line being analysed, and no absorption edges may fall between these two lines except absorption edges generated by the two elements concerned. According to the above criteria, mercury was chosen as the internal standard for platinum. A few trial solutions were prepared and the optimum mercury concentration was deter- mined.The choice of an internal standard for palladium was more difficult. On the short wave- length side, the spectrum was very crowded and full of absorption edges, the ruthenium KP line overlapped the palladium Ka line, and no suitable internal standard or background position could be found. On the long wavelength side, there were the rhodium Ka and ruthenium K a lines. If the ruthenium was removed, both the ruthenium Ka line, and theCRUICKSHANK AND MUNRO 1051 ruthenium KP line interference on the palladium Ka line, would disappear. This would have made niobium a good internal standard for palladium if the niobium K/3 line was used. However, it was found that the preparation of a 40 g 1-1 niobium solution was difficult and, as an alternative, thorium was chosen.Although thorium dissolved easily, it was found that on mixing with the sample solution a crystalline precipitate formed. This was overcome by vigorously boiling the thorium solution for 1-2 min during the original dissolu- tion process. This stabilised the thorium standard solution and precipitation no longer occurred on mixing with the sample solution, even when this mixture was allowed to stand for over 2 weeks. The usual method of removing ruthenium is by the addition of sodium bromate in the presence of sulphuric acid. Although this works well, the sulphuric acid causes precipitation of thorium and this method is therefore unsuitable. Fuming perchloric acid was also tried, but proved to be time consuming. The most suitable method was found to be the stepwise addition of sodium bromate and hydrochloric acid,3 which resulted in the complete removal of ruthenium, provided that no nitric acid was present in the sample.The analytical line used was the thorium Ly,. Proposed Method Samples and 11-19yo of palladium were analysed. Samples of platinum-bearing concentrate containing approximately 16-35y0 of platinum Reagents All reagents used were of analytical-reagent grade. Hydrojuoric acid, sp. gr. 1.16. Hydrochloric acid, sp. gr. 1.19. Nitric acid, sp. gr. 1.40. Sodium peroxide. Granular. Sodium bromate solution. Mercury internal standard solution. Tlzoriztm internal standard solution. Platinum standard solution. Palladium standard solution. A 10% solution in de-ionised water.Dissolve 30 g of mercury(I1) chloride in 1 1 of 3 M Dissolve 200 g of Th(N03),.(4-6)H20 in 1 1 of 3 M Prepared by dissolving Specpure platinum sponge in aqua Prepared by dissolving Specpure palladium wire in aqua hydrochloric acid. hydrochloric acid. regia to give solutions containing 6, 9, 13, 16 and 19 g 1-l. regia to give solutions containing 4, 5, . . ., 10 g 1-1. The solution was boiled vigorously for 1-2 min after preparation. Apparatus Philips PW 1450/20 sequential X-ray jhorescence spectrometer. sample changer and related equipment were used. Millipore filter apparatus with a 0.45-pm membrane. Vitreous carbon crucibles with 30-cm3 capacity. A 60-position automatic Any similar instrument could be used. Procedure Samples (5 g) were weighed in triplicate into PTFE beakers and 15 cm3 of hydrofluoric acid were added.The PTFE beakers were placed on a low-temperature hot-plate and the samples dried; if large amounts of silica were known to be present the dried sample was re-dissolved in another 15 cm3 of hydrofluoric acid and the drying step repeated. Aqua regia (80 cm3) was then added, the samples were digested on a steam-bath for approximately 2 h and, after cooling, filtered through a Millipore filter apparatus. The filtrate was evaporated to dryness on a steam-bath, 15 cm3 of hydrochloric acid were added and the beakers returned to the steam-bath to dissolve the residue from the evaporated filtrate. The Millipore membrane was transferred into a carbon crucible, dried and ignited at 800 "C. Sodium peroxide (4 g) was added to the cool crucible containing the residue and the sample was carefully fused.The fusion mixture was leached with approximately 80 cm3 of distilled water in a 400-cm3 squat-form beaker and the carbon crucible removed. The hydrochloric1052 CRUICKSHANK AND MUNRO: X-RAY FLUORESCENCE DETERMINATION OF Analyst, Vd. 104 acid filtrate was also transferred into the same beaker, the resultant solution tested with pH paper to ensure that it was acidic and if necessary a few more drops of hydrochloric acid were added. The solution was brought to the boil, and 60 cm3 of 10% sodium bromate solution were slowly added to the boiling solution, in 5-cm3 aliquots, followed by 50 om3 of concentrated hydrochloric acid added in the same manner. When the addition of the sodium bromate and the hydrochloric acid was complete, the solution was boiled to reduce the volume, cooled and diluted to volume in a 100-cm3 cali- brated flask.Two 10-cm3 aliquots of the solution (one for the platinum determination and the other for the palladium determination) were pipetted into 100-cm3 beakers. Mercury(I1) chloride solution (10cm3) was pipetted into the beaker containing the aliquot for the platinum measurement and thorium nitrate solution (10 cm3) was pipetted into the beaker containing the aliquot for the palladium determination. The platinum and palladium were measured by X-ray fluorescence spectrometry using the Feather and Willis* background-correction method. According to this method, background intensities at differing wavelength positions are linearly related provided that they fall between adjacent major element absorption edges.This allows for the determination of the background by a single measurement at an interference-free background position. Background-correction calibrations may be set up at the beginning of a run by using blanks. The background-correction calibration will be of the form y = mx + c , where y is the blank reading at the peak position, x the blank reading at the background position and c and m are constants. The choice of blanks for the platinum and mercury background corrections is relatively easy. At least two solutions must be used, one with a light matrix and the other with a heavy matrix such that the background intensity of the sample falls between those of the two blanks.De-ionised water and a 100 g 1-1 nickel solution would be suitable blanks for the platinum and mercury background corrections. The choice of blanks for the palladium and thorium background corrections is, however, more difficult, in that a thorium absorption edge falls between the background position and the palladium and thorium lines. For the palladium background correction, this is over- come by adding thorium to the blanks in the same ratio as to the samples. The background correction is thus always made across the thorium absorption edge, and as the amount of thorium is always constant, no matrix variations due to thorium can occur. A 1 + 1 mixture of de-ionised water and the thorium standard solution, plus a 1 + 1 mixture of a 9Og1-1 platinum solution and the thorium standard solution, would be suitable blanks for the palladium background correction.For the thorium background correction de-ionised water and a 90 g 1-1 platinum solution could be used provided that the water is “infinitely thick” to thorium radiation. Alternatively, instead of water, a slightly heavier matrix can be used, e.g., a nickel solution of sufficient concentration to be “infinitely thick” to thorium radiation. This will depend on the geometry and capacity of the sample holders used. Tube* .. . . .. Counter . . .. .. Collimator . . . . . . Discriminator setting . . Counting time, peak and background/s . . .. Peak angle, “28 . . .. Background angle, “219 . . Voltage/kV . . .. .. CurrentlmA .. .. Vacuum . . .. .. Counter H.T./V .. .. Spinner . . .. . . Crystal . . .. .. TABLE I INSTRUMENTAL PARAMETERS Platinum determination Palladium determination Mo LiF (220) Scintillation Fine (160 pm) LL = 200, window = 200 2 x 20 Hg La = 51.69; Pt La = 54.92 67.70 80 36 Off -1040 On Au LiF (220) Scintillation Fine (160 pm) LL = 200, window = 200 2 x 20 Pd Ka = 23.77; Th Lyl = 26.56 28.70 80 35 Off -1040 On * It was not possible to use the same tube for both platinum and palladium determinations owing to line interferences.November, 1979 Pt and Pd IN Pt CONCENTRATES USING A SOLUTION TECHNIQUE 1053 As the thorium content is always constant, the thorium absorption edge between the back- ground position and the thorium line can be disregarded. The instrumental parameters used in the determinations are shown in Table I.Results Values obtained by the X-ray fluorescence spectrometry solution technique compared well with values obtained by the two other laboratories using wet-chemical procedures (see Table 11). TABLE I1 X-RAY FLUORESCENCE VALUES AND VALUES OBTAINED FROM TWO OTHER LABORATORIES USING WET-CHEMICAL PROCEDURES Sample No. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 16 16 17 18 19 20 21 22 23 24 26 Average . . Platinum determined, yo Laboratory Laborator; X-ray value No. 1 No. 2 17.70 17.88 17.61 16.64 16.84 16.40 23.77 24.31 24.17 26.49 27.44 27.63 28.97 29.23 29.04 30.32 30.50 30.37 27.80 27.25 27.05 22.14 22.62 22.26 24.93 24.98 25.12 24.62 24.96 24.90 24.18 24.25 24.24 28.16 28.52 28.28 28.34 28.52 28.07 24.97 25.19 24.89 31.32 28.11 31.51 28.86 28.54 28.37 26.12 26.37 26.20 28.07 27.47 27.96 25.05 25.05 25.17 27.86 27.96 28.17 29.30 29.49 29.21 35.74 35.21 35.11 29.97 29.87 30.13 32.32 32.17 32.77 32.32 32.23 32.38 Palladium determined, yo r A > Laboratory Laboratory X-ray value No.1 No. 2 13.40 13.90 13.68 14.11 14.67 14.34 12.58 13.20 12.93 13.89 14.64 14.40 13.92 14.28 13.89 11.86 12.14 11.76 12.91 13.50 13.23 15.40 15.81 15.62 14.39 14.57 14.70 14.52 14.78 14.81 14.33 14.10 14.20 14.43 14.75 14.50 15.67 16.12 15.48 15.23 15.59 16.39 14.87 15.35 14.92 15.69 16.08 15.99 16.08 16.20 16.20 16.22 16.10 16.25 16.37 16.52 16.32 15.86 15.87 16.12 15.22 15.31 15.26 13.62 13.87 13.63 17.82 18.36 18.11 18.81 19.25 18.67 19.09 19.78 19.33 . . 27.04 27.00 27.08 15.05 15.40 16.17 Statistical comparisons of the results indicate that while there is no significant difference at the 0.05 level for platinum, there is, however, a significant difference for palladium between the three laboratories (see Table 111).The sum of squares (for X-ray fluorescence spectrometric results and results for labora- tories 1 and 2) was decomposed into two contrasts, one comparing laboratory 1 and laboratory 2, and the other comparing the X-ray fluorescence spectrometry technique with the mean values obtained from laboratory 1 and laboratory 2.5 As expected, the t values obtained from the contrasts for platinum were not significant, but for palladium the t values obtained from the contrasts were significant. It can be seen that the reproducibility of the palladium results obtained by laboratory 1 and laboratory 2, is of the same order as the reproducibility of results obtained by the X-ray fluorescence spectrometry technique and the mean of the results of laboratory 1 and laboratory 2. The removal of the ruthenium is time consuming, but with proper planning large numbers of samples can be treated simultaneously, whereas the wet-chemical procedure as used by this laboratory requires, after the dissolution of the sample, the separation of the platinum group metals and there- fore only a limited qumber of samples can be conveniently handled.The method, as described, is reasonably fast and accurate.1054 (1) Platinum Analysis of variance- Source of variation CRUICKSHANK AND MUNRO TABLE I11 STATISTICAL COMPARISON OF RESULTS Sum of squares Laboratories 0.084 Samples .. .. . . .. . . 1274.136 Residuals . . .. .. .. .. 9.971 Total . . . . .. .. .. .. 1284.191 (X-ray method and laboratories 1 and 2) Contrasts- Laboratory 1 versus laboratory 2 . . .. . . . . laboratory 2 . . . . . . .. . . .. X-ray method versus mean of laboratory 1 and (2) Palladium Analysis of variance- Sum of Source of variation squares Laboratories 1.553 Samples .. .. .. . . . . 222.673 Residuals . . .. .. .. . . 1.044 Total . . .. .. .. .. .. 226.271 (X-ray method and laboratories 1 and 2) Contrasts- Laboratory 1 versus laboratory 2 . . .. . . laboratory 2 . . .. .. .. .. .. X-ray method versus mean of laboratory 1 and Degrees of freedom 2 24 48 74 t value - 0.636 - 0.008 Degrees of freedom 2 24 48 74 t value 5.387 - 6.508 Mean square F value 0.042 0.202 53.089 255.563 0.207 - - - Significance at 0.05 level Not significant Not significant Mean square F value 0.776 35.696 9.278 426.342 0.021 - - - Significance a t 0.05 level Significant Significant The authors thank Mr. M. Laws, Matthey Rustenburg Refiners, Wadeville, Mrs. B. Fourie and Mr. D. Nicolas for their assistance in the development of the method, Dr. D. Hawkins for his assistance with the statistical evaluation of results and the Management of Johannes- burg Consolidated Investment Company (Pty) Ltd. for permission to publish this paper. References 1. Austen, C. E., and Steele, T. W., “The Determination by X-Ray Fluorescence Spectrometry of Noble and Base Metals in Matte-Leach Residues,” Report No. 1912, Laboratory Method No. 78/26, National Institute for Metallurgy, Johannesburg, September 1977. Shestakov, V. A., Arkhipov, N. A., Makarov. D. F., and Kukushkin, Yu. N., J. Analyt. Chem. USSR, 1974, 29, 1872. Schoeller, W. R., and Powell, A. R., “Analysis of Minerals and Ores of the Rarer Elements,” Second Edition, Griffin, London, 1940, pp. 240-296. Feather, C. E., and Willis, T. P., X-Ray Spectrom., 1976, 5, 41. Chatfield, C. C . , “Statistics for Technology,” Chapman and Hall, London, 1970. 2. 3. 4. 5. Received March l6th, 1979 Accepted June Sth, 1979

ISSN:0003-2654    

DOI:10.1039/AN9790401050

出版商:RSC    

年代:1979

数据来源: RSC

摘要:

Analyst, November, 1979, Vol. 104, $9. 1055-1061 1055 Determination of Iron(l1) in Rock, Soil and Clay L. Th. Begheijn Department of Soil Science and Geology, Agricultural Universz't-v Wageningen, P.O. Box 37, 6700 A A Wageningen, The Netherlands A rapid and direct method for the determination of iron(I1) in silicates is described. Redox processes frequently occurring during decomposition are suppressed satisfactorily by limiting the reaction time to 10 s while main- taining the temperature at 60-65 "C. Reproducible decomposition tempera- tures are achieved by mixing concentrated sulphuric and hydrofluoric acids (1 + 3 V / V ) in the reaction vessel. The coloured iron(I1) - 1,lO-phenanthroline complex is used in the spectrophotometric determination of the two valency states of iron, iron(I1) directly and iron(II1) by difference after subsequent reduction by hydroquinone.Mean results of duplicates of the USGS geo- logical standards G-2, AGV-1 and BCR-1 are within 0.1% and those for DTS-1 and PCC-1 within 0.3% of the quoted average values for iron(I1) oxide. Keywords : Iron(II) determination ; silicates ; hydrofluoric acid decomposition ; spectrophotometry The determination of iron(I1) in mineral material, including clay fractions of soils, suffers from many possible sources of error. Apart from incomplete decomposition, interferences frequently reported are due to the reduction of iron(III), mainly by organic matter and sulphides. The principal sources of errors in iron( 11) determinations are the mechanical dry grinding of samples in air and oxidation during decomposition.Pruden and Bloomfieldl reported the effect of organic matter and concluded that its presence vitiates the deter- mination of iron(I1). Mitsuchi and Oyama2 lowered the temperature of the decomposition to less than 80 "C in order to overcome reduction of iron(II1) by organic matter. Kiss3 showed that air oxidation may cause appreciable loss of iron(I1) oxide. Clemency and Hagner4 considered an inert atmosphere unnecessary; however, they used a system in which both reduction and oxidation may have occurred. Replacement of air by nitrogen or carbon dioxide is often advocated for eliminating air oxidation. French and Adams5 reported that the hydrogen fluoride vapour evolved by the pre-mixed equal volumes of concentrated hydrofluoric and sulphuric acids with silicates effectively prevents air oxidation.Oxidation during sample preparation or pre-treatment may also lower the apparent iron( 11) content. French and Adams5 found that grinding rock samples for 10 min with continuous moistening with acetone produced no detectable oxidation. Removal of organic matter from soil samples by hydrogen peroxide does not seem to affect the iron(I1) in the minerals directly but may cause the pH of the suspension to become very low, resulting in degradation, especially in the 2 : 1 clay minerals6 During this degrada- tion iron(I1) fixed in a non-exchangeable form within aluminium interlayers may be removed and ~xidised.~ These effects may be suppressed by the addition of a 1 M sodium acetate buffer (pH 5) as suggested by Jackson8 for the removal of carbonates before hydrogen peroxide treatment .6s7 This study, partly an extension of the work of Clemency and Hagr~er,~ was made in order to investigate whether interfering oxidation and reduction processes can be eliminated by limiting the time and temperature of digestion while, a t the same time, achieving complete decomposition.Experimental US Geological Survey StandardsgJo of powdered rock with a wide range of iron(I1) oxide contents were selected for the experiments. The decomposition procedure recommended by Clemency and Hagner4 was modified by limiting the time of reaction and avoiding external1056 BEGHEI J N : DETERMINATION OF Analyst, Vol. I04 heating. A temperature of 60-65 "C is reached spontaneously if 0.10 g of silicate sample is mixed successively with 1.0 ml of sulphuric acid (sp.gr. 1.84) and 3.0 ml of hydrofluoric acid (sp. gr. 1.19). The subsequent 2-min boiling step in the presence of boric acid was examined for the effect of complete dissolution of iron after disintegration of the silicate structure (Table I). Separate samples of some of the standard rocks were each decomposed for different periods of time. After optimum conditions of decomposition were evaluated the effect of organic matter was investigated by treating the selected geochemical standards in the presence of 25 mg of humic acid (Fluka, practicum grade, relative molecular mass 500-1 000; Table 11). The colour development of iron(I1) with 1 ,lo-phenanthroline was tested under experi- mental conditions.Separate solutions of ammonium iron(I1) sulphate and ammonium iron(II1) sulphate (analytical-reagent grade) and a mixture of both were transferred into silica beakers containing 10 ml of boric acid (4% m/V), 1.0 ml of hydrochloric acid (sp. gr. 1.16), 1.0ml of sulphuric acid (sp. gr. 1.84) and 3.0ml of hydrofluoric acid (sp. gr. 1.19), and boiled for 2 min. The contents of the silica beakers were then transferred into 100-ml polypropylene calibrated flasks, containing 40 ml of boric acid, diluted to volume and filtered. Aliquots of 4.00 ml were pipetted into 100-ml calibrated flasks that contained 20 ml of buffer solution and 8.0 ml of 1,lO-phenanthroline solution (see Reagents), diluted with water to volume and mixed. After the addition of 1 ,lo-phenanthroline solution absorbances of these solutions at 510nm were measured after periods of 0, 3, 5 , 10, 15, 23 and 30min.About 25mg of hydroquinone were then added to the remainder of each solution (about 50ml were left). The rate of colour development, after reduction of iron(II1) by the hydroquinone, was studied by measuring the absorbances at 5-min intervals from 0 to 30min after the hydroquinone addition. Apparatus Platinum - rhodium (95 + 5 ) crucibles of capacity 30 ml. Fwne cupboard. Silica beakers, 150 ml. PolyProPylene calibrated Jlasks, 100 ml. Vitatron Universal Photometer UFD with 1-cm cell. Reagents water was used to prepare all solutions. All reagents, except for hydroquinone, were of analytical-reagent grade, and de-ionised HydroJluoric acid, s#.gr. 1.19. Sulphuric acid, sp. gr. 1.84. Hydrochloric acid, sp. gr. 1.16. Boric acid solution, 4% m/V (saturated). Iron stock solution, 1.000 g 1-1 of iron. Iron standard solution, 0.010 g 1-1 of irorc(l1I) oxide. Potassium hydrogen phthalate bufer solution, 0.5 M, pH 4.1. 1 ,lO-Phenanthroline hydrochloride (monohydrate) solution, 0.25y0 m/V. Hydroquin one. Dissolve 40 g of boric acid in 1 1 of water. Dissolve 1 .OOO g of iron powder (Merck) in 25 ml Dilute 7.00 ml of the iron stock Dissolve 100 g of potassium Dissolve 0.25 g of Prepare freshly each day (solutions become coloured of 4 M hydrochloric acid. Dilute with water to a volume of 1 1. solution with water to a volume of 1 1. hydrogen phthalate in 1 1 of water and heat gently. C,,H,N,.HCl.H,O in 100 ml of water.with age). Procedure mill (TEMA). The standard USGS rock samples passed a 150-pm sieve without previous powdering. Powder soil samples by grinding for 30 s in a mechanical tungsten carbide ball (or ring) A finely powdered sample is obtained of which 80% is less than 50pm. Freeze-dried samples of clay fractions may be used after light grinding with a pestle andNovember, 1979 IRON(II) IN ROCK, SOIL AND CLAY 1057 mortar. Samples used for separation of clay fractions should be treated with hydrogen peroxide buffered at pH 5 to remove most of the organic matter8 and kept between pH 5 and 7.5 during dispersion and separation. Weigh accurately, to within 0.1 mg, about 100 mg of sample into a platinum crucible. Add 1.0 ml of sulphuric acid and homogenise.Add, from a 10-ml polypropylene measuring cylinder, 3.0 ml of hydrofluoric acid. Caution-Appropriate safety measures should be taken. Swirl gently for 10 s to allow the temperature to rise to 60-65 "C. Copious vapours now blanket the mixture. Transfer quickly into a silica (see Note 1) beaker that contains 10 ml of boric acid solution (which inactivates part of the hydrofluoric acid) and 1.0 ml of hydro- chloric acid. (Wear gloves throughout this last part of the procedure and work in a fume cupboard). Cover the beaker with a polypropylene cover, heat and boil gently for 2 min in order to dissolve the decomposition products. Cool, transfer into a 100-ml polypropylene (see Note 1) calibrated flask, which contains another 40 ml of boric acidsolution (see Note 2).Dilute to volume with water, homogenise and filter. Prepare a similar blank solution for each batch of analyses. NOTES- 1. 2. Standard laboratory glassware (Jena glass) may release appreciable amounts of metals (sodium, This additional boric acid is added in order to obtain an excess over the hydrofluoric acid. calcium, barium and aluminium). This would make the filtrate unusable for further elemental analysis. Determination of iron(II) Pipette 2.00 ml of the filtrate into a 50-ml calibrated flask. Caution-Do not pipette orally. Add 10 ml of the buffer solution and 4.0 ml of 1,lO-phenanthroline solution. If highest accuracy (1% relative or better) is desired, the solution can be made 0.01 M in nitrilotriacetic acid (NTA) as suggested by Fadrus and Malq.ll Immediately (or within 10 min) measure the absorbance at 510 nm against water as a reference.Keep the remainder of the solution for the determination of total iron (see below). If the filtrate is coloured by organic matter, a duplicate solution without phenanthroline should be used as a blank (check that the absorbance of phenanthroline is near to zero). Dilute with water to a volume of about 30 ml. Dilute to volume and mix. Determination of total iron 40ml) and homogenise. absorbance as for the iron(I1) determination. With a spatula add about 20 mg of hydroquinone to the remainder of the solution (about Leave for a period between 10 and 30min and measure the Calibration graph solution into separate 50-ml calibrated flasks. 1,lO-phenanthroline solution and 25 mg of hydroquinone.and homogenise. 20 mg of hydroquinone. concentration. make readings of absorbance less accurate. Pipette 0.0, 10.0, 20.0 and 30.0ml of the standard iron solution and 2.0ml of blank Add 10ml of buffer solution, 4.0ml of Dilute to volume with water Read the absorbance at 510 nm after 30 min against water as a reference. The absorbance of the blank solution should be read before and after the addition of Construct a calibration graph of absorbance versus iron(II1) oxide More concentrated solutions The graph is linear up to 5 mg 1-1 of iron. Results and Discussion The effect of the time of decomposition on the recoveries of iron(I1) and total iron [iron(II) As can be seen, the decomposition is complete within Extended tests with the geological standard AGV-1 show complete recovery of As shown in Table I, boiling is clearly essential for the Fig.2(a) shows that in this procedure, plus iron(III)] is shown in Fig. 1. 1 min. total iron with 15 s of swirling. complete dissolution of the decomposition products.1058 BEGHEI JN : DETERMINATION OF Analyst, Vol. 104 formation of the iron(I1) - 1 ,lo-phenanthroline complex is practically instantaneous; subse- quent addition of hydroquinone has no further effect. A negligible effect of iron(II1) at very low iron(I1) concentration is indicated in Fig. 2 ( b ) . The maximum absorbance after reduction of iron(II1) (after addition of hydroquinone) is reached in 5 min (Fig. 2 ( b ) ] . An effect of iron(II1) is seen in Fig. 2(c) in the presence of a moderate iron(I1) concentration. Seconds swirling Minutes on hot-plate Time of decomposition Fig.1. Measured iron(I1) and total iron contents as a function of time of decomposition. Solid lines : contents according to USGS (1972); A, iron(I1) plus iron(II1); and 0, iron(I1). Positive errors in the iron(I1) values may increase by up to 5% relative after 30 min standing, or 2% relative after 10min. According to Fadrus and Mal$ll this could be eliminated by addition of NTA. Results for iron(I1) and total iron (I1 plus III), as compared with USGS standards, are shown in Table 11. Recoveries of iron(I1) oxide are good for the felsic, intermediate and mafic standard rocks (G-2, AGV-1 and BCR-1, respectively), but low (97-98y0) for two ultramafic rocks (DTS-1 and PCC-1). Still lower recoveries (92-96y0) of total iron were found in the ultramafic samples indicating that iron(II1) especially, in the samples, is only partially liberated by the proposed procedure. Solid residues, after hydrofluoric acid - sulphuric acid treatments, amounting to 2 and 10% of the original mass of DTS-1 and PCC-1, respectively, were observed. Longer decomposition times (up to 1 min) or longer boiling (5 min) hardly increased the measured iron(II1) oxide content and had no effect on iron(I1) oxide.Enstatite and a nickel - chromium - iron spinel were identified by X-ray diffraction in the solid residues of PCC-1 and DTS-1, respectively. This indicates that this method is less suitable for measuring total iron in ultramafic rocks. In soil and clay samples with low iron(I1) oxide contents (0.1-0.4~0) analysed in this laboratory, the standard deviation is about 0.05-0.06% .7 TABLE I EFFECT OF BOILING ON RECOVERY OF IRON(II) OXIDE AND IRON(III) OXIDE Sample 100 mg of rock, 10 s decomposition in hydrofluoric acid - sulphuric acid. Iron(I1) oxide found, yo Iron(II1) oxide found, y-, -- With Rock boiling boiling boiling boiling BCR-1 .. . . 6.59 8.94 3.76 3.68 AGV-1 . . . . 1.55 2.09 4.34 4.35 (3-2 .. . . 1.22 1.42 1.09 0.96November, 1979 IRON(II) IN ROCK, SOIL AND CLAY 1059 The results of these experiments demonstrate that the problems with the usual recom- mended longer decomposition periods, with external heating, are avoidable. The disintegra- tion of most silicates is virtually instantaneous. This also applies to a wide range of powdered clay fractions examined in this laboratory, as checked against the determination of total iron by X-ray fluorescence spectrometry.The minimum period needed for adding reagents and transferring products into the silica beaker is about 15 s. A short time of reaction appears essential for suppressing interfering. factors. (Fig. 1 shows that oxidation starts after 10 s of decomposition, possibly due to oxidising constituents in the sample.) Although the period of mixing (swirling) should be as short as possible, it should still be reproducible and should ensure adequate mixing. Thc 10-s period of swirling appears to satisfy these con& tions. hydroqu inone hydroquinone I I I I I c 0 20 40 60 0 20 40 Timehin I Without I With - hydroquinone I Fig.2. Formation of iron(I1) - l,l0-phenanthroline complex in the presence of iron(II1) before and after the addition of hydroquinone : (a), concentration of iron(I1) 1.7 mg l-l, iron(II1) nil; (b), concentration of iron(I1) nil, iron(II1) 1.4 mg 1-l; and (c), concentration of iron(I1) 1.7 mg l-l, iron(II1) 1.4 mg 1-I. The sequence of addition of the acids may also influence oxidation. Additions of the acids in the order sulphuric and then hydrofluoric acid to a mixture of iron(I1) and iron(II1) salts resulted in almost complete recovery (96.4y0), whereas after addition in the reverse order (hydrofluoric then sulphuric acid), only 89.4% of iron(I1) was recovered (Table 111). The sequence of addition did not cause significant differences in results for rock BCR-1 or for silicate material.Apparently, oxidation takes place only after decomposition, whereas the iron salts tested are dissolved immediately. TABLE I1 RESULTS FOR IRON(I1) OXIDE AND TOTAL IRON [AS IRON(II1) OXIDE] IN INTERNATIONAL GEOLOGICAL STANDARDS Sample 100 mg of rock, 10 s decomposition in hydrofluoric acid - sulphuric acid, 2 min boiling in boric acid - hydrochloric acid. Iron(I1) oxide, yo Total iron as iron(II1) oxide, yo USGS Rock (1969) BCR- 1 . . 8.91 AGV- 1 . . 2.04 G-2 . . 1.44 DTS-1 . . 6.79 PCC- 1 . . 4.94 USGS* (1972) Found S.D.t Recovery 8.80 8.90 0.09 101.1 2.05 2.06 0.09 100.5 1.45 1.48 0.11 102.1 7.23 7.07 0.12 97.8 5.24 5.06 0.03 96.6 bSGS USGS* (1969) (1972) Found S.D.7 Recovery 13.50 13.40 13.67 0.12 102.0 6.80 6.76 6.73 0.14 99.6 2.76 2.65 2.58 0.13 97.4 8.85 8.64 8.31 0.11 96.2 8.53 8.35 7.66 0.03 91.7 m * USGS (1972) used for reference.t S.D. = standard deviation (six determinations per sample).1060 BEGHEI J N : DETERMINATION OF Analyst, Vol. 104 The effects on measured iron(I1) contents of even large amounts of organic matter (10 and 25 mg) appear to be small (Table IV); there is a slight rising trend, but none of the differences are significant at P = 0.05. At lower organic matter contents, as reported by Brinkman,' the measured iron(I1) oxide content in clay after removal of organic matter was within 0.06% of the value. TABLE I11 SPECTROPHOTOMETRIC DETERMINATION OF IRON(II) IN THE PRESENCE OF IRON(III) FROM PURE IRON SOURCES 'WITH 1 ,lo-PHENANTHROLINE Absorbances are read within 10 min after adding 1,lO-phenanthroline.Mass of iron(I1) Mass of iron(I1) Recovery, No.* Test takenlmg foundlmg % 1 30.1 mg of Fe(NH,),(SO4),.6H,O 4.29 4.27 99.5 2 30.9 mg of Fe(NH,)(S04),.12H,0 0 0.07 - 3 4 6 30.1 mg of Fe(NH4),(SO4)2.6HaO + 28.4 mg of Fe(NH,),(S04),.6H,0 + 27.3 mg of Fe(NH4),(SO4),.6H2O + 30.9 mg of Fe(NH4)(S0,)a.12H,0 4.29 4.44 103.5 22.4 mg of Fe(NH,)(SO4),.12H,O 4.04 3.61 89.4 23.5 mg of Fe(NH,)(S0,),.12H20 3.89 3.75 96.4 * 1, 2 and 3. fluoric acid and 1.0 ml of sulphuric acid, boil for 2 min. that contain 40 ml of boric acid. 4. In a platinum crucible. swirl and 2-min boil. 5. In a platinum crucible. and 2-min boil. Dissolve in 10 ml of boric acid plus 1.0 ml of hydrochloric acid, 3.0 ml of hydro- Add to 100-ml polypropylene flasks Add 3.0 ml of hydrofluoric acid and 1:O ml of sulphuric acid; 10-s Add 1.0 ml of sulphuric acid and 3.0 ml of hydrofluoric acid; 10-s swirl Dilute to volume.The widely applied and improved spectrophotometric iron(I1) - 1 ,lo-phenanthroline method, as described by Sandell,12 was slightly modified. The tedious procedure of buffering by citrate was replaced by the addition of a fixed (excess) volume of biphthalate solution of pH 4.1. This brings the pH of the final solution (including foregoing acid treatments) up to 3.1. Interferences of iron(II1) are limited, possibly due to the presence of fluoride (see Note 3), which masks iron(II1) and hence inhibits or slows down the reaction Fe3+ + 3(phen) + e- -+ Fe(phen),2+ NOTE- 3. Complexation by boric acid forms hydrofluoroboric acid (HBF,) givins ions H+ and BF,-; a The latter process proably leaves sufficient F- in solution to little BF4- dissociates into BF, and F-.13 reduce Fe3+ activity to a very low value.The use of hydroquinone may be a problem in the presence of titanium, which may cause an interfering brown colour.12 In this procedure and at the expected concentration of TABLE IV EFFECT OF ORGANIC MATTER ON RECOVERY OF IRON(II) OXIDE AND TOTAL IRON AS IRON(III) OXIDE Sample 100 mg of rock, 10 s decomposition in hydrofluoric acid - sulphuric acid, 2 min boiling in Blank determinations were carried out on 0, 10 and 25 mg of pure humic acid treated in the same way. boric acid - hydrofluoric acid. Iron(I1) oxide found, % Total iron as iron(II1) oxide found, % With 10 mg With 25 mg With 10 mg With 25 mg A I \ r A \ Without of humic of humic Without of humic of humic Rock humic acid acid acid humic acid acid acid BCR-1 .. . . 8.90 8.84 8.95 13.67 13.53 13.73 AGV-1 . . . . 2.06 2.01 2.21 6.73 6.67 6.63 G-2 .. . . 1.48 1.52 1.61 2.58 2.56 2.79November, 1979 IRON(II) IN ROCK, SOIL AND CLAY 1061 titanium (less than 1 mg 1-1 of titanium oxide) this interference appears to be of no practical importance. During the analysis of pyrite-containing soils (acid sulphate soils), iron(I1) disulphide (pyrite) is not included in the measured iron(I1) value because it is resistant to the applied concentrated acids. In fact, the latter residue is available for analysis of pyrite iron(I1) by a subsequent nitric acid di~solution.~~ An oxidimetric titration, as used by French and ad am^,^ for standard rock samples would obviate the need for filtration, but does not seem to be suitable for soils and clay fractions, because of possible interference by organic matter.Conclusion Analysis of iron(I1) in rocks, soils and clay fractions contairling some organic matter is subject to many interfering factors, which are somewhat unresolved in the literature. This paper evaluates recent published work and recommends a highly reproducible, rapid and simple method that does not require specialised apparatus. The proposed method is widely applicable to soils and clays and to most rocks; some reservation must be made, however, for the group of the ultrabasic crystalline rocks. The author thanks J. D. J. van Doesburg for X-ray diffraction analyses and N. van Breemen and R. Brinkman for their critical comments on a draft of this paper. A. Breeuwsma and R. A. Koning (Soil Survey Institute, Wageningen) are thanked for supplying the rock samples and for the useful discussions on a modified method designed for materials with high organic matter contents. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 4. References Pruden, G., and Bloomfield, C., Analyst, 1969, 94, 688. Mitsuchi, M., and Oyama, M., J . Sci. Soil Manure, Japan, 1963, 34, 23. Kiss, E., Analytica Chim. Acta, 1977, 89, 303. Clemency, C. V., and Hagner, A. F., Analyt. Chem., 1961, 33, 888. French, W. J., and Adams, S. J., Analyst, 1972, 97, 828. Douglas, L. A., and Fiessinger, F., Clays Clay Miner., 1971, 19, 67. Brinkman, R., Geoderma, 1977, 17, 111. Jackson, M. L., “Soil Chemical Analysis, Advanced Course,” published by the author, Department Flanagan, F. J., Geochim. Cosmochim. Acta, 1969, 33, 81. Flanagan, F. J., Geochim. Cosmochim. Acta, 1972, 37, 1189. Fadrus, H., and Malg, J., Analyst, 1975, 100, 549. Sandell, E. B., “Colorimetric Determination of Traces of Metals,” Third Edition, Interscience, New Mellor, J , W., “A Comprehensive Treatise on Inorganic and Theoretical Chemistry,” Volume V, Begheijn, L. Th., van Breemen, N., and Velthorst, E. J., Commun. Soil Sci. Plant Afial., 1978, 9, of Soil Science, University of Wisconsin, Madison, Wisc., 1956 (Fifth Printing 1969). York, 1959. 1946, p. 125. 873. Received April 17th, 1978 Amended June 4th, 1979 Accepted June 18th, 1978

ISSN:0003-2654    

DOI:10.1039/AN9790401055

出版商:RSC    

年代:1979

数据来源: RSC