摘要:

THE ANALYSTTHE ANALYTICAL JOURNAL OF THE CHEMICAL SOCIETYEDITORIAL ADVISORY BOARD'Chairman: H. J. Cluley ( Wembley)'L. S. Bark (Salford)R. Belcher (Birmingham)L. J. Bellamy, C.B.E. (Waltham Abbey)L. S. Birks (U.S.A.)E. Bishop (Exeter)L. R. P. Butler (South Africa)E. A. M. F. Dahmen (The Netherlands)A. C. Docherty (Billingham)D. Dyrssen (Sweden)J. Hoste (Belgium)H. M. N. H. Irving (feeds)M. T. Kelley (U.S.A.)W. Kernula (Poland)' G . F. Kirkbright (London)G . W. C. Milner (Harwell)'W. T. Elwell (Birmingham)'J. A. Hunter (Edinburgh)G. H. Morrison (U.S.A.)H. W. Nurnberg (W. Germany)"J. M. Ottaway (Glasgow)'G. E. Penketh (Wilton)'T. B. Pierce (Harwell)E. Pungor (Hungary)D. 1. Rees (London)"R. Sawyer (London)P. H. Scholes (Sheffield)'W.H. C. Shaw (Greenford)S. Siggia (U.S.A.)A. A. Smales, O.B.E. (Harwell)A. Walsh (Australia)T. S. West (Aberdeen)A. L. Wilson (Medmenham)P. Zuman (U.S.A.)'A. Townshend (Birmingham)'Members of the Board serving on The Analyst Publications CommitteeREGIONAL ADVISORY EDITORSDr. J. Aggett, Department of Chemistry, University of Auckland, Private Bag, Auckland, NEWDr. G. Ghersini, Laboratori CISE, Casella Postale 3986, 201 00 Milano, ITALY.Professor L. Gierst, Universit6 Libre de Bruxelles, Facult6 des Sciences, Avenue F.-D. Roosevelt 50,Professor R. Herrmann, Abteilung fur Med. Physik., 63 Giessen, Schlangenzahl 29, W. GERMANY.Professor W. A. E. McBryde, Dean of Faculty of Science, University of Waterloo, Waterloo, OntarioOr. W.Wayne Meinke, KMS Fusion Inc., 3941 Research Park Drive, P.O. Box 1567, Ann Arbor,Dr. 1. Rubeska, Geological Survey of Czechoslovakia, Kostelni 26, Praha 7, CZECHOSLOVAKIA.Dr. J. R i i i i k a , Chemistry Department A, Technical University of Denmark, 2800 Lyngby, DENMARK.Professor K. Saito, Department of Chemistry, Tohoku University, Sendai, JAPAN.Dr. A. Strasheim, National Physical Research Laboratory, P.O. Box 395, Pretoria, SOUTH AFRICA.ZEA LA N D.Bruxelles, BELGIUM.CANADA.Mich. 481 06, U.S.A.Published by The Chemical SocietyEditorial: The Director of Publications, The Chemical Society, Burlington House,London, W I V OBN. Telephone 01 -734 9864. Telex No. 268001Advertisements: Advertisement Department, The Chemical Society, Burlington House, Piccadilly,London, W I V OBN. Tedephone 01 -734 9864Subscriptions (non-members): The Chemical Society, Distribution Centre, Blackhorse Road,Letchworth, Herts., SG6 1 HNVolume 103 No 1228@ Ths Chemical Society 1978July 197

ISSN:0003-2654    

DOI:10.1039/AN97803FX025

出版商:RSC    

年代:1978

数据来源: RSC

摘要:

ANALAO 103 (1 228) 673-784 (I 978)ISSN 0003-2654July 197867370571 4723728734745754759766768770772775.776780784THE ANALYSTTHE ANALYTICAL JOURNAL OF THE CHEMICAL SOCIETYCONTENTSREVIEW.Simultaneous Determination of Arsenic, Antimony, Bismuth, Selenium and Tellurium in AqueousSolution by Introduction of the Gaseous Hydrides into an Inductively Coupled Plasma Sourcefor Emission Spectrometry. Part II. Intterference Studies-M. Thompson, B. Pahlavanpour,S. J. Walton and G. F. KirkbrightContributions t o Automated Trace Analysis. Rapid Method for the Automated Deter-mination of Lead in Whole Blood by Electro'thermal Atomic-absorption Spectrophotometry-M. Stoeppler, K. Brandt and T. C. RainsComparative Laboratory Evaluation of Some Reported Methods for the Determination of DDTand BHC lnsecticides in Human Blood-R.C. Gupta, A. B. Karnik, S. K. Nigam and S. K. KashyapQuantitative Analysis of Active Substances in Liquid Pharmaceutical Preparations Using High-performance Liquid Chromatography-Monir Amin and Peter W. SchnelderProblems of Accuracy and Precision in the Determination of Trace Elements in Water as Shownby Recent International Atomic Energy Agency lntercomparison Tests--R. Dybczyfiski, A.Tugsavul and 0. SuschnyAnalytical Pyrolysis in Biomedical Studies-W. J. Irwin and J. A. SlackPart I I .Spectrophotometric Determination of Boron in Natural Waters by Solvent Extraction withChromotropic Acid and Removal of the Excess of Reagent-Takashi Korenaga, Shoji Motomizuand Kyoji TBeiCause and Control of Losses of Chromium During Nitric - Perchloric Acid Oxidation of AquaticContinuous Colorimetric Monitoring of Vapour-phase Urea and Cyanates-L.E. Trimble and R. J. H.Sediments-Keith ScottVoorhoeveSHORT PAPERSDetermination of Thallium by Precipitation of TIiallium(l) Chromate from Homogeneous SolutionResponse t o Calcium Ions of the Calcium Fluoride Crystal-membrane Electrode-Malathi Ghosh,Comparison of Methods for the Determination of Residues of Organophosphorus Pesticides i nRelationship Between Heat of Combustion and Elemental Composition i n Some Biological Mole-by Replacement of Thallium(1) in a Complex with EDTA-K. N. UpadhyayaM. R. Dhaneshwar, R. G. Dhaneshwar and B. GhoshSamples Having incurred Residues-N. A.Srriart, A. R. C. Hill and Patricia A. Roughancules-J. D. Pryce and G. F. EnfieldCO M MU NICATIONSPacked Flow-through Cell Technique for the Measurement of the Room-temperature Phosphor-Mechanism of Atomisation of Molybdenurn in Carbon Furnace Atomic-absorption Spectrophoto-Book ReviewsErratum: Solid-state Mercury(1) Chloride Electrode for Determining 0.01-1 .O wg ml-1 Levels ofescence Spectra of Adsorbed Compounds-Ji. B. F. Lloydm e t r y d . Sneddon, J. M. Ottaway and W. B. RowstonChloride in Boiler Water and Other High-purity Waters-G. B. Marshal! and D. MidgleySummaries o f Papers in this Issue-Pages iv, vi, vii, x, xiiPrinted by Heffers Printers Ltd Cambridge EnglandEntered as Second Class at New York, USA, Post OfficJOURNALS BOOKSMONOGRAPHSOrders for all publicationsformerly published by the Societyfor Analytical Chemistry shouldbe sent direct or through abookseller toTHE CHEMICAL SOCIETY,Distribution Centre,Blackhorse Road, Letchworth.Herts., SG6 1 HNSelected Annual Reviewsof the Analytical SciencesVolume 4CONTENTS'Advances in Voltammetric Techniques,' byB.Fleet and R. D. Jee'High-frequency Electrodeless PlasmaSpectrometry,' by B. L. SharpPp. vi + 73 f 9.50ISBN 0 85990 204 8CS Members' price f3.00Orders should be sent direct, with remittance, orthrough your usual bookseller to-THE CHEMICAL SOCIETYDistribution Centre,Black horse Road, Letc hwort h,Herts. SG6 IHN, EnglandCS Members must write direct to the above addressenclosing the appropriate remittance.NOTICE TO SUBSCRIBERS(other than Members of the Society)Subscriptions for The Analyst, Analytical Abstracts and Proceedings shouldbe sent to:The Chemical Society, Distribution Centre,Blackhorse Road, Letchworth, Herts., SG6 1 HNRates for 1978The Analyst, Analytical Abstracts and Proceedings (including indexes) :, .f99.00Proceedings . . . . . . . . . . . . . . . . . . €105.00( a ) The Analyst, Analytical Abstracts and Proceedings . . . . ..(b) The Analyst, Analytical Abstracts printed on one side of the paper, andThe Analyst and Analytical Abstracts without Proceedings (including indexes) :The Analyst, and Analytical Abstracts printed on one side of the paper(c) The Analyst, and Analytical Abstracts . . .. * . . . .. . . f87.00(d) . . f93.00(Subscriptions are NOT accepted for The Analyst and/or for Proceedings alone)Analytical Abstracts only (two volumes per year, including indexes) :( e ) Analytical Abstracts . . .. .. . . .. .. .. . . f67.00(f) Analytical Abstracts printed on one side of the paper . . * . . . , . f73.0

ISSN:0003-2654    

DOI:10.1039/AN97803BX027

出版商:RSC    

年代:1978

数据来源: RSC

摘要:

THE ANALYST iiiTREATISE ON ANALYTICAL CHEMISTRY Part 2: AnalyticalChemistry of Inorganic and Organic Compounds VoI. I 0edited by I. M. Kolthoff, University of Minnesota; P. J. Elving, University of Michigan; and R. S.Braman, University of South Florida.This is the tenth volume in the Treatise on Analytical Chemistry series. Designed as a complete anddefinitive source of information for all analytical chemists the many volumes present a concise,critical, comprehensive and systematic treatment of all aspects of the subject.0471 49998 6 586 pages April 1978 &24.40/$33.8 5SURFACTANTS AND INTERFACIAL PHENOMENAby M. J. Rosen, City University of New Yorlc.Examines surfactants and their role in modifying interfacial phenomena. It surveys the structuraltypes of commercially available surfactants and examines, interfacial phenomena, the physico-chemical principles underlying the action of surfactants in each phenomenon, and the effect ofstructural changes in the surfactants and environment changes on their action.0471 73600 7 approx.288 pages In Press approx. f14.10/$25.40TECHNIQUES OF CHEMISTRY Vol. 12: Separation andPurification 3rd Ed. Part 1Series Editor: Arnold Weissberger; Volume Editor: E. S. Perry, Eastman Kodak Co., Rochester,New York.Provides the theoretical background for an understanding of the various methods and operations,and describes the techniques and tools, their modifications, merits and limitations, and theirhandling.0471 02655 7 approx. 400 pages In Press approx. &19.40/$34.95MEASUREMENT OF DISSOLVED OXYGENby M.L. Hitchmann, RCA Laboratories, Zurich, Switzerland.A critical assessment of the various aspects of dissolved oxygen measurements that uses a “why”rather than a “how” approach to emphasise physiochemical principles. This book illustrates themonitoring of dissolved oxygen and discusses various situations where a knowledge of oxygenconcentration is necessary. (Chemical Analysis Series)0471 03885 7 approx. 400 pages In Press approx. E 19.40/$34.90Available from allgood booksellers or from Wiley. I f you wish to use American Express, Diners Club,Barclaycard or Access, please quote your card and numberi V SUMMARIES OF PAP.ERS I N THIS ISSUE July, 1978Summaries of Papers in this IssueAnalytical Pyrolysis in Biomedical StudiesA ReviewSummary of ContentsIntroductionTechniquesPyrolysis methodsDetection methodsGas chromatographyMass spectrometryQualitative and quantitative aspectsApplicationsDrugsBiomacromoleculesAmino acids, peptides and proteinsCarbohydratesNucleic acidsLipids and other compoundsTaxonomyBacteriaFungiMiscellaneousPathologyGeological materials of biological originMechanismsData handlingConclusionKeywords : Pyrolysis - gas claroinatograplay ; pyrolysis - mass spectrcmetry ;biochemical analysis ; taxonomy ; biological analysisW.J. IRWIN and J. A. SLACKDepartment of Pharmacy, University of Aston in Birmingham, Gosta Green,Birmingham, B4 7ET.Analyst, 1978, 103, 673-704.Simultaneous Determination of Trace Concentrations of Arsenic,Antimony, Bismuth, Selenium and Tellurium in Aqueous Solutionby Introduction of the Gaseous Hydrides into an Inductively CoupledPlasma Source for Emission SpectrometryPart 11.Interference StudiesIn the simultaneous determination of arsenic, antimony, bismuth, seleniumand tellurium by measurement of the a.tomic line emission from the elementsin an inductively coupled plasma source after their introduction as the gaseoushydrides, two types of possible interferences were investigated, Mutualinterferences between the analytes were found to be negligible under theoperating conditions used. The presence of certain metal ions in solutionresulted in low recoveries of some of the analytes, especially selenium andtellurium.However, this could be overcome by prior separation of theanalytes by co-precipitation en lanthanum hydroxide.Keywords 1 Arsenic, antimony, bismuth, selenium and tellurium determination ;geochemical samples ; hydride gevzeration ; inductively coupled plasmaemission spectrometry ; interferencesM. THOMPSON, B. PAHLAVANPOUR and S. J. WALTONApplied Geochemistry Research Group, Department of Geology, Imperial College,London, SW7 2BP.and G. F. KIRKBRIGHTDepartment of Chemistry, Imperial College, London SW7 2AY.Analyst, 1978, 103, 705-713vi SUMMARIES OF PAPERS I N THIS ISSUEContributions to Automated Trace Analysis. Part 11. Rapid Methodfor the Automated Determination of Lead in Whole Blood byElectrothermal Atomic-absorption SpectrophotometryA simple, rapid method is described for the determination of lead in wholeblood by means of electrothermal atomic-absorption spectrophotometry.Aliquots (25-200 pl) of fingerprick and venous samples were treated with2 M nitric acid for deproteinisation and matrix modification.After centri-fuging, the supernatant was taken for -the automated analysis for lead. Theaccuracy of the method was checked -with independent methods and foundto be satisfactory. Thus it was established that the accuracy obtainableof <30% lies almost within the confid,ence intervals (9 = 0.05) of precision.An evaluation of 282 pairs of randomly selected routine samples indicated anacceptable precision : the relative standard deviation for, for example, thenormal level of 100 pg1-1 is 8.4%, and for an elevated level of 480 pgl-l is3.47%.The shortened temperature programme makes up to 370 measure-ments per clay possible and the computer coupling permits an immediate dataevaluation during occupational and screening programmes and makes possibleround-the-clock measurements.Keywovds : Whole blood analysis ; lead determination ; automated tvaceJuly, 1978analysis ; electrothermal atomic-abso:vption spectrophotometryM. STOEPPLER and K. BRANDTInstitute of Chemistry, Institute 4, App1it:d Physical Chemistry, Nuclear ResearchCentre, P.O. Box 1913, D-5170 Julich, West Germanyand T. C. RAINSNational Bureau of Standards, Analytical Chemistry Division, Washington, D.C.,USA.Analyst, 1978, 103, 714-722.Comparative Laboratory Evaluation of Some Reported Methods forthe Determination of DDT and BHC Insecticides in Human BloodThe relative efficiency of extraction of DDT, BHC and related insecticidesfrom spiked serum samples by different methods is reported.The insecticidewas extracted from the serum by using hexane or hesane - acetone (9 + 1)and then determined in the extract by gas - liquid chromatography using anelectron-capture detector. The highest recoveries were obtained when theserum was treated with formic acid before extraction, lower values wereobtained when sulphuric acid was used in place of formic acid and the lowestvalues were obtained when the untreated serum was extracted. Recoveriesin general were higher when the hexarie - acetone mixture was used in placeof hexane. The order of extractability was y-BHC > a-BHC > /3-BHC >pp‘-DDE > p$’-DDT > o$’-DDT.Keywords : Insecticide determination ; BHC determination ; DDT determina-tion ; human blood analysis ; gas - liquid chromatographyR. C.GUPTA, A. B. KARNIK, S. K. NIGAM and S. K. KASHYAPNational Institute of Occupational Health, Ahmedabad-380 016, Gujarat, India.Analyst, 1978, 103, 723-727July, I 9 78 SUMMARIES OF PAPERS I N THIS ISSUEQuantitative Analysis of Active Substances in Liquid PharmaceuticalPreparations Using High-performance Liquid ChromatographyviiHigh-performance liquid chromatography has been applied to pharmaceuticalpreparations containing the active ingredients in aqueous, alcoholic or oilysolution.A comparison is made of the high-performance liquid chromato-graphic method with gas and thin-layer chromatography.Keywords : High-performance liquid chromatography ; liquid pharmacmticalpreparations ; quantitative analysis of drugs ; quantitative analysis ofsteroidsMONIR AMIN and PETER W. SCHNEIDERSchering A.G., Department Galenik, Department Allgemeine Physikochemie,Mullerstrasse 170-1 78, 1000 Berlin 65, Germany.Analyst, 1978, 103, 728-733.Problems of Accuracy and Precision in the Determination of TraceElements in Water as Shown by Recent International AtomicEnergy Agency Intercomparison TestsResults of a recent International Atomic Energy Agency intercomparisontest on the determination of 16 trace elements in simulated fresh watersamples are reported and discussed from the standpoint of accuracy andprecision.Despite the considerable spread of results obtained from variouslaboratories, the over-all mean (after rejection of outliers) in most instancesapproximates closely to the true value. The importance of using sufficientlysensitive tests for the detection of outlying results is emphasised and theeffectiveness of concurrent use of several criteria for that purpose is demon-strated. The relative frequency of employment of various analyticaltechniques is discussed and an attempt to compare their accuracy is presented.Keywords 1 Water analysis ; trace-element determination ; accuracy ; precision ;outlier testsR. DYBCZYNSKI, A. TUGSAVUL and 0.SUSCHNYInternational Atomic Energy Agency, Laboratory Seibersdorf, A- 101 1 Vienna,Austria.Analyst, 1978, 103, 734-744.Spectrophotometric Determination of Boron in Natural Waters bySolvent Extraction with Chromotropic Acid and Removal of theExcess of ReagentA method of removing the excess of reagent extracted into an organic phase inthe solvent extraction of metal complex anions with quaternary ammoniumsalts was applied to the solvent extraction of the boron complex anion formedwith chromotropic acid ( 1,8-dihydroxynaphthalene-3,6-disulphonic acid) byusing zephiramine (tetradecyldimethylbenzylammonium chloride). EDTA,chromotropic acid, zephiramine and buffer (pH 3.5) are added to the samplesolution, which is then shaken with 1,Z-dichloroethane. The organic phaseis washed twice with a back-washing solution (0.6 M in sodium chloride, pH9.2) and the absorbance of the organic phase measured in a quartz cell.The boron complex is extracted quantitatively into the 1,2-dichloroethaneand its apparent molar absorptivity in 1,2-dichloroethane is 1.4 x lo41 mol-l cm-l a t 351 nm. Micro and trace amounts of boron in natural watersare determined spectrophotometrically.Keywords ; Boron detevmination ; water analysis ; spectroplaotometry ; solventextraction ; chromotropic acidTAKASHI KORENAGA, SHOJI MOTOMIZU and KYOJI TOE1Department of Chemistry, Faculty of Science, Okayama University, 3- 1-1, Tsushi-manaka, Okayama-shi, 700, Japan.Analyst, 1978, 103, 745-753

ISSN:0003-2654    

DOI:10.1039/AN97803FP057

出版商:RSC    

年代:1978

数据来源: RSC

摘要:

July, 2978 THE ANALYST ix I $fi ANN ARBORSCIENCEVIRUSES AND TRACE CONTAMINANTS IN WATER AND WASTEWATERedited by J. A. Borchardt, University of Michigan, J. K. Cleland, W. J. Redman and G. E. Oliver,Michigan Department of Public Health.This volume emanates from a conference held by the Michigan Section, AWWA, The Michigan WaterPollution Control Association and the Michigan Department of Public Health in January 1977.It examines the chemical aspects, toxicological problems and administrative difficulties involved inestablishing meaninfgul water quality standards for viruses and trace contaminants.0250 401 70 3 264 pages February 1978 f 18.60/$32.45CARBON ADSORPTIONedited by P. N. Cheremisinoff, Consulting ,Engineer, Closter, New Jersey, and F. Ellerbusch, the MITRECorporation, McL ean, Virginia.This definitive study focuses on the practical aspects of carbon adsorption use in air and water pollutioncontrol as well as manufacturing uses.0250 40236 X approx.900 pages In Press approx. f25.15/$43.95WATER CHLORINATION : Environmental Impact and Health Effects Vol. 2edited by R. L. Jolley, Oak Ridge National Laboratory, H. Gorchev, Environmental Protection Agency,Washington, and D. H. Hamilton, Jr., Department of Energy, Washington.Emphasising the latest research in the field, this important volume stresses the risks and benefits of waterchlorination to public health and the environment, analyses its use in wastewater treatment, and summarisesthe latest government regulations on chlorine treatment of drinking water.025040201 7 approx.900 pages In Press approx. f 18.90/$33.00LAGOON INFORMATION SOURCE BOOKby E. J. Middlebrooks, N. B. Jones, M. F. Torpy and J. H. Reynolds, Utah State University,A selected summary of the most pertinent articles and reports available describing the various aspects ofwastewater stabilisation lagoons; more than 650 abstracts cover all facets of the topic. This valuable sourcebook gives data on all lagoon types-aerobic, anaerobic, facultative, mechanically aerated, and combinationsof these categories. It describes many personal experiences with design, operation and maintenance. Bothindustrial and domestic waste treatment is covered. No other source gives this amount of information insuch an organised, easy-to-use form.0250 401 98 3 approx.21 6 pages In Press approx. f 12.60/$22.00Also available:AIR POLLUTION SAMPLING AND ANALYSIS DESKBOOKedited by P. N. Cheremisinoff, Consulting Engineer, Closter, New Jersey, and A. C. Morresi,Hoffman-LaRoche, New Jersey.0250402343 approx. 475 pages In Press approx. f 18.60/$32.45ELECTRON MICROSCOPY AND X-RAY APPLICATIONS TOENVIRONMENTAL AND OCCUPATIONAL HEALTH VoI. Iedited by P. A. Russell, University of Denver, and A. E. Hutchings, Philips Electronics Co. lnc.,Mah wah, New Jersey.0250 40222 X approx. 260 pages In Press approx. f 17.65/$30.90Available from all good booksellers or from Wiley. If you wish to use American Express, Diners Club,Barclaycard or Access, please quote your card and numberx SUMMARIES OF PAPERS IN THIS ISSUECause and Control of Losses of Chromium During Nitric - PerchloricAcid Oxidation of Aquatic SedimentsJuly, I 9 7 8The extraction of chromium from aquatic sediments by using boiling nitric -perchloric acid was studied.The amount of chromium extracted was foundto depend upon the time that heating was continued after the nitric acid hadboiled off and to decrease significantly with prolonged heating. The decreasewas shown to be due to adsorption on to the silica residue, as the lostchromium could be recovered by digestion of the residue with hydrofluoricacid. An extraction procedure in which the time of heating in perchloricacid was carefully controlled gave satisfactory precision for the determinationof chromium, zinc, copper, lead and ciiclmiun1, as well as aclequate recoveriesof added amounts of these metals.Keymards A quatic sediments ; chromium determination ; perchlovic acid ;chromium losses ; heavy inetals dete:vwzinationKEITH SCOTTDepartment of Biochemistry, University of Qucensland, St.Lucia, Queensland,Australia, 4067.Analyst, 1978, 103, 754-758.Continuous Colorimetric Monitoring of Vapour -phaseUrea and CyanatesR method is described for continuously monitoring the concentrations ofurea and cyanates in gases, based on a, modification of a colorimetric methodused for urea in human serum. The method was calibrated with standardsolutions of urea and potassium cyanate. The reproducibility was 2%throughout. Using this method, the vapour pressure of urea was measuredfrom 58 to 113 "C.The vapour pressiire results have an estimated accuracyo f 1 0 % in the range 3 x to 5 x nimHg.Keywords : Urea determination ; cyanate determinution ; absorption ; vapourpressureL. E. TRINBLE and R. J. H. VOORHOEVEBell I,aboratorirs, Murray Hill, N. J. 07974, I K 4 .Analyst, 1978, 103, 759-765.Determination of Thallium by Precipitation of Thallium(1) Chromatefrom Homogeneous Solution by Replacement of Thallium(1) in aComplex with EDTAShort Paperli'eymords : Thallium determination ; gvavimetvy ; chromate $recipitation ;precipitation from homogeneous sohution ; re$lacement techniqueK. N. UPADHYAYAChemistry Department, University of Dar ges Salaam, 'P.O. Box 35061, 1)ar es Salaam,Tanzania.Analyst, 1978, 103, 766-768July, 1978 THE ANALYSTInternal Rotation and InversionXIAn lntroduction to Large Amplitude Motions in MoleculesDavid G.Lister, John N. Macdonald and Noel L. OwenJune/July 1978, x i i + 246pp. fl1.50/$23.75Our knowledge of large amplitude motions within molecules has advanced enormously over the past twenty years or soyet it is often difficult for the non-specialist to gain access quickly to the now very extensive and, in some cases, verydetailed information concerning this subject. This book brings together, in a form accessible to the advanced undergraduateor interested research worker rather than just the specialist in molecular vibrations, information of a general and introduc-tory nature of large amplitude movements of atoms within molecules.0.1 2.452250.5Economic Microbiology Volume 2Primary Products of Metabolismedited by Anthony H.RoseMay/June 1978, xvi + 470pp.. f22.50/$44.00 0.1 2.596552.4Although microbial activity and its end products have for centuries been used in a pragmatic and somewhat haphazard wayin various manufacturing processes, thorough investigation of any of these processes has occurred only recently. Aschemists and microbiologists realised the commercial potential in the exploitation of microbial metabolism, they began totrace the biosynthetic pathways and the possibilities of manipulation and mutation of microbes to increase productivity.This volume is the second in a new series which arose from the Biochemistry of Industrial Micro-organisms, edited byC.Rainbow and A. H. Rose (Academic Press, 1963).The Analysis of Organic Materials -An International Series of MonographsSeries editors: R. Belcher and D. M. W. AndersonNumber 9Aldehydes-Photometric AnalysisVolume 5Eugene Sawicki and Carole R. SawickiApril/May 1978, x i i + 408pp., f22.00/$43.00 0.1 2.620505.1Volume 5 of Aldehydes-Photometric Analysis is the third and last part of a sub-set of volumes dealing with the photo-metric determination of aldehyde precursors through their derived aldehydes. Many different methods of analysis aredescribed, with emphasis on analytical methodologies in the environmental and health fields. Included is a backgrounddiscussion on the precursors in environmental mixtures, the formation of aldehydes from their precursors in the environ-ment and in tissue, the genotoxic and hazardous properties of these aldehydes and their precursors, and the reaction ofthe precursors through their aldehydes with normal physiological and environmental chemicals,Number 13Instrumental Organic Element Analysisedited by R.Belcher1977, x i i + 30Ppp., f 14.00/$27.35 0.1 2.085950.5During the last fifteen years automated elemental organic analysis has expanded to such an extent that it has almostsuperseded manual operations, for certain elements at least. The purpose of this book is to bring together the accumulatedknowledge and experience of investigators who have been dealing first hand with these new methods. To provide someflexibility, two different commercial apparatus for the determination of carbon, hydrogen, nitrogen and oxygen, have beendascribed-t he Perkin-Elmer and the,Carlo-Erba apparatus.Acomprehensive account ofshe applications of the Dohrmanncoulometric apparatus has been included and also the remarkably efficient Merz-Dumas method.Academic PressLondon New York San FranciscoA Subsidiary of flarcourt Brace Jovanovich, Publishers24-28 Oval Road, London NWI, England11 1 Fifth Avenue, New York, NY 10003, USxii SUMMARIES OF PAPERS I N THIS ISSUE July, 1978Response to Calcium Ions of the Calcium FluorideCrystal-membrane ElectrodeShort PaperKeywords : Calcium ion-selective electrode ; calcium j?uoride membrane ;MALATHI GHOSH, M. R. DHANESHWAR, R. G. DHANESHWAR andB. GHOSHBhabha Atomic Research Centre, Trombay, Bombay-400 085, India.crystal-membrane electrodeAnalyst, 1978, 103, 768-770.Comparison of Methods for the Determination of Residues ofOrganophosphorus Pesticides in Samples Having Incurred ResiduesShort PaperKeywords : Organophosphorus pesticide residues ; fruits ; vegetables ; gas -liquid chromatographyN.A. SMART, A. R. C. HILL and PATRICIA A. ROUGHANPlant Pathology Laboratory, Ministry of Agriculture, Fisheries and Food, HatchingGreen, Harpenden, Hertfordshire.Analyst, 1978, 103, 770-772.Relationship Between Heat ad Combustion and ElementalComposition in Some Biological MoleculesShort PaperKeywords : Biological nzolecules ; heat of combustion ; elemental covlzpositionJ. D. PRYCE and G. F. ENFIELDChemical Pathology Department, Ipswich Hospital, Heath Road, Ipswich, IP4 5PD.Analyst, 1978, 103, 772-774.Packed Flow- through Cell Technique for the Measurement of theRoom-temperature Phosphorescence Spectra of Adsorbed CompoundsCommunicationKeywords: Roorn-temperature phosphorescence; packed flow-through cell; tracewater ; adsorbatesJ. B. F. LLOYDHome Office Forensic Laboratory, Gooch Street North, Birmingham, B5 6QQ.Analyst, 1978, 103, 775-776.Mechanism of Atomisation of Molybdenum in Carbon FurnaceAtomic-absorption SpectrophotometryCommunicationKeywords : Molybdenum atomisation ; carbon furnace atomic-absorptionspectrophotometryJ. SNEDDON and J. M. OTTAWAYDepartment of Pure and Applied Chemistry, University of Strathclyde, CathedralStreet, Glasgow, G1 1XL.W. B. ROWSTONDepartment of Chemistry, Glasgow College of Technology, 70 Cowcaddens Road,Glasgow, G4 OBA.Analyst, 1978, 103, 776-779

ISSN:0003-2654    

DOI:10.1039/AN97803BP063

出版商:RSC    

年代:1978

数据来源: RSC

摘要:

JULY 1978 Vol. 103 No. 1228 The Analyst Analytical Pyrolysis in Biomedical A Review W. J. Irwin and J. A. Slack" Studies llepartmciit of Pharmacy, Uiziversity of A stoiz in Hirmiizghamz, Gosta Green, Birmingham, B4 7ET Summary of Contents Introduction Techniques Pyrolysis methods Iletection methods Gas chromatography Mass spectrometry Qualitative and quantitativc aspects Applications Drugs Nioxnacroniolecules Amino acids, peptides and proteins Carbohydrates Nucleic acids Lipids and other conipounds Taxonomy Bacteria Fungi Miscellaneous Pathology Geological materials of biological origin Mechanisms Data handling Conclusion Keywords : Pyvolysis - gas clwomatograplzy ; pyyolysis - mass spectvometry ; hioclieniicnl nvzalysis ; t a x o m m y ; biological analysis Introduction Analytical pyrolysis is a technique that is often applied to aid the analysis of high relative molecular mass or involatile organic samples that are difficult or tedious to assay by con- ventional methods.The methods usually involve a minimum of sample preparation, cover a broad field of application and are essentially concerned with the transformation of the sample into another substance or substances through the agency of heat a1one.l The pyrolysis products, or a portion thereof, are generally compounds of low relative molecular mass produced by thermal fission of the original molecule. These products are usually more amenable to analysis and a study of such fragments may yield valuable qualitative and quantitative information concerning the initial material.Although analytical pyrolysis of biological samples, with olfactory detection, has been credited with Biblical origins,2 much of the published work has been concerned with synthetic polymers and a variety of methods for pyrolysis and detection have been used, which has initiated many reviews on various aspects of this subject. In particular, a comprehensive monograph concerned with pyrolysis - gas chromatography3 and the published proceedings of the Third International Symposium on Analytical Pyrolysis* provide contemporary assessments on the state of the art. This review is concerned with the applica- tion of analytical pyrolysis to the analysis of molecules of biological origin and to the solution of biomedical problems. The literature up to mid-1977 has been consulted and * Present address : Department of Biochemical Pharmacology, Kings College Medical School, Denmark Hill, London, SE5 8RX.673 A guide to these sources may be found in Table I.674 IRWIN AND SLACK : ANALYTICAL Analyst, VoL. 103 although most pertinent references are quoted some further references may be found in the original articles. Analytical pyrolysis can be considered as the integration of a pyrolysis method with a detection or analytical procedure. The number of methods available in each of these areas and the variables that control their performance are manifold. As some of this material has been discussed el~ewhere,~ only a brief description, necessary to make this article complete, will be given here. TABLE I REVIEWS OF VARIOUS ASPEC'I'S OF ANALYTICAL PYROLYSIS Title Identification of Organic Substances by the GC Analysis of their Identification of Substances of Low Volatility by Pyrolysis Gas Chro- Techniques and Potentialities of Pyrolysis Gas Chromatography .. Pyrolysis Products . . .. .. .. .. .. .. .. matography . . .. .. .. . . . . .. .. .. Pyrolysis Gas Chromatography . . .. . . .. .. .. Pyrolysis Gas Chromatography. A Bibliography (1960-63) . . .. Pyrolytic Techniques . . .. .. .. .. .. . . . . Pyrolysis Gas Chromatography . . .. . . .. .. . . An Industrial View of Pyrolysis Gas Chromatography . . .. .. Identification By Means of Pyrolysis Products . . . . .. .. graphy . . .. .. .. .. .. .. .. .. Pyrolysis Gas Chromatography of Biological Macromolecules . . . . Pyrolysis Gas Chromatography of Involatile Substances .... Chromatographic Advances in Toxicology . . .. .. .. .. Characterisation (of Polymers) by Thermal Degradation .. .. Pyrolysis Gas Chromatography . . . . .. . _ .. .. Pyrolysis Gas Chromatography . . .. .. . . . . . . Pyrolytic Gas Chromatography of Polymers (in Russian) .. .. A Bibliography of Vapour Phase Thermal Fragmentation . . . . Pyrolysis Mass Spectrometry of Complex Biological Materials . . . . Laser-pyrolysis Mass Spectrometry . . . . .. . . . . .. Pyrolysis Gas Chromatography . . . . .. .. .. .. Trends and Advances in Design of Pyrolysis Units for Gas Chromato- Laser Pyrolysis Gas Chromatography for the Characterisation of Solid Materials . . . . .. .. .. .. .. . . .. Identification of Micro-organisms by Pyrolysis. Pyrolysis - Spectrometry.Automated Identification of Microbes . . The State of the Art Date 1960 1961 1964 1964 1964 1965 1966 1967 1967 1967 1968 1968 1968 1969 1969 1971 1972 1972 1973 1974 1975 1976 1976 1977 Number of references Reference 16 5 20 6 61 7 21 8 205 9 70 10 201 1 0 11 10 12 34 13 30 14 38 15 18 16 176 17 52 18 0 19 156 20 393 21 12 22 14 23 33 24 81 25 0 26 -370 3 Techniques Pyrolysis Methods The principal methods of pyrolysis have been classified13 as continuous mode, e.g., furnace and micro-reactor, or as pulse mode, e.g., filament, radiofrequency induction heating, electric discharge and laser. Continuous-mode pyrolysers link classical and analytical pyrolysis and consist frequently of an externally heated tube containing the sample, which is flushed by a carrier gas in order to remove pyrolysis products.The sample may be supported by a boat and is dropped or pushed into the heated zone. A characteristic of this type of pyrolyser is that as the pyrolysis products migrate from the sample they are subjected to an increase in temperature. This increase may initiate reactions that are secondary to the main fragmenta- tion processes and which may be an important cause of low reproducibility. Such effects are increased by the large (milligram level and above) amounts of sample used. In several instances, particularly concerning carbohydrates, large-scale pyrolyses or vacuum pyrolyses have been undertaken and in these instances pyrolysis and detection may be sequential rat her than integrated techniques. The two most common methods of pulse-mode pyrolysis are resistive heating of a platinum (or platinum - ruthenium) filament (heated-filament pyrolysers) and inductive heating of ferromagnetic material (usually a wire or a tube although other configurations are possible) by means of an applied radiofrequency field (Curie-point pyrolysers) .The pyrolysis tempera- ture of these latter units is near the Curie point of the wire and is controlled by such factors as the composition and dimensions of the wires and the applied radiofrequency field. ThisJuly, 1978 PYROLYSIS IN BIOMEDICAL STUDIES 675 means that although the pyrolysis temperature can be varied readily by using a wire of different composition, the sample cannot be subjected to a series of increased temperatures, as can be achieved with heated filaments.In both types of pyrolyser the pyrolysis conditions are similar, microgram or nanogram loadings can be achieved and as the heat source also usually acts as the sample holder (although filament pyrolysers may be used in conjunction with a sample holder) and is raised from ambient to pyrolysis temperature the volatile fragments migrate to cooler regions, which effectively reduces the possibilities for secondary react ions, As degradation by several temperature dependent and competing pathways may occur during the temperature-rise period, the final pyrolysis temperature and the temperature- rise profile should be well defined and reproducible. The pyrolysis temperature should not be too high as fragments of low characteristic value would result.Very rapid heating (millisecond range) and a low sample loading in order to eliminate temperature-lag effects within the sample and reduce secondary reactions are favo~red.~' The use of a capacitor- boosted heated filament or a radiofrequency field of sufficient power for Curie-point wires ensures adequate performance for these methods, although some examples of this equipment will not provide such a performance.28 Sensitivity and reproducibility are also dependent upon the yield of pyrolysis products (cracking severity) .29 Pyrolysis by 1aser24J0 is a further example of pulsed pyrolysis, in which a laser is focused on to a sample. A pulse duration in the millisecond range and a pulse energy of about 10-20 J ensure the capability of very rapid heating.Many samples are transparent to the beam, e.g., the ruby laser, 694.3 nm, and substances such as carbon black are added to aid the absorption of radiation. Heating is achieved through contact with the additive and gas absorbed in the powdered sample complicates the pyrogram. Often only very volatile fragments are detected without deliberate de-focusing of the beam. Significant differences may be found between the products generated from laser irradiation and by other methods. An alternative approach is to use a beam at infrared frequencies, e.g., the carbon dioxide laser, 9.1-1 1 .O pm, which excites the molecule directly. Pyrolysis products from this technique resemble more closely those from heated-filament and Curie-point methods. Although the rapid heating and the surface nature of the phenomenon (pyrolysis products may be produced at the surface and do not migrate through unreacted sample) are advantages of this method, the lack of precise temperature control and sample variability suggest that there is still room for definitive work in this field. Information concerning comparative studies undertaken with various pyrolysers has been ~ u m m a r i s e d .~ ~ ~ ~ ~ ~ ~ Other methods, less commonly used, and systems suitable for the pyrolysis of volatile samples have also been described.3933 Detection Methods Although several methods of detection and identifi~ation~~ have been used in pyrolysis studies, including the thermolysis thin-layer chromatographic meth0d~~9~6 of Stahl and direct detection of pyrolysates by use of gas-chromatographic (flame-ionisation) detectors without separation, most work has used either gas chromatography or mass spectrometry.Comments will be restricted to these techniques. Gas chromatography The ability of pyrolysis methods to produce complex mixtures of volatile products, the ability of gas chromatography to separate these mixtures effectively, and the ease with which the two techniques can be interfaced have ensured that pyrolysis - gas chromatography (Py - GC) has become the principal method in analytical pyrolysis. This subject has been covered in depth3 and for the purposes of this review it is necessary only to stress that control of the variables is essential for reproducibility, especially if work published in the literature is to be repeated.In particular, the technique can be used to detect volatile products only and the appearance of the gas-chromatographic trace (pyrogram) will depend upon the conditions, e.g., capillary or packed columns and their length, the nature and loading of the stationary phase, gas flow-rate and temperature of the analysis. In addition, the column performance often deteriorates with time. Although some workers have used isothermal gas-chromatographic analyses, more information can be obtained if temperature programming is used to permit the detection of as many resolved products as possible, in which event676 IRWIN AND SLACK : ANALYTICAL Analyst, Vol. 103 initial temperature, final temperature and heating rate are added to the list of variables to be specified. Many of the current studies using Py - GC have provided information on the identity of the fragments, and so the origin and mechanism of pyrolysis can be elucidated.Although retention-time data alone may be unreliable, various methods, including pyrolysis itself ,37 have been used for this purpose. Mass spectrornetry is now the favoured means in laboratories that have access to gas chromatographs interfaced with mass spectrometers. This technique (GC - &IS) does not require a dedicated instrurnent and utilisation can be limited to one run for each new analysis. The ultimate application of Py - GC - MS to date has been in the search for life on Mars,38 in which vaporisation of small fragments and pyrolysis of involatile components was expected.Powdered samples were subjected to a range of increased temperatures and the volatile material released at each stage subjected to GC - MS analysis. Such a procedure was used to identify over 50 organic components in the Allende meteorite and to confirm the low level of organic matter in lunar samples.19$39+40 Several applications of gas-chromatographic analysis, although requiring the application of heat, usually from the injection-port heater, are not generally regarded as examples of analytical pyrolysis. These include thermally induced reactions such as decarboxylation and other conversions to yield a more volatile product, or reaction gas chromatography, such as catalytic reduction or trans-methylation reaction^.^^,^^ These processes usually yield one volatile product only and they are not treated comprehensively in this review.Mass spectrometry Until recently43 the success of Py - GC entirely eclipsed the use of mass spectrometry as a detection system, despite the use of mass spectrometry as the detection method, without prior separation of the fragments (Py - MS), in the pioneering studies of the pyrolysis of biological molecules undertaken by Z e m a n ~ . ~ ~ The failure of Py - GC to detect large, more polar products (although these may be expected to be of importance in biomedical studies) and the difficulties in encoding gas-chromatographic pyrograms for automatic data handling, coupled with the availability of mass spectrometers with data-handling facilities has led Meuzelaar et al. to renew interest in this te~hnique.~3 In Py - MS pyrolysis is conducted in or near the ion source of a mass spectrometer and a time-averaged mass spectrum of the pyrolysis mixture is displayed as the pyrograni.To reduce the complexity of the pyrogram low-voltage (10-15 eV) ionisation of the products by electron impact (EIMS) is employed, which reduces secondary fragmentations and ensures a larger proportion of molecular ions in the pyrogram. The short analysis time (about 1 min per sample) and the uniform resolution that a stable, linear and reproducible mass scale allows, simplifies computer handling of the data and has immediate advantages over the use of Py - GC. Although Py - GC may have a slightly better success rate for discriminating between samples,45 the reproducibility obtainable with Py - MS has been shown to be satisfactory for fingerprint techniques to be applied.Structural assignments have been limited by the low resolution of the quadrupole mass spectrometer used.46 This spectrometer has enabled masses up to m/e 140 to be recorded and the range can be extended in systems with a greater r e s o l ~ t i o n . ~ ~ In particular, little information concerning isobaric or isomeric masses is obtained with low-resolution spectrometers. Some data may be obtained through a compari- son of low- and high-voltage p y r ~ g r a m s ~ ~ but most success in this field has been achieved by means of high-resolution mass spectrometry coupled with mild ionisation procedures. Field ionisation - mass spectrometry (FIMS), in which ionisation is achieved by passing the pyrolysis vapours through a high-potential electric field, yields simpler but comparable pyrograms to those obtained by means of EIMS.4s~49 Accurate mass measurement has enabled the composition of many of the fragments to be determined.This still gives no information regarding isomeric products arid collisional activation - mass spectrometry (CAMS) has been explored as a means to overcome this problem.50~5~ In a typical CAMS experiment the ions from the pyrolysis mixture are analysed using a reversed-geometry high-resolution mass spectrometer. The magnetic sector is tuned to allow selection of the ions of the desired m/e value, which are then activated by collison with helium in a chamber between the magnetic and electrostatic analysers.The fragments induced by this collison are mass analysed by scanning the electrostatic analyser. The use of two double-focusing mass spectrometers in tandem has been proposed to allow a high-resolution capability for thisJuly, 1978 PYROLYSIS I N BIOMEDICAL STUDIES 677 technique.52 Field-desorption mass s p e c t r ~ m e t r y ~ ~ ~ ~ ~ (FDMS) and simultaneous pyrolysis of the sample on the activated emitter of the field-desorption ion source48 have confirmed that in Py - MS the detection of large pyrolysis products depends more upon the transfer conditions between the pyrolysis and ionisation zones than upon the pyrolysis method. Such fragments may be degraded by secondary processes, trapped on chamber walls or destroyed by excess ionisation energy.23 In Py - FDMS very small sample sizes are needed (1-10 ng) and only a small amount of excess energy (0.2 eV) is delivered to the sample.Very short (10-11 s) residence times within the pyrolysis zone make the processes of pyrolysis and ionisation almost synchronous. Py - FDMS, of all the current techniques, allows the detection of the l q e s t and, hopefully, the most informative fragments from the pyrolysis Qf biopolymers. Disadvantages of Py - FDi'vIS methods include the high cost of the apparatus, the photoplate detection that has been employed to date, and the high resolving powers necessary for measurement of mass. Nevertheless the interplay of the various Py - MS methods will surely provide further advances in the analysis of biomedical samples. The use of empty columns to allow the production of Py - MS pyrograms from a dedicated GC - MS system has been r e p ~ r t e d .~ ~ , ~ ~ Although substantial losses of polar materials will occur with the system, mass-scaled data suitable for computer manipulation are obtained. Sample introduction is over a period of several minutes and computer collection of the data is probably essential. No attempt has been made to include within this review the large number of pyrolyses that have occurred incidentally during the production of mass spectra. Qualitative and Quantitative Aspects Analytical pyrolysis can be used as a qualitative tool for the identification or quality control of a sample, or as a quantitative method of analysis. In both modes the analyst is well advised to recognise and optimise as many of the variables as possible and so produce a standardised, defined proccd~re.~~-~O The exact requirements will depend on the equip- ment available, the nature of the information required and also on whether intra- or inter- laboratory comparisons are to be made.Several studies have been made in this area61--63 and a detailed discussion of a standard system for polymer analysis (using Py - GCJ has been given.3 The aim of such standardisation is to ensure that pyrograms are reproducible, within the limits demanded by the analysis, and are repeatable by different personnel after various intervals of ti1ne.47964 In a qualitative analysis, pyrograms from an unknown sample are compared with those held in a library. In this mode, the technique has been used widely in, for example, forensic studies3~12+~6~-68 and as a taxonomic tool for micro-organism identifi~ation.~ The differenti- ation of the samples depends upon the uniqueness (specificity) of the pyrogram.In the instance of well defined molecules, such as some drug ~eries,~~-'l each pyrogram may be characterised by a unique fragment and identification is achieved rapidly by visual inspection. In more complex instances, particularly when used as a taxonomic tool, no unique fragments are to be expected and differentiation depends upon the variation in the intensities of selected peaks in the pyrogram. If the resemblance between pyrograms from different samples is close, a high degree of reproducibility between replicate runs is demanded. It should be stressed that within-sample variance must be significantly less than between-sample variance before any reliance may be placed upon conclusions drawn.Visual comparisons are fre- quently unreliable and statistical and computer matching of the pyrograms are often performed.*5*74--$6 Py - NS data, in particular, are frequently analysed by these methods whereas Py - GC results, perhaps owing to the analogue nature of the initial record, are often analysed visually. In several of these instances very small differences in pyro- grams are claimed as significant although no data on reproducibility, particularly over a period of time, are given. No quantitative analytical applications of Py - MS have appeared7' and, in Py - GC, reproducibility and specificity (to allow the determination of a component or components in the presence of other substances) are again demanded. Further constraints, however, are imposed upon the system.In particular, assays are most favoured if each component to be measured yields a unique fragment (although corrections may be applied provided the system is not degenerate78-80), which is detected as an intense, symmetrical and well resolved peak. Further, variable but controlled amounts of sample must yield measurable peaks, the intensities of which are linearly related to the amount of sample. Ideally, the fragment678 IRWIN AND SLACK: ANALYTICAL Analyst, Vol. 103 should be produced by a unimolecular decomposition process so that the intensity is indepen- dent of concentration. Although some quantitative measurements have been undertaken on large (milligram amounts) samples,81 in practice it is recommended that small amounts of material only are handled in order to eliminate secondary reactions as far as possible.Quantitative analysis of biomedical samples was first described by Stanfords2 who developed conditions for the assay of microgram amounts of material using a heated capillary. The use of a heated filament has since been found to be particularly versatile and although the amounts used will depend upon the over-all sensitivity of the system (production and detection of the fragment of interest) low levels of material can be handled. In the instance of penicillins,70 a reproducible response to 10 n g was obtained and although this was shown to be linear up to at least 100 pg, the log - log plot that was used presumably introduces substantial errors in an assay procedure.Barbiturates can be determined using 1-5 pg per run.78 Probably the most comprehensive application of quantitative analytical pyrolysis may be found in the work of Szilagyi et aLs3 This work has been concerned with the development of a procedure for the determination of tissue levels of choline, acetylcholine and other onium corn pound^.^^ Conversion of choline to its propyl ester and the use of butyrylcholine as internal standard allows both choline and acetylcholine to be determined simultaneously.85 Acetylcholine (25 ng) can be detected by this method and the range 10-9-10-13 mol can be approached by using mass fragmentographys6 of the M + 1 ions after chemical ionisation (CI) with isobutane (Py - GC - CIMS).Sensitivity is improved if a simple quaternary compound in large excess is co-pyrolysed withi the samples7 and deuterium-labelled samples are used as internal standards. Trans-acylation between acetylcholine and choline may cause scrambling8 unless derivatisation of choline is performed. Levels of acetylcholine in normal and denervated carotid bodies have been determined recently.89 Among other applications of analytical pyrolysis to quantitative analysis are the deter- mination of mercury in fish,g0 the determination of organic oxygeng1*g2 and as a means of estimating pollution in w a t e r ~ a y s . ~ ~ ~ ~ ~ ~ ~ In this last instance pyrolysis of aqueous samples is followed by elution of the organic fragments with steam and the method can be used to detect oil contamination at levels of less than 10 p.p.m. In a further example of this applica- tion the herbicides paraquat and diquat can be determined rapidly in pond and river waters by pyrolysis to yield the corresponding bipyridykg6 Applications Drugs Since 1960 there have been numerous reports on the use of pyrolysis in the determination of pharmacologically active molecules.Many of the early applications of pyrolysis were of great interest at the time of publication but owing to rapid development of techniques in analytical chemistry many of these would not now be the methods of choice for the drugs concerned. This is particularly true for the pyrolysis studies of those antibiotics of high relative molecular mass whose complete structures have only recently been elucidated.Many reports have been made of on-column d,egradations and rearrangements but in many instances it is not clear whether these reactions are induced by heat alone or are dependent on some active site on the support material or the column itself. Wickramasinghe and Shaw have reported the decomposition of tlolazamideg7 and biguanidesg8 and reports on on-column pyrolysis have appeared on menadione bisulphite addition corn pound^,^^ pro- staglandinsloo and a thermally induced Lossen rearrangement of the drug 4-butoxyphenyl- acetohydroxamic acid.IO1 In an attempt to elucidate the chemistry of the ephedrine spot test,1O2 ephedrine was pyrolysed in a test-tube and the resulting vapours trapped in methylene chloride.The mixture was analysed by gas chromatography although the major product (75%) was not identified. A brief mention of other thermal conversions of drugs has been made in a review on gas chromatography in drug analysis.lo3 Reaction gas chromatography has also been used extensively, for example by Radecka and Nigamlo4 in the analysis of tropane alkaloids and by Pella and Colombo105 who used pyrolysis over various carbon surfaces in order to determine the amount of oxygen in the molecule. This review does not cover in detail reports of incidental thermal conversions of drugs of low relative molecular mass.July, 1978 PYROLYSIS IN BIOMEDICAL STUDIES 679 Brodasky80 used the conventional injection-port heater for low-temperature (200400 "C) pyrolysis studies of various antibiotics and also a conventional pyrolyser for high-temperature work (600-1 300 "C).Lincomycin was examined by low-temperature pyrolysis and it was possible to characterise the alkyl substituents at the three major positions of substitution. The identification of some of these pyrolysis products enabled the direct characterisation of substituents to be made. High-temperature pyrolysis (usually at 1000 "C) was used to distinguish between pairs of antibiotics whose structures were uncertain. High (900 "C) and low (380 "C) temperatures were also used in Py - GC studies of the polyene antifungal anti- biotics candicidin, levorin and trichomycin.lo6 No statistical data were given on the reproducibility and no quantification was possible.TABLE I1 JlAJOR PYROLYSIS PRODUCTS OF PENICILLINS" 0 +N+COOH Penicillin Characteristic fragment Retention Parent in olecule Radical (R) Structure timelmin Penicillin G . . . . (=JCH*- CHz-CN 6.0 Penicillin V .. O-CH2- O-cH3 .. a- Methicillin . . 0-CH3 Cloxacillin O-CH3 CI CI 4.0 2.5 5.75 The analysis by Py - GC of lasalocid, an antibiotic isolated from cultures of Streptomyces Zasalensis, has also been described.lo7 A retro-aldol rearrangement was thermally induced in the heated injection port to give quantitative yields of a ketone. A quantitative and qualitative study of penicillins and cephalosporins by Py - GC and Py - GC - MS was described by Roy and Szinai.70 On pyrolysis it was observed that frag- mentation occurred around the amido group to yield, in most instances, a characteristic fragment.This is demonstrated in Table 11, in which the major fragments are shown. The cracking severity, i.e., the percentage of material available for pyrolysis that was actually pyrolysed, was given for four antibiotics and this varied from 97.9% for methicillin to 100% for penicillin G. Quantitative studies were carried out using the most intense (and symmetrical) peak in the pyrogram. The use of Py - GC - MS as an analytical method for sulphonamides was described by680 IRWIN AND SLACK: ANALYTICAL Analyst, Vol. 103 Irwin and Sla~k.~089~0Q In most instances the sulphonamide, on pyrolysis, underwent fission about the sulphonamide group to yield aniline and a characteristic heterocyclic amine. This technique has the advantage that drugs can be analysed as the formulated product without any prior extraction as the excipients do not appear to interfere with the pyrolysis mechanism (Fig.1). Formulated mixtures h,ave also been analysed and the constituents can be clearly identified. Py - GC - MS has also been used to study the excretion of drugs and their metabolites. The application of this technique to sulphonamides has involved the production of pyrograms from lyophilised ~ r i n e . ~ ~ J ~ O The pyrogram produced from a 10 20 30 40 50 60 70 80 Retention time/min Fig. 1. Pyrogranis from sulphadimidine as (a), the pure drug and (b), a formulated prepara- tion. Peaks shown are ( l ) , aniline and (2), 2-amino-4,6-dimethylpyrimidine. (Reproduced from “Analytical Pyrolysis”lo8 with the per- mission of the publishers.) I d I I I I I I I 3 0 10 20 30 40 50 60 70 80 Retention time/m in Fig.2 Pyrograms showing the detection of sulphadimidine excreted in urine. (a), Blank urine; (b)! .patients’ urine; and ( c ) , N-acetyl- sulphadimidine. Drug-derived peaks are ( I ) , aniline ; (2), 2-amino-4,B-dimethylpyrimidine ; ant1 (3), acetanilide. Components from the urine arc A, pyrrole; 1-2, benzonitrile; C, p-cresol; D, benz- amide ; and E, phenylacetamide. (Reproduccd from “Analytical Pyrolysis”1o8 with the permission of the publishers.) control urine sample [Fig. 2 ( n ) ] is characterised by the presence of 9-cresol, with pyrrole, benzonitrile, benzamide and phenylacetamide as other major components. The last three compounds are derived from the pyrolysis of hippuric acid.The pyrograms from a sulphonamide (sulphadimidine) and from a urine sample from a patient undergoing therapy with this drug [Fig. 2 ( b ) ] show clearly the sulphonamide peaks superimposed on those of the control urine pyrogram. A new peak (peal; 3 ) is acetanilide, produced from the pyrolysis of the N-acetyl metabolite [Fig. 2(c)].lo8 The pyrimidine sulphonamides have been pyrolysed and the products compared with those obtained on hydrolysis.lll An interesting study was reported on the pyrolysis ofJzcly, 1978 PYROLYSIS I N BIOMEDICAL STUDIES 68 1 polymeric sulphonamides.l12 These polymers, produced as potential antibacterials and carriers for anti-tumour agents, were analysed by Py - MS and Py - FDMS and information on the structures of the monomer and dimer was obtained.The first description of Py - GC in the biomedical field was given by JanAk113 in 1960, who reported the analysis of a series of sodium salts of the barbiturates. He demonstrated that this technique yielded volatile hydrocarbons from which characterisation of the parent molecule was possible. He also noted that the fragments produced were not solely due to the simple fission of the parent molecule but also that thermally induced rearrangements occurred. For example, o-xylene was produced from the pyrolysis of phenobarbitone. This work was extended when the Py - GC of 27 substituted barbituric acids was described.l14 Comparisons were made of the pyrograms produced by pyrolysing the free acid, the sodium salt and a mixture of the free acid and anhydrous potassium carbonate.The peaks were digitised and represented with a log(time) scale. This treatment produced easily comparable results. The major pyrolysis products of a number of the free acids and their salts were given, with various nitriles being the predominant peaks. These were not as noted earlier113 but the authors argue that the chromatographic conditions used by JanAk probably led to the nitriles being retained on the column. When describing a new pyrolysis unit, the pyrograms of sodium seconal and sodium phenobarbitone were and the quantitative aspects of the Py - GC analysis of two barbiturates were reported.7* In this latter instance the drugs were separated from four inactive substances in the tablet and were pyrolysed after conversion to their sodium salts.Data on the reproducibility were given, which was better than &3% for both drugs. The quantitative analysis of atropines2 by Py - GC was described and 11 major peaks were obtained when the alkaloid was pyrolysed inside a glass capillary and chromatographed on Apiezon L at 80 "C. Two of the larger peaks were used for quantification and an accuracy of &5y0 was obtained when 30 pg of atropine were pyrolysed. A comprehensive account of the P y - GC of indole alkaloids has appeared116 and the fragments produced were identified by comparison of retention times. Different classes of alkaloids were differentiated by this means and it was shown that methoxy substituents on the indole ring yielded characteristic fragments.The analysis of 21 alkaloids by Py - GC has been described73 and it was shown that differentiation can be achieved by using the lower hydrocarbon (C,-C,) pyrolysis products and trimethylamine. Exceptions were morphine and heroin, which were distinguished by the production of acetic acid from heroin alone. The data obtained were treated by various statistical techniques, including multi- variate analysis. The problems of the thermal formation of polynuclear aromatic hydrocarbons are well known.lI7 Many investigators have attempted to identify and suggest the origin of the various constituents of tobacco and tobacco ~ i n o k e . l ~ * ~ ~ ~ ~ Pyrolysis provides a good model for the study of this problem and has been used by several workers in the investigation of the process.The breakdown of DDT in tobacco smoke120 has been studied by pyrolysing DDT in a nitrogen atmosphere and separating the various products by gas and liquid chromatography. In another experiment to examine the non-volatile products from the breakdown of DDT in tobacco smokes121 the condensate after pyrolysis was analysed by fractionation on florosil and then by gas chromatography. The pyrolytic degradation of nornicotine and mysomine was reported122 and the products identified included quinoline, isoquinoline and 3-cyanopyridine. The pyrolysis of p01yenesl~~ was studied in an attempt to understand the formation of aromatic hydrocarbons in cigarette smoke. For example, in the pyrolysis of 13-carotene a t 300 "C, 1,2,3,4-tetrahydro-l,l,6-trimethylnaphthalene (ionene) was produced with smaller amounts of toluene, m-xylene and 2,6-dimethyl- naphthalene.The analysis of 19 phenothiazinesl24 has been reported and a tentative identification of three of the gaseous hydrocarbons was made. The pyrograms were digitised and normalised in the same manner as those of the barbiturates114 and were thus easily compared. How- ever, no mention was made of the results obtained from pyrolysing mixtures of these drugs, as would be found in a clinical or forensic situation. The analysis of saccharin in various soft drinks and multivitamin products was carried out by Szinai and Roy using Py - GC.125 This work was extended with the identification of various peaks.71682 IRWIN AND SLACK: ANALYTICAL Analyst, VoL. 103 It was also possible to assay sodium saccharide by prior conversion to the free imide.The identification of the volatile constituents produced by pyrolysis of cannabidiol126 was reported using both air and hydrogen as the carrier gas. The volatile components were identified directly by GC - MS and the less volatile ones were derivatised. By using MS the number of hydroxy groups available for silylation was estimated. The qualitative determination of a series of substituted propionic acid derivatives, used as anti-inflammatory agents, by Py - GC - MS has been described.12' On pyrolysis two major volatile products are produced via decarboxylation and elimination to give the ethyl and vinyl derivatives. The excretion of the drugs and their metabolites was also followed by this process.A brief mention has also been made of the -possibility of using Py - GC in the determina- tion of various cardenolides, and pyrograms of digoxigenin and digitoxigenin were repro- duced.128 Biomacromolecwles The application of gas chromatography and mass spectrometry to the study of biological macromolecules has been hindered by the low vapour pressure of these naturally occurring polymers. Although techniques such as derivatisation and FDMS have increased the range of substances amenable to study by these methods it is usual to allow controlled degradation of the polymer to take place before analysis. In addition to fundamental studies on the thermal degradation of many biopolymers and the applications of pyrolysis in obtaining raw materials from organic the simplicity with which pyrolysis can generally be undertaken, coupled with small sample demands, has led to significant applications of this technique in the initial degradation of biopolymers.Structural information has been obtained from unique fingerprint pyrograms or by identification of many of the fragments produced by pyrolysis. It has been found that the various classes of biological macromolecules, with the exception of nucleic acids, give characteristic pyrolysis fragments. These have been tabulated132 and this knowledge has enabled Simmonds and co-workers to assign the origin of many pyrolysis fragments derived from micro-organi~msl~~ and organic matter.132 The literature, however, is frequently confusing or contradictory, particularly with some of the pioneering work and an object lesson in the interpretation of Py - GC data is afforded by a study of the published work pertaining to amino acids.Amino acids, peptides and proteins The early work on the Py - GC of amino acids was concerned with establishing that the technique was capable of distinguishing between the various amino acids found in proteins. The hope was that information concerning the amino-acid content of protein hydrolysates or of proteins themselves could be obtained via pyrolysis studies, particularly as no derivatisa- tion is required and sample size,may be very small (-100 ng). The field prior to 1968 has been reviewed by S t a ~ k . l ~ J ~ ~ Ulehla first studied the pyrolysis of amino acids as potassium salts in 1960.135 It was shown that 13 of 19 compounds examined yielded specific pyrograms and that pyrolysis fragments from the mononiers could be related to those produced from albumin.Kanomata and M a ~ h i k o l ~ ~ were able to extend this work by demonstrating that two proteins, casein and albumin, could be distinguished by the relative heights of five peaks found in the pyrograms of amino acids. Some pyrolysis fragments were also identified. Attempts at identifying the pyrolysis fragments from amino-acid monomers were made by Winter and A1br0.l~' Pyrolyses were carried out at 300 "C with an equilibrium step (1 10 "C) before analysis and unique pyrograms, which again were related to protein structure, were obtained. It was reported that the fragments consisted mainly of C,-C, amines and as an example phenylalanine was claimed to yield NH,, MeNH,, Me,NH, EtNH,, Pr3N, Pr2NH and Bu,N as the amine fragments.Identifications were based on retention-time data alone. The introduction of controlled Curie-point pyrolysis by Giacobbo and Simon13* enabled the degradation of very small samples to be achieved. The interfacing of the Py - GC system to a mass ~pectrometerl3~ also allowed positive identification of pyrolysis fragments to be made. Phenylalanine (30 pg), as the sodium salt, was among 28 amino acids studied and was found to yield benzene (5y0), toluene ('78y0), ethylbenzene (2%) and styrene (15yO)l4O after pyrolysis at 700 "C and capillary-column chromatography. Benzyl cyanide was also detected with slightly different chromatographic conditions.139 Nitriles were found to beJzlly, I978 PYROLYSIS I N BIOMEDICAL STUDIES 683 common fragments and acrylonitrile or acetonitrile occurred in many of the pyrograms.The pyrograms of mixtures and dipeptides were very similar to the sum of those of the component amino a c i d ~ . l ~ ~ - l ~ l In direct contrast to this, Merritt and Robertson, in a further Py - GC - MS study, found that acrylo- and acetonitrile were common products and also concluded that a unique fragment was observed from each of 17 amino acids.142 Benzene was quoted for phenyl- alanine (a low-intensity peak with Curie-point pyrolysis) and toluene was indicated to be specific for tyrosine (although this was the major component in the Curie-point pyrolysis of phenylalanine reported by Giacobbo and Simon13*), Clearly, inconsistencies of this magni- tude threw doubt on the usefulness of the technique and highlighted the need for a complete definition of all of the variables that may affect the analysis.These factors include variation in pyrolysis methods, enabling different primary and secondary products to be formed and the selective sampling of the pyrolysate caused by the inability of the column or the interfaces to transmit all products. The lack of precise identification, particularly if based only on retention-time data, increases the difficulty. In an attempt to rationalise the above studies Shulman and S i m m ~ n d s l ~ ~ have made a detailed study of the pyrolysis of some aromatic amino acids. These workers varied the pyrolysis temperature, the pyrolysis time, the pyrolysis-chamber material and the gas-chromatographic conditions and showed that the mechanism of fragmentation was similar in all instances.The detected products derived from phenylalanine are shown in Fig. 3. I t is interesting to note the fragments of high relative molecular mass that are detected and the complex redox, cyclisation and dimerisation reactions that are initiated. Analogous products have been detected by using Curie-point p y r o l y ~ i s . ~ ~ The mechanism of the conversions144 Labelling studies have also been undertaken.141 The variation in the published work may be due to several factors. Tyrosine and tryptophan behave similarly. N Attenuation 8+---- Attenuation 1 A fi Attenuation 8 40 30 20 10 0 Tirne/m in Fig.3. Pyrograrn of phenylalaninc. .I, Carbon dioxide; B, acetonitrile; C , water; D, benzene; E, toluene; F, ethylbenzene; G, styrene; H, propylbenzene; I, propenylbenzene ; J, indene; K, phenylacetylene; L, phenethylamine ; M, unknown ; N, benzyl cyanide; and 0, diphenylethane. The peak shown with a broken line was obtained with a thermal conductivity detector operating with an attenuation of 16 and those shown in full lines with a flame-ionisation detector operating with an attenuation of 1 or 8. and those of phenylethylamine and the corresponding piperazinedione have been studied further,145 as have thermal decompositions of other amino a~ids.l~G-l~~ Further studies by Simmonds et a1.149 with aliphatic amino acids identified a large range of products including nitriles, amides, aldehydes, imines and olefins, some of which may arise via piperazinedione formation.15* The more recent studies confirm that the pyrolysis pathways are affected by temperature, particularly with oven systems, (for example, phenylalanine yields appreciable amounts of bibenzyl a t 650 but none at 850 OC14j), but that detection of species of high relative molecular mass (phenylalanine gives some benzo [alpyrene a t 850 OC151) may also be a function of the system used.The large-scale pyrolysis of amino a ~ i d ~ , ~ ~ 6 - ~ ~ ~ * ~ ~ ~ for example sulph~r-containingl~~ and684 IRWIN AND SLACK: ANALYTICAL Analyst, VoE. 103 P-hydroxy has been undertaken to examine aroma formation on cooking or the development of mutagenic components.155 Substituted pyrazines and pyrroles were common fragments with serine and indicate the complex reactions that may occur with these conditions.The pyrolysis of six aliphatic amino acids initiated by ruby-laser irradiation, using carbon as an absorption aid, has also been reported.li6 Only 2-5 pg of sample were required and the very rapid temperature rises and the surface nature of the phenomenon were claimed to reduce secondary reactions to a minimum. The pyrograms obtained differed substantially from those reported earlier, many of which used furnace conditions, and in addition to olefinic and hydrocarbon fragments each amino acid gave as a unique product the corre- sponding aldehyde, e.g., alanine yielded acetaldehyde. Nitriles or other products of combina- tion reactions were not detected.In contrast, infrared-laser irradiation of phenylalanine, alanine and serine15' has yielded complex pyrograms. Alanine, for example, yielded many fragments among which were butene, acetonit rile, propionitrile and picoline. The product distribution in this instance more nearly parallels that of the thermal fragmentation and clearly the parameters controlligg this method of pyrolysis require further investigation. The pyrolysis of amino acids has been used in the study of geochemical samples in the detection of extra-terrestrial organic matter15* and in relating the geological and the thermal stability of amino acids.159$160 Derivatives of amino acids have not been systematically studied but Merritt et ~ 1 . ~ ~ ~ have shown that 22 phenylhydantoin derivatives produced via the Edman degradation could be identified.In this example 50-100 pgoof sample were pyrolysed at 1 000 "C and gas chromatography was carried out at 100-350 C. In view of the increasing number of publications concerning the identification of molecules and organisms by Py - MS, the detailed mechanistic study of the pyrolysis reactions through deuterium labelling is w e l c ~ m e . ~ ~ ~ , ~ ~ ~ Some of the products detected during the Py - &IS of rnethi~ninel~~ were also observed in a Py - GC - MS but in contrast to other Py - GC - 31s ~ o r k ~ ~ ~ 1 ~ ~ ~ no evidence was found for the production of piperazinedione fragments using In contrast to the above studies on amino acids, no comprehensive examination of the pyrolysis of peptides and proteins has been undertaken.Studies on small peptides have indicated that many of the peaks in the pyrogram can be explained in terms of the com- ponent amino a ~ i d ~ . ~ ~ ~ J ~ ~ However, the sequence of amino-acid residues142 may have an effect on the p y r ~ g r a m . ~ ~ ~ , ~ ~ ~ Thus, glycylleucine shows the presence of acetone and cyclo- pentane whereas leucylglycine yields acetone and acetic acid among the fragments.142 Although proposed as a possible method for end-group determination no progress has been forthcoming. However, one clearly documented example in structure elucidation is the detection of piperazinedione fragments derived from adjacent peptide residues in a series of actinomycins. The pyrolysis of 25 pg of actinomycin D yielded four piperazinediones (see below) which, together with the identity of the C-terminal residue, enabled the peptide sequence to be established.le4 Py - MS.A - Me Rile I CO 1 Pro Sa r Me-Val -* nn D-Val D-vY.20 The identification of octapeptides with hor.mona1 activity in transfusion fluids has been Curie-point Py - GC - MS on a capillary column was used, 100-200 ng of sample achieved.J d y , 1978 PYROLYSIS IN BIOMEDICAL STUDIES 685 (in 2-3 p1 of an aqueous solution) were required and buffer and stabilising additives caused little interference .l65 The pyrograms derived from proteins have also generally been found to be consistent with qualit at ive predict ions based upon amino-acid pyrolysis. 142 9 166 Work by S t ac k14 involving dental studies has extended the initial reports of protein pyrolysis135-137,167y168 and has shown that provided care is taken the reproducibility of retention times and peak area is sufficient to allow the differentiation of proteins such as albumin and a- and y-glob~1in.l~~ Insufficient difference was observed between collagen and gelatin to enable a distinction to be made.The comparison of the protein pyrograms with those obtained from amino acids or mixtures of amino acids allowed some interpretations of the origins of the fragments to be made. Of particular importance were the pyrrole fragments derived from proline and hydroxyproiine, which were major components in both gelatin and collagen pyrograms. The quantitative analysis of these observations was of use in determining the degree of mineralisation of calcified collagenous tissue such as tooth enamel and dentin.For example, the logarithm of the peak area of the pyrrole fragments was found to be related directly to the specific gravity of developing enamel.14 Products from a large-scale furnace pyrolysis of collagen and casein have been described.16g A nitrogen detector has been used for the enhancement of pyrolysis fragments derived from proteins,17* a related series of enzymes have been shown to have similar but distinguish- able pyrograms171 and the application of a pyrolysis technique to the identification of hair has been reported,12@ but it has not proved possible to use this as a definitive method to establish the source of individual hair samples, possibly owing to insufficient resolution.Py - MS was first used in 19524* for the analysis of proteins and this study demonstrated that different proteins yielded different mass pyrograms. It was also found possible to relate the presence of some ions directly to the structure of the protein. For example, the peak at m,/e 34 (H2S+) is significantly larger in albumin than in pepsin, thus reflecting the larger proportion of sulphur in the former sample. The more modern applications of P y - MS23 have not as yet included a comprehensive study of proteins. Nevertheless, results a ~ a i l a b l e ~ ~ . ~ ~ show that bovine insulin and trypsin can be distinguished. Insulin (methionine absent), in contrast to trypsin, shows no peak at mle 48 due to methane thiol derived from methionine. The peak at m/e 117 is also much reduced in insulin. This reduction reflects the absence of tryptophan, which yields indole (m/e 117), although benzyl cyanide (from phenylalanine) also contributes to this peak. Carboh-yclmtes The importance of carbohydrates as foodstuffs, textiles and as components of wood has initiated a considerable interest in the thermal decomposition of these compounds.17z In addition to the production of fingerprint pyrograms31,17*,173-175 many studies have been concerned with the pyrolysis of cellulosic materials as a prelude to combustion.Various furnace pyrolysers have been used in this field so that complex reactions could be observed. The effects of flame-retardant additives have also been explored. This area has been reviewed by Shafizadeh.176-178 In summary, the thermal degradation of cellulose takes place through a variety of primary and secondary reactions, which may lead either to fragmentation and the formation of volatile, flammable products (a process favoured a t high temperatures) or, at somewhat lower temperatures, to dehydration resulting in the formation of a less flammable carbonaceous residue (char).The rapid pyrolysis of untreated cellulose begins near 300 "C and the products consist of three fractions, namely a char, a tar and a volatile mixture.179 A major component of the tar fraction is laevoglucosan. With pyrolysis in vacm the proportion of this product was found to increase, owing to reduced oxidative decomposition, and the addition of a Lewis acid (SbC1,) prior to pyrolysis confirmed that an action of flame retardantslsO is to increase the yield of water, carbon monoxide, carbon dioxide, and of char, and to decrease the yield of tar and volatile fractions.The components of the volatile fraction were found to consist largely of compounds of low relative molecular mass, such as Cl-C5 aldehydes, 2-furaldehyde, glyoxal and water. 5-Hydroxymethyl-2-furaldehyde and laevoglucosan were also found. Because of the possible role of laevoglucosan as a key intermediate in the thermal decomposition of carbohydrates the pyrolysis of this compound has been f0llowed.l80-~83686 IRWIN AND SLACK: ANALYTICAL Analyst, Vol. 103 Although much evidence has been presented to show that laevoglucosan is a possible intermediate in the pyrolysis of polysaccharides, the pyrolysis of a series of mono-, di- and polysaccharides yields the same mixture of volatile products, which has been interpreted as indicating initial degradation, dehydration and condensation to form polymers that undergo secondary thermal fragmentation.ls4 In particular, the identification of 2-methyl-5- vinylfuran (a C7 compound) among many products obtained from D-glucose, including 1,4,3,6-dianhydro-~-glucopyranose as the main product, indicates the occurrence of secondary reactions.Although benzofuran was the only product detected that was not either acyclic or monocyclic186 other workl87 has shown that complex products such as benzoralpyrene may be produced from cellulose. A comparison of pyrolysis products from glucose under nitrogen or hydrogen at 400 or 1000 "C has recently appeared.18*-lg0 In recognition of the complexity of product< that may result from furnace methods Kato and co-workers have studied the Curie-point pyrolysis of ~ellulose,~~l xylanlg2 and related oligomers and monomers.Much simpler pyrograms were obtained, and from cellulose 2-furaldehyde and laevoglucosenone were the most abundant products. D-Glucose yielded mainly 2-furaldehyde and its 5-hydroxymethyl analogue. Thermal degradations of starch, important in the production of dextrins, have also been reviewed.193 Here, too, laevoglucosan is a major product, which can be quantified by direct gas chromatography.lg4 Other volatile products have been identified and are analogous t o those obtained from glucose and glucose polymers. The pyrolysis of chitin, a possible tobacco extender, has been reported.lg5 In addition to these qualitative studies, pyrolysis has also been used to simplify quantitative analysis.Cellulose esters, for example, yield the free acid a t 700 "C, which serves as a means of rapid assay.168 The pyrolysis of hydroxyethyl starch yields the volatile fragments characteristic of carbohydrates, but among these acetaldehyde is also produced from the hydroxyethyl group. Quantitative analysis using this peak (5 x 5 assay, correlation co- efficient Y = 0.99) makes it possible to determine the degree of substitution.lg6 Results were comparable to those obtained by conventional techniques. The pyrograms of eleven pectins have been shown to allow some correlation between the degree of methylation but little indication of the degree of polymerisation was obtained.lg7 A comparison of y-irradiation and pyrolysis of a- and P-D-glucose and a-D-glUCOSe hydrate by headspace vapour analysis has been undertaken.The pyrograms of N-D-f;lUCOSe and the hydrate were indistinguishable but otherwise a unique trace was obtained. About 25 products were detected, five of which were identified.lg8 Bacterial polysaccharide was used to demonstrate the long-term reproducibility of Py - MS using low-voltage i ~ n i s a t i o n , ~ ~ and the quality control of chitin extracted by different techniques from fungal cell walls has been reported.23 Although few structural assignments were made, contamination with protein (m/e 34 from H2S) and other polysaccharides (m/e 60, 84 and 128) could clearly be seen in one of the samples.High-resolution FIMS on the products of the Curie-point pyrolysis of two dextrans has enabled the compo- sition of some of the fragments to be deter~riined.~~ Although the possible occurrence of secondary reactions prevented the direct correlation of polymeric structure with pyrolysis products, Py - FDMS of glycogen yielded ions characteristic of disaccharide units which could be related to the polymer by direct bond! c l e a ~ a g e . ~ ~ ~ ~ ~ ~ Nucleic acids The pyrolysis of nucleic acids and components thereof has principally been aimed a t the identification and sequencing of units. Early I'y - GC studies demonstrated that pyrimidines and a purine gave distinctive pyrograms. Only very volatile products were detected but this was sufficient to permit the differentiation of isocytosine (acetone present) from cytosine (no acetone produced).Thymine also yielded acetone, which served as a basis for quantita- tive study.202 Initial studies with ribonuc1eos:ides and ribonucleotides found that at 800 "C (a temperature selected to maximise the production of species of low relative molecular mass) little difference was seen between the pyrograms of nucleosides and the corresponding nucleotide~.~0~ On the basis of the patterns of low-boiling and higher-boiling fractions distinction could be made between the various members of each group, but small differences Phenol and some acids have also been detected.ls5 Py - MS studies on carbohydrates have been limited.77*199 Py - MS of cellulotje has also been described.200July, 1978 PYROLYSIS I N BIOMEDICAL STUDIES 687 were found, for example, between adenosine and guanosine.A Py - GC - MS study has related the pyrolysis products to those of the pyrimidine and purine bases and ~-ribose.~O* A larger gas-chromatographic temperature range and column change gave well resolved pyrograms in which 17 volatile compounds were identified. Particularly useful fragments were cyanogen, produced by the purines only, and aceto-, propio- and acrylonitrile. The bases in some dinucleotides could be distinguished by the relative proportions of these nitrile fragments but no conclusions concerning the base sequence could be made. The loss of information owing to secondary reactions and the loss of products during analysis may be a fundamental problem with Py - GC in this area.The direct mass-spectro- metric analysis of pyrolysis products without prior separation has, however, shown con- siderable promise. The reported electron-impact mass spectrum of DNA205 is a result of probe p y r o l y ~ i s , ~ ~ ~ ~ ~ ~ and laser-induced pyrolysis yields a similar spectrum.23 Extrusion of the phosphodiester group was found to be an important process, no phosphorus-containing ions were detected and although the composition of many of the fragments was determined, little sequencing information was forthcoming. The use of carefully controlled and mild pyrolysis and ionisation conditions has enabled further information to be obtained.50 FDMS with pyrolysis directly on the field-desorption emitter in the ion source yielded many ions in the range m/e 18-500.48 Accurate mass measurement and computer analysis allowed the identification of many fragments and among important peaks detected were those from the parent bases (as M + 1 i0ns).~0~ Similar patterns at higher mass numbers indicated the presence of related but more complex units.In contrast to EIMS,205 cluster ions of inorganic phosphate (e.g., acids and monosodium salts of phosphorus oxy acids) were detected, again as M + 1 ions, and these were found to be useful as internal reference peaks. Also note- worthy is the detection [as (M + 2)2+] of dinucleotide ions, which may serve as a possible guide to base sequencing. To illustrate this achievement of Py - FDMS it may be noted that the Curie-point pyrolysis of DNA and RNA, although producing very characteristic pyrograms when followed by low- voltage EIMS or FIMS, showed no evidence for the presence of purine or pyrimidine bases.49 High-resolution FIMS enabled the identification of many of the fragments, which were found to be largely derived from the sugar residues. The differences caused by the presence of ribose or deoxyribose were clearly recognisable and were explained in terms of elimination and rearrangement of the sugar residues.Some products, e.g., acetonitrile and furan, were also common to the Py - GC analysis,204 but several substituted furans, pyrone and benzo- nitrile were also present. A degree of similarity was observed between the EIMS and FIMS results, although the latter showed fewer fragment ions and a greater abundance at lower mass values, implying that the pyrolysis method rather than the ionisation process is responsible for the lack of large fragments. CAMS has been used to identify products from DNA pyrolysis.51 Lipids aizd other compounds Thermal degradations of lipids, although frequently not analytical pyrolysis applications, have been extensively explored because of their culinary importance.208-211 Py - GC - MS has shown that phosphoglycerides yield the corresponding glycerides.212 Analytical pyrolysis studies on volatile and non-volatile products from the pyrolysis of triglycerides in air or nitrogen have shown that various acids and ketones213 or acids and hydrocarbons214 comprise the main products.A quantitative micro-determination of lipids, e.g., cholesteryl oleate, corn oil and lecithin, with pyrolysis on an uncoated gas-chromatographic column has been described215 and pyrolysis apparatus, but operated at 150 "C, has been used to determine residual hexane in vegetable oils.216 The Py - GC of vegetable oils such as coconut oil, olive oil and hardened olive oil yielded characteristic pyrograms, with Cl-C5 hydrocarbons being predominant.l13J68 The pyrolysis patterns of a series of hydrocarbons217 and fatty acid esters218 have been studied and these have been used to aid the identification of gas-chromatographic eluates3' A continuous series of alkenes and alkanes or esters up to the maximum chain length possible is a common feature of these pyrograms. Py - GC has also been used in conjunction with other physical methods in the study of surfactant composition,219 and the chain length of soaps can be determined by copyrolysis with acetates to yield methyl ketone derivatives.220 Mechanistic work in this area has appeared.221 The elucidation of the structure of several classes of lipid688 IRWIN AND SLACK: ANALYTICAL Aizalyst, Vol.103 has been achieved through a knowledge of pyrolysis pathways. Mycolic acids are produced by several genera of bacteria (e.g., mycobacteria) and are a-branched p-hydroxy acids. The pyrolysis of these compounds yields an acid and an aldehyde. Thus methyl corynomycolate (from Corymbacteriunz diphtheriae) yields palmitic aldehyde and methyl palmitate [see below, where K, = R2 = CH3(CH,),,-1. D R2-CH-CH-COOCH, -+ R,-COOCH, + R,-CHO I I This reaction has been used with effect in structural studies on mycolic acids isolated from lesions of human leprosy.222 Only small amounts of material were required and pyrolysis was conducted on-column or under reduced pressure and separate mass spectra were recorded.The position of the double bond in unsaturated fatty acids can be determined by pyrolysis of the ozonides or by other oxidative method^.^,,-^^^ The gas-chromatographic study of phospho- lipids can be undertaken following a thermal cleavage involving the phosphate ester functions.226 Glycerophospholipids yield diglycerides and the structure can be determined by silylation and mass spectrometry. Pyrolysis has also been used as a first step in the determination of the structure of cyclopropane and cyciopropene fatty acids227 and as a derivatisation technique to yield esters from tetraalkylammonium sa1ts.225~22*~22~ The infrared laser induced fragmentation of membrane components157 has involved the pyrolysis of saturated (methyl stearate, myrist ate and palmitate) and unsaturated (methyl oleate and elaidate) fatty acid esters, diglycerjdes (dipalmitin) and phospholipids (lecithin).Little difference was observed between oleate and elaidate but increased yields of diene and the presence of short-chain esters distinguished these from the saturated compounds. The diglyceride gave no peaks characteristic of glycerol and yielded methyl palmitate in addition to the fatty acid fragments. Trimetliylaminct, from choline, was the main distinguishing feature of lecithin pyrolysis.An early use of Py - GC (as successive rather than integrated techniques) was the correlation of the pyrolysis products with the side-chain structure of some steroids.230 Fission was catalysed by benzenesulphonyl chloride and yi.elded 2-methylheptane from cholesterol (and other steroids with a C8 side-chain) and 2,3 -dimethylheptane from ergost-8-en-14-01 and similar compounds. Other studies on the pyrolysis of steroids have been initiated to determine the origin of artifacts observed during gas-chromatographic analysis. The 5,8- peroxides of ergosterol and 7-dehydrocholesterol could be rapidly identified and distinguished from non-peroxide steroids by the pyrogram generated from flash v ~ l a t i l i s a t i o n . ~ ~ ~ Although no identification was accomplished 14C labelling at C4 showed that only a small amount of fission of the molecule occurred.Using cholesterol as a model compomd, the effects of variables such as pyrolysis temperature, pyrolysis time and con- centration of solution on the yields and reproducibility of the products was determined.58 It was found that the three main components were cholesta-3,5-diene (with unresolved 2,4- isomer from an isomerisation reaction), cholest-4-en-3-one, and cholesta-l,5-dien-3-one. Other compounds yielded analogous products. Pyrolysis of the acetate or the trimethyl- silyl ether of cholesterol232 yielded mainly the diene as the steroid product, an elimination noted during the attempted gas-chromatographic analysis of other derivative^.^^^,^^^ A more volatile fraction was also detected and mass-spectrometric identification was attempted on 15 peaks in this region.232 Although the composition varied with the structure of the side- chain more information was available from the catalytic pyr~lysis.~~O The Curie-point pyrolysis of some bile acids yields fingerprint pyrograms (no identification of fragments), which enabled differentiation between these compounds to be made.171 The pyrolysis of cholesterol has been used as a probe to study the effect of metal loading (nickel) to increase the absorptivity during laser pyrolysis235 and to identify the polycyclic compounds produced under various conditions.236 Porphyrinss2J41 :yield a characteristic volatile fraction23' and pyrrole derivatives.238 Thus haemotoporphyrin was distinguished from protoporphyrin by the presence of ethanol [from the CH,CH(OH)- side-chain] and could be quantitatively assayed (*3%) in this mixture.The failure to detect pyrroles during this analysis is almost It was also found possible to identify mixtures by this technique. Shorter-chain acids have been briefly mentioned.l56 Curie-point pyrolysis of steroids has also been undertaken.July, 1978 PYROLYSIS I N BIOMEDICAL STUDIES 689 certainly as much due to the gas-chromatographic conditions237 as to the pyrolysis tempera- t ~ r e . ~ ~ ~ I t has been shown that characteristic pyrroles, e g . , 2,3,5-trimethyl-4-ethylpyrrole, produced by cleavage of the porphyrin at the =CH- methene bridge positions only, are favoured at lower temperatures (about 400 "C) whereas higher temperatures result in an increased yield of less characteristic polymethylpyrroles.The thermal analysis of bone has been reported,23g and the pyrolysis of a variety of C5 compounds, isolated as components of food odours, yield characteristic pyrograms.240 The Py - MS of fungal melanins has also been reported.23,199f241 The pyrolysis of onium salts to initiate a Hofmann degradation (see below for acetylcholine) has been studied in order to establish a quantitative assay for these substances. n (CA,),6CH2CF120COOCH3 + (CH3),NCH,CH,0COOCH, + CH3X X- On-column pyrolysis of choline hydroxide satisfactorily yielded trimethylamine and acetaldehyde but acetylcholine yielded only degradation products.242 At higher temperatures (450 "C and above), the halides of these and other compounds84 underwent a quantitative elimination to yield a methyl halide and a characteristic tertiary amine, which served as the analytical fragment.Propionylcholine (Py - GC)s3 or deuterated analogues (Py - GC - MS)87,s8 can be used as internal standards and the technique is applicable to other compounds such as y-aminobutyric acid. Mass fragmentography of the pyrolysis products of carpronium chloride has been used to monitor blood and urine levels.243 The Py - GC - MS of antho- cvanins has been described.244 Taxonomy The use of pyrolysis in biomedical taxonomy was first suggested in 1952 by Zeman~4~, who used Py - MS to characterise proteins. Chemical profiling of organisms has received much attention and pyrolysis has several advantages as a taxonomic tool, especially in that results can be obtained very quickly with analysis times as short as 1 min.The methods used (Py - GC and Py - MS) can be automated and therefore the expertise normally required when, for instance, the classification of micro-organisms is undertaken is reduced. A further advantage is that the results obtained are usually in a format that is ideal for statistical analysis, although this may also be a necessity, and the equipment can normally be utilised for a wide range of material, e.g., bacteria, fungi, polysaccharides and proteins, without modification. However, there are numerous inherent pitfalls associated with the technique, which may trap an inexperienced user.A large amount of data is generated during analysis but the association of this indirect chemical information obtained by pyrolysis with the clinical properties (e.g., are the bacteria haemolytic?) is a t best obscure and often impossible. The technique therefore depends essentially upon comparison of data obtained from an unknown sample with those held in a data bank by various statistical and pattern-recognition tech- niques. In most instances this has involved comparison of the ratios of various peaks to each other, as there are few descriptions of organisms producing unique component~.~52 The statistical analysis of the pyrogranis requires that the reproducibility of duplicate runs is high. This is frequently claimed to be so when the sample is analysed in the same laboratory but would appear to be an elusive condition in interlaboratory analyses.If pyrolysis is to be universally accepted as a taxonomic method then the reasons for this lack of reproducibility must be understood and some general method accepted as a standard. I t may be that a closely defined Py - MS system should be chosen for although this may not separate isomeric fragments the rapidity of the analysis and the ease of data handling are compelling advantages. A brief resume of some of the causes of lack of reproducibility are given below but a more detailed discussion is given in the various papers in the subject. When analysis of organisms by pyrolysis is undertaken it is the chemical nature of that organism that is assessed. It is therefore of paramount importance that the organism always exhibits the same chemical characteristics. This is dependent upon s u b - ~ u l t u r i n g , ~ ~ ~ growth time254 and the medium upon which the organism is grown.255 If all other parameters are constant the results may still vary.64 The effect of the type of pyrolyser used is also large, and this subject has been discussed a t length e 1 ~ e w h e r e .l ~ ~ ~ ~ It is essential that a standard system be adopted that is as simple as possible, in order to eliminate unwanted adsorption690 IRWIN AND SLACK : ANALYTICAL Analyst, Vol. 103 and secondary reactions due to active or cool sites within the pyrolyser unit. Standardisation of the analyser is also vital. Traditionally, gas chromatography has been used to separate and characterise the pyrolysate and the conditions used have varied considerably. Metal columns have been used, which are well known for the adsorption of polar materials and catalysis of decompositions and rearrangements.However, most workers have used Carbo- wax 20M as the liquid phase. Probably the most important factor is that of column resolution. It would appear that there are more than 200 fragments formed on pyrolysis of bacteria59 and as most analytical systems fail to resolve more than 100 components, and some systems far fewer, much valuable information is lost. More recently Py - MS has been utilised for taxonomic purposes but there is scope for a multitude of conditions to be used with respect to the separation and detection. The rapid advances in mass spectrometric instrumentation also reduce the possibility of immediate standardisation in this field.It has also been stressed that a difficult problem is the interpretation of the results obtained. As most pyrograms differ only in the relative amounts of the various fragments produced on pyrolysis, it is essential that the reproducibility of duplicate analyses is thoroughly investigated and that the data are available for scrutiny. It would also appear to be essential that the interpretation of the results is automated in order to eliminate subjective judgements as to the significance of small differences in peak heights. It is encouraging to see that some of the recent work, especially using Py - utilises sophisti- cated statistical analysis that is fully computerised. It should be noted, however, that although Oyama167 first described the Py - GC of bacteria in 1963 very few routine systems are in use at this time.Bacteria Oyama167 first described the Py - GC of bacteria when discussing experiments for the detection of extra-terrestrial life. The pyrograms of three micro-organisms isolated from soil samples, although the amount pyrolysed seemed excessive (10 mg) , exhibited good chromatographic resolution, particularly when the results are compared to some more recent work. This work was extended257 when the effect of the growth media on the pyrograms was discussed in addition to comments on the handling of normalised pyrograms in computer matching of unknown samples. The origin OF certain peaks was discussed with respect to the pyrolysis of albumin and the constituent amino acids of albumin.The pioneering work in the use of Py - GC as a taxonomic tool was undertaken by Reiner and co-workers. The successful analysis of 95 coded samples of mycobacteria by Py - GC was reported,258 and advantages over current techniques were discussed. No statistical data on the significance of the differences in peak ratios were given. This information would appear to be essential when only quantitative differences in the pyrograms were observed. However, the need for stringent control to determine reproducibility was discussed later.259y260 It was felt that the possibility of unique fragments being present in the pyrograms was slight and that the data from which classifications were made should be available.The possibility of using computer matching methods was also mentioned. The use of Py - GC for the rapid characterisation of Salmonella organisms was discussed together with the comparative serology and chemistry of the organisms. All 54 coded samples were identified correctly and the possibility of utilising Py I- GC - MS to associate the differences in the pyrograms to a particular chemotype was discussed. This work was extended when results obtained from the Py - GC of some Gram-negative organisms were given.261 Attempts were made to differentiate between bacteria that differed only in flagellar antigens, using E. coli cultures of known serotype. Very small differences, described as consistent, were observed but no statistical information regarding the significance of these differences was given.Comments were made on the fact that it is difficult to place any clinical emphasis on differences in the pyrograms per se, as bacteria with great serological differences can exhibit very similar pyrograms. The Py - GC of 50 samples of mycobacteria was reported2G2 and the reproducibility of the retention times of two significant peaks in the pyrograms from five organisms was discussed. An interesting comparison was also made of the results obtained from two different gas chromatographs, with the conclusion that the organisms pyrolysed were identified correctly using both instruments. Bar graphs were used for interspecies variation comparisons and suggestions made that these bar graphs could be computerised for more objective comparisons.July, 1978 PYROLYSIS I N BIOMEDICAL STUDIES 691 The work on the mycobacteria was expanded64 when the effects of culture age and media on the pyrograms were discussed. These changes were apparently dependent on the indi- vidual organism and were therefore difficult to predict, but changing the media had more of an effect on the pyrogram than the culture age.Comment was made on the differences in the pyrograrns from drug-resistant and drug-susceptible organisms. More recent work on niycobacteria has also a~peared.~63 Menger et a.Z.76 extended the work reported earlier by Reiner by using computer matching of pyrograms from pathogenic organisms. Although the program gave the correct result in 90% of the cases it was too simplistic (as recognised by the authors).Only peaks at the end of the pyrogram were utilised and these were split into two groups: those with amplitudes of 3 or 4 mm and of greater than 4 mm (peaks of 2 mm or less were rejected). The program then compared the test pyrogram with those stored in a data library, altering the error limits on retention times if no peaks were found to match. The advantage of this technique is that subjective judgements as to a match or no-match of pyrograms should be eliminated. A general overview of this work has recently been given,12* in which the proposed appli- cations of Py - GC are extended to material such as cancer cells. The use of Py - GC for the differentiation of Clostridium botulinum, types A, €3, and E, has been rep0rted.~~4 Interesting comparisons were made on the effect of the growth media on the pyrograms and the effect of sporulation was large, with a small shoulder in the vegetative-cell pyrogram becoming the most intense peak in the pyrogram produced from the spores.It was possible, using Py - GC, to differentiate these organisms at the strain level but not always possible at the type level. An attempt was made to detect toxin production by using a modification of the dialysis sac technique and pyrograms were given for the dialysate supernatant fluid (DSF), DSF plus partially purified toxin and partially purified toxin. However, the peaks in the pyrogram due to the toxin were obscured by other fragments. The use of Curie-point Py - GC of bacteria on open tubular columns was reported in 1972265 and comments on interlaboratory reproducibility were made with reference to the wide range of techniques used and the effect of sample preparation.A large section was devoted to the experimental technique and the advantages of using low loadings (5-15 pg) were discussed. The pyrograms produced from different strains of Neisseria menigitides were given, but only small differences were observed. This work was continued by reporting the differentiation of strains of streptococci that differed by only one antigen, using the Py - GC of cell-wall fractions.266 This study of well defined bacteria was used in an attempt to understand the biochemical basis for the observed differences in bacterial pyrograms. Polysaccharides were extracted from the cell wall and the pyrograms compared; it was concluded from these results that the differences in cell-wall pyrograms were due solely to the differences in the poly- saccharide present.The automation of Curie-point Py - GC was des~ribed,~59~67 which permitted unattended operation for 24-30 h. Attempts were also made to automate the data handling but these were unsuccessful owing to unacceptable errors, by the digital integrator, when calculations were made involving fused or tailing peaks. The limitations of this technique were recognised by the authors and ideas proposed concerning the utilisation of sophisticated computer programs incorporating multi-dimensional correlation functions. The classification was accomplished by using the differences in relative peak heights. It was also shown that Vibrio cholerae could be distinguished from other aerobic Gram-negative organisms.These classifications corresponded to those obtained by more conventional methods. The develop- ment of a high-resolution Py - GC system, for the identification of micro-organisms, was described by Q ~ i n n . ~ ~ Great care was given to the optimisation of the pyrolysis and of the chromatographic conditions. This approach is to be welcomed as many of the earlier workers did not devote enough space to this type of discussion. The pyrograms from various bacteria were reproduced and it was recognised that some of the peaks might be due to secondary reactions within the pyrolyser. A brief report comparing the Py - GC and the gas chromatography of extracted, derivatised fatty acids as an aid to the identification of streptococci has appeared.269 The conclusion was drawn that Py - GC gave more satisfactory results. The use of Py - GC for the identifi- cation of oral streptococci has been briefly discussed270 and expanded.72 This later paper72 described the use of Py - GC to compare streptococci from deposits on teeth and in blood Py - GC was utilised to differentiate between types of Vibrio choZerae.268692 IRWIN AND SLACK: ANALYTICAL Analyst, Vol.103 with 30 named strains. The magnitudes of a selection of peaks in the pyrogram were used to obtain similarity coefficient^,^^ which in turn were used for the differentiation of samples. These results were also compared with those obtained in different laboratories and various elaborate statistical techniques were used, the results from which were discussed fully.Investigation of the chromatographic conditions was carried out with respect to the Py - GC of Gram-negative ba~teria.~7l Both polar and non-polar packed columns were investigated but owing to improved resolution support-coated open tubular (SCOT) columns were used. It was found possible to trap the pyrolysate at the top of the column and then chromatograph it normally when the correct flow conditions were restored. Although correct identification was achieved when six organisms were classified by a visual matching routine this was not achieved when computer matching was attempted. Problems, similar to those noted by were encountered owing to the deposition of material of high boiling-point a t the injection end of the column.The use of a disposable pre-column was unsuccessful but a modified backflush method was suggested. Several Bacillus species, which were difficult to identify morphologically, were studied by using Py - GC.272 Difficulty was encountered when the cells were taken straight from the media as peaks due to contaminants, e.g., niitrient agar, produced inconsistencies in the pyrograms. A technique was therefore developed in which the bacteria were grown on a membrane filter placed on the nutrient, enabling the cells to be harvested rapidly and cleanly from the filter. This work was e~tended~"S3~~5 when it was stated that the effect of culture age on the pyrogram becomes most pronounced after 48 h and that sporulation would appear to have a stabilising influence on the cellular constituents.In 1070 S i m m ~ n d s l ~ ~ described the use of Py - GC - MS in the identification of most of the fragments produced by the pyrolysis of Micrococcz~s luteus and Bacillus sz.tbtilis var. ghiger. These experiments were a natural extension of those described earlier for the detection of extra-terrestrial life.167y257 Forty-eight iragmeiits were identified and tabulated according to their possible biochemical origin, i.e., proteins, carbohydrates, nucleic acids, lipids and porphyrins. All the major peaks were identified but some, e.g., acetamide, were not assigned to a particular class. Certain fragments were unique to one organism (e.g., methylstyrene was seen only in M . luteus).Comparisons were made with the results obtained irom the previous pyrolysis132 of geochemical samples and 9lyL of the fragments seen on pyrolysis of desert soil were also observed in the pyrolysate of ill. luteus. A study has also been under- taken to evaluate the influence of the temperature - time profile of the pyrolysis process on the total yield of volatile organic components Erom Bacteroides f ~ a g i l i s . ~ ~ ~ Meuzelaar and co-worker~~7~77~~~~ described the development of Curie-point Py - MS using low-voltage electron-impact ionisation. Exarniples of the pyrograms from various materials were given with an emphasis on short ((1 min) analysis time. The use of pyrolysis followed by FIMS of organisms was descrihcd.275 The molecular ion was used to characterise the pyrolysis fragments.Fragments of inie 14-142 were tabulated as their high-resolution masses together with their probable identity. Many of these structures were consistent with those proposed by S i r n m o n d ~ . ~ ~ ~ , ~ ~ ~ In Fig. 4 and Table I11 the two sets of data are com- pared. These similarities are perhaps surprising as there is no close taxononiic relationship between the organisms studied in the two rcports, which eniphasises that the differences observed on pyrolysis are mainly of degree arid therefore great care should be taken when interpreting the results. The application of Py - MS in the differentiation of strains of streptococci was described75 and the samples were taken directly from the culture plate. The pyrograms were obtained via a signal-averaging unit, which permitted study of a single time-averaged pyro,pram.This was a useful development as small deviations due to changes in the diffusion rates of fragments were removed. A brief description of the algorithm for the matching program was given but the need for more refined pattern-recognition techniques was expressed. Other attempts to characterise bacteria have been made by using degradation followed by mass spectrometric analysis. The direct insertion of lyophilised bacteria into the ion source of a double-focusing mass spectrometer to produce characteristic mass spectra has been described.276 These spectra were shown to originate from the pyrolysis products of phospho- lipids and ubiquinones. The temperature of the ion source was held at 300-350 "C and direct probe heating was not used.Reproducibility tests were carried out on different preparations from a single strain. The use of linear-programmed thermal degradation massJuly, 1978 PYROLYSIS IN BIOMEDICAL STUDIES 693 3Q 40 50 60 10 20 Time/rnin > f c W m c., 0, c.. vl .- c., .- .- - GC 0 Fig. 4. Comparison of data from ( a ) , a Py - GC - MS analysis of Bacillus subtilis var. iziger and ( b ) , a Components common to both analyses are identified in Table [Pyrogram ( a ) is reproduced from Biomedical Mass S p e ~ t r o r v t e t v y ~ ~ ~ with the permission of the pub- Py - FIMS analysis of Pseudomonas putida. 111. lishers and (b) has been redrawn from data given in Analytical spectrometry as a tool for the identification of bacteria has recently been de~cribed.~7~ This approach utilises the production of certain fragments at a specific temperature and the use of chemical ionisation enables information concerning the molecular species of the fragments produced to be obtained.Ion profiles were reproduced, which were used to classify the ten bacteria studied and the algorithms used for making the identification were given. Correla- tion coefficients were supplied for replicate analyses of Citrobacter freundii. Meuzelaar and co-~orkers~78,~79 reported the extension of their Py - MS work with descrip- tions of fully automated sample-handling and data-reduction techniques. The instrument was similar to that previously described47 and utilised an automatic sample-introduction turntable that had a capacity of 36 samples.The data obtained were treated on a large computer, using sophisticated pattern-recognition techniques involving multivariate analysis. This is probably the most sophisticated technique in use at present and if the results could be processed on a mini-computer the whole apparatus could operate as an independent system, which would have great promise as a routine analytical method. I t should be interesting to note future publications concerning the long-term repr~ducibility,~~~ which is essential if large data banks, necessary for these types of comparisons, arc: to be kept effici- ently. Fuizgi Various dermotophytes have been studied by means of Py - GC280 and the effects of colony age, sample size and culture media on the pyrograms have been investigated.However, the pyrograms were poorly resolved and in an extension of this work281 the chromatography was694 IRWIN AND SLACK: ANALYTICAL TABLE I11 IDENTIFICATION OF COMPONENTS COMMON TO THE Py - GC - &IS252 AND Analyst, Vol. 103 Py - FIMS275 ANALYSIS OF CERTAIN BACTERIA (see Fig. 4) Compound Methane . . . . . . Ethylene . . . . . . Ethane . . . . , . Acetonitrile . . .. . . Propylene . . .. . . Ethylene oxide . . .. Propane . . . . .. Methanethiol . . . . Acrylonitrile .. .. Butadiene . . .. . . Propionitrile .. . . Acrylaldehyde . . . . Butene + methylpropene . . Acetone + propionaldehyde Butane . . . . . . Acetaniide . . ,. .. Dimethyl sulphide . . . . Pyrrole . . .. .. Methylbutadiene . . .. Butyronitrile .. . . Methylpropionitrile * .Methylbutene . . . . Butanone . . .. . . Methylpropionaldehyde . . Propionamide . . . . Benzene . . .. .. Pyridine . . .. . . Methylpyrroles . . . . suhstantiallv immoved. Furan .. ,. . . Py - FIMS mle value 16.030 28.028 30.048 41.027 42.044 44.022 44.064 48.004 53.025 54.047 55.043 56.024 56,060 58.043 58.076 59.040 62.023 67.039 68.023 68.059 69.057 69.057 70.075 72.058 72.058 73.057 78.047 79.043 81.060 Py - GC - MS peak number Compound 1 1 1 9 1 1 1 1 9 1 10 1 1 1 1 32 4 23 3 1 12 13 1 4 2 35 7 17 24, 25 Meth ylpentene Hexene . . Methylfuran . . Meth ylbu t yroni trile Pentanone . . Methylbutanal Toluene . . Picoline . . Phenol. . .. Dimeth ylpyrroles 2-Furaldehyde Dimethylfuran Furfuryl alcohol Benzonitrile . . Styrene . . Xylenes . . Ethylbenzene Dimethylp yridine Cresols .. Alkylpyrroles iC7) Indole . . .. Toluonitrile . . Phenylacetonitrile Prop ylbenzene Alkylbenzenes (C9) Ethylphenol . . Xylenols . . Alkylpyrroles (C8) Methylindole . . .. . . .. .. . . .. . . .. . . . . . . .. . . . . . . .. . . . . .. .. . . .. . . . . . . . . .. . . .. Py - FIMS ntle value 84.091 84.091 82.043 83.073 86.071 86.071 92.060 93.058 94.036 95.074 96.026 96.064 98.038 103.041 104.062 106.073 106.073 107.069 108.058 109.090 117.068 117.068 117.068 120.092 120.092 122.076 122.076 123.102 131.070 Py-GC-MS peak number 5 5 8 14 8 6 11 20 39 26, 27, 28 22 7 30 27 19 16 13 21 38, 40, 41 31 45 37 36 15 18 42 43,44 33, 34 46 When considering the problems associated with data evaluation it was found that insufficient information w& avaifable from the 20 most intense peaks in any one pyrogram, but better results for taxonomic purposes were obtained if all 114 peaks were included.A taxometric map was produced to indicate the relationship between the 21 strains, based on the information derived from the pyrograms. Py - GC was used as an aid to the identification of Penicillium species.282 Data were given on the reproducibility of the retention times for 29 peaks over six replicate runs, but the column resolution was not sufficient to separate many groups of fused peaks. It was noted that some peaks were more reproducible than others and particular emphasis was placed upon these. This work was extended to the Py - GC of Aspergillus species and the effects of hyphal age and sporulation were inve~tigated.~!33 The use of Py - GC - MS in the study of Stveptomyces longisporojiavus has been recently de~cribed.~5~ Most of the fragments were identified and assigned to various biochemical classes, as in a previous study, on bacteria, by S i m m ~ n d s .~ ~ ~ Comments were made on the appearance of two unique peaks in the pyrogi-am of the actinomycete. These peaks were isobutane and methylnaphthalene. The biochemical information obtained regarding the cellular structure was also discussed. Py - MS studies have been reported on the celLlulosic cell-wall fraction of Europhiztm aureum and Saprolegnia ferax200 and on whole cells, to xid the differentiation of genera and species of Rhinocladiella-like Comparisons, using non-linear maps, were undertaken and various ion series, characteristic of macromolecules, were identified.Particular emphasis has been placed upon the discrimination of chitinous and non-chitinous ~ e l l s . ~ 8 ~ Miscellaneous In a Py - GC study of chlorophycean and rhodophycean algaelg6 pyrograms were repro- duced for 15 organisms grown in culture and both qualitative and quantitative differences were seen. The use of Py - GC in the identification of cockroaches at the species level was described in 1975.286 The insects were dried at 100 "C and prepared for pyrolysis by being ground into aJuly, 1978 PYROLYSIS IN BIOMEDICAL STUDIES 695 homogeneous powder. The effect of the sex and the sample history on the pyrograms was discussed, and it was reported that specimens that had been pinned for as long as 70 years were still recognisable by their species-specific peaks.The polar products from the pyrolysis of various organisms have been analysed by silylation and temperature-programmed gas chromatography. The material studied included algae, bacteria and a mixed plankton population. Two methods of pyrolysis were used, i e . , conventional pyrolysis and a method in which the material was sealed in a glass tube that was heated by passing it through a flame. The methods were claimed to be comparable. The infrared spectra of the pyrolysis products were also discussed.287 Pathology The application of Py - GC to the detection of plant infections288 was reported shortly after the initial studies by Reiner and co-workers on bacterial diff e r e n t i a t i o ~ i ~ ~ ~ ~ * ~ ~ but, in contrast to other physical screening methods for the detection of pathological conditions, e.g., GC - MS,290--293 only a small amount of progress has occurred in this area. No informa- tion is available concerning the reproducibility and specificity of the method, particularly considering the differences in samples that may be expected, and the identification of diagnostic fragments has received little attention.Pyrograms of healthy and infected leaves of oats, tobacco and wheat have been reported.288 Thus, oats infected with barley yellow dwarf virus, tobacco infected with potato virus Y, wheat infected with yellow rust and mildew and barley infected with yellow rust can be detected rapidly and an assessment of the distribution of the infection can be made.It was suggested that the differences in the pyrograms caused by the fungal infections were due directly to the presence of the fungus rather than modification of the host. I t would be appropriate, therefore, that any application of this technique should be preceded by a study of pyrolysis with respect to the period of time following infection, in order to establish whether early detection of infection is possible. The application of Py - GC to mammalian cells has also been e ~ p l o r e d ~ ~ ~ ~ ~ ~ ~ and differences between normal and pathological cells have been reported, although interspecies variation of some tissues was not evident. Pyrograms reported included those from kidney and red blood cells of the mouse and the rabbit, normal and leukaeniic white blood cells and profiles of normal, cancer and transformed cells.A possible rapid method for the pre-natal detection of genetic disorders by the Py - GC of lyophilised amniotic fluid was described by the same Cultured skin fibroblasts allowed the differentiation of normal tissue from a range of pathological tissues but the tentative nature of these results was stressed. A Py - GC study of haemoglobin171 has shown that normal adult haemoglobin can be distinguished from that of the foetus (replacement of P-chains by y-chains) and from that of patients suffering from sickle-cell anaemia (replacement of valyl by glutamyl in position 6 of the P-chains). Py - FDMS of human blood296 and Py - GC - MS of urine69 have also been reported. Geological Materials of Biological Origin Although this area is not strictly a biomedical application the molecules of interest are mainly derived from biological sources: because of this and the importance of pyrolysis in this field a brief description of some of the work is presented here.Much of the organic matter of the geosphere consists of complex intractable polymeric substances. The analysis of such material is of importance in the examination of sediments and in the characterisation of soils, following, for example, the process of humification and the characterisation of genetic horizons and humus types. In sedimentary deposits close attention is paid to non-extractable material (kerogen), which may comprise 90-95% of the organic matter, and analysis leads t o information on potential petroleum-source rocks.The advantages of analytical pyrolysis have led to the useful application of the technique in these areas. Pyrolytic methods have also been used in the study of coals, bitumens and petroleum and to derive information concerning the palaeo-environment. Initial experiments in Py - GC were conducted in 1963 but it is probable that some of the impetus for these applications was provided by the demonstration that Py - GC - MS was a feasible technique for the study of organic matter from extraterrestrial samples.1323297696 IRWIN AhTD SLACK.: ANALYTICAL Analyst, Vol. 103 The aim of these experiments was to aid the biological exploration of Mars involving the determination of the presence and origin of any surface organic matter.Various samples of desert soil, pre-Cambrian shale and meteorites were pyrolysed at 500 "C and capillary gas chromatography yielded characteristic pyrograms. Many of the fragments were identified and on the basis of the known fragmentation of biomacromolecules the origin of these com- pounds was proposed. Californian desert soil was found to contain fragments derived from most classes of biomacromolecule and yielded a product distribution broadly comparable to that of humic acid from agricultural In contrast, the more abundant fragments derived from meteorite or shale, in which the bulk of the organic matter is a condensed aromatic polymer, were a complex mixture of aliphatic and aromatic hydrocarbons. As part of a series of mass spectrometric studies involving the Viking experiment~~~*-~~l the Py - GC - ills of two Martian surface samples were analysed.Carbon dioxide and water were detected but no significant organic fragments (at a detection level of 1 in 109) were detected.38 The presence of substantial amounts of naphthalene from the pyrolysis of meteorite samples and its absence in this Martian soil reflects the theoretical problems associated with this result. Analytical pyrolyjis has also been used to determine organic compounds in lunar samples302~303 and in other meteorites.lg An early report on the use of analytical pyro1:ysis in soil science was the study by Nagar,30* but although similarities and differences between pyrograms derived from various humic acids were noted, no structural conclusions were drawn.Pioneering Py - MS studies305-307 and later Py - GC - MS ~ ~ r k ~ ~ ~ ~ ~ by Eracewell and co-workers has enabled a classification of freely drained temperate soils to be made.310 These studies, on whole soil samples, have enabled distinctions within the genetic horizons of a soil profile to be made and variations in humus type between different soils can be seen. Many of the fragments produced on pyro- lysis were identified. Thus, methoxyphenols (p2aty surface horizon), dihydropyrone (horizon A) and benzonitrile, acetophenone and naphthalene (horizon B) were among peaks charac- teristic of a podzol profile. Dihydropyrone and furans permitted the differentiation of mor humus (in which the plant material is little altered) from mull humus (well decomposed and complexed components), which gave increased amounts of cyclopentenone and pyrroles.Such analyses allowed a better correlation with soil properties than other, more traditional techniques such as the determination of carbon content or carbon to nitrogen ratio. Relation- ships between different soils (in a climo sequence) can also be studied and a correlation between the relative intensities of characterisiic pyrolysis products and climatic variables was obtained.311 P y - MS studies in this area1999313J14 have included studies on humic acids and comparisons have been made with model substances such as soil fungal melanins and phenolic peptide polymers, e.g., phenolase oxidation of hydroquinone or catechol in the presence of an amino acid or peptone. Charac- teristic series of ions due to sulphides, pyrroles, benzenes, phenols and indoles were described.Series of ions derived from phenols and methoxy derivatives were characteristic of humic acids derived from straw compost but were only detected weakly in soil samples. Results were consistent with the formulation of melanins as mixed phenol - protein polymers.313 The association of carbohydrates with fulvic acids has been studied by Py - GC.3159316 The polysaccharide component yields many identifiable products and the measurement of the intensity of the furfural peak makes quantification possible. Thermal methods have been applied widely in the analysis of fossil fuels and analytical pyrolysis has been used for the characterisatison of samples and source rock.Thus, laser Py - GC has enabled a rapid estimate (&lo%) of various components of oil-bearing shales to be made through a correlation with the gaseous pyrolysis The amount of carbon monoxide evolved yields information on carbonate content and the amount of hydrocarbon gases enabled an assessment of the oil content to be made. The use of a P-induced luminescence detector enabled the moisture content to be determined. Other source-rock evaluations by Py - GC have also been undertaken and have enabled the charac- terisation of l i g n i t e ~ ~ ~ o and the petroleum-producing potential of kerogens to be assessed321 The release of hydrocarbons irom the rock at low temperatures and further yields of hydro- carbon from kerogen pyrolysis at higher temperatures form the basis of the studies.The pyrolysate can be fed directly through the detector and a study of the migration of petroleum, by a comparison with samples containing micro-reservoirs, has been ~ndertaken.~~Z A Work on isolated soil fractions has also been undertaken.312July, 1978 PYROLYSIS I N BIOMEDICAL STUDIES 697 Py - GC - MS comparison of kerogen-generated and solvent-extractable terpanes and steranes has a~peared,~*3 and the use of P y - GC - MS and Py - MS as fingerprint techniques in providing information concerning the composition and origin of kerogen in sediments has been de~cribed.32~ Coal has been assessed by analogous technique^,^^^^^^ by mass spectro- metry of the r e s i d ~ e s 3 ~ ~ and by Py - MS Py - GC using molten tin as the pyrolysis agent on coal of various ranks has been described.Arrlienius parameters for the production of benzene and toluene were calculated328 and catalytic conversion to carbon monoxide has enabled the oxygen content of coal to be mea~ured.9~ Pyrolysis - hydrogenation of coal leads to saturated or aromatic hydrocarbon fragments.329 Early work on the analysis of bitumens demonstrated that statistical evaluation of pyro- grams made characterisation possible.330 The identification of peaks due to saturated and unsaturated hydrocarbons has enabled an effective system to be developed for this purpose.331 I t has been shown that the ratio of the peak areas of unsaturated and saturated hydrocarbons (C3-C5) allows a distinction to be made between asplialtene (predominantly polyaromatic) and maltene (more paraffinic).Calibration of asphaltene content against this ratio has enabled the determination of asphalteiie content to be achieved. The determination of sulphur levels in petroleum has been described and use of Py - GC has been compared with the use of a dehydrosulphurisation - gas chromatographic technique,332 both with a sulphur detector. Many thiophenes and condensed analogues were resolved, with aliphatic com- ponents being converted into hydrogen sulphide. The production of carcinogenic compounds from organic matter in green sand has been monitored.333 Mechanisms In some instances of biomedical interest prior mechanistic knowledge has enabled the development of an analytical pyrolysis t e ~ h n i q u e . ~ ~ ~ ~ ~ ~ Most applications, however, have used this method as an empirical tool.Such work has involved either fingerprint comparisons or quantitative measurements on selected peaks in the pyrogram. Although the identity of the analytical peaks may have been established, a pviori conclusions concerning the structure of the material from which they arose are rare. Structures of polymeric materials have been i n v e ~ t i g a t e d ~ ~ ~ 3 ~ ~ 5 but many examples are restricted to chemically distinct species in which the origin, if not the detailed mechanism of formation, of important fragments can be estab- lished readily.70s109 Reasons for this restriction include the multiplicity of products derived from complex samples, the probability of varied secondary reactions occurring and the lack of an adequate molecular description of the organic component.These problems have not prevented the satisfactory application of analytical pyrolysis to analytical problems in the biomedical field but there is little doubt that further useful structural conclusions could be drawn if a greater insight into mechanistic aspects was available. Ideally, mechanistic studies should account for the production of the various fragments observed and also enable a prediction of the relative intensities to be made. It is doubtful Idlether this is an achievable goal in complex systems but with hydrocarbons there has been some success. A model for the pyrolysis behaviour of these compounds, based upon the thermodynamics and kinetics of homolysis and chin-propagation radical reactions, has enabled the vapour-phase pyrolysis products to be predicted ~ a t i s f a c t o r i l y .~ ~ , ~ ~ The use of filaments or Curie-point pyrolysers gave some similarities in their product distribution but was not modelled adequately. In view of the complexity of the over-all mechanistic approach to biomedical problems, work to date has been concerned essentially with the recognition of the origin of the various pyrolysis products. For simple molecules the origin of the major fragments can often be established without difficulty but detailed analysis by high-resolution mass ~ p e c t r o m e t r y ~ ~ ~ ~ ~ ~ and isotopic-labelling s t ~ d i e s l ~ ~ ~ l ~ ~ enables a greater understanding of the processes to be achieved. The main approach to the assignment of fragment origins has involved comparisons between the pyrograms from the test sample and those derived from isolated components or model compounds.For this technique to be successful a significant proportion of the pyrogram must result from the summation of the pyrograms of the components under test and must not suffer distortion due to the occurrence of new reactions.266 Assignments have been made on the basis of literature comparisons on1y1329133 but a more common practice is the com- parison of pyrograms from test and model components run under identical conditions. Many examples of this approach can be found in the literaturel*~~~~s312 and it has, for example, Reasons for these discrepancies were discussed.32698 IRWIN AND SLACK: ANALYTICAL Analyst, Vol. 103 enabled the variations in the pyrograms (Py - 11s) of streptococcal strains to be accounted for in terms of the antigenic polysaccharide components of the cell envel0pe.~~9~5 Comparisons with model synthetic polymers have been used to advantage.313 Comment has also been made concerning the :relationship between mass spectrometry and pyrolysis fragmentation reactions.336 This field has been reviewed337 and although in many instances the relationships are comparable it would appear that the different processes and conditions involved in the respective techniques would preclude a detailed ~orrelation.l6~,~~~ Indeed, in some instances pyrolysis has been sho,wn to differentiate between compounds that have very similar EI mass ~pectra.~~s However, with the various mass spectrometric tech- niques currently available differentiation by mass spectrometry also may now be possible.Data Handling Although treated last in this review, the manipulation of data obtained from pyrolysis to permit a valid interpretation of the results i:; of paramount importance. The degree of sophistication required depends upon the complexity of the pyrogram and the information required, but in simple instances of Py - GC, e.g., when unique fragments are observed or in the quantification of a major peak, the technique is comparable to a gas-chromatographic analy~is.~39J~O The separation parameters are chosen to ensure useful resolution and the system is then tuned to give adequate reproducibility and sensitivity. When comparisons of more complex pyrograms are required some numerical assessment is usually necessary and involves the selection of diagnostic peaks and their digital representation.Care must be taken to ensure a useful choice. Some peaks in a pyrogram may be more variable than others and their inclusion reduces the precision of a comparison. On the other hand there is no guarantee that peaks that show a high degree of reproducibility are the most useful in the differentiation of samples. The data can be gathered directly from the gas-chromatographic chart recorder or by means of automatic integrators or data-logging systems. In view of the complexity of a typical pyrogram care must be taken to ensure that errors due to variations in sample loading, retention times, estimation of base line and treatment of fused peaks are minimised, especially if automatic systems are in use.In all instances adequate presentation of the data is vital to the subsequent analysis. Pyrograms can be stored as the trace or as a selection of retention time - intensity data derived from the trace,3 as data on punched cards,44 a pyrogram ma~7~Yl61 or bar dia- grams,66~139,257~262 or as computerised data45,74,27L consisting of matrices of retention time - intensity measurements. With Py - MS the pyrograms are presented in a digitised form and are immediately suitable for computer manipulation. Automatic on-line treatment of such data can be undertaken readily. Computer programs that can be used to compare P y - GC data sets have been des- ~ribed.45*76,~~1#~30 In one such method45 the peaks are normalised with respect to sample mass, i.e., the ratio of peak height to total peak height is determined for each peak, and then the ratios of corresponding peaks (less intense : more intense) in test and stored pyrograms are deter- mined.For identical samples idealised ratios of unity would result whereas in non-matching samples a distribution of values considerably lower than this would be obtained. The degree of correspondence (similarity coefficient) is the mean of these ratios. A significant degree of similarity was suggested to be greater than 0.84 (using 13 peaks). has shown this to be a useful approach although doubt was expressed as to the validity of a one- dimensional parameter in the description of larger data sets.45 This method has also been used to analyse Py - MS data (with a slightly reduced ~oefficient)~~ but a more detailed multi-dimensional approach has also been descr.ibed.74 In this method the pyrograms are used to generate a Euclidean distance matrix in multi-dimensional space, which may be presented visually by means of a non-linear map.When weighting factors were introduced to account for reproducibility, i.e., within-sample variation (an inner variance term) and for the importance of the peaks in differentiation between samples, i.e., between-sample variation (an outer variance term) it was possible to differentiate strains of Listeria into one of the two serotypes although the pyrograms were veqr ~ i m i l a r . ~ ~ ~ ~ ~ ~ t-te~ts,~~O, cluster analysis, K-nearest neighbour and linear learning m e t h o d ~ , ~ 8 1 ~ 3 ~ ~ * ~ ~ ~ factor and multiple regression analyses343 and principal components and canonical variates analyses.344 Deriva- tions for quantitative interpretation have also appeared.94 Other Other methods of comparing pyrograms include the use of x2July, I978 PYROLYSIS IN BIOMEDICAL STUDIES 699 Conclusion In this review the importance of a critical appraisal of the literature on pyrolysis and the necessity for standardisation in any proposed application of analytical pyrolysis has been discussed. It is stressed that it has not been our intention to portray the effective use of analytical pyrolysis in biomedical studies as an unobtainable goal. Rather, we hope that in warning the analyst of possible problems and also demonstrating the diversity of analytical pyrolysis applications within this field we might stimulate the development of further useful applications of the technique.Clearly analytical pyrolysis has distinct advantages, not least of which concern the possibility of automatic processing of both samples and data. 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. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. References Levy, K. L., Chromat. Rev., 1966, 8, 48. Jones, C. E. R., and Guiochon, G., Chromatographia, 1972, 5, 483. May, R. W., Pearson. E. F., and Scothern, D., “Pyrolysis-Gas Chromatography,” Chenlical Jones, C. E. R., and Cramers, C. A., Editors, “Analytical Pyrolysis,” Elsevier, Amsterdam, 1977.Janbk, J., in Scott, R. P. W., Editor, “Gas Chromatography 1960,” Butterworths. London, 1960, Hewitt, G. C., and Whitham, B. T., Analyst, 1961, 86, 643. Perry, S. G., J . Gas Chromat., 1964, 2, 54. Kurashiki, H., Kogaku No Ryoiki, 1964, 18, 399. McKinney, R. W., J . Gas Chromat., 1964, 2, 432. Brauer, G. M., J . Polym. Sci. ( C ) , 1965, 8, 3. Perry, S. G., J . Gas Chromat., 1967, 5, 77. Kirk, P. L., J . Gas Chromat., 1967, 5, 11. Levy, R. L., J . Gas Chromat.. 1967, 5, 107. Stack, M. V., in Harbourn, C. L. A., Editor, “Gas Chromatography 1968,” Institute of Petroleum, Perry, S. G., Adv. Chromat., 1968, 7, 221. Kirk, P. L., Adv. Chromat., 1968, 5, 79. Stevens, M. P., “Characterisation of Polymers by Gas Chromatography,” Marcel Dekker, New McKinney, R.W., in Ettre, L. S., and McFadden, W. H., Editors, “Ancillary Techniques of Gas Wolf, C . J., Levy, R. L., and Walker, J. Q., I n d . Res., Chicago, 1971, 13, 40. Alishoev, V. R., Berezkin, V. G., and Malyshev, A. I., Zh. Analit. Khim., 1972, 27, 1035. Sarner, S. F., J . Chromat. Sci., 1972, 10, 65. Vanderborgh, N. E., and Ristau, W. T., Am. Lab., 1973, 41. Meuzelaar, H. L. C., Kistemaker, P. G., and Posthumus, M. A., Riomed. Mass Spectrom., 1974, 1, Kistemaker, P. G., Boerboom, A. J. H., and Meuzelaar, H. L. C., Dynamic Mass Spectrom., 1975, 4, Society, London, 1977. p. 387. London, 1969, p. 109. York, 1969, p. 64. Chromatography,” John Wiley, New York, 1969, p. 55. 312. 139. Quinn, P. A., in Johnston, H. H., and Newsom, S. W. B., Editors, in Microbiology,” Learned Information (Europe) , Ltd., 1976, p.Maugh, T. H., Science, N.Y., 1976, 194, 1403. Schmid, P. P., and Simon, W., Analytica Chim. Acta, 1977, 89, 1. Levy, li. L., Fanter, D. L., and Wolf, C. J., Analyt. Chem., 1972, Groenendyk, H., Levy, E. J.. and Sarner, S. F., .I. Chrotnat. Sci., Vanderborgh. N. E.. in Tones, C. E. R., and Cramers, C. A., “Rapid Methods and Automation 178. 44, 38. 1970, 8, 115. Editors, “Analytical Pyrolysis, ” “Analytical Pyrolysis,” Elsevier, R., and Cramers, C. A., Editors, Elsevier,-Amsterdam, 1577, p. 235. Amsterdam, 1977, p. 337. Crighton, J . S., in Jones, C. E. R., and Cramers, C. A., Editors, Walker, J. Q., and Wolf, C. J., J . Chromat. Sci., 1970, 8, 513. Uden, P. C., Henderson, D. E., and Lloyd, R. J., in Jones, C. E. Uden, P.C . , Henderson, D. E., and Lloyd, R. J., J . Chromat., 1976, 126, 225. Stahl, E., in Jones, C. E. R., and Cramers, C. A., Editors, “Analytical Pyrolysis,” Elsevier, Amster- Stahl, E., Accts Chem. Res., 1976, 9, 75. Levy, E. J., and Paul, D. G., J . Gas Chromat., 1967, 5, 136. Biemann, K., Oro’, J., Toulmin, P., 111, Orgel, L. E., Nier, A. O., Anderson, D. M., Simmonds, P. G., Levy, K. L., Wolf, C. J., Grayson, M. A., Gilbert, J., Gelpi, E., Updegrove, W. S., Zlatkis, A,, and Levy, R. L., Wolf, C. J., and Oro’, J., J . Chromat. Sci., 1970, 8, 524. Blau, K., and King, G. S., Editors, “Handbook of Derivatives for Chromatography,” Heyden, Langer, S. H., and Patton, J. E., in Purnell, H., Editor, “New Developments in Gas Chromato- “Analytical Pyrolysis,” Elsevier.Amsterdam, 1977, p. 351. dam, 1977, p. 29. Flory, D., Diaz, A. V., Rushnek, D. R., and Biller, J. A., Science, N.Y., 1976, 194, 72. Oro’, J., Nature, Lond., 1970, 227, 148. London, 1977. graphy,” John Wiley, New York, 1973, p. 293.700 43. 44. 45. IRWIN AND SLACK : ANALYTICAL Analyst, Vol. 103 Meuzelaar, H. I,. C., and Kistemaker, P. G., Annlyt. Chem., 1973, 45, 387. Zemany, P. D., Analyt. Chem., 1952, 24, 1709. Meuzelaar, H. L. C., Kistemaker, P. G., and Tornt, A., in HedCn, C.-G., and Illhi, T., Editors, “New Dawson, P. H., Editor, “Quadrupole Mass Spectrometry and its Applications,” Elsevier, Amsterdam, Meuzelaar, 13. L. C . , Posthumus, M. A., Kisternaker, P. G., and Kistemaker, J., Analyt. Chem., Schulten, H.-R., Beckey, H. D., Boerboom, A. J H., and Meuzelaar, H.L. C., Analyt. Chem., 1973, Posthumus, M. A., Nibbering, N. M. M., Boerboom, A. J . H., and Schulten, H.-R., Biomed. Mass Schulten, H.-R., in Jones, C. E. R., and Craniers, C. A., Editors, “Analytical Pyrolysis,” Elsevier, Levsen, K., and Schulten, H.-R., Biomed. iWass Spectrom., 1976, 3, 137. McLafferty, F. W., in Jones, C. E. R., and Cramers, C. A., Editors, “Analytical Pyrolysis,” Elsevier, Beckey, H. D., and Schulten, H.-R., Angew. Chem. Int. Edn, 1975, 14, 403. Schulten, H.-R., and Beckey, H. D., in West, A. R., Editor, “Advances in Mass Spectrometry,” Wheals, B. B., in Jones, C. E. R., and Cramers, C. A., Editovs, “Analytical Pyrolysis,” Elsevier, Hughes, J . C., Wheals, B. B., and Whitehouse, Id. J., Analyst, 1977, 102, 143. TydCn-Ericsson, E., Chromatographia, 1973, 6, 353.Gassiot-Matas, M., and JuliA-Dan&, E., Chromatt7graphia, 1972, 5, 493. Quinn, P. A., J . Chromat. S c i . , 1974, 12, 796. Biihler, Ch., and Simon, W., J . Chromat. Sci., 1970, 8, 323. Coupe, N. B., Jones, C. E. R., and Stockwell, P. B., Chronzatographia, 1973, 6, 483. Walker, J . Q., J . Chromat. Sci., 1977, 15, 267. Stewart, W. D., Jr., J . Ass. 08. Analyt. Chem., 1976, 59, 35. Reiner, E., Beam, R. E., and Kubica, G. P., Am. Rev. Resp. Dis., 1969, 99, 750. Levy, E. J., in Jones, C. $1. R., and Cramers, C. A., Editors, “Analytical Pyrolysis,” Elsevier, Wheals, B. B., and Nobel, W., Chromatographia, 1972, 5, 553. Lloyd, J. B. F., Hadley, K., and Roberts, B. R. G., J . Chromat., 1974, 101, 417. DeForest, P. R., and Kirk, P. L., Criminologist, 1973, 8, 35.Slack, J. A., PhD Thesis, University of Aston in Birmingham, 1977. Roy, T. A,, and Szinai, S. S., J . Chromat. Sci., 1976, 14, 580. Nelson, D. F., and Kirk, P. L., Analyt. Chem., 1964, 36, 875. Stack, M. V., Donoghue, H. D., Tyler, J. E., and Marshall, M., in Jones, C. E. R., and Cramers, Kingston, C. R., and Kirk, P. L., Bull. Narcot., 1965, 17, 19. Eshuis, W., Kistemaker, P. G., and Meuzelaar, H. L. C., in Jones, C . E. R., and Cramers, C. A., Editors, “Analytical Pyrolysis,” Elsevier, Amslzerdam, 1977, p. 151. Kistemaker, P. G., Meuzelaar, H. L. C., and Pxthumus, M. A., in HedBn, C.-G., and Illdni, T., Editors, “New Approaches to the Identification of Micro-organisms,” John Wiley, New York, 1975, p. 179. Menger, F. M., Epstein, G.A., Goldberg, D. A., and Reiner, E., AnaZyt. Chem., 1972, 44, 423. Posthumus, M. A., Boerboom, A. J. H., and Meuzelaar, H. L. C., in West, A. R., Editor, “-4dvances in Mass Spectrometry,” Volume 6, Applied Science, Barking, 1974, p. 397. Ericsson. I., PhD Thesis, University of Lund, Sweden, 1975. Lysyj, I., Newton, P. R., and Taylor, W. J., Analyt. Chem., 1971, 43, 1277. Brodasky, T. F., .f. Gas Chromat., 1967, 5, 311. Tai, H., Powers, R. M., and Protzman, T. F., Analyt. Chem., 1964, 36, 108. Stanford, F. G., Analyst, 1965, 90, 266. Szilagyi, P. I. A., Green, J. P., Brown, 0. M., and Margolis, S., J . Neurochem., 1972, 19, 2555. Schmidt, D. E., Szilagyi, P. I. A., and Green, I. P., J . Chromat. Sci., 1969, 7, 248. Stavinoha, W. B., and Weintraub, S. T., Analyt.Chem., 1974, 46, 757. PalmCr, L., and Holmstedt, B., Sci. Tools, 1975, 22, 25 and 38. Szilagyi, P. I. A., and Green, J. P., i n Jones, C.. E. R., and Cramers, C. A., Editors, “Analytical Polak, R. L., and Molenaar, P. C., J . Neurochem., 1974, 23, 1295. Fidone, S. J., Weintraub, S. T., and Stavinoha, W. B., J . Neurochem., 1976, 26, 1047. Thomas, R. J., Hagstrom, R. A., and Kuchar, E. J., Analyt. Chem., 1972, 44, 512. Ignasiak, B. S., Nandi, B. N., and Montgomery, D. S., Analyt. Chem., 1969, 41, 1676. Pella, E., and Andreoni, R., Mikrochim. Acta, 1976, 2, 175. Lysyj, I., and Nelson, K. H., Analyt. Chem., 1968, 40, 1365. Lysyj, I., Nelson, K. H., and Webb, S. R., Wat. Res., 1970, 4, 157. Lysyj, I., I n t . Lab., 1971, July-August, 33. Cannard, A. J , and Criddle, W.J., AnaZyst, 1975, 100, 848. Wickramasinghe, J . A. F., and Shaw, S. R., J . €’harm. Sci., 1971, 60, 1669. Wickramasinghe, J . A. F., and Shaw, S. R., J . Chromat., 1972, 71, 265. Winkler, V. W., Regnier, F. E., Yoder, J. M., an13 Macy, L. R., J . Pharm. Sci., 1972, 61, 1462. Approaches to the Identification of Micro-organisms,” John Wiley, New York, 1975, p. 165. 1976. 1973, 45, 1546. 45, 2358. Spectrom., 1974, 1, 352. Amsterdam, 1977, p. 17. Amsterdam, 1977, p. 39. Volume 6, Applied Science, Barking, 1974, p. 499. Amsterdam, 1977, p. 89. Amsterdam, 1977, p. 319. C. A., Editors, “Analytical Pyrolysis,” ElsevieI, Amsterdam, 1977, p. 57. Pyrolysis,” Elsevier, Amsterdam, 1977, p. 417. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65.66. 67. 68. 69. 70. 71. 72. 73. 74. 75. 76. 77. 78. 79. 80. 81. 82. 83. 84. 85. 86. 87. 88. 89. 90. 91. 92. 93. 94. 95. 96. 97. 98. 99.July, 1978 PYROLYSIS I N BIOMEDICAL STUDIES 701 100. 101. 102. 103. 104. 105. 106. 107. 108. 109. 110. 111. 112. 113. 114. 115. 116. 117. 118. 119. 120. 121. 122. 123. 124. 125. 126. 127. 128. 129. 130. 131. 132. 133. 134. 135. 136. 137. 138. 139. 140. 141. 142. 143. 144. 145. 146. 147. 148. 149. 150. 151. 152. 153. 154. 155. 156. 157. 158. 159. 160. 161. 162. 163. Levitt, ill. J., Josimovich, J. B., and Broskin, K. D., Prostaglandins, 1972, 1, 121. Dell, D., Uoreham, D. R.. and Martin, B. K., J . Pharm. Sci., 1971, 60, 1368. Chafetz, L., Elefant, M., and Amin, P. K., J . Pharm. Sci., 1967, 56, 1528. van den Heuvel, W. J.A., and Zaccei, A. G., Adv. Chromat., 1976, 14, 199. Radecka, C., and Nigam, I. C., J . Pharm. Sci., 1967, 56, 1608. Pella, E., and Colombo, B., Analyt. Chem.. 1972, 44, 1563. Burrows, H. J., and Calam, D. H., J . Chromat., 1970, 53, 566. Westley, J. W., Evans, R. H., Jr., and Stempel, A., Analyt. Biochem., 1974, 59, 574. Irwin, W. J., and Slack, J. A., in Jones, C. E. R., and Cramers, C. A., Editors, “Analytical Pyrolysis,” Irwin, W. J., and Slack, J. A., J . Chromat., 1977, 139, 364. Slack, J . A., and Irwin. W. J., Proc. Analyt. Div. Chem. Soc., 1977, 14, 249. Auterhoff, H., and Schmidt, U., Dt. ApothZtg, 1974, 114, 1581. Hofmann, V., Przybylski, M., Kingsdorf, H., and Ritter, H., Makromolek. Chem., 1976, 177, 1791. JanAk, J., Nature, Lond., 1960, 185, 684. Nelson, D.F., and Kirk, P. L., Analyt. Chem., 1962, 34, 899. Barbour, W. M., J . Gas Chromat., 1965, 3, 228. van Binst, G., Dewaersegger, L., and Martin, R. H., J . Chromat., 1966, 25, 15. Severson, R. F., Schlotzhauer, W. S., Arrendale, R. F., Snook, M. E.. and Higman, H. C., Beitr. Stedman, R. L., Chem. Rev., 1968, 68, 153. Schmeltz, I., and Hoffmann, D., Chem. Rev., 1977, 77, 295. Chopra, N. M., and Osborne, N. B., Analyt. Chem., 1971, 43, 849. Chopra, N. M., and Domanski, J . J . , Beitr. Tabakforsch., 1972, 6, 139. Balasubrahmanyam, S. N., and Quin, L. D., Tob. Sci., 1962, 6, 133. Edmunds, F. S., and Johnstone, R. A. W., J . Chem. Soc., 1965, 2892. Fontan, C. R., Jain, N. C., and Kirk, P. L., Mikrochim. Acta, 1964, 326. Szinai, S. A., and Roy, T. A., J . Chromat.Sci., 1976, 14, 327. Kiippers, F. J . E. M., Bercht, C. A. L., Salemink, C. A., Lousberg, 13. J. J. Ch., Terlouw, J. K., and Slack, J . A., and Irwin, W. J., Proc. Analyt. Div. Chem. Soc., 1977, 14, 216. Keiner, E., in Jones, C. E. R,, and Cramers, C. A., Editors, “Analytical Pyrolysis,” Elsevier, Brodowski, P. T., Wilson, N. B., and Scott, W. J., Analyt. Chem., 1976, 48, 1812. Hoffmann, D. A., and Fitz, R. A., Environ. Sci. Technol., 1968, 2, 1023. Sinn, II., Kaminsky, W., and Janning, J., Angew. Chem. I n t . Edn, 1976, 15, 660. Simmonds, P. G., Shulman, G. P., and Stembridge, C. H., J . Chromat. Sci., 1969, 7, 36. Simmonds, P. G., Appl. Microbiol., 1970, 20, 567. Stack, M. V., J . Gas Chromat., 1967, 5 , 22. olehla, J., ZivoEiSnd V/’3iroba, 1960, 5, 567; Chem.Abstr., 1961, 55, 5242g. Kanomsta, K., and Mashiko, Y., Nippon Kagaku Zasslai, 1966, 87,57; Chem. Abstr., 1966,65, 12542g. Winter, L. N., and Albro, P. W., J . Gas Chromat., 1964, 2, 1. Giacobbo, H., and Simon, W., Pharm. Acta Helv., 1964, 39, 162. Voellmin, J . , Icriemler, P., Omura, I., Seibl, J., and Simon, W., Microchem. J . , 1966, 11, 73. Simon, W., and Giacobbo, H., Angew. Chem. I n t . E d n , 1965, 4, 938. Simon, W., Kriemler, P., Voellmin, J. A., and Steiner, H., J . Gas Chromat., 1967, 5, 53. Merritt, C., Jr., and Robertson, D. H., J . Gas Chromat., 1967, 5, 96. Shulman, G. P., and Simmonds, P. G., Chem. Commun., 1968, 1040. Ratcliff, M. A., Jr., Medley, E. E., and Simmonds, P. G., J . Org. Chem., 1974, 39, 1481. Patterson, J. M., Haidar, N. F., Papadopoulos, E.P., and Smith, W. T., Jr., J . Org. Chem., 1973, Lien, Y. C., and Nawar, W. W., J . Fd Sci., 1974, 39, 911. Lien, Y. C., and Nawar, W. W., J . Fd Sci., 1974, 39, 914. Johnson, W. R., and Kang, J. C., J . Org. Chem., 1971, 36, 189. Simmonds, P. G., Medley, E. B., Ratcliff, M. A., Jr., and Shulman, G. P., Analyt. Chem., 1972, 44, Bier, M., and Teitelbaum, P., Ann. N.Y. Acad. Sci., 1959, 72, 641. Patterson, J. M., Chen, W.-Y., and Smith, W. T., Jr., Tob. Sci., 1971, 15, 41. Lien, Y. C., and Nawar, W. W., J , Fd Sci., 1974, 39, 917. Fujimaki. M., Kato, S., and Kurata, T., Agric. Biol. Chem., 1969, 33, 1144. Kato, S., Kurata, T., Ishitsuka, R., and Fujmaki, M., Agric. Biol. Chem., 1970, 34, 1826. Sugimura, T., Kawachi, T., Nagao, M., Yahagi, T., Seino.Y., Okamoto, T., Shudo, K., Kosuge, T., Kojima, T., and Morishita, F., J . Chromat. Sci., 1970, 8, 471. Means, J. C., and Perkins, E. G., in Jones, C. E. R., and Cramers, C. A., Editors, “Analytical Lawless, J. G., Geochim. Cosmochim. Acta, 1973, 37, 2207. Vallentyne, J. R., Geochim. Cosmochim. Acta, 1968, 32, 1353. Olafsson, P. G., and Bryan, A. M., Geochim. Cosmochim. Acta, 1971, 35, 327. Merritt, C., Jr., Dipietro, C., Robertson, D. H., and Levy, E. J., J . Ghromat. Sci., 1974, 12, 668. Posthumus, M. A., Nibbering, N. M. M., and Boerboom, A. J . H., Org. Mass Spectrom., 1976, 11, 907. Posthumus, M. A., and Nibbering, N. M. M., Oyg. Mass Spectrom., 1977, 12, 334. Elsevier, Amsterdam, 1977, p. 107. Tabakforsch., 1977, 9, 23. Heerma, W., J . Chromat., 1975, 108, 375.Amsterdam, 1977, p. 49. 38, 663. 2060. Tsuji, K., Wakabayashi, K., Iitaka, Y., and Itai, A., Proc. Japan Acad., 1977, 53, 58. Pyrolysis,” Elsevier, Amsterdam, 1977, p. 249.702 164. 165. IRWIK AND SLACK: ANALYTICAL Analyst, Vol. 103 Mauger, A. B., Chem. Commun., 1971, 39. SchInid, J. P., Schmid, I-’. P., and Simon, W., 21% Jones, C. E. It., and Cramers, C. A,, Editors, Stack, M. V., Biochem. J . , 1965, 96, 56P. Oyania, V. I., Nature, Lond., 1963, 200, 1058. Groten, B., Analyt. Chem., 1964, 36, 1206. Higman, E. B., Schmeltz, I., and Schlotzhauer, W . S., J . Agric. Fd Chem., 1970, 18, 636. Myers, A., and Smith, R. N. L., Chromatographia, 1972, 5, 521. Bayer, I;. L., Hopkins, J . J., and Menger, F. M., in Jones, C. E. R., and Cramers, C. A., Editors, “Analytical Pyrolysis,” Elsevier, Amsterdam, 1077, p.217. Fagerson, I. S., J . Agric. Fd Chem., 1969, 17, 747. Greenwood, C. T., Knox, J . H., and Milne, E., Chcmy I d . , 1961, 1878. Ettre, K., and Vgradi, P. F., Analyt. Chem., 1962, 34, 752. Berry, J . W.. Adv. Chromat., 1966, 2, 284. Shafizadeh, F., Adv. Carbohyd. Chem., 1968, 23, 419. Shafizadeh, F., and Lai, Y.-Z., Carbohyd. Res., 1975, 40, 263. Shafizadeh, F., Appl. Polym. Symp., 1975, 28, 153 Shafizadeh, F., and Fu, Y . L., Carbohyd. Res., 1973, 29, 113. Cope, J . F., Analyt. Chem., 1973, 45, 562. Shafizadeh, F., Philpot, C. W., and Ostojic, N., Carbohyd. Res., 1971, 16, 279. Shafizadeh, F., and Chin, P. P. S., Carbohyd. Xes., 1976, 46, 149. Glassner, S., and Pierce, A. K., 111, Analyt. Chem., 1965, 37, 525.Heyns, K., and Klier, M., Carbohyd. Res., 1968, 6, 436. Walter, R. H., and Fagerson, I . S., J . Fd Sci., 1968, 33, 294. Heyns, K., Stute, R., and Paulsen, H., Carbohyd. Res., 1966, 2, 132. Robb, E. W., Johnson, W. R., Westbrook, J . J., and Seligman, R. H., Beitr. Tabakforsch., 1966, 3, Prey, V., Eichberger, W., and Gruber, H., Sturke, 1977, 29, 60. Prey, V., and Gruber, H., Starke, 1977, 29, 96. Prey, V., and Gruber, II., Stiirke, 1977, 29, 135. Ohnishi, A., Kato, K., and Takagi, E., Polym. J . , 1975, 7, 431. Ohnishi, A., Takagi, E., and Kato, K., Carbohyd. 1?es., 1976, 50, 275. Greenwood, C. T., Adv. Cavbohyd. Chem., 1967, 22, 483. Sawardeker, J . S., Sloneker, J . H., and Dimler, R J., J . Chronzat., 1965, 20, 260. Schlotzhauer, W. S., Chortyk, 0. T., and Austin, P.R., J . Agric. Fd Chem., 1976, 24, 177. Nichols, H. W., Anderson, D. J., Shaw, J . I., and Sommerfeld, M. R., J . Phycol., 1968, 4, 362. Zamorani, A., Roda, G., and Lanzarini, G., Industrie Agric., 1971, 9, 35. Herlitz, E., Lofroth, G., and Widmark, G., Acta Chem. Scand., 1965, 19, 595. Haider, K., Nagar, B. R., Saiz, C., Meuzelaar, H L. C., and Martin, J . P., Proc. Fd -4gric. Org., Weijman, A. C. M., J . Microbiol. Serol. (Antonie van Leeuwenhoek), 1976, 42, 315. Schulten, H. K., i n HedBn, C.-G., and IlICni, T., Editors, “New Approaches to the Identification of Jennings, E. C., Jr., and Dimick, K. P., Analyt. Chem., 1962, 34, 1543. Turner, L. P., Analyt. Biochem.. 1968, 28, 288. Turner, L. P., and Barr, W. R., J . Chromat. Sci., 1971, 9, 176. Charnock, G.A., and Loo, J . L., Analyt. Biochem., 1970, 37, 81. Wiebers, J. L., Analyt. Biochem., 1973, 51, 542. Schulten, H.-R., and Beckey, H. D., Org. Mass Sfiectrom., 1973, 7 , 861. Nawar, W. W., J . Agric. Fd Chem., 1969, 17, 18. Perkins, E. G., Fd Technol., Champaign, 1967, 21, 125. Perkins, E. G., Fd Technol., Chamfiaign, 1960, 14, 508. Artman, N. R., Adv. Lipid Res., 1969, 7, 245. Perkins, E. G., and Johnston, P. V., Lipids, 1969, 4, 301. Crossley, A., Heyes, T. D., and Hudson, B. J. F., J . Am. Oil Chem. Soc., 1962, 39, 9. Higman, E. B., Schmeltz, I., Higman, H. C., anti Chortyk, 0. T., J . Agric. Fd Chem., 1973, 21, Karmen, A., Walker, T., and Bowman, R. C., J . Lipid Res., 1963, 4, 103. Hirayama, S., and Imai, C., J . Am. Oil Chem. SOG, 1977, 54, 190. Holman, K.T., Deubig, M., and Hayes, H., Lipids, 1966, 1, 247. Denker, W. D., and Wolf, C. J., J . Chromat. Sci., 1970, 8, 534. Liddicoet, T. H., and Smithson, L. H., J . Am. Oil Chem. Soc., 1965, 42, 1097. Nakagawa, T., Miyajima, K., and Uno, T., J . Chromat. Sci., 1970, 8, 261. Hites, R. A., and Biemann, K., J . Am. Chem. Soc., 1972, 94, 5772. Etemadi, A. H., J . Gas Chromat., 1967, 5, 447. Davison, V. L., and Dutton, H. J., Analyt. Chem., 1966, 38, 1302. Privett, 0. S., Prog. Chem. Fats, 1966, 9, 106. Downing, D. T., and Green, R. S., Lipids, 1968, 3, 96. Horning, M. G., Casparinni, G., and Horning, E. C., J . Chromat. Sci., 1969, 7, 267. Gellerman, J. L., and Schlenk, H.. Analyt. Chem., 1966, 38, 72. Schwarze, J. W.. and Gilmour, M. N., Analyt. Che:vn., 1969, 41, 1686.Lloyd, J. B. F., and Roberts, B. R. G., J . Chromai!., 1973, 77, 228. “Analytical Pyrolysis,” Elsevier, Amsterdam, 1977, p. 99. 597. 1976. Micro-organisms,” John Wiley, New York, 1975, p. 155. 202. 166. 167. 168. 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. 198. 199. 200. 201. 202. 203. 204. 205. 206. 207. 208. 209. 210. 211. 212. 213. 214. 215. 216. 217. 218. 219. 220. 221. 222. 223. 224. 225. 226. 227. 228. 229.July, 1978 PYROLYSIS I N BIOMEDICAL STUDIES 703 230. 231. 232. 233. 234. 235. 236. 237. 238. 239. 240. 241. 242. 243. 244. 245. 246. 247. 248. 249. 250. 251. 252. 253. 254. 255. 256. 257. 258. 259. 260. 261. 262. 263. 264. 365. 266. 267. 268.269. 270. 271. 272. 273. 274. 275. 276. 277. 278. 279. 280. 281. 282. 383. 284. 285. 286. 287. 288. Cox, J. S. G., High, L. B., and Jones, E. R. H., Proc. Chem. SOL, 1958, 234. Blondin, G. A., Kulkarni, B. D., John, J . P., van Aller, R. T., Russell, P. T., and Nes, W. R., Gassiot-Matas, M., and Julia-Dan&, E., Chromatographia, 1976, 9, 151. Bleau, G., and VandenHeuvel, W. J . A., Steroids, 1974, 24, 549. Poole, C. F., and Morgan, E. D., J . Chromat., 1974, 90, 380. Vanderborgh, N. E., Hanson, R. L., and Brower, C., Analyt. Chem., 1975, 47, 2277. Falk, H. L., Goldfein, S., and Steiner, P. E., Cancer Res., 1949, 9, 438. Levy, R. L., Gesser, H., Halevi, E. A., and Saidman, S., J . Gas Chvomat., 1964, 2, 254. Whitten, D. G., Bentley, K. E., and Kuwada, D., J . Org.Chem., 1966, 31, 322. Wist, A. O., J . Gus Chromat., 1967, 5, 157. Dhont, J . H.. Nature. Lond., 1963, 200, 882. Roder, E., and Stuthe, W., 2. Analyt. Chem., 1973. 266, 358. Szilagyi, P. I. A., Schmidt, D. E., and Green, J. P., Analyt. Chew., 1968, 40, 2009. Ohya, K., and Sano, M., Biomed. Mass Spectrom., 1977, 4, 241. Lanzarini, G., Morselli, L., Pifferi, P. G., and Giumanini, A. G., J . Chromat., 1977, 130, 261. HedCn, C.-G., and Illhi, T., Editors, “New Approaches to the Identification of Micro-organisms,” Johnston, H. H., and Newsom, S. W. B., Editors. “Rapid Methods and Automation in Microbiology,” Larsson, L., and Mardh, P.-A.. Acta Path. Micvobiol. Scand. Sect. B., SuppZ., 1977, 259, 5. Abel, K., Desmertzing, H., and Peterson, J . I., J . Bact., 1963, 85, 1039.Jantzen, E., Bryn, K., and Bovre, K., Acta Path. Microbiol. Scand. Sect. B., 1974, 82, 753. Mitruka, €3. M., Kundargi, R. S., and Jonas, A. M., Med. Res. Engng, 1972, 11, 7. Mitruka, B. M., “Gas Chromatographic A4pplications in Microbiology and Medicine,” John Wiley, Medley, E. E., Simmonds, P. G., and Manatt, S. L., Biomed. Mass Spectrom., 1975, 2, 261. Lambert, N. G., Petrow, S., Kasatiya, S. S., and Bouliane, M., Apznls Microbiol. (Paris), 1976, 127B, 309. Oxborrow, G. S., Fields, N. D., and Puleo, J . It, in Jones, C. E. R., and Cramers, C. A., Editors, “Analytical Pyrolysis,” Elsevier, Amsterdam, 1977, p. 69. Oxborrow, G. S., Fields, N. D., and Puleo, J . R., Appl. Environ. Microbiol., 1977, 33, 865. Walker, J. Q. Chromafographia, 1972, 5, 547. Oyania, V.I., and Carle, G. C., J . Gas Chromat., 1967, 5, 151. Reiner, E., Nature, Lond., 1965, 206, 1272. Reiner, E., J . Gas Chvomat., 1967, 5, 65. Reiner, E., Hicks, J. J., Beam, R. E., and David, H. L., A m . Rev. Resp. Dis., 1971, 104, 656. Reiner, E., and Ewing, W. H., Nature, Lond., 1968, 217, 191. Reiner, E., and Kubica, G. P., Am. Rev. Resp. Dis., 1969, 99, 42. Wickman, K., Acta Path. Microbiol. Scand. Sect. B., Suppl., 1977, 259, 49. Cone, R. D., and Lechowich, R. V., Appl. Microbiol., 1970, 19, 138. Meuzelaar, H. L. C., and Veld, R. A., J . Chromat. Sci., 1972, 10, 213. Veld, J . H. J., Meuzelaar, H. L. C., and Tom, A.. Appl. MicrobioZ., 1973, 26, 92. Meuzelaar, H. L. C., Ficke, H. G., and den Harink, H. C., J . Claromat. Sci., 1975, 13, 12. Haddadin, J.M., Stirland, R. M., Preston, N. W., and Collard, P., Appl. Microbiol., 1973, 25, 40. Drucker, D. B., Holmes, C., Stack, M. V., Donoghue, H. D., Tyler, J . E., and Marshal, M., J . Appl. Stack, M. V., Donoghue, H. D., Tyler, J. E., and Marshall, M., J . Appl. Bact., 1975, 39, viii. Needleman, M., and Stuchbery, P., in Jones, C. E. R., and Cramers, C. A., Editors, “Analytical Oxborrow, G. S., Fields, N. D., and Puleo, J . R., APpZ. Environ. MicrobioZ., 1976, 32, 306. Ericsson, I., Larsson, L., and Mardh, P.-A., Acta Path. Microbiol. Scand. Sect. B., Suppl., 1977, Meuzelaar, H. L. C., and Posthumus, M. A., Chem. Engng News, 1973, June 18th, 23. Schulten, H.-R., Beckey, H. D., Meuzelaar, H. L. C., and Boerboom, A. J. H., Anulyt. Chem., 1973, Anhalt, J. P., and Fenselau, C., Analyt.Chem., 1975, 47, 219. Risby, T. H., and Yergey, A. L., J . Phys. Chem., 1976, 80, 2839. Meuzelaar, H. L. C., Kistemaker, P. G., Eshuis, W., and Boerboom, A. J. H., Adv. Mass Spectrom., 1976, 7B, 1452. Meuzelaar, H. L. C., Kistemaker, P. G., Eshuis, W., and Engel, H. W. B., in Johnston, H. H., and Newsom, S. W. B., Editors, “Rapid Methods and Automation in Microbiology,” Learned Informa- tion (Europe) Ltd., 1976. Sekhon, A. S., and Carmichael, J. W., Can. J . Microbiol., 1972, 18, 1593. Carmichael, J. W., Sekhon, A. S., and Sigler, L., Can. J . Microbiol., 1973, 19, 403. Burns, D. T., Stretton, R. J., and Jayatilake, S.D.A.K., J . Chromat., 1976, 116, 107. Stretton, R. J., Campbell, M., and Burns, D. T., J . Chromat., 1976, 129, 321. De Hoog, G. S., Stud. Mycol., 1977, 15, 1. Weijman, A. C. M., in Jones, C. E. R., and Cramers, C. A., Editors, “Analytical Pyrolysis,” Elsevier, Hall, R. C., and Bennett, G. W., J . Chromat. Sci., 1973, 11, 439. Darenbach, J. B., and Ehrhardt, M., J . Chromat., 1975, 105, 339. Myers, A., and Watson, L., Nature, Lond., 1969, 223, 964. Analyt. Chem., 1967, 39, 36. John Wiley, New York, 1975. Learned Information (Europe) Ltd., 1976. New York, 1975. Bact., 1975, 39, viii. Pyrolysis,” Elsevier, Amsterdam, 1977, p. 77. 259, 43. 45, 191. Amsterdam, 1977, p. 225.704 IRWIN AND SLACK 289. 290. 291. 292. 293. 294. 295. 296. 297. 298. 299. 300. 301. 302. 303. 304. 305. 306. 307. 308. 309. 310. 311. 312. 313. 314. 315. 316. 31 7. 318. 319. 320. 321. 322. 323. 324. 325. 326. 327. 328. 329. 330. 331, 332. 333. 334. 335. 336. 337. 338. 339. 340. 341. 342. 343. 344. Reiner, E., Beam, K. E., and Kubica, G. P., J . Chromat., 1967, 27, 495. Jellum, E., Stokke, O., and Eldjarn, L., Clin. Chenz., 1972, 18, 800. Jellum, E., J. Chromat., 1977, 143, 427. Pierce, S. K.. Gearhart, H. L., and Payne-Bose, D., Talanta, 1977, 24, 473. Krotoszynski, B., Gabriel, G., and O’Neill, H., J . Chromat. Sci., 1977, 15, 239. Keiner, E., and Hicks, J. J., Chromatographia, 19712, 5, 525, Reiner, E., and Hicks, J. J., Chromatographia, 1972, 5, 529. Schulten, H.-R., in Glick, D., Editor, “Methods in Biochemical Analysis,” Volume 24, John Wiley, Hayes, J. M., and Biemann, K., Geochim. Cosmochdnz. Acta, 196S, 32, 239. Biemann, K., Origins Life, 1974, 5, 417. Owen, T., and Biemann, K., Science, N . Y . , 1976, 193, 801. Biemann, K., Owen, T., Rushneck, D. R., LaFleur, A. L., and Howarth, D. W., Science, N . Y . , 1976, 194, 76. Owen, T., Biemann, K., Rushneck, D. R., Biller, J. E., Hownrth, D. W., and LaFleur, A. L., Science, N.Y., 1976, 194, 1293. Science, N.Y.. 1970, 167, 418-784. Or6, J., Gibert, J., Updegrove, W., 1LlcReynolds, J., Ibmiez, J., Gil-Av, I;., Flory, D., and Zlatkis, Nagar, B. R., Nature, Lond., 1963, 199, 1213. Bracewell, J. M., Geoderma, 1971, 6, 163. Bracewell, J. M., and Robertson, G. W., J . Soil Sci., 1973, 24, 421. Bracewell, J. M., Robertson, G. W., and Stephen, (2. J . M., J . Soil Sci., 1975, 26, 62. Bracewell, J. M., and Robertson, G. W., in Jones, C:. E. K., and Cramers, C. :I., Editors, “Analytical Bracewell, J. M., and Robertson, G. W., J . Soil Sci., 1976, 27, 196. Kagg, J. M., Bracewell, J. M., Logan, J., and Robertson, L., J . Soil Sci., 1978, 29, in the press. Bracewell, J. M., Robertson, G. W., and Tate, K. R., Geoderma, 1976, 15, 209. Kimber, R. W. L., and Searle, P. L., Geoderma, 19’10, 4, 57. Meuzelaar, H. L. C., Haider, K., Nagar, R. R., and Martin, J . P., Geodevma, 1977, 17, 239. Nagar, B. K., Waight, E. S., Meuzelaar, H. L. C., and Kistemaker, P. G., PI. Soil, 1975, 43, 681. Martin, F., in Jones, C. E. R., and Cramers, C. A., Editors, “Analytical Pyrolysis,” Elsevier, Amster- Martin, F., Geoderma, 1976, 15, 253. Biscar, J. P., J . Chromat., 1971, 56, 348. Tlanson, R. L., Vanderborgh, N. E., and Brookins, D. G., Analyt. Chem., 1975, 47, 335. Hanson, R. L., Brookins, D. G., and Vanderborgh, N. E., Analyt. Chem., 1976, 48, 2210. Goodwin, J. E., Giam, P. Y., Rion, K. F., and Smith, S. G., Analyt. Chem., 1977, 49, 1540. Larter, S. R., Horsfield, B., and Douglas, A. G., in Jones, C. E. K., and Cramers, C. A,, Editors, Barker, C., Am. Ass. Pekol. GeoZ. BuZl., 1974, 58, 2349. Gallegos, E. J., AnaZyt. Chem., 1975, 47, 1524. Maters, W. L., Meent, D. V. D., Schuyl, P. J . W., De Leeuw, J. JV., Schenk, P. h., and Meuzelaar, H. L. C., in Jones, C . E. K., and Cramers, C. A., Editors, “Analytical Pyrolysis,” lclsevier, Amsterdam, 1977, p. 203. New York, 1977, p. 313. A., J , Chromat. Sci., 1970, 8, 297. Pyrolysis,” Elsevier, Amsterdam, 1977, p. 167. dam, 1977, p. 179. “Analytical Pyrolysis,” Elsevier, Amsterdam, 19’77, p. 189. Hanson, R. L., Vanderborgh, N. E., and Brookins, D. G., Analyt. Chena., 1977, 49, 390. Karn, F. S., Friedel, R. A., and Sharkey, A. G., J r . , Fuel, Lond., 1972, 51, 113. Joy, W. K., Ladner, W. I<., and Pritchard, E., Fuel, Lond., 1970, 49, 26. RomovbEek, J., and Kubat, J., Analyt. Chem., 1968, 40, 1119. McHugh, D. J., Saxby, J. D., and Tardif, J. W., Chem. Geol., 1976, 17, 243. Poxon, D. W., and Wright, R. G., J . Chromat., 1971, 61, 142. Ramljak, Z., Deur-Siftar, D., and solc, A., J. Chvoimat., 1976, 119, 445. Drushel, H. V., AnaZyt. Chem., 1969, 41, 569. Gwin, C. H., Scott, W. D., and James, K. H., Am. Ind. Hyg. Ass. J . , 1976, 37, 685. Tsuchida, H., Komoto, M., Kato, H., Kurata, T., and Fujirnaki, M., Agric. Biol. Chcin., 1976, 40, Heyns, K., and Hauber, R., Justzts Liebigs Annh Chem., 1970, 733, 159. Kelley, J. D., and Wolf, C. J., J . Chromat. Sci., 19’70, 8, 583. Dougherty, 33. C., Topics Curr. Chem., 1974, 45, 93. Fanter, D. L., Walker, J. Q.. and Wolf, C. J . , A n d y t . Chem., 1968, 40, 2168. Novsk, J., Adv. Chromat., 1974, 11, 1. Leathard, D. A., Adv. Chromat., 1975, 13, 265. Kullik, E., Kaljurand, M., and Koel, M., J . Chromat., 1975, 112, 297. Kiillik, E., Kaljurand, M., and Koel, M.. J . Chromak, 1976, 126, 249. Martens, A. J., and Glas. J., Chromatographia, 1972, 5, 508. Gutteridge, C. S., in Johnston, H. H., and Newsom, S. W. B., Editors, “Rapid Methods and Automa- Received November 14th, 1977 Accepted December Sth, 1977 2051. tion in Microbiology,” Learned Information (Europe) Ltd., 1976, p. 100.

ISSN:0003-2654    

DOI:10.1039/AN9780300673

出版商:RSC    

年代:1978

数据来源: RSC

摘要:

Analyst, July, 1978, Vol. 103, pp. 705-713 Simultaneous Determination of Trace Concentrations of Arsenic, Antimony, Bismuth, Selenium and Tellurium in Aqueous Solution by 705 Introduction of the Gaseous Hydrides into an Inductively Coupled Plasma Source for Emission Spectrometry Part 11." Interference Studies M. Thompson, B. Pahlavanpour and S . J. Walton and G. F. Kirkbright Applied Geochemistry Research Group, Department of Geology, Imperial College, London, S W7 2BP Department of Chemistry, Iinperial College, London, S W7 2A Y In the simultaneous determination of arsenic, antimony, bismuth, selenium and tellurium by measurement of the atomic line emission from the elements in an inductively coupled plasma source after their introduction as the gaseous hydrides, two types of possible interferences were investigated.Mutual inter- ferences between the analytes were found to be negligible under the operating conditions used. The presence of certain metal ions in solution resulted in low recoveries of some of the analytes, especially selenium ancl tellurium. However, this could be overcome by prior separation of the analytes by co-precipitation on lanthanum hydroxide. Keywords Arsenic, antimony, bismuth, selenium and tellurium determination ; geocheniical saiq!des ; hydride generation ; inductively coupled plasma einission spectrometry ; interferences In Part I the initial development of a system of spectrochemical analysis involving the generation of gaseous hydrides followed by excitation in an inductively coupled plasma (ICP) source was described.Although emission spectroscopy with this source is reported to be relatively free from interference effects,I at least when sample introduction by means of nebulisation is utilised, methods involving the generation of gaseous hydrides and subsequent measurement by atomic absorption have been shown to have various disadvantage^.^^^ The second part of this study was therefore devoted to investigations of these disadvantages and of means of overcoming them. The interferences involved fall into two classes, i.e., those due to changes in the efficiency of the reduction and those due to compound formation between the reduced elements in the atom cell. For those elements which form gaseous hydrides, the two effects cannot be differentiated experimentally unless separate hydride generators are used.In this study it was more convenient to consider (i), the mutual effects between the hydride-forming elements, and (ii), effects due to other substances, without regard to the cause. Interferences due to compound formation in the plasma tail flame were considered unlikely to be significant and mutual interference generally would have to be low for a multi-element method to be viable. Mutual Effects Between Volatile Hydride-forming Elements Interference studies often involve possible interferents whose effects are usually considered only singly and at a few selected concentration levels. This approach suffers from several defects, viz., possible synergistic effects between two or more interferents remain unknown, the full range of expected concentration ratios is seldom covered and, in multi-element * For Part I of this serics, see Analyst, 1978, 103, 565.Ti06 THOMPSON et al.: SIMULTANEOUS TRACE DETERMINATION OF As Analyst, VoZ. 103 mutual interference studies, a considerable number of element combinations would, in principle, need to be studied in order to effect complete coverage. We have utilised an experimental design, to our knowledge not previously applied in analytical interference studies, with which it is possible to overcome these objections efficiently when interference is slight and to ascertain where further detailed study is required in more severe instances. Method Fifty solutions were prepared, each of which contained a level of gerinanium(IV), tin(II), antimony(III), bismuth( 111), selenium(1V) and tellurium( IV) selected independently and at random from the following: 1.0, 2.0, 5.0, 10.0, 20.0 .. . 500.0 and 1000 ng ml-l. Levels of arsenic(II1) were similarly selected, but from the range 10.0-10000 ng ml-l. These levels represented ranges that might reasonably be expected in solutions resulting from geochemical samples and also correspond roughly with the established linear ranges of cali- bration from the detection limit upwards. Lead was excluded as it did not appear to be reduced under the experimental conditions. ‘The elements were then reduced to the hydrides in the manner described in Part I using the selected experimental conditions, the resulting hydrides were passed into the ICP and the response for each of the elements was recorded. The results were analysed statistically by means of stepwise multiple linear regression, using the instrument response for each element in succession as the dependent variable and the known concentrations of all the elements as independent variables.The regression was performed with a slightly modified version of the program STEPR from the IBM Scientific Subroutine Package. Stepwise regression with n independent variables carries out a regressions, starting with the most significant independent variable and each time bringing the next most significant variable in to the regression. Its usefulness terminates when no more significant regressions can be extracted. A full description of regression procedures can be found in Draper and Smith.4 With this procedure it is possible to carry out a complete statistical analysis in only the simplest instances, e.g., when the calibrations are linear and the interference is a linear function of the concentration of the interferent. In other instances more sophisticated methods, such as trend surface analysis, may afford a complete treatment. However, multiple linear regression is always a useful first approach. If all the variance in a dependent variable, i.e., the analyte response, can be accounted for by regression with the appropriate concentra- tion, plus a contribution from “pure error,” and in addition no other element makes a significant contribution to the total variance, then the determination of that analyte can be deemed free from interference.If there is a significant amount of variance that can be accounted for by another element or elements, as judged by a t-test on the appropriate regression coefficient, then a simple interference such as a spectral line overlap may be indicated. If a significant amount of the total variance other than “pure error” cannot be accounted for by the regressions and the primary calibration is linear, then a more complex type of interference may be occurring and the elements involved may be indicated by their significantly high regression coefficients. The process can therefore be regarded as a rapid test. Detailed follow-up studies may be required to elucidate the response surface if a complex type of interference is suspected. Results In this work the experimental procedure and statistical analysis described above were carried out on two independent occasions, attention being given on each occasion to any element other than the analyte that produced a.t-value outside the range k l . 9 6 and also contributed a proportion of the total variance greater than 0.5%. In the first test the only possible interference revealed was arsenic on antimony, causing a small apparent enhance- ment. However, subsequent detailed study with arsenic and antimony alone revealed antimony to be an impurity in the arsenic solution and no actual interference. In the second study the response for tellurium was very noisy and a small regression on the concentration of arsenic was apparent, with t = -3.05, representing 2.1% of the variance.Again, subsequent studies showed that this apparent interference could not be reproduced with pure solutions of arsenic and tellurium, and the original effect could be attributed to random errors. It was therefore concluded that, within the concentration ranges studied and underJdy, 1978 Sb, Bi, Se AND Te BY PLASMA EMISSION SPECTROMETRY. PART II 707 the particular conditions employed, there were no mutual interference effects between the hydride-f orming elements. This conclusion was independently confirmed by the recovery experiments described below, in which random concentration mixtures were again employed. Interference from Other Ions The findings of other workers2J and our own exploratory work on single-element deter- mination using atomic-absorption spectrophotometry indicated that considerable interference could be expected at the reduction stage from a number of elements likely to be present in significant amounts in geochemical materials.We adopted an experimental approach by which it could be rapidly established whether a number of major and trace constituents caused interference. This initial approach was followed by detailed studies where inter- ferences were found. Preliminary Studies For each potential interfering element under consideration four solutions were prepared and analysed for the five hydride-forming elements. The solutions contained: (i), the reaction matrix only, i.e., 5 M hydrochloric acid; (ii), the potential interferent at an appro- priate level; (%), the hydride-forming elements at a level of 1 pg ml-l; (iv), as (iii) plus the potential interferent at the same level as in (ii).Any significant differences that were detected between (iii) and (iv) indicated interference, requiring further elucidation. A positive difference between (ii) and (i), though not found at significant levels in this study, would indicate the presence of the analyte element as an impurity in the interferent. The results obtained are shown in Table I, expressed as percentages of the interference- free response. Differences of less than about 4% are probably not significant, arising through random fluctuations, and those of less than 10% are usually tolerable at the low concentra- tions of analyte under consideration. The ions sodium(I), potassium(I), magnesium(II), TABLE I EFFECT OF VARIOUS IONS ON THE RECOVERY OF ARSENIC, ANTIMONY, BISMUTH, SELENIUM AND TELLURIUM AT A CONCENTRATION OF 1.0 pg ml-l Results are given as the percentage deviation from the interference-free response.In terferen t Na(1) . . K(1) . . Ca(I1) . . Fe(II1) . . Fe(I1) . . Mn(I1) . . Al(II1) . . Ti(1V) . . Cr(V1) . . Cr(II1) . . Mg(I1) . . V(V) . . Mo(V1) . . Co(I1) . f Ni(I1) . . Cu(I1) . . %n(II) . . Cd(I1) . . Hg(I1) . . Pb(I1) . . . . . . . . . . .. .. . . .. .. .. . . .. .. . . .. .. .. .. . . . - Deviation, % Concentration/ , A \ pg ml-l 10 000 10 000 10 000 10 000 10 000 1000 10 000 10 000 10 000 1000 1000 100 1000 100 1000 1 000 1000 100 1000 100 1000 100 1000 100 1 000 100 1000 100 1000 100 As 0 0 0 +2 -2 0 0 1- 1 + l + 1 0 0 - 50 - 5 - 3 0 + I 0 -2 0 0 0 -6 0 +2 + 1 0 0 -2 0 Sb 0 0 f 1 -2 -5 0 +6 4- 2 + 7 - 1 - 15 0 - 34 -2 -3 0 + 1 0 - 11 0 -4 0 -7 0 - 7 +2 -t 15 + 10 - 16 0 Bi -2 0 + 1 0 -8 0 1- 10 0 0 - 1 0 0 -2 0 -1 0 -4 0 -6 - 1 - 92 - 77 - 3 0 -4 0 - 83 - 20 -5 0 Se +2 +1 0 +2 - 26 -6 - 81 0 0 - 3 -11 -1 -5 0 0 - 15 -4 0 - 12 0 - 99 - 77 -7 0 - 12 -2 0 0 - 53 -2 Te + 2 + l + 1 + 1 - 10 - 3 0 - 10 -6 -5 - 50 -1 - 60 -8 - 4 - 10 - 39 0 - 72 0 - 99 - 99 -61 - 3 - 99 - 3 - 70 - 15 - 99 - 12708 THOMPSON et d. : SIMULTANEOUS TRACE DETERMINATION OF AS AnaZyst, VOL.103 calcium( 11), manganese(I1) and aluminium(II1) had no significant effects at levels well above those usually encountered. Iron as iron(I1) or iron(II1) had little effect on arsenic, antimony and bismuth but was of importance for selenium and perhaps tellurium. Vanadium(V), chromium(VI), molybdenum(VI), cobalt(II), nickel(II), copper(II), zinc(II), cadmium(II), mercury(I1) and lead(I1) showed significant effects when present at a concentra- tion of 1000 pg ml-l, especially towards selenium and tellurium.However, at the lower concentration of 100 pg ml-l, which for these elements is unlikely to be exceeded in normal geochemical work, only copper(II), lead(I1) and mercury(I1) had significant effects. Copper seriously suppressed the response for bismuth, selenium and tellurium, leaving arsenic and antimony unaffected, while lead diminished the tellurium response. The presence of mercury even at this concentration is unlikely. Thus, only copper, iron and lead were considered in further detail, as described below.Copper The effect of copper(I1) at concentrations between 10 ng ml-l and 100 pg ml-l on the recovery of the five elements is shown in Fig. 1. Arsenic and antimony were unaffected in this range and well outside it. The response for bismuth was unaffected at levels up to about 2 pg ml-l, for selenium up to about 1 pg m1-1 and the tellurium response was affected by as little as 50 ng ml-l of copper. Clearly the presence of copper(I1) at normal levels cannot be tolerated in the determination of bismuth, selenium and tellurium, although the interference effect has been reported to diminish in the presence of iron.3 100 n -1 0.01 0.1 1 .I1 10 100 Copper conceniration/pg mI-' Fig. 1. Effect of concentration of copper(1J) on the recovery of the analytes.Iron concentration/pg mI- ' Fig. 2. Effect of concentration of iron on the recovery of seleniuiri and tellurium.July, 1978 Sb, Bi, Se AND Te BY PLASMA I.)MISSION SPECTROMETRY. PART 11 709 Iron The effect of iron(I1) and iron(II1) at concentrations between 100 pg ml-1 and 1.0% on the recovery of selenium and tellurium is shown in Fig. 2 . Iron(III), the form normally resulting from the chemical digestion of environmental materials, had little influence on tellurium, but selenium was markedly affected at high concentrations of iron. The effect of a concentration of 1 000 pg ml-l of iron(III), which is rarely exceeded in practice, was less than 10% for selenium. However, iron( 11) caused slight enhancement of tellurium, possibly by increasing the effectiveness of the final reduction with sodium tetrahydroborate( 111), and had a severe effect on selenium, apparently causing premature reduction to the element, which is not reduced by sodium tetrahydroborate(II1). Lead the five elements is shown in Fig.3. normally exceeded, the effect was acceptably small even on tellurium. The effect of lead(I1) at concentrations between 1.0 and 1000 pg ml-l on the recovery of At a concentration of lOOpgml-l, which is not ------.--.---1 10 100 1 000 Lead concentration/pg ml-’ Fig. 3. Effect of concentration of lead on the recovery of the 0 analytes. Masking Studies In view of the serious interference of copper in the production of the hydrides of bismuth, selenium and tellurium, and the less marked effect of iron on selenium, potential methods for masking these elements were investigated. Preliminary work with atomic-absorption spectrophotometry had shown that, as expected, the usual sequestering agents for iron and copper, such as EDTA, tartaric acid and 2,3-dimercaptopropan-l-o1, were ineffective at the high acidity of the reduction medium or were themselves reduced by the sodium tetra- hydroborate( 111).Other potential masking agents, such as thiocyanate, triethanolamine and cupferron, caused very low recoveries of some elements, particularly selenium and tellurium. Potassium iodide was studied in some detail as it precipitates copper as copper(1) iodide and reduces iron(II1) to iron(I1). In addition it has been used in liydride-generation methods to improve the production of arsine in reductions with zinc metal5 and as a pre-reducer for arsenic(V) and antimony(V) in reductions with sodium tetrahydroborate( 111) .3 9 6 9 7 The effects on the reduction of the elements in their lower oxidation states, investigated by adding various concentrations of potassium iodide to the 5 M hydrochloric acid medium, are shown in Fig. 4(a). While the responses for arsenic(II1) and antimony(II1) were marginally enhanced by small concentrations (0.1%) of the salt, bismuth was suppressed and selenium and tellurium were severely suppressed. Selenium is reduced to the element, which is not subsequently reducible, the response being almost completely suppressed by potassium iodide at concentrations as low at 10 pug ml-l. Tellurium forms a yellow complex, which is only partially susceptible to reduction.7 10 Analyst, voj. 103 The effect of potassium iodide at various concentrations on the higher oxidation states of arsenic and antimony is shown in Fig.4(b). The efficiency of reduction of arsenic(V) and antimony(V) in pure hydrochloric acid is lower than that of the elements in the +3 oxida- dation state, but in the presence of a small concentration of potassium iodide both oxida- tion states of arsenic give the same response. THOMPSON et al. : SIMULTANEOUS TRACE DETERMINATION OF As A.ntimony behaves similarly. 0 0.1 0.2 0.3 0.4 0.5 0 0.05 0.1 0.15 0.2 0.0 0.05 0.10 0.15 0.2 Potassium iodide concentration, % Potassium iodide concentration, 56 Fig. 4. (a) Relative efficiency of hydride evolution as a function of potassium iodide concentration.(p) Comparison of the reduction efficiencies of arsenic and antimony when present in the -1-3 state (open circles) and the + 5 state (closed circles). Separation from Copper on Lanthanum Hydroxide Bkdard and Kerbyson have reported the successful separation from a copper matrix, firstly of traces of bismuth* and subsequently of traces of arsenic, selenium, tellurium and tin.* After dissolution of the sample, lanthanum(II1) is added to the solution and precipi- tated by the addition of ammonia. The copper remains in solution but the elements sought are co-precipitated by lanthanum hydroxide. After a double precipitation the hydroxide is filtered off and redissolved and the elements are determined by hydride evolution and flame atomic-absorption spectrophotometry.This method has been modified to provide a rapid procedure that is applicable to geochemical .work. The double precipitation has been replaced by a single operation, as the copper concentrations involved are much smaller, and the filtration has been replaced by centrifugation. In addition, the final concentration of lanthanum used has been reduced from 0.25 to O.lyo. Procedure To 10ml of the solution under study in a polypropylene centrifuge tube was added 1 ml of 2.5% m/V lanthanum nitrate solution, which was admixed by stirring, followed by 10ml of 1 + 1 ammonia solution. The precipitate was separated by clmtrifuging and the supernatant liquid dis- carded. The precipitate was washed with 3 ml of 1 + 1 ammonia solution, centrifuged again and the liquid discarded.The precipitate was redissolved in 10 ml of 5 M hydro- chloric acid and this solution was used directly for hydride generation. Greater final dilutions could be used if the concentrations were expected to be above the linear calibration ranges, the evolution efficiency of the hydrides being unaffected by the concentration of lanthanum . In recovery experiments the following method was adopted. Recovery from Copper Solutions Recovery from copper solutions was studied by application of the above procedure to 20 solutions containing 100 pg ml-l of copper as well as all of the analyte elements at concentrations selected at random, independently for each element, from the levels 1.0, 2.0,J d y , 1978 Sb, Bi, Se AND Te BY PLASMA EMISSION SPECTROMETRY. PART II 711 5.0, 10.0, 20.0, 50.0 and 100.0 ng ml-l.The recoveries are shown in Fig. 5, which shows the analyte concentration found plotted against the concentration present in solution. Recoveries of arsenic, antimony, bismuth and selenium are excellent. Erratic results at the lower end of the scale are due to the proximity of the detection limit. Results for tellurium are promising but the determination of tellurium is generally less precise than that of the other analytes, the lack of precision arising in the determination stage. 100 c I - E cn C -. -0 0 u c II: 10 +. .- v1 h 1 100 c I - E cn U 0 \ 8 10 5 rn LI- .- i 1 10 100 1 Arsenic present/ng ml-’ ” 1 00 c I - E 10 100 Antimony presenting rnl-’ 10 100 1 Bismuth presenthg ml-’ 10 100 Selenium present/ng mi-’ Tellurium present/ng mI-’ Fig.5. Recovery of the analytes from solutions containing randomly selected concentrations plus 100 pg ml-l of copper. The lines represent 100~o recovery: ( a ) , arsenic; ( b ) , antimony; (c), bismuth; ( d ) , selenium; and ( e ) , tellurium. Recovery from Solutions of Other Ions Although the lanthanum hydroxide precipitation was originally designed to alleviate the effect of interference by copper, it was recognised that it might be useful with other ions. This possibility was studied by attempting to recover the analyte elements at a concentra-712 THOMPSON et al. : SIMULTANEOUS TRACE DETERMINATION OF As Analyst, VoZ. 103 tion of 1 .O pg ml-l from solutions containing LOO0 pg ml-l of the interfering metal ions (see Table I).The results obtained are shown in Table 11, which gives the percentage devi- ation from the interference-free response by the precipitation procedure and by the direct method. Interference in the determination of selenium and tellurium produced by 1000 pg ml-l of vanadium(V), cobalt (11), nickel( 11) and copper( 11) was reduced t o negligible pro- portions, the effect of zinc(I1) and cadmium(I1) was substantially reduced and that of lead(I1) somewhat alleviated. Interference from iron( 111) remained unchanged as the ion is collected with the precipitate of lanthanum hydroxide. TABLE I1 RECOVERY OF BISMUTH, SELENIUM AND TELLURIUM FROM SOLUTIONS CONTAINING THE ANALYTE AT 1 .O pg ml-l PLUS 1000 pg ml-l OF VARIOUS IONS, AFTER SEPARATION ON LANTHANUM HYDROXIDE The results are given as the percentage 'deviation from the interference-free (see Table I).response. The values in parentheses were obtained without separation Deviation, yo Interferent V(V) Cr(V1) Blo(V1) CO(I1) Ni(I1) Cu(I1) Zn(I1) Cd(I1) Hg(II) Pb(1I) In Part I it was shown .. .. .. .. .. .. .. .. .. .. .. .. . . .. . . .. * . .. Conclusion Se - 5 0 0 0 -2 -2 - 5 - 10 0 0 - 20 (-11) ( - 5 ) (-16) (-4) (-12) (-99) (-7) (- 13) ( - 53) that sensitive determinations of arsenic, antimonv, bismuth, selenium and tellurium can be achieved by reduction of the elements to the-ir hydrides, which are introduced into an ICP source for emission spectrometry. In Part I1 it has been demonstrated that mutual interference effects between the analyte elements are negligible, while the effects of other cations at specific concentrations are either negligible or can be reduced to insignificant levels by a simple and rapid precipitation technique.The scope of this technique, coupled with the good linear calibration range of the ICP method, leads to applications spanning wide concentration ranges. It can be used as a concentration pro- cedure when levels are expected to be very low, as in natural waters. At high levels, in addition to the wide calibration range available, further dilution with the sample reaction medium (5 M hydrochloric acid) is possible without any control of the concentration of lanthanum being necessary. In Part I11 applications in the field of applied geochemistry will be dealt with and com- parison with other methods will be made. One of us (B.P.) is indebted to the Government of Iran for financial support during this Computing was carried out on the CDC Facility at the Imperial College Computer work. Centre.WY 1 . 2. 3. 4. 5. 6. 7 . 8. 9. 1978 Sb, Bi, Se AND Te BY PLASMA EMISSION SPECTROMETRY. PART 11 713 References Larson, G. F., Fassel, V. A., Scott, R. H., and Kniseley, R. N., Analyt. Chem., 1975, 47, 238. Smith, A. E., Analyst, 1975, 100, 300. Fleming, H. D., and Ide, R. G., Analytica Chim. Acta, 1976, 83, 67. Draper, N. R., and Smith, H., “Applied Regression Analysis,” John Wiley, New York, 1966. Sandell, E. B., “Colorimetric Determination of Traces of Metals,” Third Edition, Interscience, New Slemer, D. D., Koteel, P., and Jarlwala, V., Analyt. Chem., 1976, 48, 836. Fiorino, J. A,, Jones, J. W., and Capar, S. G., Analyt. Chem., 1976, 48, 120. RBdard, M., and Kerbyson, J . D., Analyt. Chem., 1975, 47, 1441. BCdard, M., and Kerbyson, J . D., Can. J . Spectrosc., 1976, 21, 64. York, 1959. Received November 7th, 1977 Accepted Februavy 21st, 1977

ISSN:0003-2654    

DOI:10.1039/AN9780300705

出版商:RSC    

年代:1978

数据来源: RSC

摘要:

714 Analyst, July, 1978, Vol. 103, pp. 714-722 Contributions to Automated Trace Analysis. Part I I . * Rapid Method for the Automated Determination of Lead in Whole Blood by Electrothermal Atomic-absorption S pect ro p hotometry M. Stoeppler and K. Brandt and T. C. Rains Institute of Chemistry, Institute 4, Applied Physical Chtmistry, Nuclear Research Centre, P.O. Box 19 13, 0-61 70 Jiilich, West Germany iiational Bureau of Standards, Analytical Chemistry Diui.;ion, Washington, D.C., USA h simple, rapid method is described for the determination of lead in whole blood by means of electrothermal atomic-absorption spectrophotometry. Aliquots (25-200 pl) of fingerprick and venous samples were treated with 2 M nitric acid for deproteinisation and matrix modification. After centri- fuging, the supernatant was taken for the automated analysis for lead. The accuracy of the method was checked with independent methods and found to be satisfactory.Thus it was established that the accuracy obtainable of <30% lies almost within the confidence intervals ( p = 0.05) of precision. An evaluation of 282 pairs of randomly selected routine samples indicated an acceptable precision : the relative standard deviation for, for example, the normal level of 100 pg 1-1 is 8.4Oj,, and for an elevated level of 480 pg 1-1 is 3.47%. The shortened temperature programme makes up to 370 measure- ments per day possible and the computer coupling permits an immediate data evaluation during occupational and screening programmes and makes possible round-the-clock measurements. Keywords : Whole blood analysis ; lead determination ; automated trace analysis ; electrothermal atomic-absorption spectrclphotometry Recently conducted round-robin exercises on the determination of lead and cadmium in biological fluids indicated clearly that the necessary high level of reliability in this field is frequently not achieved.’ To improve this situ:ation and the technical possibilities for the analysis of numerous other materials of medical, toxicological and environmental significance critical studies in this field are urgently needed.Thus, a few years ago the German Federal Department of Research and Technology (BMFT) started some research projects on “Toxic Elements and Biocides and their Automated Determination.” In the programme of auto- mation we carried out a series of basic studies.The results of our investigation showed the excellent performance of automated sample-injection systems for graphite Within the same project several methods for the determination of trace elements in whole blood by electrothermal atomic-absorption spectrophotometry were studied comparatively and in some instances impro~ed.~ Promising pre-treatment techniques for blood axe deproteinisation and matrix modification. Methods of this type were advocated by Einarsson and Lindstedt,s and more recently by Baily and Kilroe-Smith7 and Andersen and Zac:hariasen.8 These methods involve the use of mixtures of trichloroacetic acid and perchloric acid or sulphuric acid and were reported partly in order to show the good results obtaina‘ble with graphite furnace techniques7 However, from our experience the direct application even of dilute perchloric acid and sulphuric acid leads to a considerable attack on the surface of the graphite tubes, causing fluctuating results and decreasing tube lifetimes.:3-5 Thus a more advantageous and simpler deproteinisation technique had to be developed.From some preliminary studies carried out at. the National Bureau of Standards the use of nitric acid seemed promising because of its excellent deproteinisation properties, a highly * For details of Part I of this series, see reference list, p. 722.STOEPPLER, BRAATDT AND RAINS 715 improved charring process and the minimisation of spectral matrix interference^.^ Numerous measurements carried out in the meantime showed that it is advantageous to use dilute nitric acid for deproteinisation as well as for matrix modification.This method is described and discussed in this paper. Experimental Apparatus The atomic-absorption devices used were Perkin-Elmer Rf 400 instruments equipped with HGA-74 and HGA-76 furnace units, the Perkin-Elmer Autosampler AS-1, the UP-1 printer and a PE 56 recorder. One M 400 device was, in the course of the study, combined with a PDP 11/10 system for data evaluation. Some series of measurements were carried out with a Jarrell-Ash 811 spectrophotometer - HGA-74 - AS-1 system, coupled with a Kipp and Zonen BD9 recorder. All laboratory ware used was carefully cleaned by rinsing with dilute acids ( e g . , 2 M nitric acid) or chelating agents prior to use and was checked for lead content.Reagents reagent grade (p.a.) or sometimes Merck Suprapur quality. Argon of 99.999% purity was used throughout. Other chemicals were of analytical- Procedure Sampling and sample preparatio?a Samples were taken by means of venipuncture with sterile, practically lead-free syringes, equipped with stainless-steel tips, to be used once only (DIN 13098, Braun, Melsungen, West Germany). Immediately after sampling the blood was thoroughly mixed in pre-cleaned polystyrene tubes with 4 mg of sodium citrate per millilitre of whole blood, but other anticoagulants such as ethylenediaminetetraacetic acid ammonium salt or sodium salt and heparin can also be used. Into the polystyrene tubes were pipetted 150 or 300 pl of 2 M nitric acid, 100 or 200 pl of fresh or stored blood and 150 or 3OOpl of distilled water.The tubes and contents were then vigorously shaken by means of an electrical (e.g., vortex) mixer for 30 s. After centrifuging the supernatant was transferred to a 1-ml AS-1 polyolefin sample cup (cleaned together with the other laboratory ware) and measured. It is also possible, in principle, to obtain reliable results from fingerprick sampling if the skin is thoroughly cleaned before cutting with a solution of a complexing agent, e.g., ethylenediaminetetraacetic acid sodium sa1t.l') However, there are some problems with people who smoke or are exposed in their occupation, but it is the method of choice for children and non-smokers if venipuncture is difficult or a frequent control is required.Into the AS-1 sample cup 30 or 60 pl of 1 M nitric acid were pipetted, and 20 or 40 p1 of blood and 30 or 60 pl of distilled lead-free water were added. The contents of the cup were vigorously mixed and then centrifuged. The supernatant was transferred to another cup, as described above. The volume of supernatant obtained allows at least three 10-p1 injec- tions. Venous sampks. The samples were stored in a refrigerator at about 4-6 "C. The resulting solution was 0.75 M in nitrate. Fingerprick samples. Measure?nent, temperuture programme and background minimisation as the temperature programme of the HGA-74 furnace unit are shown in Table I. of matrix niodification on the non-specific background is demonstrated in Fig. 1. The operating conditions for the M 400 and the Jarrell-Ash 811 spectrophotometers as well The effect Calibration and evaluation During the development of the method a within-run calibration was carried out either by means of standard addition or against matrix-matched calibration graphs from lead-spiked blood samples with relatively low lead contents.As there was no statistically significant716 STOEPPLER, BRANDT AND RAINS: CONTRIBUTIONS Analyst, VoZ. 103 TABLE I: SPECTROPHOTOMETER OPERATING CONDITIONS AND HGA-74 PROGRAMME Instrument Conditions Perkin-Elmer 400 . . PE Lead Intensitron lamp, 283.3 nm, 10 mA Deuterium compensation, slit 7 Absorbance, peak-height mode Hamamatsu lamp, 283.3 nm, 15 mA, channel A, slit 3 Compensation with 282.0-nm non-absorbing line, channel B, slit 4 Absorbance, evaluation A-B mode Continuous registration of A-B and B Programme Jarrell-Ash 811 .. A r > Time/s Step ternperaturel'c 30 Charring 600 2 Cleaning 2 600 Approximate HGA-74 . . . . 30-60* Drying 100 5 Atomisation? 2 200 * Depending on tube properties and injected volume (10 or 20 pl). f Use of gas stop. difference between the average values and the precision data obtained by the two evaluation modes, for routine purposes the much simpler and faster method involving the preparation of a calibration graph was chosen. A typical matrix-matched calibration graph, dynamic range at least 400 digits (divisions), is shown in Fig. 2. Thus the evaluation mainly consists in the daily preparation of at least three calibration graphs in such a manner that every fifth to tenth sample is a spiked or standard solution.Normally, and if no drift effects occur, during the course of a working day (about 9-10 h operation) it is possible to evaluate the data with the respective calibration graph for each set of sample^.^ Data evaluation is carried out with the PDP 11 system, which also includes an auto-zero adjustment. Fig. 1. Comparison between direct determination of lead in whole blood and determination after deproteinisation - matrix modification: (a), direct, 1 + 4 diluted, 0.25 M hydrochloric acid (programme duration 135 s) ; (b) , nitric acid method, 1 + 3 diluted, 0.75 M nitrate (programme d.uration 67 s). 1, Non-specific absorption. 2,Total absorption - non- specific absorption. Spectrophotometer, Jarrell-Ash 810, double channel with double-channel recorder.(Because of the different start positions of the two channels, the maxima of the first peaks are recorded a t different positions, but in fact they appear simultaneously in channels A and B of the instrument.) 400 300 VI r 0) .- .- 200 1 00 I I I 0 50 100 150 2 Concentration of lead/pg I-' Fig. 2. Matrix-matched calibra- tion graph. Blood supernatant 0 . 7 5 ~ in nitrate, dilution 1 + 3, injection volume 1 0 ~ 1 . Each point repre- sents the mean of the results of three firings from the same solution. One digit m 0.001 absorbance unit. 0JULY, 1978 TO AUTOMATED TRACE ANALYSIS. PART I1 Remarks 717 If levels of lead are well above 600 pg 1-1, the dilution factor has to be increased. The centrifugation should be carried out by using a centrifuge with high speed in order to attain a good phase separation. If good separation is not achieved very small particles will be found in the supernatant, but from our experience these particles do not influence the precision of the spectrophotometric measurements.Occasionally, owing to a biochemical degradation of stored or otherwise altered sample it has been observed that signal suppression or enhancement occurs. In this instance individual evaluation by means of standard additions and/or checking measurements with other atomic-absorption spectrophotometric and electrochemical methods can be carried out. The limits of determination for 10- and 20-pl injections are about 20 and 10 pg of lead per litre of whole blood, respectively, with readings around six digits (see Fig.2). Results and Discussion Accuracy spectrophotometric methods and instruments and by voltammetry. The accuracy of the method was checked by the application of different atomic-absorption Atomic-absorption spectrophotometry Sets of relatively fresh samples with different concentrations of lead, e.g., blood diluted 1 + 19 with 0.5 M hydrochloric acid5 and blood diluted and mixed with perchloric acid - acetone,ll were analysed by direct atomic-absorption spectrophotometry, using standard additions, in comparison with the nitric acid method. The slopes of the correlation graphs and the resulting mean values showed only minor differences and therefore no statistically significant systematic deviations (see Table 11). TABLE I1 CORRELATION STUDIES FOR THE MEASUREMENT OF LEAD I N WHOLE BLOOD (50-900 pg 1-l) Number of samples = 33.Correlation Instrument Method Equation coefficient Evaluation JA 811 - HGA-74 Direct, HC1 y = 1.03% + 2.6 0.975 Standard additions PE 400 - HGA-74 HNO, Standard additions PE 400 - HGA-74 Direct, HCI 3’ = 1 . 0 1 ~ + 0.5 0.989 Standard additions PE 400 - HGA-74 HNO, Calibration graph From our results, despite some obvious disadvantages of direct spectrophotometric deter- minations in whole blood (e.g., a prolonged temperature programme, relatively poor precision and lifetime-decreasing carbon residues on inner tube walls), these techniques nevertheless showed an average relative standard deviation of about 15% at the lOOp~gl-~ level and no severe deviations compared with voltammetry.Thus we cannot confirm recent state- ments about virtually useless direct determinations of lead in whole blood.12 Voltammetry Fifty-six samples from occupationally exposed and control groups, ranging in concentration from 35 to 900 pg 1-1 of lead, were analysed by diff erential-pulse anodic-stripping voltammetry using a hanging mercury drop electrode (DPASV/HMDE) as well as by the nitric acid method. The DPASV/HMDE method, which was developed in this Institute,13 starts with718 STOEPPLER, BRANDT AND RAINS : CONTRIBUTIONS Analyst, Vol. 103 low-temperature ashing and is very precise even in the lower concentration range. The results show an acceptable agreement and demonstrate that within random variations of both methods no systematic deviation, i.e., deviation above a low uncertainty, seems to occur (see Fig.3). The comparison of independent methods is, if (certified standard reference materials are not available, the most efficient way to check the accuracy of a particular analytical pro- cedure.14 1 / I . r 2 - ! L ~ I L 100 200 300 400 500 600 700 800 901 Concentration of lead by atomic-absorption spectrophotoimetry/pg I -' Comparison of the determination of lead in whole blood by the nitric acid method, with evaluation by a matrix-adapted calibration graph, and by voltam- metry (DPASV/HMDE), with evaluation by the standard-additions technique. Xumber of samples = 56. Equation: y = 1.03% - 2.17. Correlation coefficient. 0.974 7. Mean concentration f standard deviation (pgl-1): by DPASV, 271.23 & 247.41 (error of the mean, 33.06) ; by atomic-absorption spectrophotometry , 264.52 &- 233.06 (error of the mean, 31.14).Fig. 3. Precision and Estimations of Error The mean precision, computed from 12 randomly selected series of samples (lead concentra- tion 20-100 pg 1-l) analysed up to October 1977 and based on 1008 duplicate measurements from the same solution, was 2.1 digits (range 1.2-3.7). If a level of 350 pg 1-1 of lead is considered as the highest permissible level, at least for adults,l5 at this concentration the average reproducibility is about 1 % (range 0.6--1.8%). Nevertheless, as could be shown from numerous measurements with other elements and solutions, influences of the whole instrumental arrangement, the graphite-tube properties and the matrix constituents lead to marked fluctuation in the signal with time.3s4*16 In addition, the sample pre-treatment introduces another source of random errors.Hence, the relative standard deviation for the whole analytical procedure is considerably poorer. In order to establish the relative standard deviation, a total of 564 sample aliquots, analysed from May 1976 to October 1977, were chosen, representing independent pre- treatments and analyses of 282 pairs of sub-samples. The selected series covered the concentra- tion range from 36 to about 700 pg 1-1 of lead in whole blood. In order to demonstrate the influence of concentration seven groups of samples with increasing levels of lead were arranged in order (see Table 111). Further, a series of 12 samples from unexposed persons was analysed six times during aJuly, 1978 TO AUTOMATED TRACE AXALI'SIS.PART I1 719 fortnight. The evaluation was carried out by different modes. The data obtained, together with a short description of the evaluation techniques used, are given in Table IV and provide a realistic picture of this type of analysis. The estimation of the probable total error of a blood-lead determination must be based on precision data for the over-all procedure as well as on comparison measurements. From Fig. 3 it can be seen that sometimes large differences between results obtained by means of atomic-absorption spectrophotometry and DPASV/HMDE appear. Thus, at present, the occasional occurrence of hitherto unrecognised systematic deviations cannot be absolutely excluded, and we recommend the addition of an uncertainty value of about 10% to the confidence intervals.Hence, for p = 0.05 and for concentrations of about 100 and 350 pg l-l, the total error may be <30 or <ZOyo, respectively. A possible source of error, the use of an aqueous calibration solution that is 0.75 M in nitric acid and that has some- times been found to cause errors if introduced in different electrothermal ~ystems,~ was avoided by using the matrix-matched calibration described above. Owing to the non-linearity of calibration graphs and the necessary dilution in instances of lead concentrations 2600 p g l-l, the best attainable over-all precision is at present 3.5% (see Table 111) in graphite furnace atomic-absorption spectrophotometry, because the uncertainties discussed above give a probable total error of about 15% at the 95% confi- dence level.If the given precision is too low for a certain investigation, repeated determination at different times (ie., different runs) improves the data substantially. For example, from five carefully performed independent determinations with a relative standard deviation of 10% the relative standard deviation of the mean decreases to 4.5%. TABLE 111 DETERMINATION OF LEAD I N PAIRS OF SAMPLES OF WHOLE BLOOD USING NITRIC ACID AND AUTOMATED ATOMIC-ABSORPTION SPECTROPHOTOMETRY Calculation of standard deviations, relative standard deviations and confidence intervals from pairs of sub-samples arranged in seven groups in order of increasing lead concentration. Number of pairs 24 45 45 40 40 40 48 Mean PLg I-' rangelpg 1-l concentration*/ Concentration 52.4 36-60 73.2 61-80 100.8 81-120 136.6 121-160 174.5 16 1-200 232.7 201-300 479.0 301-700 Standard deviationt I 4.88 6.55 8.50 7.68 8.72 10.38 16.63 CLg 1-' Relative standard deviation, % 9.3 8.9 8.4 5.6 5.0 4.47 3.47 Confidence intervals, yo r p = 0.1 p = 0.06 16.9 19.2 14.9 17.8 14.1 16.8 9.4 11.2 8.4 10.0 7.5 9.0 5.8 7.0 * Mean values from 2-3 firings, evaluated against a calibration graph (at least two matrix additions).t Computed for each group, from the difference in the values obtained in two parallel determinations a t different timcs, according to the following equation: Standard deviation = x ( b ) 2 J T where Ax is the difference between the two single values and n is the number of chosen pairs as given above.Performance and Reliability For environmental screening programmes as well as for the surveillance of exposed workers relatively high measuring capacities are desirable in respect of speed and costs. From former and recently performed investigations with the HGA-74 it is evident that, under routine conditions, even with nitric acid concentrations >0.7 M, average lifetimes of the graphite tubes of >500 firings are to be expected.394 This lifetime, together with the relatively high technical quality of modern atomic-absorption spectrophotometric instru- mentation, enables the analyst to perform round-the-clock measurements. Thus the number of samples that could be analysed for lead per day was determined.720 STOEPPLER, BRANDT AND RAINS : CONTRIBUTIONS Analyst, ‘VOZ.103 TABLE TV DETERMINATION OF LEAD I N TWELVE SAMPLES OF WHOLE BLOOD FROhi UNEXPOSED PERSONS USING NITRIC ACID AND AUTOMATED ATOMIC-ABSORPTION SPECTROPHOTOMETRY Concentration of lead measured on six different dayslpg 1-1 Mean Standard h I -------> concentration/ deviation] Sample Xo. 1 2 3 4 5 6 Pg 1-1 PQ 1-l (a)* Three firings from each sub-sample solu.tion, evaluation by individual stafidard additions, i.e., a total of 1 66 85 98 105 80 87 86.8 13.7 2 98 74 84 85 80 73 82.3 9.2 3 137 135 160 165 170 140 151.8 16.6 4 80 93 74 81 74 75 79.5 7.3 5 109 112 122 113 119 121 116.0 5.4 6 105 105 110 111 102 110 107.2 3.7 7 137 119 127 134 135 131 130.5 6.6 8 78 84 81 84 84 100 85.2 7.6 9 129 101 109 96 90 86 101.7 13.8 10 109 100 99 95 88 85 96.0 8.7 11 96 102 112 107 106 135 109.6 13.5 12 48 52 65 52 56 48 53.5 6.4 Mean 99.3 96.8 103.4 102.3 98.7 99.3 100.0 9.38 15 firings per sub-sample- (b) * Three firings from each sub-sample solution, evaluatioxi by a n average matrix-matched calibration graph 1 76 81 90 91 82 82 83.7 5.7 2 93 73 79 83 77 74 79.8 7.4 3 134 142 162 162 149 158 151.2 11.5 4 81 81 81 86 76 79 80.7 3.3 6 117 120 122 133 114 117 120.5 6.7 6 104 107 113 117 101 103 107.5 6.3 7 125 128 131 141 131 126 130.3 5.8 8 78 84 81 89 84 82 83.0 3.7 9 111 107 110 102 102 101 105.5 4.4 10 86 97 99 93 88 88 91.8 5.3 11 87 100 108 106 106 100 101.2 7.7 12 44 57 64 55 59 54 55.5 6.7 Mean 94.7 98.1 103.3 104.8 97.4 97 99.2 6.21 per day, i.e., a total of < 4 firings per sub-sample- (c)* One firiazg from each sub-sample solution, evaluatioaz by one single matrix-matched addition (mean of 1 78 81 88 84 85 82 83.0 3.5 2 74 71 76 73 101 74 78.2 11.3 3 146 155 155 149 151 147 150.5 3.9 4 73 79 77 76 85 82 78.7 4.3 5 107 117 121 117 134 125 120.2 9.0 6 95 103 108 104 113 109 105.3 6.2 7 126 123 127 122 135 133 127.7 5.3 8 79 80 77 75 86 86 80.5 4.6 9 100 101 108 104 121 109 107.2 7.7 10 86 86 95 91 92 98 91.3 4.8 11 103 99 104 100 98 102 101.0 2.4 12 57 52 60 58 47 54 54.7 3.7 Mean 93.7 95.7 99.7 96.1 104 100 98.2 5.64 about 10 additions pev working day), i.e., a total of Q 1.2 firings per sub-sample- * 2 f standard deviation of mean values 1-6: (a), 100.0 f 2.44; (b), 99.2 f 3.94; (c), 98.2 f 3.77 pg 1-1.Table IV shows that owing to the reliability of the automated injection for the processing of large sample numbers, only a single firing and a simple evaluation seem to be sufficient in a first approach and if only statistical averages are desired.Of course, in this instance a sufficient number of controls and all values above a distinct level (e.g., 350 &- 20%, range 280420 pg 1-1 of lead in whole blood) have to be re-analysed and additionally checked by DPASV/HMDE if threshold values have to be established. The total time required for the AS-1 operation, including a fixed cooling time of 30 s, ranges from 137 to 97 s, depending on the tube properties and volume pipetted. ThisJdy, 1978 TO AUTOMATED TRACE ANALYSIS. PART I1 721 permits a rate of 28-37 firings per hour, i.e., 280-370 per working day (10 h).If 30% are estimated as calibration, control and repeat measurements, the daily workload reaches at least 230 samples if one firing per sub-sample is carried out. If from every sample two independent sub-samples are analysed, about 10-15% of the controls are avoided and about 140 samples could be analysed per day. For round-the-clock operation, including tube replacement, an average of about 540 measurements on about 320 samples can be made per 24 h. > I BCD -binary r M400 4- Auto zero converter Read Start CAMAC FHTG-1 controller Fig. 4. Schematic diagram of the instrumental arrangement and the computer coupling. CAMAC = Computer Applications t o hleasurement and Control (see also ref. 16). I, Converts binary coded decimal data t o binary data; 11, data output from the PDP 11; 111, LAM (“look a t me”) control function input into the PDP 11; IV, interface between CAMAC modules and the PDP 11.Fig. 4 shows scliernatically the instrumental arrangement and the computer coupling using CAMAC modules. A detailed description of the applied system and the software is in preparation.17 Thus, with computer control and evaluation a relatively small staff is able to conduct, without great difficulty, a large number of measurements per day. I t is also obvious that in the course of extended screening programmes even a highly sophisticated and therefore relatively expensive device reaches a satisfactory cost/benefit ratio. The principles of CAMAC have been discussed elsewhere.16 Conclusion The method described proved, within the limits discussed, very reliable in routine applica- tionsls and round-robin exercises. The sample pre-treatment is fast and easy, the temperature programme short, and matrix influences as well as the attack of tube walls are minimised.The approach described is in principle also useful for the determination of cadmium in whole blood, as will be discussed e1se~here.l~ We thank Professor Dr. H. W. Niirnberg, Director of the Institute of Applied Physical Chemistry of the Nuclear Research Centre, Jiilich, and Dr. P. D. LaFleur, Head of the Analytical Chemistry Division of the National Bureau of Standards, Washington, D.C., for their continuous interest, and Ing. grad. M. Kampel, Ing. grad. H. Rutzel, Miss Ing. grad. H. Schinke and Ing.grad. Van Dy Nguyen for valuable technical assistance. The support of these investigations by the German Federal Department of Research and Technology (BMFT), Bonn, in the framework of the joint research project “Automation of Methods for the Determination of Toxic Substances in the Environment,” Contracts Mt 405 g/AUT 03, is gratefully acknowledged.722 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. STOEPI’LER, BRANDT AND RAINS References Lauwerys, K., Buchet, J. P., Roels, H., Berlin, A., and Smeets, J.. Cliiz. Chem., 1975, 21, 551. Dahl, K., and Stoeppler, M., Bey. Kenzforschungsankce Julich, Jul-1254, November 1975. Stoeppler, M., Kampel, M., and Welz, B., 2. Analyt. Chem., 1976, 282, 369. Stoeppler, M., and Kampel, M., Ber. Kernforschungsanlage Julich, Jiil-1360, November 1976. Brandt, K., and Stoeppler, M., Bey. Kernforschungsanlage Jiilich, in preparation. Einarsson, O., and Lindstedt, G., Scand. J . Clin. Lab. Invest., 1969, 23, 367. Baily, P., and Kilroe-Smith, T. A., Aizalytica Chinz. Acta, 1975, 77, 29. Andersen, I., and Zachariasen, H., “Proceedings of the International Symposium on Clinical Chemistry and Chemical Toxiciology of Metals, Monte Carlo, March 2-5, 1977,” to be published. Ottaway, J. M., Proc. Analyt. Div. Chem. Soc., 1976, 13, 185. Knutti, R., Balsiger, Ch., and Schlatter, Ch., Mitt. Geb. Lebensmittelunters. u. Hyg., 1977, 68, 78. Schumacher, E., and Umland, F., Z . Analyt. Chew, 1974, 270, 285. Stahel, 0 . F., Wernli, R., and Leemann, W., Beckman AAS Application, Bulletin M-614 Klin, Valenta, P., Rutzel, H., Nurnberg, 11. W., and Stoeppler, M., Z. Analyt. Chem., 1977, 285, 25. Nurnberg, H. W., Stoeppler, M., and Valenta, P., Thalassia Jugosl., 1975, 11, 85. Zielhuis, R. L., Iizt. Arch. Arbeitsmed.. 1974, 37, 103. Esone Committee, Eur 4100e, revised version 1972, Joint Nuclear Research Centre, Ispra, Italy. Maeckelburg, D., Frerichs, M., and Stoeppler, M., BCY. Kernforschztngsanlage Julich, in preparation. Stoeppler, M., and Brandt, K., “Proceedings of a Symposium on Clinical Biochemistry, Marburg, Stoeppler, M., Brandt, K., Rutzel, H., and Valenta, P., Sci. Total Enviv., in preparation. August, 1976. July 1977,” Springer-Verlag, Berlin, in the press. NOTE-Reference 3 is to Part 1 of this series. Received August 24th, 1977 Accepted December 9tA, 1977

ISSN:0003-2654    

DOI:10.1039/AN9780300714

出版商:RSC    

年代:1978

数据来源: RSC

摘要:

Analyst, July, 1978, Vol. 103, p p . 723-727 723 Corn pa rat ive Laboratory Eva I uat i on of Some Reported Methods for the Determination of DDT and BHC Insecticides in Human Blood R. C. Gupta, A. B. Karnik, S. K. Nigam and S. K. Kashyap Natioizal Institatte of Occupational Health, A hmedabad-380 016, Gujarat, I%dia The rclative efficiency of extraction of DDT, HHC and related insecticides from spiked serum samples by different methods is reported. The insecticide was extracted from the serum by using hexane or hexane - acetone (9 + 1) and then determined in the extract by gas - liquid chromatography using an electron-capture detector. The highest recoveries were obtained when the serum was treated with formic acid before extraction, lower values were obtained when sulphuric acid was used in place of formic acid and the lowest values were obtained when the untreated serum was extracted.Recoveries in general were higher when the hexane - acetone mixture was used in place of hexane. The order of extractability was y-BHC > a-BHC > S-BHC > Pp’-l>UE > pp’-DDT > o$’-DDT. Ii-~yiu~orCIs 1 Insecticide determination ; BHC determination ; DDT determina- tion ; lmiwin blood analysis ; gas - liquid clwcmatography There is wide interest in the amounts of chlorinated hydrocarbon pesticides in body fluids and, because of its availability and the reported relati~nshipl-~ between exposure and the concentrations found in it, blood has served as the tissue of choice in these studies. Dale et a1.l observed that insecticides in the blood are present in substantial proportions in the serum, and for most toxicological evaluations in which the extent of exposure to insecticides is to be evaluated, determination of the level of these materials in plasma or serum would be satisfactory .Subsequently, Laws et aL3 observed a I iigh positive relationship between the blood serum levels and fat storage levels of DDT. Jensen et aL4 reported the analysis of plasma for determination of DDT and PCB residues using chromatographic methods and Palmer and Kolmodin-Hedman5 have also utilised plasma in epidemiological studies for multiple analysis of chlorinated pesticides. In this work blood serum has been used for toxicant addition and subsequent extraction and analysis. Although analysis of serum for residues of Chlorinated insecticides may not be entirely satisfactory, it is nevertheless widely used, especially when epidemiological and clinical surveillance studies are carried out together with residue analysis for monitoring occupational exposure to pesticides.Dale et al.1 used liexane for the extraction of chlorinated insecticides from blood but found that the recoveries were not quantitative, probably owing to binding of the chlorinated insecticides to serum proteins. Gunther et aLG postulated the binding of DDT and similar compounds to protein molecules. Pre-treatment of the serum with equal volumes of formic acid was found by Dale et aL7 to release the DDT and DDT metabolites, permitting the extraction of these materials with hexane, 95% recoveries being obtained.Palmer and Kolmodin-Hedman5 reported a method suitable for multiple analysis of chlorinated pesticides in human plasma in epidemiological studies. In this method, the plasma was extracted with hexane after pre-treatment with formic acid. An internal standard was added and the identity of the peaks confirmed by mass spectrometry. Stretz and StahrS evaluated a method in which sulphuric acid was used to liberate the pesticides from the blood matrix; they were then extracted with a mixture of acetone and hexane (1 + 9). These workers recommended further evaluation of the sulphuric acid method by comparison with other methods. Saha et aL9 used a hexane - acetone mixture for extraction of pesticides from soil and Henderson et al.1° optimised the concentration of acetone in this solvent mixture for the extraction of pesticides from a sulphuric acid - blood mixture. The object of our work was to determine the relative efficiency of extraction of the different724 GUPTA et al.: COMPARATIVE LABORATORY EVALUATION OF Analyst, “02. 103 isomers of DDT and BHC by some reported methods or modifications of them. Blood serum samples were spiked with known concentrations of DDT and BHC isomers and allowed to stand for 48 h to allow an opportunity for in vitro binding to occur. This period was chosen as Dale et aZ.l have shown that the in vitro interaction or binding of the added insecticides with the constituents of serum was not increased b y storage of the spiked samples for 5-7 d. The samples were then extracted by following various extraction procedures and the insecticide residue determined by gas - liquid chromatography using an electron-capture detector.Experimental Apparatus Gas chromatografih. Toshniwal gas - liquid chromatograph (Toshniwal Instruments, Bombay) with electron-capture detector, tritium source and 10-mV full-scale chart recorder. Column. A 6 f t x 5 mm 0.d. coiled column, packed with 5% of AC-QF1 on Chromosorb W treated with hexamethyldisilazane in dichloromethane. Operating conditions : column temperature, 190 “C; injection port, 230 “C; detector, 190 “C; nitrogen carrier gas flow- rate, 40mlmin-l. The instrument was found to give a linear response to the various insecticides over the following ranges: a-BHC 0.2-20 ng; y-BHC 0.2-20 ng; P-BHC 2.0- 100 ng; pp’-DDT 5.0-200 ng; @’-DDE 5.0-200 ng; and of-DDT 2.0-100 ng.EvaPorative concentrator. Reagents Hexaize. Hexane - acetone, (9 + 1). Both components of high-purity grade. SuZphuric acid, 60% m/m. Analytical-reagent grade. Formic acid. Analytical-reagen t grade. Potassium carbonate solution, 5% m/V in distilled water. 2,2,4-Trimethyl$entane. Pesticide quality . Pesticide stock solutions. High-purity grade, distilled from all-glass apparatus. High-purity reference-standard materials obtained through the courtesy of the World Health Organization were used. Stock solutions in 2,2,4-trimethyl- pentane, each containing 100 pg ml-l of insecticides, were prepared. Appropriate amounts of stock solutions were mixed and diluted with 2,2,4-trimethylpentane to obtain mixed standard solutions (50 ml) containing a-BHC, y-BHC, /3-BHC, Pp’-DDT, @’-DDE and op’-DDT in the proportions of 1 + 1 + 5 + 10 + 10 + 5 .Working standard solutions. Standardisation of the Gas - Liquid Chromatograph solution and measurement of the response. The apparatus was standardised by injection of idiquots (1-10 pl) of the working standard Preparation of Test Solutions Test solutions having the compositions given in Table I were prepared by pipetting the required amounts of the working standard solution into test-tubes and evaporating the solvent by passing a current of dry air over the liquid. Serum was then added to the dry residue and the tube and contents were shaken to dissolve the solid. These solutions were stored for 48 h in a refrigerator before use. This procedure was adopted in order to avoid TABLE I COMPOSITION OF TEST SOLUTIONS OF INSECTICIDES IN SERUM 7 Insecticide a U-BHC .. . . 0.25 y-BHC ,. . . 0.25 8-BHC . . . . 1.25 +P’-DDT .. .. 2.50 pp’-DDE . . . . 2.50 op’-DDT .. .. 1.25 Spiking level/pg ml-1 ‘b C 0.50 0.75 0.50 0.75 2.50 3.75 5.00 7.50 5.00 7.50 2.50 3.75 --- 7 d 1.00 1.00 5.00 10.00 10.00 5.00Jdy, 1978 SOME METHODS FOR THE DETERMINATION OF DDT AND BHC IN HUMAN BLOOD 725 the possibility of errors caused by the presence of organic solvents in the spiked serum samples. The levels chosen for spiking were higher than those normally found in serum so that the amounts present in the serum before spiking could be ignored in the calculations. This procedure follows that suggested by Dale et aZ.l Extraction Methods Spiked samples were extracted by each of the following methods for each level of spiking.Method 1 To a test-tube containing 2 ml of spiked serum sample, 6 ml of hexane were added. The tube was stoppered and the contents were mixed thoroughly by vigorous shaking for about 2 min. The two phases were then separated by centrifuging at 2 0oO rev min-1 for 5 min and the upper hexane phase was then carefully transferred into a 25-ml evaporative concentrator tube. The extraction was repeated with two additional 6-ml portions of hexane and the three extracts were combined in the concentra- tor tube. The tube was placed in a water-bath at 3540 "C and the solvent was removed with the aid of a gentle stream of clean, dry air. The residue was dissolved in 0.25-1 ml of hexane, depending upon the level of spiking, and an aliquot of 10 11.1 was injected into the gas - liquid chromatograph.This method was a simple hexane extraction. Method 1A As for method 1 except that hexane - acetone (9 + 1) was used instead of hexane. Method 2 This method involved pre-treatment of the spiked serum with sulphuric acid before extrac- tion with hexane. To a test-tube containing 2 ml of sample, 1.5 ml of 60% m/m sulphuric acid were added and the tube was shaken for 30 s. A second 1.5-ml volume of acid was added and mixed for 30 s and finally 2 ml of the acid were added and mixed. The contents of the tube were then cooled and 5 ml of hexane were added. The tube and contents were shaken for 2 rnin and then centrifuged a t 2 000 rev min-1 for 10 min.The hexane layer was removed with a disposable capillary pipette and placed in a concentrator tube. The extraction was repeated twice, using 3 ml of hexane each time, the extracts were combined and the determination was completed as described for method 1. i%fethod 2A As for method 2 except that hexane - acetone (9 + 1) was used instead of hexane. Method 3 This method involved pre-treatment of the spiked serum with formic acid before extraction with hexane. To a test-tube containing 2 ml of sample, 2 ml of 97% formic acid were added and the tube and contents were shaken vigorously for 1 min. Five millilitres of hexane were added, the tube was again shaken for 1 min and the phases were separated by centri- fuging the tube a t 2000 revmin-l for 5min. The hexane phase was carefully removed and placed in a second centrifuge tube.A second 5-ml volume of hexane was pipetted into the tube containing the serum and the extraction repeated. The combined extract was washed once with 1 ml of 5% potassium carbonate solution for 1 min by vigorous shaking and the aqueous phase was discarded. The determination was completed as described for method 1. Method 3A As for method 3 except that hexane - acetone (9 + 1) was used instead of hexane.726 GUPTA et al. : COMPARATIVE LABORATORY EVALUATION OF Analyst, VoZ. 103 Recovery The recoveries of the insecticides were calculated by comparing the chromatographic peak heights (BHC isomets) or areas (DDT isomers and DDE) with those obtained by injection of aliquots of the working standard scllution. The percentage recoveries were calculated and an analysis of variance was carried outll in order to find the critical difference at the 5% level for the recovery of individual insecticides by different methods and for the recovery by individual methods of different insecticides.Results and Discussion When extracting the spiked samples by methods 2 and 2A problems caused by the forrna- tion of emulsions were encountered and these were solved by increasing the volume of the hexane or hexane - acetone mixture and lengthening the centrifuging time. Some of the TABLE 11: RECOVERIES OF INSECTICIDES BY DIFFERENT METHODS OF EXTRACTION FROM BLOOD SERUM Values given are means of three determinations. Recovery, SDikinrr I __- A \ Insecticide N-BHC . . 7-BHC .. p-BHC . . pp'- D DT p p '-D D E o@'-DDT I " . level Method : 1 .. a 43.7 b 40.1 C 37.7 d 41.7 Mean: 40.8 . . a 65.0 b 62.4 C 59.3 d 63.8 Mean: 62.6 . . a 47.3 b 41.2 C 38.7 d 42.5 Mean: 42.4 . . a 50.1 b 37.9 C 33.4 d 43.2 Mean: 41.1 .. a 50.1 b 42.7 C 40.8 d 47. I Mean: 45.1 .. a 34.3 b 30.1 C 27.3 d 31.5 Mean: 30.8 C.D.* 6.09 Method Ih 51.3 44.2 38.7 46.2 45.1 75.8 71.2 70.6 63.7 TO. 3 51.2 45.8 43.0 49.6 47.4 53.3 45.1 42.3 51.2 47.9 59.2 45.0 38.7 51.3 48.5 32.5 28.3 26.8 31.4 29.7 7.93 Method Method 2 2A 59.8 71.6 57.6 70.8 45.3 63.7 55.1 67.4 54.4 68.3 C.I).* = 9.19 78.5 73.2 71.8 71.2 61.4 67.3 68.3 69.2 70.0 70.2 C.D.* = 6.51 54.7 58.9 55.3 55.2 50.1 49.2 53.7 51.7 53.4 53.7 C.D.* = 4.99 41.8 53.1 39.4 49.3 27.6 37.8 32.9 48.2 35.4 47.1 c.I).* = 10.02 58.1 60.2 52.5 58.3 40.3 51.2 47.3 51.5 49.5 55.3 C.D.* = 8.61 38.3 48.7 35.3 36.3 29.7 34.6 32.3 35.2 33.9 38.7 C.D.* = 7.28 8.67 7.24 Method 3 98.7 94.1 73.7 85.8 88.0 98.7 93.7 88.4 89.4 92.5 78.6 75.2 73.2 68.7 73.9 73.2 59.8 48.3 62.7 61.0 71.3 61.8 64.7 67.8 66.4 69.3 65.2 58.2 51.7 61.1 11.06 Method 3A 93.7 91.9 81.7 88.9 89.0 95.3 90.3 94.6 89.7 92.4 77.1 74.8 74.6 73.2 74.9 68.3 59.3 64.1 67.3 64.7 68.7 67.2 59.5 64.3 64.9 67.3 62.3 58.2 63.0 62.7 5.52 * C.D.= critical difference a t the 5% level.Juh, 1978 SOME METHODS FOR THE DETERMINATION OF DDT AND BHC IN HUMAN BLOOD 727 collaborators in the studies of Stretz and Stahr* also reported problems with the formation of emulsions. The separation of the phases in our studies was most distinct with the methods involving formic acid (methods 3 and 3A).Palmer and Kolmodin-Hedman5 had also reported a distinct separation of the phases when using formic acid but not when extracting with hexane in the absence of formic acid. We, however, observed a satisfactory separation of the phases even when hexane or hexane - acetone alone were employed (methods 1 and 1A). Percentage recoveries and mean percentage recoveries are reported for individual insecticides and for individual methods. The critical difference at the 5% level is given for each insecticide by all methods and for all insecticides by each method. Firstly, treatment of the serum with formic acid before extraction (methods 3 and 3A) gave the best recoveries, which were significantly higher than those obtained by the other methods.Dale et a1.’ observed an increase in the recovery of total insecticide as DDT from 13.0 to 39.8% when using the formic acid- hexane method instead of the hexane extraction method. The reason for the action of formic acid is at present unknown but it could possibly be the libera- tion of the chlorinated hydrocarbons from binding sites on the plasma proteins by denatura- tion of the proteins. Secondly, treatment of the serum with sulphuric acid prior to extraction (methods 3 and 2A) gave recoveries significantly better than those obtained by extracting the serum with solvent only (methods 1 and 1A). Thirdly, recoveries in general were higher when using the hexane - acetone mixture (lA, 2A, 3A) than when using hexane alone (1, 2, 3), but were statistically significantly so only for /3-BHC by method 1A against method 1 and for a-BHC and $9’-DDT by method 2A against method 2.Fourthly, yBHC showed the highest extractability by all of the methods, suggesting a poor iut vitro binding of this isomer, and 09’-DDT gave the lowest recoveries, suggesting the maximum binding. Also, a-BHC and fl-BHC showed higher extractability than did $$’-DDE and $$’-DDT. Thus, the order of extractability was y-BHC > a-BHC > P-BHC > pp’- Fifthly, it was observed that the highest percentage recoveries were obtained for all of the methods and for all of the chemicals for the lowest spiking levels. The extracted residue from the sample spiked at level (a) was dissolved in 0.25 ml of hexane [0.50, 0.75 and 1.0 ml for levels (b), (c) and (d), respectively] from which 10 p1 were injected on to the gas chromato- graph.This concentration factor might have played an important part in this observation. The results of the investigations are presented in Table 11. The following conclusions can be drawn from the tabulated results. DDE > Pp’-DDT > o$‘-DDT. The authors thank Mr. P. K. Kulkarni and Mr. V. G. Patel for their assistance with the statistical analysis and Miss Hansa R. Patel for her assistance in the preparation of the manuscript. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. References Dale, W, E., Curley, A., and Cueto, C., Jr., Life Sci., 1966, 5, 47. Dale, W. E., Curley, A., and Hayes, W. J., Ind. Med. Surg., 1967, 36, 275. Laws, E. R., Curley, A., and Biros, I. J., Archs Ewviv. Hlth, 1967, 15, 766. Jensen, S., Renberg, L., and Vaz, R., F.A.O. Fish. Tech. Pap., No. 137, 1075, p. 231. Palmer, L., and Kolmodin-Hedman, B., J . Chromat., 1972, 74, 21. Gunther, F. A., Blinn, R. C., Carmen, G. E., and Metcalf, R. C., Archs Biochenz. Biophys., 1954, Dale, W. E., Miles, J. W., and Gaines, T. B., J . Ass. Off. Analyt. Chem., 1970, 53, 1287. Stretz, P. E., and Stahr, H. M., J . Ass. 08. Analyt. Chem.. 1973, 56, 1173. Saha, J. G., Bhavaraju, B., and Lee, Y. W., J . Agric. Fd Chem., 1969, 17, 874. Henderson, S. J., DeBoer, J. G., and Stahr, H. M., Analyt. Chenz., 1971, 43, 445. Burington, R. S., and May, D. C . , “Handbook of Probability and Statistics with Tables,” Second Received September 26th. 1977 Accepted January 4th, 1978 50, 504. Edition, McGraw-Hill. New York, 1970, p. 276.

ISSN:0003-2654    

DOI:10.1039/AN9780300723

出版商:RSC    

年代:1978

数据来源: RSC

摘要:

728 Analyst, July, 1978, Vol. 103, Pjb. 728-733 Quantitative Analysis of Active Substances in Liquid Pharmaceutical Preparations Using H igh-performance Liquid Chromatography Monir Amin and Peter W. Schneider Schering A .G., Department Galenik, Department Allgemeine Physikochemie, Mullerstrasse 170-178, 1000 Berlin 65, Germany High-performance liquid Chromatography has been applied to pharmaceutical Preparations containing the active ingredients in aqueous, alcoholic or oily solution. A comparison is made of the high-performance liquid chromato- graphic method with gas and thin-layer chromatography. Keywords : High-performance liquid chromatogyaphiy ; liqwid pharmaceutical preparations ; quantitative analysis of drugs ; quantitative analysis of steroids A central problem in the qualitative and quantitative analysis of active substances in pharmaceutical preparations is the avoidance of interferences imposed by the galenical excipients.This problem is amplified with decreasing concentrations of the active substances and increasing complexity of the excipients. Thus, elaborate separation and concentration methods must frequently precede the actual analytical determinations, Such techniques are time consuming and tend to decrease the accuracy of the analytical results. High- performance liquid chromatography (HPLC) has recently emerged as a method that allows in many instances a direct analysis or, at least, one that minimises the efforts necessary for sample preparation. The separation efficiency and accuracy of HPLC in the determina- tion of a wide range of compounds is now well estab1ished.l Examples of the analysis of pure compounds of therapeutic interest include steroid^,^-^ sulphonamides,5 vitamins,6 barbiturate^,^ alkaloids,8,9 ana1gesicsl0,l1 and antibiotics.12 The ability of HPLC systems to handle mixtures containing components of greatly varying polarities and in wide ranges of concentration is expected to simplify sample preparation considerably for such mixtures.In fact, pharmaceutical preparations have already been successfully analysed by using HPLC. Examples include tablets,l3~l4 cough mixtures,15 creams or ointrnentsl6J7 and oil- based injection preparations.l* In this paper we report on further selected examples of HPLC analyses of galenical preparations, the preparations varying in complexity.Examples include aqueous, alcoholic and oily solutions containing only a single active substance. The ease of sample preparation, accuracy, sensitivity and time required for analysis are einphasised and comparison is made with the corresponding parameters in gas and thin-layer chromatography. Experimental Materials Aqueous solution, 0.12% m/m of l-(isopropylamino)-3-(2-methylindol-4- yloxy)propan-2-01 as the sulphate salt (compound 1). Ethanolic solution, 0.1 % m/m of 6a,9-difluoro-l1 p-hydroxy-l6a-methyl- 21-valeryloxypregna-1,4-diene-3,20-dione (diflucortolone valerate) (compound 2). Oily solution of compound 2,0.174 mlm, with more hydrophilic excipients, e.g., esters. Oily solution, 0.51% m/m of 17a-ethinyloestra-l,3,5( lO)-triene-3,17/3-diol (ethinyloestradiol) (compound 3), with more hydrophobic excipients, e.g., natural oils. For preparation A 2-methylindole, for preparations B and C 9-chloro-6a-fluoro-21-heptanoyloxy-ll/3-hydroxy- pregna-l,4-diene-3,20-dione (clocortolone caproate) and for preparation D 17P-methoxy-4- androsten-3-one.Preparation A . Preparation B. Preparation C. Preparation D. Reference compounds (internal standards) for ywantitative analysis.AMIN AND SCHNEIDER 729 Apparatus and Conditions High-performance liquid chromatography A Du Pont, Model 841, liquid chromatograph with an ultraviolet (254 nm) detector and a column of Du Pont Zipax ETH Permaphase, 100 cm x 2.1 mm i.d., particle size 37 pm, was used at 25 "C. Mobile phase: preparation A, water - methanol (1 + 4 V / V ) ; preparations B and C, water - methanol (45 + 55 VlV).A Hupe-Busch - Hewlett-Packard, Model 1000, liquid chromatograph, with an ultraviolet (254 nm) detector and a column of Merckosorb Si60, 100 cm x 4 mm id., particle size 20 pm, operated at a temperature of 25 "C, was used. Mobile phase: chloro- form - cyclohexane (3 + 2 V / V ) . Preparations A-C. PreParation D. Gas chromatography A Hewlett-Packard, Model 7620A, gas chromatograph with a flame- ionisation detector was used. Columns and temperature programmes : preparation A, 0.5% OV-17 on Chromosorb G, 1.8 m (2 rnin at 220 "C, 6 "C min-l to 260 "C, 4 rnin at 260 "C and 4 "C min-l to 280 "C) ; preparations B and C, 1% SE-30 on Chromosorb G, 1.8 m (2 min at 230 "C and 4 "C min-l to 300 "C). Preflaration D.An F & M, Model 400, gas chromatograph with a flame-ionisation detector was used. Column and temperature programme (after HPLC separation) : 1% XE-60 on Chromosorb G, 1.5 m (215 "C). Preparations A-C. Thin-Zayer chromatography Silica gel 60 F,,, plates (Merck), 0.25-mm layer, 20 x 20 cm were used. Preparation A . Benzene - chloroform (1 + 1 V / V ) . Development time : approximately 90 rnin plus drying for 5 rnin in a stream of warm air followed by development for 90 rnin in dichloromethane - methanol - concentrated ammonia solution (50 + 50 + 1 V/V). Development time: approxi- mately 60 min. Develop- ment time: approximately 150 min plus drying for 5 rnin in a stream of warm air followed by development for 60 rnin in cyclohexane - ethyl acetate (1 + 1 V/V).Preparation B. Cyclohexane - ethyl acetate (1 + 1 V / V ) . Preparation C. Carbon tetrachloride - hexane - benzene (70 + 25 + 5 V/V). Procedure The HPLC and gas-chromatographic analyses were performed according to the method of internal standardisation, while the thin-layer chromatographic analyses were carried out by using the method of external standardisation. The aliquots were weighed to &0.5y0. For preparation A, approxi- mately 1 mg of 2-methylindole was added to 1 ml of A and the volume was adjusted with water to 2 ml. A 2 - 4 volume (containing 1.2 pg of compound 1) was injected on to the column. For preparations B and C, approximately 5 mg of clocortolone caproate were added to 1 ml of B or C and the volume was adjusted to 10 ml with methanol.A 2-pl volume (containing 0.2 pg of compound 2) was injected on to the column. Finally, for preparation D, approximately 5 mg of 17~-methoxy-4-androsten-3-one were added to 1 ml of D and the volume was adjusted to 2 ml with chloroform. A 0.1-ml volume (containing 0.26mg of compound 3) was injected on to the column. The fraction indicated in Fig. 5 was collected, evaporated under reduced pressure to approximately 1 ml and then analysed by means of gas chromatography. For HPLC analyses, the samples were prepared as follows. Results and Discussion The determination of the active substances in pharmaceutical preparations A and B, i.e., the aqueous solution of compound 1 and the alcoholic solution of compound 2, respectively, by use of either thin-layer chromatography, gas chromatography or HPLC are straight- forward and without particular difficulties.Nevertheless, the HPLC method offers advan- tages over the gas-chromatographic method and particularly over the thin-layer chromato- graphic method with respect to analysis time and sensitivity (Table I).730 AMIN AND SCHNEIDER: QUANTITATIVE ANALYSIS OF ACTIVE Analyst, l’d. 103 TABLE I COMPARISON OF METHODS : HPLC, GAS CHROMATOGRAPHY AND THIN-LAYER CHROMATOGRAPHY Gas chromato- Thin-la yer Preparation graPhY chromatography* HPLC Sample application . . .. A Approx. 5 s Approx. 5 mint Approx. 5 s B Approx. 5 s Approx. 5 rnin Approx. 5 s C Not applicable Approx. 5 rnin Approx. 5 s D Not applicable Not applicable Approx. 10 minf Chromatographic separation A Approx. 20 rnin Approx.120 rnin Approx. 18 rnin B Approx. 15 rnin Approx. 120 rnin Approx. 15 rnin C Not applicable Approx. 180 rnin Approx. 15 rnin D Not applicable Not applicable Approx. 12 mins Limit of detection . . .. A Approx. 0.5 pg Approx. 1 pg Approx. 0.1 pg B Approx. 0.5 pg Approx. 1 pg Approx. 0.1 pg C Not applicable Approx. 1 pg Approx. 0.1 pg D Not applicable Not applicable Not determined Time requirement : * It should be remembered, however, that several tracks are developed simultaneously on one thin-layer This is a great advantage of thin-layer chromatography as compared with the other methods, plate. particularly in routine analysis. t Including the time required for evaporation of the solvent spot. $ Including HPLC injection, evaporation of collected fraction, gas-chromatographic injection.5 HPLC approximately 6 min, gas chromatography approximately 6 min. Typical chromatograms obtained by HPLC for. preparations A and B are shown in Figs. 1 and 2. The results of calibration runs and quantitative analyses, obtained by use of the method of internal standardisation, are summarised in Table 11. The corresponding quantitative data obtained by use of gas and thin-layer chromatography are given in Table 111. Inject 1-( Isopropylamino)-3- (2-methylindol-4-yloxy)- p ropa n -2-0 I h e n t 2-Methylindole L I I I I I I 0 5 10 15 2c Time/min Fig. 1. Typical liquid chromatogram of preparation A. For conditions see under Experimental. Inject Clocortolone caproate Diflucortolone valerate Solvent 1 0 5 10 15 20 Time/m in Fig. 2. Typical liquid chromatogram of preparation B.For conditions see under Experimental.Jzc,!~, 1978 SUBSTANCES I N LIQUID PH+4RMACEUTICAL PREPARATIONS BY HPLC 73 1 TABLE I1 QUANTITATIVE ANALYTICAL RESULTS FROM HPLC .4ctive substance present in Prepara- preparation, tion % W V N A 0.12 8 B 0.10 8 C 0.10 8 D 0.51 S Calibration internal standard* A $ 7 Coefficient of vari- f x s ation, yo N 1.43 0.045 3.2 8 0.97 0.418 1.9 8 0.97 0.018 1 .Y 8 1.25 0.20 1.6 8 Analysis of preparation* A 7 Coefficient % s ation, % Content, of vari- 0.122 0.003 2.5 0.100 0.003 3.2 0.101 0.002 2.7 0.505 0.029 5.9 * A' = number of determinations; fx = calibration factor; s = standard deviation of the individual values. In comparison with the analyses of preparations A and B, considerable problems are encountered in the analysis of preparations C and D, i.e., the oily solutions of compounds 2 and 3, respectively.Direct analyses by gas chromatography are not possible because of a two-fold interference by the components of the excipients : firstly, these excipients contain components the chromatographic peaks of which coincide with the peaks of the active substances ; and secondly, owing to their low volatility, these compounds tend to accumulate on the gas-chromatographic column and thus irreversibly alter the separation characteristics of the gas-chromatographic system. Similarly, a direct analysis of preparation C by thin- layer chromatography is not possible because of the severe tailing exhibited by the hydro- philic excipients, with some components not migrating at all.Thus, a large portion of the possible migration path on the thin-layer plate is rendered void. Although this problem can be avoided by multiple development techniques, excessive time is required for this type of thin-layer chromatographic analysis (Table I). Finally, for preparation D, no multiple development system could be found. However, by using a reversed-phase HPLC system with a chemically bonded ether as stationary phase, reliable analytical results could be obtained upon direct injection of the diluted oily solution. There was no accumulation of hydrophilic excipients on the HPLC column and excellent short- and long-term performance of the separation system was noted (Fig. 3). TABLE I11 QUANTITATIVE ANALYTICAL RESULTS FROM GAS AND THIN-LAYER CHROMATOGRAPHY Analysis of preparation* Active substance present in Prepara- preparation, A 0.12 B 0.10 C 0.10 D 0.51 tion Yo m/V Standard deviation of the Content, % individual values & -7 N GC TLC GC TLC 8 0.126 0.123 0.005 0.005 8 0.102 0.100 0.003 0.005 8 - 0.105 - 0.004 - - - - - Coefficient of vari- ation, % & GC TLC 3.1 4.06 2.9 5.0 - 3.8 - - * iV = number of determinations; GC = gas chromatography; TLC = thin-layer chromatography.The components of the hydrophilic excipients are eluted from the column quantitatively within the first few millilitres following injection, as is shown by a thin-layer chromatogram (Fig. 4). As these components do not absorb ultraviolet light at 254 nm there is no inter- ference with the detection of the ultraviolet light absorbing active substances in the pharma- ceutical preparation.Analytical results for preparation C are summarised in Table I1 and a comparison of the HPLC method with gas and thin-layer chromatography is made in Tables I and 111. Attempts to use a similar HPLC system for the direct analysis of preparation D were not successful. This was mainly a result of the strong ultraviolet absorption of the aromatic components of the excipient and the low concentration of compound3. By using a reversed-732 Diflucortolone valerate Solvent + AMIN AND SCHNEIDER: QUANTITATWE ANALYSIS OF ACTIVE Anai'yst, VoZ. 103 li I -- Fraction b Fraction c Fractions d and e 0 5 110 15 20 Time/m in Fig. 3. Typical liquid chroniatogram of preparation C. For conditions see under Experimental.The fractions indicated refer to the corresponding thin-layer chromato- graphic traces (see Fig. 4.). phase HPLC system, e.g., a chemically bonded hydrocarbon, as stationary phase and water - methanol as the mobile phase, compound 3 is eluted as a sharp peak prior to the main portion of the excipients. However, a minor component of the excipients interferes considerably by producing a peak of about 100-fold intensity close to the peak given by compound 3. In addition, more than 1 h is required in order to elute all of the excipients from the column. In a normal-phase HPLC system, using silica gel and chloroform - cyclohexane, the excipients of preparation D are eluted first, followed by compound 3. Because of the broad peak and the low concentration no direct quantitative determination is possible.However, the good separation of the active substance from the excipients provided the basis for subsequent analysis by gas chromatography. For this purpose, an internal standard was selected such that its retention time on the HPLC column was close to that of compound 3. A preparative column, which could accommodate 0.1 ml of the diluted preparation D was used. The fraction containing the internal standard and compound 3 was collected and analysed by means of gas chromatography (Fig. 5). The results of typical analyses are summarised in Table 11. Conclusions By taking advantage of the high separating power of HPLC and its capability of handling compounds of widely varying properties, sample preparation procedures for complex mixtures can be reduced to a minimum.The method compares favourably with gas and thin-layer chromatography with respect to the time required for analysis, accuracy and the sensitivity. Of course, the choice of the method must depend on the analytical problem. In instances where conditions for a direct analysis do not exist, because of interferences bya b c d e f 9 h Fig. 4. Thin-laycr chromatogram of fractions b-e (see Fig. 3). Trace a corresponds t o the HPLC eluting agent background. Traces f-h are of reference solutions con- taining individual components of the more hydrophilic excipients. [To face p. 732JZCt!j!, 1978 SUBSTANCES IN LIQUID PHARMACEUTICAL PREPARATIONS BY HPLC 733 Fraction collected I I I I I 1 I 0 1 2 3 4 5 6 Time/m in Fig.5. Typical liquid chromato- gram of preparation D. The fraction collected contains both the active substance and the internal standard for subsequent analysis by GC. The tall peak following the injection mark corresponds to the excipients. For conditions see under Experimental. the pharmaceutical excipients, HPLC can be used efficiently in a preparative manner in order to prepare the sample for subsequent analysis by some other quantitative method. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. References Michaelis, A. F., Cornish, D. W., and Vivilecchia, R., J . Pharm. Sci., 1973, 62, 1399. Edwards, D. P., O’Conner, J. L., Bransome, E. D., Jr., and Braselton, W. E., Jr., J . Biol. Chem., Jefferson, W. E., Jr., and Chang, F. C., Analyt. Lett., 1976, 9, 429. Okuyama, S., Uemura, D., and Hirata, Y., Chem. Lett., 1976, 679. Kram, F. C., J . Pharm. Sci., 1972, 61, 254. Tomkins, F. D., and Tscherne, J. R., Analyt. Chesn., 1974, 46, 1602. Roos, R. W., J . Pharm. Sci., 1972, 61, 1979. Wildanger, W., Dt. LebensmittRdsch., 1976, 72, 160. Verpoorte, R., and Svendsen, A. B., J . Chromat., 1976, 120, 203. Rosenbaum, D., Analyt. Chem., 1074, 46, 2226. Ascione, P. P., and Chrekian, G. P., J . Pharm. Sci., 1075, 64, 1029. Butterfield, A. G., Hughes, D. W., Pround, N. J., and Wilson, W. C., Antimicrob. A g . Chemother., Huettemann, E. I<., and Shroff, A. P., J . Chromat. Sci., 1975, 13, 357. Gilpin, R. I<., Koppi, J. A, and Janicki, C. A., J . Chromat., 1975, 107, 115. Honigberg, I. L., Steward, J. T., and Schmit, A. P., .I. Pharm. Sci., 1974, 63, 766. Mollica, J. A., and Strusz, K. F., J . Pharm. Sci., 1972, 61, 445. Gordon, G., and Wood, P. R., Analyst. 1976, 101, 876. Evans, F. J., J . Chromat., 1971, 60, 179. 1976, 251, 1632, 1974, 4, 11. Received December 2nd. 1977 Accepted January 20th. 1978

ISSN:0003-2654    

DOI:10.1039/AN9780300728

出版商:RSC    

年代:1978

数据来源: RSC

摘要:

734 Analyst, July, 19’18, Vol. 103,pp. 7’33-744 Problems of Accuracy and Precision in the Determination of Trace Elements in Water as Shown by Recent International Atomic Energy Agency Intercomparison Tests* R. Dybczynski, A. Tugsavul and 0. Susctiny International Atomic Energy Agency, Laboratory Seibersdovj, A -101 1 Vienna, Austria Results of a recent International Atomic Energy Agency intercom parison test on the determination of 16 trace elements in simulated fresh water samples are reported and discussed from the standpoint of accuracy and precision. Despite the considerable spread of results obtained froni various laboratories, the over-all mean (after rejection of outliers) in most instances approximates closely to the true value. The importance o€ using sufficiently sensitive tests for the detection of outlying results is emphasised and the effectiveness of concurrent use of several criteria for that purpose is demon- strated.The relative frequency of employment of various analytical techniques is discussed and an attempt to compare their accuracy is presented. Keywords : Water analysis ; trace-element determination ; accuracy ; precision ; outlier tests Water is one of the most common materials that are analysed for the purpose of monitoring environmental pollution. While being an apparently easy matrix, it can in fact pose serious problems even for an experienced analyst. Trace elements and radioiiuclides may disappear from aqueous solutions by adsorption on container or by volatili~ation,~~7~~ biological activity and bacterial action,8,1° etc.Pre-concentration techniques such as evaporationltll and even freeze-dryingll can also lead to considerable losses of some elements. On the other hand, if necessary precautions are not taken some trace elements can be added to the sample from the vessels1s7Jo and from the laboratory air.l?l2 In 1975-76 the International Atomic Energy Agency (IAEA) organised an intercomparison exercise for the determination of trace amounts of several elements in simulated fresh water samples, the results of which are presented in this paper. As in the previous exercises of that kind,l3 the primary objective was to give the laboratories in various countries an opportunity to check their analytical performance. At the same time it was hoped that the availability of a large number of experimental results would enable us to draw some general conclusions on the following questions : What accuracy of determination of the “true content” of a trace element can be expected from the results of a collaborative study, including those obtained by laboratories in various countries recruited on a voluntary basis ? (2) What is the effect of variations in the statistical treatment of data, including the procedures for detection and rejection of outliers, on the accuracy of the final results? (3) Is there any direct relationship between.the precision of the results reported by a. given laboratory and their accuracy? These problems are of general interest for an analytical chemist and were considered to deserve a detailed study.(I) Preparation of the Samples The spike solutions of 16 trace elements (arsenic, barium, cadmium, cobalt, chromium, copper, iron, mercury, manganese, molybdenum, nickel, lead, selenium, uranium, vanadium and zinc) and the main ionic constituents of fresh water (K+, Na+, Mg2+, Ca2+, A13+, C1-, SO,”, PO,3- and SiO,4-) were supplied to participants in three sealed glass ampoules (4 ml * Presented a t the Fourth SAC Conference, Birmingham, July 17th to 22nd, 1977.DYBCZYNSKI, TUGSAVUL AND SUSCHNY 735 each). The solutions were prepared from analytical-reagent grade reagents of the highest purity obtainable commercially and doubly distilled water. The solutions were approxi- mately 1 N in nitric acid and the amounts of trace and main constituents were such as to give concentrations after dilution ranging from 1 to 32 pg 1-1 for trace elements and 1 to 35 mg 1-1 for the main ionic constituents.The participants were instructed to dilute the contents of the three ampoules with doubly distilled water acidified with 50 ml of concentra- ted nitric acid to a final volume of 5 1. It was requested that several (up to six) replicate determinations should be made on aliquots of this final solution and the results reported as net values (after correcting for the blank). The over-all inaccuracy with which the “true concentrations” of trace elements were known was estimated to be smaller than &3%. Results and Discussion General Information and Handling of Data Thirty-five laboratories from 19 countries participated in this intercomparison exercise, returning altogether 357 laboratory means (1 162 individual results).The data were punched on t o cards and processed by a computer program written especially for this purpose. A typical output for a given trace element is shown in Table I. To secure anonymity, the laboratories were coded by numbers, which appear in column 1 Methods used by particular laboratories were also coded and are shown in The numbers appearing before the point refer to separation and/or in Table I. column 2 in the table. TABLE I DETERMINATION OF CADMIUM IN WATER SAMPLE W-3 Significance level used for outlier tests, 0.01 ; true value, 4.0 pg 1-’. Laboratory code 1 3 4 5 6 8 9 13A 14 15B 16 17 18 21 22 23 24 26 27 28 29 30 31 32A 33A 33B 34 Method code* 53.1 4.2 9.2 4.2 .2 .2 5.1 1.2 9.6 0.2 45.2 0.2 0.2 .2 0.2 0.1 2.5 0.2 0.8 1.2 0.1 9.1 9.2 0.2 0.2 0.2 0.2 Estimated laboratory error, %t .10 .03 .10 25.05 5.15 .05 .01 .07 8..10 .10 20. .05 .16 .14 Number of determin- ations 4 2 1 2 3 5 3 2 4 6 3 6 1 6 2 2 2 6 1 1 5 2 6 5 4 2 1 Laboratory mean/ 4.362 3.700 13 .OOO: 4.450 9.333 1.900 11.433; 5.050 4.500 5.250 3.907 2.700 4.000 4.150 3.000 7.375 3.950 4.700 7.000 4.000 5.140 17.000$. 6.000 3.020 5.000 5.700 3.000 Pg 1-1 Deviation from true value/ 0.362 - 0.300 9.000 0.450 5.333 -2.100 7.433 1.050 0.500 1.250 tLg 1-l - 0.093 - 1.300 0.000 0.150 - 1.000 3.375 - 0.050 0.700 3.000 0.000 1.140 13.000 2.000 -0.980 1 .ooo 1.700 - 1.000 Laboratory standard deviation Relative: Absolute yo 0.080 1.8 - - - - - - 0.057 0.6 0.168 8.3 3.300 28.8 1.161 25.6 5.0 0.266 0.100 2.5 0.109 4.0 0.064 1.3 - - - - - - - - - - 0.141 3.0 - - - - 1.863 36.2 0.894 14.9 0.083 2.7 0.711 14.2 - - - - - - R 1.548 - - - 16.120 6.460 0.393 0.148 2.849 0.161 7.209 1.663 - - - - - 3.007 - - 0.297 1.358 6.688 0.481 - - - * The numbers before the point refer to separation and/or pre-concentration methods, those after the point to the methods used for the quantitative determination.t The number after the point gives the laboratory’s own estimate of the maximum systematic error and the number before the point gives the random error due to counting statistics for radiometric methods. : These results were rejected when calculating the over-all mean and other parameters.736 DYBCZY~SKI et al. : PROBLEMS OF ACCURACY AND PRECISION Analyst, Vol.103 pre-concentration methods, those after the point to the methods used for the quantitative determination. In column 3 the laboratory’s own estimate of the maximum systematic error (number after the point) and the random error due to counting statistics (number before the point, for radiometric methods only) are indicated. For each laboratory the program computed the laboratory mean and its deviation from the true value. The laboratory standard deviation (both absolute and relative) was also calculated in instances in which at least three results for a given element were supplied. In addition the R value, defined below, was calculated whenever possible. where A is the deviation of the laboratory mean from the true value, s is the standard deviation of an individual determination and is given by f i (x, - 5)Z s = 12, 4 - It - 1 n is the number of determinations performed and to.ol is the Student’s factor for a significance level of 0.01.The R value relates the actually observed difference between the laboratory mean and the true value to the maximum reasonable deviation from the true value that can occur in practice (in 99 out of 100 instances), provided that random errors only are operative. Thus, a value of R <1 can normally be expected. If R >1 the existence of a systematic error may be postulated. Firstly, when calculating R, an allowance might have been made for maximum uncertainty of the true value. Some laboratories producing highly precise results with the mean close to the true value (e.g., laboratory No.21 in Table I) are unlikely to have made a systematic error. However, as will be shown later (Table 111), such instances are seldom encountered and they do not change the conclusions that can be drawn as to the status of analytical work in general. Secondly, for some laboratories reporting mean values far from the true value, the value of R may be lower than unity if the laboratory’s standard deviation is very high. While in this instance there is no internal discrepancy between the inaccuracy actually revealed and that expected on the basis of the laboratory’s precision, it raises the question as to whether the method used was the correct choice for the purpose. A survey of the analytical methods used by participants has shown a strong preference for atomic-absorption spectrophotometry and neutron-activation analysis, which contri- buted jointly over 80% of the results for most of the elements. X-ray fluorescence occupied the third place followed by colorimetry plus spectrophotometry and emission spectroscopy.Other methods were seldom used except in special instances (e.g., fluorimetry for uranium). For some of the elements (arsenic, barium, selenium and uranium) neutron-activation analysis contributed over 40% of the results and several laboratories were using this technique for the determina- tion of all of the elements except lead. A survey of pre-concentration and separation methods is not complete, as unfortunately over one third of the participants did not send the relevant information. Direct analysis of the sample without any pre-concentration or separation was applied in approximately 10% of all determinations.Pre-concentration by evaporation was used in an equal number of instances. Ion exchange, precipitation and extraction, in that order, were the most popular separation methods. Some comments seem to be necessary here. The firm position of neutron-activation analysis is worth emphasising. Details are given in an IAEA report.l* Detection and Rejection of Outlying Results In every set of observations there may be one or more outliers, i.e., observations thatJuly, 197s I N THE DETERMINATION OF TRACE ELEMENTS I N WATER 737 deviate markedly from the population of other observations. In trace-element analysis, because of the very low concentrations being determined, such gross errors may be caused by uncontrolled losses or contamination of the sample from the environment.A spread of results reaching up to two orders of magnitude was found at times in this and in previous IAEA intercom par is on^.^^^^^ When attempting, for instance, to check how well the results of a collaborative study approximate to the true value, the outlying results should be rejected before calculating statistical parameters that characterise the given population, as otherwise they can lead to false conclusions. Various approaches and criteria for the detection of outliers have been discussed in the literat~re.~~l*9~4 Four of them were selected for use in this study and are summarised in Table 11. Test Dixon’s Gr u bbs’ * Coefficient of skewness Coefficient of kurtosis TABLE I1 STATISTICAL CRITERIA USED FOR REJECTION OF OUTLYING RESULTS Statistic calculated Decision X _ - - 14 QB 4 3 0 Reject x1 or Xn if Y,, >critical value if smallest value *3 - XI yz2 r= - - X* - Xn-2 Xn-2 - x1 is suspect xu - x3 is suspect if largest value -- Reject x1 (or x,) if 3 <critical value ss n c ( X i - 533 2 ( X I - 3 4 i = l 1 2 - b, = n x - I- References 15, 17, 20 18, 19 Reject the result 16, 17 furthest from the mean if dL/b, >critical value Reiect the result 16, 17 iurthest from the mean if b2 >critical value * Only the formula for the instance when the smallest value ( x l ) is suspect is explicitly shown in the table.In the preliminary stage of this investigation it was found that the concurrent use of several criteria (at the same significance level, a) gave better results than the use of any single on@ This improvement arises because of the fact that when several outliers are present in a set of results, the detection of some of them may be masked by the presence of other spurious results.The degree of this masking effect depends on the distribution of results within the given set and the particular criterion used. As it was desired to eliminate distinct outliers but at the same time not to reject too easily, the procedure adopted here involved the concurrent use of all four criteria listed in Table I1 and calculating values at the conservative and moderate significance levels of 0.01 and 0.05, respectively.TABLE I11 41 SUMMARY OF RESULTS ON WATER SAMPLE W-3 Element determined/vg 1-I Cobalt Chromium 12.1 12.0 30 27 98 84 ----*---- ----- Cadmium 4.0 27 87 24 (23) 81 (78) 1.9-17.0 1.9-9.3 (1.9-7.4) 11 (15) 33 [30] 48 7 .______-____L____C__- Copper Iron Mercury 12.0 32.0 1 .o 33 31 21 106 99 63 Parameter Number of Laboratory averages , .reported results{ Individual determinations. . Number of Laboratory averages accepted results { Individual determinations ' Total range of laboratory averages . . . . Range of accepted laboratory averages* . . True value, 2 . . .. .. . . Arsenic 16.0 14 52 14 (14) 52 (52) 1.98-30.0 1.98-30.0 (1.98-30.0) 0 (0) 36 [29] 36 Barium 32.0 16 50 16 (16) 50 (50) 5.0-104.6 5 .O-104.6 (5.0-1 04.6) 0 (0) 38 [38] 50 29 (29) 27 (27) 95 (95) 84 (84) 4.0-54.3 2.97-21.9 30 (29) 27 (25) i S (16) 102 (97) 91 (84) 50 (50) 2.9-230.0 1.4-150.0 0.13-218.8 4.0-23.0 2.97-21.9 (4.0-23.0) (2.97-2 1.9) 3 (3) 0 (0) 33 [27] 26 [26] 43 48 2.9-21 .O 1.4-90.0 0.13-3.0 (6.86-21.0) (1.4-61.3) (0.13-3.9 9 (12) 13 (19) 24 (24) 30 [30] 35 [35] 33 [29] 42 48 52 Percentage of outlying laboratories* Percentage of laboratories with the R vaiue ' .>It Percentage of laboratories'ior which R dould ' ' not be calculated . . . . . . . . Over-all mean of accepted laboratory averages, x* , . . . . . . . . . . . 15.6 (15.6) 6.76 (6.76) 1.81 (1.81) 43.3 (43.3) 11.6 (11.6) -0.4 (-0.4) -2.6 (-2.5) 10.1 <1*<21.0 (11.7 <p<19.5) 47.6 (47.6) 27.5 (27.5) 57.8 (57.8) 14.5 (14.5) 6.88 (6.88) 4.63 (4.43) 11.9 (11.9) 11.0 (11.0) 1.63 (1.32) 4.16 (4.16) 4.07 (4.97) 35.2 (29.8) 36.0 (35.0) 45.2 (45.2) 0.333 (0.276) 0.772 (0.772) 0.956 10.956) 12.5 (12.8) 37.6 (33.6) 1.62 (1.62) 3.47 (3.01) 19.36 (13.46) 0.643 (0.643) 27.8 (23.4) 51.6 (40.1) 39.7 (39.7) Standard deviation 1 % : .. Standard error* f :iz$ "/b ; ; ; ; Deviation of the rabsolute,z - 2 . . over-all mean from4 relative, % 0.633 (0.558) 3.726 (2.691) 0.16i (0.i61) 5.1 (4.3) 9.9 (8.0) 9.9 (9.9) 7.2 (6.2) ' 6.5 (6.5) ' 8.7 (8.7) ' 15.6 (15.6) 48.8 (48.8) 0.63 (0.43) 15.8 (10.7) - 0.2 (- 0.2) - 1.0 (- 1.0) -1.7 (-1.7) -8.3 (-8.3) 5.55 (1.6) 0.62 (0.62) 0.5 (0.8) 4.2 (6.7) 62.3 (62.3) 17.3 (5.0) . . the true value v, for a probability level of 0.99$ [loo (z - z ) / ~ Confidence limits for the mean of the population, . . . . 27.3 < ~ < 6 7 . 9 (32.9 <y<62.2) 3.7 < ~ < 5 .6 (3.9 <v < 5.0) 9.8 <EL <14.1 (10.3 <g < 13.5) 8.4 <v <13.7 (9.1 < p < 13.0) Element determined/vg 1-l 10.8 < ~ < 1 4 . 3 27.2 < <47.9 1.1 <v < 2.1 (11.7 <p<14.0) (28.0<~<39.1) (1.3 < ~ < 2 . 0 ) Parameter True value, x^ . . . . . . . . Number of Laboratory averages. . . . reported results Individual determinations . . Number of Laboratory averages . . accepted results Individual determinations Total range of laboratory averages . . . . Range of accepted laboratory averages* . . r hlanganese Molybdenum Nickel Lead Selenium 12.0 8.0 12.0 16.0 8.0 30 13 21 25 10 105 43 76 86 24 30 (30) 12 (12) 19 (19) 24 (24) 10 (10) 105 (105) 40 (40) 70 (70) 82 (82) 24 (24) 2.8-18.0 1.0-31.7 5.8-261.5 12.7-80.0 2.0-10.3 2.8-18.0 1.0-10.4 5.8-16.9 12.7-49.7 2.0-10.3 (2.8-18.0) (1.0-10.4) (5.8-16.9) (1 2.7-49.7) (2.0-10.3) 0 (0) 8 (8) 10 (10) 4 (4) 0 (0) Uranium Vanadium 3.2 8.0 17 13 51 45 48 16 (16) 48) 44 (44) 2.5-50.0 0.57-1 8.7 0.57-7.8 2.5-16.7 (0.57-7.8) (2.5-16.7) 12 (12) 6 (6) 8 (8) - 7 Zinc 12.0 29 94 26 (26) 86 (86) 2.4-90.3 2.4-30.0 (2.P30.0) 10 (10) Percentage of outlying laboratories* Percentage of laboratories with the >It .... .. Percentage of laboratories for which not be calculated Over-all mean of accepted' laboratory R value .. .. 43 38 38 44 50 41 38 41 averages, 0 33 [30] 23 [23] 29 [19] 40 [40] 0 P I 18 (181 31 [31] 24 1211 R could 2 z . . .. 10.8 (10.8) 7.23 (7.23) 10.8 (10.8) 29.5 (29.5) 7.22 (7.22) 3.51 (3.54) 8.42 (8.42) 13.6 (13.6) 2 .., . 5.4 (5.4) 13.7 (13.7) 5.8 (5.8) 6.1 (6.1) 14.3 (14.3) 13.3 (13.3) 14.5 (14.5) 9.7 (9.7) .. . . 3.21 (3.21) 3.45 (3.45) 2.70 (2.70) 8.87 (8.87) 3.26 (3.26) 1.87 (1.87) 4.21 (4.21) 6.79 (6.79) .. . , 0.586 (0.586) 0.995 (0.995) 0.620 (0.620) 1.81 (1.81) 1.03 (1.03) 0.468 (0.468) 1.22 (1.22) 1.32 (1.32) . . 29.7 (29.7) 47.6 (47.6) 25.1 (25.1) 30.0 (30.0) 45.1 (45.1) 53.1 (53.1) 50.0 (50.0) 49.8 (49.8) x* .. .. 1.6 (1.6) 13.3 (13.3) 0.42 (0.42) . . -1.2 (-1.2) -0.77 (-0.77) -1.24 (-1.24) 13.5 (13.5) -0.78 (-0.78) 0.32 (0.32) . . . - 10.0 (-10.0) - 9.7 (- 9.7) - 10.3 ( - 10.3) 84.6 (84.6) - 9.7 (- 9.7) 10.0 (10.0) 5.2 (5.2) Confidence limits for the mean of the population, 9.1<y<12.4) 4.l<y<10.3 9.0<1~.<12.0 24.5<~<34.6 3.9<y<10.6 2 .2 < ~ < 4 . 9 4 . 6 < ~ < 1 2 . 2 9 . 9 < ~ < 1 7 . 2 v, for aprobability lcvel of 0.99$ .. . . 9.6<y<11.3) (5.0<~<9.4) (9.5<1~<12.1) (%.8<y<33.3) (4.9<@<9.6) ( 2 . 5 < ~ < 4 . 5 ) (6.7<[~<11.1) (10.9<p<16.4) 3 ? k, 0 k * These results were calculated using significance levels of 0.01 and the values in parentheses using levels of 0.06 for the rejection of outlying results. t The values in brackets were calculated after making allowance for the maximum estimated uncertainty in the true value. $ The values in parentheses were calculated for a probability level of 0.95.July, 1978 IN THE DETERMINATION OF TRACE ELEMENTS IN WATER 739 If a result was declared to be an outlier by any criterion it was rejected and the whole The outlying results (e.g., those marked by 3 in Table I) were not taken into account in procedure repeated until no more outliers could be identified.computing the over-all mean and other parameters. Accuracy and Precision The values of over-all mean, range of accepted laboratory averages, etc., refer to those obtained when using significance levels of 0.01 and those in parentheses when using levels of 0.05 for the rejection of outlying results as described above. When calculating the percentage of laboratories with R >1, the values in brackets were obtained after making allowance for the maximum estimated uncertainty in the true value. Although the spread of results supplied by particular laboratories is high (in some instances exceeding two orders of magnitude), the over-all mean calculated after rejection of outliers usually approximates closely to the true value (i.e., the amount added to the solution).A wide spread of results in trace-element analysis seems to be typical at present. In some other intercomparison tests in which all participating laboratories had a well established reputation in trace-metal analysis, values differing by a factor larger than 10 and sometimes larger than 100 were obtained for some elements.21v22 For most of the elements determined in this study (arsenic, cadmium, cobalt, chromium, copper, iron, manganese, molybdenum, nickel, selenium, uranium, vanadium and zinc) the true value is well within the confidence limits of the mean of the population of all laboratory means, defined as The results of this intercomparison are summarised in Table 111.where x SE to.ol(o.05) For these elements the percentage deviation of the over-all mean from the true value, 1OO(x - z)/x^ (where 2 is the true value), is relatively small (2-17% or 2-ll% for the two L- over-all mean (non-weighted meam of all accepted laboratory averages) ; = standard error (standard deviation of the mean); and = Student's factor for a significance level of 0.01 or 0.05. 54.3 X T X 1 I i + ! I * 4.8 4.03 2i" i i x X X * 1 . 4.9 '' ; ' ' 8 ' 1'1 ' 1; 1; ' 1 6 18 ' 2'0 ' i1 ' i 3 ' 2'5 ' 2'8 ' 3'0 ' i 3 ' 2 G 0 13 14 15 17 19 20 22 24 26 29 32 35 Laboratory code number Fig. 1. Results for cobalt content of water sample W-3. Individual laboratory means are shown as crosses or short horizontal lines.Within-laboratory standard deviations and standard errors are shown by thin and thick vertical lines ,respectively.740 DYBCZYfiSKI et d: PROBLEMS OF ACCURACY AND PRECISION Analyst, Vol. 103 sets of results shown in Table 111). Both positive and negative deviations from the true value were observed, as shown in Figs. 1 and 2, in which laboratory means are shown as crosses or short horizontal lines accompanied by the within-laboratory standard deviation and standard error (thin and thick vertical lines, respectively). Arrows indicate results that lie outside the scale of the diagram. 64.4 90.3 48.3 - T X I X 1 3 6 10 13 16 18 20 22 24 26 29 31 33 34 2 4 9 11 14 17 19 21 23 25 28 30 32 33 0 Laboratory code number Fig.2. Results for zinc content of water sample W-3. Individual laboratory means are shown as crosses or short horizontal lines. Within-laboratory standard deviations and standard errors are shown by thin and thick vertical lines, respec- tively. For barium, mercury and lead the positive deviations from the true value were much larger (Table 111) and for the last two elements the true values lie distinctly below the confidence limits for the mean of the population. Either impurities in the reagents added as macro-constituents of the water sample or environmental contamination in most of the laboratories participating in this exercise might account for the high results for these elements. The latter explanation would agree with the latest report on the wide occurrence of gross positive errors in the analysis of water and other matrices for lead.23 Inspection of Figs.1 and 2 reveals that one can find hardly any correlation between the precision of the results produced by a given laboratory (for example as expressed in terms of laboratory standard deviation) and their accuracy, i.e., the deviation of the laboratory mean from the true value. It follows from Table I11 that for most of the elements over 30% of the laboratories showed an R value [as defined by equation (l)] greater than unity. Allowance for the uncertainty in the true value does not change these figures significantly. The R value could not be calculated in more than 40% of the sets of results as only one or two results were available; it can therefore be inferred that approximately 50% of all results are subject to systematic errors.Table IV uses examples of a few elements to show how the statistical parameters of the population of laboratory averages (over-all mean and standard deviation) can change with the progression of rejecting the outliers. The complementary action of the four criteria in certain instances can be seen distinctly. For instance, if Dixon’s criterion alone were used for treatment of mercury results the over-all mean would be 3.33 pg 1-1 and the accuracy error would amount to 233%. In this instance the result of 15.0 pg ml-l has not been detected as an outlier because of masking action, but once it has been removed by other criteria, the three following highest results could be rejected on the basis of Dixon’s test.Similarly, the Grubbs’ test rejects the first and third highest results for cadmium (at a = O.Ol), but not the second one and does not reject the highest result for iron at bothJdy, 197’8 741 significance levels. The coefficient of skewness and especially the coefficient of kurtosis are more effective as single criteria. However, as follows from this and other, still unpublished IAEA intercomparison tests, they also fail sometimes (even at 01 = 0.05) to detect some outliers that may be identified by other tests. On the other hand, it was found that the application of criteria that are too severe for the rejection of outliers (e.g., the at higher significance levels, but not exceeding a = 0.10) can result sometimes in significant deviations of the over-all mean from the true value, rejection of most results and unrealistic estimates of the standard deviation.IN THE DETERMINATION OF TRACE ELEMENTS IN WATER TABLE IV USE OF VARIOUS CRITERIA FOR THE REJECTION OF OUTLIERS Number of Over-all Standard laboratory mean/ deviation1 Element averages vg 1-1 wg 1-l Cadmium 27 5.65 3.41 26 5.22 2.60 2B 4.91 2.10 24 4.63t 1.63t 23 4.43: 1.32: Deviation of a = 0.01 a = 0.06 the over-all 7 7- A \ mean from Detected by* Detected by* the true Outliers/ value, % pg 1-1 Outliers’ pgl-1 -- D G S K 41‘3 17.00 - + + + 17.00 + + + + i::: 13.00 - - + + 13.00 - + + + + + + 11.43 + + + + 9.33 - + + + 15.8t - 10.82 39.34 7**8 230.9 0 f + + 230.9 0 + + 4- 11.75 75.0 0 + + + 75.0 0 + + 4- 4.2t 27.0 0 + + + 27.0 0 + + + + Copper 33 21.45 32 14.91 31 12.97 4.29 30 12.50f 3.47i 29 12.84: 3.01: 7.0: 2.86 - - - Iron Mercury 31 30 29 28 ~.27 26 28 21 20 19 18 17 16 49.64 46.30 42.89 40.21 35.54 33.56: 13.59 3.33 37.55t ~.~~ 2.72 2.16 1.76 1.62: 37.16 32.70 27.34 23.64 19.36f 16.60 13.463 ;;:; 150.0 0 - + - 150.0 0 - + + 34.0 145.0 - - + - 145.0 - + + + 118.0 - - f - 118.0 - + + + + + 112.0 - + + + 90.0 - + f + 85.0 + + - + 11.1 4.9: 47.18 1259 4.03 233 3.03 172 1.87 116 0.848 76 0.643t 62: 218.8 + + + + 218.8 + + + + 15.0 - + + + 16.0 - + + + 12.75 + + + + 12.75 + + + + 8.88 + + + + 8.88 + + + + 4.00 + - - - 4.0 + + - + 18.82 !::; 90.33 + + + + 90.33 + + + + 64.4 + + + + 64.4 + + + + f!t:z2 48.25 - + + + 48.25 + + + + Zinc 29 19.21 28 16.67 13.16 27 14.90 9.43 26 13.622 6.80: * D = Dixon’s test; G = Grubbs’ test; S = coefficient of skewness test; and K = coefficient of kurtosis test.t Fjnal values for significance level of 0.01. 1 Final values for significance level of 0.05. Tables of critical 0 appears in the table when Dixon’s criterion could not be used. values for Dixon’s criterion are available only up to N = 30. - Signifies that the outlier was not rejected by this test and + that it was rejected. Comparison of Analytical Methods An attempt was made to assess the relative accuracy provided by various analytical methods with respect to particular elements. The results, reported as mean & standard error followed by number of accepted laboratory means in parentheses, are shown in Table V. Only those elements were considered for which at least two methods, each with not less than three accepted laboratory means, were available.As can be seen, there is generally fairly good agreement between the various methods, especially neutron-activation analysis and atomic-absorption spectrophotometry. It appears that X-ray fluorescence tends to produce results that are too low for several elements. Similarly, for iron and zinc, neutron-activation analysis seems to yield results that are too low and atomic-absorption spectrophotometry results that are too high. These conclusions, however, should be considered with caution and treated rather as an indication of a possible trend because they are based on an insufficient number of results. Calculated confidence limits for a significance level of 01 = 0.05 and such a small number of degrees of freedom usually encompass the true value.742 DYBCZYNSKI et d.: PROBLEMS OF ACCURACY AND PRECISION Analyst, V d 103 The effect of a pre-concentration step on the accuracy and precision of results could not be established, mainly because many laboratories did not send the relevant information.TABLE V COMPARISON OF ACCURACY OF INDIVIDUAL METHODS FOR THE DETEHMINATION OF VARIOUS ELEMENTS I N WATER SAMPLE W-3 Results are reported as mean & standard error followed by the number oi accepted laboratory means in parentheses. Concentration of element/pg 1-1 - A 7 Element Cadmium .. Cobalt . . Chromium Copper .. Iron .. Manganese Molybdenum Nickel . . Uranium . . Zinc .. .. .. .. .. .. .. .. .. .. .. ~ Mean of accepted .. results by named Method* method NAA AAS NAA AAS XRF NAA AAS AAS XRF NAA AAS NAA AAS XRF NAA AAS AAS XRF NAA Fluorimetry NAA AAS 5.6 f 0.9 ( 3) 4.4 f 0.4 (18) 13.5 f 1.1 ( 7) 12.0 9 1.0 (17) 8.8 f 2.5 ( 3) 11.4 f 2.3 ( 7) 11.0 f 1.3 (15) 12.2 f 0.7 (23) 13.0 f 4.2 ( 3) 28.2 f 13.6 ( 3) 37.8 f 4.1 (20) 11.2 f 2.7 ( 6) 11.1 f 0.5 (20) 6.5 & 2.8 ( 3) 7.0 f 1.4 ( 6) 11.2 f 0.8 (14) 2.7 f 0.6 ( 6) 4.1 f 0.6 ( 7) 7.0 & 2.3 ( 6) 15.2 f 1.5 (18) 8.1 f 1.0 ( 3) 8.9 f 1.6 ( 3) Mean of accepted results by all methods 4.6 & 0.3 (24) 11.9 f 0.8 (29) 11.1 & 1.0 (27) 12.5 -f 0.6 (30) 37.6 f 3.7 (27) 10.8 & 3.2 (30) 7.2 & 1.0 (12) 10.8 f 0.6 (19) 3.5 & 0.5 ( l G ) 13.6 & 1.3 (26) True value 4.0 12.1 12.0 12.0 32.0 12.0 8.0 12.0 3.2 12.0 * NAA = neutron-activation analysis, AAS = atomic-absorption spectrophotometry and XRF = X-ray fluorescence spectrometry. Conclusions The results of this intercomparison exercise clearly show the need for frequently checking the accuracy of analytical procedures employed in laboratories dealing with the determina- tion of trace elements in water. It seems that many analysts concentrate their efforts mainly on improving the precision (repeatability) of their results, whereas it is their accuracy that requires more attention.On the other hand, despite the considerable spread of results obtained by various labora- tories using different techniques, the over-all mean in general approximates closely to the true value, provided that gross errors are eliminated by the use of appropriately selective procedures for the rejection of outliers.This work shows a good mutual compensation of negative and positive errors and suggests that systematic errors in the general population of results obtained by various techniques may be normally or a t least symmetrically distributed. Rejection of outlying results is essential for obtaining good accuracy of determination of a true value. Concurrent use of several criteria for the rejection of outliers instead of any single criterion, at small or moderate significance levels, proved to be a good method for that purpose. Participants in the Intercomparison Exercise Prof. J. Korkisch, Institute for Analytical Chemistry, University of Vienna, Vienna, Austria; Prof. R. Van Grieken and Prof. F. C. Adams, Department of Chemistry, UniversityJuly, 1978 IN THE DETERMINATION OF TRACE ELEMENTS I N WATER 743 of Antwerp, Wilrijk, Belgium; Messrs.C. Boelen and A. Bekaert, Institut d’Hygiene et d’Epidkmiologie, Brussels, Belgium ; Miss E. Peeters, Institut Royal des Sciences Naturelles de Belgique, Brussels, Belgium ; Dr. M. Mertens, IRScNB lab0 d’ocdanographie physique, Brussels, Belgium; Dr. E. P. Mignonsin, Laboratoire d’Application des RadioCldments, Universitd de Li&ge, Lihge, Belgium ; Mr. M. Alvarenga, Instituto de Pesquisas Radioativas, Bela Horizonte, Brazil; Mr. F. J. Krug, Centro de Energia Nuclear na Agricultura, Piracicaba, Brazil; Mr. A. Zdrojewski, Air Pollution Control Directorate, Ottawa, Canada; Mr. G. J. Day, Eldorado Nuclear Ltd., Port Hope, Ontario, Canada; Mr.N. H. Bastholm, Danish Atomic Energy Commission, Roskilde , Denmark; Dr. G. Baudin, Commissariat B 1’Energie Atomique, Fontenay-aux-Roses, France ; Mrs. M. Archimbaud, Laboratoire de Surveillance des Nuisances de 1’Homme et de son Environment, Pierrelatte, France; Mr. H. Bouzigues, CEA Centre de Pierrelatte, Pierrelatte, France; Gesellschaft zur Wiederaufarbeitung von Kernbrennstoffen mbH, Eggenstein-Leopoldshafen, Federal Republic of Germany; Mr. K. Coy, Bayrisches Landesamt fiir Umweltschutz, Munich, Federal Republic of Germany; Mr. U. Schleichert and co-workers, Bundesanstalt fur Gewasserkunde, Koblenz, Federal Republic of Germany; Dr. P. H. Koske and Dr. K. Weiler, GKSS, Institut fur Werkstoff- technologie und Chemie, Geesthacht-Tesperhude, Federal Republic of Germany; Mrs.U. Schaefer, Zentralinstitut fur Kernforschung, Dresden, German Democratic Republic ; Dr. F. Moldenhawer, Staatliches Amt fur Atomischerheit und Strahlenschutz der DDR, Berlin, German Democratic Republic; Mr. I. S. Bhat, Environmental Survey Laboratory, TAPS Colony of H.P. Division, Barc, Trombay, Bombay, India; Bhabha Atomic Research Centre, Trombay, Bombay, India; Mr. A. Barocas, Laboratorio Contaminazione Continentale, CNEN CSN Casaccia, Italy; Dr. T. Mamuro, Radiation Centre of Osaka Prefecture, Osaka, Japan; Dr. J. G. van Raaphorst, Reactor Centrum Nederland, The Netherlands; Dr. W. G. de Ruig, Government Dairy Station, Leiden, The Netherlands; Mr. B. Salbu, Nuclear Chemistry Division, University of Oslo, Oslo, Norway; Prof. Dr.2. Marczenko, Zaklad Chemii Analitycznej, PW, Warsaw, Poland; Mrs. B. Domaniecki, National Institute for Water Research, S.W.A. Regional Laboratory, Republic of South Africa; Mr. R. Smith, National Institute for Water Research, CSIR, Pretoria, Republic of South Africa; Dr. C. Capdevila, Junta de Energia Nuclear, Madrid, Spain; Mr. H. G. Bart, Anorganische Chemische Institut/Radiochemie, Universitat Bern, Bern, Switzerland ; Mr. A. L. Wilson, Water Research Centre, Medmenham Laboratory, Medmenham, United Kingdom; Dr. G. A. Welford, U.S. Energy Research and Development Administration, Health and Safety Laboratory, New York, N.Y., USA; and Dr. J. Sedlet, Argonne National Laboratory, Argonne, Ill., USA. The authors are indebted to Dr. G. B. Cook for reading the manuscript and making valuable comments.Thanks are also due to Mrs. K. Goettler for writing the computer program, to the staff of the IAEA Computer Section for performing the calculations and to Mr. P. Siwy €or discussions on the statistics. 1. 2. 3. 4. 6. 0. 7. 8. 9. 10. 11. 12. 13. References Minczewski, J., Chwastowska, J., and Dybczynski, R., “Analiza Sladowa,” WNT, Warsaw, 1973. Rook, H. L., and Moody, J., in Vogt, J. R., and Meyer, W., Editors, “Proceedings of the Second International Conference on Nuclear Methods in Environmental Research,” University of Missouri, Columbia, Mo., 1974, p. 44. Shenderikar, A. D., Dharmarajan, V., Walker-Merrick, H., and West, P. W., Analytica Chim. Acta, 1976, 84, 409. Eicholz, G. G., Nagel, A. E., and Hughes, R. B., Analyt. Chsm., 1965, 37, 863.Robertson, D. E., Analytica Chim. Acta, 1968, 42, 633. Crouthamel, E. R., Heinrich, R., and Satrousis, C., Talanla, 1958, 1, 396. Struempler, A. V., Analyt. Chem., 1973, 45, 2251. Baudin, G., “Proceedings of the 1976 International Conference on Modern Trends in Activation Toribara, T. Y., Shields, C. P., and Koval, L., Talanta, 1970, 17, 1025. Baier, R. W., Wojnowich, L., and Petric, L., Analyt. Chem., 1975, 47, 2464. Litman, R., Finston, H. L., and Williams, E. T., Analyt. Chem., 1975, 47, 2364. Bothner, M. H., and Robertson, D. E., Analyt. Chem., 1975, 47, 592. Gbrski, L., Heinonen, J., and Suschny, O., in La Fleur, P. D.. Editor, “Accuracy in Trace Analysis Proceedings of the 7th IMR Symposium,” NBS Special Publication No. 422, Volume 1, National Bureau of Standards, Washington, D.C., 1976, p. 189. Analysis,” Volume 1, Munich, 1976, p. 94.744 14. 16. 16. 17. 18. 19. 20. 21. 22. 23. 24. DYBCZY~~SKI, TUGSAVUL AND SUSCHNY Dybczyfiski, R., Tugsavul, A., and Suschny, O., “Report on the Intercomparison of Determina- tions of Selected Radionuclides and Trace Elements in Water [W-I, W-2 and W-3 (1975-76)],” IAEA/RL/40, International Atomic Energy Agency, Vienna, December 1976. Natrella, M. G., “Experimental Statistics,” NBS Handbook 91, National Bureau of Standards, Washington, D.C., 1963. Ferguson, T. S., in Neyman, J., Editor, “Fourth Berkeley Symposium on Mathematical Statistics and Probability,” University of California Press, Berkeley, Calif., 1961, p. 253. Grubbs, F. E., Technometrics, 1969, 11, 1. Grubbs, F. E., Ann. Math. Statist., 1950, 21, 27. “Model 10 Hewlett-Packard Calculator 9810 A, Stat. Pac.,” Hewlett-Packard Calculator Products Division, Loveland, Colo., p. 216. Johnson, N. L., and Leone, F. C., “Statistics and Experimental Design in Engineering and thc Physical Sciences,” Volume 1, Wilcy, New Yorlr, 1964. von Lehmden, D. J., Jungers, R. H., and Lee, R. E., Jr., Analyt. Chern.., 1974, 46, 239. Wagemann, R., in La Fleur, P. D., Editor, “Accuracy in Trace Analysis, Proceedings of a Symposium,” NBS Special Publication No. 422, Volume 1 , National Bureau of Standards, Washington, D.C., 1976, p. 247. Patterson, C. C., and Settle, D. M., i~ La Fleur, P. D., Editor, “Accuracy in Trace Analysis, Pro- ceedings of a Symposium,” NBS Special Publication No. 422, Volume 1, National Bureau of Standards, Washington, D.C., 1976, p. 321. Recherche des Valeurs Aberrantes, in “Methodes Statistiques en Chimie Analitique,” Volume 3, Part 4, CETAMA, Paris, 1966, pp. 7-9. Received September 26th, 1977 Accepted November 30th, 1977

ISSN:0003-2654    

DOI:10.1039/AN9780300734

出版商:RSC    

年代:1978

数据来源: RSC